Systemic corticosteroids for the treatment of COVID‐19: Equity‐related analyses and update on evidence

Abstract

Background

Systemic corticosteroids are used to treat people with COVID‐19 because they counter hyper‐inflammation. Existing evidence syntheses suggest a slight benefit on mortality. Nonetheless, size of effect, optimal therapy regimen, and selection of patients who are likely to benefit most are factors that remain to be evaluated.

Objectives

To assess whether and at which doses systemic corticosteroids are effective and safe in the treatment of people with COVID‐19, to explore equity‐related aspects in subgroup analyses, and to keep up to date with the evolving evidence base using a living systematic review approach.

Search methods

We searched the Cochrane COVID‐19 Study Register (which includes PubMed, Embase, CENTRAL, ClinicalTrials.gov, WHO ICTRP, and medRxiv), Web of Science (Science Citation Index, Emerging Citation Index), and the WHO COVID‐19 Global literature on coronavirus disease to identify completed and ongoing studies to 6 January 2022.

Selection criteria

We included randomised controlled trials (RCTs) that evaluated systemic corticosteroids for people with COVID‐19.

We included any type or dose of systemic corticosteroids and the following comparisons: systemic corticosteroids plus standard care versus standard care, different types, doses and timings (early versus late) of corticosteroids.

We excluded corticosteroids in combination with other active substances versus standard care, topical or inhaled corticosteroids, and corticosteroids for long‐COVID treatment.

Data collection and analysis

We followed standard Cochrane methodology. To assess the risk of bias in included studies, we used the Cochrane 'Risk of bias' 2 tool for RCTs. We rated the certainty of the evidence using the GRADE approach for the following outcomes: all‐cause mortality up to 30 and 120 days, discharged alive (clinical improvement), new need for invasive mechanical ventilation or death (clinical worsening), serious adverse events, adverse events, hospital‐acquired infections, and invasive fungal infections.

Main results

We included 16 RCTs in 9549 participants, of whom 8271 (87%) originated from high‐income countries. A total of 4532 participants were randomised to corticosteroid arms and the majority received dexamethasone (n = 3766). These studies included participants mostly older than 50 years and male. We also identified 42 ongoing and 23 completed studies lacking published results or relevant information on the study design.

Hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19

Systemic corticosteroids plus standard care versus standard care plus/minus placebo

We included 11 RCTs (8019 participants), one of which did not report any of our pre‐specified outcomes and thus our analyses included outcome data from 10 studies.

Systemic corticosteroids plus standard care compared to standard care probably reduce all‐cause mortality (up to 30 days) slightly (risk ratio (RR) 0.90, 95% confidence interval (CI) 0.84 to 0.97; 7898 participants; estimated absolute effect: 274 deaths per 1000 people not receiving systemic corticosteroids compared to 246 deaths per 1000 people receiving the intervention (95% CI 230 to 265 per 1000 people); moderate‐certainty evidence).

The evidence is very uncertain about the effect on all‐cause mortality (up to 120 days) (RR 0.74, 95% CI 0.23 to 2.34; 485 participants). The chance of clinical improvement (discharged alive at day 28) may slightly increase (RR 1.07, 95% CI 1.03 to 1.11; 6786 participants; low‐certainty evidence) while the risk of clinical worsening (new need for invasive mechanical ventilation or death) may slightly decrease (RR 0.92, 95% CI 0.84 to 1.01; 5586 participants; low‐certainty evidence).

For serious adverse events (two RCTs, 678 participants), adverse events (three RCTs, 447 participants), hospital‐acquired infections (four RCTs, 598 participants), and invasive fungal infections (one study, 64 participants), we did not perform any analyses beyond the presentation of descriptive statistics due to very low‐certainty evidence (high risk of bias, heterogeneous definitions, and underreporting).

Different types, dosages or timing of systemic corticosteroids

We identified one RCT (86 participants) comparing methylprednisolone to dexamethasone, thus the evidence is very uncertain about the effect of methylprednisolone on all‐cause mortality (up to 30 days) (RR 0.51, 95% CI 0.24 to 1.07; 86 participants). None of the other outcomes of interest were reported in this study.

We included four RCTs (1383 participants) comparing high‐dose dexamethasone (12 mg or higher) to low‐dose dexamethasone (6 mg to 8 mg).

High‐dose dexamethasone compared to low‐dose dexamethasone may reduce all‐cause mortality (up to 30 days) (RR 0.87, 95% CI 0.73 to 1.04; 1269 participants; low‐certainty evidence), but the evidence is very uncertain about the effect of high‐dose dexamethasone on all‐cause mortality (up to 120 days) (RR 0.93, 95% CI 0.79 to 1.08; 1383 participants) and it may have little or no impact on clinical improvement (discharged alive at 28 days) (RR 0.98, 95% CI 0.89 to 1.09; 200 participants; low‐certainty evidence). Studies did not report data on clinical worsening (new need for invasive mechanical ventilation or death).

For serious adverse events, adverse events, hospital‐acquired infections, and invasive fungal infections, we did not perform analyses beyond the presentation of descriptive statistics due to very low‐certainty evidence.

We could not identify studies for comparisons of different timing and systemic corticosteroids versus other active substances.

Equity‐related subgroup analyses

We conducted the following subgroup analyses to explore equity‐related factors: sex, age (< 70 years; ≥ 70 years), ethnicity (Black, Asian or other versus White versus unknown) and place of residence (high‐income versus low‐ and middle‐income countries). Except for age and ethnicity, no evidence for differences could be identified. For all‐cause mortality up to 30 days, participants younger than 70 years seemed to benefit from systemic corticosteroids in comparison to those aged 70 years and older. The few participants from a Black, Asian, or other minority ethnic group showed a larger estimated effect than the many White participants.

Outpatients with asymptomatic or mild disease

There are no studies published in populations with asymptomatic infection or mild disease.

Authors' conclusions

Systemic corticosteroids probably slightly reduce all‐cause mortality up to 30 days in people hospitalised because of symptomatic COVID‐19, while the evidence is very uncertain about the effect on all‐cause mortality up to 120 days. For younger people (under 70 years of age) there was a potential advantage, as well as for Black, Asian, or people of a minority ethnic group; further subgroup analyses showed no relevant effects. Evidence related to the most effective type, dose, or timing of systemic corticosteroids remains immature. Currently, there is no evidence on asymptomatic or mild disease (non‐hospitalised participants). Due to the low to very low certainty of the current evidence, we cannot assess safety adequately to rule out harmful effects of the treatment, therefore there is an urgent need for good‐quality safety data. Findings of equity‐related subgroup analyses should be interpreted with caution because of their explorative nature, low precision, and missing data.

We identified 42 ongoing and 23 completed studies lacking published results or relevant information on the study design, suggesting there may be possible changes of the effect estimates and certainty of the evidence in the future.

Plain language summary

Are corticosteroids (anti‐inflammatory medicines) given orally or by injection an effective treatment for people with COVID‐19?

Key messages

• Corticosteroids (anti‐inflammatory medicines) given orally or by injection (systemic) are evaluated for the treatment of coronavirus disease 2019 (COVID‐19).

• Corticosteroids are effective in reducing mortality slightly.

• We do not know whether a specific type or dose of corticosteroid is effective.

• There are no data for people who were not hospitalised.

• We found 42 ongoing and 23 completed studies lacking published results or relevant information on the study design, so our findings may change in the future.

What are corticosteroids?

Corticosteroids are anti‐inflammatory drugs that reduce redness and swelling that arises due to an insult (e.g. injury, irritation) to the body. They also reduce the activity of the immune system, which defends the body against disease and infection. Corticosteroids are used to treat a variety of conditions, such as asthma, eczema, joint strains, and rheumatoid arthritis. Systemic corticosteroids can be swallowed or taken as an injection to treat problems anywhere in the body. Short‐term intake of high doses can increase the risk of further infections (including fungal infections) as well as high blood sugar and blood pressure. Furthermore, it can cause swelling of the body and psychiatric side effects such as steroid psychosis and delirium.

Why are corticosteroids possible treatments for COVID‐19?

COVID‐19 affects the lungs and airways. As the immune system fights the virus, the lungs and airways become injured and inflamed, causing breathing difficulties, and hinder oxygen transport to other vital organs. Some patients’ immune systems overreact against the invading virus causing further inflammation and tissue damage in the whole body; corticosteroids may help to control this response.

We wanted to know:

• whether and in which doses systemic corticosteroids are an effective treatment for people with COVID‐19;

• whether they cause unwanted effects; and

• whether the benefits and harms differ with respect to equity‐related aspects (e.g. age, sex, ethnicity, income by country).

We were interested in:

• deaths from any cause up to 30 and 120 days after treatment start;

• whether people got better or worse after treatment;

• unwanted effects, for example infections caught in hospital.

What did we do?

We looked for studies that investigated systemic corticosteroids for people with COVID‐19. People could be of any age, sex, or ethnicity.

Studies could compare corticosteroids:

• plus usual care versus usual care with or without placebo (sham medicine);
• versus another type of corticosteroid;
• versus a different medicine;
• in different doses;
• given as early versus late treatment.

We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

We found 16 studies with 9549 people. About 4532 people received corticosteroids, mostly dexamethasone (3766 people). These studies included participants mostly older than 50 years, male, and from high‐income countries.

We also found 42 ongoing and 23 completed studies lacking published results or relevant information on the study design.

Main results

Eleven studies compared corticosteroids plus usual care versus usual care with or without a placebo. Only one study compared two types of corticosteroids. Four studies compared different dosing of a corticosteroid named dexamethasone. The studies included only hospitalised people with confirmed or suspected COVID‐19. No studies looked at non‐hospitalised people or different timing of treatment.

Systemic corticosteroids plus usual care compared to usual care with or without placebo:

• probably reduce the number of deaths from any cause slightly, up to 30 days after treatment;

• may slightly increase the chance of being discharged alive from hospital and may slightly decrease the risk of needing breathing support or dying;

• we don't know if corticosteroids increase or decrease the number of deaths from any cause up to 120 days after treatment, any unwanted effects, or infections caught in the hospital.

Methylprednisolone versus dexamethasone:

The evidence for the number of deaths up to 30 days is very uncertain (one small study only).

High‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8 mg)

High‐dose dexamethasone:

• may reduce the number of deaths from any cause up to 30 days after treatment;
• may make little to no difference to the chance of being discharged alive from hospital;
• we don't know if high‐dose dexamethasone increases or decreases the number of deaths from any cause up to 120 days after treatment, any unwanted effects, or infections caught in the hospital.

Equity‐related subgroup analyses

We examined the following equity‐related aspects: ethnicity (Black, Asian or other versus White versus unknown) and place of residence (high‐income versus low‐ and middle‐income countries). For most of the subgroups, except for age and ethnicity, no evidence for differences could be identified. For death from any cause up to 30 days, participants younger than 70 years seem to benefit from corticosteroids in contrast to participants who were aged 70 years and older. Furthermore, the few participants from a Black, Asian, or minority ethnic group had a larger estimated effect than the many White participants, but these subgroup results need cautious interpretation.

What are the limitations of the evidence?

We are moderately confident in the evidence about the effect of corticosteroids on deaths from any cause within 30 days after treatment. However, our confidence in the other evidence is low to very low, because studies did not use the most robust methods, and the way results were recorded and reported differed across studies.

How up‐to‐date is this evidence?

This review updates our previous review. The evidence is up‐to‐date to 6 January 2022.

Summary of findings

Summary of findings 1. Summary of findings table ‐ Corticosteroids plus standard care compared to standard care (plus/minus placebo) for hospitalised and unvaccinated individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Corticosteroids plus standard care compared to standard care (plus/minus placebo) for hospitalised and unvaccinated individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19
Patient or population: hospitalised and unvaccinated individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19 Setting: inpatient, ICU Intervention: corticosteroids plus standard care Comparison: standard care (plus/minus placebo)
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with standard care (plus/minus placebo)	Risk with corticosteroids plus standard care
All‐cause mortality up to 30 days	274 per 1000	246 per 1000 (230 to 265)	RR 0.90 (0.84 to 0.97)	7898 (9 RCTs)	⊕⊕⊕⊝ Moderatea	Systemic corticosteroids probably reduce all‐cause mortality up to 30 days slightly.
All‐cause mortality up to 120 days	402 per 1000	298 per 1000 (93 to 942)	RR 0.74 (0.23 to 2.34)	485 (3 RCTs)	⊕⊝⊝⊝ Very lowb	The evidence is very uncertain about the effect of systemic corticosteroids on all‐cause mortality up to 120 days.
Clinical improvement: discharged alive (follow‐up: 28 days)	620 per 1000	664 per 1000 (639 to 688)	RR 1.07 (1.03 to 1.11)	6786 (3 RCTs)	⊕⊕⊝⊝ Lowc	Systemic corticosteroids may slightly increase the chance of clinical improvement: discharged alive.
Clinical worsening: new need for invasive mechanical ventilation or death	282 per 1000	260 per 1000 (237 to 285)	RR 0.92 (0.84 to 1.01)	5586 (2 RCTs)	⊕⊕⊝⊝ Lowd	Systemic corticosteroids may slightly decrease the risk of clinical deterioration: new need for invasive mechanical ventilation or death.
Serious adverse events (follow‐up: during treatment)	We did not perform meta‐analyses because of high risk of bias, heterogeneous definitions, and underreporting. Therefore, we only present descriptive statistics: Angus 2020 shock‐dependent hydrocortisone: RR 4.11 (95% CI 0.23 to 72.98); Angus 2020 fixed‐dose hydrocortisone: RR 1.43 (95% CI 0.16 to 12.49); Tomazini 2020: RR 0.54 (95% CI 0.19 to 1.59).			678 (2 RCTs)	⊕⊝⊝⊝ Very lowe	The evidence is very uncertain about the effect of systemic corticosteroids on serious adverse events.
Adverse events (any grade) (follow‐up: during treatment)	We did not perform meta‐analyses because of high risk of bias, heterogeneous definitions, and underreporting. We only present descriptive statistics: Edalatifard 2020: RR 0.82 (95% CI 0.12 to 5.48); Tang 2021: RR 0.63 (95% CI 0.22 to 1.76); Tomazini 2020: RR 0.99 (95% CI 0.89 to 1.10).			447 (3 RCTs)	⊕⊝⊝⊝ Very lowf	The evidence is very uncertain about the effect of systemic corticosteroids on adverse events.
Hospital‐acquired infections (follow up: during treatment)	We did not perform meta‐analyses because of high risk of bias, heterogeneous definitions, and underreporting. We present descriptive statistics only: Corral‐Gudino 2021: RR 4.14 (95% CI 0.51 to 33.49); Dequin 2020: RR 0.90 (95% CI 0.60 to 1.34); Tang 2021: RR 2.00 (95% CI 0.19 to 21.24); Tomazini 2020: RR 0.75 (95% CI 0.50 to 1.15).			598 (4 RCTs)	⊕⊝⊝⊝ Very lowf	The evidence is very uncertain about the effect of systemic corticosteroids on hospital‐acquired infections.
Invasive fungal infections (follow‐up: during treatment)	We present descriptive statistics only because of high risk of bias: Corral‐Gudino 2021: RR 2.50 (95% CI 0.11 to 59.15).			64 (1 RCT)	⊕⊝⊝⊝ Very lowf	The evidence is very uncertain about the effect of systemic corticosteroids on invasive fungal infections.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_424096272361428897.

^a We downgraded one level for serious risk of bias (partly deviations from the intended intervention, selection of the reported results, missing information about the allocation concealment, baseline differences)
^b We downgraded one level for serious risk of bias (partly deviations from the intended intervention, selection of the reported results, missing information about the allocation concealment), one level for serious inconsistency and one level for serious imprecision (wide confidence interval, low number of participants)
^c We downgraded one level for serious risk of bias (partly deviations from the intended intervention, selection of the reported results, missing information about the allocation concealment) and one level for serious inconsistency.
^d We downgraded one level for serious risk of bias (deviations from the intended intervention, measurement of the outcome) and one level for serious inconsistency.
^e We downgraded two levels for very serious risk of bias (deviations from the intended intervention, missing adjustment for competing risk of death, reporting bias (the safety‐relevant outcome was not reported)) and one level for serious imprecision (fewer than 500 events).
^f We downgraded two levels for very serious risk of bias (deviations from the intended intervention, missing adjustment for competing risk of death, missing information about the allocation concealment, reporting bias (the safety‐relevant outcome was not reported) and one level for serious imprecision (fewer than 500 events).

Summary of findings 2. Summary of findings table ‐ Methylprednisolone compared to dexamethasone for hospitalised and unvaccinated individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Methylprednisolone compared to dexamethasone for hospitalised and unvaccinated individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19
Patient or population: hospitalised and unvaccinated individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19 Setting: inpatient, ICU Intervention: methylprednisolone Comparison: dexamethasone
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with dexamethasone	Risk with methylprednisolone
All‐cause mortality up to 30 days	357 per 1000	182 per 1000 (86 to 382)	RR 0.51 (0.24 to 1.07)	86 (1 RCT)	⊕⊝⊝⊝ Very lowa	The evidence is very uncertain about the effect of methylprednisolone on all‐cause mortality up to 30 days.
All‐cause mortality up to 120 days ‐ not reported	‐	‐	‐	‐	‐	No study reported this outcome.
Clinical improvement: discharged alive ‐ not reported	‐	‐	‐	‐	‐	No study reported this outcome.
Clinical worsening: new need for invasive mechanical ventilation or death ‐ not reported	‐	‐	‐	‐	‐	No study reported this outcome.
Serious adverse events ‐ not reported	‐	‐	‐	‐	‐	No study reported this outcome.
Adverse events ‐ not reported	‐	‐	‐	‐	‐	No study reported this outcome.
Hospital‐acquired infections ‐ not reported	‐	‐	‐	‐	‐	No study reported this outcome.
Invasive fungal infections ‐ not reported	‐	‐	‐	‐	‐	No study reported this outcome.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_424391243004646489.

^a We downgraded one level for serious risk of bias for missing pre‐specification/protocol/statistical analysis plan and two levels for very serious imprecision (fewer than 50 events).

Summary of findings 3. Summary of findings table ‐ High‐dose dexamethasone (12 mg or higher) compared to low‐dose dexamethasone (6 mg to 8 mg) for hospitalised individuals with unknown vaccination status and a confirmed diagnosis of symptomatic COVID‐19.

High‐dose dexamethasone (12 mg or higher) compared to low‐dose dexamethasone (6 mg to 8 mg) for hospitalised individuals with unknown vaccination status and a confirmed diagnosis of symptomatic COVID‐19
Patient or population: hospitalised individuals with unknown vaccination status and a confirmed diagnosis of symptomatic COVID‐19 Setting: inpatient Intervention: high‐dose dexamethasone (12 mg or higher) Comparison: low‐dose dexamethasone (6 mg to 8 mg)
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with low‐dose dexamethasone (6 mg to 8 mg)	Risk with high‐dose dexamethasone (12 mg or higher)
All‐cause mortality up to 30 days	285 per 1000	248 per 1000 (208 to 297)	RR 0.87 (0.73 to 1.04)	1269 (3 RCTs)	⊕⊕⊝⊝ Lowa	High‐dose dexamethasone may further reduce all‐cause mortality up to 30 days compared to low‐dose dexamethasone (6 mg to 8 mg).
All‐cause mortality up to 120 days	329 per 1000	306 per 1000 (260 to 355)	RR 0.93 (0.79 to 1.08)	1383 (4 RCTs)	⊕⊝⊝⊝ Very lowb	The evidence is very uncertain about the effect of high‐dose dexamethasone (12 mg or higher) on all‐cause mortality up to 120 days compared to low‐dose dexamethasone (6 mg to 8 mg).
Clinical improvement: discharged alive (follow‐up: 28 days)	882 per 1000	865 per 1000 (785 to 962)	RR 0.98 (0.89 to 1.09)	200 (1 RCT)	⊕⊕⊝⊝ Lowc	High‐dose dexamethasone (12 mg or higher) may have little or no impact on the chance of clinical improvement: discharged alive compared to low‐dose dexamethasone (6 mg to 8 mg).
Clinical worsening: new need for invasive mechanical ventilation or death ‐ not reported	‐	‐	‐	‐	‐	No study reported this outcome.
Serious adverse events (follow‐up: during treatment)	We did not perform meta‐analyses because of high risk of bias arising from the missing adjustment for competing risk of death. We present descriptive data only: Munch 2021b: RR 0.80 (95% CI 0.60 to 1.07); Maskin 2021: RR 1.05 (95% CI 0.88 to 1.25).			1080 (2 RCTs)	⊕⊝⊝⊝ Very lowd	The evidence is very uncertain about the effect of high‐dose dexamethasone (12 mg or higher) on serious adverse events compared to low‐dose dexamethasone (6 mg to 8 mg).
Adverse events (any grade) (follow‐up: during treatment)	We did not perform meta‐analyses because of high risk of bias arising from the missing adjustment for competing risk of death. We present descriptive data only: Maskin 2021: RR 1.02 (95% CI 0.96 to 1.08).			98 (1 RCT)	⊕⊝⊝⊝ Very lowe	The evidence is very uncertain about the effect of high‐dose dexamethasone (12 mg or higher) on adverse events compared to low‐dose dexamethasone (6 mg to 8 mg).
Hospital‐acquired infections (follow‐up: during treatment)	We did not perform meta‐analyses because of high risk of bias arising from the missing adjustment for competing risk of death. We present descriptive data only: Maskin 2021: RR 0.89 (95% CI 0.70 to 1.14); Munch 2021b: RR 0.80 (95% CI 0.56 to 1.14).			1080 (2 RCTs)	⊕⊝⊝⊝ Very lowf	The evidence is very uncertain about the effect of high‐dose dexamethasone (12 mg or higher) on hospital‐acquired infections compared to low‐dose dexamethasone (6 mg to 8 mg).
Invasive fungal infections (follow‐up: during treatment)	We did not perform meta‐analyses because of high risk of bias arising from the missing adjustment for competing risk of death. We present descriptive data only: Munch 2021b: RR 0.70 (95% CI 0.36 to 1.34); Maskin 2021: RR 1.00 (95% CI 0.21 to 4.71).			1080 (2 RCTs)	⊕⊝⊝⊝ Very lowf	The evidence is very uncertain about the effect of high‐dose dexamethasone (12 mg or higher) on invasive fungal infections compared to low‐dose dexamethasone (6 mg to 8 mg).
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_431226044092715996.

^a We downgraded one level for serious risk of bias (deviations from the intended intervention, no information about the allocation concealment) and one level for serious imprecision (wide confidence interval, low number of participants/events). 
^b We downgraded one level for serious risk of bias (no information about the allocation concealment; deviations from the protocol), one level for serious imprecision (wide confidence interval, low number of participants/events) and one level for serious inconsistency (Toroghi 2021 differs from the other studies). 
^c We downgraded one level for serious risk of bias (no information about the allocation concealment) and one level for serious imprecision (wide confidence interval, low number of participants/events). 
^d We downgraded one level for serious risk of bias (missing adjustment of competing risk of death) and two levels for very serious imprecision (very low number of events/participants).
^e We downgraded one level for serious risk of bias (missing adjustment for competing risk of death, deviations from the intended intervention), two levels for very serious imprecision (very low number of events/participants).
^f We downgraded one level for serious risk of bias (missing adjustment for competing risk of death, protocol deviations, measurement of the outcome, no information about the allocation concealment), two levels for very serious imprecision (very low number of events/participants).

Background

This work is part of a series of Cochrane Reviews investigating treatments and therapies for coronavirus disease 2019 (COVID‐19). Reviews in this series share information in the background section and methodology with the first published reviews about monoclonal antibodies (Kreuzberger 2021), and convalescent plasma (Piechotta 2021).

Description of the condition

COVID‐19 is a rapidly spreading infectious disease caused by severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) (WHO 2020a). On 11 March 2020, the World Health Organization (WHO) declared the current COVID‐19 outbreak a pandemic, which is unprecedented in comparison to previous coronavirus outbreaks (severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS), which each caused fewer than 1000 deaths; WHO 2007; WHO 2019). Despite intensive international efforts to contain its spread, SARS‐CoV‐2 has resulted in an ongoing increase of new weekly cases and deaths in several regions around the globe (WHO 2021a; WHO 2022a). In the meantime, the emergence of SARS‐CoV‐2 variants, with the potential for altered transmission or disease characteristics, or to impact the effectiveness of vaccines, therapeutics, diagnostics, or public health and social measures, challenge strategies to control disease spread (WHO 2022b).

Apart from age and co‐morbidities, vaccination status substantially influences the risk of a severe course of disease, hospitalisation, and mortality. In patients without effective immunisation (individuals who are unvaccinated, incompletely vaccinated, or individuals who fail to develop an immunological response despite being fully vaccinated) this risk is higher among those aged 65 years or older, smokers and those with certain underlying medical conditions such as cancer, chronic kidney disease, chronic obstructive pulmonary disease (COPD), heart conditions, immunocompromised state, obesity, sickle cell disease or type 2 diabetes mellitus (Huang 2020; Liang 2020; WHO 2020a; Williamson 2020). COVID‐19 case fatality ratios vary widely between countries and reporting periods, from 0.0% to more than 18% (Johns Hopkins University 2022). However, these numbers may be misleading as variability in reporting and testing certainly contributed to inaccuracy in case fatality ratios due to varying testing frequency, a lack of reporting dates, and variations in case definitions, especially at the beginning of the pandemic when the main focus was on severe cases (WHO 2020b).

The median incubation time and time to symptom onset depends on the virus variant and is estimated to be three days (range zero to eight days) in the case of the Omicron variant of concern, which is shorter compared with previous reports for the Delta variant and other previously circulating non‐Delta SARS‐CoV‐2 (five to six days) (Brandal 2021; Lauer 2020). Sore throat, cough, fever, headache, fatigue, and myalgia or arthralgia are the most commonly reported symptoms (Brandal 2021; Struyf 2020). Other symptoms include dyspnoea, chills, nausea or vomiting, diarrhoea, and nasal congestion (WHO 2020a). The reported frequency of asymptomatic infections varies greatly, depending on the time of the investigation, the cohort investigated, and the virus variant, and ranges between 6% and 96% (Buitrago‐Garcia 2020; Funk 2021; Lewnard 2022; Oran 2020; Wolter 2022).

A smaller proportion of people are affected by severe (approximately 11% to 20%) or critical (approximately 1% to 5%) disease with hospitalisation and intensive care unit (ICU) admission due to respiratory failure, septic shock, or multiple organ dysfunction syndrome (Ferguson 2021; Funk 2021; Lewnard 2022; Wolter 2022; Wu 2020). In one systematic review and meta‐analysis of international studies, the proportion of patients who died among those treated in the ICU was estimated to be 34% and for those who received invasive mechanical ventilation it was 83% (Potere 2020). However, the hospitalisation and ICU treatment rates seem to depend on the virus variant. Analyses from the United Kingdom show a significant reduction in the relative risk of hospitalisation for adult Omicron cases compared to Delta (Ferguson 2021). There may also have been a different threshold for admission to hospital or ICU during the course of the pandemic. Depending on the local pressure on ICU resources, some normal wards will have learned to provide continuous positive airway pressure (CPAP) therapy equivalent to ICU support in other healthcare systems. It is unclear whether triage criteria in some healthcare systems may have influenced admission to hospital or ICU (or both).

As the evidence on many of the substances that were investigated for the treatment of COVID‐19 increased over the course of the pandemic, national and international guidelines emerged to support daily clinical decisions (NICE 2021; NIH 2021; WHO 2021b). However, so far there are only a few substances with clearly proven benefits and clear recommendations as well as approval by national and international authorities for the treatment of COVID‐19 (EMA 2022; FDA 2022; WHO 2021b; WHO 2021c). In light of the extent of the COVID‐19 pandemic and the scarcity of effective treatments, there is still an urgent need for effective therapies to save lives and to reduce the high burden on healthcare systems (either with a high workload caused by COVID‐19 or staff shortages due to infected health care providers), especially in the face of evolving variants of the virus with the potential for increased transmissibility and the limited global availability of vaccines.

Description of the intervention

Corticosteroids are a group of stress hormones produced from the adrenal cortex. In addition to their stress‐mediated mechanisms for generating energy substrates, corticosteroids have anti‐inflammatory and immunosuppressive properties in higher doses and are applied in a wide variety of ways in almost all fields of medicine (Barnes 2006; Rhen 2005). For example, corticosteroids are used at high doses of more than 6 mg/kg up to 30 mg/kg methylprednisolone corresponding to more than 30 mg/kg up to 150 mg/kg hydrocortisone equivalents daily for short‐term, high‐dose pulse therapy against solid organ transplant rejection, or about 0.5 mg/kg hydrocortisone equivalents daily for prolonged therapy in different inflammatory lung diseases. A major representative of synthetic corticosteroids is the long‐acting compound dexamethasone. Examples of other synthetic corticosteroids with weaker and shorter activity are methylprednisolone and hydrocortisone (Bourdeau 2003). To obtain comparable effects, dosage equivalents are needed for the different corticosteroids. However, patients may have a higher risk of infections (including fungal infections) and can suffer from blood glucose problems, hypertension, oedema, and psychiatric side effects such as steroid psychosis and delirium (Schreiber 2014; Waljee 2017). The therapeutic use of higher doses of corticosteroids over a longer time suppresses the hypothalamic‐pituitary‐adrenal axis such that dosage‐tapering may be needed (Taves 2011; Yasir 2022). 

How the intervention might work

It has been proposed that corticosteroids could be clinically effective against severe and critical COVID‐19. A substantial percentage of patients develop severe and critical COVID‐19 that requires hospitalisation, with dyspnoea, hypoxia, or relevant lung involvement based on imaging, as well as respiratory failure, shock, or multi‐organ dysfunction requiring ventilator support (Thibeault 2021; Wu 2020). In COVID‐19, an insufficient host defence and unbalanced inflammation is thought to play a key role in the pathophysiology of hypoxaemic respiratory failure (Schulte‐Schrepping 2020). A systemic inflammatory response with the excessive release of cytokines and inflammation mediators can lead to lung injury with the development of acute respiratory distress syndrome (ARDS). The potent anti‐inflammatory effects of corticosteroids might prevent or mitigate these deleterious effects by modulating cytokine release (Villar 2020). Corticosteroids have been widely used in syndromes closely related to COVID‐19, including SARS, MERS, severe influenza, and community‐acquired pneumonia. The evidence to support or discourage the use of corticosteroids under these conditions has been weak. Corticosteroids can induce harm through immunosuppressive effects during the treatment of infection. In SARS‐CoV‐2 infection, viral shedding appears early in the illness and declines thereafter. The effect of corticosteroid therapy on virus clearance in COVID‐19 needs to be taken into consideration as well as the unwanted adverse events. In acutely critically ill people, dexamethasone has comparatively few side effects (Rochwerg 2018). However, patients may suffer from blood glucose variations and potential invasive fungal infections. The therapeutic use of higher doses of corticosteroids over a longer time (more than 21 days) suppresses the hypothalamic‐pituitary‐adrenal axis such that dosage‐tapering may be needed.

Why it is important to do this review

Systemic corticosteroids are now part of the standard therapy for COVID‐19 in patients having additional oxygen supply and/or mechanical ventilation, and are recommended in national and international guidelines (NICE 2021; NIH 2021; WHO 2021b), based on systematic reviews regarding interventions for COVID‐19, including corticosteroids. For example, several systematic reviews investigated the association between the use of corticosteroids and COVID‐19‐related mortality based on randomised controlled trials (RCTs) and non‐randomised studies (e.g. Chaudhuri 2021; Sterne 2020; Van Paassen 2020). Comparison of different corticosteroids, dosages, and time points of administration in terms of clinical progression, or comparison with other active substances, especially other anti‐inflammatory substances, appears necessary in order to define the role of corticosteroids in the treatment of COVID‐19. This not only relates to effectiveness but also the risk profile, especially with regard to a potential risk of systemic fungal infections. This, moreover, is interconnected with the equity considerations that this updated systematic review is addressing: do participant characteristics such as age, sex, place of residence, or ethnicity stratify clinical outcomes and influence how large a benefit a participant can expect in terms of mortality reduction, or should certain subgroups not receive systemic corticosteroids at all? 

We were encouraged by Tomlinson 2021 to not only present the baseline characteristics of participants but also to analyse outcome data with respect to equity‐related aspects. We aimed to perform subgroup analyses stratified by age group, sex, income group of the country of origin, and ethnicity for mortality outcomes in all comparisons. We used PROGRESS‐Plus to consider the equity‐related stratifying factors below (Welch 2012):

• Sex: sex of study participants might influence the outcome and potentially access to care (Ahrenfeldt 2020).

• Age: age is a known risk factor for severe disease and potentially access to care (Ahrenfeldt 2020).

• Ethnicity: existing studies have indicated that some ethnicities might have a worse outcome (e.g. people of colour, including in a high‐income country), therefore it is important to evaluate whether intervention effects are consistent across all ethnicities (Navar 2021).

• Place of residence: we did not evaluate at country‐level, but by income ranking (low‐ and middle‐income versus high‐income countries). One reason for this was to see whether this effective treatment has also been evaluated in low‐ and middle‐income countries and whether the results are comparable, as this intervention is cheap and available almost everywhere (Oke 2020).

The 'living' approach in evidence synthesis is necessary in order to incorporate new knowledge into the data evaluation, to keep the evidence assessment up‐to‐date and thus to offer a better basis for recommendations and everyday clinical decisions. This Cochrane Review is the first update of our published review (Wagner 2021a). The update was necessary because important new studies have been published in the meantime. It will fill current evidence gaps by identifying, describing, evaluating, and meta‐analysing RCTs of systemic corticosteroids in relation to clinical outcomes in COVID‐19. Unlike other systematic reviews in this field, it considers the outcome clinical improvement (defined as Discharged alive) and worsening (defined as a combined endpoint New need for invasive ventilation or death), and provides equity‐relevant subgroup analyses and methodologically critical sensitivity analyses. This living systematic review will be updated once new evidence becomes available.

Objectives

To assess whether and at which doses systemic corticosteroids are effective and safe in the treatment of people with COVID‐19, and to keep up‐to‐date with the evolving evidence base using a living systematic review approach. Apart from adding newly published evidence to existing and new comparisons, we would like to assess equity‐related aspects quantitatively in subgroup analyses where possible.

Methods

Criteria for considering studies for this review

Types of studies

The main description of methods is based on Cochrane Haematology's standard template and is in line with a series of Cochrane Reviews investigating treatments and therapies against COVID‐19. We made specific adaptations related to the research question if necessary. The protocol for this review was registered with PROSPERO on 21 December 2020 (Wagner 2021b).

To assess the efficacy and safety of systemic corticosteroids against COVID‐19, we included RCTs, as this study design, if performed appropriately, provides the best evidence for experimental therapies in highly controlled therapeutic settings. We used the methods recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022a). We would also have accepted cluster‐randomised trials for inclusion, if we had found any.

We included the following formats if sufficient information was available on study design, characteristics of participants, interventions, and outcomes:

• full‐text publications;

• preprint articles.

For our previous version of the review (Wagner 2021a), we included preprints for a complete overview of ongoing research activity, especially for tracking newly emerging studies about the use of systemic corticosteroids against COVID‐19, but in this updated version all studies that contributed to data/analyses are peer‐reviewed RCTs. 

We did not apply any limitation with respect to the length of follow‐up. The types of study designs remained the same as in the first version of this review, and we did not include any additional study designs that may be more suited for conducting equity‐related analyses.

Types of participants

We included adults with a suspected or confirmed diagnosis of COVID‐19 (as described in the study) and we did not exclude any studies based on sex, ethnicity, disease severity, or setting.

We excluded studies evaluating the use of corticosteroids against coronavirus diseases such as SARS or MERS, or other viral diseases, such as influenza. If studies enrolled populations with or exposed to mixed viral diseases, we had planned to only include these if the study authors provided subgroup data for SARS‐CoV‐2 infection.

Types of interventions

We included the following interventions:

• any type or dose of systemic corticosteroids;

• oral or intravenous application.

We had planned to include the following comparisons:

• systemic corticosteroids plus standard care versus standard care (plus/minus placebo);

• dose comparisons;

• timing comparisons (early versus late);

• different types of corticosteroids;

• systemic corticosteroid versus other active substances.

However, comparisons of different timings (early versus late) and systemic corticosteroids versus other active substances were not available.

Standard care in both arms should be similar.

We excluded the following interventions:

• corticosteroid plus other active substance versus standard care;

• topical corticosteroids;

• inhaled corticosteroids;

• corticosteroids for long‐COVID treatment.

Types of outcome measures

We evaluated core outcomes in accordance with the Core Outcome Measures in Effectiveness Trials (COMET) Initiative for COVID‐19 patients (COMET 2020; Marshall 2020), and additional outcomes that have been prioritised by consumer representatives and the panel of the German 'National Treatment Guidance for Hospitalised COVID‐19 Patients' (Kluge 2022).

We defined this outcome set for hospitalised individuals with a confirmed or suspected diagnosis of COVID‐19 and moderate to severe disease, according to the WHO clinical progression scale stage 4 to 9 (Marshall 2020); that is, all patients who were hospitalised because of symptomatic COVID‐19 treated with all different levels of respiratory support (no additional oxygen, low‐flow oxygen prongs or mask ('low‐flow oxygen' only hereafter), high‐flow oxygen or non‐invasive ventilation (NIV), invasive mechanical ventilation including extracorporeal membrane oxygenation (ECMO)), and individuals with a confirmed or suspected diagnosis of SARS‐CoV‐2 infection and asymptomatic or mild disease, according to the WHO clinical progression scale (Marshall 2020). Of note, readers may encounter respiratory support both as a baseline characteristic and as an outcome measure ‐ in the latter case we used changes in the level of support.

Individuals with a suspected or confirmed diagnosis of COVID‐19 and moderate to severe disease

Prioritised outcomes (included in the summary of findings tables)

• Mortality: 

  • All‐cause mortality

    • Up to day 30 (from here on for simplicity in this version of the review All‐cause mortality up to 30 days)

    • Any longer observation period from day 31 on (from here on All‐cause mortality up to 120 days)

• Improvement of clinical status during the longest observation period available:

  • Participants discharged alive (without clinical worsening or death)

• Worsening of clinical status during the longest observation period available:

  • New need for invasive mechanical ventilation or death

• Serious adverse events, defined as the number of participants with any serious adverse event (serious as defined according to CTCAE (Common Terminology Criteria for Adverse Events))

• Adverse events (any grade), defined as the number of participants with any adverse event

• Specific adverse events: hospital‐acquired infections, defined as the number of participants with an event

• Specific adverse events: invasive fungal infections, defined as the number of participants with an event

Prioritised outcomes (not included in the summary of findings tables):

• Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available

Additional outcomes (not included in the summary of findings tables)

• New need for dialysis during the longest period available

• Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days

Individuals with a suspected or confirmed diagnosis of SARS‐CoV‐2 infection and asymptomatic or mild disease

Prioritised outcomes (included in the summary of findings tables)

• Mortality: all‐cause mortality up to day 30 or any longer observation period

• Admission to hospital or death within 28 days

• Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available

• Serious adverse events, defined as the number of participants with any serious adverse event (serious as defined according to CTCAE (Common Terminology Criteria for Adverse Events)

• Adverse events (any grade), defined as the number of participants with any adverse event

• Specific adverse events: infections, defined as the number of participants with an event

• Specific adverse events: invasive fungal infections, defined as the number of participants with an event

Additional outcomes (not included in the summary of findings tables)

• Viral clearance, assessed with RT‐PCR test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days

Timing of outcome measurement

In the case of time‐to‐event analysis, for example, for time to clinical improvement, we included the outcome measure based on the longest follow‐up time. We also collected information on outcomes from all other time points reported in the publications.

Search methods for identification of studies

Electronic searches

Our information specialist (MIM) conducted systematic searches in the following sources from the inception of each database to 6 January 2022 (search date for all databases) and did not place restrictions on the language of publication.

• Cochrane COVID‐19 Study Register (CCSR) (www.covid-19.cochrane.org), comprising:

  • MEDLINE (PubMed), weekly updates;

  • Embase.com, weekly updates;

  • ClinicalTrials.gov (www.clinicaltrials.gov), daily updates;

  • WHO International Clinical Trials Registry Platform (ICTRP) (www.who.int/trialsearch), weekly updates;

  • medRxiv (www.medrxiv.org), weekly updates;

  • Cochrane Central Register of Controlled Trials (CENTRAL), monthly updates.

• Web of Science Core Collection (Clarivate), from 1 January 2020 onwards:

  • Science Citation Index Expanded (1945 to present);

  • Emerging Sources Citation Index (2015 to present).

• WHO COVID‐19 Global literature on coronavirus disease (search.bvsalud.org/global-literature-on-novel-coronavirus-2019-ncov/).

Database search results for Web of Science were restricted to publications from 2020 to the present date, as no treatment studies on COVID‐19 were registered prior to January 2020. For detailed search strategies, see Appendix 1 (previous review version) and Appendix 2 (current review version).

We did not conduct separate searches of the databases required by the Methodological Expectations of Cochrane Intervention Reviews (MECIR) standards (Higgins 2022), since these databases are already being regularly searched for the production of the CCSR.

Searching other resources

We identified other potentially eligible studies or ancillary publications by searching the reference lists of included studies and systematic reviews.

Data collection and analysis

Selection of studies

Two out of three review authors (MSp, CW, ALF) independently screened the results of the search strategies for eligibility for the review by reading the titles and abstracts using Ouzzani 2016. We coded the abstracts as either 'include' or 'exclude'. In the case of disagreement or if it was unclear whether we should retrieve the abstract or not, we obtained the full‐text publication for further discussion. Two review authors assessed the full‐text articles of selected studies. If the two review authors were unable to reach a consensus, they consulted a senior review author to reach a final decision.

We documented the study selection process in a flow chart, as recommended in the PRISMA statement (Moher 2009), and showed the total numbers of retrieved references and the numbers of included and excluded studies. We listed all studies that we excluded after full‐text assessment and the reasons for their exclusion in the Characteristics of excluded studies section.

Data extraction and management

We conducted data extraction according to the guidelines proposed by Cochrane (Li 2021). Two out of five review authors (CW, MSp, JD, ALF, MG) extracted data independently and in duplicate, using a customised data extraction form developed in Microsoft Excel (Microsoft 2018). We resolved disagreements by discussion. If we were unable to reach agreement, we involved a third review author.

Two review authors (CW, MSp) independently assessed eligible studies obtained in the process of study selection (as described above) for methodological quality and risk of bias. If the review authors were unable to reach a consensus, they consulted a third review author.

We extracted the following information if reported:

• General information: author, title, source, publication date, country, language, duplicate publications

• Study characteristics: trial design, setting and dates, source of participants, inclusion/exclusion criteria, comparability of groups, treatment cross‐overs, compliance with assigned treatment, length of follow‐up

• Participant characteristics: age, sex, ethnicity, number of participants recruited/allocated/evaluated, number of participants with positive, negative or unknown RT‐PCR test result, additional diagnoses, severity of disease, previous treatments, concurrent treatments, co‐morbidities (e.g. diabetes, immunosuppression), vaccination status

• Interventions: type of corticosteroid, dose, frequency, timing, duration and route of administration, setting (e.g. inpatient, outpatient), duration of follow‐up

• Control interventions: placebo, no treatment or other intervention; dose, frequency, timing, duration and route of administration, setting, duration of follow‐up

• Outcomes: as specified under Types of outcome measures

• Risk of bias assessment: randomisation process, deviations from the intended interventions, missing outcome data, measurement of the outcome, selection of the reported results

• Equity‐related aspects as per the PROGRESS‐Plus approach (Welch 2012): place of residence, race/ethnicity, occupation, sex, religion, education, socioeconomic status, social capital and personal characteristics associated with discrimination (e.g. age, disability); features of relationships; time‐dependent relationships 

  • Countries were classified socio‐economically based on the World Bank's latest version of the Country and Lending Groups (World Bank 2021)

Assessment of risk of bias in included studies

We used the RoB 2 tool (version of 22 August 2019) to analyse the risk of bias of study results (Sterne 2019). Of interest for this review is the effect of the assignment to the intervention (the intention‐to‐treat (ITT) effect); thus, we performed all assessments with RoB 2 on this effect. The outcomes that we assessed are those specified for inclusion in the summary of findings table.

Two out of six review authors (CW, MSp, MG, ALF, AAN, JD) independently assessed the risk of bias for each outcome. In case of discrepancies among their judgements and inability to reach consensus, we consulted the third review author to reach a final decision. We assessed the following types of bias as outlined in Chapter 8 (Higgins 2022b) and for cluster‐RCTs as outlined in Chapter 23 (Table 23.1.a; Higgins 2022c) of the Cochrane Handbook for Systematic Reviews of Interventions:

For RCTs:

• bias arising from the randomisation process;

• bias due to deviations from the intended interventions;

• bias due to missing outcome data;

• bias in measurement of the outcome;

• bias in selection of the reported result.

For cluster‐RCTs:

• bias arising from the randomisation process;

• bias arising from the timing of identification and recruitment of participants;

• bias due to deviations from intended interventions;

• bias due to missing outcome data;

• bias in measurement of the outcome;

• bias in selection of the reported result.

To address these types of bias we used the signalling questions recommended in RoB 2 and made a judgement using the following options.

• 'Yes': if there is firm evidence that the question is fulfilled in the study (i.e. the study is at low or high risk of bias for the given the direction of the question).

• 'Probably yes': a judgement has been made that the question is fulfilled in the study (i.e. the study is at low or high risk of bias given the direction of the question).

• 'No': if there is firm evidence that the question is unfulfilled in the study (i.e. the study is at low or high risk of bias for the given the direction of the question).

• 'Probably no': a judgement has been made that the question is unfulfilled in the study (i.e. the study is at low or high risk of bias given the direction of the question).

• 'No information': if the study report does not provide sufficient information to allow any judgement.

We used the algorithms proposed by RoB 2 to assign each domain one of the following levels of bias:

• low risk of bias;

• some concerns;

• high risk of bias.

Subsequently, we derived an overall risk of bias rating for each pre‐specified outcome in each study in accordance with the following suggestions.

• 'Low risk of bias': we judge the trial to be at low risk of bias for all domains for this result.

• 'Some concerns': we judge the trial to raise some concerns in at least one domain for this result, but not to be at high risk of bias for any domain.

• 'High risk of bias': we judge the trial to be at high risk of bias in at least one domain for the result, or we judge the trial to have some concerns for multiple domains in a way that substantially lowers confidence in the results.

We used the RoB 2 Excel tool to implement RoB 2 (available on the riskofbias.info website), and stored and presented our detailed RoB 2 assessments in the analyses section and as supplementary online material.

As we collected the data from the studies and assessed RoB 2, we noticed an issue with competing risk of death (Columbia Public Health 2021 as easily accessible introduction), which we discussed in Quality of the evidence of the first version of the review. We dealt with this issue within domain 3 of RoB 2 (Higgins 2019). For risk of bias in the subgroup analyses of mortality we applied the same assessments as for the respective studies' main mortality result, irrespective of whether data were primarily published or sent in upon request.

Additionally, we pioneered critical appraisal of specific sources of bias in platform trials using the Park 2020 checklist. Sources of bias were not assessed on result but at study level and this did not have a direct impact on GRADEing of the evidence.

Measures of treatment effect

For continuous outcomes, we recorded the mean, standard deviation (SD), and total number of participants in both treatment and control groups. Where continuous outcomes used the same scale, we performed analyses using the mean difference (MD) with 95% confidence intervals (CIs). For continuous outcomes measured with different scales, we had planned to perform analyses using the standardised mean difference (SMD). For interpreting SMDs, we would have re‐expressed SMDs in the original units of a particular scale with the most clinical relevance and impact.

For dichotomous outcomes, we recorded the number of events and total number of participants in both treatment and control groups. We reported the pooled risk ratio (RR) with a 95% CI (Deeks 2022).

If available, we had planned to extract and report hazard ratios (HRs) for time‐to‐event outcomes (e.g. time to liberation from invasive ventilation), but we did not find data to estimate them using the methods proposed by Parmar and Tierney (Parmar 1998; Tierney 2007).

Unit of analysis issues

The aim of this review is to summarise studies that analyse data at the level of the individual. We would also have accepted cluster‐randomised trials for inclusion, if we had found any. We collated multiple reports of one study so that the study, and not the report, is the unit of analysis. In case of adverse events, serious adverse events, hospital‐acquired infections, and invasive fungal infections only the number of events and not the number of participants are counted in most of the studies. We therefore asked the authors if they could provide us with the data as the number of participants with at least one event, i.e. as dichotomous data.

Studies with multiple treatment groups

As recommended in Chapter 6 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022d), for studies with multiple treatment groups of the same intervention (i.e. dose, route of administration), we planned to evaluate whether study arms were sufficiently homogeneous to be combined. If arms could not be pooled, we planned to compare each arm with the common comparator separately. For pair‐wise meta‐analysis, we planned to split the ‘shared’ group into two or more groups with smaller sample size, and include two or more (reasonably independent) comparisons. For this purpose, for dichotomous outcomes, we planned to divide both the number of events and the total number of participants, and for continuous outcomes, we planned to divide the total number of participants with unchanged means and SDs.

Dealing with missing data

Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions suggests a number of potential sources for missing data, which we took into account at study level, at outcome level, and at summary data level (Deeks 2022). At all levels, it is important to differentiate between data 'missing at random', which may often be unbiased, and 'not missing at random', which may bias study and thus review results.

Missing outcome data on general safety and corticosteroid‐specific safety, and missing adjustment for competing risk of death, had a negative impact on the certainty of the evidence in the first version of this review. Moreover, the influence of equity‐related aspects on mortality had not been examined quantitatively. Therefore, we requested from all corresponding authors data for Serious adverse events, Adverse events, Hospital‐acquired infections, and Invasive fungal infections standardised on the nominal scale as a dichotomous outcome, Participants with at least one event/patients at risk. Additionally, we requested safety data adjusted for competing risk of death and mortality data stratified by age group, sex, and ethnicity to explore equity‐related aspects. Finally, we asked corresponding authors of studies awaiting classification to clarify registration, protocol, and randomisation issues (Gautam 2021; Ghanei 2021; Rashad 2021), and make available the manuscript where only an abstract was published (Montalvan 2021).

Assessment of heterogeneity

We assessed heterogeneity of treatment effects between trials using a Chi² test with a significance level at P < 0.1. We used the I² statistic (Higgins 2003), and visual examination, to assess possible heterogeneity (I² statistic > 30% to signify moderate heterogeneity, I² statistic > 75% to signify considerable heterogeneity; Deeks 2022). If the I² statistic was above 80%, we had planned to explore potential causes through sensitivity and subgroup analyses (see Sensitivity analysis and Subgroup analysis and investigation of heterogeneity). For future updates, if we cannot identify reasons for heterogeneity in subgroup or sensitivity analysis, we will not perform a meta‐analysis but, instead, provide outcome data for all studies without an overall effect estimate.

Assessment of reporting biases

As mentioned above, we searched trials registries to identify completed studies that have not been published elsewhere, to minimise or determine publication bias. We intended to explore potential publication bias by generating a funnel plot and statistically testing this by conducting a linear regression test for meta‐analyses involving at least 10 trials (Sterne 2019). We considered P < 0.1 as significant for this test. We planned to generate a funnel plot, but had fewer than 10 studies reporting a comparable outcome.

Data synthesis

If the clinical and methodological characteristics of individual studies were sufficiently homogeneous, we pooled the data in a meta‐analysis. We performed analyses according to the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022). Additionally, we analysed studies that included different severities of disease separately, grouping them with respect to disease severity according to need for respiratory support at randomisation (see Types of outcome measures). We treated placebo and standard care as the same intervention, as well as standard care at different institutions and time points.

We used Review Manager Web (RevMan Web) software for analyses (RevMan Web 2019). One review author entered the data into RevMan Web, and a second review author checked the data for accuracy. For most of our analyses we used the random‐effects model as planned and as required to take into account differences, for example, in settings, disease severity, and co‐medications. However, in rare cases (i.e. Analysis 1.3; Analysis 1.4; Analysis 9.2), we decided to report the fixed‐effect model as the primary analysis so that very small studies with few events did not receive extraordinary weight compared to very large studies with many events. If we deemed meta‐analysis inappropriate for a certain outcome because of heterogeneity of the included studies both statistically or conceptually or due to too high a risk of bias, we presented descriptive statistics only.

1.3. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 3: Clinical improvement: discharged alive

1.4. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 4: Clinical worsening: new need for IMV or death

9.2. Analysis.

Comparison 9: High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 2: All‐cause mortality up to 120 days

If meta‐analysis was possible, we assessed the effects of potential biases in sensitivity analyses (see Sensitivity analysis). For binary outcomes, we based the estimation of the between‐study variance on the Mantel‐Haenszel method. We explored heterogeneity above 80% with subgroup and sensitivity analyses.

Subgroup analysis and investigation of heterogeneity

Our focus was on both clinical relevance and equity, therefore we performed subgroup analyses for all‐cause mortality based on participant characteristics that may stratify health outcomes for every comparison in which data were available, irrespective of observed statistical heterogeneity.

Clinical relevance

• Respiratory support at randomisation

• Type of systemic corticosteroid: dexamethasone versus (methyl‐)prednisolone versus hydrocortisone

Equity‐related aspects

• Sex: female versus male

• Age: < 70 years versus ≥ 70 years

• Ethnicity: Black, Asian, or other versus White versus unknown

• Place of residence: high‐income countries (HIC) versus low‐ and middle‐income countries (LMIC)

We also performed subgroup analyses for clinical improvement (discharged alive) because the I² statistic was found to be above 80%. We made our decision on subgroup definitions before performing the first analyses. For future updates, if the I² statistic is found to be above 80% for the other outcomes, we will also conduct subgroup analyses for these outcomes (see also Assessment of heterogeneity).

Sensitivity analysis

We performed the following sensitivity analysis for all‐cause mortality up to 30 and 120 days as well as for clinical improvement (discharged alive) because I² statistic for the latter was found to be above 80%:

• Risk of bias assessment components (studies with a low risk of bias or some concerns versus studies with a high risk of bias).

• Platform trials versus no platform trials.

• Fixed‐effect model versus random‐effects model.

• Preprint versus journal publication.

Summary of findings and assessment of the certainty of the evidence

We used the GRADE approach to assess the certainty of the evidence for the following outcomes, and prepared one summary of findings table per population.

Summary of findings

We used the GRADEpro GDT software to create summary of findings tables. For time‐to‐event outcomes, we would have calculated absolute effects at specific time points, as recommended in the GRADE guidance (Skoetz 2020).

According to Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions, the “most critical and/or important health outcomes, both desirable and undesirable, limited to seven or fewer outcomes” should be included in the summary of findings table(s) (Schünemann 2021). We included outcomes prioritised according to the core outcome sets for studies for the treatment of patients with confirmed or suspected COVID‐19 (COMET 2020), and patient relevance. These outcomes were as follows.

Individuals with a suspected or confirmed diagnosis of COVID‐19 and moderate to severe disease

• All‐cause mortality

  • Up to day 30 (from here on for simplicity in this version of the review All‐cause mortality up to 30 days)

  • Any longer observation period from day 31 on (from here on for simplicity in this version of the review All‐cause mortality up to 120 days)

• Improvement of clinical status during the longest observation period available:

  • Participants discharged alive (without clinical worsening or death)

• Worsening of clinical status during the longest observation period available:

  • New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (Figure 1)

• Serious adverse events, defined as the number of participants with any serious adverse event (serious as defined according to CTCAE (Common Terminology Criteria for Adverse Events)

• Adverse events (any grade), defined as the number of participants with any adverse event

• Specific adverse events: hospital‐acquired infections

• Specific adverse events: invasive fungal infections

1.

WHO Clinical Progression Scale (Marshall 2020). Copyright © 2020 Elsevier Ltd. All rights reserved: reproduced with permission.

ECMO = extracorporeal membrane oxygenation; FiO₂ = fraction of inspired oxygen; NIV = non‐invasive ventilation; pO₂ = partial pressure of oxygen; RNA = ribonucleic acid; SpO₂ = oxygen saturation.

*If hospitalised for isolation only, record status for ambulatory patients.

Individuals with a suspected or confirmed diagnosis of SARS‐CoV‐2 infection and asymptomatic or mild disease

• All‐cause mortality

  • Up to day 30

  • Any longer observation period

• Admission to hospital or death within 28 days

• Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available

• Serious adverse events, defined as the number of participants with any serious adverse event (serious as defined according to CTCAE (Common Terminology Criteria for Adverse Events)

• Adverse events (any grade), defined as the number of participants with any adverse event

• Specific adverse events: infections

Assessment of the certainty of the evidence

We used the GRADE approach to assess the certainty of the evidence for the outcomes listed in the previous section.

The GRADE approach uses five domains (risk of bias, inconsistency, imprecision, indirectness, and publication bias) to assess certainty in the body of evidence for each prioritised outcome.

We downgraded our certainty of the evidence for:

• serious (‐1) or very serious (‐2) risk of bias;

• serious (‐1) or very serious (‐2) inconsistency;

• serious (‐1) or very serious (‐2) uncertainty about directness;

• serious (‐1) or very serious (‐2) imprecise or sparse data;

• serious (‐1) or very serious (‐2) probability of reporting bias.

The GRADE system uses the following criteria for assigning grade of evidence.

• 'High': we are very confident that the true effect lies close to that of the estimate of the effect.

• 'Moderate': we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different.

• 'Low': our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.

• 'Very low': we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

We followed the current GRADE guidance for these assessments in its entirety as recommended in the Cochrane Handbook for Systematic Reviews of Interventions, Chapter 14 (Schünemann 2021).

We used the overall risk of bias judgement, derived from the RoB 2 Excel tool, to inform our decision on downgrading for risk of bias. We phrased the findings and certainty in the evidence as suggested in the informative statement guidance (Santesso 2020).

Methods for future updates

Living systematic review considerations

Our Information Specialist (MIM) will provide us with new search records each week, which two review authors will screen, extract, evaluate, and integrate following the guidance for Cochrane living systematic reviews (Living Evidence Network 2019).

We will manually check platform trials that were previously identified and listed as 'studies awaiting classification' for additional treatment arms.

We will wait until the accumulating evidence changes our conclusions in the implications for research and practice before republishing the review. We will consider one or more of the following components to inform this decision:

• findings of one or more prioritised outcomes;

• credibility (e.g. GRADE rating) of one or more prioritised outcomes;

• new settings, populations, interventions, comparisons, or outcomes studied.

In case of emerging policy relevance because of global controversies around the intervention, we will consider republishing an updated review even though our conclusions remain unchanged. We will review the review scope and methods approximately monthly, or more frequently if appropriate, in light of potential changes in COVID‐19 research (for example, when additional comparisons, interventions, subgroups or outcomes, or new review methods become available).

Results

Description of studies

Results of the search

We searched all databases and screened the resulting records up to 6 January 2022. We identified 3552 records. After removing duplicates, we screened 2682 records based on their titles and abstracts. We excluded 2546 records that did not meet the inclusion criteria. Of the remaining 136 records, we included 105 records:

• 16 RCTs (in 32 records) for inclusion in this review of which five were newly included in this second version (Maskin 2021; Munch 2021a; Munch 2021b; Taboada 2021; Toroghi 2021);

• 42 RCTs (in 47 records) are ongoing;

• 23 RCTs (in 26 records) are awaiting classification as they have been reported as being completed, but the results have not yet been published or they lack relevant information on the study design.

The study flow diagram in Figure 2 illustrates the study selection process according to the PRISMA guidelines (Moher 2009).

2.

PRISMA flow diagram illustrating our study selection process.

Included studies

Design and sample size

We included 16 RCTs, of which two were multicentre platform RCTs (Horby 2021; Angus 2020), eight were multicentre RCTs (Corral‐Gudino 2021; Dequin 2020; Edalatifard 2020; Maskin 2021; Munch 2021a; Munch 2021b; Tang 2021; Tomazini 2020), and six were single‐centre RCTs (Farahani 2021; Jamaati 2021; Jeronimo 2020; Ranjbar 2021; Taboada 2021; Toroghi 2021).

Setting

Of 9549 participants in the included studies, 8418 (88%) originated from seven studies from high‐income countries (Angus 2020; Corral‐Gudino 2021; Dequin 2020; Horby 2021; Munch 2021a; Munch 2021b; Taboada 2021); 1131 (12%) participants originated from nine studies from upper‐ and lower‐middle‐income countries (Edalatifard 2020; Farahani 2021; Jamaati 2021; Jeronimo 2020; Maskin 2021; Ranjbar 2021; Tang 2021; Tomazini 2020; Toroghi 2021). There were no studies from low‐income countries.

Participants

All participants were adults hospitalised for either acute proven COVID‐19 or, as in the case of Angus 2020, Dequin 2020, Horby 2021, Jeronimo 2020, and Tomazini 2020, suspected acute COVID‐19. Positive RT‐PCR rates within the studies ranged from 100% in Maskin 2021, Munch 2021a, and Munch 2021b to about 80% in Angus 2020. All included participants were hospitalised because of symptomatic (suspected) COVID‐19 and were treated with different levels of respiratory support (no oxygen, low‐flow oxygen, high‐flow oxygen or non‐invasive ventilation (NIV), or invasive mechanical ventilation including extracorporeal membrane oxygenation (ECMO)). Based on the different levels of respiratory support at baseline, disease severity ranged from 4 to 9 on the WHO Clinical Progression Scale (Marshall 2020).

Shares of female participants ranged between 20% in Munch 2021a and 52% in Tang 2021. Means of age ranged between 54 years (standard deviation (SD) 15) in the intervention group in Jeronimo 2020 and 73 years (SD 11) in the intervention group in Corral‐Gudino 2021. No explicit information could be found regarding the proportion of vaccinated participants in any included study. However, recruitment of all studies in the comparisons 1 and 2 had ended before the first human was injected a vaccine dose outside a clinical trial worldwide on 8 December 2020 so that we deem it reasonable to declare all those participants in comparison 1 and 2 (Table 1; Table 2) "unvaccinated" with consequences for the applicability of the evidence today (Watson 2022). However, recruitment in comparison 3 (Table 3) took place until May 2021 so that we declared the vaccination status of these participants "unknown".

Interventions

Of the included completed studies, 11 compared systemic corticosteroids, which were hydrocortisone, prednisolone, methylprednisolone, and dexamethasone, to standard care (plus/minus placebo), one compared methylprednisolone to dexamethasone (Ranjbar 2021), and four directly compared different doses of dexamethasone (Maskin 2021; Munch 2021b; Taboada 2021; Toroghi 2021). The route of administration was intravenous except in Horby 2021, where both oral and intravenous administration were allowed, and Farahani 2021, with oral dose‐tapering after intravenous initiation. Doses as hydrocortisone equivalents were ≤ 200 mg/day in Angus 2020, Dequin 2020, Horby 2021, and Munch 2021a; 201 to 500 mg/day in Corral‐Gudino 2021, Jamaati 2021, Jeronimo 2020, Tang 2021, and Tomazini 2020; 1250 mg/day hydrocortisone equivalent and more as pulse dose regimen in Edalatifard 2020 and Farahani 2021. Durations of treatment differed between the studies, depending on the type of corticosteroid used, in a range between 3 days (methylprednisolone 250 mg/day, in Edalatifard 2020) and 14 days (hydrocortisone 200 mg/day tapered to 50 mg/day, in Dequin 2020). All the studies investigating the effects of dexamethasone reported treatment durations limited to 10 days (Horby 2021; Jamaati 2021; Maskin 2021; Munch 2021b; Taboada 2021; Tomazini 2020; Toroghi 2021). None of the included studies stratified data analyses according to treatment duration.

Equity‐related aspects

No study data were available for the following equity‐related elements: occupation, religion, education, socioeconomic status, and social capital. Data regarding place of residence, age, and sex were reported in all studies. Angus 2020, Horby 2021, and Jeronimo 2020 are the only trials reporting ethnicity. For details please see Table 4.

1. Equity elements of the included studies.

Study/equity‐element	Place of residence (World Bank 2021)	Age	Ethnicity (%)	Occupation	Sex (% female)	Religion	Education	Socioeconomic status	Social capital
Angus 2020	High‐income country (Australia, Canada, France, Ireland, the Netherlands, New Zealand, UK, USA)	Mean (years, SD) Fixed‐dose intervention group: 60.4 (11.6) Shock‐dependent intervention group: 59.5 (12.7) Control group: 59.9 (14.6)	Fixed‐dose intervention group: White: 79 (72.2%) Asian: 18 (16.2%) Black: 4 (3.6%) Mixed: 4 (3.6%) Other: 6 (5.4%) Shock‐dependent intervention group: White: 80 (76.2%) Asian: 11 (10.5%) Black: 7 (6.7%) Mixed: 0 Other: 7 (6.7%)   Control group: White: 45 (57%) Asian: 22 (27.9%) Black: 4 (5.1%) Mixed: 2 (2.5%) Other: 6 (7.6%)	NR	Fixed‐dose intervention group: 39 (28.5%) Shock‐dependent intervention group: 43 (29.5%) Control group: 29 (28.7%)	NR	NR	NR	NR
Corral‐Gudino 2021	High‐income country (Spain)	Mean (years, SD) Intervention group: 73 (11) Control group: 66 (12)	NR	NR	Intervention group: 11 (32%) Control group: 13 (45%)	NR	NR	NR	NR
Dequin 2020	High‐income country (France)	Median (years, IQR) Intervention group: 63.1 (51.5 to 70.8) Control group: 66.3 (53.5 to 72.7)	NR	NR	Intervention group: 22 (28.9%) Control group: 23 (31.5%)	NR	NR	NR	NR
Edalatifard 2020	Middle‐income country (Iran)	Mean (years, SD) Intervention group: 55.8 (16.35) Control group: 61.7 (16.62)	NR	NR	Intervention group: 10 (29.4%) Control group: 13 (46.4%)	NR	NR	NR	NR
Farahani 2021	Middle‐income country (Iran)	Mean (years, SD): Intervention group: 61.07 (12.83) Control group: 66.80  (14.03)	NR	NR	Intervention group: 4 (28.6%) Control group: 6 (40%)	NR	NR	NR	NR
Horby 2021	High‐income country (UK)	Mean (years, SD): Intervention group: 66.9 (15.4) Control group: 65.8 (15.8)	Intervention group: White: 1550 (74%) Black, Asian or minority ethnic group: 364 (17%) Unknown: 190 (9%) Control group: White: 3139 (73%) Black, Asian or minority ethnic group: 783 (18%) Unknown: 399 (9%)	NR	Intervention group: 766 (36%) Control group: 1572 (36%)	NR	NR	NR	NR
Jamaati 2021	Middle‐income country (Iran)	Median (years, IQR) Intervention group survivor: 54 (37 to 63) Intervention group non‐survivor: 63 (55.5 to 72.5) Control group survivor: 61.5 (54 to 62) Control group non‐survivor: 67 (48 to 73)	NR	NR	Intervention group: 7 (28%) Control group: 7 (28%)	NR	NR	NR	NR
Jeronimo 2020	Middle‐income country (Brazil)	Mean (years, SD) Intervention group: 54 (15) Control group: 57 (15)	Intervention group: White: 30 (15.5%) Mixed: 147 (75.8%) Black: 6 (3.1%) Asian: 3 (1.5%) Amerindian: 8 (4.1%) Control group: White: 28 (14.1%) Mixed: 147 (73.9%) Black: 17 (8.5%) Asian: 3 (1.5%) Amerindian: 4 (2.0%)	NR	Intervention group: 68 (35.1%) Control group: 71 (35.7%)	NR	NR	NR	NR
Maskin 2021	Middle‐income country (Argentina)	Mean age (years, SD) Low‐dose group: 60.04 (13.08) High‐dose group: 63.57 (13.59)	NR	NR	Low‐dose dexamethasone: 16 (33%) High‐dose dexamethasone: 13 (26%)	NR	NR	NR	NR
Munch 2021a	High‐income country (Denmark)	Median (years, IQR) Intervention group: 59 (52 to 74) Control group: 62 (55 to 71)	NR	NR	Intervention group: 2 (12%) Control group: 4 (29%)	NR	NR	NR	NR
Munch 2021b	High‐income country (Denmark, India, Sweden, Switzerland)	Median (years, IQR) Low‐dose group: 64 (54 to 72) High‐dose group: 65 (56 to 74)	NR	NR	Low‐dose dexamethasone: 154 (32%) High‐dose dexamethasone: 151 (30%)	NR	NR	NR	NR
Ranjbar 2021	Middle‐income country (Iran)	Mean (years, SD) Methylprednisolone group: 56.2 (17.5) Dexamethasone group: 61.3 (17.3)	NR	NR	Methylprednisolone group: 17 (38.6%) Dexamethasone group: 20 (47.6%)	NR	NR	NR	NR
Taboada 2021	High‐income country (Spain)	Mean age (years, SD) Low‐dose group: 64.8 (14.1) High‐dose group: 63.9 (14.5)	NR	NR	Low‐dose dexamethasone: 41 (39.6%) High‐dose dexamethasone: 36 (36.7%)	NR	NR	NR	NR
Tang 2021	Middle‐income country (China)	Median (years, IQR) Intervention group: 57 (49 to 67) Control group: 55 (38 to 65)	NR	NR	Intervention group: 22 (51.2%) Control group: 23 (53.5%)	NR	NR	NR	NR
Tomazini 2020	Middle‐income country (Brazil)	Mean (years, SD) Intervention group: 60.1 (15.8) Control group: 62.7 (13.1)	NR	NR	Intervention group: 61 (40.4%) Control group: 51 (34.5%)	NR	NR	NR	NR
Toroghi 2021	Middle‐income country (Iran)	Mean age (years, SD) Low‐dose group (8 mg once a day): 59 (14) Intermediate‐dose group (8 mg twice a day): 59 (17) High‐dose group (8 mg 3 times a day): 56 (16)	NR	NR	Low‐dose dexamethasone: 19 (40.4%) High‐dose dexamethasone: 34 (39.5%)	NR	NR	NR	NR

NR: not reported; SD: standard deviation; IQR: interquartile range

Included studies for comparison of systemic corticosteroids plus standard care to standard care (plus/minus placebo)

We included 11 studies describing 8019 participants in this comparison, of whom 3002 were randomised to corticosteroids and 5017 to standard care (plus/minus placebo). Daily hydrocortisone equivalents of the initial doses ranged from 150 mg to 5000 mg and durations of treatment ranged from zero to approximately 20 days. The majority of participants (n = 2577; 86%) randomised to corticosteroids received equivalents of 200 mg/day or less, 463 (15%) received 201 mg/day to 500 mg/day, and 48 (2%) received 501 mg/day to 5000 mg/day. Please see Table 5 for details, but note that no outcome data from Farahani 2021 (29 participants) were applicable for further analysis, resulting in quantitative analysis of 10 trials only in this comparison.

2. Characteristics of the included studies for the comparison: systemic corticosteroids versus placebo or standard care for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Study ID	Intervention and regimen	Hydrocortisone equivalent of initial dose: for 80 kg bodyweight if applicable (Stoelting 2006)	Control	Randomised to corticosteroids	Randomised to control	Study design	Place of residence (World Bank 2021) and recruitment period	Age	Sex (% female)	Population/disease severity at randomisation on WHO Clinical Progression Scale (Marshall 2020)
Angus 2020	Hydrocortisone, IV, 150 mg daily for 7 days	150 mg/d	Standard care	143 (fixed‐dose) and 152 (shock‐dependent dose)a	108	Platform Open‐label Multicentre	High‐income country (Australia, Canada, France, Ireland, the Netherlands, New Zealand, United Kingdom, USA) between March and June 2020	Mean (years, SD) Fixed‐dose intervention group: 60.4 (11.6) Shock‐dependent intervention group: 59.5 (12.7) Control group: 59.9 (14.6)	Fixed‐dose intervention group: 39 (28.5%) Shock‐dependent intervention group: 43 (29.5%) Control group: 29 (28.7%)	Severe ≥ 6
Corral‐Gudino 2021	Methylprednisolone, IV 80 mg for 3 days + 40 mg for 3 days	400 mg/d	Standard care	35	29	Open‐label Multicentre	High‐income country (Spain) between April and May 2020	Mean (years, SD) Intervention group: 73 (11) Control group: 66 (12)	Intervention group: 11 (32%) Control group: 13 (45%)	Moderate to severe 5 to 6
Dequin 2020	Hydrocortisone, IV 200 mg for 7 days, 100 mg for 4 days + 50 mg for 3 days	200 mg/d	Placebo	76	73	Double‐blind Multicentre	High‐income country (France) between March and June 2020	Median (years, IQR) Intervention group: 63.1 (51.5 to 70.8) Control group: 66.3 (53.5 to 72.7)	Intervention group: 22 (28.9%) Control group: 23 (31.5%)	Moderate to severe ≥ 5
Edalatifard 2020	Methylprednisolone, IV, 250 mg for 3 days	1250 mg/d	Standard care	34	34	Multicentre	Middle‐income country (Iran) between March and May 2020	Mean (years, SD) Intervention group: 55.8 (16.35) Control group: 61.7 (16.62)	Intervention group: 10 (29.4%) Control group: 13 (46.4%)	Moderate to severe 5 to 6
Farahani 2021	Methylprednisolone, IV 1000 mg/d for 3 days + tapering with 1 mg/kg prednisolone for 10 days	5000 mg/d	Standard care	14	15	Open‐label Single‐centre	Middle‐income country (Iran) between March and May 2020	Mean (years, SD): Intervention group: 61.07 (12.83) Control group: 66.80  (14.03)	Intervention group: 4 (28.6%) Control group: 6 (40%)	Moderate to severe 5 to 6
Horby 2021	Dexamethasone, IV or oral 6 mg daily for 10 days	150 mg/d	Standard care	2104	4321	Platform Open‐label Multi‐centre	High‐income country (United Kingdom) between May and June 2020	Mean (years, SD): Intervention group: 66.9 (15.4) Control group: 65.8 (15.8)	Intervention group: 766 (36%) Control group: 1572 (36%)	Moderate to severe 4 to 9
Jamaati 2021	Dexamethasone, IV, 20 mg for 5 days + 10 mg for 5 days	500 mg/d	Standard care	25	25	Open‐label Single‐centre	Middle‐income country (Iran) in March 2020	Median (years, IQR) Intervention group survivor: 54 (37 to 63) Intervention group non‐survivor: 63 (55.5 to 72.5) Control group survivor: 61.5 (54 to 62) Control group non‐survivor: 67 (48 to 73)	Intervention group: 7 (28%) Control group: 7 (28%)	Most likely moderate 5; no IMV at randomisation
Jeronimo 2020	Methylprednisolone (as sodium succinate), IV 1 mg/kg for 5 days	400 mg/d	Placebo	209	207	Double‐blind Single‐centre	Middle‐income country (Brazil) between April and June 2020	Mean (years, SD) Intervention group: 54 (15) Control group: 57 (15)	Intervention group: 68 (35.1%) Control group: 71 (35.7%)	Moderate to severe 5 to 9
Munch 2021a	Hydrocortisone, IV, 200 mg per day, for 7 days or until hospital discharge	200 mg/d	Placebo	16	14	Triple‐blind Multicentre	High‐income country (Denmark) between April and June 2020	Median (years, IQR) Intervention group: 59 (52 to 74) Control group: 62 (55 to 71)	Intervention group: 2 (12%) Control group: 4 (29%)	Severe ≥ 6
Tang 2021	Methylprednisolone, IV, 1 mg/kg for 7 days	400 mg/d	Placebo	43	43	Single‐blind Multicentre	Middle‐income country (China) between February and March 2020	Median (years, IQR) Intervention group: 57 (49 to 67) Control group: 55 (38 to 65)	Intervention group: 22 (51.2%) Control group: 23 (53.5%)	Moderate 4 to 5
Tomazini 2020	Dexamethasone, IV, 20 mg for 5 days + 10 mg for 5 days	500 mg/d	Standard care	151	148	Open‐label Multicentre	Middle‐income country (Brazil) between April and June 2020	Mean (years, SD) Intervention group: 60.1 (15.8) Control group: 62.7 (13.1)	Intervention group: 61 (40.4%) Control group: 51 (34.5%)	Severe 7 to 9
d: day;IMV: invasive mechanical ventilation; IV: intravenous; SD: standard deviation; IQR: interquartile range

^a Shock‐dependent dose: shock‐dependent dosing strategy was that restricting hydrocortisone to the period when the patient had overt shock would maximise the risk‐benefit ratio. Shock was defined as the requirement for intravenous vasopressor infusion for the treatment of shock presumed due to COVID‐19. Hydrocortisone was discontinued in the shock‐dependent group once shock was considered to have resolved or vasopressors had been discontinued for 24 hours.

Included studies for comparison of different types of systemic corticosteroids

We included Ranjbar 2021 describing 86 participants in this comparison, of whom 44 were randomised to methylprednisolone and 42 to dexamethasone. For details please see Table 6.

3. Characteristics of the included studies for the comparison: methylprednisolone versus dexamethasone for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19.

Study ID	Intervention A	Hydrocortisone equivalent of initial dose (for 80 kg bodyweight if applicable)	Intervention B	Randomised to intervention A	Randomised to Intervention B	Study design	Place of residence (World Bank 2021) and recruitment period	Age	Sex (% female)	Population/disease severity at randomisation on WHO Clinical Progression Scale (Marshall 2020)
Ranjbar 2021	Methylprednisolone 2 mg/kg, i.e. for an 80 kg participant: IV 160 mg for 5 days + 80 mg for 5 days + 40 mg for 5 days + 20 mg for 5 days (approximation of tapering scheme)	800 mg/d in the methylprednisolone arm for an 80 kg participant   1000 mg/d in the dexamethasone arm fixed	Dexamethasone, IV, 6 mg for 10 days	44	42	Triple‐blind Single‐centre	Middle‐income country (Iran) between August and November 2020	Mean (years, SD) Methylprednisolone group: 56.2 (17.5) Dexamethasone group: 61.3 (17.3)	Intervention A group: 17 (38.6%) Intervention B group: 20 (47.6%)	Moderate 4 to 5
IV: intravenous, d: day, SD: standard deviation

Included studies for comparison of high‐dose dexamethasone to low‐dose dexamethasone

We included Maskin 2021, Munch 2021b, Taboada 2021, and Toroghi 2021 describing 1444 participants in this comparison, of whom 746 were randomised to high‐dose (12 mg/day or higher) dexamethasone and 698 to low‐dose (6 mg to 8 mg/day) dexamethasone. For details please see Table 7.

4. Characteristics of the included studies for the comparison: high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8 mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19.

Study ID	Intervention dose and regimen	Hydrocortisone equivalent of initial dose: for 80 kg bodyweight if applicable (Stoelting 2006)	Control	Randomised to high‐dose dexamethasone	Randomised to low‐dose dexamethasone	Study design	Place of residence (World Bank 2021) and recruitment date	Age	Sex (% female)	Population/disease severity at randomisation on WHO Clinical Progression Scale (Marshall 2020)
Munch 2021b	Dexamethasone, IV, 12 mg/day for up to 10 days	300 mg/d vs 150 mg/d	Dexamethasone, IV, 6 mg/day for up to 10 days	503	497	Multicentre Triple‐blind	Mainly high‐income countries (Denmark, India, Sweden, Switzerland) between August 2020 and May 2021	Median age (years, IQR) Low‐dose group: 64 (54 to 72) High‐dose group: 65 (56 to 74)	Low‐dose dexamethasone: 154 (32%) High‐dose dexamethasone: 151 (30%)	Moderate to severe ≥5
Maskin 2021	Dexamethasone + standard care, IV, 16 mg once daily for 5 days, followed by 8 mg administered once daily for additional 5 days	400 mg/d vs 150 mg/d	Dexamethasone + standard care, IV, 6 mg per day for 10 days	49	51	Multicentre Open‐label	Middle‐income country (Argentina) between June 2020 and March 2021	Mean age (years, SD) Low‐dose group: 60.04 (13.08) High‐dose group: 63.57 (13.59)	Low‐dose dexamethasone: 16 (33%) High‐dose dexamethasone: 13 (26%)	Severe ≥ 7
Toroghi 2021	Arm 2: Dexamethasone, IV, 8 mg twice a day for up to 10 days or until hospital discharge Arm 3: 8 mg 3 times a day for up to 10 days or until hospital discharge	Arm 2: 400 mg/d or Arm 3: 600 mg/d vs 200 mg/d	Dexamethasone, IV, 8 mg once a day for up to 10 days or until hospital discharge	96	48	Open‐label Single‐centre	Middle‐income country (Iran) between October 2020 and January 2021	Mean age (years, SD) Low‐dose group (8 mg once a day): 59 (14) Intermediate‐dose group (8 mg twice a day): 59 (17) High‐dose group (8 mg 3 times a day): 56 (16)	Low‐dose dexamethasone: 19 (40.4%) High‐dose dexamethasone: 34 (39.5%)	Moderate to severe ≥ 5
Taboada 2021	Dexamethasone, IV, 20 mg once daily for 5 days, followed by 10 mg once daily for additional 5 days	500 mg/d vs 150  mg/d	Dexamethasone, IV, 6 mg once daily for 10 days	98	102	Open‐label Single‐centre	High‐income country (Spain) between January and May 2021	Mean age (years, SD) Low‐dose group: 64.8 (14.1) High‐dose group: 63.9 (14.5)	Low‐dose dexamethasone: 41 (39.6%) High‐dose dexamethasone: 36 (36.7%)	Moderate to severe ≥ 5
IV: intravenous; d: day; SD: standard deviation; IQR: interquartile range

Outcome summary

All studies except Farahani 2021 reported utilisable dichotomous mortality data: for 7898 participants with follow‐up of up to 30 days and 485 participants with follow‐up between 31 and 120 days in the comparison of corticosteroids versus standard care (plus/minus placebo), 86 participants in the direct comparison of methylprednisolone and dexamethasone, and 1269 participants with follow‐up up to 30 days and 1399 participants with follow‐up between 31 and 120 days in the comparison of different corticosteroid doses. In the setting of acute COVID‐19 with immediate risk of death, we assumed in‐hospital mortality and all‐cause mortality with follow‐up of 14 to 30 days similar enough for meta‐analysis. For intermediate‐term survival we considered observation periods of 31 to 120 days equivalent and ready for meta‐analysis.

Other efficacy outcomes were reported heterogeneously.

The reporting of safety data was heterogeneous both in terms of whether reported at all (missing outcome data) and the underlying definitions. Including recent data requests for all studies, only four studies explicitly reported adverse events regardless of their nature and suspected relation to the intervention (Angus 2020; Edalatifard 2020; Tomazini 2020; Toroghi 2021). Another eight studies reported specific serious adverse events and adverse events related to the expected side effects of corticosteroids in both arms (Corral‐Gudino 2021; Dequin 2020; Jeronimo 2020; Maskin 2021; Munch 2021a; Munch 2021b; Taboada 2021; Tang 2021).

Apart from that, by far the largest study with 6425 participants reported safety outcomes only for the intervention arm as suspected drug reactions (Horby 2021), and three studies with 165 participants did not report safety outcomes at all (Farahani 2021; Jamaati 2021; Ranjbar 2021).

Beware that only dichotomous safety data (participants with at least one event/participants at risk) is presented narratively in the summary of findings tables, while additional discrete continuous data (number of events/participants at risk) is presented in an overview of safety data reporting with additional information (Table 8).

5. Reporting of safety data.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo)
Study	Definition as published	Way of counting	Intervention group	Control group	Study design
Angus 2020	Any SAE	SAE: participants with at least one event/participants at risk	Shock‐dependent hydrocortisone: SAE: 5/141   Fixed‐dose hydrocortisone: SAE: 4/137	SAE: 0/52     SAE: 1/49	Open‐label
Corral‐Gudino 2021	Microbiology‐proven infection and hyperglycaemia	AE, HAI: events/participants at risk   HAI, FI: participants with at least one event/participants at risk	AE: 14/35 HAI: 5/35 (events and participants) FI: 1/35	AE: 1/29 HAI: 1/29 (events and participants) FI: 0/29	Open‐label
Dequin 2020	Nosocomial infections until day 28 defined by need for antibiotics. No other SAEs/AEs.	AE, HAI: events/participants at risk   HAI: participants with at least one event/participants at risk	AE: 40/76 HAI: 40/76 (events), 28/76 (participants)	AE: 50/73 HAI: 50/73 (events), 30/73 (participants)	Double‐blind
Edalatifard 2020	All undesirable effects (adverse events)	HAI: events/participants at risk   AE: participants with at least one event/participants at risk	AE: 2/34 HAI: 1/34	AE: 2/28 HAI: 0/28	Single‐blind
Farahani 2021	Not reported	Not applicable	—	—	Double‐blind
Horby 2021	Suspected drug reactions reported	Not applicable	—	—	Open‐label
Jamaati 2021	Not reported	Not applicable	—	—	Open‐label
Jeronimo 2020	AE/SAE not explicitly reported Positive blood culture, need for insulin therapy, sepsis reported	Not applicable	—	—	Double‐blind
Munch 2021a	Serious adverse reaction reported	Not applicable	—	—	Triple‐blind
Tang 2021	Hyperglycaemia, ventilator‐associated pneumonia, stress ulcer, gastrointestinal haemorrhage	AE, HAI: participants with at least one event/participants at risk	AE: 5/43 HAI: 2/43	AE: 8/43 HAI: 1/43	Single‐blind
Tomazini 2020	Glycaemic control, nosocomial infection, other AEs	SAE, AE, HAI: participants with at least one event/participants at risk	SAE: 5/151 AE: 122/151 HAI: 30/151	SAE: 9/148 AE: 121/148 HAI: 39/148	Open‐label
Comparison 2: Methylprednisolone versus dexamethasone
Study	Definition as published	Way of counting	Methylprednisolone	Dexamethasone	Study design
Ranjbar 2021	Not reported	Not applicable	—	—	Triple‐blind
Comparison 3: High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d)
Study	Definition as published	Way of counting	High‐dose	Low‐dose	Study design
Maskin 2021	SAE: effects causing disability (defined as muscular weakness with a MRC scale < 48) or death   AE: hyperglycaemia (≥ 200 mg/dL), delirium (positive CAM‐ICU or requirement of neuroleptic drugs), hospital‐acquired infections (according to investigator judgement)   HAI: as defined by the attending physician   FI: candidemia (documented as Candida sp. in blood cultures), aspergillosis (documented as Aspergillus sp. in cultures)	SAE, FI, AE, HAI: participants with at least one event/participants at risk	SAE: 42/49 AE: 49/49 HAI: 34/49 FI: 3/49	SAE: 40/49 AE: 48/49 HAI: 38/49 FI: 3/49	Open‐label
Munch 2021b	Serious adverse reaction (i.e. new episodes of septic shock, invasive fungal infection, clinically important gastrointestinal bleeding, or anaphylactic reaction to dexamethasone)   HAI: either new episode of septic shock OR invasive fungal infection, as per the elements of the composite outcome of ≥ 1 serious adverse reaction at day 28   FI: not further defined	SAE, FI, HAI: participants with at least one event/participants at risk   HAI, FI: events/participants at risk	SAE: 70/497 HAI: 27/497 (events), 50/497 (participants) FI: 15/497	SAE: 85/485 HAI: 35/485 (events), 61/485 (participants) FI: 21/485	Triple‐blind
Taboada 2021	AE: nosocomial infection, insuline use for hyperglycaemia, thrombosis, death at day 28 and death at day 60	AE, HAI: events/participants at risk	AE: 58/98 HAI: 10/98	AE: 65/102 HAI: 10/102	Open‐label
Toroghi 2021	AE: not further defined HAI: secondary infections	AE, HAI: events/participants at risk	AE: 252/86 HAI: 5/86	AE: 122/47 HAI: 1/47	Open‐label
AE: adverse event; ICU: intensive care unit; FI: fungal infection; SAE: serious adverse event; HAI: hospital‐acquired infection; SAP: statistical analysis plan

Ongoing studies

We identified 42 ongoing RCTs with systemic application of steroids for acute COVID‐19 (details listed in Table 9), of which 25 were classified as 'recruiting' or 'ongoing' according to the study registrations. One was classified as 'active, not recruiting', for one study recruitment was completed, and one was 'temporarily halted'. Thirteen were classified as 'not yet recruiting' and one as 'not recruiting'. The majority of studies were conducted in high‐income countries (about 62%). For further characteristics of ongoing studies please see Table 9.

6. Characteristics of ongoing studies.

Study	Sponsor/developer	Design	Place of residence (World Bank 2021)	Population/disease severity	Setting	Drug	Route of administration	Number of participants	Status
ACTRN12621001200875	Royal Prince Alfred Hospital	RCT	High‐income country (Australia)	Exhibiting mild/moderate COVID‐19 symptoms (fever, respiratory tract symptoms, dyspnoea, headache, gastrointestinal symptoms), with no requirement for oxygen or hospitalisation	Outpatient	Dexamethasone	Oral	650	Recruiting
ACTRN12621001603808	Monash Health Hospital Victoria	RCT	High‐income country (Australia)	Women pregnant > 16 weeks with COVID‐19 hospitalised with an oxygen requirement	Inpatient	Prednisolone, dexamethasone	Oral	192	Not yet recruiting
ChiCTR2000029386	Chongqing Public Health Medical Center	RCT	Middle‐income country (China)	Severe	Inpatient	Methylprednisolone	IV	48	Recruiting
ChiCTR2000029656	Wuhan Pulmonary Hospital	RCT	Middle‐income country (China)	Severe	Inpatient	Methylprednisolone	IV	100	Not yet recruiting
ChiCTR2000030481	Zhongnan Hospital of Wuhan University	RCT	Middle‐income country (China)	Diagnosed COVID‐19 infection	Inpatient	Early corticosteroid intervention, middle‐late corticosteroid intervention	Unclear, most likely systemic	200	Recruiting
CTRI/2020/07/026608	Dr Ananthakumar PK, Chettinad Hospital and Research Institute Kelambakkam Kancheepuram Dist Pin 603103	RCT	Middle‐income country (India)	Diagnosed COVID‐infection + ARDS	Inpatient	Dexamethasone, methylprednisolone	IV	40	Not yet recruiting
CTRI/2020/12/029894	SRM Medical College Hospital and Research Centre	RCT	Middle‐income country (India)	SpO2 < 94% under room air and requiring supplemental oxygen for hypoxaemia, respiratory rate 24 to 30/min	Inpatient	Dexamethasone, methylprednisolone	IV	50	Not yet recruiting
CTRI/2020/12/030143	Maulana Azad Medical College and associated Lok Nayak Hospital	RCT	Middle‐income country (India)	Admitted to ICU within 14 days of onset of symptoms; receiving invasive or non‐invasive positive pressure ventilation or respiratory support through HFNC	Inpatient	Dexamethasone, methylprednisolone	IV	500	Not yet recruiting
CTRI/2021/05/033873	JK Hospital and LN Medical College	RCT	Middle‐income country (India)	Excluded: moderate to severe type of disease; eligible for home isolation	Outpatient	Methylprednisolone	Oral	500	Not yet recruiting
CTRI/2021/08/035822	Animesh Ray	RCT	Middle‐income country (India)	Severe COVID‐19 pneumonia (SpO2 < 94%; PaO2/FiO2 < 300 mm Hg or respiratory rate (RR) > 30 breaths/min) with lack of response to dexamethasone 6 mg after 48 hours (defined as similar or worsening oxygen requirement (margin of error is 5% Fio2 for high‐flow nasal cannula, 2 L/min for NRBM, and 1 L/min for low flow oxygen devices)	Inpatient	Dexamethasone high‐dose, dexamethasone low‐dose	IV or oral	120	Not yet recruiting
EUCTR2020‐001413‐20‐ES	Fundació Clínic per a la Recerca Biomèdica	RCT	High‐income country (Spain)	Non‐critical patient with pneumonia in radiological progression and/or patient with progressive respiratory failure in the last 24 to 48 h	Inpatient	Methylprednisolone	IV	100	Temporarily halted
EUCTR2020‐001457‐43‐FR	APHP	RCT	High‐income country (France)	Admitted to ICU	Inpatient	Dexamethasone	IV	550	Ongoing
EUCTR2020‐001622‐64‐ES	Dra Ana Pueyo Bastida	RCT	High‐income country (Spain)	Clinical diagnosis of pulmonary involvement (respiratory symptoms ± pathological auscultation ± O2 desaturation) + chest X‐ray with mild‐moderate or normal alterations	Outpatient	Prednisone	Oral	200	Ongoing
EUCTR2020‐001707‐16‐ES	Iis Biodonostia	RCT	High‐income country (Spain)	Bilateral pneumonia caused by SARS‐CoV‐2 without response to the treatment: defined as persistence of fever (above 37.5 ºC without other focus) and respiratory worsening (more dyspnoea, more cough, oxygen therapy at increasing doses, worsening of the degree of respiratory distress according to the PaO2/FiO2 ratio in categories 'mild, moderate or serious') or absence of improvement with respect to the previous state	Inpatient	Methylprednisolone	IV	60	Ongoing
EUCTR2020‐001921‐30	Azienda Ospedaliero‐Universitaria Policlinico di Modena	RCT	High‐income country (Italy)	Positive pressure ventilation (either non‐invasive or invasive) from > 24 h, IMV from < 96 h, PaO2/FiO2 ratio < 150	Inpatient	Methylprednisolone	IV	200	Ongoing
EUCTR2020‐003363‐25‐DK	Department of Intensive Care, Rigshospitalet	RCT	High‐income country (Denmark)	Severe, IMV/NIV	Inpatient	Dexamethasone (high dose and low dose)	IV	1000	Ongoing
EUCTR2020‐006054‐43‐IT	Università degli Studi di Trieste	RCT	High‐income country (Italy)	PaO2 ≤ 60 mmHg or SpO2 ≤ 90% or on HFNC, CPAP or NPPV at randomisation	Inpatient	Methylprednisolone, dexamethasone	IV	680	Ongoing
EUCTR2021‐001416‐29‐ES	Fundación para la Investigación e Innovación Biomédica (FIIB) del Hospital Universitario Infanta Leonor y Hospital Unive	RCT	High‐income country (Spain)	With need for oxygen therapy in NG = 1 lpm to maintain saturation = 94%	Inpatient	Dexamethasone high‐dose, dexamethasone low‐dose	IV	200	Ongoing
EUCTR2021‐004021‐71	University Medical Center Utrecht	RCT	High‐income country (Netherlands)	Recent COVID‐19 infection (< 3 months)	Outpatient	Prednisolone	Oral	116	Ongoing
IRCT20190606043826N2	Esfahan University of Medical Sciences	RCT	Middle‐income country (Iran)	Hospitalised patients with COVID‐19	Inpatient	Dexamethasone high‐dose, dexamethasone low‐dose	IV	60	Recruitment completed
NCT04329650	Judit Pich Martínez, Fundacion Clinic per a la Recerca Biomédica	RCT	High‐income country (Spain)	Non‐critical patient with pneumonia in radiological progression and/or patient with progressive respiratory failure in the last 24 to 48 h	Inpatient	Methylprednisolone	IV	200	Recruiting
NCT04344730	Assistance Publique ‐ Hôpitaux de Paris	RCT	High‐income country (France)	Admitted to ICU	Inpatient	Dexamethasone	IV	Actual enrolment 550	Not recruiting
NCT04345445	University of Malaya	RCT	Middle‐income country (Malaysia)	Excluded: receipt of mechanical ventilation	Inpatient	Methylprednisolone	IV	310	Not yet recruiting
NCT04377503	Hospital Sao Domingos	RCT	Middle‐income country (Brazil)	COVID diagnosis confirmed by real time PCR, PaO2/FIO2 < 200, laboratory: high sensitivity CRP > 5 mg/L; LDH > 245 U/L; ferritin > 300; D‐dimer > 1500; interleukin‐6 > 7.0 pg/mL	Inpatient	Methylprednisolone	Oral	40	Not yet recruiting
NCT04452565	NeuroActiva, Inc.	RCT	High‐income country (USA)	Excluded: IMV	Inpatient	Dexamethasone	Oral	525	Recruiting
NCT04499313	Chattogram General Hospital	RCT	Middle‐income country (Bangladesh)	Moderate to severe COVID‐19 infection	Inpatient	Dexamethasone, methylprednisolone	IV	60	Recruiting
NCT04509973	Scandinavian Critical Care Trials Group	RCT	High‐income country (Denmark)	IMV OR NIV or continuous use of CPAP for hypoxia OR oxygen supplementation with an oxygen flow of at least 10 L/min independent of delivery system	Inpatient	Dexamethasone	IV	1000	Active, not recruiting
NCT04513184	Edda Sciutto Conde	RCT	Middle‐income country (Mexico)	Hospitalised patients with moderate to severe respiratory complications that have not received mechanical ventilation	Inpatient	Dexamethasone	IV vs nasal	60	Recruiting
NCT04528329	ClinAmygate	RCT	Middle‐income country (Egypt)	Mild to moderate severity	Unclear	Dexamethasone	Unclear, most likely systemic	300	Recruiting
NCT04528888	Massimo Girardis, University of Modena and Reggio Emilia	RCT	High‐income country (Italy)	Included: positive pressure ventilation (IMV/NIV) for > 24 h, IMV from < 96 h, PaO2/FiO2 ratio < 150	Inpatient	Methylprednisolone	IV	210	Recruiting
NCT04545242	Dr. Negrin University Hospital	RCT	High‐income country (Spain)	Intubated and mechanically ventilated	Inpatient	Dexamethasone	IV	980	Not yet recruiting
NCT04636671	University of Trieste	RCT	High‐income country (Italy)	Excluded: on IMV Included: PaO2 ≤ 60 mmHg or SpO2 ≤ 90% or on HFNC, CPAP or NPPV at randomisation	Inpatient	Dexamethasone, methylprednisolone	IV	680	Recruiting
NCT04663555	Brno University Hospital	RCT	High‐income country (Czech Republic)	Intubation/mechanical ventilation or ongoing HFNC oxygen therapy; admission to ICU	Inpatient	Dexamethasone	IV	300	Recruiting
NCT04673162	Azienda Unità Sanitaria Locale Reggio Emilia	RCT	High‐income country (Italy)	Need for supplemental oxygen in any delivery mode with the exception of IMV	Inpatient	Methylprednisolone	IV	260	Not yet recruiting
NCT04707534	University of Oklahoma	RCT	High‐income country (USA)	Positive pressure ventilation (non‐invasive or invasive) or HFNC or need supplemental oxygen with oxygen mask or nasal cannula	Inpatient	Dexamethasone	Unclear, most likely systemic	300	Recruiting
NCT04765371	Centre Hospitalier René Dubos	RCT	High‐income country (France)	Patient with SpaO2 ≤ 94 % in room air (90% for patient with respiratory failure) and requiring an oxygen therapy	Inpatient	Dexamethasone, prednisolone	Unclear, most likely systemic	220	Recruiting
NCT04780581	Fundación Instituto de Estudios de Ciencias de la Salud de Castilla y León	RCT	High‐income country (Spain)	Requires supplementary oxygen due to basal saturation ≤ 93% (with ambient O2, 21%), excluded if IMV, NIV, HFNC	Inpatient	Dexamethasone, methylprednisolone	Unclear, most likely systemic	290	Recruiting
NCT04795583	University of Alberta	RCT	Middle‐income country (Colombia, Mexico)	Ambulatory, confirmed SARS‐CoV‐2. Clinical symptoms compatible with COVID‐19 for ≤ 14 days before randomisation. Oxygen saturation ≥ 95%	Outpatient	Prednisone	Oral	1526	Not yet recruiting
NCT04834375	Northwell Health	RCT	High‐income country (USA)	Hypoxaemia defined by an oxygen saturation < 94% or the need for supplemental oxygen	Inpatient	Dexamethasone	IV	142	Recruiting
NCT04836780	Hospital Universitario Infanta Leonor	RCT	High‐income country (Spain)	Peripheral capillary oxygen saturation (SpO2) ≥ 94% and < 22 breaths per minute (bpm) breathing room air. High risk of developing ARDS	Inpatient	Dexamethasone	IV	126	Recruiting
NCT04860518	Faron Pharmaceuticals Ltd	RCT	High‐income country (USA)	Admitted to hospital with respiratory symptoms of COVID‐19 requiring hospital care and oxygen supplementation (≤ 8L/min)	Inpatient	Dexamethasone and human intravenous interferon beta‐Ia	IV	140	Recruiting
TCTR20211017001	Faculty of Medicine Ramathibodi Hospital	RCT	Middle‐income country (Thailand)	Resting oxygen saturation between 90% and 94%	Inpatient	Methylprednisolone, dexamethasone	IV	120	Not yet recruiting
ARDS: acute respiratory distress syndrome; CPAP: continuous positive airway pressure; CRP: c‐reactive protein; FiO2: fraction of inspired oxygen; HFNC: high‐flow nasal cannula; ICU: intensive care unit; IMV: invasive mechanical ventilation; IV: intravenous; LDH: lactic dehydrogenase; NIV: non‐invasive ventilation; NPPV: non‐invasive positive pressure ventilation; PaO2: partial pressure oxygen; PCR: polymerase chain reaction RCT: randomised; controlled trial; RT‐PCR: reverse transcription polymerase chain reaction; SpO2: blood oxygen saturation

On excluding the studies that were 'not recruiting', 'not yet recruiting', and 'active, not recruiting' the 27 studies that were recruiting, ongoing, temporarily halted and that completed recruitment comprised a total of 7617 expected participants. Most of the potentially eligible ongoing studies identified intend to recruit people who are admitted to hospital and require varying levels of respiratory support. Of the 42 ongoing studies, 15 planned to test dexamethasone, 11 methylprednisolone, and three prednisolone or prednisone. Four studies planned to compare different dexamethasone dosing regimens. Eight studies planned to compare dexamethasone to methylprednisolone, and one study dexamethasone to prednisolone, another dexamethasone to human intravenous interferon beta‐Ia. One study planned to compare corticosteroids at different time points.

Studies awaiting classification

We identified 23 RCTs with systemic application of steroids for acute COVID‐19 (details listed in Table 10). Of the 23 studies comprising 5608 expected participants, 15 were classified as ‘completed’, four as ‘prematurely ended’, and four as 'terminated' (due to lack of enrolment, too few participants, local recommendations or corticosteroid use approval), according to the study registrations. For four studies, the full‐text publications are available, but the methodology is unclear. For one study only the abstract is available, which does not provide any information on the randomisation process. The majority of studies were conducted in low‐ and middle‐income countries (about 56%; for details see Table 10).

7. Characteristics of studies awaiting classification.

Study	Sponsor/developer	Design	Place of residence (World Bank 2021)	Population/disease severity	Setting	Drug	Route of administration	Number of participants	Status
EUCTR2020‐001333‐13‐FR	Groupe Hospitalier Paris Saint‐Joseph	RCT	High‐income country (France)	Included: patient diagnosed COVID positive by RT‐PCR and/or scanner (patients admitted with already mechanical ventilation and sedation, or with acute respiratory failure evolving very quickly)	Inpatient	Dexamethasone	IV	122	Prematurely ended
EUCTR2020‐001307‐16‐ES	Fundación para la Investigación Biomédica Hospital Ramón y Cajal	RCT	High‐income country (Spain)	ARDS	Inpatient	Methylprednisolone	IV	104	Prematurely ended
EUCTR2020‐001553‐48‐FR	Hospices Civils de Lyon	RCT	High‐income country (France)	Peripheral saturation by pulse oximeter SpO2 ≤ 94% in ambient air measured twice at 5‐ to 15‐min intervals, or PaO2/FiO2 < 300 mmHg	Inpatient	Prednisone	Oral	304	Prematurely ended
IRCT20081027001411N3	Teheran University of Medical Sciences	RCT	Middle‐income country (Iran)	Blood oxygen saturation < 93%; with ARDS	Inpatient	Prednisolone	Not stated	60	Completed
IRCT20120215009014N354	Hamedan University of Medical Sciences	RCT	Middle‐income country (Iran)	Hospitalised in ICU, bilateral pulmonary infiltration in chest X‐ray or CT‐scan; respiratory distress with > 24 breaths per minute	Inpatient	Hydrocortisone, methylprednisolone, dexamethasone	IV	81	Completed
IRCT20160118026097N4	Ghoum University of Medical Sciences	RCT	Middle‐income country (Iran)	Hypoxia requires supplemental oxygen to maintain oxygen saturation > 90%	Inpatient	Dexamethasone	Not stated	64	Completed
IRCT20200611047727N3	Shahid Beheshti University of Medical Sciences	RCT	Middle‐income country (Iran)	Oxygen saturation level < 93	Inpatient	Methylprednisolone	IV	60	Completed
IRCT20201015049030N1	Teheran University of Medical Sciences	RCT	Middle‐income country (Iran)	Blood oxygen saturation between 90% and 95%	Outpatient	Dexamethasone	Not stated	200	Completed
ISRCTN33037282	Clínica Medellín ‐ Grupo Quirónsalud	RCT	Middle‐income country (Colombia)	PaO2 ≤ 60 mmHg or SpO2 ≤ 90% or on HFNC, CPAP or NPPV at randomisation Excluded: on IMV	Inpatient	Methylprednisolone, dexamethasone	IV	680	Completed
NCT04244591	Peking Union Medical College Hospital	RCT	Middle‐income country (China)	PaO2/FiO2 < 200 mmHg; positive pressure ventilation (non‐invasive or invasive) or HFNC > 45 L/min for < 48 h; requiring ICU admission	Inpatient	Methylprednisolone	Not stated	80	Completed
NCT04325061	Dr. Negrin University Hospital	RCT	High‐income country (Spain)	Intubated and mechanically ventilated	Inpatient	Dexamethasone	IV	19	Terminated (lack of enrolment)
NCT04530409	ClinAmygate	RCT	Middle‐income country (Egypt)	Mild and moderate severity	Unclear	Dexamethasone	Unclear, most likely systemic	450	Completed
NCT04746430	General Practitioners Research Institute	RCT	High‐income country (Netherlands)	Exercise‐induced desaturation, defined as SpO2 < 92% (< 90% for COPD patients) and/or an absolute drop of ≥ 4% in SpO2 after a 1‐min sit‐to‐stand test or SpO2 < 92% (< 90% for COPD patients) at rest with GP's and patient's shared decision to keep patient at home despite this in itself being an indication for referral to hospital	Outpatient	Dexamethasone	Unclear, most likely systemic	2000	Terminated (too few patients)
EUCTR2020‐002186‐34‐ES	Fundació Hospital Universitari Vall d'Hebron ‐ Institut de Recerca (VHIR)	RCT	High‐income country (Spain)	Air oxygen saturation > 90 and < 94%; PaO2/FiO2 > 200 and ≤ 300 mmHg; Sa:FiO2 (O2 saturation measured with pulse oximeter/inspired O2 fraction) ≤ 350	Inpatient	Methylprednisolone	IV	100	Prematurely ended
EUCTR2020‐004323‐16	Azienda Ospedaliera Arcispedale Santa Maria Nuova/IRCCS di Reggio Emilia	RCT	High‐income country (Italy)	Need for supplemental oxygen in any delivery mode with the exception of IMV	Inpatient	Methylprednisolone	IV	260	Completed
NCT04347980	Centre Chirurgical Marie Lannelongue	RCT	High‐income country (France)	Admitted to ICU	Inpatient	Dexamethasone	IV	122	Terminated (ANSM (Agence nationale de sécurité du médicament et des produits de santé) Recommendation)
NCT04438980	Fundacion Miguel Servet	RCT	High‐income country (Spain)	Hospitalised Excluded: SpO2 < 90% (in air ambient) or PaO2 < 60 mmHg (in ambient air) or PaO2/FiO2 < 300 mmHg	Inpatient	Methylprednisolone	IV	72	Completed
NCT04451174	University of Chile	RCT	High‐income country (Chile)	Excluded: requirements of mechanical ventilation (IMV/NIV) Included: oxygen requirements until 35 % by Venturi mask or 5 L/min by nasal cannula	Inpatient	Prednisone	IV	184	Terminated (corticosteroid use approval)
Salukhov 2021	Not stated	RCT (unclear randomisation process)	Middle‐income country (Russia)	Moderate to severe course with process prevalence according to computed tomography without hypoxaemia (saturation SpO2 > 93%) with a duration of hyperthermia > 38 °C for 3 days or more and C‐reactive protein levels of 15 to 50 mg/L	Inpatient	Methylprednisolone	oral	40	Completed and published (but unclear randomisation process)
Ghanei 2021	Not stated	RCT	Middle‐income country (Iran)	Oxygen saturation (Spo2) less than 94%	Inpatient	Prednisolone	Not stated	236	Completed and published (not pre‐registered, no ethics vote named, no patient informed consent)
Gautam 2021	None	RCT	Middle‐income country (India)	Moderate to severe COVID‐19, oxygen saturation < 93% on room air	Inpatient	Methylprednisolone, dexamethasone	IV	140	Completed and published (not pre‐registered, no ethics vote named, no patient informed consent, no information about randomisation)
Montalvan 2021	Not stated	RCT (unclear randomisation process)	Middle‐income country (Honduras)	Not stated	Inpatient	Dexamethasone	Not stated	81	Completed and abstract published (no registration number, no information about randomisation process), only abstract available
Rashad 2021	South Valley University	RCT (unclear randomisation process)	Middle‐income country (Egypt)	Moderate to severe ≥ 5	Inpatient	Dexamethasone	IV	149	Completed and published (no pre‐registration, unclear randomisation process)
ARDS: acute respiratory distress syndrome; COPD: chronic obstructive pulmonary disease; CT: computed tomography; HFNC: high‐flow nasal cannula;ICU: intensive care unit; RCT: randomised controlled trial; RT‐PCR: reverse transcription polymerase chain reaction.

Only one of the studies awaiting classification was placebo‐controlled and compared methylprednisolone in addition to standard care to standard care plus placebo. Four studies planned to compare dexamethasone, four methylprednisolone, and three prednisolone or prednisone (each in addition to standard care) to standard care only. One open‐label study planned to compare methylprednisolone without specification of the control. Two studies planned to compare dexamethasone to methylprednisolone, one study methylprednisolone to prednisolone, and one study hydrocortisone to methylprednisolone to dexamethasone. Two other studies planned to compare dexamethasone plus hydroxychloroquine to hydroxychloroquine only and one prednisolone plus hydroxychloroquine plus azithromycin plus naproxen to hydroxychloroquine plus azithromycin plus naproxen. One study planned to compare two different treatment regimes: an early versus a late treatment with dexamethasone. One study planned to compare different doses of dexamethasone. Another study compared dexamethasone to tocilizumab.

Excluded studies

We excluded 28 studies (31 references) that did not meet our inclusion criteria.

• Nine studies investigated corticosteroids plus other active substances versus standard care (EUCTR2020‐001445‐39‐ES; IRCT20120225009124N4; IRCT20190312043030N2; NCT04341038; NCT04411667; NCT04468646; NCT04561180; NCT04640168; NCT04826822).

• Six studies examined inhaled corticosteroids (EUCTR2020‐001616‐18‐ES; EUCTR2020‐001889‐10; ISRCTN86534580; NCT04355637; NCT04381364; NCT04416399).

• Four studies investigated topical corticosteroids (IRCT20200522047542N1; NCT04361474; NCT04484493; NCT04569825).

• Three studies considered corticosteroids for long‐COVID treatment (NCT04551781; NCT04534478; NCT04657484).

• One study stopped early before enrolling its first participant (NCT04359511).

• One study was withdrawn due to a recent study suggesting a new corticosteroid regime for intensive care unit patients, so that no participants were recruited (NCT05133635).

• In one study the administered antibody (foralumab) was not standard care (Moreira 2021).

• In one study the participants received low‐dose dexamethasone in addition to standard care (Naik 2021).

• One study compared bio‐marker adjusted corticosteroid dosing with usual care (corticosteroid use and dosing determined by the physician) (Odeyemi 2021).

• For one study it was not possible to perform the trial due to the availability and logistics of porcine heparin (NCT04485429).

Risk of bias in included studies

We assessed the risk of bias of results from 15 RCTs that contributed to our analyses (Angus 2020; Corral‐Gudino 2021; Dequin 2020; Edalatifard 2020; Horby 2021; Jamaati 2021; Jeronimo 2020; Maskin 2021; Munch 2021a; Munch 2021b; Ranjbar 2021;Taboada 2021; Tang 2021; Tomazini 2020; Toroghi 2021), using the RoB 2 tool (version 22 August 2019) recommended in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022b). We made no assessment of bias for Farahani 2021 as it did not report any outcomes relevant to this review.

The 15 studies contributed 47 study results to eight outcomes in the respective summary of findings tables and three more in additional analyses for inpatient individuals. The completed RoB 2 tool with responses to all assessed signalling questions is available online at: https://zenodo.org/record/6500842.

Overall judgements for studies that included individuals with a confirmed diagnosis of moderate to severe COVID‐19

Overall risk of bias by study

From 47 study results, we rated 5 (11%) at low risk of bias, noted some concerns in 20 (43%) study results, and rated 22 (47%) at high risk of bias. Regarding the respective studies we included the following:

• Angus 2020 ‐ two study results with some concerns and one with high risk.

• Corral‐Gudino 2021 ‐ one study result with some concerns and two with high risk.

• Dequin 2020 ‐ one study result with some concerns and one with high risk.

• Edalatifard 2020 ‐ two study results with some concerns and one with high risk.

• Horby 2021 ‐ three study results with some concerns and one with high risk.

• Jamaati 2021 ‐ one study result with some concerns.

• Jeronimo 2020 ‐ two study results with some concerns and three with high risk.

• Maskin 2021 ‐ two study results with some concerns and four with high risk.

• Munch 2021a ‐ two study results with low risk.

• Munch 2021b ‐ two study results with low risk and three with high risk.

• Ranjbar 2021 ‐ one study result with some concerns.

• Taboada 2021 ‐ three study results with some concerns.

• Tang 2021 ‐ one study result with low risk and two with high risk.

• Tomazini 2020 ‐ two study results with some concerns and three with high risk.

• Toroghi 2021 ‐ one study result with high risk.

Overall risk of bias by outcome

Systemic corticosteroids (plus standard care) versus standard care (plus/minus placebo)

We had some concerns about all‐cause‐mortality up to 30 days (Table 84) and all‐cause‐mortality up to 120 days (Table 85) because of issues with randomisation, deviations from intended interventions, and pre‐specification of the outcomes. We also applied the respective risk of bias assessments to the six subgroup analyses (Table 95; Table 96; Table 97; Table 98; Table 99; Table 100; Table 101; Table 102). We also had some concerns about clinical improvement (discharged alive) because of issues with randomisation, deviations from intended intervention, assessment, and pre‐specification of the outcome (Table 86). There were some concerns about clinical worsening (new need for invasive mechanical ventilation (IMV) or death) because of deviations from intended interventions and assessment of the outcome only (Table 87). We judged all safety outcomes (serious adverse events, adverse events, hospital‐acquired infections, and invasive fungal infections), need for dialysis, and viral clearance to be at high risk of bias mainly due to missing adjustment for competing risk of death (Table 88; Table 89; Table 90; Table 91; Table 93; Table 94).

Risk of bias for analysis 1.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

15% of the participants in the no hydrocortisone group received hydrocortisone

Low risk of bias

Data were available for this outcome. 238 participants were randomised and 238 were analysed.

Low risk of bias

Mortality is an observer‐reported outcome not involving judgement.

Low risk of bias

Protocol and SAP are available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

15% of the participants in the no hydrocortisone group received hydrocortisone

Low risk of bias

Data were available for this outcome. 238 participants were randomised and 238 were analysed.

Low risk of bias

Mortality is an observer‐reported outcome not involving judgement.

Low risk of bias

Protocol and SAP are available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

Open‐label design (participants and clinicians were aware of the assigned treatment). ITT data presented.

Low risk of bias

No substantial data missing.

Low risk of bias

Both groups were measured at the same time and the measurements were similar between groups.

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results and missing information about the allocation concealment.

Dequin 2020

Low risk of bias

Randomisation was centralised and performed electronically. Allocation sequences were generated in a 1:1 ratio by a computer‐generated random number using a blocking schema. There were no baseline differences between the groups.

Low risk of bias

Protocol: Patients, investigators and care providers will be blinded for the patient‐arm. ITT was used.

Low risk of bias

Data were available for this outcome.

Low risk of bias

The measurements were similar between groups.

Some concerns

The protocol and statistical analysis plan were available. 21‐day mortality was a post hoc outcome.

Some concerns

Overall judged some concerns due to selection of reported results.

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone.

Low risk of bias

6425 participants were randomised and 6425 were analysed.

Low risk of bias

Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Jamaati 2021

Some concerns

The selected patients were allocated to either the dexamethasone group or the control group by block randomisation. Ten blocks were generated by the Online Randomiser website. Pulmonary disease was highly significantly more prevalent in the control group. Additionally, baseline level of respiratory support was not reported so that we can only assume the most likely circumstance, i.e. that no support beyond oxygen insufflation had been given until randomisation.

Low risk of bias

No deviations mentioned. ITT data reported.

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups.

Some concerns

Trial registered after recruitment stop. No analysis plan published beyond registry entry.

Some concerns

Potential bias through probably problematic block randomisation with significant baseline imbalance for pulmonary disease, very delayed registration, and potentially not pre‐specified stop for futility raise some concerns. The issues were not discussed. Potential mild bias towards experimental.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

416 participants randomised and 416 participants analysed.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

Some concerns

Overall judged some concerns due to protocol deviations.

Munch 2021a

Low risk of bias

No issues with randomisation process.

Low risk of bias

Withdrawals with consecutive open‐label steroids may have affected the outcome, but withdrawals and non‐compliance was rather balanced after all.

Low risk of bias

No missing data.

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with the reported result.

Low risk of bias

Withdrawals because of the experimental context lead to some concerns (but it is less than 10% of all patients).

Tang 2021

Low risk of bias

Randomisation was stratified by the statistician of the leading site, who produced computer‐generated block randomisation lists with a block size of 4 patients.

Low risk of bias

Single‐blind design (participants), ITT data presented. No deviations reported.

Low risk of bias

No missing data.

Low risk of bias

The data collection and end point judgement were blinded, and the statisticians were also blinded during the statistical analysis. The measurements were similar between groups

Low risk of bias

30‐day mortality as pre‐specified in the trial registration was not explicitly reported, but could reliably derive from a synopsis of full text, figures, and supplemental material.

Low risk of bias

All domains were rated low risk of bias.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

No missing data.

Low risk of bias

Individuals who assessed the outcomes were not blinded for the assigned treatment. Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall judged some concerns because of protocol deviations

Risk of bias for analysis 1.2 All‐cause mortality up to 120 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Edalatifard 2020

Some concerns

No information about the allocation concealment.

Some concerns

6 patients in the control group received the intervention drug and were excluded from the analyses (17%).

Low risk of bias

The data were requested from the authors because the follow‐up time was not clearly visible from the publication.

Low risk of bias

The measurements were similar between groups.

Some concerns

Neither the protocol nor the SAP were available

Some concerns

Overall judged some concerns due to the randomisation process, protocol deviations and the selection of the reported results.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

416 participants randomised and 416 participants analysed.

Low risk of bias

The measurements were similar between groups.

Some concerns

Protocol and statistical plan available. Outcome was not pre‐specified.

Some concerns

Overall judged some concerns due to protocol deviations and selective reporting.

Munch 2021a

Low risk of bias

No issues with randomisation process.

Low risk of bias

Withdrawals with consecutive open‐label steroids may have affected the outcome, but withdrawals and non‐compliance was rather balanced after all.

Low risk of bias

No missing data.

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with the reported result.

Low risk of bias

Withdrawals because of the experimental context lead to some concerns (below 10% of all patients).

Risk of bias for analysis 2.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 2.1.1 No oxygen

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

6425 randomised, 6425 analysed

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Subgroup 2.1.2 Low‐flow oxygen only

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

No protocol deviations. ITT data presented.

Low risk of bias

We requested the data from the authors, as the data was not broken down by our subgroups of interest in the publication

Low risk of bias

The measurements were similar between groups

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

416 participants randomised and 416 participants analysed.

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

Some concerns

Overall judged some concerns due to protocol deviations.

Subgroup 2.1.3 NIV, high‐flow, and low‐flow oxygen combined

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

6425 randomised, 6425 analysed

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Subgroup 2.1.4 NIV or high‐flow oxygen only

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

No protocol deviations. ITT data presented.

Low risk of bias

We requested the data from the authors, as the data was not broken down by our subgroups of interest in the publication

Low risk of bias

The measurements were similar between groups

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

416 participants randomised and 416 participants analysed.

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

Some concerns

Overall judged some concerns due to protocol deviations.

Subgroup 2.1.5 Invasive ventilation

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

6425 randomised, 6425 analysed

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

416 participants randomised and 416 participants analysed.

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

Some concerns

Overall judged some concerns due to protocol deviations.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall judged some concerns because of protocol deviations.

Risk of bias for analysis 3.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 3.1.1 Dexamethasone

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

6425 randomised, 6425 analysed

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available

Some concerns

Overall judged some concerns due to protocol deviations.

Jamaati 2021

Some concerns

The selected patients were allocated to either the dexamethasone group or the control group by block randomisation. Ten blocks were generated by the Online Randomiser website. Pulmonary disease was highly significantly more prevalent in the control group. Additionally, baseline level of respiratory support was not reported so that we can only assume the most likely circumstance, i.e. that no support beyond oxygen insufflation had been given until randomisation.

Low risk of bias

No deviations mentioned. ITT data reported.

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups.

Some concerns

Trial registered after recruitment stop. No analysis plan published beyond registry entry.

Some concerns

Overall judged some concerns due to selective reporting and baseline differences.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall judged some concerns due to protocol deviations.

Subgroup 3.1.2 Methylprednisolone

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

Open‐label study. ITT data presented.

Low risk of bias

No substantial data missing.

Low risk of bias

The measurements were similar between groups

Some concerns

Endpoint was registered in Eudra‐CT as in‐hospital mortality but reported as 28d‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selective reporting.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

416 participants randomised and 416 participants analysed.

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

Some concerns

Overall judged some concerns due to protocol deviations.

Tang 2021

Low risk of bias

Randomisation was stratified by the statistician of the leading site, who produced computer‐generated block randomisation lists with a block size of 4 patients.

Low risk of bias

ITT data presented. No deviations reported.

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

30‐day mortality as pre‐specified in the trial registration could was not explicitly reported, but could reliably derive from a synopsis of full text, figures, and supplemental material.

Low risk of bias

All domains were rated low.

Subgroup 3.1.3 Hydrocortisone

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

15% of the participants in the no hydrocortisone group received corticosteroids

Low risk of bias

No missing data.

Low risk of bias

Mortality is an observer‐reported outcome not involving judgement.

Low risk of bias

Protocol and SAP available

Some concerns

Overall judged some concerns due to protocol deviations

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

15% of the participants in the no hydrocortisone group received corticosteroids

Low risk of bias

No missing data.

Low risk of bias

Mortality is an observer‐reported outcome not involving judgement.

Low risk of bias

Protocol and SAP available

Some concerns

Overall judged some concerns due to protocol deviations

Dequin 2020

Low risk of bias

Randomisation was centralised and performed electronically. Allocation sequences were generated in a 1:1 ratio by a computer‐generated random number using a blocking schema. There were no baseline differences between the groups.

Low risk of bias

Protocol: Patients, investigators and care providers will be blinded for the patient‐arm ITT was used

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups.

Some concerns

The protocol and statistical analysis plan were available. 21‐day mortality was a post hoc outcome

Some concerns

Overall judged some concerns due to selective reporting.

Munch 2021a

Low risk of bias

No issues with randomisation process.

Low risk of bias

Withdrawals with consecutive open‐label steroids may have affected the outcome, but withdrawals and non‐compliance was rather balanced after all.

Low risk of bias

No missing data.

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with the reported result.

Low risk of bias

Withdrawals because of the experimental context lead to some concerns (below 10% of all patients).

Risk of bias for analysis 3.2 All‐cause mortality up to 120 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 3.2.1 Hydrocortisone

Munch 2021a

Low risk of bias

No issues with randomisation process.

Low risk of bias

Withdrawals with consecutive open‐label steroids may have affected the outcome, but withdrawals and non‐compliance was rather balanced after all.

Low risk of bias

No missing data

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with the reported result.

Low risk of bias

Withdrawals because of the experimental context lead to some concerns (below 10% of all patients).

Subgroup 3.2.2 Methylprednisolone

Edalatifard 2020

Some concerns

No information about the allocation concealment.

Some concerns

6 patients in the control group received the intervention drug and were excluded from the analyses (17%).

Low risk of bias

The data were requested from the authors because the follow‐up time was not clearly visible from the publication.

Low risk of bias

The measurements were similar between groups.

Some concerns

Neither the protocol nor the SAP were available

Some concerns

Overall judged some concerns due to the randomisation process, protocol deviations and the selection of the reported results.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment ; Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

Data is available for all participants.

Low risk of bias

The measurements were similar between groups,

Some concerns

No pre‐specification of the outcome in the protocol.

Some concerns

Overall we judged some concern due to protocol deviations and selective reporting

Risk of bias for analysis 3.3 Clinical improvement: discharged alive.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 3.3.1 Dexamethasone

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

Data for this outcome was available for all participants randomised.

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.

Low risk of bias

No issue with selective reporting.

Some concerns

Overall judged some concerns due to the randomisation process, protocol deviations and limited bias in measurement.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

Data for this outcome was available for all participants randomised

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall rated some concerns due to protocol deviations and measurement of the outcome.

Subgroup 3.3.2 Methylprednisolone

Edalatifard 2020

Some concerns

No information about the allocation concealment.

Some concerns

In this study, patients did not know which group of them used medicine Physicians and clinicians team know about the medicine and intervention groups. 6 patients in the control group received the intervention drug and were excluded from the analyses Intention‐to‐treat

Low risk of bias

Data for this outcome was available for all participants randomised

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.No protocol or SAP available.

Some concerns

Neither the protocol nor the SAP were available

Some concerns

Overall judged some concerns due to missing information about the allocation concealment, deviations from the intended interventions, measurement of the outcome and selection of the reported result

Risk of bias for analysis 4.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 4.1.1 Female

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

Open‐label design (participants and clinicians were aware of the assigned treatment). ITT data presented.

Low risk of bias

No substantial data missing.

Low risk of bias

Both groups were measured at the same time and the measurements were similar between groups.

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results and missing information about the allocation concealment.

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone.

Low risk of bias

6425 participants were randomised and 6425 were analysed.

Low risk of bias

Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

No missing data. We requested the data from the authors.

Low risk of bias

Individuals who assessed the outcomes were not blinded for the assigned treatment. Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall judged some concerns because of protocol deviations

Subgroup 4.1.2 Male

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

Open‐label design (participants and clinicians were aware of the assigned treatment). ITT data presented.

Low risk of bias

No substantial data missing.

Low risk of bias

Both groups were measured at the same time and the measurements were similar between groups.

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results and missing information about the allocation concealment.

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone.

Low risk of bias

6425 participants were randomised and 6425 were analysed.

Low risk of bias

Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

No missing data. We requested the data from the authors.

Low risk of bias

Individuals who assessed the outcomes were not blinded for the assigned treatment. Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall judged some concerns because of protocol deviations

Risk of bias for analysis 5.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 5.1.1 < 70 years

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

Open‐label design (participants and clinicians were aware of the assigned treatment). ITT data presented.

Low risk of bias

We requested data from the authors. No substantial data missing.

Low risk of bias

Both groups were measured at the same time and the measurements were similar between groups.

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results and missing information about the allocation concealment.

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone.

Low risk of bias

6425 participants were randomised and 6425 were analysed.

Low risk of bias

Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

No missing data. We requested the data from the authors.

Low risk of bias

Individuals who assessed the outcomes were not blinded for the assigned treatment. Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall judged some concerns because of protocol deviations

Subgroup 5.1.2 ≥ 70 years

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

Open‐label design (participants and clinicians were aware of the assigned treatment). ITT data presented.

Low risk of bias

We requested data from the authors. No substantial data missing.

Low risk of bias

Both groups were measured at the same time and the measurements were similar between groups.

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results and missing information about the allocation concealment.

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone.

Low risk of bias

6425 participants were randomised and 6425 were analysed.

Low risk of bias

Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

No missing data. We requested the data from the authors.

Low risk of bias

Individuals who assessed the outcomes were not blinded for the assigned treatment. Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall judged some concerns because of protocol deviations

Risk of bias for analysis 6.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 6.1.1 White

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

Open‐label design (participants and clinicians were aware of the assigned treatment). ITT data presented.

Low risk of bias

No substantial data missing.

Low risk of bias

Both groups were measured at the same time and the measurements were similar between groups.

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results and missing information about the allocation concealment.

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone.

Low risk of bias

6425 participants were randomised and 6425 were analysed.

Low risk of bias

Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Subgroup 6.1.2 Black, Asian or minority ethnic group

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone.

Low risk of bias

6425 participants were randomised and 6425 were analysed.

Low risk of bias

Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Subgroup 6.1.3 Unknown

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone.

Low risk of bias

6425 participants were randomised and 6425 were analysed.

Low risk of bias

Both groups were measured at the same time. The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Risk of bias for analysis 7.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 7.1.1 High‐income countries

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

15% of the participants in the no hydrocortisone group received hydrocortisone

Low risk of bias

Data were available for this outcome. 238 participants were randomised and 238 were analysed.

Low risk of bias

Mortality is an observer‐reported outcome not involving judgement.

Low risk of bias

Protocol and SAP are available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

15% of the participants in the no hydrocortisone group received hydrocortisone

Low risk of bias

Data were available for this outcome. 238 participants were randomised and 238 were analysed.

Low risk of bias

Mortality is an observer‐reported outcome not involving judgement.

Low risk of bias

Protocol and SAP are available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

No protocol deviations. ITT data presented.

Low risk of bias

We requested the data from the authors, as the data was not broken down by our subgroups of interest in the publication

Low risk of bias

The measurements were similar between groups

Some concerns

Outcome was registered in Eudra‐CT as in‐hospital mortality but reported as 28‐day‐mortality in the supplemental material.

Some concerns

Overall judged some concerns due to selection of reported results.

Dequin 2020

Low risk of bias

Randomisation was centralised and performed electronically. Allocation sequences were generated in a 1:1 ratio by a computer‐generated random number using a blocking schema. There were no baseline differences between the groups.

Low risk of bias

Protocol: Patients, investigators and care providers will be blinded for the patient‐arm. ITT was used.

Low risk of bias

Data were available for this outcome.

Low risk of bias

The measurements were similar between groups.

Some concerns

The protocol and statistical analysis plan were available. 21‐day mortality was a post hoc outcome.

Some concerns

Overall judged some concerns due to selection of reported results.

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

6425 randomised, 6425 analysed

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available.

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group.

Munch 2021a

Low risk of bias

No issues with randomisation process.

Low risk of bias

Withdrawals with consecutive open‐label steroids may have affected the outcome, but withdrawals and non‐compliance was rather balanced after all.

Low risk of bias

No missing data.

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with the reported result.

Low risk of bias

Withdrawals because of the experimental context lead to some concerns (but it is less than 10% of all patients).

Subgroup 7.1.2 Low‐ and middle‐income countries

Jamaati 2021

Some concerns

The selected patients were allocated to either the dexamethasone group or the control group by block randomisation. Ten blocks were generated by the Online Randomiser website. Pulmonary disease was highly significantly more prevalent in the control group. Additionally, baseline level of respiratory support was not reported so that we can only assume the most likely circumstance, i.e. that no support beyond oxygen insufflation had been given until randomisation.

Low risk of bias

No deviations mentioned. ITT data reported.

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups.

Some concerns

Trial registered after recruitment stop. No analysis plan published beyond registry entry.

Some concerns

Potential bias through probably problematic block randomisation with significant baseline imbalance for pulmonary disease, very delayed registration, and potentially not pre‐specified stop for futility raise some concerns. The issues were not discussed. Potential mild bias towards experimental.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

416 participants randomised and 416 participants analysed.

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

Some concerns

Overall judged some concerns due to protocol deviations.

Tang 2021

Low risk of bias

Randomisation was stratified by the statistician of the leading site, who produced computer‐generated block randomisation lists with a block size of 4 patients.

Low risk of bias

Single‐blind design (participants), ITT data presented. No deviations reported.

Low risk of bias

No missing data.

Low risk of bias

The data collection and end point judgement were blinded, and the statisticians were also blinded during the statistical analysis. The measurements were similar between groups

Low risk of bias

30‐day mortality as pre‐specified in the trial registration was not explicitly reported, but could reliably derive from a synopsis of full text, figures, and supplemental material.

Low risk of bias

All domains were rated low risk of bias.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall judged some concerns because of protocol deviations.

Risk of bias for analysis 1.3 Clinical improvement: discharged alive.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Edalatifard 2020

Some concerns

No information about the allocation concealment.

Some concerns

In this study, patients did not know which group of them used medicine Physicians and clinicians team know about the medicine and intervention groups. 6 patients in the control group received the intervention drug and were excluded from the analyses Intention‐to‐treat

Low risk of bias

Data for this outcome was available for all participants randomised

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.No protocol or SAP available.

Some concerns

Neither the protocol nor the SAP were available

Some concerns

Overall judged some concerns due to missing information about the allocation concealment, deviations from the intended interventions, measurement of the outcome and selection of the reported result

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

Data for this outcome was available for all participants randomised

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.

Low risk of bias

No issue with selective reporting

Some concerns

Overall judged some concerns due to the randomisation process, protocol deviations and limited bias in measurement.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

Data for this outcome was available for all participants randomised

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall rated some concerns due to protocol deviations and measurement of the outcome.

Risk of bias for analysis 1.4 Clinical worsening: new need for IMV or death.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

Participants and clinicians were aware of the assigned treatment. Furthermore, 15% of the participants in the no hydrocortisone group received hydrocortisone, and although we deem it rather unlikely that a larger part of these deviations arose from the experimental context, we see some concerns.

Low risk of bias

Since the data were available for all participants, we arrive at the following assessment of domain 3: Low

Some concerns

Assessment of new need for IMV is partly subjective, but we deem it unlikely that large efforts were made or feasible to treat one arm different from another.

Low risk of bias

Because the data were analysed in accordance with a pre‐specified analysis plan we judged this domain low.

Some concerns

Overall judged some concerns because of some degree of deviations and some concerns arising from assessment of the outcome.

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

Participants and clinicians were aware of the assigned treatment. Furthermore, 15% of the participants in the no hydrocortisone group received hydrocortisone, and although we deem it rather unlikely that a larger part of these deviations arose from the experimental context, we see some concerns.

Low risk of bias

Since the data were available for all participants, we arrive at the following assessment of domain 3: Low

Some concerns

Assessment of new need for IMV is partly subjective, but we deem it unlikely that large efforts were made or feasible to treat one arm different from another.

Low risk of bias

Because the data were analysed in accordance with a pre‐specified analysis plan we judged this domain low.

Some concerns

Overall judged some concerns because of some degree of deviations and some concerns arising from assessment of the outcome.

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

Because data is available for all participants we judged this domain low.

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.

Low risk of bias

The analysis for this result was prespecified in the protocol

Some concerns

Overall judged some concerns due to deviations from the intended intervention in the control group and limited bias in measurement.

Risk of bias for analysis 1.5 Serious adverse events.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

Participants and clinicians were aware of the assigned treatment. Further 15% of the participants in the no hydrocortisone group received hydrocortisone, and although we deem it rather unlikely that a larger part of these deviations arose from the experimental context, we see some concerns.

High risk of bias

Since data were not adjusted for competing risk of death in the presence of a relevant death rate, risk of bias is high.

Low risk of bias

The measurement method was appropriate.

Low risk of bias

Protocol and SAP available.

High risk of bias

Because of the issue of competing risk of death all data for this outcome might not be available. There was no adjustment in the analysis for competing risk of death.

Angus 2020

Low risk of bias

Participants were randomised to each locally available group using balanced assignment. Participants were randomly assigned via a computer software program to each locally available group using proportional assignment (e.g., 1:1 if 2 groups available and 1:1:1 if 3 groups available) Protocol: The RAR will occur centrally as part of the computerised randomisation process. Sites will receive the allocation status and will not be informed of the randomisation proportions. Each region will maintain its own computer‐based randomisation program that is accessed by sites in that region but the RAR proportions will be determined by a SAC and provided monthly to the administrator of each region's randomisation program who will update the RAR proportions Baseline characteristics were similar across groups.

Some concerns

Participants and clinicians were aware of the assigned treatment. Further 15% of the participants in the no hydrocortisone group received hydrocortisone, and although we deem it rather unlikely that a larger part of these deviations arose from the experimental context, we see some concerns.

High risk of bias

Since data were not adjusted for competing risk of death in the presence of a relevant death rate, risk of bias is high.

Low risk of bias

The measurement method was appropriate.

Low risk of bias

Protocol and SAP available.

High risk of bias

Because of the issue of competing risk of death all data for this outcome might not be available. There was no adjustment in the analysis for competing risk of death.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

High risk of bias

Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

High risk of bias

Occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias, so overall we judged high risk of bias.

Risk of bias for analysis 1.6 Adverse events.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Edalatifard 2020

Some concerns

No information about the allocation concealment.

Some concerns

6 patients in the control group received the intervention drug and were excluded from the analyses (17%)

High risk of bias

6 patients in the control group were not included in the analysis because of deviations from the protocol (17%) Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Some concerns

Neither the protocol nor statistical analysis plan were available

High risk of bias

Overall judged high because 6 patients in the control group received the intervention drug and were excluded from the analyses. Moreover, occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias.

Tang 2021

Low risk of bias

Randomisation was stratified by the statistician of the leading site, who produced computer‐generated block randomisation lists with a block size of 4 patients.

Low risk of bias

ITT presented. The participants were blinded, and the physicians were aware of the treatment assignment. No protocol deviations.

High risk of bias

Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Some concerns

Endpoint including its subitems were not named in NCT registration.

High risk of bias

Occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias, so overall we judged high risk of bias.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

High risk of bias

We requested data from the authors. Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

High risk of bias

Occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias, so overall we judged high risk of bias.

Risk of bias for analysis 1.7 Hospital‐acquired infections.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

No deviations from the protocol.

High risk of bias

We request data from the authors. Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Some concerns

Endpoint and definition is not pre‐specified in Eudra‐CT registration.

High risk of bias

Occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias, so overall we judged high risk of bias.

Dequin 2020

Low risk of bias

Randomisation was centralised and performed electronically. Allocation sequences were generated in a 1:1 ratio by a computer‐generated random number using a blocking schema. There were no baseline differences between the groups.

Low risk of bias

Protocol: Patients, investigators and care providers will be blinded for the patient‐arm

High risk of bias

Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available.

High risk of bias

Occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias, so overall we judged high risk of bias.

Tang 2021

Low risk of bias

Randomisation was stratified by the statistician of the leading site, who produced computer‐generated block randomisation lists with a block size of 4 patients.

Low risk of bias

Single‐blind design, ITT data presented. No deviations reported.

High risk of bias

Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Some concerns

Endpoint was not named in NCT registration.

High risk of bias

Occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias, so overall we judged high risk of bias.

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

High risk of bias

We requested data from the authors. Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

High risk of bias

Occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias, so overall we judged high risk of bias.

Risk of bias for analysis 1.8 Invasive fungal infections.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Corral‐Gudino 2021

Some concerns

We judged the domain some concerns due to missing information about the allocation concealment.

Low risk of bias

Participants and clinicians were aware of the assigned treatment. The intention‐to‐treat‐analysis was used to estimate the effect of the assignment to intervention. All this leads us to the following judgement for domain 2: Low.

High risk of bias

Because of the issue of competing risk of death all data for this outcome might not be available. There was no adjustment in the analysis for competing risk of death.

Low risk of bias

No issues with measurement.

Some concerns

We judged some concerns as the outcome was not pre‐specified but requested from the authors.

High risk of bias

Overall we judged risk of bias to be high mainly because of missing adjustment for competing risk of death in domain 3.

Risk of bias for analysis 1.10 New need for dialysis.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

High risk of bias

Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available.

High risk of bias

Occurrence of non‐terminal events can be precluded by death as a competing risk. Without adjustment this leads to high risk of bias, so overall we judged high risk of bias.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment; Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

High risk of bias

Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

High risk of bias

Overall judged high due to missing outcome data.

Risk of bias for analysis 1.11 Viral clearance.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment; Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

High risk of bias

Unknown amount of missing data because of competing risk, and no analysis to adjust for this was made.

Low risk of bias

The measurements were similar between groups.

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

High risk of bias

Overall judged high due to missing outcome data.

Methylprednisolone versus dexamethasone

We had some concerns about all‐cause‐mortality up to 30 days because of questionable pre‐specification, the only outcome included in this version of the review (Table 105).

Risk of bias for analysis 8.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ranjbar 2021

Low risk of bias

Random allocation using the block randomisation method was performed in all four branches of the strata, based on two prognostic factors such as age (< 55 and ≥55) and disease severity based on O2 saturation (<85 and ≥85). The patient, assessor, and analyser in the two groups did not have access to the randomisation list and type of administered drug (Triple blind). No major differences in the baseline characteristics.

Low risk of bias

The patient, assessor, and analyser in the two groups did not have access to the randomisation list and type of administered drug (Triple blind). ITT analysis was used.

Low risk of bias

No missing data.

Low risk of bias

The measurements were similar between groups

Some concerns

Protocol and SAP are not available.

Some concerns

Overall judged some concerns due to selective reporting.

High‐dose dexamethasone versus low‐dose dexamethasone

We had some concerns about all‐cause‐mortality up to 30 days because of issues with randomisation and deviations from intended interventions (Table 106). We judged all‐cause mortality up to 120 days to be at high risk of bias mainly because of deviations from intended interventions (Table 107). We also applied the respective risk of bias assessments to the four subgroup analyses, although it has to be noted that thereby the age‐stratified analysis could be assessed to be at low risk of bias (Table 113; Table 114; Table 115; Table 116; Table 117; Table 118). We had some concerns about clinical improvement (discharged alive) because of issues with randomisation (Table 108). We judged all safety outcomes (serious adverse events, adverse events, hospital‐acquired infections, and invasive fungal infections) to be at high risk of bias due to missing adjustment for competing risk of death (Table 109; Table 110; Table 111; Table 112).

Risk of bias for analysis 9.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

Few, balanced, and blinded exclcusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Taboada 2021

Some concerns

There are no further information regarding the concealment of the allocation sequence

Low risk of bias

Both participants and those delivering the intervention were aware of intervention received, but there was no information on deviations from intended intervention. The analysis was appropriate.

Low risk of bias

Data for this outcome was available for all 200 participants randomised.

Low risk of bias

The measurement of the outcome was appropriate, and it is unlikely that it differed between intervention groups. The outcome assessors were aware of the intervention received, but it is unlikely that knowledge of intervention received could have affected outcome measurement

Low risk of bias

Outcome is pre‐specified.

Some concerns

However, there are some concerns for bias in the randomisation process.

Risk of bias for analysis 9.2 All‐cause mortality up to 120 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

Few, balanced, and blinded exclcusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Taboada 2021

Some concerns

There are no further information regarding the concealment of the allocation sequence

Low risk of bias

Both participants and those delivering the intervention were aware of intervention received, but there was no information on deviations from intended intervention. The analysis was appropriate.

Low risk of bias

Data for this outcome was available for all 200 participants randomised.

Low risk of bias

The measurement of the outcome was appropriate and it is unlikely that it differed between intervention groups. The outcome assessors were aware of the intervention received, but it is unlikely that knowledge of intervention received could have affected outcome measurement

Low risk of bias

Protocol available. Outcome pre‐specified.

Some concerns

For this outcome, there is low risk of bias due to deviations from intended interventions, due to missing outcome data and in measurement of the outcome. However, there are some concerns for bias in the randomisation process.

Toroghi 2021

Low risk of bias

Participants were block randomised, and the allocation sequence was concealed. There are no baseline differences that would suggest a problem with randomisation.

High risk of bias

The study design was blind on patient level, but there were no arrangements that ensured blinding. Therefore, we do not know whether the higher dropout rate in the intermediate/high‐dose groups occurred due to the assigned intervention. No information on treatment adherence was reported.

Low risk of bias

Data was available for nearly all participants.

Low risk of bias

The measurement of the outcome was appropriate, and it is unlikely that it differed between intervention groups. The outcome assessors were unaware of the intervention received.

Some concerns

There was a trial register entry, which provided details of the pre‐specified outcomes, but "mortality" was not pre‐specified at trial registration and in the protocol

High risk of bias

The study design was blind on patient level, but there were no arrangements that ensured blinding. Therefore, we do not know whether the higher dropout rate in the intermediate/high‐dose groups occurred due to the assigned intervention. No information on treatment adherence was reported.

Risk of bias for analysis 10.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 10.1.1 Female

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

We requested the data from the authors. No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

We requested data from the authors. Few, balanced, and blinded exclusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Subgroup 10.1.2 Male

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

We requested the data from the authors. No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

We requested data from the authors. Few, balanced, and blinded exclusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Risk of bias for analysis 10.2 All‐cause mortality up to 120 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 10.2.1 Female

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

We requested data from the authors. Few, balanced, and blinded exclcusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Subgroup 10.2.2 Male

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

We requested data from the authors. Few, balanced, and blinded exclcusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Risk of bias for analysis 11.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 11.1.1 <70 years

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

We requested data from the authors. Few, balanced, and blinded exclusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Subgroup 11.1.2 ≥ 70 years

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

We requested data from the authors. Few, balanced, and blinded exclusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Risk of bias for analysis 11.2 All‐cause mortality up to 120 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 11.2.1 <70 years

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

We requested data from the authors. No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

We requested data from the authors. Few, balanced, and blinded exclusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Subgroup 11.2.2 ≥ 70 years

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

We requested data from the authors. No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

We requested data from the authors. Few, balanced, and blinded exclusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Risk of bias for analysis 12.1 All‐cause mortality up to 30 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 12.1.1 High‐income countries

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

Few, balanced, and blinded exclusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Taboada 2021

Some concerns

There are no further information regarding the concealment of the allocation sequence

Low risk of bias

Both participants and those delivering the intervention were aware of intervention received, but there was no information on deviations from intended intervention. The analysis was appropriate.

Low risk of bias

Data for this outcome was available for all 200 participants randomised.

Low risk of bias

The measurement of the outcome was appropriate, and it is unlikely that it differed between intervention groups. The outcome assessors were aware of the intervention received, but it is unlikely that knowledge of intervention received could have affected outcome measurement

Low risk of bias

Outcome is pre‐specified.

Some concerns

However, there are some concerns for bias in the randomisation process.

Subgroup 12.1.2 Low‐ and middle‐income countries

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Risk of bias for analysis 12.2 All‐cause mortality up to 120 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 12.2.1 High‐income countries

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

Low risk of bias

Few, balanced, and blinded exclusions for similar reasons in both arms.

Low risk of bias

No issues with measurement

Low risk of bias

No issues with potential selection.

Low risk of bias

Low risk of bias in all domains.

Taboada 2021

Some concerns

There are no further information regarding the concealment of the allocation sequence

Low risk of bias

Both participants and those delivering the intervention were aware of intervention received, but there was no information on deviations from intended intervention. The analysis was appropriate.

Low risk of bias

Data for this outcome was available for all 200 participants randomised.

Low risk of bias

The measurement of the outcome was appropriate, and it is unlikely that it differed between intervention groups. The outcome assessors were aware of the intervention received, but it is unlikely that knowledge of intervention received could have affected outcome measurement

Low risk of bias

Protocol available. Outcome pre‐specified.

Some concerns

For this outcome, there is low risk of bias due to deviations from intended interventions, due to missing outcome data and in measurement of the outcome. However, there are some concerns for bias in the randomisation process.

Subgroup 12.2.2 Low‐ and middle‐income countries

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with deviations (2/51 patients from the control received not intended intervention) and appropriate analysis.

Low risk of bias

No relevant amount of missing data.

Low risk of bias

No relevant issues with measurement.

Low risk of bias

No relevant issues with reporting.

Some concerns

Some minor concerns regarding deviations in about 5% of patients and their exclusion from analysis.

Toroghi 2021

Low risk of bias

Participants were block randomised, and the allocation sequence was concealed. There are no baseline differences that would suggest a problem with randomisation.

High risk of bias

The study design was blind on patient level, but there were no arrangements that ensured blinding. Therefore, we do not know whether the higher dropout rate in the intermediate/high‐dose groups occurred due to the assigned intervention. No information on treatment adherence was reported.

Low risk of bias

Data was available for nearly all participants.

Low risk of bias

The measurement of the outcome was appropriate, and it is unlikely that it differed between intervention groups. The outcome assessors were unaware of the intervention received.

Some concerns

There was a trial register entry, which provided details of the pre‐specified outcomes, but "mortality" was not pre‐specified at trial registration and in the protocol

High risk of bias

The study design was blind on patient level, but there were no arrangements that ensured blinding. Therefore, we do not know whether the higher dropout rate in the intermediate/high‐dose groups occurred due to the assigned intervention. No information on treatment adherence was reported.

Risk of bias for analysis 9.3 Clinical improvement: discharged alive.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Taboada 2021

Some concerns

There are no further information regarding the concealment of the allocation sequence

Low risk of bias

Both participants and those delivering the intervention were aware of intervention received, but there was no information on deviations from intended intervention. The analysis was appropriate.

Low risk of bias

Data for this outcome was available for all 200 participants randomised.

Low risk of bias

The measurement of the outcome was appropriate and it is unlikely that it differed between intervention groups. The outcome assessors were aware of the intervention received, but it is unlikely that knowledge of intervention received could have affected outcome measurement

Low risk of bias

Protocol available. Outcome pre‐specified.

Some concerns

For this outcome, there is low risk of bias due to deviations from intended interventions, due to missing outcome data and in measurement of the outcome. However, there are some concerns for bias in selection of the reported result and the randomisation process.

Risk of bias for analysis 9.4 Serious adverse events.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with appropriate analysis but no relevant number of deviations or exclusions from analysis.

High risk of bias

Missing adjustment for competing risk of death in a trial with relevant death rate.

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with reporting

High risk of bias

High risk of bias due to missing adjustment for competing risk of death in domain 3.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

High risk of bias

Few, balanced, and blinded exclusions for similar reasons in both arms. Missing adjustment for competing risk of death.

Low risk of bias

No issues with measurement

Low risk of bias

No extraordinary issues with potential selection.

High risk of bias

High risk of bias due to missing adjustment for competing risk of death.

Risk of bias for analysis 9.5 Adverse events.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with appropriate analysis but no relevant number of deviations or exclusions from analysis.

High risk of bias

Missing adjustment for competing risk of death in a trial with relevant death rate.

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with reporting.

High risk of bias

High risk of bias due to missing adjustment for competing risk of death in domain 3.

Risk of bias for analysis 9.6 Hospital‐acquired infections.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with appropriate analysis but no relevant number of deviations or exclusions from analysis.

High risk of bias

Missing adjustment for competing risk of death in a trial with relevant death rate.

Some concerns

Some concerns because also subjective clinical assessment was allowed but not deemed of major concern.

Low risk of bias

No relevant issues with reporting

High risk of bias

High risk of bias due to missing adjustment for competing risk of death in domain 3.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions

High risk of bias

We requested data from the authors. No adjustment for competing risk of death done or supported.

Low risk of bias

No issues with measurement

Low risk of bias

No extraordinary issues with potential selection.

High risk of bias

High risk of bias due to missing adjustment for competing risk of death.

Risk of bias for analysis 9.7 Invasive fungal infections.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Maskin 2021

Low risk of bias

No issues with randomisation.

Some concerns

Minor issues with appropriate analysis but no relevant number of deviations or exclusions from analysis.

High risk of bias

Missing adjustment for competing risk of death in a trial with relevant death rate.

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with reporting

High risk of bias

High risk of bias due to missing adjustment for competing risk of death in domain 3.

Munch 2021b

Low risk of bias

No issues with randomisation process.

Low risk of bias

No issues with deviations from intended interventions.

High risk of bias

Adjustment for competing risk of death missing or not possible on time‐to‐event scale.

Low risk of bias

No issues with measurement.

Low risk of bias

No issues with potential selection, extraction initiated by review authors.

High risk of bias

High risk of bias due to missing data or adjustment regarding competing risk of death in circumstances of relevant mortality.

Effects of interventions

See: Table 1; Table 2; Table 3

Hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19

From 16 RCTs in this population (Angus 2020; Corral‐Gudino 2021; Dequin 2020; Edalatifard 2020; Farahani 2021; Horby 2021; Jamaati 2021; Jeronimo 2020; Maskin 2021; Munch 2021a; Munch 2021b; Ranjbar 2021; Taboada 2021; Tang 2021; Tomazini 2020; Toroghi 2021), the study of Farahani 2021 did not report any of the prioritised outcomes of interest.

We received data for serious adverse events, adverse events, hospital‐acquired infections, invasive fungal infections, age‐stratified, sex‐stratified and ethnicity‐stratified mortality from four studies upon request (Corral‐Gudino 2021; Maskin 2021; Munch 2021b; Tomazini 2020).

Systemic corticosteroids plus standard care versus standard care (plus/minus placebo)

The evidence profile is presented in Table 1.

All‐cause mortality up to 30 days

Data on all‐cause mortality up to 30 days were available from nine studies with a total of 7898 participants (Angus 2020; Corral‐Gudino 2021; Dequin 2020; Horby 2021; Jamaati 2021; Jeronimo 2020; Munch 2021a; Tang 2021; Tomazini 2020). The observation periods ranged from 21 to 28 days. Overall, 764 of 2937 participants in the intervention groups died compared with 1357 of 4961 participants in the control groups. The risk ratio (RR) of this dichotomous outcome of death was 0.90 (95% confidence interval (CI) 0.84 to 0.97; I² = 0%; random‐effects model; Analysis 1.1). We downgraded the certainty of the evidence for this outcome from high to moderate due to serious risk of bias.

1.1. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

Subgroup analyses: clinical relevance

• Respiratory support at randomisation: the test for subgroup differences indicated no difference (P = 0.11) (Analysis 2.1). In contrast with the beneficial effect on mortality in all those subgroups needing respiratory support, a higher risk of death with systemic corticosteroids was seen among symptomatic COVID‐19 participants who did not need any respiratory support (RR 1.27, 95% CI 1.00 to 1.61). A single large study contributed data for participants without the need for respiratory support (Horby 2021).

• Dexamethasone versus methylprednisolone versus hydrocortisone indirectly compared for their effects relative to placebo/standard care: the test for subgroup differences indicated no difference (P = 0.71) (Analysis 3.1). However, the effect estimates for each subgroup were in favour of corticosteroid use.

2.1. Analysis.

Comparison 2: Subgroup analysis: respiratory support for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

3.1. Analysis.

Comparison 3: Subgroup analysis: dexamethasone versus methylprednisolone versus hydrocortisone for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

Subgroup analyses: equity‐related aspects

• Sex ‐ female versus male: the test for subgroup differences indicated no difference (P = 0.44) (Analysis 4.1). However, the effect estimates for each subgroup were in favour of corticosteroid use.

• Age ‐ < 70 years versus ≥ 70 years: the test for subgroup differences indicated differences (P = < 0.0001) (Analysis 5.1). Participants younger than 70 years seem to benefit from corticosteroids in contrast to participants who were aged 70 years and older. Furthermore, the degree of heterogeneity in each of the subgroups is much lower (< 10 %) than in the set of the combined subgroups (77%). This might be a sign for identifying one of the causes of overall heterogeneity.

• Ethnicity ‐ White versus Black, Asian or minority ethnic group versus unknown: the test for subgroup differences indicated differences (P = 0.01) (Analysis 6.1). The few participants from a Black, Asian, or minority ethnic group had a larger estimated effect than the many White participants.

• Place of residence ‐ high‐income versus low‐ and middle‐income countries: the test for subgroup differences indicated no difference (P = 0.42) (Analysis 7.1). However, the effect estimates for both subgroups were in favour of corticosteroid use.

4.1. Analysis.

Comparison 4: Subgroup analysis: female versus male for the comparison of systemic corticosteroids versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

5.1. Analysis.

Comparison 5: Subgroup analysis: < 70 years versus ≥ 70 years for the comparison of systemic corticosteroids versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

6.1. Analysis.

Comparison 6: Subgroup analysis: White versus Black, Asian or minority ethnic group versus unknown for the comparison of systemic corticosteroids versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

7.1. Analysis.

Comparison 7: Subgroup analysis: high‐income countries versus low‐ and middle‐income countries for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

Sensitivity analyses

We summarised the effects of sensitivity analyses in Table 11. Reported effects in our main analysis were robust when removing platform trials and using the fixed‐effect model. As we found only a low risk of bias or some concerns in our included RCTs for this outcome, there was no opportunity to exclude studies at high risk of bias for sensitivity analysis. Furthermore, none of the included studies for this comparison were preprints.

8. Sensitivity analyses for the comparison: systemic corticosteroids plus standard care versus standard care (plus/minus placebo).

Outcome	Main analyses	Risk of bias (excluding studies at high risk of bias)	Fixed‐effect model or random‐effects model	Preprint (excluding preprints)	Platform trial (excluding platform trialsa)
All‐cause mortality up to 30 days	RR 0.90, 95% CI 0.84 to 0.97; 9 studies, 7898 participants	There were no RCTs with high risk of bias concerning this outcome	RR 0.90, 95% CI 0.83 to 0.97; 9 studies, 7898 participants	There were no RCTs as a preprint concerning this outcome	RR 0.94, 95% CI 0.82 to 1.08; 7 studies, 1094 participants
All‐cause mortality up to 120 days	RR 0.74, 95% CI 0.23 to 2.34; 3 studies, 485 participants	There were no RCTs with high risk of bias concerning this outcome	RR 0.92, 95% CI 0.73 to 1.15; 3 studies, 485 participants	There were no RCTs as a preprint concerning this outcome	There were no platform trials concerning this outcome
Clinical improvement: discharged alive	RR 1.07, 95% CI 1.03 to 1.11; 3 studies, 6786 participantsb	There were no RCTs with high risk of bias concerning this outcome	RR 1.36, 95% CI 0.95 to 1.96; 3 studies, 6786 participantsc	There were no RCTs as a preprint concerning this outcome	RR 1.65, 95% CI 1.23 to 2.21; 2 studies, 361 participants

^aExcluding platform trials: Angus 2020; Horby 2021.

^bUsing the fixed‐effect model.

^c Using the random‐effects model.

CI: confidence interval; RR: risk ratio

All‐cause mortality up to 120 days

Data on all‐cause mortality up to 120 days were available from three studies with a total of 485 participants (Edalatifard 2020; Jeronimo 2020 (matched with long‐term data from Barros 2021); Munch 2021a). Overall, 89 of 244 participants in the intervention groups died compared with 97 of 241 participants in the control groups. The RR of death was 0.74 (95% CI 0.23 to 2.34; I² = 79%; random‐effects model; Analysis 1.2). We downgraded the certainty of the evidence for this outcome from high to very low due to serious risk of bias, serious inconsistency, and serious imprecision.

1.2. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 2: All‐cause mortality up to 120 days

Subgroup analyses: clinical relevance

• Respiratory support at randomisation: We refrained from performing this analysis because only a single study would have contributed 62 participants.

• Methylprednisolone versus hydrocortisone indirectly compared for their effect relative to placebo/standard care: the test for subgroup differences indicated no difference (P = 0.17) (Analysis 3.2) and description of different directions of effects appears misleading, taking into account the small sample size.

3.2. Analysis.

Comparison 3: Subgroup analysis: dexamethasone versus methylprednisolone versus hydrocortisone for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 2: All‐cause mortality up to 120 days

Subgroup analyses: equity‐related aspects

• High‐income versus low‐ and middle‐income countries: the test for subgroup differences indicated no difference (P = 0.14) (Analysis 7.2) and the description of different directions of effects appears misleading, taking into account the small sample size.

7.2. Analysis.

Comparison 7: Subgroup analysis: high‐income countries versus low‐ and middle‐income countries for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 2: All‐cause mortality up to 120 days

We could not perform the subgroup analyses for sex, age, and ethnicity due to missing data.

Sensitivity analyses

We summarised the effects of the two sensitivity analyses in Table 11. When using the random‐effects and fixed‐effect model, we found different effects with broad confidence intervals pointing in the same direction (random‐effects model: RR 0.74, 95% CI 0.23 to 2.34; fixed‐effect model: RR 0.92, 95% CI 0.73 to 1.15). There were no study results with high risk of bias and none of the included RCTs were published as preprints for this comparison. Hence, we did not perform the further two planned analyses.

Clinical improvement: discharged alive (at 28 days)

Data on clinical improvement were available from three studies with a total of 6786 participants (Edalatifard 2020; Horby 2021; Tomazini 2020). Overall, 1490 of 2289 participants in the intervention groups were discharged alive compared with 2789 of 4497 participants in the control groups. The RR of being discharged alive was 1.07 (95% CI 1.03 to 1.11; I² = 81%; fixed‐effect model; Analysis 1.3). We downgraded the certainty of the evidence for this outcome from high to low due to serious risk of bias, serious inconsistency, and serious imprecision. Although this is not a critical outcome, because of the high statistical heterogeneity we performed subgroup and sensitivity analyses as pre‐specified:

Subgroup analyses: clinical relevance

• Methylprednisolone versus hydrocortisone indirectly compared for their effect relative to placebo/standard care: the test for subgroup differences indicated a difference (P = 0.01) (Analysis 3.3), but large heterogeneity (I² = 85%). In both groups, the risk ratio favoured treatment with glucocorticoids, while the effect estimate in the trial evaluating methylprednisolone was stronger, but the number of the included participants very low.

3.3. Analysis.

Comparison 3: Subgroup analysis: dexamethasone versus methylprednisolone versus hydrocortisone for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 3: Clinical improvement: discharged alive

Subgroup analyses: equity‐related aspects

• High‐income versus low‐ and middle‐income countries: the test for subgroup differences indicated a difference (P = 0.003) (Analysis 7.3), but large heterogeneity (I² = 88%). In both groups, the risk ratio favoured treatment with glucocorticoids, while the effect estimate in the trial evaluating low‐ and middle‐income countries was stronger, but the number of the included participants very low.

7.3. Analysis.

Comparison 7: Subgroup analysis: high‐income countries versus low‐ and middle‐income countries for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 3: Clinical improvement: discharged alive

We could not perform the subgroup analyses for sex, age, and ethnicity because of missing data.

Sensitivity analyses

We summarised the effects of the sensitivity analyses in Table 11. When using the random‐effects and fixed‐effect model, we found different effects with broad confidence intervals pointing in the same direction (random‐effects model: RR 1.36, 95% CI 0.95 to 1.96; fixed‐effect model: RR 1.07, 95% CI 1.03 to 1.11). When removing the platform trial we found a different effect with a broader confidence interval pointing in the same direction as the main analysis (main analysis: RR 1.07, 95% CI 1.03 to 1.11; excluding Horby 2021: RR 1.65, 95% CI 1.23 to 2.21). As we found only a risk of bias of 'some concerns' in our included RCTs for this outcome, there was no opportunity to exclude studies at high risk of bias for sensitivity analysis. Furthermore, all included RCTs were published in full in peer‐reviewed journals for this comparison.

Clinical worsening: new need for invasive mechanical ventilation or death

Data on clinical worsening were available from two studies with a total of 5586 participants (Angus 2020; Horby 2021). Overall, 527 of 1900 participants in the intervention groups needed invasive mechanical ventilation or died compared with 1040 of 3686 participants in the control groups. The RR of needing invasive mechanical ventilation or death was 0.92 (95% CI 0.84 to 1.01; I² = 72%; fixed‐effect model; Analysis 1.4). We downgraded the certainty of the evidence for this outcome from high to low due to serious risk of bias and serious inconsistency.

Serious adverse events

Two studies reported serious adverse events for 678 participants (Angus 2020; Tomazini 2020). In Angus 2020 (shock‐dependent hydrocortisone group) 5 of 141 participants in the intervention group and 0 of 52 participants in the control group developed a serious adverse event (RR 4.11, 95% CI 0.23 to 72.98). In Angus 2020 (fixed‐dose hydrocortisone group), 4 of 137 participants in the intervention group and 1 of 49 participants in the control group developed a serious adverse event (RR 1.43, 95% CI 0.16 to 12.49). In Tomazini 2020, 5 of 151 participants in the intervention group and 9 of 148 participants in the control group developed a serious adverse event (RR 0.54, 95% CI 0.19 to 1.59). Data on serious adverse events can be taken from Analysis 1.5 and additional information from Table 8. We decided not to carry out a meta‐analysis because we are highly uncertain about the size and direction of the effects not adjusted for competing risk of death. However, a cautious analysis of the descriptive statistics suggested little to no difference between the groups. We downgraded the certainty of the evidence for this outcome from high to very low due to very serious risk of bias and serious imprecision.

1.5. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 5: Serious adverse events

Adverse events (any grade)

Three studies reported adverse events amongst 447 participants (Edalatifard 2020; Tang 2021; Tomazini 2020). In Edalatifard 2020, 2 of 34 participants in the intervention group and 2 of 28 participants in the control group developed an adverse event (RR 0.82, 95% CI 0.12 to 5.48). In Tang 2021, 5 of 43 participants in the intervention group and 8 of 43 participants in the control group developed an adverse event (RR 0.63, 95% CI 0.22 to 1.76). In Tomazini 2020, 122 of 151 participants in the intervention group and 121 of 148 participants in the control group developed an adverse event (RR 0.99, 95% CI 0.89 to 1.10). Data on adverse events are presented in Analysis 1.6 and Table 8. We decided not to carry out meta‐analysis because we are highly uncertain about size and the direction of the effects not adjusted for competing risk of death. However, a cautious analysis of the descriptive statistics suggested little to no difference between the groups. We downgraded the certainty of the evidence for this outcome from high to very low due to very serious risk of bias and serious imprecision.

1.6. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 6: Adverse events

Hospital‐acquired infections

Four studies involving 598 participants reported hospital‐acquired infections (Corral‐Gudino 2021; Dequin 2020; Tang 2021; Tomazini 2020). In Corral‐Gudino 2021, 5 of 35 participants in the intervention group and 1 of 29 participants in the control group developed a hospital‐acquired infection (RR 4.14, 95% CI 0.51 to 33.49). In Dequin 2020, 28 of 76 participants in the intervention group and 30 of 73 participants in the control group developed a hospital‐acquired infection (RR 0.90, 95% CI 0.60 to 1.34). In Tang 2021, 2 of 43 participants in the intervention group and 1 of 43 participants in the control group developed a hospital‐acquired infection (RR 2.00, 95% CI 0.19 to 21.24). In Tomazini 2020, 30 of 151 participants in the intervention group and 39 of 148 participants in the control group developed a hospital‐acquired infection (RR 0.75, 95% CI 0.50 to 1.15). Data on hospital‐acquired infections can be taken from Analysis 1.7 and additional information from Table 8. We decided not to carry out meta‐analysis because we are highly uncertain about size and the direction of the effects not adjusted for competing risk of death. However, a cautious analysis of the descriptive statistics suggested little to no difference between the groups. We downgraded the certainty of the evidence for this outcome from high to very low due to very serious risk of bias and serious imprecision.

1.7. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 7: Hospital‐acquired infections

Invasive fungal infections

One study reported this outcome for 64 participants (Corral‐Gudino 2021), where 1 of 35 participants in the intervention group and 0 of 29 participants in the control group developed an invasive fungal infection (RR 2.50, 95% CI 0.11 to 59.15). Data on invasive fungal infections can be taken from Analysis 1.8 and additional information from Table 8. We are highly uncertain about size and the direction of the effects not adjusted for competing risk of death. We downgraded the certainty of the evidence for this outcome from high to very low due to very serious risk of bias and serious imprecision.

1.8. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 8: Invasive fungal infections

Quality of life (up to 120 days)

One study reported quality of life for 118 participants with the Duke activity status index (DASI) (Jeronimo 2020). The score ranged from 0 to 58.2, with higher scores indicating greater levels of fitness. Participants in the intervention group reached a mean score of 43.3 (SD 14.2) and participants in the control group reached a mean score of 40.6 (SD 13.9). The mean difference was 2.70 (95% CI ‐2.38 to 7.78). Data on quality of life can be taken from Analysis 1.9. However, a cautious analysis of the descriptive statistics suggested little to no difference between the groups. We would downgrade the certainty of the evidence from high to very low due to serious risk of bias and very serious imprecision.

1.9. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 9: Quality of life up to 120 days

New need for dialysis

Two studies reported data on new need for dialysis for 6613 participants (Horby 2021; Jeronimo 2020). In Horby 2021, 89 of 2034 participants in the intervention group and 314 of 4194 participants in the control group needed dialysis (RR 0.58, 95% CI 0.46 to 0.74). In Jeronimo 2020, 29 of 191 participants in the intervention group and 33 of 194 participants in the control group needed dialysis (RR 0.89, 95% CI 0.57 to 1.41). Data on the new need for dialysis can be taken from Analysis 1.10. We decided not to carry out a meta‐analysis because we are highly uncertain about the size and direction of the unadjusted effects. We would downgrade the certainty of the evidence from high to very low due to serious risk of bias and very serious imprecision.

1.10. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 10: New need for dialysis

Viral clearance

One study reported data on viral clearance (Jeronimo 2020). Viral clearance was defined as the presence of viral RNA detected through RT‐PCR in the naso‐/oropharyngeal swab on day seven. In Jeronimo 2020, no viral load was detected in 61 of 117 participants in the intervention group and in 50 of 95 in the control group (RR 0.99, 95% CI 0.77 to 1.28). Data on viral clearance can be taken from Analysis 1.11. We would downgrade the certainty of the evidence from high to very low due to serious risk of bias and very serious imprecision.

1.11. Analysis.

Comparison 1: Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 11: Viral clearance

Different types of systemic corticosteroids: methylprednisolone versus dexamethasone

The evidence profile is presented in Table 2.

All‐cause mortality up to 30 days

Data on all‐cause mortality up to 30 days were available from one trial with 86 participants (Ranjbar 2021), where 8 of 44 participants in the intervention group died compared with 15 of 42 participants in the control group. The RR of death was 0.51 (95% CI 0.24 to 1.07; Analysis 8.1), favouring methylprednisolone. We downgraded the certainty of the evidence for this outcome to very low due to serious risk of bias and very serious imprecision.

8.1. Analysis.

Comparison 8: Methylprednisolone versus dexamethasone for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

Subgroup and sensitivity analyses

We could not perform any of our planned subgroup and sensitivity analyses for the comparison of methylprednisolone versus dexamethasone and the outcome of all‐cause mortality up to 30 days.

Other outcomes

No further outcome of interest was available from Ranjbar 2021 nor could we identify any other study for this comparison.

High‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8 mg)

The evidence profile is presented in Table 3.

All‐cause mortality up to 30 days

Data on all‐cause mortality up to 30 days were available from three studies involving a total of 1269 participants (Maskin 2021; Munch 2021b; Taboada 2021). The observation period in the studies was 28 days. Overall, 159 of 638 participants in the intervention group died compared with 180 of 631 participants in the control group. The RR of death was 0.87 (95% CI 0.73 to 1.04; I² = 0%; random‐effects model; Analysis 9.1). We downgraded the certainty of the evidence for this outcome from high to low due to serious risk of bias and serious imprecision.

9.1. Analysis.

Comparison 9: High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

Subgroup analyses: equity‐related aspects

• Female versus male: the test for subgroup differences indicated no difference (P = 0.92) between subgroups stratified by the sex of the participants (Analysis 10.1). However, the effect estimates for each subgroup were in favour of high‐dose dexamethasone.

• < 70 years versus ≥ 70 years: the test for subgroup differences indicated no difference (P = 0.55) between subgroups stratified by the age of the participants (Analysis 11.1). However, the effect estimates for each subgroup were in favour of high‐dose dexamethasone.

• High‐income versus low‐ and middle‐income countries: the test for subgroup differences indicated no difference (P = 0.41) between subgroups stratified by income of the country (Analysis 12.1).

10.1. Analysis.

Comparison 10: Subgroup analysis: female versus male for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

11.1. Analysis.

Comparison 11: Subgroup analysis: < 70 years versus ≥ 70 years for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

12.1. Analysis.

Comparison 12: Subgroup analysis: high‐income countries versus low‐ and middle‐income countries for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 1: All‐cause mortality up to 30 days

We could not perform the subgroup analyses for respiratory support at randomisation and ethnicity due to insufficient data.

Sensitivity analyses

We summarised the effects of sensitivity analyses in Table 12. Reported effects of our main analysis were robust when using the fixed‐effect model. We found no preprints, platform trials, or trials with high risk of bias concerning this outcome, so we could not exclude them.

9. Sensitivity analyses for the comparison: high‐dose dexamethasone versus low‐dose dexamethasone.

Outcome	Main analyses	Risk of bias (excluding studies at high risk of bias)b	Fixed‐effect model or random‐effects model	Preprint (excluding preprints)	Platform trials (excluding platform trials)
All‐cause mortality up to 30 days	RR 0.87, 95% CI 0.73 to 1.04; 3 studies, 1269 participants	There were no RCTs with high risk of bias concerning this outcome	RR 0.87, 95% CI 0.73 to 1.04; 3 studies, 1269 participants	There were no RCTs as a preprint concerning this outcome	There were no platform trials concerning this outcome
All‐cause mortality up to 120 days	RR 1.05, 95% CI 0.74 to 1.48; 4 studies, 1399 participantsa	RR 0.87, 95% CI 0.74 to 1.02; 3 studies, 1266 participants	RR 0.93, 95% CI 0.79 to 1.08; 4 studies, 1399 participantsc	There were no RCTs as a preprint concerning this outcome	There were no platform trials concerning this outcome

^aUsing the fixed‐effect model.

^bExcluded studies with high risk of bias: Toroghi 2021.

^cUsing the random‐effects model.

CI: confidence interval; RR: risk ratio

All‐cause mortality up to 120 days

Data on all‐cause mortality up to 120 days were available from four studies with a total of 1383 participants (Maskin 2021; Munch 2021b; Taboada 2021; Toroghi 2021). The observation periods ranged from 60 to 90 days. Overall, 218 of 717 participants in the intervention group died compared with 219 of 666 participants in the control group. The RR of death was 0.93 (95% CI 0.79 to 1.08; I² = 54%; fixed‐effect model; Analysis 9.2). We downgraded the certainty of the evidence for this outcome from high to very low due to serious risk of bias, serious inconsistency, and serious imprecision.

Subgroup analyses: equity‐related aspects

• Female versus male: the test for subgroup differences indicated no difference (P = 0.59) between subgroups stratified by the sex of the participants (Analysis 10.2). However, the effect estimates for each subgroup were in favour of high‐dose dexamethasone.

• < 70 years versus ≥ 70 years: the test for subgroup differences indicated no difference (P = 0.30) between subgroups stratified by the age of the participants (Analysis 11.2). However, the effect estimates for each subgroup were in favour of high‐dose dexamethasone.

• High‐income versus low‐ and middle‐income countries: the test for subgroup differences indicated differences (P = 0.03) between subgroups stratified by income of the country (Analysis 12.2) favouring high‐dose dexamethasone in high‐income countries, while the effect estimate in low‐income countries favoured low‐dose dexamethasone. However, the estimated effect for low‐ and middle‐income countries was fully carried by Toroghi 2021, the only study result at high risk of bias in this analysis.

10.2. Analysis.

Comparison 10: Subgroup analysis: female versus male for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 2: All‐cause mortality up to 120 days

11.2. Analysis.

Comparison 11: Subgroup analysis: < 70 years versus ≥ 70 years for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 2: All‐cause mortality up to 120 days

12.2. Analysis.

Comparison 12: Subgroup analysis: high‐income countries versus low‐ and middle‐income countries for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 2: All‐cause mortality up to 120 days

We could not perform the subgroup analyses for respiratory support at randomisation and ethnicity due to insufficient data.

Sensitivity analyses

We summarised the effects of sensitivity analyses in Table 12. When using the random‐effects and fixed‐effect models, we found different effects with broad confidence intervals pointing in opposite directions (random‐effects model: RR 0.93, 95% CI 0.79 to 1.08; fixed‐effect model: RR 1.05, 95% CI 0.74 to 1.48). When excluding the study with high risk of bias (Toroghi 2021), we found effects with broad confidence intervals also pointing in different directions (main analysis: RR 1.05, 95% CI 0.74 to 1.48; excluding study at high risk of bias: RR 0.87, 95% CI 0.74 to 1.02). No preprints nor platform trials contributed to this outcome for this comparison, and thus we did not pursue the respective sensitivity analyses.

Clinical improvement: discharged alive

Data on clinical improvement were available from one study with a total of 200 participants (Taboada 2021). Overall, 85 of 98 participants in the intervention group were discharged alive compared with 90 of 102 participants in the control group. The RR of being discharged alive was 0.98 (95% CI 0.89 to 1.09; I² = not applicable; random‐effects model; Analysis 9.3). We downgraded the certainty of the evidence for this outcome from high to low due to serious risk of bias and serious imprecision.

9.3. Analysis.

Comparison 9: High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 3: Clinical improvement: discharged alive

Clinical worsening: new need for invasive mechanical ventilation or death

We did not identify any study reporting this outcome.

Serious adverse events

Two studies reported serious adverse events for 1080 participants (Munch 2021b; Maskin 2021). In Maskin 2021, 42 of 49 participants in the high‐dose group and 40 of 49 participants in the low‐dose group developed a serious adverse event (RR 1.05, 95% CI 0.88 to 1.25). In Munch 2021b, 70 of 497 participants in the high‐dose group and 85 of 485 participants in the low‐dose group developed a serious adverse event (RR 0.80, 95% CI 0.60 to 1.07). Data on serious adverse events can be taken from Analysis 9.4 and additional information from Table 8. We decided not to carry out a meta‐analysis because we are highly uncertain about the size and direction of the effects not adjusted for competing risk of death. However, a cautious analysis of the descriptive statistics suggested little to no difference between the groups. We downgraded the certainty of the evidence for this outcome from high to very low due to serious risk of bias and very serious imprecision.

9.4. Analysis.

Comparison 9: High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 4: Serious adverse events

Adverse events (any grade)

One study reported data for 98 participants for this outcome (Maskin 2021), where all 49 participants in the high‐dose group and 48 of 49 participants in the low‐dose group developed an adverse event (RR 1.02, 95% CI 0.96 to 1.08). Data on adverse events can be taken from Analysis 9.5 and Table 8. Without adjustment for competing risk of death, a cautious analysis of the descriptive statistics nonetheless suggested little to no difference between the groups. We downgraded the certainty of the evidence for this outcome from high to very low due to serious risk of bias and very serious imprecision.

9.5. Analysis.

Comparison 9: High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 5: Adverse events

Hospital‐acquired infections

Two studies reported this outcome for 1080 participants (Maskin 2021; Munch 2021b). In Maskin 2021, 34 of 49 participants in the high‐dose group and 38 of 49 participants in the low‐dose group developed a hospital‐acquired infection (RR 0.89, 95% CI 0.70 to 1.14). In Munch 2021b, 50 of 497 participants in the high‐dose group and 61 of 485 participants in the low‐dose group developed a hospital‐acquired infection (RR 0.80, 95% CI 0.56 to 1.14). Data on hospital‐acquired infections can be taken from Analysis 9.6 and additional information from Table 8. We decided not to carry out a meta‐analysis because we are highly uncertain about the size and direction of the effects not adjusted for competing risk of death. However, a cautious analysis of the descriptive statistics suggested little to no difference between the groups. We downgraded the certainty of the evidence for this outcome from high to very low due to serious risk of bias and very serious imprecision.

9.6. Analysis.

Comparison 9: High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 6: Hospital‐acquired infections

Invasive fungal infections

Two studies reported this outcome for 1080 participants (Maskin 2021; Munch 2021b). In Maskin 2021, 3 of 49 participants in the high‐dose group and 3 of 49 participants in the low‐dose group developed an invasive fungal infection (RR 1.00, 95% CI 0.21 to 4.71). In Munch 2021b, 15 of 497 participants in the high‐dose group and 21 of 485 participants in the low‐dose group developed an invasive fungal infection (RR 0.70, 95% CI 0.36 to 1.34). Data on invasive fungal infections can be taken from Analysis 9.7 and Table 8. We decided not to carry out a meta‐analysis because we are highly uncertain about the size and direction of the effects not adjusted for competing risk of death. However, a cautious analysis of the descriptive statistics suggested little to no difference between the groups. We downgraded the certainty of the evidence for this outcome from high to very low due to serious risk of bias and very serious imprecision.

9.7. Analysis.

Comparison 9: High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19, Outcome 7: Invasive fungal infections

Other outcomes

No further outcome of interest was available from Maskin 2021, Munch 2021b, Taboada 2021, and Toroghi 2021.

Timing comparisons (early versus late)

No studies provided data for this comparison.

Systemic corticosteroids versus other active substances

No studies provided data for this comparison.

Outpatients with asymptomatic or mild disease

We did not identify any study that investigated the effects of systemic corticosteroids in people with asymptomatic infection or mild disease (i.e. non‐hospitalised individuals).

Discussion

Summary of main results

This review aimed to assess the efficacy and safety of systemic corticosteroids for the treatment of COVID‐19, as well as equity‐related aspects quantitatively in subgroup analyses where possible. We found no studies for people with asymptomatic infection or mild COVID‐19 disease.

For hospitalised people with a confirmed or suspected diagnosis of symptomatic COVID‐19, we identified 16 RCTs (9549 participants). Eleven trials (8019 participants) evaluated systemic corticosteroids plus standard care compared to standard care (with or without placebo), one trial compared different types of corticosteroids (86 participants), and four studies compared different doses of dexamethasone (1444 participants). We identified 42 ongoing studies evaluating systemic corticosteroids and 23 completed studies lacking published results or relevant information on the study design, which we categorised as 'studies awaiting classification'. We checked the proportion of PCR‐positive tests in each study, as some studies included confirmed or suspected COVID‐19 infections, or both. The study with the lowest PCR‐positive rate of approximately 80% was Angus 2020.

Effects of interventions

Hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19

Systemic corticosteroids plus standard care versus standard care (plus/minus placebo)

• Systemic corticosteroids probably slightly reduce all‐cause mortality up to 30 days (moderate‐certainty evidence).

• The evidence is very uncertain about the effect of systemic corticosteroids on all‐cause mortality up to 120 days (very low‐certainty evidence).

• Systemic corticosteroids may slightly increase the chance of clinical improvement (discharged alive) and may slightly decrease the risk of clinical worsening (new need for invasive mechanical ventilation or death) (both low‐certainty evidence).

• We are uncertain of the effects of systemic corticosteroids on serious adverse events, adverse events, hospital‐acquired infections, and invasive fungal infections (very low‐certainty evidence).

Direct comparisons of different types and different dosages

The evidence is of low and mostly very low certainty.

Overall completeness and applicability of evidence

Diagnosis of COVID‐19 was confirmed by positive SARS‐CoV‐2 PCR testing in 93.8% of the participants. Regarding virological aspects, all participants included in this review were enrolled up to 26 May 2021 and none of the studies reported information on vaccination status. Thus, it is important to consider that the findings of this review might no longer be directly applicable to the current (i.e. late 2022) treatment situation of patients who are infected with later variants of SARS‐CoV‐2 including Delta and Omicron, and, at the same time, with an increasing proportion being vaccinated with different types and dosage regimens of SARS‐CoV‐2 vaccines or having recovered from prior infections. Specifically, these limitations regarding variants and vaccination apply most importantly to data comparison 1 (Systemic corticosteroids plus standard care versus standard care (plus/minus placebo)) for which recruitment of the majority of patients had ended in mid 2020 (Table 5).

 In addition, a considerable share of participants also received various COVID‐19 treatment options such as antibiotics with potential antiviral and anti‐inflammatory properties (i.e. azithromycin), established antiviral drugs (e.g. remdesivir, lopinavir/ritonavir), hydroxychloroquine, convalescent plasma, or combinations of these drugs, mostly balanced between the arms. None of the trials evaluated participants with asymptomatic infections or mild disease (non‐hospitalised participants, WHO clinical progression scale 0 to 3); at this point we would like to refer the reader to the Cochrane Review Griesel 2022, addressing the use of inhaled corticosteroids against COVID‐19, which were mostly studied in outpatient participants with mild disease.

Considering the severity of COVID‐19 at the beginning of corticosteroid treatment, the difficult issue remains of the different scales of severity and progression used across studies and in the different guidelines and consensus statements (Marshall 2020; WHO 2021b). Similar to the first version of this review (Wagner 2021a), we used the need for respiratory support (no oxygen, low‐flow oxygen support, non‐invasive ventilation/high‐flow nasal cannula, and invasive ventilation) at randomisation and its changes as a surrogate for COVID‐19 disease severity. Among the 9314 of 9549 participants from all three comparisons for whom we could extract respiratory support information at randomisation, 16% did not receive any additional oxygen, 61% received any non‐invasive respiratory support, and 22% received invasive ventilation.

Regarding our equity‐related analyses, firstly we would like to highlight the fact that study participants in this version of the review mainly originated from high‐income countries. The share of participants from middle‐income countries has sunk from 13% in the first version of this review (Wagner 2021a) to 12% in this version, and again none came from low‐income countries although about 80% of the world's population lives in a low‐ or middle‐income country (Allen 2017). These numbers are quite similar to the findings of a systematic review on representation of low‐ and middle‐income countries in all COVID‐19‐related RCTs in high‐ranking journals, so the inequality extends beyond corticosteroids to the structural level (Ramanan 2022). Regarding the applicability of the evidence in this version of the review, unfortunately we must reiterate that it might only partially apply to people with COVID‐19 who are treated under different circumstances to those of most studies included in our review: there might be a more severe shortage of hospital beds in both ordinary wards and intensive care units, shortage of oxygen, and other resource constraints on the delivery of respiratory support, and other aspects of care labelled as standard care in this review.

Benefits 

Effects on all‐cause mortality

As the main result of our analysis, we found that systemic corticosteroid treatment probably reduces all‐cause mortality (up to 30 days) in hospitalised patients with moderate to severe COVID‐19 from estimated absolute mortality of 27.4 % (standard of care, plus/minus placebo) to 24.6% (95% CI 23% to 26.5%). Considering this small effect size on all‐cause mortality (absolute mortality reduction 2.8%, RR 0.90), we aimed by additional subgroup analyses to identify subsets of patients potentially benefiting most and, further, to ascertain the types and doses of steroids that are more or less beneficial in treating COVID‐19 patients.

Disease severity by level of respiratory support

Since none of the newly included studies contained corresponding data subsets, the result of the subgroup analysis regarding respiratory support as an indicator for disease severity at baseline did not change: subgroup analysis did not reveal a significant difference in all‐cause mortality up to 30 days between the different levels of respiratory support. Despite missing statistical significance for a subgroup difference on 5% alpha error level, we observed a gradient of harm to benefit as the level of respiratory support increased at randomisation, which could be plausible from a biological perspective and might indicate harm in COVID‐19 patients without clinical signs of COVID‐19‐associated pulmonary gas exchange impairment, i.e. no oxygen at baseline. Nonetheless, for scientific rigour we kept our approach to subgroup analysis as per the first version of this review (Wagner 2021a), i.e. with the highest possible level of respiratory support (reported in each trial or through data request) at randomisation and refrained from a posteriori data‐driven pooling of all participants with any need for respiratory support. We acknowledge that this approach may lead to some overlap of subgroups and reduce statistical power; however, we would like to point out again that the observed consistent trend along with the increasing need for respiratory support and the negative effect among those without respiratory support is of informative value, too. However, there is a widely read pre‐planned analysis of combined respiratory support subgroups of the many participants of Horby 2021 in their referenced publication showing a clearer subgroup difference.

Comparison of different types of corticosteroids

Compared to the first version of this review (Wagner 2021a), our subgroup analysis of different types of corticosteroids included only one additional study with a very low number of participants treated with hydrocortisone (Munch 2021a). Hence, the majority of the participants (about 86%) included in this analysis were treated with dexamethasone. Again, we did not detect any significant differences between the estimated effects of dexamethasone, methylprednisolone, or hydrocortisone on all‐cause mortality in COVID‐19. We did an additional subgroup analysis regarding the different types of corticosteroids for the newly included outcomes all‐cause mortality up to 120 days and clinical improvement (discharged alive), but the certainty of the evidence was low to very low. Furthermore, equivalence of dosages could not be ensured. Hence, we refrain from further interpretation of those analyses. Further, unchanged from the first version of this review, in the direct comparison of two different types of corticosteroids the certainty of the evidence remains very low and interpretation of data unwarranted.

Dose escalation

Most participants in the RCTs included for the comparison of systemic corticosteroids versus standard care received daily doses of corticosteroids of hydrocortisone equivalent ≤ 200 mg/day, and we again refrained from subgroup analyses stratified by dosing because data for the high‐dose pulse regimens were very few. However, in this version of the review we were able to include a new comparison encompassing 1444 participants from four RCTs, in which different doses of dexamethasone were directly compared ‐ this had been identified as a research gap in the first version of this review (Wagner 2021a). Low‐certainty evidence showed that dexamethasone dosed at least twice as high (12 mg/day and higher, 300 mg hydrocortisone equivalent) as in Horby 2021 (6 mg/day, 150 mg hydrocortisone equivalent) may add to the probable slight reduction in All‐cause mortality up to 30 days from Analysis 1.1 when compared to low‐dose hydrocortisone equivalents (6 mg to 8 mg dexamethasone/day). Still, the decision for dose escalation based on our evidence should be critically scrutinised, taking into account the low‐certainty evidence basis for the cited result and that safety data on dose escalation are even more scarce in this comparison and of very low certainty. Moreover, large patient groups and settings peculiar to constrained resources and climate are again underrepresented, and harm through fungal secondary infections as discussed below can increase. Finally, the benefits and harms of corticosteroids do not only depend on dose but on treatment duration and regimens; the included studies only examined pulse dose and short‐term regimens of up to 10 days (or up to 14 days when dose tapering was part of the regimen).

Equity‐related subgroup analyses

Here we discuss the equity‐related subgroup analyses of the only moderate‐certainty outcome, All‐cause mortality up to 30 days in the comparison of systemic corticosteroids plus standard care versus standard care (plus/minus placebo).

It should be noted that all the subgroup analyses are only useful for generation of hypotheses, are not adjusted for multiple testing, suffer heavily from missing outcome data, and most likely lack statistical power.

Sex 

In the analyses of sex, estimated effects of both groups were in favour of corticosteroid use without a large difference (Analysis 4.1).

Age

Participants younger than 70 years seem to benefit from corticosteroids in contrast to participants who are aged 70 years and older (Analysis 5.1), which might be attributable to the fact that the immune system's activity generally declines with entering the senium (Aw 2007). Consequently, targetting hyper‐inflammation in COVID‐19 in this older population with systemic corticosteroids could be less effective but more harmful due to known adverse effects of corticosteroids, including hyperglycaemia, further immunosuppression, and consequent vulnerability to secondary infection, which might outweigh the modest benefits as seen in our meta‐analysis (Analysis 1.1). Moreover, the degree of heterogeneity in each of the subgroups is much lower (below 10%) than in the analysis of the two subgroups combined (77%), which could be a hint for identifying one of the causes of overall heterogeneity.

Ethnicity

Black, Asian or minority ethnic group participants seem to benefit from corticosteroids in contrast to White participants, which we struggle to explain based on our own data without the possibility of multiple regression analysis (Analysis 6.1). However, these subgroup data were mainly from Horby 2021 and it is known from a pre‐COVID‐19, large prospective cohort study in the UK that first emergency hospital admissions in patients of Black and Asian ethnicity took place at a significantly earlier age than in Whites (Wan 2021), which might link this finding with the above discussed age‐stratified analysis. It is unclear why details on ethnicity were unavailable for some study participants, or whether they were missing at random or because of specific ethnicity. 

Place of residence 

In our own subgroup analysis, estimated effects were slightly in favour of corticosteroid use without a large difference (Analysis 7.1), but there are compelling signs from non‐RCT studies that widespread corticosteroid use in COVID‐19 might do substantial harm. Data from low‐ and middle‐income countries were relatively scarce, but we felt it was important to explore geographical and economical inequities and hence hypothesised beyond our approach that systemic corticosteroids might have a different impact on outcomes amongst people with COVID‐19 in low‐ and middle‐income countries. Related factors such as ethnicity and climatic zone have been shown to influence COVID‐19 outcomes (Liu 2021). Studies from the UK have consistently shown that ethnic minorities have a higher chance of having poorer COVID‐19 outcomes ‐ including hospitalisation and mortality (Aldridge 2020; Mathur 2021). Similar findings were made in Brazil where the Pardo and Preto ethnicities had a higher risk of COVID‐19‐related mortality (Baqui 2020). There were possible confounding factors such as larger family size, poor access to quality health care, and higher prevalence of co‐morbidities, which could have played a role in the association between ethnicity and COVID‐19 outcomes (Morales 2021). Regarding co‐morbidities, type 2 diabetes mellitus is of high importance: the prevalence varies significantly among various ethnicities and is exceptionally high among Asians and middle‐eastern countries (El‐Kebbi 2021; Ramachandran 2012). Hence, administration of systemic (gluco‐)corticosteroids like the one under study in this review could be associated with higher diabetes‐related adverse effects in populations with a higher prevalence of diabetes and less resources to diagnose and control the condition both acutely and chronically (Alessi 2020; Caughey 2013). Before the second wave of COVID‐19 began globally in March 2021 (WHO 2022a), it was hypothesised that countries with higher temperatures, relative humidity, and those closer to the equator would have fewer daily cases of COVID‐19, possibly due to unfavourable aerosol transmission, particularly in summer (Chen 2021). However, at the end of the second wave (July 2021), no such apparent dissimilarities across countries were noted in COVID‐19 prevalence. Despite the climatic factors, the implementation of public health policies such as social distancing and vaccination programmes could have played a part in the prevention of SARS‐CoV‐2 infections. This argument continues in the following discussion of the potential harms of systemic corticosteroid use.

Potential harms

In general, there is still a relevant underreporting of safety data (serious adverse events, adverse events, hospital‐acquired infections and fungal infections) from the included studies: only two evaluated serious adverse events for 678 participants, five reported partly selected adverse events for 660 participants, five studies provided information on hospital‐acquired infections for 660 participants, and only one study including 64 participants reported data on invasive fungal infections. Unfortunately, the largest included study did not report any adverse event data, but only four suspected drug reactions from the 1996 participants that received at least one dose of dexamethasone. Hence, the certainty of the evidence about potential harmful effects of corticosteroid treatment remains low and the few results from individual trials regarding safety outcomes should not be readily applied to the current clinical treatment situation of COVID‐19 patients.

Invasive fungal infections

As mentioned above, the certainty of the evidence about the safety outcome fungal infections and hospital‐acquired infections is very low. Nevertheless, specific difficult to treat fungal infections like mucormycosis and invasive aspergillosis increasingly draw attention in the context of widespread usage of corticosteroids in critical ill COVID‐19 patients. Even before the onset of COVID‐19, the prevalence of mucormycosis was high in Asian countries, particularly India (Prakash 2019). However, mucormycosis cases were on an unprecedented rapid rise in India during the second wave of COVID in April 2021. It throttled the already struggling health care machinery as the disease was invariably fatal either due to morbid surgery or highly nephrotoxic antifungal therapy. After COVID‐19, there was a sudden increase in the prevalence of mucormycosis. A single‐centre study from South India reported that poorly controlled diabetes (odds ratio of 4.6), previously undiagnosed diabetes (odds ratio of 3.3), and unwarranted steroid use (odds ratio of 28.4) were the three main risk factors that were associated with increased risk of mucormycosis in COVID‐19. The authors have highlighted that almost all the patients who had mucormycosis only had mild COVID‐19, which did not warrant corticosteroids in the management of COVID‐19 (Bhanuprasad 2021). From a global perspective, mucormycosis cases have been reported in various countries, including Europe. The high prevalence of diabetes among mucormycosis patients was a common factor across all countries (Hoenigl 2022). It is well known that corticosteroid therapy is associated with a higher risk of fungal infections. However, a retrospective study from Berlin, Germany reported more than twice the prevalence of aspergillosis in severe and critical COVID‐19 patients receiving corticosteroids compared to those who did not receive them. Diabetes, inappropriate use of steroids, oxygen, and antibiotics are all associated with mucormycosis in COVID‐19 patients (Aaranjani 2021). We tried to obtain these data for our review, particularly from Gautam 2021; however, the data were unavailable. In summary, while the certainty of the evidence about fungal and other secondary infections from prospective randomised trials remains very low, observational data from various countries indicate potential clinically relevant harmful effects of corticosteroid therapy in COVID‐19 patients.

Prospect of studies ongoing and awaiting classification

With regard to additional evidence from future publications, we identified 42 ongoing, recruiting, or temporarily halted studies (compared to 31 as identified in the first version of the review) (Wagner 2021a), encompassing approximately 13,500 participants. The number of studies described as completed, prematurely ended, or terminated grew from 16 to 23, encompassing approximately 5500 participants. Most of the studies intend to recruit people treated with different levels of respiratory support and a growing number seek to investigate different dexamethasone doses or different types of steroids. With the publication of even parts of these data, possible publication bias would be countered and changes to the precision and effect estimates themselves are not unlikely, although we do not expect a relevant difference or even a change of direction in the effect of mortality in the first comparison but rather more precise estimates for subgroup analyses, subordinate and safety outcomes, regarding dosing, and for outpatients. At this point we would like to emphasise that there are five completed and partially published studies from low‐ and middle‐income countries, for which we have so far unsuccessfully contacted the authors for full publication or additional information (Gautam 2021; Ghanei 2021; Montalvan 2021; Rashad 2021).

Quality of the evidence

Systemic corticosteroids plus standard care versus standard care (plus/minus placebo)

We included data from 10 RCTs in the analysis of the efficacy and safety of systemic corticosteroids. The population of interest was hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19 when compared to treatment with standard care (plus/minus placebo).

We rated the certainty of the evidence as moderate to very low (see Table 1). We downgraded the certainty of the evidence due to risk of bias arising from deviations from the intended intervention for all outcomes, missing adjustment for competing risk of death for all safety outcomes, selection of the reported results (all‐cause mortality up to 30 and 120 days and clinical improvement: discharged alive), missing information about allocation concealment (all‐cause mortality up to 30 and 120 days, clinical improvement: discharged alive, adverse events, hospital‐acquired infections, invasive fungal infections), baseline differences (all‐cause mortality up to 30 days), measurement of the outcome (new need for invasive mechanical ventilation or death), and reporting bias (serious adverse events, adverse events, hospital‐acquired infections, invasive fungal infections). Moreover, we downgraded for inconsistency for the following outcomes: all‐cause mortality up to 120 days, clinical improvement (discharged alive), and new need for invasive mechanical ventilation or death. Low numbers of events/imprecision affected all‐cause mortality up to 120 days and all safety outcomes.

Comparison of different types of corticosteroids

Methylprednisolone versus dexamethasone

We included data from one RCT assessing the efficacy and safety of methylprednisolone for individuals with a confirmed diagnosis of COVID‐19 when compared to treatment with dexamethasone.

We rated the certainty of the evidence as very low (see Table 2). We downgraded the certainty of the evidence due to risk of bias arising from the missing pre‐specification of the outcome and the missing protocol, and due to very serious imprecision.

High‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8 mg)

We included data from four RCTs assessing the efficacy and safety of different doses of dexamethasone for individuals with a confirmed diagnosis of COVID‐19.

We rated the certainty of the evidence as low to very low (see Table 3). We downgraded the certainty of the evidence due to risk of bias arising from deviations from the intended intervention (All‐cause mortality up to 30 and 120 days, Adverse events, Hospital‐acquired infections, Invasive fungal infections), missing information about the allocation concealment (All‐cause mortality up to 30 and 120 days, Clinical improvement: Discharged alive, Hospital‐acquired infections, Invasive fungal infections), missing adjustment for competing risk of death (all safety outcomes) and measurement of the outcome (Hospital‐acquired infections, Invasive fungal infections). Furthermore, we downgraded for inconsistency for All‐cause mortality up to 120 days and Hospital‐acquired infections. Low numbers of events/imprecision affected all outcomes.

Critical appraisal of selected outcome parameters

Outcomes with death as competing risk

Death and our exploratory endpoints in the domain of clinical improvement and worsening, as well as safety data, have to be seen as semi‐competing risks for each other: death precludes the occurrence of non‐terminal outcomes if participants die early. We discussed the issue extensively in the first version of this review, and it still has an impact on the risk of bias assessment and decisions not to meta‐analyse non‐terminal endpoints that were not adjusted by design or modelling (Brock 2011; Columbia Public Health 2021; Wu 2020).

Composite outcomes

Conceptual peculiarity of composite outcomes like New need for invasive mechanical ventilation or death and Discharged alive with inherent quasi‐adjustment for competing risk of death were discussed in the first version of this review.

Importance of platform trials and their specific sources of bias

Without a doubt, platform trials such as the included studies Angus 2020 and Horby 2021 offered great chances for rapid and adaptive generation of evidence involving huge numbers of participants in the pandemic. A few platform trials have extraordinarily contributed to the evidence base of pharmacotherapy in COVID‐19, which can be illustrated by their weight in the main RCT‐only mortality analyses in selected Cochrane Reviews: 99.6% regarding colchicine (Mikolajewska 2021), 77% regarding remdesivir (Ansems 2021), > 90% overall regarding convalescent plasma (Piechotta 2021), 100% regarding inhaled corticosteroids (Griesel 2022), and 51% regarding systemic corticosteroids in the first version of this review (Wagner 2021a). Thus, by impacting evidence synthesis their results have significantly influenced treatment guideline recommendations on COVID‐19 treatment. Nevertheless, they inherently have specific sources of risk of bias through their adaptive, multi‐arm, and often pragmatic (e.g. missing source data verification in Horby 2021) and open‐label designs, which are under‐ or unrepresented in the latest tools for risk of bias assessment in randomised trials, e.g. RoB 2 (Sterne 2019). In an attempt to address these issues and since there is not yet an established tool for critically appraising platform trials, we pioneered a checklist (Park 2020), with results available at https://zenodo.org/record/7015269#.YwOn43HP2Ul. In summary, easily detectable bias risk arose from the combination of open‐label and pragmatic design, leading to questionable allocation concealment. However, it was not feasible to sufficiently review and judge the different protocol versions and statistical architecture with its main pillars of multiple interim analyses, adaptive randomisation, and case‐count reducing compilation of control groups within factorial trial design from the outside without employing additional expert resources. The scientific community needs to build up expertise in reviewing platform trials and research groups conducting them should ensure high‐quality standards of data acquisition despite the pragmatic approaches and support the understanding of their architecture in comparison to traditionally performed RCTs in their respective fields of research.

Potential biases in the review process

In addition to peer‐reviewed, full‐text articles, we also included preprints. We are aware of the potentially lower quality of preprint publications, and that the results could change once the peer‐reviewed journal publications are available. In cases of missing data, we contacted study authors for additional data or relevant details if we needed more information. We are confident that we identified all relevant studies and will monitor ongoing studies as well as full publications of preprints closely after the publication of this review. 

We decided on definitions of subgroups before running the first analyses based on the availability of data and based on clinical considerations. Subgroup analyses are exploratory only and confounding variables may have impacted their influence through biology, equity, or both. Our equity‐related subgroup analyses, like the other meta‐analyses, were not adjusted for multiple testing.

Agreements and disagreements with other studies or reviews

In the first version of this review (Wagner 2021a), we compared our findings to the four RCT‐only reviews: Chaudhuri 2021; Pasin 2021; Siemieniuk 2020; Sterne 2020. Since then no updates on these nor other high‐quality, RCT‐only systematic reviews of compelling relevance for discussion in this field have been published.

Once again, the open‐label adaptive and pragmatic platform study Horby 2021 must be discussed due to its huge impact. Not only was it the largest contributor to our review in terms of events and participants, but it also had an immense influence on treatment guidelines and ongoing studies in 2020 when its preliminary report became available (WHO 2021c). The direction and size of the effect on 30‐day mortality is congruent with our overall findings as well as the findings in the subgroup analysis stratified by respiratory support at randomisation. It is worth highlighting that this statement is based on analyses of unadjusted dichotomous 30‐day mortality data, including Horby 2021, in this review, while the authors of Horby 2021 in their publication presented rate ratios of time‐to‐event data from age‐adjusted Cox regression.

Authors' conclusions

Implications for practice.

Based on the current evidence, we are moderately certain that systemic corticosteroids probably slightly reduce All‐cause mortality up to 30 days amongst hospitalised, symptomatic COVID‐19 patients while level of respiratory support might play a critical role in patient selection. This finding is supported by low‐certainty evidence for a slightly increased chance of clinical improvement and a slightly decreased risk of worsening. Subgroup analyses suggest that the beneficial effect of the intervention might vanish for patients aged 70 years and older, but could be stronger in patients of Black, Asian, or other non‐White ethnicity. Long‐term survival data up to 120 days is of very low certainty and hence inconclusive.

Direct evidence on types of corticosteroids in the second comparison is of very low certainty and hence also inconclusive. Dose escalation may add to the slight benefit on All‐cause mortality up to 30 days, but the certainty of the evidence is low. It is most important to state that evidence on safety throughout the comparisons is of very low certainty because of underreporting and missing adjustment for competing risk of death.

The applicability of the data is limited. Data primarily came from European high‐income countries and are therefore limited in their applicability to settings different from that with regard to resources, climate, and other infectious diseases. Our own subgroup analyses are inconclusive in that regard but non‐RCT data raise suspicion of a potential for harm. Moreover, even in this update of the review the data were generated from unvaccinated participants and before the first occurrence of the recent variants of SARS‐CoV‐2, so the evidence may not be fully transferable to patients in 2022 and beyond.

Currently, there is no evidence to characterise the benefits and harms of corticosteroids in patients with asymptomatic or mild disease (non‐hospitalised) and regarding timing of the intervention.

Implications for research.

To address large uncertainties in the already existing analyses, we would prioritise the following:

• Applicability: RCT data from low‐ and middle‐income countries and diverse environments is needed. All publishers should support timely and adequate reporting of already completed studies.

• Safety data (adverse events/serious adverse events, detailed reporting of secondary infections): study authors should consistently measure and report adverse events, serious adverse events, and infections for both study arms.

• Survival data: there is an urgent need for more long‐term data (up to 120 days and beyond) to improve the certainty of the evidence.

• Patient‐centred outcomes: quality of life, neurological function, and independence in daily activities should be examined and reported.

• Regimens: treatment dosing and duration are likely to influence both benefits and harms and the latter remains unaddressed in RCTs.

Interpretation of the current evidence warrants a call for new RCTs in the main comparison of systemic steroids plus standard care to standard care alone (plus/minus placebo) directly investigating the following:

• Age: as there are some hints in the subgroup analysis that older people (above 70 years) might benefit less, study authors should report subgroup analyses for different age groups.

• Immunological effects of recent variants and vaccination: re‐applying study designs from 2020 in contemporary and vaccinated populations.

We identified 42 ongoing and 23 completed RCTs lacking publication or important information in trials registries and databases, which will probably increase the certainty of the evidence in the future. In accordance with the living systematic review approach, we will continually update our search and include eligible trials.

What's new

Date	Event	Description
29 April 2022	New citation required and conclusions have changed	New studies included, new comparisons, equity aspects considered.
28 April 2022	New search has been performed	Update search on January 6, 2022, inclusion of new studies, and introduction of new methods and focus on health equity.

History

Review first published: Issue 8, 2021

Notes

Parts of the review's methods section and of the background were adopted from Cochrane Haematology templates (Kreuzberger 2021; Piechotta 2021).

Risk of bias

Risk of bias for analysis 1.9 Quality of life up to 120 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment; Control group: 5 excluded before starting treatment, 3 excluded after starting treatment As the study is meant to detect a positive effect of corticosteroids, participants in the control group receiving the experimental medication lead to expected bias towards null. Hence, we only see a slight tendency of under‐estimating the effect, but there is no increase in the risk to mistakenly introduce a medication for its beneficial effect.

High risk of bias

Because of the issue of competing risk of death all data for this outcome might not be available. There was no adjustment in the analysis for competing risk of death.

Low risk of bias

We judged this domain low because the measurements were similar between groups.

Some concerns

The outcome was presented on a single scale and with one analysis only, but was not prespecified which leads to some concerns.

High risk of bias

Overall we see high risk bias due to missing adjustment for competing risks

Risk of bias for analysis 7.2 All‐cause mortality up to 120 days.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 7.2.1 High‐income countries

Munch 2021a

Low risk of bias

No issues with randomisation process.

Low risk of bias

Withdrawals with consecutive open‐label steroids may have affected the outcome, but withdrawals and non‐compliance was rather balanced after all.

Low risk of bias

No missing data.

Low risk of bias

No issues with measurement.

Low risk of bias

No relevant issues with the reported result.

Low risk of bias

Withdrawals because of the experimental context lead to some concerns (but it is less than 10% of all patients).

Subgroup 7.2.2 Low‐ and middle‐income countries

Edalatifard 2020

Some concerns

No information about the allocation concealment.

Some concerns

6 patients in the control group received the intervention drug and were excluded from the analyses (17%).

Low risk of bias

The data were requested from the authors because the follow‐up time was not clearly visible from the publication.

Low risk of bias

The measurements were similar between groups.

Some concerns

Neither the protocol nor the SAP were available

Some concerns

Overall judged some concerns due to the randomisation process, protocol deviations and the selection of the reported results.

Jeronimo 2020

Low risk of bias

An independent statistician prepared an electronically generated randomisation list with 14 blocks of 30 participants per block, generated via R software version 3.6.1 (blockrand package). The list was accessible only to non‐blinded pharmacists in the study. Participants were randomised by the study pharmacist to their designated treatment regimen at the time of inclusion and were subsequently identified throughout the study only by their allocated study number. There were no major differences in baseline characteristics between the intervention and placebo groups

Some concerns

Intervention group: 14 excluded before starting treatment, 1 excluded after starting treatment Control group: 5 excluded before starting treatment, 3 excluded after starting treatment

Low risk of bias

416 participants randomised and 416 participants analysed.

Low risk of bias

The measurements were similar between groups

Low risk of bias

Protocol and statistical plan available. Data analysed and presented according to a pre‐specified plan.

Some concerns

Overall judged some concerns due to protocol deviations.

Risk of bias for analysis 7.3 Clinical improvement: discharged alive.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 7.3.1 High‐income country

Horby 2021

Low risk of bias

Randomisation was performed with the use of a web‐based system with concealment of the trial‐group assignment. No baseline differences between the groups.

Some concerns

8% of the participants in the control group received dexamethasone

Low risk of bias

Data for this outcome was available for all participants randomised.

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.

Low risk of bias

No issue with selective reporting.

Some concerns

Overall judged some concerns due to the randomisation process, protocol deviations and limited bias in measurement.

Subgroup 7.3.2 Low‐ and middle‐income country

Edalatifard 2020

Some concerns

No information about the allocation concealment.

Some concerns

In this study, patients did not know which group of them used medicine Physicians and clinicians team know about the medicine and intervention groups. 6 patients in the control group received the intervention drug and were excluded from the analyses Intention‐to‐treat

Low risk of bias

Data for this outcome was available for all participants randomised

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.No protocol or SAP available.

Some concerns

Neither the protocol nor the SAP were available

Some concerns

Overall judged some concerns due to missing information about the allocation concealment, deviations from the intended interventions, measurement of the outcome and selection of the reported result

Tomazini 2020

Low risk of bias

Randomisation was performed through an online web‐based system using computer‐generated random numbers and blocks of 2 and 4, unknown to the investigators, and was stratified by centre. The group treatment was disclosed to the investigator only after all information regarding patient enrolment was recorded in the online system. Baseline characteristics were well‐balanced between groups.

Some concerns

25 deviations from protocol in the intervention arm (16.55%); 1 patient received a corticosteroid other than dexamethasone. In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%).

Low risk of bias

Data for this outcome was available for all participants randomised

Some concerns

Outcome assessors were aware of the treatment assignments. Knowledge of intervention received could have affected ascertainment only in patients who had not died. So the influence of the unblinded assessor is limited but existing ‐ therefore deviation between algorithm and judgement.

Low risk of bias

The protocol and statistical analysis plan were available. Outcomes reported as prespecified.

Some concerns

Overall rated some concerns due to protocol deviations and measurement of the outcome.

Appendices

Appendix 1. Search strategies (previous review published on 16 August 2021)

Cochrane COVID‐19 Study Register

Search string:

corticosteroid* OR corticoid* OR prednison* OR dehydrocortison* OR deltason* OR decortin* OR orasone* OR deltra* OR meticorten* OR cortancyl* OR deltacorten* OR dacortin* OR adasone* OR "delta‐cortison" OR panasol* OR decorton* OR metacortandracin* OR paracort* OR predicor* OR decortisyl* OR delta‐1‐cortison* OR "delta‐dome" OR deltadehydrocortison* OR ofisolon* OR panafcort* OR predicorten* OR predni* OR econonson* OR promifen* OR servison* OR deltison* OR lisacort* OR meproson* OR rayos OR sterapred* OR "liquid pred" OR cortan* OR rectodelt* OR predeltin* OR prednisolon* OR methylprednisolon* OR medrol OR "pred forte" OR medrone OR urbason OR wyacort OR "Delta‐F" OR duralon* OR medrate OR omnipred OR adlone OR caberdelta OR depmedalon* OR "Depo Moderin" OR "Depo‐Nisolone" OR Emmetipi OR esameton* OR firmacort OR medlon* OR "Mega‐Star" OR meprolon* OR metilbetason* OR metrocort OR metypresol OR metysolon* OR orapred OR "Predni‐M‐Tablinen" OR radilem OR sieropresol OR solpredon* OR "A‐MethaPred" OR prelone OR medrone OR aprednislon OR pediapred OR hostacortin OR "Di‐Adreson‐F" OR adnisolon* OR capsoid OR cortalon* OR cortisolon* OR deltacortril OR estilsona OR panafcortelone OR sterane OR "Delta‐Cortef" OR econopred OR dacortin OR decaprednil OR "Delta‐Diona" OR "Delta‐Phoricol" OR deltahydrocortison* OR deltasolon* OR deltidrosol OR dhasolone OR fisopred OR frisolona OR gupison* OR hydeltra OR hydeltrasol OR klismacort OR kuhlprednon OR lenisolon* OR "Lepi‐Cortinolo" OR "Linola‐H" OR longiprednil OR metacortandralon* OR "Meti Derm" OR meticortelon* OR opredsone Or precortisyl OR "Pred‐Clysma" OR predeltilon* OR prenilone OR hydrocortancyl OR "Solu Moderin" OR predonin* OR metypred OR prednisol OR dexamethason* OR "BB 1101" OR decadron OR hexadrol OR fortecortin OR dexameth OR dexone OR hexadecadrol OR desamethason* OR ozurdex OR deronil OR baycuten OR aacidexam OR spersadex OR dexacortal OR gammacorten OR visumetazon* OR adexone OR "Alba‐Dex" OR cortidexason OR decacort OR decadrol OR dectancyl OR desameton OR loverine OR millicorten OR orgadrone OR alin OR auxiloson OR cortisumman OR decalix OR decameth OR decasone OR dekacort OR deltafluorene OR "Dexa‐Mamallet" OR dexafluorene OR dexalocal OR dexamecortin OR dexamonozon OR dexapos OR dexinoral OR fluorodelta OR lokalison OR methylfluorprednisolon* OR mymethason* OR "Dexa‐Rhinosan" OR "Dexa‐Scheroson" OR "Dexa‐sine" OR dexacortin OR dexafarma OR dinormon OR baycadron OR "Aeroseb‐Dex" OR Maxidex OR Dextenza OR dexasone OR dexpak OR hydrocortison* OR cortisol OR cortef OR hydrocorton* OR cetacort OR barseb OR aeroseb OR "Cort‐Dome" OR cortenema OR cortril OR cortifan OR cortispray OR dermacort OR domolene OR eldecort OR hautosone OR "Heb‐Cort" OR hytone OR Komed OR Nutracort OR Proctocort OR Rectoid OR Hydrocort OR locoid OR Solu‐Glyc

Study characteristics:
1) "Intervention assignment": “Randomised” OR
2) "Study type": "Interventional" AND "Study design": "Parallel/Crossover" OR 
3) "Study type": "Interventional" AND "Study design": "Unclear"

Web of Science Core Collection (Advanced search)

#1 TI=( corticosteroid* OR corticoid* OR prednison* OR dehydrocortison* OR deltason* OR decortin* OR orasone* OR deltra* OR meticorten* OR cortancyl* OR deltacorten* OR dacortin* OR adasone* OR "delta‐cortison" OR panasol* OR decorton* OR metacortandracin* OR paracort* OR predicor* OR decortisyl* OR delta‐1‐cortison* OR "delta‐dome" OR deltadehydrocortison* OR ofisolon* OR panafcort* OR predicorten* OR predni* OR econonson* OR promifen* OR servison* OR deltison* OR lisacort* OR meproson* OR rayos OR sterapred* OR "liquid pred" OR cortan* OR rectodelt* OR predeltin* OR prednisolon* OR methylprednisolon* OR medrol OR "pred forte" OR medrone OR urbason OR wyacort OR "Delta‐F" OR duralon* OR medrate OR omnipred OR adlone OR caberdelta OR depmedalon* OR "Depo Moderin" OR "Depo‐Nisolone" OR Emmetipi OR esameton* OR firmacort OR medlon* OR "Mega‐Star" OR meprolon* OR metilbetason* OR metrocort OR metypresol OR metysolon* OR orapred OR "Predni‐M‐Tablinen" OR radilem OR sieropresol OR solpredon* OR "A‐MethaPred" OR prelone OR medrone OR aprednislon OR pediapred OR hostacortin OR "Di‐Adreson‐F" OR adnisolon* OR capsoid OR cortalon* OR cortisolon* OR deltacortril OR estilsona OR panafcortelone OR sterane OR "Delta‐Cortef" OR econopred OR dacortin OR decaprednil OR "Delta‐Diona" OR "Delta‐Phoricol" OR deltahydrocortison* OR deltasolon* OR deltidrosol OR dhasolone OR fisopred OR frisolona OR gupison* OR hydeltra OR hydeltrasol OR klismacort OR kuhlprednon OR lenisolon* OR "Lepi‐Cortinolo" OR "Linola‐H" OR longiprednil OR metacortandralon* OR "Meti Derm" OR meticortelon* OR opredsone Or precortisyl OR "Pred‐Clysma" OR predeltilon* OR prenilone OR hydrocortancyl OR "Solu Moderin" OR predonin* OR metypred OR prednisol OR dexamethason* OR "BB 1101" OR decadron OR hexadrol OR fortecortin OR dexameth OR dexone OR hexadecadrol OR desamethason* OR ozurdex OR deronil OR baycuten OR aacidexam OR spersadex OR dexacortal OR gammacorten OR visumetazon* OR adexone OR "Alba‐Dex" OR cortidexason OR decacort OR decadrol OR dectancyl OR desameton OR loverine OR millicorten OR orgadrone OR alin OR auxiloson OR cortisumman OR decalix OR decameth OR decasone OR dekacort OR deltafluorene OR "Dexa‐Mamallet" OR dexafluorene OR dexalocal OR dexamecortin OR dexamonozon OR dexapos OR dexinoral OR fluorodelta OR lokalison OR methylfluorprednisolon* OR mymethason* OR "Dexa‐Rhinosan" OR "Dexa‐Scheroson" OR "Dexa‐sine" OR dexacortin OR dexafarma OR dinormon OR baycadron OR "Aeroseb‐Dex" OR Maxidex OR Dextenza OR dexasone OR dexpak OR hydrocortison* OR cortisol OR cortef OR hydrocorton* OR cetacort OR barseb OR aeroseb OR "Cort‐Dome" OR cortenema OR cortril OR cortifan OR cortispray OR dermacort OR domolene OR eldecort OR hautosone OR "Heb‐Cort" OR hytone OR Komed OR Nutracort OR Proctocort OR Rectoid OR Hydrocort OR locoid OR Solu‐Glyc) OR AB=( corticosteroid* OR corticoid* OR prednison* OR dehydrocortison* OR deltason* OR decortin* OR orasone* OR deltra* OR meticorten* OR cortancyl* OR deltacorten* OR dacortin* OR adasone* OR "delta‐cortison" OR panasol* OR decorton* OR metacortandracin* OR paracort* OR predicor* OR decortisyl* OR delta‐1‐cortison* OR "delta‐dome" OR deltadehydrocortison* OR ofisolon* OR panafcort* OR predicorten* OR predni* OR econonson* OR promifen* OR servison* OR deltison* OR lisacort* OR meproson* OR rayos OR sterapred* OR "liquid pred" OR cortan* OR rectodelt* OR predeltin* OR prednisolon* OR methylprednisolon* OR medrol OR "pred forte" OR medrone OR urbason OR wyacort OR "Delta‐F" OR duralon* OR medrate OR omnipred OR adlone OR caberdelta OR depmedalon* OR "Depo Moderin" OR "Depo‐Nisolone" OR Emmetipi OR esameton* OR firmacort OR medlon* OR "Mega‐Star" OR meprolon* OR metilbetason* OR metrocort OR metypresol OR metysolon* OR orapred OR "Predni‐M‐Tablinen" OR radilem OR sieropresol OR solpredon* OR "A‐MethaPred" OR prelone OR medrone OR aprednislon OR pediapred OR hostacortin OR "Di‐Adreson‐F" OR adnisolon* OR capsoid OR cortalon* OR cortisolon* OR deltacortril OR estilsona OR panafcortelone OR sterane OR "Delta‐Cortef" OR econopred OR dacortin OR decaprednil OR "Delta‐Diona" OR "Delta‐Phoricol" OR deltahydrocortison* OR deltasolon* OR deltidrosol OR dhasolone OR fisopred OR frisolona OR gupison* OR hydeltra OR hydeltrasol OR klismacort OR kuhlprednon OR lenisolon* OR "Lepi‐Cortinolo" OR "Linola‐H" OR longiprednil OR metacortandralon* OR "Meti Derm" OR meticortelon* OR opredsone Or precortisyl OR "Pred‐Clysma" OR predeltilon* OR prenilone OR hydrocortancyl OR "Solu Moderin" OR predonin* OR metypred OR prednisol OR dexamethason* OR "BB 1101" OR decadron OR hexadrol OR fortecortin OR dexameth OR dexone OR hexadecadrol OR desamethason* OR ozurdex OR deronil OR baycuten OR aacidexam OR spersadex OR dexacortal OR gammacorten OR visumetazon* OR adexone OR "Alba‐Dex" OR cortidexason OR decacort OR decadrol OR dectancyl OR desameton OR loverine OR millicorten OR orgadrone OR alin OR auxiloson OR cortisumman OR decalix OR decameth OR decasone OR dekacort OR deltafluorene OR "Dexa‐Mamallet" OR dexafluorene OR dexalocal OR dexamecortin OR dexamonozon OR dexapos OR dexinoral OR fluorodelta OR lokalison OR methylfluorprednisolon* OR mymethason* OR "Dexa‐Rhinosan" OR "Dexa‐Scheroson" OR "Dexa‐sine" OR dexacortin OR dexafarma OR dinormon OR baycadron OR "Aeroseb‐Dex" OR Maxidex OR Dextenza OR dexasone OR dexpak OR hydrocortison* OR cortisol OR cortef OR hydrocorton* OR cetacort OR barseb OR aeroseb OR "Cort‐Dome" OR cortenema OR cortril OR cortifan OR cortispray OR dermacort OR domolene OR eldecort OR hautosone OR "Heb‐Cort" OR hytone OR Komed OR Nutracort OR Proctocort OR Rectoid OR Hydrocort OR locoid OR Solu‐Glyc)

#2 TI=(COVID OR COVID19 OR "SARS‐CoV‐2" OR "SARS‐CoV2" OR SARSCoV2 OR "SARSCoV‐2" OR "SARS coronavirus 2" OR "2019 nCoV" OR "2019nCoV" OR "2019‐novel CoV" OR "nCov 2019" OR "nCov 19" OR "severe acute respiratory syndrome coronavirus 2" OR "novel coronavirus disease" OR "novel corona virus disease" OR "corona virus disease 2019" OR "coronavirus disease 2019" OR "novel coronavirus pneumonia" OR "novel corona virus pneumonia" OR "severe acute respiratory syndrome coronavirus 2") OR AB=(COVID OR COVID19 OR "SARS‐CoV‐2" OR "SARS‐CoV2" OR SARSCoV2 OR "SARSCoV‐2" OR "SARS coronavirus 2" OR "2019 nCoV" OR "2019nCoV" OR "2019‐novel CoV" OR "nCov 2019" OR "nCov 19" OR "severe acute respiratory syndrome coronavirus 2" OR "novel coronavirus disease" OR "novel corona virus disease" OR "corona virus disease 2019" OR "coronavirus disease 2019" OR "novel coronavirus pneumonia" OR "novel corona virus pneumonia" OR "severe acute respiratory syndrome coronavirus 2")

#3 #1 AND #2

#4 TI=(random* OR placebo OR trial OR groups OR "phase 3" or "phase3" or p3 or "pIII") OR AB=(random* OR placebo OR trial OR groups OR "phase 3" or "phase3" or p3 or "pIII")

#5 #3 AND #4

Indexes=SCI‐EXPANDED, ESCI Timespan=2020‐2021

WHO COVID‐19 Global literature on coronavirus disease

(corticosteroid* OR corticoid* OR prednis* OR hydrocorti* OR methylpredni* OR deltahydrocorti* OR dehydrocorti* OR dexameth* OR desameth*) AND (random* OR placebo OR trial OR groups OR "phase 3" or "phase3" or p3 or "pIII")

Appendix 2. Search strategies (update)

Cochrane COVID‐19 Study Register

Search string:

(corticosteroid* OR glucocorticoid* OR corticoid* OR cloprednol* OR methylpredni* OR metilpredni* OR predni* OR triamcinol* OR beclomet* OR betamet* OR dexamethas* OR dexametas* OR paramethas* OR parametas* OR prednyli* OR budesonid* OR deflazacort OR hydrocortis* OR cortison* OR fludrocortison*)

Study characteristics:
1) "Intervention assignment": “Randomised” OR “Quasi‐Randomised” OR 
2) "Study design": "Parallel/Crossover" OR "Unclear"
= 356 studies (621 references)


Web of Science Core Collection (Advanced search)

#1 TI=(corticosteroid* OR glucocorticoid* OR corticoid* OR cloprednol* OR methylpredni* OR metilpredni* OR predni* OR triamcinol* OR beclomet* OR betamet* OR dexamethas* OR dexametas* OR paramethas* OR parametas* OR prednyli* OR budesonid* OR deflazacort OR hydrocortis* OR cortison* OR fludrocortison*) OR AB=(corticosteroid* OR glucocorticoid* OR corticoid* OR cloprednol* OR methylpredni* OR metilpredni* OR predni* OR triamcinol* OR beclomet* OR betamet* OR dexamethas* OR dexametas* OR paramethas* OR parametas* OR prednyli* OR budesonid* OR deflazacort OR hydrocortis* OR cortison* OR fludrocortison*)

#2 TI=(COVID OR COVID19 OR "SARS‐CoV‐2" OR "SARS‐CoV2" OR SARSCoV2 OR "SARSCoV‐2" OR "SARS coronavirus 2" OR "2019 nCoV" OR "2019nCoV" OR "2019‐novel CoV" OR "nCov 2019" OR "nCov 19" OR "severe acute respiratory syndrome coronavirus 2" OR "novel coronavirus disease" OR "novel corona virus disease" OR "corona virus disease 2019" OR "coronavirus disease 2019" OR "novel coronavirus pneumonia" OR "novel corona virus pneumonia" OR "severe acute respiratory syndrome coronavirus 2") OR AB=(COVID OR COVID19 OR "SARS‐CoV‐2" OR "SARS‐CoV2" OR SARSCoV2 OR "SARSCoV‐2" OR "SARS coronavirus 2" OR "2019 nCoV" OR "2019nCoV" OR "2019‐novel CoV" OR "nCov 2019" OR "nCov 19" OR "severe acute respiratory syndrome coronavirus 2" OR "novel coronavirus disease" OR "novel corona virus disease" OR "corona virus disease 2019" OR "coronavirus disease 2019" OR "novel coronavirus pneumonia" OR "novel corona virus pneumonia" OR "severe acute respiratory syndrome coronavirus 2")

#3 #1 AND #2

#4 TI=(random* OR placebo OR trial OR groups OR "phase 3" or "phase3" or p3 or "pIII") OR AB=(random* OR placebo OR trial OR groups OR "phase 3" or "phase3" or p3 or "pIII")

#5 #3 AND #4
Indexes=SCI‐EXPANDED, ESCI Timespan=2020‐2022
= 967


WHO COVID‐19 Global literature on coronavirus disease

Title, abstract, subject: (corticosteroid* OR glucocorticoid* OR corticoid* OR cloprednol* OR methylpredni* OR metilpredni* OR predni* OR triamcinol* OR beclomet* OR betamet* OR dexamethas* OR dexametas* OR paramethas* OR parametas* OR prednyli* OR budesonid* OR deflazacort OR hydrocortis* OR cortison* OR fludrocortison*) AND (random* OR placebo OR trial OR groups OR "phase 3" or "phase3" or p3 or "pIII")
=1964

Data and analyses

Comparison 1. Systemic corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 All‐cause mortality up to 30 days	9	7898	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.84, 0.97]
1.2 All‐cause mortality up to 120 days	3	485	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.23, 2.34]
1.3 Clinical improvement: discharged alive	3	6786	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [1.03, 1.11]
1.4 Clinical worsening: new need for IMV or death	2	5586	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.84, 1.01]
1.5 Serious adverse events	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.6 Adverse events	3		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.7 Hospital‐acquired infections	4		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.8 Invasive fungal infections	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.9 Quality of life up to 120 days	1		Mean Difference (IV, Random, 95% CI)	Totals not selected
1.10 New need for dialysis	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.11 Viral clearance	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 2. Subgroup analysis: respiratory support for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 All‐cause mortality up to 30 days	4	7137	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.81, 1.03]
2.1.1 No oxygen	1	1535	Risk Ratio (M‐H, Random, 95% CI)	1.27 [1.00, 1.61]
2.1.2 Low‐flow oxygen only	2	85	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.50, 1.63]
2.1.3 NIV, high‐flow, and low‐flow oxygen combined	1	3883	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.79, 1.00]
2.1.4 NIV or high‐flow oxygen only	2	195	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.53, 1.58]
2.1.5 Invasive ventilation	3	1439	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.70, 1.04]

Comparison 3. Subgroup analysis: dexamethasone versus methylprednisolone versus hydrocortisone for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 All‐cause mortality up to 30 days	9	7898	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.84, 0.97]
3.1.1 Dexamethasone	3	6774	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.83, 0.98]
3.1.2 Methylprednisolone	3	566	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.77, 1.24]
3.1.3 Hydrocortisone	3	558	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.56, 1.21]
3.2 All‐cause mortality up to 120 days	3	485	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.23, 2.34]
3.2.1 Hydrocortisone	1	30	Risk Ratio (M‐H, Random, 95% CI)	2.04 [0.65, 6.43]
3.2.2 Methylprednisolone	2	455	Risk Ratio (M‐H, Random, 95% CI)	0.42 [0.06, 3.04]
3.3 Clinical improvement: discharged alive	3	6786	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [1.03, 1.11]
3.3.1 Dexamethasone	2	6724	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [1.03, 1.11]
3.3.2 Methylprednisolone	1	62	Risk Ratio (M‐H, Fixed, 95% CI)	1.65 [1.18, 2.29]

Comparison 4. Subgroup analysis: female versus male for the comparison of systemic corticosteroids versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 All‐cause mortality up to 30 days	3	6788	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.83, 0.98]
4.1.1 Female	3	2475	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.81, 1.10]
4.1.2 Male	3	4313	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.80, 0.97]

Comparison 5. Subgroup analysis: < 70 years versus ≥ 70 years for the comparison of systemic corticosteroids versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 All‐cause mortality up to 30 days	3	6781	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.71, 1.10]
5.1.1 < 70 years	3	3885	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.60, 0.82]
5.1.2 ≥ 70 years	3	2896	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.92, 1.12]

Comparison 6. Subgroup analysis: White versus Black, Asian or minority ethnic group versus unknown for the comparison of systemic corticosteroids versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 All‐cause mortality up to 30 days	2	6489	Risk Ratio (M‐H, Random, 95% CI)	0.79 [0.60, 1.04]
6.1.1 White	2	4753	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.87, 1.06]
6.1.2 Black, Asian or minority ethnic group	1	1147	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.52, 0.93]
6.1.3 Unknown	1	589	Risk Ratio (M‐H, Random, 95% CI)	0.60 [0.41, 0.88]

Comparison 7. Subgroup analysis: high‐income countries versus low‐ and middle‐income countries for the comparison of corticosteroids plus standard care versus standard care (plus/minus placebo) for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 All‐cause mortality up to 30 days	9	7898	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.84, 0.97]
7.1.1 High‐income countries	5	7047	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.81, 0.97]
7.1.2 Low‐ and middle‐income countries	4	851	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.82, 1.09]
7.2 All‐cause mortality up to 120 days	3	485	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.20, 2.75]
7.2.1 High‐income countries	1	30	Risk Ratio (M‐H, Random, 95% CI)	2.62 [0.63, 10.98]
7.2.2 Low‐ and middle‐income countries	2	455	Risk Ratio (M‐H, Random, 95% CI)	0.42 [0.06, 3.04]
7.3 Clinical improvement: discharged alive	3	6786	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [1.03, 1.11]
7.3.1 High‐income country	1	6425	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [1.02, 1.10]
7.3.2 Low‐ and middle‐income country	2	361	Risk Ratio (M‐H, Fixed, 95% CI)	1.65 [1.23, 2.21]

Comparison 8. Methylprednisolone versus dexamethasone for hospitalised individuals with a confirmed or suspected diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 All‐cause mortality up to 30 days	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 9. High‐dose dexamethasone (12 mg/d or higher) versus low‐dose dexamethasone (6 to 8 mg/d) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 All‐cause mortality up to 30 days	3	1269	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.73, 1.04]
9.2 All‐cause mortality up to 120 days	4	1383	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.79, 1.08]
9.3 Clinical improvement: discharged alive	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
9.4 Serious adverse events	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
9.5 Adverse events	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
9.6 Hospital‐acquired infections	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
9.7 Invasive fungal infections	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 10. Subgroup analysis: female versus male for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
10.1 All‐cause mortality up to 30 days	2	1069	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.72, 1.04]
10.1.1 Female	2	333	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.65, 1.22]
10.1.2 Male	2	736	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.62, 1.34]
10.2 All‐cause mortality up to 120 days	2	1066	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.74, 1.02]
10.2.1 Female	2	333	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.70, 1.23]
10.2.2 Male	2	733	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.70, 1.03]

Comparison 11. Subgroup analysis: < 70 years versus ≥ 70 years for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
11.1 All‐cause mortality up to 30 days	2	1069	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.73, 1.04]
11.1.1 <70 years	2	707	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.63, 1.07]
11.1.2 ≥ 70 years	2	362	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.73, 1.15]
11.2 All‐cause mortality up to 120 days	2	1066	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.76, 1.03]
11.2.1 <70 years	2	706	Risk Ratio (M‐H, Random, 95% CI)	0.80 [0.63, 1.02]
11.2.2 ≥ 70 years	2	360	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.78, 1.14]

Comparison 12. Subgroup analysis: high‐income countries versus low‐ and middle‐income countries for the comparison of high‐dose dexamethasone (12 mg or higher) versus low‐dose dexamethasone (6 mg to 8mg) for hospitalised individuals with a confirmed diagnosis of symptomatic COVID‐19.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
12.1 All‐cause mortality up to 30 days	3	1269	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.73, 1.04]
12.1.1 High‐income countries	2	1171	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.70, 1.02]
12.1.2 Low‐ and middle‐income countries	1	98	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.65, 1.71]
12.2 All‐cause mortality up to 120 days	4	1383	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.79, 1.08]
12.2.1 High‐income countries	2	1152	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.72, 1.01]
12.2.2 Low‐ and middle‐income countries	2	231	Risk Ratio (M‐H, Fixed, 95% CI)	1.35 [0.93, 1.94]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Angus 2020.

Study characteristics
Methods	Trial design: multicentre, open‐label, platform RCT Type of publication: journal publication Setting: inpatient Recruitment dates: 9 March to 17 June 2020 Country: Australia, Canada, France, Ireland, the Netherlands, New Zealand, UK, USA Language: English Number of centres: 121 clinical sites Trial registration number: NCT02735707 Date first posted: 13 April 2016
Participants	Age: mean age of 60.4 years (SD 11.6) in the fixed‐dose intervention group; 59.5 years (SD 12.7) in the shock‐dependent intervention group; 59.9 years (SD 14.6) in the control group. Gender 98 male (71.5%) and 39 female (28.5%) in the fixed‐dose intervention group; 103 male (70.6%) and 43 female (29.5%) in the shock‐dependent intervention group; 72 male (71.3%) and 29 female (28.7%) in the control group. Proportion of confirmed infections Positive: 81.3% fixed‐dose intervention arm 69.6%; shock‐dependent intervention; 79% in the control arm Negative: not reported Unclear: not reported Ethnicity White: 71.2% in fixed‐dose intervention group; 76.2% in the shock‐dependent intervention group; 57% in the control group Asian: 16.2% in fixed‐dose intervention group; 10.5% in the shock‐dependent intervention group; 27.9% in the control group Black: 3.6% in fixed‐dose intervention group; 6.7% in the shock‐dependent intervention group; 5.1% in the control group Mixed: 3.6% in fixed‐dose intervention group; 0% in the shock‐dependent intervention group; 2.5% in the control group Other: 5.4% in fixed‐dose intervention group; 6.7% in the shock‐dependent intervention group; 7.6% in the control group Number of participants (recruited/allocated/evaluated) Recruited: 614 Allocated: 143 fixed‐dose intervention group; 152 shock‐dependent intervention group; 108 control group Evaluated: 137 fixed‐dose intervention group; 141 shock‐dependent group; 101 control group Severity of condition according to study definition Fixed‐dose intervention group None/supplemental oxygen only: 0% HFNC: 12.4% IV only: 24.1% IMV: 63.5% ECMO: 0.7% Vasopressor support: 40.9% Shock‐dependent intervention group None/supplemental oxygen only: 0.7% HFNC: 15.8% NIV only: 33.6% IMV: 50% ECMO: 0% Vasopressor support: 32.2% Control group None/supplemental oxygen only: 0% HFNC: 15.8% NIV only: 31.7% IMV: 52.5% ECMO: 2.0% Vasopressor support: 29.7% Severity of condition according to WHO score: severe ≥ 6 Co‐morbidities: diabetes, respiratory disease, kidney disease, severe cardiovascular disease, immunosuppressive disease Inclusion criteria Adult patient admitted to hospital with acute illness due to suspected or proven pandemic (COVID‐19) infection Severe disease state, defined by receiving respiratory or cardiovascular organ failure support in an ICU Exclusion criteria Death is deemed to be imminent and inevitable during the next 24 h AND one or more of the patient, substitute decision maker, or attending physician are not committed to full active treatment Patient is expected to be discharged from hospital today or tomorrow > 14 days have elapsed while admitted to hospital with symptoms of an acute illness due to suspected or proven pandemic infection Previous participation in this REMAP within the last 90 days Known hypersensitivity to hydrocortisone Intention to prescribe systemic corticosteroids for a reason that is unrelated to the current episode of CAP/COVID‐19 (or direct complications of CAP/COVID‐19), such as chronic corticosteroid use before admission, acute severe asthma, or suspected or proven Pneumocystis jiroveci pneumonia > 36 h have elapsed since ICU admission (noting that this may be operationalised as > 24 h has elapsed since commencement of sustained organ failure support) Patient has been randomised in a trial evaluating corticosteroids, where the protocol of that trial requires ongoing administration of study drug The treating clinician believes that participation in the domain would not be in the best interests of the patient Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): yes
Interventions	Treatment details of intervention group (e.g dose, route of administration, number of doses) Type of corticosteroid: hydrocortisone Dose: Fixed‐dose: 50 mg every 6 h for 7 days Shock‐dependent: 50 mg every 6 h for up to 28 days if in shock Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses) No hydrocortisone Concomitant therapy (e.g. description of standard care): not reported Duration of follow‐up: follow‐up ended 12 August 2020 Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: respiratory and cardiovascular organ support‐free days up to day 21, with subcomponents in‐hospital deaths and organ support‐free days among survivors Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): reported Serious adverse events, defined as the number of participants with any event: reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: not reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: time to death, cardiovascular organ support–free days, length of ICU stay, WHO scale at day 14
Identification
Notes	Date of publication: 2 September 2020 Sponsor/funding: Platform for European Preparedness Against (Re‐) emerging Epidemics (PREPARE) consortium by the European Union, FP7‐HEALTH‐2013‐INNOVATION‐1 (grant 602525), the Australian National Health and Medical Research Council (grant APP1101719), the New Zealand Health Research Council (grant 16/ 631), the Canadian Institute of Health Research Strategy for Patient‐Oriented Research Innovative Clinical Trials Program (grant 158584), the UK National Institute for Health Research (NIHR) and the NIHR Imperial Biomedical Research Centre, the Health Research Board of Ireland (grant CTN 2014‐012), the UPMC Learning While Doing Program, the Breast Cancer Research Foundation, the French Ministry of Health (grant PHRC‐20‐0147), and the Minderoo Foundation Risk of bias table presents two entries for analysis 1.1 All‐cause mortality and analysis 1.5 Serious adverse events: one entry for fixed‐dose arm and one for shock‐dependent arm

Corral‐Gudino 2021.

Study characteristics
Methods	Trial design: multicentre, open‐label, RCT Type of publication: journal publication Setting: inpatient Recruitment dates: April to May 2020 Country: Spain Language: English Number of centres: 5 hospitals Trial registration number: EUCTR 2020‐001934‐37 Date of trial registration: 8 May 2020
Participants	Age: mean age of: 73 years (SD +/‐ 11) in the intervention group 66 years (SD +/‐ 12) in the control group Gender: 23 male (66%) in the intervention group 16 male (55%) in the control group Proportion of confirmed infections: PCR positivity inclusion criterion Ethnicity: not reported Number of participants (recruited/allocated/evaluated): Recruited: 86 Allocated: 35 intervention group and 29 control group Evaluated: 35 intervention group and 29 control group Severity of condition according to study definition PaO2/FiO2 or PaFi < 300 SaO2/FiO2 or SaFi < 400 or At least 2 criteria of the BRESCIA‐COVID Respiratory Severity Scale (BCRSS) Severity of condition according to WHO score: moderate to severe 5 to 6 Co‐morbidities: hypertension, cardiac disease, respiratory disease, diabetes Inclusion criteria Confirmed SARS‐CoV‐2 infection Symptom duration of at least 7 days Radiological evidence of lung disease on chest X‐ray or CT scan Moderate to severe disease with abnormal gas exchange: PaO2/FiO2 or PaFi < 300 SaO2/FiO2 or SaFi < 400 or At least 2 criteria of the BRESCIA‐COVID Respiratory Severity Scale (BCRSS) Exclusion criteria: Mechanical ventilation Hospitalised in the ICU Treated with corticosteroids or immunosuppressive drugs Chronic kidney disease on dialysis Pregnant Previous treatments: not reported
Interventions	Treatment details of intervention group (e.g dose, route of administration, number of doses) Type of corticosteroid: methylprednisolone Dose: 40 mg for 3 days and then 20 mg for 3 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses): standard care Concomitant therapy (e.g. description of standard care): standard care included acetaminophen, oxygen therapy, low molecular weight heparin and antibiotics; azithromycin, hydroxychloroquine and lopinavir plus ritonavir Duration of follow‐up: until hospital discharge or day 28 after inclusion Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: composite endpoint including in‐hospital all‐cause mortality, escalation to ICU admission or progression of respiratory insufficiency that required noninvasive ventilation Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death, that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: composite endpoint included in‐hospital all‐cause mortality, escalation to ICU admission, or progression of respiratory insufficiency that required noninvasive ventilation
Identification
Notes	Date of publication: 3 February 2021 Sponsor/funding: IDIVAL Instituto de Investigación Sanitaria Valdecilla

Dequin 2020.

Study characteristics
Methods	Trial design: multicentre, double‐blind, randomised trial Type of publication: journal publication Setting: inpatient Recruitment dates: 7 March to 1 June 2020 Country: France Language: English Number of centres: 9 Trial registration number: NCT02517489 Date first posted: 7 August 2015
Participants	Age: Median age 63.1 years (IQR 51.5 to 70.8) in the intervention group Median age 66.3 years (IQR 53.5 to 72.7) in the control group Gender: 54 male (71.1%) and 22 female (28.9%) in the intervention group 50 male (68.5%) and 23 female (31.5%) in the control group Proportion of PCR‐confirmed infections Positive: 94.7% intervention arm; 98.8% control arm Negative: not reported Unclear: not reported Ethnicity: not reported Number of participants (recruited/allocated/evaluated): Recruited: 40 Allocated: 76 intervention group and 73 control group Evaluated: 76 in intervention group and 73 in control group Severity of condition according to study definition Mechanical ventilation (included noninvasive ventilation): 81.6% in the intervention group 59% in the control group High‐flow oxygen therapy: 13.2% in the intervention group 12.3% in the control group Nonrebreathing mask with reservoir bag: 5.3%% in the intervention group 6.8% in the control group Severity of condition according to WHO score: moderate to severe ≥ 5 Co‐morbidities: diabetes, COPD/asthma, immunosuppression Inclusion criteria Admitted to 1 of the 9 participating French ICUs for acute respiratory failure At least 18 years of age Biologically confirmed (RT‐PCR) or suspected (suggestive chest CT scan result in the absence of any other cause of pneumonia) COVID‐19 1 of 4 severity criteria had to be present: Need for mechanical ventilation with a PEEP of ≥ 5 cm H2O A ratio of PaO2 to FIo2 < 300 on high‐flow oxygen therapy with an FIO2 value of at least 50% For participants receiving oxygen through a reservoir mask, a PaO2:FIO2 ratio < 300, estimated using prespecified charts, or Pulmonary Severity Index > 130 Exclusion criteria Unable to meet inclusion deadlines Included in another interventional trial Septic shock Long‐term corticosteroid therapy Did not meet severity criteria Transferred to another ICU Medical team declined enrolment Under judicial protection Do‐not‐intubate order Moribund Declined to participate Required hydrocortisone for other medical condition Miscellaneous Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): yes
Interventions	Treatment details of intervention group (e.g dose, route of administration, number of doses): Type of corticosteroid: hydrocortisone Dose: 200 mg/day until day 7, then 100 mg/day for 4 days and 50 mg/day for 3 days, for a total of 14 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses): placebo Concomitant therapy (e.g. description of standard care): not reported Duration of follow‐up: last follow‐up on 29 June 2020 Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: treatment failure on day 21 (death or persistent dependence of mechanical ventilation or high‐flow oxygen therapy) Review outcomes: Inpatient setting: Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: reported Adverse events (any grade), defined as the number of participants with any event: reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: endotracheal intubation (for patients noninvasively ventilated at inclusion), prone position, ECMO, inhaled nitric oxide
Identification
Notes	Date of publication: 6 October 2020 Sponsor/funding: University Hospital, Tours

Edalatifard 2020.

Study characteristics
Methods	Trial design: RCT Type of publication: journal publication Setting: inpatient Recruitment dates: 28 March to 28 May 2020 (study register entry) Country: Iran Language: English Number of centres: 4 Trial registration number: IRCT20200404046947N1 Date of trial registration: 15 April 2020
Participants	Age Mean age 55.8 ± 16.35 years in the intervention group Mean age 61.7 ± 16.62 years in the control group Gender: 4 male (70.6%) and 10 female (29.4%) in the intervention group 15 male (53.6%) and 13 female (46.4%) in the control group Proportion of confirmed infections: PCR positivity inclusion criterion Ethnicity: not reported Number of participants (recruited/allocated/evaluated) Recruited: 151 recruited Allocated: 34 in the intervention group and 34 in the control group Evaluated: 34 in the intervention group and 28 in the control group Severity of condition according to study definition Nasal cannula: 11.8% in the intervention group 32.1% in the control group Simple mask: 14.7% in the intervention group 7.1% in the control group Reserve mask: 35.3% in the intervention group 21.4% in the control group NIV: 38.2% in the intervention group 35.7% in the control group Severity of condition according to WHO score: moderate to severe 5 to 6 Co‐morbidities: diabetes, hypothyroidism, cancer, respiratory disorder, renal disorder, cardiovascular disorder, hypertension, autoimmune, and neurodegenerative diseases Inclusion criteria Aged ≥ 18 years Confirmed COVID‐19 (RT‐PCR) with blood oxygen saturation < 90%, elevated C‐reactive protein (CRP > 10), and interleukin (IL)‐6 (> 6) at the early pulmonary phase of disease before connecting to the ventilator and intubation Agreed to give informed consent Exclusion criteria Patients were intolerant or allergic to any therapeutic agents used in this research Pregnant or lactating women Patients with blood oxygen saturation < 75%, positive pro‐calcitonin (PCT) and troponin test, ARDS, uncontrolled hypertension (HTN), uncontrolled diabetes mellitus (DM), gastrointestinal problems or gastrointestinal bleeding (GIB) history, heart failure (HF), active malignancies, and received any immune‐suppressor agents Previous treatments: not reported
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses): Type of corticosteroid: methylprednisolone Dose: 250 mg/day for 3 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses): standard care Concomitant therapy (e.g. description of standard care): standard of care included hydroxychloroquine sulphate, lopinavir and naproxen Duration of follow‐up: 3 days Treatment cross‐overs: no Compliance with assigned treatment: 6 patients in the control group received the intervention drug
Outcomes	Primary study outcome: time to event (discharge or death), time to improvement Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: not reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: blood SO2 level, BORG score, heart rate, temperature, respiratory rate
Identification
Notes	Date of publication: 7 September 2020 Sponsor/funding: Tehran University of Medical Sciences

Farahani 2021.

Study characteristics
Methods	Trial design: open‐label, single‐centre RCT Type of publication: preprint Setting: inpatient Recruitment dates: 30 March to 18 May 2020 (only estimated dates from registry entry) Country: Iran Language: English Number of centres: 1 Trial registration number: IRCT20200406046963N1 Date of trial registration: 22 April 2020
Participants	Age (mean, SD): intervention group: 61.07 ± 12.83, control group: 66.80 ± 14.03 Gender (female, n (%)): intervention group: 4 (28,6), control group: 6 (40) Proportion of confirmed infections: PCR positivity inclusion criterion Ethnicity: no ethnicities excluded Number of participants (recruited/allocated/evaluated): 14 intervention group, 15 control group Severity of condition according to study definition: Moderate to severe COVID‐19 admitted to ICU PaO2/FiO2 < 300 Progression of disease severity and not responding to standard treatment Prediction of intubation for next 24 h Severity of condition according to WHO score: moderate‐severe 5 to 6 Co‐morbidities: not reported Inclusion criteria Confirmed SARS‐CoV‐2 infection Moderate‐severe COVID‐19 Admitted to ICU PaO2/FiO2 < 300 Progression of disease severity and not responding to standard treatment Prediction of intubation for next 24 h Exclusion criteria Uncontrolled diabetes mellitus Active GI bleeding History of corticosteroid hypersensitivity Severe electrolyte imbalances Procalcitonin > 0.5 active bacterial Viral (HIV, hepatitis) and fungal infection Previous treatments: not specified
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses): 1000 mg methylprednisolone IV for 3 days followed by 1 mg/kg oral prednisolone with dose tapering for 7 days + standard care Treatment details of control group (e.g. dose, route of administration, number of doses): standard care Concomitant therapy (e.g. description of standard care): Kaletra (lopinavir/ritonavir) daily Hydroxychloroquine 400 mg daily Azithromycin 500 mg daily Duration of follow‐up: not specified Treatment cross‐overs: none reported Compliance with assigned treatment: no deviations reported
Outcomes	Primary study outcome: mortality rate, blood O2 saturation and need for further oxygen therapy Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: not reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: not reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes Glasgow Coma Scale: daily for 10 days mean, although scale is not metric Means of SpO2, FiO2, blood pressure, PEEP, CPK, LDH for 10 days
Identification
Notes	Date of publication: 9 September 2020 Sponsor/funding: Artesh University of Medical Sciences

Horby 2021.

Study characteristics
Methods	Trial design: open‐label RCT Type of publication: journal publication Setting: inpatient Recruitment dates: recruitment ended on 8 June 2020 Country: UK Language: English Number of centres: 176 Trial registration number: NCT04381936 Date of trial registration: 11 May 2020
Participants	Age: mean age 66.9 ± 15.4 years in the intervention group 65.8 ± 15.8 years in the control group Gender 1338 male (64%) and 766 female (36%) in the intervention group 2749 male (64%) and 1572 female (36%) in the control group Proportion of PCR test results Positive: 89% intervention arm, 90% control arm Negative: 11% intervention arm, 10% control arm Unclear: 1% intervention arm, < 1% control arm Ethnicity: not reported Number of participants (recruited/allocated/evaluated): Recruited: 11,303 Allocated: 2104 in the intervention group and 4321 in the control group Evaluated: 2104 in the intervention group and 4321 in the control group Severity of condition according to study definition No oxygen: 501 (24%) intervention group; 1034 (24%) control group Oxygen only: 1279 (61%) intervention group; 2604 (60%) control group IMV: 324 (15%) intervention group; 683 (16%) control group Severity of condition according to WHO score: moderate to severe 4 to 9 Co‐morbidities: diabetes, heart disease, chronic lung disease, tuberculosis, HIV infection, severe liver disease, severe kidney impairment Inclusion criteria Clinically suspected or laboratory confirmed SARS‐CoV‐2 infection Hospitalised patients: no medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial Exclusion criteria If the attending clinician believes that there is a specific contra‐indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial Previous treatments: not reported
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroid: dexamethasone Dose: 6 mg once daily for up to 10 days (or until hospital discharge if sooner) Route of administration: IV or oral Treatment details of control group (e.g. dose, route of administration, number of doses) Standard care Concomitant therapy (e.g. description of standard care): none Duration of follow‐up: until discharge or death, or 28 days after randomisation Treatment cross‐overs: no Compliance with assigned treatment: 8% in the control group received intervention drug
Outcomes	Primary study outcome: 28‐day mortality Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: not reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: composite outcome IMV or death
Identification
Notes	Date of publication: 17 July 2020 Sponsor/funding: University of Oxford

Jamaati 2021.

Study characteristics
Methods	Trial design: RCT Type of publication: journal publication Setting: inpatient Recruitment dates: March 2020 Country: Iran Language: English Number of centres: 1 Trial registration number: IRCT20151227025726N17 Date of trial registration: 31 May 2020
Participants	Age: median age Intervention group survivor: 54 years (IQR 37 to 63) Intervention group non‐survivor: 63 years (IQR 55.5 to 72.5) Control group survivor: 61.5 years (IQR 54 to 62) Control group non‐survivor: 67 years (IQR 48 to 73) Gender (male, n(%)): Intervention group: 18 (72) Control group: 18 (72) Proportion of confirmed infections: PCR positivity inclusion criterion Ethnicity: not reported Number of participants (recruited/allocated/evaluated): Recruited: no information Allocated: 25 intervention group and 25 control group Evaluated: 25 intervention group and 25 control group Severity of condition according to study definition: PaO2/FiO2 between 100 and 300 mmHg Severity of condition according to WHO score: most likely 5, no invasive ventilation at randomisation Co‐morbidities: diabetes, hypertension, cardiovascular disease Inclusion criteria Age > 18 years SARS‐CoV‐2 infection confirmed by RT‐PCR PaO2/FiO2 between 100 and 300 mmHg Bilateral lung infiltration Provision of written informed consent by the patient Exclusion criteria Patients with chronic kidney diseases Patients with chronic liver diseases Patients with hyperglycaemia Pregnant or breastfeeding women Previous treatments: not reported
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroid: dexamethasone Dose: 20 mg/day from day 1 to 5 and then at 10 mg/day from day 6 to 10 Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses) Standard care Concomitant therapy (e.g. description of standard care): oxygen support, fluid support, lopinavir/ritonavir (200/50 mg, 2 tablets twice a day) Duration of follow‐up: 28 days Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: need for IMV, death rate Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: not reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: duration of clinical improvement, radiological changes in the CT scan
Identification
Notes	Date of publication: 16 February 2021 Sponsor/funding: Shahid Beheshti University of Medical Sciences

Jeronimo 2020.

Study characteristics
Methods	Trial design: double‐blind RCT Type of publication: journal publication Setting: inpatient Recruitment dates: 18 April to 16 June 2020 Country: Brazil Language: English Number of centres: 1 Trial registration number: NCT04343729 Date of trial registration: 13 April 2020
Participants	Age: mean age 54 years (SD 15) in the intervention group 57 years (SD 15) in the control group Gender 68 female (35.1%) in the intervention group 71 female (35.7%) in the control group Proportion of PCR test results Positive: 83.4% intervention arm, 79.3% control arm Negative: not reported Unclear: not reported Ethnicity: white, black, admixed, Asian, Amerindian Number of participants (recruited/allocated/evaluated) Recruited: 425 Allocated: 209 intervention group and 207 in the control group Evaluated: 195 intervention group and 202 control group Severity of condition according to study definition IMV at baseline: 66 (34%) intervention group; 67 (33.7%) control group Non‐invasive oxygen therapy at baseline: 98 (50.5%) intervention group; (45.2%) 90 control group Severity of condition according to WHO score: moderate to severe: 5 to 9 Co‐morbidities: diabetes, hypertension, alcohol use disorder, heart disease, asthma, rheumatic disease, liver disease, previous tuberculosis, COPD Inclusion criteria Clinical and/or radiological suspicion of COVID‐19 (history of fever and any respiratory symptom; eg, cough or dyspnoea and/or ground glass opacity or pulmonary consolidation on CT scan) Aged ≥ 18 years Either had SpO2 ≤ 94% with room air, required supplementary oxygen, or required IMV Exclusion criteria History of hypersensitivity to methylprednisolone Living with HIV or AIDS Had a history of chronic use of corticosteroids or immunosuppressive agents Were pregnant or breastfeeding Had decompensated cirrhosis or chronic renal failure Previous treatments: not reported
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses): Type of corticosteroid: methylprednisolone Dose: 0.5 mg/kg twice daily for 5 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses): placebo Concomitant therapy (e.g. description of standard care): all patients meeting ARDS criteria used pre‐emptive IV ceftriaxone (1 g twice a day for 7 days) plus azithromycin (500 mg once a day for 5 days) or clarithromycin (500 mg twice a day for 7 days), starting on day 1 Duration of follow‐up: 28 days Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: 28‐day mortality Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: not reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: reported Additional study outcomes: none
Identification
Notes	Date of publication: 12 August 2020 Sponsor/funding: Fundação de Medicina Tropical Dr. Heitor Vieira Dourado

Maskin 2021.

Study characteristics
Methods	Trial design: multicentre, open‐label, randomised controlled trial Type of publication: journal publication Setting: inpatient (ICU) Recruitment dates: 17 June 2020 and 27 March 2021 Country: Argentina Language: English Number of centres: 5 Trial registration number: NCT04395105 Date of trial registration: 16 May 2020
Participants	Age: Mean age (SD): 60.04 years (± 13.08) in the low‐dose group Mean age (SD): 63.57 years (± 13.59) in the high‐dose group Gender: 16 female (33%) in the low‐dose group 13 female (26%) in the high‐dose group Ethnicity: not stated Number of participants (recruited/allocated/evaluated): 100/49 in the high‐dose group and 51 in the low‐dose group/49 in the high‐dose group and 49 in the low‐dose group Severity of condition according to study definition: receiving mechanical ventilation Severity of condition according to WHO score: severe ≥ 7 Co‐morbidities: Charlson’s comorbidity index is used, otherwise not stated Inclusion criteria: aged 18 years old or more, who had ARDS according to the Berlin Definition criteria, who had confirmed SARS‐CoV‐2 infection by reverse transcription polymerase chain reaction and were receiving mechanical ventilation for less than 72 hours Exclusion criteria: pregnant or breastfeeding women, terminal disease, therapeutic limitation, severe immunosuppression, chronic treatment with glucocorticoids, participation in another randomised clinical trial, prior use of dexamethasone for COVID‐19 (> 5 days), or consent refusal Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): mechanical ventilation
Interventions	Details of intervention: dexamethasone + standard care Dose: 16 mg once daily for 5 days, followed by 8 mg administered once daily for additional 5 days Route of administration: intravenous Treatment details of control group (e.g. dose, route of administration): 6 mg of dexamethasone intravenously per day for 10 days + standard care Concomitant therapy: —
Outcomes	Primary study outcome: ventilator‐free days during the first 28 days (defined as the number of days alive and free from mechanical ventilation up to the 28th day from randomisation), co‐primary outcome: the time to complete and successful discontinuation of mechanical ventilation or death Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: the daily value of glucose and insulin dose, muscle strength score, and the frequency of delirium within 28 days of randomisation
Identification
Notes	Date of publication: 29 November 2021 Sponsor/funding: —

Munch 2021a.

Study characteristics
Methods	Trial design: multicentre, triple‐blind, randomised controlled trial Type of publication: journal publication Setting: inpatient (ICU) Recruitment dates: 15 April 2020; paused on 16 June 2020; and terminated early on 4 September 2020 Country: Denmark Language: English Number of centres: 12 Trial registration number: NCT04348305, EudraCT 2020‐001395‐15 Date of trial registration: 11 April 2020
Participants	Age: Median age (IQR): 59 years (52 to 74) in the intervention group Median age (IQR): 62 years (55 to 71) in the control group Gender: 14 male (88%) in the intervention group 10 male (71%) in the control group Ethnicity: not stated Number of participants (recruited/allocated/evaluated): 30/16 in the intervention group and 14 in the control group/16 in the intervention group and 14 in the control group Severity of condition according to study definition: use of either invasive mechanical ventilation, non‐invasive ventilation, or continuous use of continuous positive airway pressure (CPAP) for hypoxia, or oxygen supplementation with an oxygen flow of at least 10 L/min Severity of condition according to WHO score: severe ≥ 5 Co‐morbidities: ischaemic heart disease or heart failure, hypertension, chronic pulmonary disease, diabetes mellitus Inclusion criteria: (18 years or above) with confirmed SARS‐CoV‐2 infection and severe hypoxia defined as use of either invasive mechanical ventilation, non‐invasive ventilation, or continuous use of continuous positive airway pressure (CPAP) for hypoxia, or oxygen supplementation with an oxygen flow of at least 10 L/min independent of delivery system Exclusion criteria: use of systemic corticosteroids, invasive mechanical ventilation for more than 48 hours prior to screening, invasive fungal infection, fertile women (< 60 years of age) with a positive urine or plasma human gonadotropin (hCG), known hypersensitivity to hydrocortisone, a patient for whom the clinical team had decided not to use invasive mechanical ventilation, previously randomised into the COVID STEROID trial, informed consent unobtainable Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): invasive mechanical ventilation, non‐invasive ventilation, or continuous use of continuous positive airway pressure (CPAP) for hypoxia, or oxygen supplementation with an oxygen flow of at least 10 L/min
Interventions	Details of intervention: hydrocortisone Dose: 200 mg per day, for seven days or until hospital discharge (whichever is first) Route of administration: intravenous, as 24‐h infusion or as a bolus Treatment details of control group (e.g. dose, route of administration): placebo, continuous infusion over 24 h or as bolus injections every 6 hours Concomitant therapy: —
Outcomes	Primary study outcome: days alive without the use of life support (i.e. invasive mechanical ventilation, circulatory support, or renal replacement therapy, including days in between intermittent renal replacement therapy) at day 28 Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: not reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: clinically important gastrointestinal bleeding, or anaphylactic reaction; days alive without life support at day 90; days alive and out of hospital at day 90
Identification
Notes	Date of publication: 17 June 2021 Sponsor/funding: Novo Nordisk Foundation

Munch 2021b.

Study characteristics
Methods	Trial design: multicentre, triple‐blind, randomised controlled trial Type of publication: journal publication Setting: inpatient (ICU) Recruitment dates: August 2020 and May 2021 Country: Denmark, India, Sweden, Switzerland Language: English Number of centres: 31 Trial registration number: NCT04509973, CTRI/2020/10/028731 Date of trial registration: 11 August 2020
Participants	Age: Median age (IQR): 65 years (56 to 74) in the high‐dose group Median age (IQR): 64 years (54 to 72) in the low‐dose group Gender: 346 male (70%) in the high‐dose group, 331 male (68%) in the low‐dose group 151 female (30%) in the high‐dose group, 154 female (32%) in the low‐dose group Ethnicity: not stated, but 369 participants were recruited in India, an LMIC Number of participants (recruited/allocated/evaluated): 1000/503 in the high‐dose group and 497 in the low‐dose group/491 in the high‐dose group and 480 in the low‐dose group Severity of condition according to study definition: required supplementary oxygen at a flow rate of at least 10 L/min, noninvasive ventilation or continuous positive airway pressure for hypoxaemia, or invasive mechanical ventilation Severity of condition according to WHO score: moderate to severe ≥ 5 Co‐morbidities: diabetes, ischaemic heart disease or heart failure, chronic obstructive pulmonary disease Inclusion criteria: 18 years or older, hospitalised with confirmed SARS‐CoV‐2 infection, and required (1) supplementary oxygen at a flow rate of at least 10 L/min (independent of delivery system), (2) noninvasive ventilation or continuous positive airway pressure for hypoxaemia, or (3) invasive mechanical ventilation Exclusion criteria: (1) were treated with systemic glucocorticoids in doses higher than 6 mg of dexamethasone equivalents for indications other than COVID‐19 or had been treated with systemic glucocorticoids for COVID‐19 for 5 days or longer, (2) had invasive fungal infection or active tuberculosis, (3) had known hypersensitivity to dexamethasone, and (4) were pregnant Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): dexamethasone, remdesivir, convalescent plasma, systemic antibacterial agents, vasopressors or inotropes, IL‐6 receptor antagonists, janus kinase inhibitors
Interventions	Details of intervention: dexamethasone Dose: 12 mg/day for up to 10 days Route of administration: intravenous Treatment details of control group (e.g. dose, route of administration): 6 mg/day intravenous dexamethasone for up to 10 days Concomitant therapy: no
Outcomes	Primary study outcome: the number of days alive without life support (invasive mechanical ventilation, circulatory support, or kidney replacement therapy) at 28 days after randomisation Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: the number of days alive without life support at 90 days, the number of days alive out of the hospital at 90 days, health‐related quality of life measured using the 5‐dimension, 5‐level European Quality of Life questionnaire and the European Quality of Life visual analogue scale, will be assessed at 180 days after randomisation
Identification
Notes	Date of publication: 21 October 2021 Sponsor/funding: Novo Nordisk Foundation

Ranjbar 2021.

Study characteristics
Methods	Trial design: triple‐blind RCT Type of publication: preprint Setting: inpatient Recruitment dates: 10 August 2020 to 15 November 2020 Country: Iran Language: English Number of centres: 1 Trial registration number: IRCT20200204046369N1 Date of trial registration: 8 April 2020
Participants	Age: mean age 56.2 years (SD ± 17.5) in the intervention group 61.3 years (SD ± 17.3) in the control group Gender 27 male (61.4%) and 17 female (38.6%) in the intervention group 22 male (52.4%) and 20 female (47.6%) in the control group Proportion of confirmed infections: PCR positivity inclusion criterion Ethnicity: not reported Number of participants (recruited/allocated/evaluated): Recruited: 86 Allocated: 44 in the intervention group and 42 in the control group Evaluated: 44 in the intervention group and 42 in the control group Severity of condition according to study definition: patients with SpO2 < 92 in room air Severity of condition according to WHO score: moderate 4 to 5 Co‐morbidities: diabetes, cardiovascular disease, hypertension, renal diseases, liver diseases Inclusion criteria Age > 18 years Confirmed SARS‐CoV‐2 infection Hospitalised SpO2 saturation < 92 in room air Exclusion criteria: Pregnancy Uncontrolled diabetes mellitus Uncontrolled hypertension Previously been treated with steroids Any contraindication of steroid administration Immunodeficiency disorders SpO2 > 92 in room air Previous treatments: not reported
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroid: methylprednisolone Dose: 2 mg/kg daily infused over 60 min, tapered to half dosage every 5 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses): 6 mg of dexamethasone IV daily for 10 days Concomitant therapy (e.g. description of standard care): no Duration of follow‐up: 28 days Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: all‐cause mortality in 28 days, clinical status after 5 and 10 days after enrolment with 9‐point WHO scale Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: not reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: intubation, admission to ICU, hospital death during 28 days after enrolment
Identification
Notes	Date of publication: 1 February 2021 Sponsor/funding: Shiraz University of Medical Sciences Dr. Mohsen Moghadami

Taboada 2021.

Study characteristics
Methods	Trial design: open‐label, randomised controlled trial Type of publication: journal publication Setting: inpatient Recruitment dates: 15 January to 26 May 2021 Country: Spain Language: English Number of centres: 1 Trial registration number: NCT04726098, EUCTR2020‐ 005702‐25 Date first posted: 27 January 2021
Participants	Age: mean age 64.8 years (SD ± 14.1) in the low‐dose dexamethasone group 63.9 years (SD ± 14.5) in the high‐dose dexamethasone group Gender: 61 male (60.4%) in the low‐dose dexamethasone group 62 male (63.3%) in the high‐dose dexamethasone group Proportion of confirmed infections: confirmed SARS‐CoV‐2 infection by nasopharyngeal swab polymerase chain reaction was inclusion criterion Ethnicity: not stated Number of participants (recruited/allocated/evaluated) Recruited: 681 Allocated: 98 high‐dose dexamethasone group, 102 in the low‐dose dexamethasone group Evaluated: 98 high‐dose dexamethasone group, 102 in the low‐dose dexamethasone group Severity of condition according to study definition: receiving supplemental oxygen in order to maintain an oxygen saturation greater than 92% Severity of condition according to WHO score: moderate to severe ≥ 5 Co‐morbidities: hypertension, hyperlipidaemia, obesity, diabetes, chronic pulmonary disease, asthma, cardiovascular disease, history of cancer, chronic kidney disease Inclusion criteria: At least 18 years old Had SARS‐CoV‐2 infection confirmed by nasopharyngeal swab polymerase chain reaction Were receiving supplemental oxygen in order to maintain an oxygen saturation greater than 92% (Level 4 using the World Health Organization 7‐point Ordinal Scale for clinical improvement (WHO‐CIS). Scores on the 7‐point ordinal scale WHO‐CIS were defined as follows: (1) Not hospitalised; (2) Hospitalised, not requiring supplemental oxygen, no longer requires ongoing medical care (independent); (3) Hospitalised, not requiring supplemental oxygen, but in need of ongoing medical care (COVID‐19 related or otherwise); (4) Hospitalised, requiring supplemental oxygen; (5) Hospitalised, requiring non‐invasive ventilation or high flow nasal cannula; (6) Hospitalised, requiring ICU admission and invasive mechanical ventilation or extracorporeal membrane oxygenation (ECMO); (7) Death Exclusion criteria: Pregnancy or active lactation Known history of dexamethasone allergy or known contraindication to the use of corticosteroids Indication for corticosteroids use for other clinical conditions (e.g. refractory septic shock) Daily use of oral or intravenous corticosteroids in the past 15 days Expected death within the next 48 hours Different level of 4 using the 7‐point ordinal scale WHO‐CIS, need of supplemental oxygen with fraction of inspired oxygen > 0.5 in order to maintain an oxygen saturation greater than 92% Consent refusal for participating in the trial Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): ACE inhibitors, antihypertensive, anticoagulants, antiplatelet, inhaled corticosteroids, statins, immunosuppressants, insulin
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroid: dexamethasone Dose: 20 mg once daily for 5 days, followed by 10 mg once daily for additional 5 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses): dexamethasone, IV, 6 mg once daily for 10 days Concomitant therapy (e.g. description of standard care): no Duration of follow‐up: 28 days Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: clinical worsening within 11 days since randomisation, defined as worsening of the patient's condition during treatment (need to increase fraction of inspired oxygen > 0.2, need for fraction of inspired oxygen > 0.5, respiratory rate > 25) or score higher than 4 on the 7‐point ordinal scale WHO‐CIS Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1). If new need was not available directly, we used death as a proxy for assumed intubation counted together with patients alive and ventilated: not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: time to recovery (defined as the first day after enrollment, on which a patient attained category 1, 2, or 3 on the 7‐point ordinal scale WHO‐CIS), clinical status of patients using the 7‐point ordinal scale WHO‐CIS at day 5, 11, 14, 28, and 60 days after randomisation, adverse drug reactions, number of patients admitted to the ICU, number of patients who needed mechanical ventilation, duration of mechanical ventilation, length of ICU and hospital stay, mortality during hospitalisation
Identification
Notes	Date of publication: 16 December 2021 Sponsor/funding: Manuel Taboada Muñiz

Tang 2021.

Study characteristics
Methods	Trial design: prospective, multicentre, single‐blind RCT Type of publication: journal publication Setting: inpatient Recruitment dates: 19 February 2020 to 31 March 2020 Country: China Language: English Number of centres: 7 Trial registration number: NCT04273321 Date of trial registration: 15 February 2020
Participants	Age: median age 57 years (IQR 49 to 67) in the intervention group 55 years (IQR 38 to 65) in the control group Gender 21 male (48.8%) in the intervention group 20 male (46.5%) in the control group Proportion of confirmed infections: PCR positivity inclusion criterion Ethnicity: not reported Number of participants (recruited/allocated/evaluated): 213 Allocated: 43 in the intervention group and 43 in the control group Evaluated: 43 in the intervention group and 43 in the control group Severity of condition according to study definition: COVID‐19 pneumonia (confirmed by chest‐CT) Admitted to the general wards Severity of condition according to WHO score: moderate to severe 4 to 6 Co‐morbidities: COPD, asthma, hypertension, coronary heart disease, diabetes, chronic renal failure Inclusion criteria Age > 18 years old Confirmed SARS‐CoV‐2 infection Admitted in the general wards Able to sign informed consent Exclusion criteria Severe immunosuppression (HIV infection, long‐term use of immunosuppressive agents) Pregnant or lactation period women Glucocorticoids are needed for other diseases Unwilling or unable to participate or complete the study Participating in other study Previous treatments: not reported
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroid: methylprednisolone Dose: 1 mg/kg/day in 100 mL 0.9% NaCl for 7 days Route of administration: intravenous Treatment details of control group (e.g. dose, route of administration, number of doses) 100 mL 0.9% NaCl IV and standard care Concomitant therapy (e.g. description of standard care): standard therapy of COVID‐19: according to the Chinese Diagnosis and Treatment Plan for COVID‐19 (trial version 6); antivirals: 67 (77.9%) of patients, antibiotics: 61 (70.9%) of patients Duration of follow‐up: at least 14 days after randomisation or until hospital discharge Treatment cross‐overs: none documented Compliance with assigned treatment: yes
Outcomes	Primary study outcome: clinical deterioration 14 days after randomisation Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: reported Additional study outcomes: clinical deterioration 14 days after randomisation (defined as deterioration of clinical signs and symptoms, new pulmonary or extrapulmonary lesions, progress in chest CT, ICU admission or death); clinical cure 14 days after randomisation (defined as improvement of clinical signs and symptoms of COVID‐19 and no need of additional therapy); time from randomisation to clinical cure, median (IQR), days; ICU admission
Identification
Notes	Date of publication: 22 January 2021 Sponsor/funding: Beijing Chao Yang Hospital

Tomazini 2020.

Study characteristics
Methods	Trial design: multicentre, open‐label RCT Type of publication: journal publication Setting: inpatient Recruitment dates: 17 April to 23 June 2020 Country: Brazil Language: English Number of centres: 41 Trial registration number: NCT04327401 Date of trial registration: 31 March 2020
Participants	Age: mean 60.1 years (SD 15.8) intervention group and 62.7 years (SD 13.1) control group Gender: 90 (59.6%) male and 61 (40.4%) female in the intervention group; 97 (65.6%) male and 51 (34.5%) female in the control group Proportion of PCR test results: Positive: 95.4% intervention arm, 95.9% control arm Negative: 0% intervention arm, 0.7% control arm Unclear: 4.6% intervention arm, 3.4% control arm Ethnicity: not reported Number of participants (recruited/allocated/evaluated): 545/151 intervention group and 148 control group/151 intervention group and 148 control group Severity of condition according to study definition: all participants were mechanically ventilated Severity of condition according to WHO score: severe 7 to 9 Co‐morbidities: hypertension, diabetes, obesity, heart failure, chronic kidney failure Inclusion criteria: At least 18 years old Confirmed or suspected COVID‐19 infection Receiving mechanical ventilation within 48 hours of meeting criteria for moderate to severe ARDS with partial pressure of arterial blood oxygen to fraction of inspired oxygen (PaO2:FIO2) ratio of 200 or less Exclusion criteria: Pregnancy or active lactation Known history of dexamethasone allergy Corticosteroid use in the past 15 days for non‐hospitalised patients Use of corticosteroids during the present hospital stay for more than 1 day Indication for corticosteroid use for other clinical conditions (e.g. refractory septic shock) Use of immunosuppressive drugs Cytotoxic chemotherapy in the past 21 days Neutropenia due to haematological or solid malignancies with bone marrow invasion Consent refusal Expected death in the next 24 hours Previous treatments: no
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses): Type of corticosteroid: dexamethasone Dose: 20 mg once daily for 5 days, followed by 10 mg intravenously once daily for additional 5 days or until ICU discharge Route of administration: intravenous Treatment details of control group (e.g. dose, route of administration, number of doses): standard care Concomitant therapy (e.g. description of standard care): hydroxychloroquine, azithromycin, other antibiotics, oseltamivir Duration of follow‐up: 28 days Treatment cross‐overs: no Compliance with assigned treatment 25 deviations from protocol in the intervention arm (16.55%) 1 patient received a corticosteroid other than dexamethasone In the control arm, 52 patients received corticosteroids, of which 14 were protocol deviations (9.4%)
Outcomes	Primary study outcome: number of days alive and free from mechanical ventilation for at least 48 consecutive hours Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: reported Adverse events (any grade), defined as the number of participants with any event: reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: not reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: Sequential Organ Failure Assessment (SOFA) scores
Identification
Notes	Date of publication: 2 September 2020 Sponsor/funding: this trial was funded and supported by the Coalition COVID‐19 Brazil. The Laboratórios Farmacêuticos provided the study drug, distribution logistics, and insurance for the study patients

Toroghi 2021.

Study characteristics
Methods	Trial design: open‐label, randomised clinical trial Type of publication: journal publication Setting: inpatient Recruitment dates: 26 October 2020 to 25 January 2021 Country: Iran Language: English Number of centres: 1 Trial registration number: IRCT20100228003449N31 Date of trial registration: 8 October 2020
Participants	Age: mean 59 years (SD 14) in the low‐dose group, mean 59 years (SD 17) in the intermediate‐dose group and mean 56 years (SD 16) in the high‐dose group Gender: 28 (59.6%) male in the low‐dose group, 21 (52.5%) male in the intermediate‐dose group, and 31 (58.6%) male in the high‐dose group Proportion of confirmed infections: PCR positivity inclusion criterion Ethnicity: not reported Number of participants (recruited/allocated/evaluated): Recruited: 144 Allocated: 48 low‐dose group, 48 intermediate‐dose group, 48 high‐dose group Evaluated: 47 low‐dose group, 40 intermediate‐dose group, 46 high‐dose group Severity of condition according to study definition: moderate to severe; moderate COVID‐19 was considered when clinical signs of pneumonia (fever, cough, dyspnoea, and high respiratory rate) were positive along with SpO2 between 90% and 93% on room air. Severe COVID‐19 was described as clinical signs of pneumonia plus respiratory rate > 30 breaths/min or SpO2 < 90% on room air. Severity of condition according to WHO score: moderate to severe ≥ 5 Co‐morbidities: hypertension, diabetes mellitus, ischaemic heart disease, hypothyroidism, respiratory disorders, cerebrovascular accident, dyslipidaemia, neuropsychiatric disorders, rheumatoid arthritis, Parkinson's disease, depression, malignancy, renal disorders, liver disorders, heart failure Inclusion criteria Hospitalised adult patients With moderate to severe COVID‐19 Requiring supplemental oxygen Reverse transcriptase‐polymerase chain reaction (RT‐PCR) of nasopharyngeal samples and a lung computed tomography (CT) scan Positive RT‐PCR test or compatible lung involvement was considered for diagnosis of COVID‐19 Exclusion criteria Contraindication of corticosteroids (uncontrolled diabetes mellitus, active bacteria, fungal or parasite infections, hypersensitivity reactions, close angle glaucoma, uncontrolled neuropsychiatric disorders, unstable cardiovascular disorders including acute myocardial infarction, acute thrombosis, uncontrolled hypertension, viral hepatitis, history of corticosteroids induced myopathy) Pregnancy lactation History of recent corticosteroids Other immunosuppressant drugs use Previous treatments: aspirin, angiotensin receptor blockers, statin, beta blocker, metformin, azithromycin, levothyroxine, sofosbuvir‐ledipasvir, insulin, doxycycline, hydroxychloroquine, immunosuppressants, supplements, other antibiotics, angiotensinogen converting enzyme inhibitors
Interventions	3 groups comparing different doses of dexamethasone Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroids: dexamethasone 8 mg twice a day for up to 10 days or until hospital discharge 8 mg ≥ 3 times a day for up to 10 days or until hospital discharge Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses): Type of corticosteroids: dexamethasone Dose: 8 mg once a day for up to 10 days or until hospital discharge Route of administration: IV Concomitant therapy (e.g. description of standard care): administration of antivirals, anticoagulants, antibiotics, analgesics, fluids, electrolytes, supplemental oxygen, vitamins, minerals, nutritional supports, and stress ulcer prophylaxis Duration of follow‐up: probably 60 days Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: time to a clinical response that was described as improvement of at least 2 scores in the 8‐category ordinal scale of the National Institute of Health (NIH): (1) discharge, with no limitations in usual activity, (2) discharge, with some limitations in usual activity, (3) hospital admission without the requirement of supplemental oxygen, (4) hospital admission, requiring oxygen by mask or nasal cannula, (5) hospital admission requiring non‐invasive ventilation or high‐flow oxygen, (6) intubation and mechanical ventilation, (7) mechanical ventilation and additional organ support like vasopressors, renal replacement therapy (RRT) or extracorporeal membrane oxygenation (ECMO), and (8) death. Review outcomes: inpatient setting Mortality: all‐cause mortality at day 14 or any longer observation period, in‐hospital all‐cause mortality: reported Improvement of clinical status during the longest observation period available: Ventilator‐free days (defined as days alive and free from mechanical ventilation): not reported Participants discharged alive. Participants should be discharged without clinical deterioration or death: not reported Deterioration of clinical status during the longest observation period available: New need for invasive mechanical ventilation or death; that is, transition to WHO 7 to 9 if 6 or lower at baseline (see Figure 1): not reported Serious adverse events, defined as the number of participants with any event: not reported Adverse events (any grade), defined as the number of participants with any event: not reported Specific adverse events: hospital‐acquired infections: reported Fungal infections: reported Quality of life, including fatigue and neurological status, assessed with standardised scales (e.g. WHOQOL‐100) during the longest period available: not reported New need for dialysis during the longest period available: not reported Viral clearance, assessed with reverse transcription polymerase chain reaction (RT‐PCR) test for SARS‐CoV‐2 at baseline, up to 3, 7, and 15 days: not reported Additional study outcomes: time to 50% decrease in serum CRP level, time to respiratory rate ≤ 20 breaths per minute, time to peripheral oxygen saturation ≥ 93%, hospital readmission, need for ICU admission, duration of hospital and ICU stay, need for mechanical ventilation
Identification
Notes	Date of publication: 27 November 2021 Sponsor/funding: the authors did not receive any funds for this work

AE: adverse event; ARDS: acute respiratory distress syndrome; COPD: chronic obstructive pulmonary disease; CPK: creatine phosphokinase; CT: computed tomography; ECMO: extracorporeal membrane oxygenation; FiO2: fraction of inspired oxygen; GI: gastrointestinal; h: hours; ICU: intensive care unit; IMV: invasive mechanical ventilation; IQR: interquartile range: IV: intravenous; LDH: lactate dehydrogenase; LMIC: low‐ and middle‐income countries; NIV: non‐invasive ventilation; PaO2: partial pressure of oxygen; PEEP: positive end‐expiratory pressure; RCT: randomised controlled trial; RT‐PCR: reverse transcription polymerase chain reaction; SAE: serious adverse event; SaO2: arterial oxygen saturation; SpO2: blood oxygen saturation; SD: standard deviation; WHO: World Health Organization

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
EUCTR2020‐001445‐39‐ES	Corticosteroid plus other active substances versus standard care
EUCTR2020‐001616‐18‐ES	Inhaled corticosteroids
EUCTR2020‐001889‐10	Inhaled corticosteroids
IRCT20120225009124N4	Corticosteroid plus other active substances versus standard care
IRCT20190312043030N2	Corticosteroid plus other active substances versus standard care
IRCT20200522047542N1	Topical corticosteroids
ISRCTN86534580	Inhaled corticosteroids
Moreira 2021	The administered antibody (foralumab) was not standard care
Naik 2021	The participants received low‐dose dexamethasone in addition to standard care
NCT04341038	Corticosteroid plus other active substances versus standard care
NCT04355637	Inhaled corticosteroids
NCT04359511	Withdrawn (competitor test RECOVERY)
NCT04361474	Topical corticosteroids
NCT04381364	Inhaled corticosteroids
NCT04411667	Corticosteroid plus other active substances versus standard care
NCT04416399	Inhaled corticosteroids
NCT04468646	Corticosteroid plus other active substances versus standard care
NCT04484493	Topical corticosteroids
NCT04485429	Withdrawn (it was not possible to perform the trial due to the availability and logistics of porcine heparin)
NCT04534478	Corticosteroids for long‐COVID treatment
NCT04551781	Corticosteroids for long‐COVID treatment
NCT04561180	Corticosteroid plus other active substances versus standard care
NCT04569825	Topical corticosteroids
NCT04640168	Corticosteroid plus other active substances versus standard care
NCT04657484	Corticosteroids for long‐COVID treatment
NCT04826822	Corticosteroid plus other active substance versus standard care
NCT05133635	Study was withdrawn because a recent study suggested a new corticosteroid regime for intensive care unit patients, so that no participants were recruited
Odeyemi 2021	Compared bio‐marker adjusted corticosteroid dosing with usual care (corticosteroid use and dosing determined by the physician)

Characteristics of studies awaiting classification [ordered by study ID]

EUCTR2020‐001307‐16‐ES.

Methods	Trial design: open RCT Sample size: 104 Setting: inpatient Language: Spanish, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Diagnosis of SARS‐CoV‐2 by RT‐PCR tested on a respiratory sample Pneumonia confirmed by radiological imaging test ARDS criteria: Bilateral infiltrates; PO2/FiO2 < 300 mmHg Reasonable clinical exclusion of heart cause (requires all) Verbal consent of the patient Exclusion criteria Age < 18 years < 5 days from the onset of symptoms to randomisation Pregnancy Hypersensitivity or known allergy to methylprednisolone Bacterial infection: not drained abscess, intravascular infection, bacterial pneumonia, septic shock, disseminated fungal infection Participation in another trial in the previous 30 days Acquired immunodeficiency syndrome Previous use of corticosteroids (cumulative dose of prednisone (or equivalent) of > 300 mg in the last 21 days; or > 15 mg/d in the last 7 days before randomisation) Cytotoxic treatment in the last 3 weeks Known or suspected adrenal insufficiency Lung or bone marrow transplant Severe liver disease
Interventions	Details of intervention: methylprednisolone Dose: dosage unclear Route of administration: IV Treatment details of control group (e.g. dose, route of administration): no information Concomitant therapy: no information
Outcomes	Primary study outcome: death for any cause in the first 28 days after randomisation
Notes	Recruitment status: prematurely ended Prospective completion date: 6‐month duration Date last update was posted: unclear Sponsor/funding: Fundación para la Investigación Biomédica Hospital Ramón y Cajal

EUCTR2020‐001333‐13‐FR.

Methods	Trial design: open RCT Sample size: 122 Setting: inpatient Language: French, English Number of centres: 18 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Patient aged > 18 Patient affiliated to a health insurance plan Patient who has given their free, informed and written consent or patient for whom an independent doctor has given their signed consent as part of an emergency procedure Serum potassium > 3.5 mmol/L Patient diagnosed COVID‐positive by RT‐PCR and/or scanner (patients admitted with already IMV and sedation, or with acute respiratory failure evolving very quickly) Exclusion criteria Patient under guardianship or curatorship Patient with plausible alternate diagnosis ARDS evolving for > 4 days Contraindication to hydroxychloroquine Contraindication to dexamethasone Uncontrolled septic shock Untreated active infection or treated < 24 h Long‐term patient treated with corticosteroids (> 20 mg/d) or hydroxychloroquine Immunocompromised patients: AIDS, bone marrow or solid organ transplant recipients Pregnant women
Interventions	Details of intervention Dose: 20 mg dexamethasone + hydroxychloroquine Route of administration: dexamethasone IV, hydroxychloroquine orally Treatment details of control group (e.g. dose, route of administration): hydroxychloroquine Concomitant therapy: no information
Outcomes	Primary study outcome: mortality on day 28
Notes	Recruitment status: prematurely ended Date of the global end of the trial: 7 August 2020 Date last update was posted: unclear Sponsor/funding: Groupe Hospitalier Paris Saint‐Joseph

EUCTR2020‐001553‐48‐FR.

Methods	Trial design: open RCT Sample size: 304 Setting: inpatient Language: French Number of centres: 17 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years old Hospitalisation for COVID‐19 infection confirmed by RT‐PCR or other virological method SpO2 ≤ 94% in ambient air measured twice at 5‐ to 15‐min intervals, or PaO2/FiO2 < 300 mmHg Abnormalities on the chest X‐ray or CT scan suggesting viral pneumonia Signature of a free and informed consent by the patient Exclusion criteria COVID‐19 infection with first symptoms > 9 days ago (depending on the patient’s questioning; the day of symptoms is defined as the first day with fever, cough, shortness of breath, and/or chills related to the COVID‐19 infection) Patients with primary or secondary immunodeficiency, including: HIV, chronic haematological disease, solid organ transplant, ongoing immunosuppressive treatment Long‐term corticosteroid therapy defined by taking > 10 mg/d (prednisone equivalent) Infection suspected or confirmed or with bacteria, fungal agents, or viruses (in addition to COVID‐19) Known contraindication to systemic corticosteroids Systolic blood pressure < 80 mmHg SpO2 < 90% under 5 L/min of oxygen in the mask at medium concentration, or higher oxygen requirements Patient on long‐term oxygen therapy Mechanical ventilation in progress Septic shock in progress Multi‐organ failure in progress Pregnant or breastfeeding woman (oral diagnosis) Lack of affiliation or beneficiary of a Social Security scheme Guardianship, curatorship or safeguard of justice
Interventions	Details of intervention Dose: prednisone 0.75 mg (concrete dosage/time frame missing) Route of administration: oral Treatment details of control group (e.g. dose, route of administration): standard care Concomitant therapy: no information
Outcomes	Primary study outcome Number of patients on day 7 of randomisation (i.e. on day 14 of symptoms ± 5 days) presenting a theoretical indication for transfer to ICU with a respiratory indication evaluated by an SpO2 < 90% stabilised at rest and under 5 L/min of oxygen at mask at medium concentration measured twice 5 to 15 min apart. The average value of the two measurements will be used.
Notes	Recruitment status: prematurely ended Prospective completion date: 7 months predicted Date last update was posted: unclear Sponsor/funding: Hospices Civils de Lyon

EUCTR2020‐002186‐34‐ES.

Methods	Trial design: open‐label, randomised trial Sample size: 100 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age 18 to 80 Coronavirus‐19 infection (SARS‐CoV‐2) demonstrated by nasopharyngeal smears PCR or any other biological sample; COVID‐19 described as: nasopharyngeal smear with SARS‐CoV‐2 positive PCR, lung infiltrates by radiography (or other imaging technique) consistent with pneumonia, punctuation of 3 or 4 in the WHO Ordinal Scale for Clinical Improvement for COVID‐19; one of the following criteria: ambient air oxygen, saturation > 90% and < 94%, Pa: FiO2 (partial pressure O2/fraction of inspired O2) > 200 and ≤ 300 mmHg, Sa: FiO2 (O2 saturation measured with pulse oximeter/inspired O2 fraction) ≤ 350 Exclusion criteria Previous treatment with oral or IV corticosteroids for > 5 days in a row or alternate days (6 previous months) Treatment during the previous 12 months with biological drugs such as monoclonal antibodies including anti‐TNFα, anti‐interleukins, interferons type I Any other contraindication for the use of individualised corticosteroid pulses according to the clinical criteria of the patient medical team Contraindications to treatment with methylprednisolone (limited to known hypersensitivity to the active substance and its excipients), as well as receiving treatment in a post‐vaccination period (with live or live attenuated micro‐organism vaccines) Patients with severe ARDS, defined as SaFi < 150 Patients with COPD requiring home oxygen
Interventions	Details of intervention: methylprednisolone Dose: 250 mg; frequency not stated Route of administration: IV Treatment details of control group (e.g. dose, route of administration): standard care Concomitant therapy: no
Outcomes	Primary outcome: incidence of a combined variable made up of the variables death, ICU admission, non‐IMV, or need for high‐flow oxygen therapy (defined as SaFi < 200 with FiO2 ≥ 50%) (day 90)
Notes	Recruitment status: prematurely ended Prospective completion date: estimated duration 1 year Date last update was posted: no information Sponsor/funding: Fundació Hospital Universitari Vall d'Hebron ‐ Institut de Recerca (VHIR)

EUCTR2020‐004323‐16.

Methods	Trial design: randomised, multicentre, double‐blind study Sample size: 260 Setting: inpatient Language: Italian, English Number of centres: 5 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age = 18 years Informed consent for participation in the study and for data processing Molecular diagnosis with PCR test of Sars‐CoV2 infection Hospitalisation in a specialist ward for COVID‐19 patient care (e.g. Infectious Diseases, Pulmonology or Internal Medicine) Need for supplemental oxygen in any delivery mode with the exception of IMV PaO2/FiO2 between 100 and 300 mmHg Clinical/instrumental diagnosis (high‐resolution chest CT scan or chest X‐ray or lung ultrasound) of interstitial pneumonia for no more than 3 days Serum CRP > 5 mg/dL Interval from onset of SARS‐CoV2 infection symptoms to randomisation > 5 days Exclusion criteria IMV Presence of shock or concomitant organ failure that requires admission to the ICU Pregnancy or breastfeeding Severe heart or kidney failure Known hypersensitivity to methylprednisolone, to dexamethasone or to an exception Diabetes not compensated according to the doctor's judgement Other clinical conditions that contraindicate methylprednisolone and cannot be treated or resolved according to the doctor's judgement Steroid bolus therapy in the week prior to enrolment for the study Enrolment in another clinical trial Patient already randomised in this study
Interventions	Details of intervention: Dose: methylprednisolone (1 g/d for 3 days) Route of administration: IV Treatment details of control group (e.g. dose, route of administration): placebo, IV Concomitant therapy: no information
Outcomes	Primary study outcome: length of hospitalisation, calculated as the interval between randomisation and discharge from the hospital without the need for supplemental oxygen
Notes	Recruitment status: completed Prospective completion date: 4‐month trial duration planned Date last update was posted: 25 November 2020 Sponsor/funding: Azienda Ospedaliera Arcispedale Santa Maria Nuova/IRCCS Di Reggio Emilia

Gautam 2021.

Methods	Trial design: randomised controlled trial Type of publication: journal publication Setting: inpatient Recruitment dates: 15 April to 15 June 2021 Country: India Language: English Number of centres: 1 Trial registration number: not stated Date first posted: not stated
Participants	Age: mean age 45.50 years in the methylprednisolone group 45.34 years in the dexamethasone group Gender: 46 male (65.71%) and 24 (34.28%) in the methylprednisolone group 50 male (71.42%) and 20 female (28.57%) in the dexamethasone group Proportion of confirmed infections: PCR positivity inclusion criterion Ethnicity: not stated Number of participants (recruited/allocated/evaluated) Recruited: 140 Allocated: 70 methylprednisolone group, 70 dexamethasone group Evaluated: 70 methylprednisolone group, 70 dexamethasone group Severity of condition according to study definition: moderate‐to‐severe COVID‐19 cases according to definition, patients having oxygen saturation < 93% on room air Severity of condition according to WHO score: moderate to severe ≥ 5 Co‐morbidities: diabetes mellitus, chronic kidney disease, hypertension, chronic obstructive pulmonary disease, hypothyroidism, heart disease, obesity, malignancy Inclusion criteria: Age: 18 to 75 years COVID‐19 RT‐PCR positive Moderate‐to‐severe COVID‐19 cases according to definition Patients having oxygen saturation < 93% on room air, regardless of chest X‐ray infiltrates Patients given informed consent Exclusion criteria: Pregnant or lactating females Immunocompromised conditions such as HIV or long‐term use of immunosuppressant for any chronic illness Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): not stated
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroid: methylprednisolone Dose: 2.0 mg/kg bodyweight for 3 days followed by 1.0 mg/kg for 3 days in the divided dose Route of administration: IV Treatment details of control group (e.g. dose, route of administration, number of doses): dexamethasone, 8 mg/day intravenously daily in divided dose up to 10 days Concomitant therapy (e.g. description of standard care): patients were given oxygen by nasal cannula, face mask, and non‐rebreathing mask Duration of follow‐up: day 0, 5, 10 Treatment cross‐overs: no Compliance with assigned treatment: yes
Outcomes	Primary study outcome: not stated Additional study outcomes: all‐cause mortality, hospital‐acquired infections, radiological improvement, transfer to intensive care unit, hyperglycaemic coma, clinical improvement
Notes	Date of publication: 1 December 2021 Sponsor/funding: none

Ghanei 2021.

Methods	Trial design: open‐label, multicentre, 3‐arm, randomised controlled trial Type of publication: journal publication Setting: inpatient Recruitment dates: 13 April and 9 August 2020 Country: Iran Language: English Number of centres: 6 Trial registration number: IRCT20200318046812N2 Date of trial registration: 8 April 2020
Participants	Age: mean age (SD): 58.2 years (17.2) in the intervention group (prednisolone plus azithromycin) Mean age (SD): 57.6 years (15.6) in the control group (azithromycin) Gender: 57 male (49.1%) in the intervention group 55 male (50.0%) in the control group Ethnicity: not stated Number of participants (recruited/allocated/evaluated): 336/120 in the intervention group and 116 in the control group/116 in the intervention group and 110 in the control group Severity of condition according to study definition: oxygen saturation (SpO2) 94% or less Severity of condition according to WHO score: moderate to severe Co‐morbidities: hypertension, diabetes, chronic heart disease, chronic lung disease, chronic kidney disease, mild liver disease, rheumatologic disease, chronic neurologic disease Inclusion criteria: hospitalised patients, 16 years of age or older, had a positive polymerase chain‐reaction (PCR) assay and oxygen saturation (SpO2) 94% or less Exclusion criteria: history of receiving any medications (i.e. immunosuppressive drugs, systemic steroids, chemotherapy drugs, hydroxychloroquine, lopinavir/ritonavir, ribavirin, and oseltamivir) for COVID‐19 in the last month, patients with uncontrolled diabetes or asthma, patients with chronic renal or liver disease, gastrointestinal haemorrhage, untreated bacterial infection, pregnancy or breast‐feeding, and QT interval ≥ 500 ms; COVID‐19 patients who were ill for less than 48 h Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): not clearly defined
Interventions	Details of intervention: 400 mg hydroxychloroquine stat + prednisolone (25 mg prednisolone daily) + 250 mg azithromycin (2 tablets on the first day and then 250 mg daily) + 50 mg naproxen (twice a day) for 5 days. Prednisolone was gradually tapered to 5 mg per week after discharge for reduction of readmission. Route of administration: not stated Treatment details of control group (e.g. dose, route of administration): 400 mg hydroxychloroquine stat + 250 mg azithromycin (2 tablets on the first day and then 250 mg daily) + 250 mg naproxen (twice a day) for 5 days Concomitant therapy: 40 mg of pantoprazole tablets or capsules daily during treatment to prevent gastrointestinal complications; oral steroid, intravenous steroid, steroid pulse therapy, interferon, NSAID, anticoagulant, bronchodilator, plasmaphaeresis, favipiravir, hyperimmune plasma, CinnoRA, haemoperfusion, ACTEMRA
Outcomes	Primary study outcome: number of admissions to intensive care unit Additional study outcomes: all‐cause mortality, the length of hospital stay (LOS), death during admission, intubation in ICU, and time to clinical recovery. Clinical recovery was defined as being medically stable and ready for discharge from the hospital, determined by the attending physician
Notes	Date of publication: 15 September 2021 Sponsor/funding: not applicable

IRCT20081027001411N3.

Methods	Trial design: single‐blinded RCT Sample size: 60 Setting: inpatient Language: English Number of centres: 4 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria COVID‐19 patient with ARDS Confirmed by positive PCR test for SARS‐CoV‐2 or confirmed by abnormal CT scan finding (bilateral, subpleural, peripheral ground glass opacities) SpO2 < 93% Not responding to standard COVID‐19 treatment after 48 to 72 h Exclusion criteria Type I diabetes Asthma and lung diseases Malignancies Kidney and heart failure Uncontrolled high blood pressure Positive pro‐calcitonin and active infection Taking immunosuppressive drugs and corticosteroids Pregnant or lactating women Prescribed antibiotics due to a bacterial infection
Interventions	Details of intervention Dose: 0.5 mg/kg prednisolone in 3 divided doses up to 30 mg/d for 5 to 7 days Route of administration: not stated Treatment details of control group (e.g. dose, route of administration): control group receives standard treatment for COVID‐19 disease Concomitant therapy: not stated
Outcomes	Primary study outcome Radiographic features findings (CT scan day 8 + 14) Mortality rate O2 saturation (day 8 + 14) Need for an oxygen therapy
Notes	Recruitment status: completed Prospective completion date: 30 June 2020 Date last update was posted: 1 June 2020 Sponsor/funding: Teheran University of Medical Sciences

IRCT20120215009014N354.

Methods	Trial design: double‐blind, phase II RCT Sample size: 81 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age 18 to 70 years Hospitalised in ICU < past 48 h Mild to moderate ARDS due to COVID‐19 Bilateral pulmonary infiltration in chest X‐ray or CT‐scan Respiratory distress with > 24 breaths/min Exclusion criteria Pregnancy or breastfeeding Cardiopulmonary oedema Severe acute respiratory distress syndrome Using antioxidant drugs Chronic liver or renal disease Contraindication of N‐acetyl cysteine
Interventions	Details of intervention group 1 Type of corticosteroid: hydrocortisone Dose: 50 mg every 6 h for 5 days Route of administration: IV Details of intervention group 2 Type of corticosteroid: methylprednisolone Dose: 40 mg every 12 h for 5 days Route of administration: IV Details of intervention group 3 Type of corticosteroid: dexamethasone Dose: 20 mg daily for 5 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): Concomitant therapy: routine care
Outcomes	Primary study outcome Need for mechanical ventilation (at day 28) Patients' clinical status (at day 28) Mortality (at day 28)
Notes	Recruitment status: completed Prospective completion date: 5 August 2020 Date last update was posted: 1 May 2020 Sponsor/funding: Hamedan University of Medical Sciences

IRCT20160118026097N4.

Methods	Trial design: unblinded, RCT Sample size: 60 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age between 18 and 70 years Laboratory confirmation of COVID‐19 infection with RT‐PCR from any diagnostic sampling source New organ dysfunction, which is related to COVID‐19, includes: Hypoxia requires supplemental oxygen to maintain oxygen saturation > 90% Hypotension (systolic blood pressure < 90 mmHg) or need for vasopressor/inotropic drug Renal impairment (especially creatinine) 50% of baseline, onset, received based on glomerular filtration film Reduce the Glasgow scale by ≥ 2 Thrombocytopenia (< 150,000 platelets per millimetre) Symptoms of gastrointestinal upset requiring hospitalisation (e.g. severe nausea, vomiting, diarrhoea, or abdominal pain) Exclusion criteria Sensitivity or sensitivity to lopinavir/ritonavir or recombinant IFN‐β1b, including toxic epidermal necrolysis, Stevens‐Johnson syndrome, erythema or angioedema syndrome ALT above 5 times normal Use of drugs that are contraindicated with lopinavir/ritonavir and do not replace or stop during the study period, such as CYP3A inhibitors Pregnancy ‐ eligible female participants are tested at gestational age before enrolling in a pregnancy study HIV infection is known to cause concern about the resistance to lopinavir/ritonavir if used in combination with other anti‐HIV drugs Uncontrolled diabetes (prohibition of prednisolone) According to the 31st National Guide, all vulnerable groups, such as the mentally disabled, emergency patients, or inmates, are excluded from the study
Interventions	Details of intervention Type of corticosteroids: dexamethasone Dose: daily dexamethasone dose of 0.1 mg/kg body weight (our plausible correction from "dexamethasone is treated daily 0/1mg/kg for a week" as stated in the trial register) Route of administration: no information Treatment details of control group (e.g. dose, route of administration): treatment according to Ministry of Health's protocol Concomitant therapy: no information
Outcomes	Primary study outcome: mortality rate or recovery within 30 days after hospitalisation
Notes	Recruitment status: completed Prospective completion date: no information Date last update was posted: 13 September 2020 Sponsor/funding: Ghoum University of Medical Sciences

IRCT20200611047727N3.

Methods	Trial design: single‐blinded, RCT Sample size: 60 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Oxygen saturation level < 93 At least 7 days have passed since the onset of symptoms At least 5 days of antiviral treatment Exclusion criteria Patients who are unable to go for follow‐up scans
Interventions	Details of intervention: Type of corticosteroids: methylprednisolone Dose: 0.75 to 1 mg/kg for 5 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): standard care Concomitant therapy: no
Outcomes	Primary study outcome: radiological changes (before the intervention and 6 weeks later)
Notes	Recruitment status: completed Prospective completion date: no information Date last update was posted: 3 January 2021 Sponsor/funding: Shahid Beheshti University of Medical Sciences

IRCT20201015049030N1.

Methods	Trial design: single‐blind, RCT Sample size: 200 Setting: outpatient Language: English Number of centres: 4 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria COVID‐19 confirmed by positive PCR test for SARS‐CoV‐2 or confirmed by abnormal CT scan finding (bilateral, subpleural, peripheral ground glass opacities) SpO2 between 90% and 95% Patients with severe respiratory symptoms such as cough, shortness of breath, and severe shortness of breath or CRP > 20 or after 3 days of standard treatment worsened symptoms of the disease, including exacerbated fevers, aggravated weakness or aggravated shortness of breath based on the physician's clinical judgement Exclusion criteria Patients with a history of underlying disease such as diabetes, malignancies, renal and heart failure, uncontrolled hypertension Patients taking immunosuppressive drugs and corticosteroids Patients with other active infections Pregnant or lactating women
Interventions	Details of intervention Type of corticosteroids: dexamethasone Dose: 8 mg (if patient's condition does not change, the dose will repeat after 48 h) Route of administration: no information Treatment details of control group (e.g. dose, route of administration): standard care Concomitant therapy: no
Outcomes	Primary study outcome O2 saturation Mortality rate Need for an oxygen therapy Constitutional. (The review authors are unsure about the meaning of "Constitutional" and the study registry entry does not provide further information. The characteristics of this study will be updated once a publication is available.) Needs of hospitalisation
Notes	Recruitment status: completed Prospective completion date: no information Date last update was posted: 7 November 2020 Sponsor/funding: Teheran University of Medical Sciences

ISRCTN33037282.

Methods	Trial design: open‐label, RCT Sample size: 680 Setting: inpatient Language: English Number of centres: 3 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Able to understand and sign the informed consent SARS‐CoV‐2 positive on at least 1 upper respiratory swab or bronchoalveolar lavage PaO2 ≤ 60 mmHg or SpO2 ≤ 90% or on HFNC, CPAP or NPPV at randomisation Age ≥ 18 years old at randomisation Exclusion criteria On IMV (either intubated or tracheostomised) Heart failure as the main cause of acute respiratory failure On long‐term oxygen or home mechanical ventilation Decompensated liver cirrhosis Immunosuppression (i.e. cancer on treatment, post‐organ transplantation, HIV‐positive, on immunosuppressant therapy) Chronic renal failure with dialysis dependence Progressive neuromuscular disorders Cognitively impaired, dementia or decompensated psychiatric disorder Quadriplegia/hemiplegia or quadriparesis/hemiparesis Do‐not‐resuscitate order Previous or current use of remdesivir Participating in other clinical trial including experimental compound with proved or expected activity against SARS‐CoV‐2 infection Any other condition that in the opinion of the investigator may significantly impact with patient's capability to comply with protocol intervention
Interventions	Details of intervention: per‐protocol methylprednisolone administration and tapering A. On day 1, loading dose of methylprednisolone 80 mg IV in 30 min, promptly followed by continuous infusion of methylprednisolone 80 mg/d in 240 mL of normal saline at 10 mL/h B. From day 2 to day 8: infusion of methylprednisolone 80 mg/d in 240 mL of normal saline at 10 mL/h C. From day 9 and beyond: If not intubated patient and PaO2/FiO2 > 200, taper to methylprednisolone 20 mg IV in 30 min 3/d for 3 days, then methylprednisolone 20 mg IV twice daily for 3 days, then methylprednisolone 20 mg IV once daily for 2 days, then switch to methylprednisolone 16 mg/d orally for 2 days, then methylprednisolone 8 mg/d orally for 2 days, then MP 4 mg/d orally for 2 days If intubated patient or PaO2/FiO2 ≤ 200 with at least 5 cm H2O CPAP, continue infusion of methylprednisolone 80 mg/d in 240 mL of normal saline at 10 mL/h until PaO2/FiO2 > 200 then taper as in A Treatment details of control group: per‐protocol dexamethasone administration A. Dexamethasone 6 mg IV in 30 min or orally from day 1 to day 10 or until hospital discharge (if sooner) B. After day 10 study treatment is interrupted Concomitant therapy: no
Outcomes	Primary study outcome Recovery time measured in days using patient records Recovery time determined as the time until hospital discharge when each of the following criteria were met: Decrease in laboratory severity markers Improvement in symptoms Decrease in oxygen requirement until nasal cannula or supplementary oxygen removal and at least 2 doses of the respective treatment have been received
Notes	Recruitment status: completed Prospective completion date: 30 April 2021 Date last update was posted: 19 November 2020 Sponsor/funding: Clínica Medellín ‐ Grupo Quirónsalud

Montalvan 2021.

Methods	Trial design: single‐centre, randomised, non‐blinded control trial pilot study Type of publication: abstract only Setting: hospitalised participants Recruitment dates: 81 randomised (40 low‐dose, 41 high‐dose dexamethasone) Country: Honduras Language: English Number of centres: single‐centre Trial registration number: not reported Date first posted: October 2021
Participants	Age: (mean ± SD) of 56.9 ± 14.9 and 57.5 ± 16.5 for low and high‐dose dexamethasone respectively Gender: low‐dose n = 17 (42.5%) male and n = 23 (57.5%) female vs the high‐dose n = 29(70.7%) male and n = 12(29.3%) female Proportion of confirmed infections: not reported Ethnicity: not reported Number of participants (recruited/allocated/evaluated): not reported/81/81 (40 low‐dose, 41 high‐dose dexamethasone) Severity of condition according to study definition: not reported Severity of condition according to WHO score: not reported Co‐morbidities: not reported Inclusion criteria: not reported Exclusion criteria: not reported Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): not reported
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroid: dexamethasone Dose: 24 mg/day Route of administration: not reported Treatment details of control group (e.g. dose, route of administration, number of doses): dexamethasone 6 mg /day Concomitant therapy (e.g. description of standard care): not reported Duration of follow‐up: not reported Treatment cross‐overs: not reported Compliance with assigned treatment: not reported
Outcomes	Primary study outcome: reduction in mortality and intubation Additional study outcomes: hospital‐acquired infections, risk of becoming critically ill and requiring intubation
Notes	Date of publication: October 2021 Sponsor/funding: not reported

NCT04244591.

Methods	Trial design: randomised, open‐label Sample size: 80 Setting: inpatient Language: English Number of centres: multicentre Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: Adult PCR confirmed COVID‐19 infection Symptoms developed > 7 days PaO2/FiO2 < 200 mmHg Positive pressure ventilation (non‐invasive or invasive) or HFNC > 45 L/min for < 48 h Requiring ICU admission Exclusion criteria Pregnancy Patients currently taking corticosteroids (cumulative 400 mg prednisone or equivalent) Severe underlying disease, i.e. end stage of malignancy disease or end stage of pulmonary disease Severe adverse events before ICU admission, i.e. cardiac arrest Underlying disease requiring corticosteroids Contraindication for corticosteroids Recruited in other clinical intervention trial
Interventions	Details of intervention Dose: methylprednisolone 40 mg every 12 h for 5 days Route of administration: no information Treatment details of control group (e.g. dose, route of administration): standard care Concomitant therapy: no information
Outcomes	Primary study outcome Lower Murray lung injury score (time frame: 7 days after randomisation) Lower Murray lung injury score (time frame: 14 days after randomisation)
Notes	Recruitment status: completed Prospective completion date: 13 April 2020 Date last update was posted: 13 April 2020 Sponsor/funding: Peking Union Medical College Hospital

NCT04325061.

Methods	Trial design: open‐label RCT Sample size: 19 Setting: inpatient Language: English Number of centres: multicentre Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years Positive RT‐PCR assay for COVID‐19 in a respiratory tract sample Intubated and mechanically ventilated Acute onset of ARDS, as defined by Berlin criteria as moderate‐to‐severe ARDS, which includes: having pneumonia or worsening respiratory symptoms; bilateral pulmonary infiltrates on chest imaging (X‐ray or CT scan) absence of left atrial hypertension, pulmonary capillary wedge pressure < 18 mmHg, or no clinical signs of left heart failure; hypoxaemia, as defined by a PaO2/FiO2 ratio of ≤ 200 mmHg on PEEP of ≥ 5 cm H2O, regardless of FiO2. Exclusion criteria Routine treatment with corticosteroids during the previous week irrespective of dose Corticosteroid use within the previous 24 h of > 20 mg of dexamethasone or equivalent Patients with a known contraindication to corticosteroids Decision by a physician that involvement in the study is not in the patient's best interest Pregnancy and breast‐feeding Participation in another therapeutic trial
Interventions	Details of intervention Dose: dexamethasone (20 mg/daily/from day 1 of randomisation for 5 days, followed by 10 mg/daily from day 6 to 10 of randomisation Route of administration: IV Treatment details of control group (e.g. dose, route of administration): patients will be treated with standard intensive care Concomitant therapy: no information
Outcomes	Primary study outcome: all‐cause mortality at 60 days after enrolment
Notes	Recruitment status: terminated (lack of enrolment) Prospective completion date: June 2020 Date last update was posted: February 2021 Sponsor/funding: Dr. Negrin University Hospital

NCT04347980.

Methods	Trial design: single‐blinded (participants) RCT Sample size: 122 Setting: inpatient Language: French, English Number of centres: multicentre, no concrete information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Patient > 18 years old Patient affiliated to a health insurance plan Patient who has given their free, informed and written consent or patient for whom an independent doctor has given their signed consent as part of an emergency procedure Kaliaemia > 3.5 mmol/L Patient diagnosed COVID‐19‐positive by RT‐PCR and/or CT The diagnosis of COVID‐19 will be made if: patient with radiological images strongly suggestive of a chest scan associated with respiratory symptoms, without other obvious aetiologies; OR patient with suggestive respiratory symptoms associated with a positive RT‐PCR. Patients admitted to ICU with acute respiratory distress syndrome secondary to COVID‐19, intubated for < 5 days with one of: hypoxaemia defined by PaO2/FiO2 ratio < 100 after 2 sessions of prone position an alteration in pulmonary compliance (tidal volume divided by plateau pressure minus positive expiratory pressure) immediately or over the first 96 h after the start of ARDS defined by: immediately: impossibility of maintaining a plateau pressure < 30 cm of water in a ventilated patient with a tidal volume of 6 mL/kg of weight predicted by the size and a positive expiratory pressure at 10 cm of water, during the course of the evolution: decrease in compliance by 20% compared to the initial compliance (day of treatment of the intubated and ventilated patient) We define the start date of ARDS by the day and time when the patient is intubated and ventilated with regard to our definition of COVID‐19 Exclusion criteria Patient under guardianship or curator Patient with plausible alternate diagnosis ARDS evolving for > 4 days Contraindication to hydroxychloroquine: known allergy or intolerance to the hydroxychloroquine or to one of the excipients of the drug, in particular to lactose; documented QT prolongation and/or known risk factors for QT prolongation (including ongoing treatment with citalopram, escitalopram, hydroxyzine, domperidone or piperaquine), retinopathies Contraindication to dexamethasone: known allergy or intolerance to dexamethasone or to one of the excipients of the drug, another evolving virosis (hepatitis, herpes, chickenpox, shingles), severe coagulation disorder Uncontrolled septic shock Untreated active infection or treated < 24 h Long‐term patient treated with corticosteroids (> 20 mg/day) or hydroxychloroquine Immunocompromised patients: AIDS, bone marrow or solid organ transplant recipients Pregnant women Glucose‐6‐phosphate dehydrogenase (G6PD) deficiency
Interventions	Details of intervention Dose: dexamethasone (20 mg for 5 days followed by 10 mg for 5 days) combined with 600 mg/d dose of hydroxychloroquine for 10 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): 200 mg x 3/d enterally from J1 of the hydroxychloroquine for 10 days. If the patient is extubated before the 10th day, he will receive his last dose of hydroxychloroquine before. Concomitant therapy: no information
Outcomes	Primary study outcome: day 28 mortality
Notes	Recruitment status: terminated (ANSM (Agence nationale de sécurité du médicament et des produits de santé) Recommendation) Prospective completion date: August 2020 Date last update was posted: 17 April 2020 Sponsor/funding: Centre Chirurgical Marie Lannelongue

NCT04438980.

Methods	Trial design: double‐blinded RCT Sample size: 72 Setting: inpatient Language: Spanish, English Number of centres: 2 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years old Diagnosis of SARS‐CoV‐2 pneumonia confirmed by RT‐PCR on nasopharyngeal swab or sputum according to the recommendations of the Spanish Ministry of Health Length of symptoms consistent with COVID‐19 ≥ 7 days Hospital admission At least one of the following: CRP > 60 mg/L IL‐6 > 40 pg/mL Ferritin > 1000 μg/L. Acceptance of informed consent Exclusion criteria Allergy or contraindication to any of the drugs under study SpO2 < 90% (in air ambient) or PaO2 < 60 mmHg (in ambient air) or PaO2/FiO2 < 300 mmHg Ongoing treatment with glucocorticoids, immunosuppressive, or biologic drugs with another indication Decompensated diabetes mellitus Uncontrolled hypertension Psychotic or manic disorder Active cancer Pregnancy or lactation Clinical or biochemical suspicion (procalcitonin > 0.5 ng/mL) of active infection other than SARS‐CoV‐2 Out‐of‐hospital management patient Conservative or palliative management patient Participation in another clinical trial Any important and uncontrolled medical, psychological, psychiatric, geographic, or social problem that contraindicates the patient's participation in the trial or that does not allow adequate follow‐up and adherence to the protocol and evaluation of the study results.
Interventions	Details of intervention Dose: methylprednisolone 120 mg/d for 3 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): standard care and infusion bag of 100 mL of 0.9% saline Concomitant therapy: no information
Outcomes	Primary outcome Death Need for admission in an ICU Need for mechanical ventilation Decrease in SpO2 < 90% (in ambient air) or PaO2 < 60 mmHg (in ambient air) or PaO2/FiO2 < 300 mmHg, associated with radiological impairment (time frame: at 14 days after randomisation)
Notes	Recruitment status: completed Prospective completion date: February 2021 Date last update was posted: 22 July 2020 Sponsor/funding: Fundacion Miguel Servet

NCT04451174.

Methods	Trial design: single‐blinded (outcomes assessor) RCT Sample size: 184 Setting: inpatient Language: Spanish, English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria ≥ 18 years COVID‐19 confirmed by PCR Oxygen requirements until 35 % by Venturi mask or 5 L/min by nasal cannula Consent form signed Exclusion criteria Previous steroid use ≥ 48 h Pregnancy Chronic respiratory failure Requirements of mechanical ventilation (invasive or non‐invasive) Chronic liver damage Child Pugh B or C Chronic kidney disease stage IV or V Immunosuppressed Participation on other trial
Interventions	Details of intervention Dose: prednisone 40 mg days 1 to 4 then prednisone 20 mg days 5 to 8 Route of administration: IV Treatment details of control group (e.g. dose, route of administration): no intervention Concomitant therapy: no information
Outcomes	Primary outcome Admission to ICU Need for IMV or All‐cause death by day 28 (Time frame: 28 days)
Notes	Recruitment status: terminated (corticosteroid use approval) Prospective completion date: 3 December 2020 Date last update was posted: 20 October 2020 Sponsor/funding: University of Chile

NCT04530409.

Methods	Trial design: open‐label RCT Sample size: 450 Setting: no information Language: Arabic, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Any case with COVID‐19 ≥ 18 years, mild and moderate severity Exclusion criteria Any contra‐indication for the interventional drug Mentally disabled cases
Interventions	Details of intervention Dose: early use of dexamethasone as early as the laboratory confirmation of inflammation Route of administration: most likely systemic Treatment details of control group (e.g. dose, route of administration): dexamethasone is to be used upon the deterioration of cases Concomitant therapy: no information
Outcomes	Primary outcome Deterioration in the clinical picture of cases that necessitate hospitalisation Percentage of cases whose clinical status deteriorate to ARDS Percentage of cases that will need hospitalisation (time frame: 1 to 2 weeks) Percentage of cases that deteriorate to ARDS (time frame: 1 to 2 weeks)
Notes	Recruitment status: completed Prospective completion date: estimated primary completion date 1 April 2021 Estimated study completion date: 1 May 2021 Date last update was posted: 16 February 2021 Sponsor/funding: ClinAmygate

NCT04746430.

Methods	Trial design: open‐label RCT Sample size: 2000 Setting: outpatient Language: Dutch, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years A positive test for SARS‐CoV‐2 GP consultation for deteriorating COVID‐19 symptoms Additional inclusion criteria in order to be eligible for randomisation to the trial: Exercise‐induced desaturation, defined as SpO2 < 92% (< 90% for COPD patients) and/or an absolute drop of ≥ 4% in SpO2 after a 1‐min sit‐to‐stand test OR SpO2 < 92% (< 90% for COPD patients) at rest with GP's and patient's shared decision to keep patient at home despite this in itself being an indication for referral to hospital Exclusion criteria Inability to understand and sign the written consent form Inability to perform saturation measurements or sit‐to‐stand test Not willing to be admitted to hospital On the discretion of the recruiting clinician if he or she deems a patient not eligible Contra‐indication for dexamethasone
Interventions	Details of intervention: Dose: 6 mg dexamethasone prescribed for 10 days and as a precaution combined with electronic monitoring of saturation and other signs and symptoms Route of administration: most likely systemic Treatment details of control group (e.g. dose, route of administration): only remote monitoring Concomitant therapy: no information
Outcomes	Primary outcome: time to first hospital admission or death (time frame: 28 days)
Notes	Recruitment status: terminated (too few patients) Prospective completion date: March 2022 Date last update was posted: 5 April 2021 Sponsor/funding: General Practitioners Research Institute

Rashad 2021.

Methods	Trial design: open‐label, randomised controlled trial Type of publication: journal publication Setting: inpatient (ICU) Recruitment dates: no information Country: Egypt Language: English Number of centres: 1 Trial registration number: NCT04519385 Date of trial registration: 17 August 2020
Participants	Age: Median age (IQR): 60.5 years (49.5 to 66.5) in the tocilizumab group Median age (IQR): 64 years (55 to 72) in the dexamethasone group Gender: 26 male (57%) in the tocilizumab group, 36 male (57%) in the dexamethasone group 20 female (43%) in the tocilizumab group, 27 female (43%) in the dexamethasone group Ethnicity: not stated Number of participants (recruited/allocated/evaluated): 149/74 in the tocilizumab group, 75 in the dexamethasone group/46 in the tocilizumab group and 63 in the dexamethasone group Severity of condition according to study definition: significant deterioration in respiratory clinical status with respiratory rate > 30 cycle/minute, bilateral chest computed tomography (CT) infiltration > 30%, PaO2/FiO2 ratio < 150 or saturation < 90 on > 6 L/min; 2 positive laboratory tests of the following: (CRP > 10 g/L, lymphocytes < 600/mm3, D‐dimer > 500 ng/mL, ferritin > 500 ng/mL) Severity of condition according to WHO score: moderate to severe ≥ 5 Co‐morbidities: diabetes mellitus, hypertension, heart disease, kidney disease, cancer, hyperthyroidism, hepatitis C virus, COPD, asthma Inclusion criteria: patients with significant deterioration in respiratory clinical status with respiratory rate > 30 cycle/minute, bilateral chest computed tomography (CT) infiltration > 30%, PaO2/FiO2 ratio < 150 or saturation < 90 on > 6 L/min, 2 positive laboratory tests of the following: CRP > 10 g/L, lymphocytes < 600/mm3, D‐dimer >500 ng/mL, ferritin > 500 ng/mL Exclusion criteria: paediatric patients < 18 years old, patients with an active bacterial or fungal infection, patients on chemotherapy, patients with interstitial lung disease and patients who were not requiring supplemental oxygen, patients who died before the 3rd day of ICU admission Previous treatments (e.g. experimental drug therapies, oxygen therapy, ventilation): not stated
Interventions	Details of intervention: dexamethasone Dose: 4 mg/kg/day for 3 days, followed by a maintenance dose of 8 mg/day for 10 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): tocilizumab 4 mg/kg/dose in 100 cc normal saline over 1 hour repeated after 24 h, then patient continue symptomatic treatment and oxygen therapy and/or assisted ventilation as needed Concomitant therapy: azithromycin 500 mg/day for 7 days, oxygen therapy, and non‐invasive or mechanical ventilation when needed
Outcomes	Primary study outcome: time to failure, defined as death, within 14 days from ICU admission Additional study outcomes: all‐cause mortality, hospital‐acquired infections
Notes	Date of publication: 23 April 2021 Sponsor/funding: South Valley University

Salukhov 2021.

Methods	Trial design: probably randomised controlled trial (uncertainties regarding the randomisation process) Sample size: 40 Setting: inpatient Language: Russian Number of centres: 1 Trial registration number: not stated Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: Age older than 18 years Diagnosis of COVID‐19 induced pneumonia confirmed by nasopharyngeal/oropharyngeal smears using polymerase chain reaction with reverse transcription Moderate to severe course with process prevalence according to computed tomography without hypoxaemia (saturation (SpO2) > 93%) with a duration of hyperthermia > 38 °C for 3 days or more and C‐reactive protein levels of 15 to 50 mg/L Exclusion criteria: Individual intolerance to glucocorticosteroids Presence of concomitant diseases requiring other glucocorticosteroids and/or other immunomodulatory therapy as well as Janus kinase inhibitors and anti‐cytokine drugs Patients participating in any clinical trials within the previous calendar month
Interventions	Treatment details of intervention group (e.g. dose, route of administration, number of doses) Type of corticosteroid: methylprednisolone plus standard care Dose: 28 mg/day (7 tablets with 4 mg), divided into 2 doses (4 in the morning and 3 in the afternoon), for 7 days Route of administration: oral Treatment details of control group (e.g. dose, route of administration, number of doses): standard care Concomitant therapy (e.g. description of standard care): no
Outcomes	Primary study outcome: composite score of: COVID‐19 progression to extreme severity, a decrease in SpO2 ≤ 93% and the need for oxygen therapy, occurrence of pulmonary and extrapulmonary complications requiring transfer to ICU, and death
Notes	Date of publication: 2021 Sponsor/funding: not stated

ALT: alanine transaminase; ARDS: acute respiratory distress syndrome; COPD: chronic obstructive pulmonary disease; CPAP: continuous positive airway pressure; CPK: creatine phosphokinase; CRP: C‐reactive protein; CT: computed tomography; d: days; FiO2: fraction of inspired oxygen; h: hours; HFNC: high‐flow nasal cannula; ICU: intensive care unit; IMV: invasive mechanical ventilation; IQR: interquartile range: IV: intravenous; LDH: lactate dehydrogenase; NPPV: non‐invasive positive pressure ventilation; PaO2: partial pressure of oxygen; PEEP: positive end‐expiratory pressure; RCT: randomised controlled trial; RT‐PCR: reverse transcription polymerase chain reaction; SD: standard deviation; SpO2: blood oxygen saturation

Characteristics of ongoing studies [ordered by study ID]

ACTRN12621001200875.

Study name	A randomised controlled trial of dexamethasone for emergency and life‐threatening admissions due to COVID‐19 in virtual care: the DELTA study
Methods	Trial design: triple‐blind, randomised controlled trial Sample size: 650 Setting: outpatient Language: English Number of centres: not stated Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: Adult patients (age greater than or equal to 18 years) Currently in Special Health Accommodation (SHA) Confirmed COVID‐19 on polymerase chain reaction (PCR) testing within the last 14 days Exhibiting mild/moderate COVID‐19 symptoms (fever, respiratory tract symptoms, dyspnoea, headache, gastrointestinal symptoms), with no requirement for oxygen or hospitalisation Within 72 hours of onset of mild/moderate COVID‐19 symptoms Requiring medical consultation by RPA Virtual Hospital doctors Exclusion criteria: Currently taking oral or inhaled corticosteroids Major medical or psychiatric co‐morbidities precluding the use of corticosteroids (e.g. poorly controlled diabetes mellitus, chronic heart disease, chronic renal disease, chronic liver failure, schizophrenia or thought disorder, bipolar disorder) Immunosuppression or treatment for malignancy Hypersensitivity to corticosteroids, thiamine, wheat, lactose, povidone, maize starch or magnesium stearate History of alcohol dependence or at risk alcohol intake At risk of thiamine deficiency (e.g. eating disorders, chronic malabsorption) Requires urgent transfer to the emergency department at the time of assessment Patient declines or is unable to provide consent Lactating or pregnant women No access to a personal or loaned electronic device Unable to read or write English
Interventions	Details of intervention: dexamethasone plus usual care Dose: 6 mg per day, for 2 days Route of administration: oral Treatment details of control group (e.g. dose, route of administration): probably placebo (not explicitly explained) Concomitant therapy: not stated
Outcomes	Primary study outcome: COVID‐19 related hospitalisation, defined as COVID‐19 related emergency presentation (excluding injuries and presentations for social reasons), or hospital admission or intensive care unit admission or death
Starting date	20 September 2021
Contact information	Prof Michael Dinh Royal Prince Alfred Hospital Level 10, King George V Building, Camperdown, New South Wales 2050 Australia Telephone: +61 419620654 Email: Michael.Dinh@health.nsw.gov.au
Notes	Recruitment status: recruiting Prospective completion date: not stated Date last update was posted: 10 December 2021 Sponsor/funding: Royal Prince Alfred Hospital

ACTRN12621001603808.

Study name	The effect of prednisolone vs dexamethasone on Covid‐19 in pregnancy: an open labelled randomised control trial
Methods	Trial design: open‐label, randomised controlled trial Sample size: 192 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: women aged 18 to 50 years, pregnant > 16 weeks with COVID‐19, hospitalised with an oxygen requirement Co‐morbidities are included, not limited to obesity, hypertension, diabetes, pre‐existing illnesses Exclusion criteria: pregnancy < 16 weeks gestation, not requiring oxygen, contraindications to receiving corticosteroids
Interventions	Details of intervention: prednisolone Dose: 50 mg once per day for up to 10 days Route of administration: oral Treatment details of control group (e.g. dose, route of administration): dexamethasone, dose and route of administration not explicitly explained Concomitant therapy: not stated
Outcomes	Primary study outcome: WHO ordinal scale of clinical improvement compared at day 5 and at day 10
Starting date	1 December 2021
Contact information	Dr Carole‐Anne Whigham, Monash Health 246 Clayton Road Clayton 3168 Vic, Australia Phone: +61 404644029 Email: carole‐anne.whigham@monashhealth.org
Notes	Recruitment status: not yet recruiting Prospective completion date: not stated Date last update was posted: 24 November 2021 Sponsor/funding: Monash Health Hospital Victoria, Australia

ChiCTR2000029386.

Study name	Effectiveness of glucocorticoid therapy in patients with severe coronavirus disease 2019: protocol of a randomised controlled trial
Methods	Trial design: open‐label RCT Sample size: 48 Setting: inpatient Language: Chinese, English Number of centres: single‐centre Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Aged ≥ 18 years Severe COVID‐19 Willing to give informed consent Exclusion criteria Allergic or intolerant to any therapeutic drugs used in this study Pregnant or lactating women Presence of severe systemic illness that may affect the effectiveness or safety evaluation for this study
Interventions	Details of intervention: Dose: methylprednisolone at a dose of 1 to 2 mg/kg/d for 3 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): no glucocorticoid use Concomitant therapy: no information
Outcomes	Primary study outcome Change in sequential organ failure assessment (SOFA) at 3 days after randomisation Clinical improvement rate
Starting date	No information
Contact information	Dr. Yao‐Kai Chen, Division of Infectious Diseases, Chongqing Public Health Medical Center, Chongqing 400036, China  E‐Mail: yaokaichen@hotmail.com
Notes	Recruitment status: recruiting Prospective completion date: unclear Date last update was posted: 5 May 2020 Sponsor/funding: Chongqing Special Research Project for Prevention and Control of Novel Coronavirus Pneumonia

ChiCTR2000029656.

Study name	A randomised, open‐label study to evaluate the efficacy and safety of low‐dose corticosteroids in hospitalised patients with novel coronavirus pneumonia (COVID‐19)
Methods	Trial design: randomised, open‐label Sample size: 100 Setting: inpatient Language: Chinese, English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Adults (defined as age ≥ 18 years) Patients with new type of coronavirus infection confirmed by PCR/serum antibodies The time interval between symptom onset and random enrolment is within 10 days. The onset of symptoms is mainly based on fever. If there is no fever, cough or other related symptoms can be used. Imaging‐confirmed pneumonia In the state of no oxygen at rest, the patient's SPO2 ≤ 94% or shortness of breath (breathing frequency ≥ 24) or oxygenation index ≤ 300 mmHg Exclusion criteria Known to receive hormone therapy orally or intravenously Hormone therapy is needed due to concomitant disease upon admission Patients with diabetes are receiving oral medication or insulin therapy Known contraindications to dexamethasone or other excipients (such as refractory hypertension; epilepsy or delirium and glaucoma) Known active gastrointestinal bleeding in the past 3 months Known difficulties in correcting hypokalaemia Known secondary bacterial or fungal infections Known immunosuppressive status (such as chemotherapy/radiotherapy/HIV infection within 1 month after surgery) The clinician thinks that participating in the trial may cause patient damage (such as severe lymphocyte reduction) The patient may be transferred to a non‐participating hospital within 72 h
Interventions	Details of intervention Dose: standard treatment and methylprednisolone for injection (no dosage information) Route of administration: IV Treatment details of control group (e.g. dose, route of administration): standard treatment Concomitant therapy: no information
Outcomes	Primary study outcome ECG Chest imaging Complications Vital signs NEWS2 score
Starting date	14 February 2020
Contact information	Ronghui Du: +86 15337110926 Email: bluesearh006@sina.com
Notes	Recruitment status: not yet recruiting Prospective completion date: 14 April 2020 Date last update was posted: 12 February 2020 Sponsor/funding: Wuhan Pulmonary Hospital

ChiCTR2000030481.

Study name	The clinical value of corticosteroid therapy timing in the treatment of novel coronavirus pneumonia (COVID‐19): a prospective randomised controlled trial
Methods	Trial design: open‐label RCT Sample size: 200 Setting: no information, probably inpatient Language: Chinese, English Number of centres: 3 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Patients who are > 18 years Definitely diagnosed with COVID‐19 (i.e. the diagnosis of 2019‐nCoV‐infected pneumonia patients was diagnosed according to the diagnostic criteria for novel coronavirus pneumonia diagnosis and treatment programme (trial version 5) issued by the National Health and Health Commission on 5 February 2020) Exclusion criteria Patients who are allergic to corticosteroids Patients who are diagnosed with adrenal insufficiency Severe immunosuppression, one of the following: infection with HIV and CD4 cell count below 350 cells per microlitre; immunosuppressive therapy after solid organ transplantation; neutropenia (< 500 cells/microlitre) and so on. Patients with cystic fibrosis or active tuberculosis Patients with gastrointestinal bleeding in the past 3 months Incomplete clinical data Participated in other clinical research Patients who cannot understand and execute the survey plan Patients who abandoned treatment
Interventions	Details of intervention Group 1: early corticosteroid intervention group Group 2: middle‐late corticosteroid intervention group Dosage and route of administration: no information, most likely systemic Treatment details of control group (e.g. dose, route of administration): no corticosteroid Concomitant therapy: no information
Outcomes	Primary study outcome: the time of duration of COVID‐19 nucleic acid RT‐PCR test results of respiratory specimens (such as throat swabs) or blood specimens change to negative Among secondary outcomes: 21‐day all‐cause mortality
Starting date	1 March 2020
Contact information	Chen Zhenshun: +86 13627288300 Email: chzs1990@163.com
Notes	Recruitment status: recruiting Prospective completion date: 30 April 2020 Date last update was posted: 3 March 2020 Sponsor/funding: Science and Technology Department of Hubei Province

CTRI/2020/07/026608.

Study name	A clinical trial to study the effects of two drugs methylprednisolone and dexamethasone in patients with severe COVID‐19
Methods	Trial design: randomised, parallel‐group trial Sample size: 40 Setting: inpatient Language: English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Aged 18 to 80 years Men and women Laboratory‐confirmed COVID‐19 cases with ARDS Exclusion criteria Mild and moderate COVID‐19 Severe immunosuppression (HIV infection, long‐term use of immunosuppressive agents) Pregnant or lactating women Patients already on steroids Patients with high procalcitonin level
Interventions	Details of intervention: Dose: methylprednisolone 1 mg/kg once a day for 3 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): injection dexamethasone 6 mg IV once a day for 3 days Concomitant therapy: no information
Outcomes	Primary study outcome Difference in IL‐6 level from baseline Days to ventilator liberation Length of hospital stay In‐hospital all‐cause mortality
Starting date	27 July 2020
Contact information	Prof V R Mohan Rao, India: 9841210011 Email: medicinehod@chettinadhealthcity.com
Notes	Recruitment status: not yet recruiting Prospective completion date: estimated duration of trial 3 months Date last update was posted: 15 July 2020 Sponsor/funding: Chettinad Hospital and Research Institute Kelambakkam, Dr Ananthakumar PK

CTRI/2020/12/029894.

Study name	Comparing the effectiveness of dexamethasone versus methylprednisolone in patients with moderate COVID 19 ‐ a randomised controlled trial
Methods	Trial design: open‐label, RCT Sample size: 50 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Hospitalised patients with moderate SARS‐COV‐2 infection laboratory confirmed by RT‐PCR (moderate COVID‐19: SPO2 < 94% under room air and requiring supplemental oxygen for hypoxaemia, respiratory rate 24 to 30/min, NLR > 5, CRP 50 to 100 mg/L, serum ferritin 600 to 1500 ng/mL, serum LDH 300 to 500 IU/L, D Dimer 0.5 to 1.0 mic/mL, IL‐6 20 to 100 pg/mL, CT chest showing 25% to 75% infiltrates) Exclusion criteria Patients with mild and severe COVID pneumonia Patients who are already on steroid treatment for any underlying condition Patients already started on antivirals Pregnant or lactating women Any known contraindication to short‐term corticosteroid like severe immunosuppression, HIV infection or any other immunosuppressant usage
Interventions	Details of intervention: dexamethasone Dose: 6 mg/8 mg, 7 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): methylprednisolone 32 mg/60 mg intravenous, 6 days Concomitant therapy: no
Outcomes	Primary outcome: mortality during hospital stay (day 1 to day 7)
Starting date	No information
Contact information	DR R Nivetha, Department of General Medicine 1st floor SRM Medical College Hospital and Research Centre SRM University Potheri Kattankulathur, India Email: nivethamdr@gmail.com
Notes	Recruitment status: not yet recruiting Prospective completion date: estimated duration 6 months Date last update was posted: no information Sponsor/funding: SRM Medical College Hospital and Research Centre

CTRI/2020/12/030143.

Study name	Evaluation of different steroid regimes in critically ill adult patients of COVID‐19 admitted to intensive care units
Methods	Trial design: randomised controlled trial Sample size: 500 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Confirmed COVID‐19 infection with pulmonary involvement Admitted to ICU within 14 days of onset of symptoms Receiving invasive or non‐invasive positive pressure ventilation or respiratory support through HFNC Exclusion criteria Septicaemia, active malignancy or patient on immunosuppressive therapy within last 3 months Uncontrolled hyperglycaemia Clinically important gastrointestinal bleed Pregnancy History of hypersensitivity to steroid preparations
Interventions	Details of intervention: dexamethasone Dose: 6 mg, frequency: once daily for 10 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): methylprednisolone, 1 to 2 mg/kg body weight, IV, frequency: daily for 10 days Concomitant therapy: no
Outcomes	Primary outcome: 28‐day mortality
Starting date	No information
Contact information	Dr Sukhyanti Kerai, Maulana Azad Medical College New Delhi, India Email: drsukhi25@gmail.com
Notes	Recruitment status: not yet recruiting Prospective completion date: estimated duration 6 months Date last update was posted: no information Sponsor/funding: Maulana Azad Medical College and associated Lok Nayak Hospital

CTRI/2021/05/033873.

Study name	Pre‐emptive steroids to alter the disease course in COVID‐19 patients
Methods	Trial design: open‐label, randomised controlled trial Sample size: 500 Setting: outpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: Male or non‐pregnant female adult ≥ 18 years of age at the time of enrolment Confirmed diagnosis of SARS‐CoV‐2 infection, defined as either: a) SARS‐CoV‐2 PCR positive within 72 h before randomisation or b) in patients without PCR‐confirmed diagnosis at inclusion, all efforts will be made to confirm definite SARS‐CoV‐2 infection (e.g. PCR on bronchial aspirate, PCR on BAL fluid or serologic testing), but patients who despite all efforts do not have a confirmed diagnosis of COVID‐19 will be excluded from the analysis Patients eligible for home isolation Exclusion criteria: Moderate to severe type of disease, with at least one of the following criteria: Frequency of breath > 24 per minute, which does not decrease after the body temperature drops to normal or subfebrile values Blood oxygen saturation (SpO2) < 94% at rest Chest involvement on HRCT Partial pressure of oxygen in arterial blood (PaO2) < 60 mm Hg Oxygenation index (RaO2/FiO2) ≤ 200 mm Hg Partial pressure of CO2 in arterial blood (PaCO2) < 60 mm Hg Septic shock Patients treated with lopinavir/ritonavir, ribavirin, arbidol, chloroquine, hydroxychloroquine, mefloquine, favipiravir within 7 days prior to screening Severe cardiovascular diseases currently or 6 months prior to trial Uncontrolled arterial hypertension with systolic blood pressure > 180 mm Hg and diastolic blood pressure > 110 mm Hg, pulmonary embolism or deep vein thrombosis Severe chronic renal impairment (GFR < 30 mL/min) or continuous renal replacement therapy, haemodialysis or peritoneal dialysis A history of cirrhosis or an increase in alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) > 5 times × upper limit of normal (ULN) Significant uncontrolled concomitant disease, e.g. neurological, renal, hepatic, endocrinological or gastrointestinal disorder which according to the investigator, could prevent the patient from participating in the study Malignancies that require chemotherapy within 6 months prior to screening Known HIV infection Uncontrolled diabetes mellitus Participation in other clinical studies or taking other study drugs within 28 days prior to screening Pregnant or lactating women Patients not giving consent
Interventions	Details of intervention: methylprednisolone plus standard care Dose: given pre‐emptively at day 5 in addition to the standard care: administered for 14 days: 2 phases ‐ prevention phase for 1 week followed by maintenance and tapering phase for 1 week, Category 1: 4 mg to 8 mg per day if minimal initial symptoms (fever 100.5 °F, body ache, headache, sore throat, HRCT 5%), upgrade to Category 2 if symptoms persist beyond 3 days or increasing, Category 2: 12 mg to 32 mg per day if moderate symptoms (fever 100.5 °F to 103 °F, loss of smell/ taste, lethargy, cough) Route of administration: oral Treatment details of control group (e.g. dose, route of administration): standard care Concomitant therapy: not stated
Outcomes	Primary study outcome: number of patients requiring hospitalisation, supplemental oxygen and/or ICU care as compared to standard of care protocol (time points: baseline, 7 days, 14 days and, if applicable, ICU status/hospitalisation time)
Starting date	2 May 2022
Contact information	Prakhar Gupta, JK Hospital and LN Medical College, Department of Medicine, Third floor JK Hospital and LN Medical college, Bhopal MADHYA PRADESH 462042, India Email: itsme.prakhar@gmail.com
Notes	Recruitment status: not yet recruiting Prospective completion date: not stated Date last update was posted: 28 May 2021 Sponsor/funding: JK Hospital and LN Medical College

CTRI/2021/08/035822.

Study name	Assessment of doubling dose of dexamethasone in progressively worsening severe COVID‐19 pneumonia ‐ a randomised controlled trial
Methods	Trial design: open‐label, randomised controlled trial Sample size: 120 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: patients aged 18 to 99 years, hospitalised, confirmed SARS‐CoV‐2 infection by nucleic acid based testing (RT‐PCR, CB‐NAAT, or TrueNAT) or antigen testing, severe COVID‐19 pneumonia (SpO2 < 94%; PaO2/FiO2 < 300 mm Hg or respiratory rate (RR) > 30 breaths/min) with lack of response to dexamethasone 6 mg after 48 hours (defined as similar or worsening oxygen requirement (margin of error is 5% Fio2 for high flow nasal cannula, 2 L/min for NRBM, and 1 L/min for low flow oxygen devices)) Exclusion criteria: patient already on corticosteroid therapy for an unrelated indication; patient with impending death or respiratory failure necessitating ICU care within 24 hours including inability to maintain SpO2 ≥ 90% despite HFNC with flow 60 L/min and FiO2 1.0 or, if available, NIV with PEEP of up to 8 cm H2O and FiO2 1.0; patients who have received ≥ 2 day of steroids outside hospital care or within the hospital outside of wards that are involved in the study. These doses must be no greater than 12 mg dexamethasone or 64 mg methylprednisolone cumulatively; patients with a known contraindication to corticosteroids including untreated bacterial sepsis, diabetic keto‐acidosis, and invasive fungal infections such as mucormycosis; medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial; pregnancy; recruitment in another therapeutic trial; use of immunosuppressive drugs, cytotoxic chemotherapy in the past 21 days; neutropenia due to haematological or solid malignancies with bone marrow invasion; refusal of consent.
Interventions	Details of intervention: dexamethasone Dose: 12 mg once daily, for 10 days or until day of discharge, whichever is earlier Route of administration: IV or oral Treatment details of control group (e.g. dose, route of administration): dexamethasone, 6 mg, once daily, through intravenous or oral route, for 10 days or till discharge, whichever is earlier Concomitant therapy: not stated
Outcomes	Primary study outcome: supplemental oxygen‐free days at day 28 from hospitalisation, proportion of patients requiring non‐invasive ventilation by NIV mask or invasive mechanical ventilation
Starting date	1 September 2021
Contact information	Animesh Ray, All India Institute of Medical Sciences, New Delhi, Room no. 3070A, Department of Medicine, 3rd floor teaching block, All India Institute of Medical Sciences, New Delhi‐110029 South West DELHI 110029, India Phone: 01126593963 Email: doctoranimeshray@gmail.com
Notes	Recruitment status: not yet recruiting Prospective completion date: not stated Date last update was posted: 16 August 2021 Sponsor/funding: All India Institute of Medical Sciences, New Delhi

EUCTR2020‐001413‐20‐ES.

Study name	Efficacy and safety of siltuximab vs. corticosteroids in hospitalised patients with COVID‐19 pneumonia
Methods	Trial design: phase 2, randomised, open‐label Sample size: 100 Setting: inpatient Language: Spanish, English Number of centres: single‐centre Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years old Hospitalised patient (or documentation of a hospitalisation plan if the patient is in an emergency department) with illness of > 5 days of duration with evidence of pneumonia by chest radiography/CT and meets at least 1 of the following requirements: Non‐critical patient with pneumonia in radiological progression Patient with progressive respiratory failure in the last 24 to 48 h Laboratory‐confirmed SARS‐CoV‐2 infection (by PCR) or other commercialised analysis or public health in any sample collected 4 days before the randomisation or COVID‐19 criteria following the defined diagnostic criteria at that time in the centre Patient with a maximum O2 support of 35% Willing and able to comply with the study‐related procedures/evaluations Women of childbearing potential should have a negative serum pregnancy test before enrolment in the study and must commit to using methods highly effective contraceptives (intrauterine device, bilateral tubal occlusion, vasectomised couple and sexual abstinence) Written informed consent. In case of inability of the patient to sign the informed consent, a verbal informed consent from the legal representative or family witness (or failing this, an impartial witness outside the investigator team) will be obtained by phone. When circumstances so allow, participants should sign the consent form. The confirmation of the verbal informed consent will be documented in a document as evidence that verbal consent has been obtained. Exclusion criteria Patient who, in the investigator's opinion, is unlikely to survive > 48 h after inclusion in the study Presence of any of the following abnormal analytical values at the time of the inclusion in the study: Absolute neutrophil count (RAN) < 2000/mm3 AST or ALT > 5 times the ULN Platelets < 50,000/mm3 In active treatment with immunosuppressants or previous prolonged treatment (> 3 months) of oral corticosteroids for a disease not related to COVID‐19 at a dose > 10 mg of prednisone or equivalent per day Known active tuberculosis (TB) or known history of TB uncompleted treatment Patients with active systemic bacterial and/or fungal infections Participants who, at the investigator's discretion, are not eligible to participate, regardless of the reason, including medical or clinical conditions, or participants potentially at risk of not following study procedures Patients who do not have entry criteria in the ICU Pregnancy or lactation Known hypersensitivity to siltuximab or to any of its excipients (histidine, histidine hydrochloride, polysorbate 80 and sucrose)
Interventions	Details of intervention: siltuximab Dose: 11 mg/kg Route of administration: intravenous Treatment details of control group (e.g. dose, route of administration): methylprednisolone 250 mg/kg Concomitant therapy: no information
Outcomes	Primary study outcome: proportion of patients requiring ICU admission at any time within the study period (time frame 29 days)
Starting date	No information
Contact information	Felipe García: +349322754002884 Email: fgarcia@clinic.cat
Notes	Recruitment status: temporarily halted Prospective completion date: prospective duration of trial 45 days Date last update was posted: no information Sponsor/funding: Fundació Clínic per a la Recerca Biomèdica

EUCTR2020‐001457‐43‐FR.

Study name	Dexamethasone and oxygen support strategies in ICU patients with COVID‐19 pneumonia
Methods	Trial design: double‐blind (phase III) RCT Sample size: 550 Setting: inpatient Language: French, English Number of centres: 12 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years Admitted to ICU within 48 h Confirmed or highly suspected COVID‐19 infection Acute hypoxaemic respiratory failure (PaO2 < 70 mmHg or SpO2 < 90% on room air or tachypnoea > 30/min or laboured breathing or respiratory distress; need for oxygen flow ≥ 6L/min) Any treatment intended to treat the SARS‐CoV‐2 infection (either as a compassionate use or in the context of a clinical trial, i.e remdesivir, lopinavir/ritonavir, favipiravir, hydroxychloroquine and any other new drug with potential activity) Exclusion criteria Moribund patient Pregnancy or lactation Long‐term therapy with corticosteroids with dosage ≥ 0.5 mg/kg/d Active bacterial, fungal or parasitic infection without treatment Missing option of obtaining informed consent from patient or legal representative Allergy or intolerance to dexamethasone or one of its derivatives For patients without mechanical ventilation other exclusion criteria are: Anatomic factors that inhibit nasal cannulation Hypercapny (paCO2 ≥ 50 mmHg)
Interventions	Details of intervention: dexamethasone Dose: 4 mg Route of administration: IV Treatment details of control group (e.g. dose, route of administration): placebo, IV Concomitant therapy: no information
Outcomes	Primary study outcome Time‐to‐death from all causes within the first 60 days after randomisation Time to need for mechanical ventilation (MV) (the additional objective is to assess whether oxygen support based on either HFNO or CPAP modality in COVID‐19‐related AHRF reduces the need for mechanical ventilation at day 28)
Starting date	No information
Contact information	Email: fadila.amerali@aphp.fr
Notes	Recruitment status: ongoing Prospective completion date: October 2020 Date last update was posted: 2 April 2020 Sponsor/funding: APHP (An ancillary study CACAO (Covidicus air contamination) will be performed in 4 centres aiming at assessing the environmental contamination by SARS‐CoV‐2 according to the oxygen support modality. Additional funding will be searched for these analyses (submitted for ANR call))

EUCTR2020‐001622‐64‐ES.

Study name	Outpatient treatment of COVID‐19 with early pulmonary corticosteroids as an opportunity to modify the course of the disease (TAC‐COVID‐19)
Methods	Trial design: open (phase IV) RCT Sample size: 200 Setting: outpatient Language: Spanish, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age 18 to 75 Both sexes Diagnosis of SARS‐CoV‐2 infection, by PCR and/or Ac (IgM+) and/or Ag test Clinical diagnosis of pulmonary involvement (respiratory symptoms ± pathological auscultation ± O2 desaturation) + Chest X‐ray with mild‐moderate or normal alterations Verbal informed consent in front of witnesses, reflected in medical records Exclusion criteria Desaturation < 93% or PO2 < 62 Moderate‐severe dyspnoea or significant respiratory or general deterioration that makes admission advisable Chest X‐ray with multifocal cotton infiltrates Insulin‐dependent diabetes with poor control or glycaemia in the emergency analysis > 300 mg/mL (fasting or not) Other significant co‐morbidities: Severe renal failure CrCl < 30 mL/min Cirrhosis or chronic liver disease Poorly controlled hypertension Heart rhythm disturbances (including prolonged QT) Severe immunosuppression (HIV infection, long‐term use of immunosuppressive agents); cancer Pregnant or lactating women Use of glucocorticoids for other diseases Unwilling or unable to participate until the study is complete Participate in another study Allergy or intolerance to any of the study drugs (prednisone, azithromycin or hydroxychloroquine) Taking any of the drugs being tested within 7 days of being included in the study Non‐suppressible drugs with risk of QT prolongation or significant interactions
Interventions	Details of intervention: prednisone Dose: concrete dosage/duration not mentioned Route of administration: oral Treatment details of control group (e.g. dose, route of administration): concomitant therapy Concomitant therapy: symptomatic treatment + hydroxychloroquine + azithromycin
Outcomes	Primary study outcome: the aim of this study is to explore the effectiveness and safety of oral corticosteroids (prednisone) in the treatment of early stage SARS‐Cov‐2 pneumonia in patients who do not yet meet hospital admission criteria: admission after 30 days
Starting date	19 April 2020
Contact information	María Jesús Coma: 0034610620180 Email: mjcoma@hubu.es
Notes	Recruitment status: ongoing Prospective completion date: 3‐month estimated duration of trial Date last update was posted: 20 April 2020 Sponsor/funding: Dra Ana Pueyo Bastida, Hospital Universitario de Burgos

EUCTR2020‐001707‐16‐ES.

Study name	Outpatient treatment of COVID‐19 with early pulmonary corticosteroids as an opportunity to modify the course of the disease (TOCICOVID)
Methods	Trial design: open (phase IV) RCT Sample size: 60 Setting: inpatient Language: Spanish, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Patient > 18 years old Ability to grant consent Bilateral pneumonia caused by SARS‐CoV‐2 without response to the treatment used according to local protocol. This is defined as persistence of fever (above 37.5ºC without other focus) and respiratory worsening (more dyspnoea, more cough, oxygen therapy at increasing doses, worsening of the degree of respiratory distress according to the PaO2/FiO2 ratio in categories 'mild, moderate or serious') or absence of improvement with respect to the previous state Persistently elevated inflammatory markers, among which must be met: ferritin > 1000 ng/mL and/or D‐dimer > 1500 ng/mL and/or IL‐6 > 40 pg/mL Exclusion criteria Pregnancy and lactation Terminal situation or life expectancy < 30 days in the judgement of the researcher Allergy or intolerance to any of the drugs under study or to any of the excipients of the preparations (e.g. polysorbate 80) Non‐tolerable interaction of the study drugs with some essential chronic medication of the patient ALT/AST > 5 x ULN Severe neutropenia (< 500 cells / mm3 Plateletpenia < 50,000/mm3 Sepsis (clinical suspicion of active infection at another level with a value on the qSOFA scale of ≥ 2 points) or septic shock (need for vasopressors to maintain a mean arterial pressure ≥ 65 mmHg with a lactate > 2 mmol/L, despite adequate volume replacement Another active infection at any level Complicated diverticulitis or intestinal perforation Kidney failure with estimated glomerular filtration < 30 mL/min Liver failure (Child B onwards) Previous use (during the acute process or as chronic medication for another reason) of medication with potential effect in this phase of the disease (Janus kinase inhibitors, interleukin‐1 inhibitors, other immunosuppressants or immunomodulators that, in the investigator's judgement, could have an effect on the disease based on pathophysiological criteria or previous research or started up in this same period) Be included in another clinical trial Patients who, due to their current situation, their baseline situation or other aspects, in the opinion of the researcher, are not considered candidates to enter the study
Interventions	Details of intervention: tocilizumab Dose: 20 mg/mL Route of administration: IV Treatment details of control group (e.g. dose, route of administration): methylprednisolone, 331 mg, IV Concomitant therapy: no information
Outcomes	Primary study outcome: respiratory situation at 24 hours, 3 and 7 days based on PaO2/FiO2 ratio that graduates respiratory distress from mild (200 to 300), moderate (100 to 200) and severe (< 100). In addition, it will include: presence of dyspnoea and grade according to the New York Health Association (NYHA) scale, presence of respiratory work and respiratory rate (FR)
Starting date	22 July 2020
Contact information	IIS BIODONOSTIA 943006288
Notes	Recruitment status: ongoing Prospective completion date: 20‐month estimated duration of trial Date last update was posted: 20 April 2020 Sponsor/funding: IIS BIODONOSTIA

EUCTR2020‐001921‐30.

Study name	Steroids and unfractionated heparin in critically‐ill patients with pneumonia from COVID‐19 infection. A multicenter, interventional, randomised, three arms study design
Methods	Trial design: open RCT Sample size: 200 Setting: inpatient Language: Italian, English Number of centres: 9 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Positive SARS‐CoV‐2 diagnostic (on pharyngeal swab of deep airways material) Positive pressure ventilation (either non‐invasive or invasive) from > 24 h Invasive mechanical ventilation from < 96 h P/F ratio < 150 D‐dimer level > 6 x upper limit of local reference range PCR > 6‐fold upper limit of local reference range Exclusion criteria Age < 18 years Ongoing treatment with anticoagulant drugs Platelet count < 100,000/mm3 History of heparin‐induced thrombocytopenia Allergy to sodium enoxaparine or other LMWH, unfractionated heparin or methylprednisolone Active bleeding or ongoing clinical condition deemed at high risk of bleeding contraindicating anticoagulant treatment Recent (in the last 1 month prior to randomisation) brain, spinal or ophthalmic surgery Chronic intake of corticosteroids (we corrected "Chronic assumption or oral corticosteroids" as stated in the trial register by translating "Assunzione cronica di corticosteroidi") Pregnancy or breastfeeding or positive pregnancy test. In childbearing age women, before inclusion, a pregnancy test will be performed if not available Clinical decision to withhold life‐sustaining treatment or “too sick to benefit” Presence of other severe diseases impairing life expectancy (e.g. patients are not expected to survive 28 days given their pre‐existing medical condition) Lack or withdrawal of informed consent
Interventions	Details of intervention: methylprednisolone + unfractionated heparin (Eparina) Dose: methylprednisolone (125 to 1000 mg/mL) + unfractionated heparin (Eparina) (25,000 international units) Route of administration: IV Treatment details of control group (e.g. dose, route of administration): heparin subcutaneous (25,000 international units) Concomitant therapy: no information
Outcomes	Primary study outcome: all‐cause mortality at day 28
Starting date	14 May 2020
Contact information	Clinical Trials Quality Team: 0594225868 Email: mighali.pasquale@aou.mo.it
Notes	Recruitment status: ongoing Prospective completion date: 1 year later, so May 2021 Date last update was posted: 26 June 2020 Sponsor/funding: AZIENDA OSPEDALIERO‐UNIVERSITARIA POLICLINICO DI MODENA

EUCTR2020‐003363‐25‐DK.

Study name	Higher vs. lower doses of dexamethasone in patients with COVID‐19 and severe hypoxia: the COVID STEROID 2 trial
Methods	Trial design: double‐blinded RCT Sample size: 1000 Setting: inpatient Language: English Number of centres: 36 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Aged ≥ 18 years or above; AND Confirmed SARS‐CoV‐2 (COVID‐19) requiring hospitalisation; AND Use of one of the following: IMV; OR NIV or continuous use of CPAP for hypoxia; OR oxygen supplementation with an oxygen flow of at least 10 L/min independent of delivery system. Exclusion criteria Use of systemic corticosteroids in doses > 6 mg dexamethasone equivalents for other indications than COVID‐19 Use of systemic corticosteroids for COVID‐19 for ≥ 5 days or more Invasive fungal infection Active tuberculosis Fertile woman (< 60 years of age) with positive urine human gonadotropin (hCG) or plasma‐hCG Known hypersensitivity to dexamethasone Previously randomised into the COVID STEROID 2 trial Informed consent not obtainable
Interventions	Details of intervention: Dose: dexamethasone 12 mg (timeframe not mentioned) Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone 6 mg, IV Concomitant therapy: no information
Outcomes	Primary study outcome: days alive without life support (i.e. IMV, circulatory support or renal replacement therapy) from randomisation to day 28
Starting date	18 August 2020
Contact information	Department of Intensive Care, Rigshospitalet: +4535457237 Email: covid‐steroid@cric.nu
Notes	Recruitment status: ongoing Prospective completion date: 18‐month duration planned Date last update was posted: 18 August 2020 Sponsor/funding: Department of Intensive Care, Rigshospitalet, Novo Nordisk Foundation

EUCTR2020‐006054‐43‐IT.

Study name	Randomised controlled trial of methylprednisolone versus dexamethasone in COVID‐19 pneumonia ‐ methylprednisolone vs dexamethasone in COVID‐19
Methods	Trial design: open‐label, randomised controlled trial Sample size: 680 Setting: inpatient Language: English Number of centres: not stated Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: Able to understand and sign the informed consent form SARS‐CoV‐2 positive on at least one upper respiratory swab or bronchoalveolar lavage PaO2 ≤ 60 mmHg or SpO2 ≤ 90% or on HFNC, CPAP or NPPV at randomisation Age ≥ 18 years old at randomisation Exclusion criteria: On invasive mechanical ventilation (either intubated or tracheostomised) Heart failure as the main cause of acute respiratory failure On long‐term oxygen or home mechanical ventilation Decompensated liver cirrhosis Immunosuppression (i.e. cancer on treatment, post‐organ transplantation, HIV‐positive, on immunosuppressant therapy) On chronic steroid therapy or other immunomodulant therapy (e.g. azathioprine, methotrexate, mycophenolate, convalescent/hyperimmune plasma) Chronic renal failure with dialysis dependence Progressive neuromuscular disorders Cognitively impaired, dementia or decompensated psychiatric disorder Quadriplegia/hemiplegia or quadriparesis/hemiparesis Do‐not‐resuscitate order Participating in other clinical trial including experimental compound with proved or expected activity against SARS‐CoV‐2 infection Any other condition that in the opinion of the investigator may significantly impact with patient’s capability to comply with protocol intervention Refuse to participate in the study or absence of signed informed consent form
Interventions	Details of intervention: methylprednisolone Dose: 1000 mg Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone, 8 mg, IV Concomitant therapy: not stated
Outcomes	Primary study outcome: survival proportion at 28 days in both arms
Starting date	3 March 2021
Contact information	Università degli Studi di Trieste, Strada di Fiume, 447 Trieste 34129, Italy Email: mconfalonieri@units.it
Notes	Recruitment status: ongoing Prospective completion date: not stated Date last update was posted: 26 July 2021 Sponsor/funding: Università degli Studi di Trieste

EUCTR2021‐001416‐29‐ES.

Study name	Early treatment strategy with high‐dose versus standard‐dose dexamethasone in patients with SARS‐CoV‐2 pneumonia (COVID‐19)
Methods	Trial design: open‐label, randomised controlled trial Sample size: 200 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: patients admitted with SARS‐CoV‐2 pneumonia (COVID‐19) confirmed by antigenic test or PCR, age = 18 years, presents with a PCR = 66 mg/L and = 150 mg/L at inclusion or pandemic score at admission > 200 with PCR 9.7 to 149 mg/L at inclusion, WHO scale level 4, with need for oxygen therapy in NG = 1 lpm to maintain saturation = 94%, onset of symptoms = 10 days before the date of inclusion Exclusion criteria: patients with criteria of respiratory distress at the time of randomisation, understood as need for OCNAF/NIMV/MIV (levels 5 and 6 of the WHO scale) or O2 saturation = 92% and/or RF = 30 despite oxygen in NG at 4 litres, patients with allergy or contraindication to the use of systemic corticosteroids, patients with severe asthma or chronic lung disease with home oxygen requirements and active corticosteroid therapy, patients on chronic corticosteroid therapy, use of corticosteroids daily in the 15 days prior to hospital admission, indication for steroid use due to other clinical conditions of the patient (e.g. septic shock), pregnant or actively breastfeeding women, patients with suspected or confirmed bacterial, fungal, or viral infection other than SARS‐CoV‐2 itself at time of randomisation, patients with confirmed past or latent tuberculosis infection prior to inclusion, patients with known HIV infection with CD4 below 500 cells/mm3 or on active treatment with protease inhibitors or boosters such as cobicistat or ritonavir, patients with active oncological processes in the last year or on active treatment with chemotherapy, patients with life expectancy < 3 months at inclusion due to clinical conditions other than SARS‐CoV‐2 pneumonia, patients expected to die within 48 to 72 hours, patients included in another clinical trial
Interventions	Details of intervention: dexamethasone Dose: 20 mg Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone, IV, 4 mg Concomitant therapy: not stated
Outcomes	Primary study outcome: requirement of respiratory therapy with NIV/MIV/OCNAF, need for respiratory support with NIMV/MIV/OCNAF, days of hospitalisation counted from signature of informed consent to time of discharge from hospital
Starting date	31 August 2021
Contact information	Ángel Pueyo, Calle Puerto de Lumbreras, 5 28031 Madrid Spain, Fundación para la Investigación e Innovación Biomédica (FIIB) del Hospital Universitario Infanta Leonor Telephone: +34911919855 Email: pueyo.angel@investiganet.es
Notes	Recruitment status: ongoing Prospective completion date: not stated Date last update was posted: 7 September 2021 Sponsor/funding: Fundación para la Investigación e Innovación Biomédica (FIIB) del Hospital Universitario Infanta Leonor y Hospital Unive

EUCTR2021‐004021‐71.

Study name	Corticosteroids for COVID‐19 induced loss of smell – COCOS trial
Methods	Trial design: double‐blind, randomised controlled trial Sample size: 116 Setting: outpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: recent COVID‐19 infection (< 3 months), confirmed with a positive test (PCR of antigen by GGD), persistent loss of smell after one month, objectified by TDI < 30.5 on Sniffin’ Stick test, age 18 years or older, capable of giving informed consent, good understanding of the Dutch language Exclusion criteria: pre‐existing olfactory disorders, chronic rhinitis or rhinosinusitis (with or without nasal polyps), corticosteroids (nasal, oral or intravenously) since positive test, pregnancy, contra‐indications of steroid use (Insulin dependent diabetes mellitus, Ulcus pepticum)
Interventions	Details of intervention: prednisolone Dose: not stated Route of administration: oral (capsule) Treatment details of control group (e.g. dose, route of administration): placebo, oral (capsule) Concomitant therapy: not stated
Outcomes	Primary study outcome: objective olfactory function means of Sniffin’ Sticks, combining olfactory detection threshold (T), discrimination (D) and identification (I) ability into a composite TDI score (ranging from 1 to 48 points) that can be categorised into normosmia, hyposmia and anosmia, clinical improvement is set at > 5.5 points
Starting date	Not stated
Contact information	University Medical Center Utrecht, Otorhinolaryngology Department, Heidelberglaan 100, Utrecht, 3584 CX, Netherlands Telephone number: +310887555555
Notes	Recruitment status: ongoing Prospective completion date: not stated Date last update was posted: not stated Sponsor/funding: University Medical Center Utrecht

IRCT20190606043826N2.

Study name	Comparison of the effectiveness and complication of dexamethasone at doses of 8 and 24 mg in the treatment of in hospitalised patients with Covid‐19
Methods	Trial design: double‐blind, randomised controlled trial Sample size: 60 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: hospitalised patients with COVID‐19 Exclusion criteria: pregnancy, immunodeficiency diseases, chronic liver and kidney diseases, or gastrointestinal bleeding, glucocorticoid use in the last month
Interventions	Details of intervention: dexamethasone Dose: 24 mg daily for up to 3 days, change to 8 mg daily after 3 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone, 8 mg daily, IV Concomitant therapy: not stated
Outcomes	Primary study outcome: determination of mean arterial blood oxygen saturation with and without oxygen therapy (at the beginning of hospitalisation and 3, 5, 7, and 10 days later), shortness of breath score based on modified Borg scale (at the beginning of hospitalisation and 3, 5, 7, and 10 days later), number of days required for mechanical ventilation, measurement of serum levels of C‐reactive protein (at the beginning of hospitalisation and 3, 5, 7, and 10 days later)
Starting date	6 July 2021
Contact information	Marzieh Mollaei Ardestani Phone: +98 31 5558 0931 Email: mollaei‐m@kaums.ac.ir
Notes	Recruitment status: recruitment completed Prospective completion date: not stated Date last update was posted: July 1, 2021 Sponsor/funding: Esfahan University of Medical Sciences

NCT04329650.

Study name	Efficacy and safety of siltuximab vs. corticosteroids in hospitalised patients with COVID‐19 pneumonia
Methods	Trial design: phase 2, randomised, open‐label Sample size: 200 Setting: inpatient Language: Spanish, English Number of centres: 4 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years old Hospitalised patient (or documentation of a hospitalisation plan if the patient is in an emergency department) with illness of > 5 days of duration with evidence of pneumonia by chest radiography/CT and meets at least 1 of the following requirements: Non‐critical patient with pneumonia in radiological progression and/or Patient with progressive respiratory failure at the last 24 to 48 hours Laboratory‐confirmed SARS‐CoV‐2 infection (by PCR) or other commercialised analysis or public health in any sample collected 4 days before the randomisation or COVID‐19 criteria following the defined diagnostic criteria at that time in the centre Patient with a maximum O2 support of 35% Willing and able to comply with the study‐related procedures/evaluations Women of childbearing potential should have a negative serum pregnancy test before enrolment in the study and must commit to using methods highly effective contraceptives (intrauterine device, bilateral tubal occlusion, vasectomised couple and sexual abstinence). Written informed consent. In case of inability of the patient to sign the informed consent, a verbal informed consent from the legal representative or family witness (or failing this, an impartial witness outside the investigator team) will be obtained by phone. When circumstances so allow, participants should sign the consent form. The confirmation of the verbal informed consent will be documented in a document as evidence that verbal consent has been obtained. Exclusion criteria Patient who, in the investigator's opinion, is unlikely to survive > 48 h after the inclusion in the study Presence of any of the following abnormal analytical values at the time of the inclusion in the study: Absolute neutrophil count < 2000/mm3 AST or ALT > 5 times the upper limit of normality Platelets < 50,000 per mm3 In active treatment with immunosuppressants or previous prolonged treatment (> 3 months) of oral corticosteroids for a disease not related to COVID‐19 at a dose > 10 mg of prednisone or equivalent per day Known active tuberculosis or known history of tuberculosis uncompleted treatment Patients with active systemic bacterial and/or fungal infections Patients who have received previous treatment with IL6 inhibitor (tocilizumab, sarilumab) Participants who, at the investigator's discretion, are not eligible to participate, regardless of the reason, including medical or clinical conditions, or participants potentially at risk of not following study procedures Patients who do not have entry criteria in the ICU Pregnancy or lactation Known hypersensitivity to siltuximab or to any of its excipients (histidine, histidine hydrochloride, polysorbate 80 and sucrose)
Interventions	Details of intervention Dose: single‐dose of 11 mg/kg of siltuximab Route of administration: IV Treatment details of control group (e.g. dose, route of administration) A dose of 250 mg/24 h of methylprednisolone for 3 days followed by 30 mg/24 h for 3 days will be administered by IV infusion If the patient is taking lopinavir/ritonavir, the dose will be 125 mg/24 h for 3 days followed by 15 mg/24 h for 3 days Concomitant therapy: no information
Outcomes	Primary study outcome: proportion of patients requiring ICU admission at any time within the study period (time frame: 29 days)
Starting date
Contact information	Contact: Felipe García, MD+34932275400 ext 2884  Email: fgarcia@clinic.ca
Notes	Recruitment status: recruiting Prospective completion date: 20 May 2020 Date last update was posted: 17 April 2020 Sponsor/funding: Judit Pich Martínez

NCT04344730.

Study name	Dexamethasone and oxygen support strategies in ICU patients with COVID‐19 pneumonia (COVIDICUS)
Methods	Trial design: quadruple‐masked RCT Sample size: 550 Setting: inpatient Language: French, English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years Admitted to ICU within 48 h Confirmed or highly suspected COVID‐19 infection Acute hypoxaemic respiratory failure (PaO2 < 70 mmHg or SpO2 < 90% on room air or tachypnoea > 30/min or laboured breathing or respiratory distress; need for oxygen flow ≥ 6 L/min) Any treatment intended to treat the SARS‐CoV‐2 infection in the absence of contraindications (either as a compassionate use or in the context of a clinical trial, i.e. remdesivir, lopinavir/ritonavir, favipiravir, hydroxychloroquine and any other new drug with potential activity) Exclusion criteria Moribund status Pregnancy or breastfeeding Long‐term corticotherapy at a dose of 0.5 mg/kg/d or higher Active and untreated bacterial, fungal or parasitic infection No written informed consent from the patient or a legal representative if appropriate. If absence a legal representative of the patient may be included in emergency procedure Hypersensitivity to dexamethasone or to any of the excipients No affiliation to French social security
Interventions	Details of intervention Dose: dexamethasone 20 mg in 5 mL Route of administration: IV Treatment details of control group (e.g. dose, route of administration): placebo Concomitant therapy: study stratified according to subgroups (only oxygen therapy, CPAP, HFNC, IMV)
Outcomes	Primary study outcome Time to death from all causes (time frame: day 60) Time to need for mechanical ventilation (time frame: day 28) Time to death from all causes within the first 60 days after randomisation
Starting date	10 April 2020
Contact information	Jean François TIMSIT, Pr
Notes	Recruitment status: active, not recruiting Prospective completion date: 31 December 2021 Date last update was posted: 9 February 2021 Sponsor/funding: Assistance Publique ‐ Hôpitaux de Paris

NCT04345445.

Study name	Study to evaluate the efficacy and safety of tocilizumab versus corticosteroids in hospitalised COVID‐19 patients with high risk of progression
Methods	Trial design: open‐label, randomised, cross‐over interventional study Sample size: 310 Setting: inpatient Language: English Number of centres: 4 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Hospitalised symptomatic COVID‐19 patients Presence of clinical and radiological signs of progressive disease, AND laboratory evidence indicative of risk of cytokine storm complications: Clinical: dyspnoea or respiratory rate > 20 breaths/min AND O2 saturation < 93% on room air or increasing need for O2 supplementation to maintain O2 saturation > 95% on room air, WITH Radiological: chest X‐ray or CT indicative of pneumonia or worsening findings over time AND Laboratory: CRP levels > 60 or an increase of CRP > 20 over 12 h, WITH an increasing ferritin level or declining lymphocyte counts Age > 18 years Able to give consent Exclusion criteria Known sensitivity/allergy to tocilizumab or other monoclonal antibodies AST/ALT > 5 times ULN Platelet counts < 50,000 or neutrophil counts < 500 Active TB Pregnant Receipt of mechanical ventilation Has received other immunomodulatory drugs (including tocilizumab) in the past for the treatment of other conditions Individuals, in the opinion of the investigator, where progression to death is imminent and inevitable in the next 24 h irrespective of treatment provision or who have signed a do not resuscitate order Participating in other clinical trials (subject to approval) Any serious medical condition or abnormal clinical laboratory tests which, in the judgement of the investigator, may compromise patient safety should he/she participate in the study
Interventions	Details of intervention Dose: 8 mg/kg (body weight) tocilizumab Route of administration: IV Treatment details of control group (e.g. dose, route of administration): methylprednisolone 120 mg/day for 3 days IV Concomitant therapy: no information
Outcomes	Primary study outcome The proportion of patients requiring mechanical ventilation (time frame: through study completion, and average of 6 months) Mean days of ventilation (time frame: through study completion, and average of 6 months)
Starting date	14 April 2020
Contact information	Adeeba Kamarulzaman, MBBS: +603‐79492050 Email: adeeba@um.edu.my
Notes	Recruitment status: not yet recruiting Prospective completion date: 31 October 2020 Date last update was posted: 14 April 2020 Sponsor/funding: University of Malaya

NCT04377503.

Study name	Tocilizumab versus methylprednisolone in the cytokine release syndrome of patients with COVID‐19
Methods	Trial design: phase II trial (open‐label), RCT Sample size: 40 Setting: inpatient Language: English Number of centres: 1 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Men and non‐pregnant women > 18 years old COVID diagnosis confirmed by RT‐PCR Pao2/FIO2 < 200 Laboratory: High‐sensitivity CRP > 5 mg /L LDH > 245 U/L Ferritin > 300 D‐dimer > 1500 IL‐6 > 7.0 pg/mL Exclusion criteria Known sensitivity/allergy to tocilizumab Active tuberculosis Pregnancy Individuals, in the opinion of the investigators where progression to death is imminent and inevitable in the next 24 h
Interventions	Details of intervention Dose: tocilizumab, 8 mg/kg diluted in 100 mL of saline, dose will be repeated only once 12 h after the first dose Route of administration: IV Treatment details of control group (e.g. dose, route of administration): methylprednisolone at a dose of 1.5 mg/kg/d divided into 2 daily doses for 7 days. Then 1 mg/kg/day for another 7 days in 2 daily doses. Finally, 0.5 mg/kg/d for another 7 days. Concomitant therapy: no information
Outcomes	Primary outcome: patient clinical status 15 days after randomisation on a 7‐category ordinal scale (time frame: 15 days after randomisation)
Starting date	No information
Contact information	Jose A Azevedo, MD, PhD: +559832168110 Email: jrazevedo47@gmail.com
Notes	Recruitment status: not yet recruiting Prospective completion date: no information Date last update was posted: 6 May 020 Sponsor/funding: Hospital Sao Domingos

NCT04452565.

Study name	NA‐831, atazanavir and dexamethasone combination therapy for the treatment of COVID‐19 infection (NATADEX)
Methods	Trial design: phase 2/3, triple‐blinded (care provider, investigator, outcomes assessor), RCT Sample size: 525 Setting: inpatient Language: English Number of centres: 31 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Hospitalisation for management of SARS CoV‐2 infection Positive SARS CoV‐2 test Age ≥ 18 years Provision of informed consent ECG ≤ 48 h prior to enrolment Complete blood count, glucose‐6 phosphate‐dehydrogenase (G6PD), comprehensive metabolic panel and magnesium ≤ 48 h prior to enrolment from standard care If participating in sexual activity that could lead to pregnancy, individuals of reproductive potential who can become pregnant must agree to use contraception throughout the study. At least one of the following must be used throughout the study: condom (male or female) with or without spermicide, diaphragm or cervical cap with spermicide, intrauterine device (IUD), hormone‐based contraceptive Exclusion criteria Contraindication or allergy to NA‐831, atazanavir, dexamethasone Current use any antiviral drug or anti‐inflammatory drug Concurrent use of another investigational agent IMV Participants who have any severe and/or uncontrolled medical conditions such as: Unstable angina pectoris Symptomatic congestive heart failure Myocardial infarction Cardiac arrhythmias or Known prolonged QTc > 470 men, > 480 women on ECG pulmonary insufficiency Epilepsy (interaction with chloroquine) Prior retinal eye disease Concurrent malignancy requiring chemotherapy Known chronic kidney disease, eGFR < 10 or dialysis, G‐6‐PD deficiency, if unknown requires G6PD testing prior to enrolment Known porphyria Known myasthenia gravis Currently pregnant or planning on getting pregnant while on study Breastfeeding AST/ALT > 5 x ULN Bilirubin > 5 x ULN Magnesium < 1.4 mEq/L Calcium < 8.4 mg/dL > 10.6 mg/dL Potassium < 3.3 > 5.5 mEg/L
Interventions	Arm 1: NA‐831 30 mg orally twice a day for 1 day, followed by 30 mg once day for 4 consecutive days (5 days in total) Arm 2: NA‐831 60 mg orally twice a day for 1 day, followed by 30 mg once a day for 4 consecutive days (5 days in total). The drug will be supplied in 30 mg capsule AND atazanavir 400 mg orally twice a day for 1 day, followed by 200 mg daily for 4 consecutive days (5 days total). Arm 3: NA‐831 60 mg orally twice a day for 1 day, followed by 30 mg once a day for 4 consecutive days (5 days in total) AND dexamethasone 8 mg orally twice a day for 1 day, followed by 4 mg daily for 4 consecutive days (5 days total) Arm 4: atazanavir 400 mg orally twice a day for 1 day, followed by 200 mg daily for 4 consecutive days (5 days total) AND dexamethasone 8 mg orally twice a day for 1 day, followed by 4 mg daily for 4 consecutive days (5 days total). The drug will be supplied in 4 mg tablets.
Outcomes	Primary outcome Time (hours) from randomisation to recovery defined as: absence of fever, as defined as at least 48 h since last temperature ≥ 38.0 °C without the use of fever‐reducing medications; AND absence of symptoms of > mild severity for 24 h; AND not requiring supplemental oxygen beyond pre‐COVID baseline; AND freedom from mechanical ventilation or death (time frame: 36 days).
Starting date	First posted: 30 June 2020
Contact information	Brian Tran, MD: 1‐415‐941‐3133  Email: BTran@neuroactiva.com
Notes	Recruitment status: recruiting Prospective completion date: 15 February 2021 Date last update was posted: 7 September 2020 Sponsor/funding: NeuroActiva, Inc.

NCT04499313.

Study name	Dexamethasone versus methylprednisolone for the treatment of patients with ARDS caused by COVID‐19
Methods	Trial design: open‐label RCT Sample size: 60 Setting: inpatient Language: English Number of centres: 2 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Moderate to severe COVID‐19 requires hospitalisation SARS‐CoV‐2 infection will be confirmed by RT PCR/CT chest in every case Exclusion criteria Participants with uncontrolled clinical status who were hospitalised from before Contraindication/possible drug interaction Participants who have any severe and/or uncontrolled medical conditions like severe ischaemic heart disease, epilepsy, malignancy, pulmonary/renal/hepatic disease, pregnancy, cor pulmonale, etc.
Interventions	Details of intervention: Dose: dexamethasone (20 mg/daily/from day 1 of randomisation, followed by a tapering dose according to the patient's condition) Route of administration: IV Treatment details of control group (e.g. dose, route of administration): methylprednisolone sodium succinate at a dose of 0.5 mg/kg (injectable solution) Concomitant therapy: no information
Outcomes	Primary outcome The number of participants with 'clinical improvement' determined by the improvement of individual presenting symptoms of COVID‐19 Changes in radiological and laboratory values Patient admitted in general bed requiring high dependency unit (HDU), and an HDU patient requiring ventilator or intensive care support Mortality rate (in hospital) (time frame: following randomisation 30 days) Clinical improvement (time frame: following randomisation 30 days)
Starting date	2 August 2020
Contact information	Abu Taiub Mohammed Mohiuddin Chowdhury, MBBS, MD: +88 01817711079 Email: dr_mohiuddinchy@yahoo.com
Notes	Recruitment status: recruiting Prospective completion date: 30 November 2020 Date last update was posted: 18 August 2020 Sponsor/funding: Chattogram General Hospital

NCT04509973.

Study name	Higher vs. lower doses of dexamethasone for COVID‐19 and severe hypoxia (COVIDSTEROID2)
Methods	Trial design: quadruple‐blinded, multicentre, clinical RCT Sample size: 1000 Setting: inpatient Language: Swedish, Danish, English Number of centres: 53 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Aged ≥ 18 years AND Confirmed SARS‐CoV‐2 (COVID‐19) requiring hospitalisation; AND Use of one of the following: IMV; OR NIV or continuous use of continuous positive airway pressure (CPAP) for hypoxia; OR oxygen supplementation with an oxygen flow of at least 10 L/min independent of delivery system. Exclusion criteria Use of systemic corticosteroids for other indications than COVID‐19 in doses > 6 mg dexamethasone equivalents Use of systemic corticosteroids for COVID‐19 for ≥ 5 consecutive days Invasive fungal infection Active tuberculosis Fertile woman (< 60 years of age) with positive urine human gonadotropin (hCG) or plasma‐hCG Known hypersensitivity to dexamethasone Previously randomised into the COVID STEROID 2 trial Informed consent not obtainable
Interventions	Details of intervention Dose: dexamethasone 12 mg once daily in addition to standard care for up to 10 days. We will allow the use of betamethasone 12 mg at sites where dexamethasone is not available Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone 6 mg once daily in addition to standard care for up to 10 days. We will allow the use of betamethasone 6 mg at sites where dexamethasone is not available. Concomitant therapy: no information
Outcomes	Primary outcome: days alive without life support (i.e. IMV, circulatory support or renal replacement therapy) from randomisation to day 28
Starting date	27 August 2020
Contact information	Anders Perner, MD, PhD, Professor: +4535458333  Email: anders.perner@regionh.dk
Notes	Recruitment status: active, not recruiting Prospective completion date: 17 February 2022 Date last update was posted: 1 September 2020 Sponsor/funding: Scandinavian Critical Care Trials Group

NCT04513184.

Study name	Randomised clinical trial of intranasal dexamethasone as an adjuvant in patients with COVID‐19
Methods	Trial design: multicentre, double‐masked (participant, care provider), RCT Sample size: 60 Setting: inpatient Language: Spanish, English Number of centres: 3 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria 18 to 75 years old Positive diagnosis of SARS‐CoV‐2 by real‐time RT‐PCR in oropharyngeal sample ≥ 7 days after the start of the infection Hospitalised patients with moderate to severe respiratory complications that have not received mechanical ventilation Patients receiving standard therapy at the Hospital General de México Eduardo Liceaga Signing of the informed consent form Patients of both sexes (non‐pregnant female) ≥ 18 years of age will be eligible if they have a positive diagnostic sample by RT‐PCR, pneumonia confirmed by chest imaging and oxygen saturation (SaO2) < 93% at ambient air or PaO2: FiO2 at ≤ 300 mmHg Exclusion criteria Patients participating in another research protocol Patients receiving oral or IV glucocorticoids Immunosuppressed patients (including HIV infection) Glaucoma patients Patients with allergy to dexamethasone Pregnant or lactating women Concomitant autoimmune diseases Refusal by the patient or family to participate in the study
Interventions	Details of intervention Dose: 6 mg dexamethasone from day 1 to 10 after randomisation Route of administration: IV Treatment details of control group (e.g. dose, route of administration): nasal dexamethasone 0.12 mg/kg/daily for 3 days from day 1, followed by 0.06 mg/kg/daily from day 4 to 10 after randomisation Concomitant therapy: no information
Outcomes	Primary outcome Evaluation of the clinical status of patients after randomisation, defined as a 2‐point improvement in the WHO 7‐point Ordinal Scale Time of clinical improvement (time frame: 10 days after randomisation)
Starting date	14 July 2020
Contact information	Graciela A Cárdenas‐Hernández, PhD: +525556063822 ext 2012  Email: gracielacardenas@yahoo.com.mx
Notes	Recruitment status: recruiting Prospective completion date: primary completion date 30 March 2021 estimated study completion date 31 July 2021 Date last update was posted: 12 November 2020 Sponsor/funding: Edda Sciutto Conde

NCT04528329.

Study name	Anosmia and / or ageusia in COVID‐19: timeline, treatment with early corticosteroid and recovery
Methods	Trial design: open‐label RCT Sample size: 300 Setting: no information Language: Arabic, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Any case with COVID‐19 Age ≥ 18 years Mild to moderate severity Exclusion criteria Diabetes Any contra‐indication for the interventional drug Mentally disabled cases
Interventions	Details of intervention: Dose: early use of dexamethasone as early as the laboratory confirmation of inflammation Route of administration: no information Treatment details of control group (e.g. dose, route of administration): dexamethasone is to be used lately upon the deterioration of cases Concomitant therapy: no information
Outcomes	Time to recovery (time frame: 1 to 6 weeks) from anosmia and/or ageusia
Starting date	30 August 2020
Contact information	Emad R Issak, MD: 01272228989  Email: dr.emad.r.h.issak@gmail.com
Notes	Recruitment status: recruiting Prospective completion date: 15 April 2021 Date last update was posted: 29 March 2021 Sponsor/funding: ClinAmygate

NCT04528888.

Study name	Steroids and unfractionated heparin in critically ill patients with pneumonia from COVID‐19 infection (STAUNCH‐19)
Methods	Trial design: multicentre, national, interventional, randomised, open‐label Sample size: 210 Setting: inpatient Language: Italian, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Positive SARS‐CoV‐2 diagnostic (on pharyngeal swab of deep airways material) Positive pressure ventilation (either non‐invasive or invasive) from > 24 h IMV from < 96 h P/F ratio < 150 D‐dimer level > 6 x upper limit of local reference range PCR > 6‐fold upper limit of local reference range Exclusion criteria Age < 18 years Ongoing treatment with anticoagulant drugs Platelet count < 100,000/mm3 History of heparin‐induced thrombocytopenia Allergy to sodium enoxaparin or other LMWH, unfractionated heparin or methylprednisolone Active bleeding or ongoing clinical condition deemed at high risk of bleeding contraindicating anticoagulant treatment Recent (in the last 1 month prior to randomisation) brain, spinal or ophthalmic surgery Chronic intake of corticosteroids (we corrected "Chronic assumption or oral corticosteroids" as stated in the trial register by translating "Assunzione cronica di corticosteroidi") Pregnancy or breastfeeding or positive pregnancy test. In childbearing age women, before inclusion, a pregnancy test will be performed if not available Clinical decision to withhold life‐sustaining treatment or 'too sick to benefit' Presence of other severe diseases impairing life expectancy (e.g. patients are not expected to survive 28 days given their pre‐existing medical condition) Lack or withdrawal of informed consent
Interventions	Arm 1: enoxaparin SC at standard prophylactic dose (i.e. 4000 UI once day, increased to 6000 UI once day for patients weighing > 90 kg), treatment administered daily up to ICU discharge, destination ward decides whether discontinued Arm 2: enoxaparin SC at standard prophylactic dose (i.e. 4000 UI once day, increased to 6000 UI once day for patients weighing > 90 kg), treatment administered daily up to ICU discharge, destination ward decides whether discontinued AND methylprednisolone IV initial bolus of 0.5 mg/kg followed by administration of 0.5 mg/kg 4 times daily for 7 days, 0.5 mg/kg 3 times daily from day 8 to 10, 0.5 mg/kg 2 times daily at days 11 and 12 and 0.5 mg/kg once daily at days 13 and 14 Arm 3: methylprednisolone IV initial bolus of 0.5 mg/kg followed by administration of 0.5 mg/kg 4 times daily for 7 days, 0.5 mg/kg 3 times daily from day 8 to 10, 0.5 mg/kg 2 times daily at days 11 and 12 and 0.5 mg/kg once daily at days 13 and 14 AND unfractionated heparin IV at therapeutic doses. Infusion started at an infusion rate of 18 IU/kg/h and then modified to attain APTT ratio in the range 1.5 to 2.0. APTT will be periodically checked at intervals no longer than 12 h. Treatment with unfractionated heparin will be administered up to ICU discharge. After ICU discharge anticoagulant therapy may be interrupted or switched to prophylaxis with LMWH in the destination ward, on the clinical judgement of the attending physician.
Outcomes	Primary outcome: all‐cause mortality at day 28, defined as the comparison of proportions of patients' death for any cause at day 28 from randomisation
Starting date	1 September 2020
Contact information	Massimo Girardis, PD: 0594225878 ext 0039  Email: massimo.girardis@unimore.it
Notes	Recruitment status: recruiting Prospective completion date: 30 July 2021 Date last update was posted: 27 August 2020 Sponsor/funding: Massimo Girardis, University of Modena and Reggio Emilia

NCT04545242.

Study name	Efficacy of dexamethasone in patients with acute hypoxemic respiratory failure caused by infections (DEXA‐REFINE)
Methods	Trial design: multicentre, open‐label, clinical RCT Sample size: 980 Setting: inpatient Language: English Number of centres: 40 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Patients with acute hypoxaemic respiratory failure (including ARDS) caused by Infections (including COVID‐19) Age ≥ 18 years Intubated and mechanically ventilated Acute onset of AHRF (as defined by a PaO2/FiO2 ≤ 300 mmHg during at least 6 h from diagnosis). For the measurement of PaO2 and calculation of PaO2/FiO2 ratio, the minimum accepted value for PEEP is 5 cm H2O and for FiO2 is 0.3. ARDS is defined by Berlin criteria 4, which includes: having pneumonia or worsening respiratory symptoms; bilateral pulmonary infiltrates on chest imaging (X‐ray or CT scan); absence of left atrial hypertension or no clinical signs of left heart failure; and hypoxaemia, as defined by a PaO2/FiO2 ≤ 300 mmHg on PEEP of ≥ 5 cmH2O, regardless of FiO2. Pulmonary or systemic infectious aetiology of AHRF. Exclusion criteria Patients with a known contraindication to corticosteroids Patient included in another therapeutic clinical trial Lack of informed consent
Interventions	Details of intervention: Dose: dexamethasone: 6 mg/day during 10 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone: 20 mg/IV/ daily from day of randomisation (day 1) during 5 days, followed by 10 mg/IV/ daily from day 6 to 10 of randomisation Concomitant therapy: no information
Outcomes	Primary outcome: all‐cause mortality at 60 days after randomisation
Starting date	8 February 2021
Contact information	Jesús Villar, MD: +34606860027  Email: jesus.villar54@gmail.com
Notes	Recruitment status: not yet recruiting Prospective completion date: 30 December 2023 Date last update was posted: 22 January 2021 Sponsor/funding: Dr. Negrin University Hospital

NCT04636671.

Study name	Methylprednisolone vs. dexamethasone in COVID‐19 pneumonia (MEDEAS RCT) (MEDEAS)
Methods	Trial design: open‐label RCT Sample size: 680 Setting: inpatient Language: Italian, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Able to understand and sign the informed consent SARS‐CoV‐2 positive on at least 1 upper respiratory swab or bronchoalveolar lavage PaO2 ≤ 60 mmHg or SpO2 ≤ 90% or on HFNC, CPAP or NPPV at randomisation Age ≥ 18 years old at randomisation Exclusion criteria On IMV (either intubated or tracheostomised) Heart failure as the main cause of acute respiratory failure On long‐term oxygen or home mechanical ventilation Decompensated liver cirrhosis Immunosuppression (i.e. cancer on treatment, post‐organ transplantation, HIV‐positive, on immunosuppressant therapy) Chronic renal failure with dialysis dependence Progressive neuromuscular disorders Cognitively impaired, dementia or decompensated psychiatric disorder Quadriplegia/hemiplegia or quadriparesis/hemiparesis Do‐not‐resuscitate order Previous or current use of remdesivir Participating in other clinical trial including experimental compound with proved or expected activity against SARS‐CoV‐2 infection Any other condition that in the opinion of the investigator may significantly impact with patient's capability to comply with protocol intervention
Interventions	Details of intervention Dose: A. On day 1, loading dose of methylprednisolone 80 mg IV in 30 min, promptly followed by continuous infusion of methylprednisolone 80 mg/d in 240 mL of normal saline at 10 mL/h B. From day 2 to 8: infusion of methylprednisolone 80 mg/day in 240 mL of normal saline at 10 mL/h. C. From day 9 and beyond: if not intubated patient and PaO2/FiO2 > 200, taper to methylprednisolone 20 mg IV in 30 min 3 times a day for 3 days, then methylprednisolone 20 mg IV twice daily for 3 days, then methylprednisolone 20 mg IV once daily for 2 days, then switch to methylprednisolone 16 mg/day orally for 2 days, then methylprednisolone 8 mg/day orally for 2 days, then methylprednisolone 4 mg/day orally for 2 days; if intubated patient or PaO2/FiO2 ≤ 200 with at least 5 cmH2O CPAP, continue infusion of methylprednisolone 80 mg/day in 240 mL of normal saline at 10 mL/h until PaO2/FiO2 > 200 then taper as in a) Route of administration: IV Treatment details of control group (e.g. dose, route of administration): A. dexamethasone 6 mg IV in 30 min or orally from day 1 to 10 or until hospital discharge (if sooner). B. After day 10 study treatment is interrupted. Concomitant therapy: no information
Outcomes	Primary outcome: survival proportion at 28 days in both arms
Starting date	25 November 2020
Contact information	Marco Confalonieri, MD: +390403994667  Email: mconfalonieri@units.it
Notes	Recruitment status: recruiting Prospective completion date: estimated primary completion date 31 March 2021; estimated study completion date 30 April 2021 Date last update was posted: 19 November 2020 Sponsor/funding: University of Trieste and Centro di Riferimento Oncologico ‐ Aviano and National Institute for the Infectious Diseases (L. Spallanzani) ‐ Rome

NCT04663555.

Study name	Effect of two different doses of dexamethasone in patients with ARDS and COVID‐19 (REMED)
Methods	Trial design: prospective, phase II, open‐label, RCT Sample size: 300 Setting: inpatient Language: Czech, English Number of centres: 11 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Adult (≥ 18 years of age) at time of enrolment Present COVID‐19 (infection confirmed by RT‐PCR or antigen testing) Intubation/mechanical ventilation or ongoing HFNC oxygen therapy Moderate or severe ARDS according to Berlin criteria: moderate ‐ PaO2/FiO2 100 to 200 mmHg; severe ‐ PaO2/FiO2 < 100 mmHg Admission to ICU in the last 24 h Exclusion criteria Known allergy/hypersensitivity to dexamethasone or excipients of the investigational medicinal product (e.g. parabens, benzyl alcohol) Fulfilled criteria for ARDS for ≥ 14 days at enrolment Pregnancy or breastfeeding Unwillingness to comply with contraception measurements from enrolment to at least 1 week after the last dose of dexamethasone (sexual abstinence is considered as adequate contraception method) End‐of‐life decision or patient is expected to die within next 24 h Decision not to intubate or ceilings of treatment in place Immunosuppression and/or immunosuppressive drugs in medical history: Systemic immunosuppressive drugs or chemotherapy in the past 30 days Systemic corticosteroid use before hospitalisation Any dose of dexamethasone during the present hospital stay for COVID‐19 for ≥ last 5 days before enrolment Systemic corticosteroids during present hospital stay for other conditions than COVID‐19 (e.g. septic shock) Present haematological or generalised solid malignancy Any of contraindications of corticosteroids, e.g. intractable hyperglycaemia; active gastrointestinal bleeding; adrenal gland disorders; a presence of superinfection diagnosed with locally established clinical and laboratory criteria without adequate antimicrobial treatment Cardiac arrest before ICU admission Participation in another interventional trial in the last 30 days
Interventions	Details of intervention Dose: dexamethasone 20 mg once daily on day 1 to 5, followed by dexamethasone 10 mg IV once daily on day 6 to 10 Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone 6 mg day 1 to 10 Concomitant therapy: no information
Outcomes	Primary outcome: number of ventilator‐free days at 28 days after randomisation, defined as being alive and free from mechanical ventilation (> 48 h)
Starting date
Contact information	Jan Maláska, MD, PhD, EDIC: +420723784101  Email: jan.malaska@gmail.com
Notes	Recruitment status: recruiting Prospective completion date: 31 March 2021 Date last update was posted: 4 February 2021 Sponsor/funding: Brno University Hospital

NCT04673162.

Study name	Evaluation of the efficacy of high doses of methylprednisolone in SARS‐CoV2 (COVID‐19) pneumonia patients
Methods	Trial design: quadruple‐blind, multicentre, randomised study Sample size: 260 Setting: inpatient Language: Italian, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age = 18 years Informed consent for participation in the study and for data processing Molecular diagnosis with PCR test of SARS‐CoV‐2 infection Hospitalisation in a specialist ward for COVID‐19 patient care (e.g. Infectious Diseases, Pulmonology or Internal Medicine) Need for supplemental oxygen in any delivery mode with the exception of IMV PaO2/FiO2 between 100 and 300 mmHg Clinical/instrumental diagnosis (high‐resolution chest CT scan or chest X‐ray or lung ultrasound) of interstitial pneumonia for no more than 3 days Serum CRP > 5 mg/dL Interval from onset of SARS‐CoV2 infection symptoms to randomisation > 5 days Exclusion criteria IMV Presence of shock or concomitant organ failure that requires admission to the ICU Pregnancy or breastfeeding Severe heart or kidney failure Known hypersensitivity to methylprednisolone, to dexamethasone or to an exception Diabetes not compensated according to the doctor's judgement Other clinical conditions that contraindicate methylprednisolone and cannot be treated or resolved according to the doctor's judgement Steroid bolus therapy in the week prior to enrolment for the study Enrolment in another clinical trial Patient already randomised in this study
Interventions	Details of intervention Dose: standard treatment (currently dexamethasone 6 mg/daily for 10 days) plus methylprednisolone 1 g daily on days 1, 2, 3 Route of administration: IV Treatment details of control group (e.g. dose, route of administration): standard treatment (currently dexamethasone 6 mg/daily for 10 days) plus placebo Concomitant therapy: no information
Outcomes	Primary study outcome Time to recovery (discharge from hospital) Invasive ventilation prevention Survival
Starting date
Contact information	Massimo Costantini, MD: +390522296986 Email: massimo.costantini@ausl.re.it
Notes	Recruitment status: not yet recruiting Prospective completion date: estimated primary completion date April 2021; estimated study completion date June 2021 Date last update was posted: 17 December 2020 Sponsor/funding: Azienda Unità Sanitaria Locale Reggio Emilia

NCT04707534.

Study name	Dexamethasone for COVID‐19
Methods	Trial design: open‐label RCT Sample size: 300 Setting: inpatient Language: English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years old RT‐PCR confirmed COVID‐19 infection Positive pressure ventilation (non‐invasive or invasive) or HFNC or need supplemental oxygen with oxygen mask or nasal cannula Exclusion criteria Underlying disease requiring chronic corticosteroids Severe adverse events before admission, i.e. cardiac arrest Contraindication for corticosteroids Death is deemed to be imminent and inevitable during the next 24 h Recruited in other clinical intervention trial Pregnancy Patient on judicial protection
Interventions	Details of intervention Dose: dexamethasone 20 mg daily for 5 days, followed by dexamethasone 10 mg daily for 5 days Route of administration: no information Treatment details of control group (e.g. dose, route of administration): dexamethasone 6 mg daily for 10 days Concomitant therapy: no information
Outcomes	Primary outcome: 8‐point World Health Organization ordinal scale at day 28
Starting date	21 January 2021
Contact information	Huimin Wu, MD MPH: (405) 271‐6173  Email: Huimin‐Wu@ouhsc.edu
Notes	Recruitment status: recruiting Prospective completion date: 21 June 2021 Date last update was posted: 8 January 2021 Sponsor/funding: University of Oklahoma

NCT04765371.

Study name	Comparison between prednisolone and dexamethasone on mortality in patients on oxygen therapy, with COVID‐19 (COPreDex)
Methods	Trial design: open‐label RCT Sample size: 220 Setting: inpatients Language: French, English Number of centres: 6 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Patient ≥ 18 years old Patient with SARS‐CoV‐2 pneumopathy documented by nasopharyngeal or bronchoalveolar lavage fluid RT‐PCR or any documented clinical symptoms support by CT scan Patient with SpaO2 ≤ 94 % in room air (90% for patient with respiratory failure) and requiring an oxygen therapy Negative pregnancy test for women of childbearing age Informed and written informed consent (IC) obtained Patients with affiliation to the social security system Exclusion criteria Patient with corticosteroids as background treatment (≥ 10 mg equivalent) Patient under supplemental oxygen > 6 L/min Immunocompromised patient (AIDS, bone marrow or solid organ transplants, etc.) Patient who received a corticosteroid dose within 3 days for COVID‐19 Medical history of hypersensitivity to prednisolone or dexamethasone; or lactose/galactose (excipients with known effect) Another active virus such hepatitis, herpes, varicella, shingles, etc. Psychotic state not controlled by treatment
Interventions	Details of intervention Dose: 6 mg/d of dexamethasone during 10 days Route of administration: most likely systemic Treatment details of control group (e.g. dose, route of administration): 60 mg/d of prednisolone during 10 days Concomitant therapy: no information
Outcomes	Primary outcome: mortality assessment at day 28
Starting date	March 2021
Contact information	Maryline Delattre: +33 130754131  Email: maryline.delattre@ght‐novo.fr
Notes	Recruitment status: recruiting Prospective completion date: October 2021 Date last update was posted: 1 March 2021 Sponsor/funding: Centre Hospitalier René Dubos

NCT04780581.

Study name	Glucocorticoid therapy in coronavirus disease COVID‐19 patients
Methods	Trial design: open‐label RCT Sample size: 290 Setting: inpatient Language: Spanish, English Number of centres: 5 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria > 18 years of age Inpatient Diagnosis of SARS‐CoV‐2 infection confirmed by RT‐PCR or antigen Present evidence in CT of pulmonary involvement attributed to the infection by COVID Patients in whom CT scans are not performed must have suspected pulmonary involvement by clinical examination with simple compatible or suggestive radiology Requires supplementary oxygen due to basal saturation ≤ 93% (with ambient O2, 21%) Exclusion criteria Patient's situation is so serious that the doctor in charge thinks they could die within 24 At the time of randomisation, patients require one of the following 4 ventilatory supports: high‐flow oxygen devices, mechanical NIV, IMV, ECMO The patient is or has been treated in the 2 weeks prior to randomisation with glucocorticoids or inflammation‐modifying drugs, both conventional (thiopurines, cyclophosphamide, cyclosporine, tacrolimus, leflunomide, methotrexate, mycophenolate mofetil/mycophenolic acid, sulfasalazine, hydroxychloroquine or chloroquine) or synthetics or biologics directed against therapeutic targets (abatacept, belimumab, CD‐20, IL1, IL6, Il12. 23, IL‐23, Il.17, TNF, integrin α4β7 or Janus kinase inhibitors JAK). Patients who are only on maintenance treatment with doses of steroids less than or equal to 7.5 mg of prednisone or equivalent per day will not be excluded. The patient is pregnant or breastfeeding The patient has a chronic renal disease in stage 4 or 5 (CCr < 30 mL/min) Moderate to severe dementia at the investigator's discretion Hypersensitivity to any of the active ingredients or to any of the excipients included in its formulation Untreated systemic infections not caused by COVID‐19 Active stomach or duodenal ulcer Recent vaccination with live vaccines Other infection or disease that explains the lung disorder Inability of the patient to understand the study or to sign the informed consent unless consent is delegated to a legal representative Active participation in another clinical study in the last 15 days
Interventions	Details of intervention Dose: intermediate‐dose dexamethasone (6 mg/24 h, 10 days) Route of administration: most likely systemic Treatment details of control group (e.g. dose, route of administration): high‐dose methylprednisolone bolus (250 mg/4 h, 3 days) Concomitant therapy: no information
Outcomes	Primary outcome: mortality rate in COVID‐19 patients after high‐dose methylprednisolone bolus administration versus mortality rate intermediate‐dose dexamethasone pattern (time frame: 28 days)
Starting date	1 February 2021
Contact information	Luis Corral Gudino: 983 420400  Email: lcorral@saludcastillayleon.es
Notes	Recruitment status: recruiting Prospective completion date: 31 December 2021 Date last update was posted: 3 March 2021 Sponsor/funding: Fundación Instituto de Estudios de Ciencias de la Salud de Castilla y León

NCT04795583.

Study name	Corticosteroids for COVID‐19 (CORE‐COVID)
Methods	Trial design: interventional, randomised, placebo‐controlled, triple‐blinded, adaptive clinical trial Sample size: 1526 Setting: outpatient Language: English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Age ≥ 18 years Microbiologically‐confirmed SARS‐CoV‐2 Clinical symptoms compatible with COVID‐19 for ≤ 14 days before randomisation Oxygen saturation ≥ 95% Protein C‐reactive in blood performed by point‐of‐care testing ≥ 20 mg/L Signed informed consent Exclusion criteria Oxygen requirement at home due to chronic lung disease Patients with immunosuppression or immunosuppressive therapies defined as: Cancer on active chemotherapy Stem cell transplant in the previous 6 months Neutrophil count < 1000 cells/mm3 Chronic treatment with immunosuppressive therapy, except for low‐dose (≤ 10 mg daily) prednisone or equivalent dose of other corticosteroids HIV‐infected patients with CD4 < 200 x 106/L Diagnosis with primary immunodeficiencies Chronic liver damage Child‐Pugh C Chronic underlying process with suspected life expectancy < 12 weeks Uncontrolled diabetes mellitus at screening, defined as no blood glucose testing or any blood glucose level > 14 mmol/L (or 250 mg/dL) in the last 14 days Diagnosis of any form of psychosis without receiving appropriate treatment Pregnancy at time of randomisation Patients who received approved or trial vaccination for SARS‐CoV‐2
Interventions	Details of intervention: Dose: prednisone 25 mg capsules ≤ 50 kg = 2 capsules every day for 7 days (maximum dose = 50 mg/d) 50 kg to 80 kg = 3 capsules every day for 7 days (maximum dose = 75 mg/d) > 80 kg = 4 capsules every day for 7 days (maximum dose = 100 mg/d) Route of administration: oral Treatment details of control group (e.g. dose, route of administration): capsules with the same appearance as prednisone Concomitant therapy: no information
Outcomes	Primary outcome Need to be admitted to hospital Death during the first 2 weeks after randomisation
Starting date	April 2021
Contact information	Carlos Cervera, MD, PhD: 780‐492‐5346 Email: cerveraa@ualberta.ca
Notes	Recruitment status: not yet recruiting Prospective completion date: August 2022 Date last update was posted: 18 March 2021 Sponsor/funding: University of Alberta

NCT04834375.

Study name	Randomised open investigation determining steroid dose (ROIDS‐Dose)
Methods	Trial design: randomised, open‐label trial Sample size: 142 Setting: probably inpatient Language: English Number of centres: single‐centre Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Adults ≥ 18 years old COVID‐19 infection confirmed by positive PCR test Hypoxaemia defined by an oxygen saturation < 94% or the need for supplemental oxygen Exclusion criteria Corticosteroid use for > 48 h within the past 15 days prior to enrolment Use of steroids with doses > the equivalent to dexamethasone 6 mg Use of immunosuppressive drugs Pregnant women Chronic oxygen use Known history of dexamethasone allergy Do‐not‐resuscitate/do‐not‐intubate orders Patient or proxy cannot consent
Interventions	Details of intervention Dose: dexamethasone 6 mg daily for 10 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone 0.2 mg/kg/d IV (maximum 20 mg daily) for 10 days Concomitant therapy: no information
Outcomes	Primary outcome: all‐cause mortality at 28 days
Starting date	19 March 2021
Contact information	Carlos X Rabascall, MD: 5164655400 Email: crabascallay@northwell.edu
Notes	Recruitment status: recruiting Prospective completion date: 19 April 2022 Date last update was posted: 8 April 2021 Sponsor/funding: Northwell Health

NCT04836780.

Study name	Dexamethasone early administration in hospitalised patients with COVID‐19 pneumonia (EARLYDEXCoV2)
Methods	Trial design: prospective, multicentre, phase‐4, parallel‐group, open‐label RCT Sample size: 126 Setting: inpatient Language: Spanish, English Number of centres: no information Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria Adults (age ≥ 18 years) Diagnosed with SARS‐CoV‐2 infection by PCR or rapid antigen test on upper respiratory tract (nasopharyngeal and oropharyngeal) specimens Evidence of infiltrates on chest radiography or CT Peripheral capillary oxygen saturation (SpO2) ≥ 94% and < 22 breaths per min (bpm) breathing room air High risk of developing ARDS defined by a LDH > 245 U/L, C‐RP > 100 mg/L, and absolute lymphocytes < 800 cells/µL Eligible participants will meet 2/3 above analytical criteria associated with severe COVID‐19 Patients will provide written informed consent or who have a legally authorised representative available to do so. In these exceptional circumstances and following the recommendations of the Spanish Agency of Medicines and Medical Devices, the National Competent Authority of clinical trials, during the coronavirus crisis to avoid the risk of contagion, consent will be possible to obtained orally in the presence of at least one impartial witness. Exclusion criteria Patients with a history of allergy to dexamethasone Pregnant or lactating women Oral or inhaled corticosteroids treatment within 15 days before randomisation Immunosuppressive agent or cytotoxic drug therapy within 30 days before randomisation Neutropenia < 1000 cells/µL HIV infection with CD4 cell counts < 500 cells within 90 days after randomisation Dementia Chronic liver disease defined by ALT or AST ≥ 5 times the ULN Chronic kidney injury defined by a glomerular filtration rate ≤ 30 mL/min, haemodialysis or peritoneal dialysis Uncontrolled infection Patients who are already enrolled in another clinical trial
Interventions	Details of intervention: Dose: dexamethasone base 6 mg once daily for 7 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): standard care therapy Concomitant therapy: no information
Outcomes	Primary outcome The primary trial outcome is the development of moderate‐severe ARDS (time frame: 7 days), based on the Berlin criteria The collected data as outcome measure will be general vital signs, Sequential Organ Failure Assessment (SOFA) score, the clinical status of the patient using the ordinal scale of the WHO, SpO2, PaO2/FiO2 ratio calculated from SpO2/FiO2, blood routine tests and chest radiography Concomitant drugs and adverse event monitoring will be collected Data will be measured during admission Participants will schedule for a follow‐up visit on the 30th and 90th day to track their long‐term prognosis, clinical status and sequelae
Starting date	No information
Contact information	Anabel Franco Moreno, MD, PhD: +34 911 918000 Email: afranco278@hotmail.com
Notes	Recruitment status: recruiting Prospective completion date: 30 June 2021 Date last update was posted: 8 April 2021 Sponsor/funding: Hospital Universitario Infanta Leonor

NCT04860518.

Study name	Human intravenous interferon beta‐Ia safety and preliminary efficacy in hospitalised subjects with Coronavirus (HIBISCUS)
Methods	Trial design: double‐blind, randomised controlled trial Sample size: 140 Setting: inpatient Language: English Number of centres: 4 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: Age ≥ 18 years Positive SARS‐CoV‐2 test by PCR (polymerase chain reaction) or other diagnostic method within the past 7 days Admission to hospital with respiratory symptoms of COVID‐19 requiring hospital care and oxygen supplementation (≤ 8L/min) Respiratory symptom onset no more than 7 days prior to hospital arrival Informed consent from the subject or the subject's personal legal representative or a professional legal representative must be available Exclusion criteria: Unable to screen, randomise, and administer study drug within 48 hours from arrival to hospital Systemic corticosteroid, baricitinib or tofacitinib (or other JAK‐STAT signalling pathway inhibitors) therapy within 7 days prior to arrival to hospital or planned for the next days Known hypersensitivity or contraindication to natural or recombinant IFN‐beta‐1a or its excipients, or to dexamethasone or its excipients Currently receiving IFN‐beta‐1a therapy Home assisted ventilation (via tracheotomy or non‐invasive) except for continuous positive airway pressure (CPAP)/bilevel positive airway pressure (BIPAP) used only for sleep‐disordered breathing Participation in another concurrent interventional pharmacotherapy trial during the study period Decision to withhold life‐sustaining treatment; patient not committed to full support (except DNR after cardiac arrest only) Woman known to be pregnant, lactating or with a positive pregnancy test (urine or serum test) Subject is not expected to survive for 24 hours Subject has liver failure (Child‐Pugh grade C) Any clinical condition that in the opinion of the attending clinician or Investigator would present a risk for the subject to participate in the study
Interventions	Details of intervention: interferon beta‐Ia Dose: 10 μg for 6 days while hospitalised Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone, IV, daily for 6 days while hospitalised, dose not stated Concomitant therapy: n.i.
Outcomes	Primary study outcome: clinical status at day 14 (first day of study drug is day 1) as measured by WHO 9‐point ordinal scale WHO 9‐point ordinal scale: 0 ‐ no detectable infection Not hospitalised, no limitations on activities Not hospitalised, limitation on activities Hospitalised, not requiring supplemental oxygen Hospitalised, requiring supplemental oxygen Hospitalised, on non‐invasive ventilation or high flow oxygen devices Hospitalised, on invasive mechanical ventilation Hospitalised, on mechanical ventilation plus additional organ support: renal replacement therapy (RRT), extracorporeal membrane oxygenation (ECMO) Death
Starting date	23 August 2021
Contact information	Jarna Hannukainen, PhD: +358 02 469 5151 Email: jarna.hannukainen@faron.com Gerd Tötterman: +358 02 469 5151 Email: gerd.totterman@faron.com
Notes	Recruitment status: recruiting Prospective completion date: April 2023 Date last update was posted: 3 January 2022 Sponsor/funding: Faron Pharmaceuticals Ltd

TCTR20211017001.

Study name	A comparative study of the effectiveness between pulse regimen methylprednisolone versus high dose dexamethasone as the initial treatment of moderate Covid‐19 pneumonia: an open‐label randomised controlled trial
Methods	Trial design: open‐label, randomised controlled trial Sample size: 120 Setting: inpatient Language: English Number of centres: 2 Type of intervention (treatment/prevention): treatment
Participants	Inclusion criteria: age over 18 years old, visited within 48 hours, respiratory sample was confirmed Covid‐19 by RT‐PCR, evidence of pneumonia by chest imaging, resting oxygen saturation between 90% and 94%, voluntary consent Exclusion criteria: received high flow nasal cannula, non‐invasive mechanical ventilation or invasive mechanical ventilation, high risk for corticosteroid such as current active infection or poor glycaemic control, immunocompromised host such as end stage liver disease, end stage renal disease without renal replacement therapy, HIV CD4 less than 200 or cancer with ongoing chemotherapy, pregnant women, psychiatric problems, current use of corticosteroid (more than 20 mg of prednisolone equivalent dose and consecutively more than 14 day)
Interventions	Details of intervention: methylprednisolone Dose: 250 mg per day for 3 days, then intravenous or oral dexamethasone 20 mg per day for 2 days, then intravenous or oral dexamethasone 10 mg per day for 5 days Route of administration: IV Treatment details of control group (e.g. dose, route of administration): dexamethasone, IV, 20 mg per day for 3 days, then intravenous or oral dexamethasone 20 mg per day for 2 days, then intravenous or oral dexamethasone 10 mg per day for 5 days Concomitant therapy: not stated
Outcomes	Primary study outcome: mean of WHO clinical progression scale at day 5
Starting date	18 October 2021
Contact information	Jakkrit Laikitmongkhon: 270 Rama 6 Rd. Phayatai Ratchathewi Bangkok 10400, Thailand
Notes	Recruitment status: not yet recruiting Prospective completion date: 31 December 2022 Date last update was posted: 16 November 2021 Sponsor/funding: Faculty of Medicine Ramathibodi Hospital

AHRF: acute hypercapnic respiratory failure; ALT: alanine transaminase; APPT: activated partial thromboplastin time; ARDS: acute respiratory distress syndrome; AST: aspartate transaminase; CT: computed tomography; d: days; ECG: echocardiogram; ECMO: extracorporeal membrane oxygenation; FiO2: fraction of inspired oxygen; HFNC: high‐flow nasal cannula; HFNO: high‐flow nasal oxygen; ICU: intensive care unit; IMV: invasive mechanical ventilation; IV: intravenous; LMWH: low molecular weight heparin; NLR: neutrophil‐lymphocyte ratio; NPPV: non‐invasive positive pressure ventilation; PCR: polymerase chain reaction; PEEP: positive end‐expiratory pressure; RCT: randomised controlled trial; RT‐PCR: reverse transcription polymerase chain reaction; SC: subcutaneously; SpO2: blood oxygen saturation; ULN: upper limit of normal; WHO: World Health Organization

Differences between protocol and review

Differences between protocol and first review version

Types of participants

For the different endpoints of interest and treatment settings, we excluded participants treated for symptoms of long‐COVID. However, treatments for long‐COVID should soon be addressed outside this review.

Types of outcome measures

Due to heterogeneous reporting and high bias through death as competing risk, we omitted most endpoints with regard to clinical improvement and worsening, and length of hospital stay. As a compromise and because of their importance to patients, data availability in trials, and strong implications for resource usage, we kept in the first publication of this living review the following outcome measures: new need for invasive ventilation, liberation from invasive ventilation, ventilator‐free days, need for dialysis, viral clearance, quality of life/neurological outcome.

We counted adverse events regardless of their grades because the included studies did not report grades for adverse events.

If not specified otherwise, the observation period for outcomes other than mortality was the longest period available. Where this differed between arms of a trial, risk of bias assessment would be adjusted.

Types of intervention

After discussion with clinical experts we added comparisons to meet questions arising in daily routine practice:

• dose comparisons;

• time comparisons (early versus late) but indirectly in terms of disease severity as described below;

• two different types of corticosteroids;

• corticosteroid versus another active substance (e.g. remdesivir, tocilizumab).

On the other hand, we excluded topical and inhaled steroids from this review because of inherently different pharmacokinetics and treatment settings. However, their role is critical and inhaled steroids were addressed in another review by our review group (Griesel 2022).  

Analysis

We made calculations with RevMan Web instead of RevMan 5.4 software.

We omitted subgroup analysis for treatment settings; that is, outpatient, inpatient, and intensive care unit, because we deemed triage criteria, definition of intensive care or high‐dependency units, and available resources too heterogeneous. Taking away a degree of indirectness, we instead performed subgroup analysis stratified by the level of respiratory support needed at randomisation. This allowed for the fact that levels of respiratory support can at least partially be delivered independent of the treatment setting and hence rendered a more valid conclusion about disease severity.

Differences between first and second review version

This version of the review puts more focus on health equity‐related aspects. We examined participant characteristics more closely and analysed mortality data with respect to equity (see below). Moreover, we put emphasis on possible harm through secondary infections especially in tropical low‐ and middle‐income settings in the discussion section. Finally, we strengthened the authoring team (previously solely authors from Germany and therefore with a high‐income background) with two Indian clinician scientists and guideline authors, and we adjusted the title of the review.

In addition, competing risk of death was extensively discussed in the first version of the review. The respective section was very much shortened but the impact on risk of bias assessments remains where the outcomes could not be changed or adjusted (see below).

To address additional sources of bias arising from platform trials we pioneered a checklist for their critical appraisal and added sensitivity analyses (see below).

Because we suspect a different immune response in vaccinated participants, we added vaccination status as an item for data extraction.

Types of outcome measures

Individuals with a suspected or confirmed diagnosis of COVID‐19 and moderate to severe disease: In line with the research gap identified in the first version of the review we added All‐cause mortality up to 120 days and retained All‐cause mortality up to 30 days for optimal evidence robustness. To reach a certain level of adjustment for competing risk of death, we added Discharged alive and changed New need for invasive mechanical ventilation (IMV) to New need for IMV or death. Because of compelling clinical and equity implications, we added Invasive fungal infections as a prioritised outcome included in the summary of findings tables while Quality of life is a prioritised outcome but no longer included in the summary of findings table. Need for dialysis and Viral clearance are categorised as additional outcomes. We deleted ventilator‐free days from the outcome set, because there is now New need for IMV or death as another outcome quasi‐adjusted for competing risk of death, which is more widely reported.

Certain outcomes are no longer additionally defined as changes in the COVID‐19‐specific WHO clinical progression scale (WHO 2020c), but as simple descriptions.

The outcome Need for dialysis was renamed New need for dialysis to clarify that participants with pre‐existing need for dialysis were excluded from this analysis. Data remained unchanged.

Individuals with a suspected or confirmed diagnosis of SARS‐CoV‐2 infection and asymptomatic or mild disease: To better adjust our review to the priorities of outpatients after publication of a Cochrane Review on inhaled corticosteroids in COVID‐19 (Griesel 2022), we added Admission to hospital or death within 28 days and Quality of life as prioritised outcomes.

Subgroup analysis

We added all‐cause mortality subgroup analyses stratified by age group, sex, and ethnicity in all comparisons.

Sensitivity analysis

We added a new sensitivity analysis excluding platform trials from the mortality outcomes' analyses. Moreover, following Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022), we performed sensitivity analysis with a fixed‐effect versus a random‐effects model in this version of the review.

Contributions of authors

CW: screening, data extraction, characteristics of included studies/ongoing studies/studies awaiting classification, risk of bias assessment, meta‐analysis, conception and writing of the review, taking responsibility for reading and checking the review before submission

MG: data extraction, risk of bias assessment, appraisal of platform trials, meta‐analysis, conception and writing of the review, taking responsibility for reading and checking the review before submission

AM: clinical expertise, writing of the review, taking responsibility for reading and checking the review before submission

MIM: design and conduct of searches, drafting of search methods section, taking responsibility for reading and checking the review before submission

MSp: screening, data extraction, risk of bias assessment, appraisal of platform trials, taking responsibility for reading and checking the review before submission

ALF: data extraction, risk of bias assessment, clinical expertise, writing of the review, taking responsibility for reading and checking the review before submission

AAN: data extraction, risk of bias assessment, clinical expertise, writing of the review, taking responsibility for reading and checking the review before submission

JD: data extraction, risk of bias assessment, clinical expertise, writing of the review, taking responsibility for reading and checking the review before submission

MS: clinical expertise, writing of the review, taking responsibility for reading and checking the review before submission

NS: methodological expertise and advice, conception and writing of the review, taking responsibility for reading and checking the review before submission

FF: clinical expertise, writing of the review, taking responsibility for reading and checking the review before submission

Sources of support

Internal sources

• University Hospital of Cologne, Germany

  Cochrane Cancer, Department I of Internal Medicine

• University of Leipzig Medical Center, Germany

  Department of Anesthesiology and Intensive Care

• Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt‐Universität zu Berlin, Berlin, Germany, Germany

  Department of Infectious Diseases and Respiratory Medicine

• Christian Medical College, Vellore, Tamil Nadu, India

  Department of Respiratory Medicine and Department of Pulmonary Medicine

External sources

• Federal Ministry of Education and Research, Germany

  NaFoUniMedCovid19 (funding number: 01KX2021) part of the project „CEO‐Sys“

Declarations of interest

CW: Federal Ministry of Education and Research (grant/contract); staff of Cochrane Haematology.

MG: Bundesministerium für Bildung und Forschung (grant/contract); published Cochrane‐initiated Twitter posts and an upcoming Cochrane‐initiated podcast on systemic corticosteroids; Resident at the Department of Anesthesiology and Critical Care, University of Leipzig Medical Service; Member of the German Society of Anaesthesia and Intensive Care Medicine (Deutsche Gesellschaft für Anästhesiologie & Intensivmedizin, DGAI), which supports and promotes the German Clinical Practice Guideline on COVID‐19 Inpatient Therapy.

AM: no relevant interests; co‐ordination of Section COVRIIN and work in the office of STAKOB (Competence and Treatment Centres for high‐consequence infectious diseases) at Robert Koch‐Institute Centre for Biological Threats and Special Pathogens (ZBS), Section Clinical Management and Infection Control.

MIM: no relevant interests; performs editorial activities for reviews overseen by Cochrane Metabolic and Endocrine Disorders.

MSp: no relevant interests; Resident, Universitätsklinikum Leipzig.

ALF: Universitätsklinikum Leipzig AöR (employment); Fellowship in University Hospital Leipzig, 04103 Leipzig, Germany.

AAN: no relevant interests; part of Indian COVID guidelines, worked in the evidence synthesis team for inhaled steroids and HFNC versus NIV in COVID; works at the Department of Respiratory Medicine, Christian Medical College, Vellore.

JD: no relevant interests; Pulmonologist, Department of Pulmonary Medicine, CMC Vellore, India.

MS: no relevant interests; Medical Doctor, Charité – Universitätsmedizin Berlin, Germany.

NS: no relevant interests; Editor of Cochrane Haematology but was not involved in the editorial process for this review.

FF: no relevant interests; Intensive Care Consultant, University Hospital, University of Leipzig Medical Faculty.

contributed equally (first author)

contributed equally (last author)

New search for studies and content updated (conclusions changed)