Obesity as an independent risk factor for COVID‐19 severity and mortality

Abstract

Background

Since December 2019, the world has struggled with the COVID‐19 pandemic. Even after the introduction of various vaccines, this disease still takes a considerable toll. In order to improve the optimal allocation of resources and communication of prognosis, healthcare providers and patients need an accurate understanding of factors (such as obesity) that are associated with a higher risk of adverse outcomes from the COVID‐19 infection.

Objectives

To evaluate obesity as an independent prognostic factor for COVID‐19 severity and mortality among adult patients in whom infection with the COVID‐19 virus is confirmed.

Search methods

MEDLINE, Embase, two COVID‐19 reference collections, and four Chinese biomedical databases were searched up to April 2021.

Selection criteria

We included case‐control, case‐series, prospective and retrospective cohort studies, and secondary analyses of randomised controlled trials if they evaluated associations between obesity and COVID‐19 adverse outcomes including mortality, mechanical ventilation, intensive care unit (ICU) admission, hospitalisation, severe COVID, and COVID pneumonia. Given our interest in ascertaining the independent association between obesity and these outcomes, we selected studies that adjusted for at least one factor other than obesity. Studies were evaluated for inclusion by two independent reviewers working in duplicate. 

Data collection and analysis

Using standardised data extraction forms, we extracted relevant information from the included studies. When appropriate, we pooled the estimates of association across studies with the use of random‐effects meta‐analyses. The Quality in Prognostic Studies (QUIPS) tool provided the platform for assessing the risk of bias across each included study. In our main comparison, we conducted meta‐analyses for each obesity class separately. We also meta‐analysed unclassified obesity and obesity as a continuous variable (5 kg/m² increase in BMI (body mass index)). We used the GRADE framework to rate our certainty in the importance of the association observed between obesity and each outcome. As obesity is closely associated with other comorbidities, we decided to prespecify the minimum adjustment set of variables including age, sex, diabetes, hypertension, and cardiovascular disease for subgroup analysis. 

Main results

We identified 171 studies, 149 of which were included in meta‐analyses.  As compared to 'normal' BMI (18.5 to 24.9 kg/m²) or patients without obesity, those with obesity classes I (BMI 30 to 35 kg/m²), and II (BMI 35 to 40 kg/m²) were not at increased odds for mortality (Class I: odds ratio [OR] 1.04, 95% confidence interval [CI] 0.94 to 1.16, high certainty (15 studies, 335,209 participants); Class II: OR 1.16, 95% CI 0.99 to 1.36, high certainty (11 studies, 317,925 participants)). However, those with class III obesity (BMI 40 kg/m² and above) may be at increased odds for mortality (Class III: OR 1.67, 95% CI 1.39 to 2.00, low certainty, (19 studies, 354,967 participants)) compared to normal BMI or patients without obesity. For mechanical ventilation, we observed increasing odds with higher classes of obesity in comparison to normal BMI or patients without obesity (class I: OR 1.38, 95% CI 1.20 to 1.59, 10 studies, 187,895 participants, moderate certainty; class II: OR 1.67, 95% CI 1.42 to 1.96, 6 studies, 171,149 participants, high certainty; class III: OR 2.17, 95% CI 1.59 to 2.97, 12 studies, 174,520 participants, high certainty). However, we did not observe a dose‐response relationship across increasing obesity classifications for ICU admission and hospitalisation.

Authors' conclusions

Our findings suggest that obesity is an important independent prognostic factor in the setting of COVID‐19. Consideration of obesity may inform the optimal management and allocation of limited resources in the care of COVID‐19 patients.

Plain language summary

Obesity and adverse COVID‐19 outcomes

What are the effects of obesity on COVID‐19 outcomes?

Key messages

• There is enough evidence to support the finding that extreme obesity (BMI > 40 kg/m²) increases the chance of a person dying, requiring a breathing tube, being hospitalised, and being admitted to the ICU due to COVID‐19. 

• Obesity in general will result in a person requiring a breathing tube.

• The higher one's BMI gets, the higher the chance that a person will suffer from severe COVID‐19 disease.  

What is obesity?

Obesity is defined as abnormal or excessive fat accumulation in different parts of the human body and it presents a risk to health. To assess obesity, different indices such as body mass index (BMI) can be used, which is one's weight in kilograms divided by the square of height in metres. The WHO has classified obesity into three classes. According to this classification, class I obesity includes a BMI ranging from 30 to 35 kg/m², class II from 35 to 40 kg/m², and class III from 40 kg/m² and more.

What did we want to find out?

We wanted to find out whether obesity has any effects on mortality, requiring a breathing tube, hospitalisation, ICU admission, severe disease or pneumonia due to COVID‐19 disease.  

What did we do?

We conducted a systematic search in medical databases for evidence looking at the association of obesity and mortality and other outcomes from December 2019 to April 2021. We then categorised and rated these findings based on our confidence in the evidence, study size, and quality.

What did we find?

We identified 171 eligible studies, with 149 studies (12,045,976 participants) providing quantitative data for at least one of our meta‐analyses. In terms of the outcomes, 111 studies reported on mortality, 48 on requiring a breathing tube, 47 on ICU admission, 34 on hospitalisation, 32 on severe COVID‐19, six on pneumonia, five on length of hospitalisation, two on length of ICU admission, and one on the duration of the requirement of a breathing tube. 

Main results

Our findings indicate that there is a high certainty of evidence that class III obesity is associated with an increased risk of mortality among COVID‐19 patients. However, we found that, in mild cases of obesity (classes I and II), this factor might not be independently associated with increased risk of mortality in COVID‐19 patients. Similarly, we are very certain that obesity is an independent important factor associated with the risk of requiring a breathing tube in COVID‐19 patients. However, the effect estimate sizes were not consistent with a dose‐response relationship across increasing obesity classes for ICU admission, hospitalisation, severe COVID‐19 disease and pneumonia. To conclude, this review investigated the potential association between obesity and adverse COVID‐19 outcomes. We were able to gather evidence from multiple studies and concluded that the association of obesity with mortality and requiring a breathing tube is of high certainty.

What are the limitations of the evidence?

Although BMI is a widely used measurement, the relationship between BMI and body fat is non‐linear. Moreover, our review did not discriminate against self‐reported and measured BMI. Finally, we were unable to keep up with the rapid pace of publications on COVID‐19 despite our best efforts. 

How up‐to‐date is the evidence?
The evidence is up‐to‐date to April 2021.

Summary of findings

Summary of findings 1. Obesity Class I compared to Normal Weight or Non‐Obese for Adults with COVID‐19.

Obesity class I (30 kg/m2 ≤ BMI < 35 kg/m2) compared to normal BMI or patients with a BMI < 30 kg/m2 for adults with COVID‐19
Patient or population: Adults with COVID‐19  Settings: Community and in‐hospital
Outcomes Time frame of absolute effects	Absolute effects from study(ies)* (95% CI)			Relative effect 95% CI	No of Participants (studies)	Quality of the evidence (GRADE)	Plain language summary
	Normal BMI or Non‐Obese	Obesity Class I	Difference with Obesity Class I
Mortality (in‐hospital)	180 per 1000	186 per 1000	6 more per 1000 (9 fewer to 23 more)	Odds Ratio: 1.04(CI 95% 0.94 to 1.16)1	335,209 (15)	⊕ ⊕ ⊕ ⊕ HIGH2	Obesity class I has little or no difference on mortality.
Mechanical ventilation (in‐hospital)	198 per 1000	254 per 1000	56 more per 1000 (31 more to 84 more)	Odds Ratio: 1.38(CI 95% 1.2 to 1.59)3	187,895 (10)	⊕ ⊕ ⊕ ⊖ MODERATE4	Obesity class I probably increases the risk of mechanical ventilation.
ICU admission (in‐hospital)	208 per 1000	263 per 1000	55 more per 1000 (10 more to 107 more)	Odds Ratio: 1.36(CI 95% 1.06 to 1.75)5	162,741 (7)	⊕ ⊕ ⊕ ⊖ MODERATE6	Obesity class I probably increases the risk of ICU admission.
Hospitalisation (30‐days, community)	146 per 1000	141 per  1000	5 fewer per 1000 (23 fewer to 17 more)	Odds Ratio: 0.96(CI 95% 0.82 to 1.14)7	515,155 (5)	⊕ ⊕ ⊕ ⊖ MODERATE8	Obesity class I probably has little or no difference on the risk of hospitalisation.
Severe COVID‐19 (in‐hospital)	158 per 1000	217 per  1000	59 more per 1000 (21 more to 102 more)	Odds Ratio: 1.48(CI 95% 1.16 to 1.87)9	1040 (3)	⊕ ⊕ ⊖ ⊖ LOW10	Obesity class I may increase the risk of severe Covid‐19.
Pneumonia	‐	‐	‐	‐	‐	‐	No studies were found that looked at the impact of obesity class I on pneumonia.
*The basis for the control group absolute risks from the study(ies) is mean risk across study(ies) unless otherwise stated in comments. The intervention absolute risk and difference is based on the risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working User Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

1.Systematic review. Baseline/comparator control arm of reference used for intervention. 

2. No reasons to rate down.

3. Systematic review. Baseline/comparator control arm of reference used for intervention. 

4. Imprecision: serious. The lower bound 95% CI crossed our absolute risk difference of 50 per 1000 patients followed.

5. Systematic review. Baseline/comparator control arm of reference used for intervention.

6. Inconsistency: serious. Lack of overlap in point estimates and 95% CI across studies. Our subgroup analyses failed to explain the observed heterogeneity. Low credibility (based on ICEMAN) for one statistically significant subgroup analysis based on the reference group. 

7. Systematic review. Baseline/comparator control arm of reference used for intervention.

8. Inconsistency: serious. A significant subgroup effect was observed based on the adjustment criteria. The credibility of this subgroup effect, however, was low (based on ICEMAN). 

9. Systematic review. Baseline/comparator control arm of reference used for intervention. 

10. Risk of Bias: serious. 2 of 3 studies were at an overall high risk of bias. Imprecision: serious. The lower bound 95% CI crossed our prespecified absolute risk difference threshold of 50 per 1000 patients followed. 

Note: BMI ‐ Body Mass Index; CI ‐ Confidence Interval; GRADE ‐ Grading of Recommendations, Assessment, Development and Evaluations

Summary of findings 2. Obesity Class II compared to Normal Weight or Non‐Obese for Adults with COVID‐19.

Obesity Class II (35 kg/m2 ≤ BMI < 40 kg/m2) compared to normal BMI or patients with a BMI < 30 kg/m2 for adults with COVID‐19
Patient or population: Adults with COVID‐19 Settings: Community and in‐hospital
Outcomes Time frame of absolute effects	Absolute effects from study(ies)* (95% CI)			Relative effect 95% CI	No of Participants (studies)	Quality of the evidence (GRADE)	Plain language summary
	Normal BMI or Non‐Obese	Obesity Class II	Difference with Obesity Class II
Mortality (in‐hospital)	180 per 1000	203 per 1000 (178 to 244)	23 more per 1000 (1 fewer to 50 more)	Odds Ratio: 1.16(CI 95% 0.99 to 1.36)1	317,925 (11)	⊕ ⊕ ⊕ ⊕ HIGH2	Obesity class II has little or no difference on the risk of mortality.
Mechanical ventilation (in‐hospital)	198 per 1000	292 per 1000 (281 to 388)	94 more per 1000 (62 more to 128 more)	Odds Ratio: 1.67(CI 95% 1.42 to 1.96)3	171,149 (6)	⊕ ⊕ ⊕ ⊕ HIGH4	Obesity class II increases the risk of mechanical ventilation.
ICU admission (in‐hospital)	208 per 1000	211 per 1000 (187 to 239)	3 more per 1000 (17 fewer to 24 more)	Odds Ratio: 1.02(CI 95% 0.9 to 1.15)7	157,665 (4)	⊕ ⊕ ⊖ ⊖ LOW6	Obesity class II may have little or no difference in the risk of ICU admission.
Hospitalisation (30‐days, community)	209 per 1000	216 per 1000 (188 to 250)	7 more per 1000 (17 fewer to 32 more)	Odds Ratio: 1.04(CI 95% 0.9 to 1.2)9	293,707 (3)	⊕ ⊕ ⊕ ⊖ MODERATE8	Obesity class II probably has little or no difference on the risk of hospitalisation.
Severe COVID	‐	‐	‐	‐	‐	‐	No studies were found that looked at the impact of obesity class II on severe COVID.
Pneumonia	‐	‐	‐	‐	‐	‐	No studies were found that looked at the impact of obesity class II on pneumonia.
*The basis for the control group absolute risks from the study(ies) is mean risk across study(ies) unless otherwise stated in comments. The intervention absolute risk and difference is based on the risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working UserGroup grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

1. Systematic review. Baseline/comparator control arm of reference used for intervention. 

2. No reasons to rate down. 

3. Systematic review. Baseline/comparator control arm of reference used for intervention. 

4. No reasons to rate down. 

5. Systematic review. Baseline/comparator control arm of reference used for intervention. 

6. Risk of Bias: very serious. There were only 4 studies in the meta‐analysis and about 70% of the total study weights came from the 3 studies at a high risk of bias.

7. Systematic review. Baseline/comparator control arm of reference used for intervention. 

8. Inconsistency: serious. One study that did not adjust for the prespecified adjustment set and was at low risk of bias showed a statistically significant larger effect size. The credibility of this subgroup effect, however, was low (based on ICEMAN). Therefore, we decided to rate down for inconsistency.

Note: BMI ‐ Body Mass Index; CI ‐ Confidence Interval; GRADE ‐ Grading of Recommendations, Assessment, Development and Evaluations

Summary of findings 3. Obesity Class III compared to Normal BMI, Non‐obese, or BMI < 40 for Adults with COVID‐19.

Obesity class III (BMI ≥ 40 kg/m2) compared to normal weight, patients with a BMI < 30 kg/m2, or BMI < 40 for adults with COVID‐19
Patient or population: Adults with COVID‐19 Settings: Community and in‐hospital
Outcomes Time frame of absolute effects	Absolute effects from study(ies)* (95% CI)			Relative effect 95% CI	No of Participants (studies)	Quality of the evidence (GRADE)	Plain language summary
	Normal BMI, Non‐obese, or BMI < 40	Obesity Class III	Difference with Obesity Class III
Mortality (in‐hospital)	180 per 1000	268 per 1000 (250 to 360)	88 more per 1000 (54 more to 125 more)	Odds Ratio: 1.67 (CI 95% 1.39 to 2.0)1	354,967 (19)	⊕ ⊕ ⊖ ⊖ LOW2	Obesity class III may increase the risk of mortality.
Mechanical ventilation (in‐hospital)	198 per 1000	349 per 1000 (314 to 588)	151 more per 1000 (84 more to 225 more)	Odds Ratio: 2.17 (CI 95% 1.59 to 2.97)3	174,520 (11)	⊕ ⊕ ⊕ ⊕ HIGH4	Obesity class III increases the risk of mechanical ventilation.
ICU admission (adjusted for at least DM, HTN, cardiovascular disease, age, and sex) (in‐hospital)	208 per 1000	240 per 1000 (201 to 309)	32 more per 1000 (5 fewer to 73 more)	Odds Ratio: 1.2 (CI 95% 0.97 to 1.49)5	155,405 (3)	⊕ ⊕ ⊕ ⊖ MODERATE6	Obesity class III probably has little or no difference on the risk of ICU admission when adjusted for DM, HTN, cardiovascular disease, age, and sex.
Hospitalisation (adjusted for DM, HTN, cardiovascular disease, age, and sex) (30‐days, community)	243 per 1000	302 per 1000 (289 to 369)	59 more per 1000 (33 more to 85 more)	Odds Ratio: 1.35 (CI 95% 1.19 to 1.52)7	293,004 (4)	⊕ ⊕ ⊖ ⊖ LOW8	Obesity class III may increase the risk of hospitalisation when adjusted for DM, HTN, cardiovascular disease, age, and sex.
ICU admission (in‐hospital)	208 per 1000	372 per 1000 (270 to 488 )	164 more per 1000 (62 more to 280 more)	Odds Ratio: 2.26 (CI 95% 1.41 to 3.63)	15,691 (7)	⊕ ⊕ ⊖ ⊖ LOW9	Obesity class III may increase the risk of ICU admission.
Hospitalisation (30‐days, community)	243 per 1000	362 per 1000 (310 to 420 )	119 more per 1000 (67 more to 177 more)	Odds Ratio: 1.77 (CI 95% 1.4 to 2.26)	747,176 (7)	⊕ ⊕ ⊖ ⊖ LOW10	Obesity class III may increase the risk of hospitalisation.
Severe COVID	‐	‐	‐	‐	‐	‐	No studies were found that looked at the impact of obesity class III on severe COVID.
.Pneumonia	‐	‐	‐	‐	‐	‐	No studies were found that looked at the impact of obesity class III on pneumonia.
*The basis for the control group absolute risks from the study(ies) is mean risk across study(ies) unless otherwise stated in comments. The intervention absolute risk and difference is based on the risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working UserGroup grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

1. Systematic review. Baseline/comparator control arm of reference used for intervention.

2. Inconsistency: serious. The visual inspection of the forest plot indicated considerable CIs not overlapping with a high I squared value. None of the prespecified subgroup analyses could explain the heterogeneity. Publication bias: serious. Asymmetrical funnel plot.

3. Systematic review. Baseline/comparator control arm of reference used for intervention.

4. No reasons to rate down.

5. Systematic review. Baseline/comparator control arm of reference used for intervention. 

6. Imprecision: serious. The confidence interval for absolute risk difference crossed the prespecified threshold of 50 per 1000 COVID patients.

7. Systematic review. Baseline/comparator control arm of reference used for intervention. 

8. Inconsistency: serious. Two studies had non‐overlapping CIs. The I‐squared value was considerably high. Imprecision: serious. A sensitivity analysis that omits the two studies with different CIs changes the interpretation of the results, therefore, we decided to rate down for imprecision and inconsistency twice. The CI also crossed the prespecified threshold.

9. Risk of Bias: serious. More than 50% of the total weight of the analysis fell on the studies at a high risk of bias. Furthermore, the effect estimate from the low risk of bias studies was too wide to provide a basis for comparison with the overall effect estimate. Inconsistency: serious. The effect estimates from different studies had minimal overlapping confidence intervals with very different indications for interpretation of the results.

10. Inconsistency: very serious. The effect estimate CIs from different studies were very far apart from each other with each warranting different interpretations of the results. Furthermore, the pooled effects for subgroups based on the adjustment set showed very different estimates.

Note: BMI ‐ Body Mass Index; CI ‐ Confidence Interval; GRADE ‐ Grading of Recommendations, Assessment, Development and Evaluations

Summary of findings 4. Obesity (Unclassified) compared to Normal Weight or Non‐Obese for Adults with COVID‐19.

Obesity (unclassified) compared to normal BMI or  patients with a BMI < 30 kg/m2 for adults with COVID‐19
Patient or population: Adults with COVID‐19 Settings: Community and in‐hospital
Outcomes Time frame of absolute effects	Absolute effects from study(ies)* (95% CI)			Relative effect 95% CI	No of Participants (studies)	Quality of the evidence (GRADE)	Plain language summary
	Normal BMI or Non‐Obese	Obesity (Unclassified)	Difference with Obesity (Unclassified)
Mortality (in‐hospital)	180 per 1000	229 per 1000 (230 to 255)	49 more per 1000 (39 more to 58 more)	Odds Ratio: 1.35 (CI 95% 1.28 to 1.42)1	1,307,520 (54)	⊕ ⊕ ⊖ ⊖ LOW2	Obesity (unclassified) may have little or no difference on the risk of mortality.
Mechanical ventilation (in‐hospital)	198 per 1000	294 per 1000 (285 to 394)	96 more per 1000 (64 more to 131 more)	Odds Ratio: 1.69 (CI 95%1.44 to 1.99)3	62,348 (21)	⊕ ⊕ ⊖ ⊖ LOW4	Obesity (unclassified) may increase the risk of mechanical ventilation.
ICU admission (in‐hospital)	208 per 1000	328 per 1000 (291 to 367)	120 more per 1000 (83 more to 159 more)	Odds Ratio: 1.86 (CI 95% 1.56 to 2.21)5	70,529 (21)	⊕ ⊕ ⊕ ⊖ MODERATE6	Obesity (unclassified) probably increases the risk of ICU admission.
Hospitalisation (adjusted for age, sex, DM, HTN, and cardiovascular disease) (30 days, community)	257 per 1000	312 per 1000 (308 to 370)	55 more per 1000 (36 more to 75 more)	Odds Ratio: 1.31 (CI 95% 1.2 to 1.44)7	510,405 (14)	⊕ ⊕ ⊕ ⊖ MODERATE8	Obesity (unclassified) probably increases the risk of hospitalisation.
Severe COVID (30‐days, community)	191 per 1000	314 per 1000 (309 to 443)	123 more per 1000 (86 more to 163 more)	Odds Ratio: 1.94 (CI 95% 1.62 to 2.32)9	878,804 (19)	⊕ ⊕ ⊕ ⊕ HIGH10	Obesity (unclassified) increases the risk of severe COVID.
Pneumonia (30‐days, community)	300 per 1000	382 per 1000 (363 to 516)	82 more per 1000 (41 more to 124 more)	Odds Ratio: 1.44 (CI 95% 1.21 to 1.72)11	35,924 (5)	⊕ ⊕ ⊕ ⊖ MODERATE12	Obesity (unclassified) probably increases the chance of pneumonia due to COVID.
Hospitalisation  (30 days, community)	257 per 1000	340 per 1000 (317 to 362)	83 more per 1000 (60 more to 105 more)	Odds Ratio: 1.31 (CI 95% 1.2 to 1.44)7	515,517 (20)	⊕ ⊕ ⊖ ⊖ LOW13	Obesity (unclassified) may increase the risk of hospitalisation.
*The basis for the control group absolute risks from the study(ies) is mean risk across study(ies) unless otherwise stated in comments. The intervention absolute risk and difference is based on the risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working UserGroup grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

1. Systematic review. Baseline/comparator control arm of reference used for intervention. 

2. Inconsistency: serious. There was a considerable I‐squared value. We decided to rate down twice for inconsistency and imprecision. Imprecision: serious. The pooled effect estimate was right below our prespecified threshold of absolute risk difference of 50 in every 1000 COVID patients.

3. Systematic review. Baseline/comparator control arm of reference used for intervention. 

4. Inconsistency: serious. Many CIs did not overlap and there was a high I‐squared value. The subgroup analysis did not explain the heterogeneity. Publication bias: serious. Asymmetrical funnel plot. 

5. Systematic review. Baseline/comparator control arm of reference used for intervention. 

6. Inconsistency: serious. The confidence interval of some of the studies did not overlap with those of most included studies/the point estimate of some of the included studies. The magnitude of statistical heterogeneity was high, with an I‐squared of about 80%.

7. Systematic review. Baseline/comparator control arm of reference used for intervention. 

8. Imprecision: serious. Even though a part of the width of the confidence interval was due to the heterogeneity, the pooled confidence interval spanned rather symmetrically around the absolute risk difference threshold of 50 in 1000.

9. Systematic review. Baseline/comparator control arm of reference used for intervention.

10. No reasons to rate down. 

11. Systematic review. Baseline/comparator control arm of reference used for intervention. 

12.Imprecision: serious. The lower bound of the confidence interval crossed the prespecified threshold of 50 in 1000 for the absolute risk difference.

13. Inconsistency: serious. Some of the confidence intervals did not overlap with the pooled confidence interval. These different studies require different and conflicting interpretation of their results. Publication bias: serious. An asymmetric funnel plot was observed.

Note: BMI ‐ Body Mass Index; CI ‐ Confidence Interval; GRADE ‐ Grading of Recommendations, Assessment, Development and Evaluations

Summary of findings 5. Every 5 units (kg/m²) increase in BMI compared to N/A for Adults with COVID.

Every 5 Units (kg/m2) Increase in BMI compared to N/A for Adults with COVID
Patient or population: Adults with COVID Settings: Community and in‐hospital
Outcomes Time frame of absolute effects	Absolute effects from study(ies)* (95% CI)			Relative effect 95% CI	No of Participants (studies), follow‐up	Quality of the evidence (GRADE)	Plain language summary
	N/A	Every 5 Units (Kg/m2) Increase in BMI	Difference with Every 5 Units (Kg/m2) Increase in BMI
Mortality (in‐hospital)	180 per 1000	203 per 1000 (194 to 223)	23 more per 1000 (12 more to 34 more)	Odds Ratio: 1.16(CI 95% 1.08 to 1.24)1	6,937,150 (10)	⊕ ⊕ ⊕ ⊕ HIGH2	Every 5 units (kg/m2) increase in BMI increases the risk of mortality.
Mechanical ventilation (in‐hospital)	195 per 1000	237 per 1000 (241 to 255)	42 more per 1000 (36 more to 46 more)	Odds Ratio: 1.28(CI 95% 1.24 to 1.31)3	13,527 (2)	⊕ ⊕ ⊕ ⊖ MODERATE4	Every 5 units (kg/m2) increase in BMI probably increases the risk of mechanical ventilation.
ICU admission	‐	‐	‐	‐	‐	‐	No studies were found that looked at the impact of continuous obesity on ICU admission.
Hospitalisation (30 days, community)	200 per 1000	226 per 1000 (209 to 226)	26 more per 1000 (8 more to 20 more)	Odds Ratio: 1.17(CI 95% 1.05 to 1.31)5	6,911,600 (3)	⊕ ⊕ ⊖ ⊖ LOW6	Every 5 units (kg/m2) increase in BMI may increase the risk of hospitalisation.
Severe COVID (community)	175 per 1000	292 per 1000 (243 to 472)	117 more per 1000 (53 more to 189 more)	Odds Ratio: 1.94(CI 95% 1.39 to 2.7)7	1041 (5)	⊕ ⊕ ⊖ ⊖ LOW8	Every 5 units (kg/m2) increase in BMI may increase the risk of severe COVID.
Pneumonia	‐	‐	‐	‐	‐	‐	No studies were found that looked at the impact of continuous obesity on pneumonia.
*The basis for the control group absolute risks from the study(ies) is mean risk across study(ies) unless otherwise stated in comments. The intervention absolute risk and difference is based on the risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working UserGroup grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

1. Systematic review. Baseline/comparator control arm of reference used for intervention. 

2. No reasons to rate down.

3. Systematic review. Baseline/comparator control arm of reference used for intervention.

4. Imprecision: serious. For this analysis, as the exposure is measured was a continuous variable, we considered no effect as the threshold. However, the limited number of studies and patients compelled us to rate down by one level.

5. Systematic review. Baseline/comparator control arm of reference used for intervention.

6. Inconsistency: serious. The direction of the effect was not consistent between the included studies. Imprecision: serious. Even though the total number of patients in the analysis was considerable, most came from one study.

7. Systematic review. Baseline/comparator control arm of reference used for intervention.

8. Risk of Bias: serious. Most of the weight of the analysis was built up of studies at a high risk of bias. Imprecision: serious. The number of participants in the analysis was low.

Note: BMI ‐ Body Mass Index; CI ‐ Confidence Interval; GRADE ‐ Grading of Recommendations, Assessment, Development and Evaluations

Background

            1.1. Brief description of the condition and context

In December 2019, a novel coronavirus (SARS‐CoV‐2) began causing respiratory infections in Wuhan, China. On February 11 2020, the World Health Organization (WHO) classified the virus causing these severe infections as COVID‐19 and declared a global pandemic on March 11, 2020 (Cascella 2022). Due to the novelty of COVID‐19, researchers around the world are working to understand the prognostic factors that are associated with COVID‐19 severity and mortality in order to develop suggestions and guidelines to promote the safety of the public. As of now, various vaccines that are effective against the virus have been introduced (CDC 2022). In many developed countries such as the United States (US), Canada, and the United Kingdom (UK), a considerable proportion of the population has been vaccinated. However, this has not yet stopped the emergence of new COVID‐19 variants of concern (WHO 2022). In fact, more than 2 years after the beginning of the pandemic, the high number of infections has imposed a crippling burden on healthcare systems around the world. As such, the COVID‐19 pandemic continues to utilise a great proportion of healthcare resources without a clear end in sight. Factors that can help in identifying the most vulnerable individuals may assist in better allocating the limited resources. To this end, the goal of our review is to evaluate the independent association between obesity with COVID‐19 outcomes.  

            1.2. Description of the prognostic factor

Obesity is a complex chronic condition associated with numerous predisposing factors such as genetics, social determinants (e.g. income, family eating patterns), environmental conditions (e.g. geographical region, access to transportation), and behavioural factors (e.g. sleep, sedentary lifestyles). The World Health Organization (WHO) defines obesity as a body mass index (BMI) of 30 kg/m² or higher (WHO 2021). This organisation, further, classifies obesity into three categories with increasing BMI (class I: BMI from 30 to 34.9 kg/m², Class II: BMI from 35 to 39.9 kg/m², and class III: BMI of 40 kg/m² and above). Importantly, the global prevalence of obesity is continually rising (Inoue 2018; WHO 2021; Wong 2020). In 2016, 13% of adults (aged 18 years and over) had obesity, and this is predicted to rise to 16.1% by 2025 (WHO 2021; World Obesity 2022). Moreover, obesity is classified by the WHO as a disease, and is also a known prognostic and risk factor for many health conditions, including diabetes, cardiovascular diseases, and cancer (Ayoub 2019; Cefalu 2015; Fruh 2017; Guh 2009). Patients with obesity have been shown to have dysregulated immune responses and chronic low‐grade inflammation accompanied by elevated levels of pro‐inflammatory cytokines, suggesting that these individuals may be more susceptible to hyperinflammation (Ritter 2020). Likewise, obesity is associated with reduced pulmonary function, including expiratory reserve volume and functional capacity (Dietz 2020). Both of these factors may contribute to poorer prognosis from COVID‐19 in patients with obesity (Muscogiuri 2022).

            1.3. Health outcomes

Even though the rate of severe adverse events among those who contract the virus remains relatively low, the very contagious nature of the disease has resulted in billions of infections and, consequently, a heavy burden on healthcare systems worldwide (CDC 2022; Chavez‐MacGregor 2021; Schneider 2021). The range of COVID‐19 symptoms varies considerably. About a third of the infected individuals remain asymptomatic, while the other two‐thirds could develop symptoms ranging from mild flu‐like symptoms to severe respiratory distress syndromes requiring oxygenation, hospitalisation, intensive care unit (ICU) admission, mechanical ventilation, and even death (Thevarajan 2020). Other serious complications following this infection include cardiac, thromboembolic, neurologic, and inflammatory complications. According to the data available from the United States Center for Disease Control (CDC), amongst the approximately 70 million reported cases of COVID disease in January 2022 in the US, more than 860,000 patients died (CDC 2022).

The time to recovery following initial symptoms can be quite different based on the patient's age, comorbidities, and severity of disease. Usually, healthy young adults would be symptom‐free in as soon as two weeks, while others might suffer from symptoms for much longer (Sykes 2021). 

            1.4. Why it is important to do this review

Emerging evidence suggests that obesity increases the severity of COVID‐19 and the risk of mortality. Several systematic reviews and meta‐analyses have reviewed studies assessing the prognostic value of obesity for COVID‐19 severity (Huang 2020; Hussain 2020; Izcovich 2020; Popkin 2020; Tamara 2020; Tocalini 2020; Yang 2020). Nonetheless, the studies in these reviews provide conflicting conclusions about the prognostic value of obesity. Some reviews suggest that individuals with obesity are at greater risk for adverse outcomes as compared to individuals with normal weight (Huang 2020; Izcovich 2020; Tamara 2020; Yang 2020). In contrast, a review by Tocalini 2020 noted that there is no significant association between obesity and in‐ICU COVID‐19 mortality. Thus far, however, few of these reviews considered the role of adjusting for covariates in the association between obesity and the COVID‐19 adverse outcomes. The observed heterogeneity and the consequent conflicting conclusions amongst the previous primary studies may be related to the extent of adjustment for other covariates. We conducted our systematic review with the aim of clarifying the inconsistency observed across studies by 1) using a more comprehensive search to identify all studies evaluating the association between obesity and adverse outcomes related to COVID‐19, and 2) investigating the role of obesity as an independent prognostic factor for COVID‐19 mortality and disease severity. Furthermore, to evaluate how reliable the results are and to assess the certainty in the body of evidence, we used the GRADE methodology (Guyatt 2008). The results of this review may provide evidence for guidelines on the management of COVID‐19 patients. 

Objectives

The main objective of this review was to evaluate obesity as an independent prognostic factor for COVID‐19 severity and mortality among adult patients in whom infection with the COVID‐19 virus is confirmed.

Methods

3.1 Protocol Registration

 We registered the protocol of our systematic review and meta‐analysis in the International Prospective Register of Systematic Reviews (CRD42020190687). This review is reported according to the Preferred Reporting Items for Systematic Review and Meta‐Analysis (PRISMA) guidelines (Page 2021).

3.2 Criteria for Considering Studies for This Review

            3.2.1. Eligibility Criteria

We used the following inclusion criteria for study selection: (i) Study design: case‐control, case‐series, prospective and retrospective cohort studies, registry data, and secondary analyses of randomised controlled trials; (ii) study participants over 18 years of age who had confirmed COVID‐19 infections; (iii) outcomes reported were at least one of the following: mortality, mechanical ventilation, ICU admission, severe COVID, and pneumonia; (iv) statistical analysis included multivariable analyses. 

According to the knowledge accumulated to date from the pandemic, children are at minimal risk of contracting the virus and experiencing its adverse outcomes. Therefore, we decided to include only adults above 18 years of age in our review (Drouin 2021). We specified that if more than 10% of an original study population consists of participants below 18 years of age, the study would be excluded. Furthermore, we only included studies that report the prognostic value of obesity among patients with confirmed COVID‐19. However, we did not limit the inclusion to the confirmation method used in the study. 

There are numerous methods used for measuring and classifying obesity. The most commonly used indicator is BMI, with further classification of obesity categories based on BMI thresholds. BMI was used to classify obesity into obesity classes I, II, and III.  According to the WHO classification, class I obesity includes the BMI range from 30 to 35 kg/m², class II from 35 to 40 kg/m², and class III 40 kg/m² and more (WHO 2000). Other indicators include waist circumference, waist‐to‐hip circumference ratio, or other adiposity indicators. We did not exclude the studies that did not use BMI. However, for meta‐analyses, we only used obesity categories based on BMI and BMI as a continuous variable. The main reason for this decision was to minimise the degree of expected clinical heterogeneity. Since some countries use lower thresholds for classifying obesity (lower than 30 kg/m²), we accepted the threshold set out in the original studies for our obesity categories. Finally, we included studies that measured and classified obesity prior to the occurrence of our outcomes of interest.

The primary outcomes prespecified in our study protocol were mortality (key outcome), mechanical ventilation, ICU admission, hospitalisation, oxygen requirement, and profound health complications resulting from COVID‐19. Due to the availability of resources and practice variations in the administration of oxygen, we decided to omit oxygen requirement as an outcome. We used severe COVID and pneumonia as profound COVID‐19 complications because of the availability of data. Furthermore, we included both the binary reporting of hospitalisation and ICU admission, as well as their continuous measures, length of hospitalisation, and ICU admission. We did not place any restriction on the timing of outcome assessment, and included studies with any follow‐up period.

This study aimed to further add to the current understanding of the effects of obesity by investigating its independent association with the outcomes. To report independent associations, the most common method that researchers use in observational studies is to statistically adjust the effect estimate for other possible prognostic factors. Therefore, we decided to only include original studies that at least adjusted their reported obesity effect estimate for another factor. In other words, we only included studies that investigated the (more) independent association between obesity and the outcomes by conducting a multivariable analysis. Due to the large number of included studies, we did not request data that we could not find from the articles from the study authors, for feasibility reasons. 

3.3 Search Methods for Identification of Studies

            3.3.1. Electronic Databases and Other Sources

Two information specialists developed and conducted systematic searches in the following English and Chinese databases in April 2021 for studies without language or publication status restrictions:

• MEDLINE (via PubMed) (1 Dec 2019 to 23 April 2021)

• Embase (via Ovid) (1 Dec 2019 to 22 April 2021)

• Cochrane COVID‐19 Study Register  (https://covid-19.cochrane.org/) (searched on 23 April 2021)

• COVID‐19 Global literature on coronavirus disease (https://search.bvsalud.org/global-literature-on-novel-coronavirus-2019-ncov/) (searched on 23 April 2021)

• China Network Knowledge Infrastructure (CNKI) (http://www.cnki.net/) (until 22 April 2021)

• Chinese Scientific Journals Database (VIP) (http://www.cqvip.com/) (until 22 April 2021)

• Wanfang data (http://www.wanfangdata.com.cn/index.html) (until 22 April 2021)

• SinoMed (http://www.sinomed.ac.cn/) (until 22 April 2021)

Our detailed search strategy is available as supplemental material in Appendix 1.  For greater precision, we searched MEDLINE and Embase from 1 December 2019 onwards as this is the time when COVID‐19 became well known. All other sources were searched from inception and without date limits. We elected to employ a Chinese search strategy in four Chinese biomedical databases due to a substantial proportion of the evidence stemming from China at the time of the search.

            3.3.2. Other Sources

We also searched the indexed conference abstracts for any additional published papers that were not identified in our electronic search. We included abstracts (otherwise not published) that provided sufficient data for inclusion in our review. 

            3.3.3. Screening

The results of the systematic searches were screened independently in duplicate according to the eligibility criteria in two stages: title and abstract screening, and full‐text screening. We used prespecified question sets for checking the eligibility of studies. These sets were piloted before the screening to ensure a similar understanding among the screeners. In case of any disagreement between the initial screeners, they were instructed to discuss the reasons for their decisions. If a unanimous decision was not reached after the discussion, an adjudicator with expertise in methodology would settle the disagreement. To facilitate the screening process, we used Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia. Available at www.covidence.org (Covidence).

             3.3.4. Inclusion of non‐English language studies

We did not restrict our inclusion to any specific language. When the English text of a paper or abstract was not available, they were screened by authors with adequate knowledge of the relevant language for screening and data extraction.

3.4 Data collection and analysis

We designed an electronic data extraction sheet in Microsoft Excel based on the CHARMS‐PF checklist. This checklist is modified by Riley 2019 from the CHARMS checklist for extracting data from prediction model studies (Moons 2014). The modified checklist is specifically geared toward prognostic factor reviews. Before starting data extraction, we piloted the extraction sheet among the authors to ensure a similar understanding. Data extractors were divided into pairs. Each pair extracted data from the studies in duplicate and independently. Similar to the screening phase, after individual data extraction, disagreements were resolved through pair discussion or third member adjudication, if a unanimous decision was not reached. 

For each included study, reviewers extracted the information about study publication, study design and settings, study recruitment and follow‐up, study population characteristics, prognostic factor measurement and definition, outcome definition, prognostic factor and outcome prevalence, the unadjusted and adjusted magnitude of effect, adjustment method, adjusted covariates, and missing data. A blank copy of the extraction sheet is provided in the supplementary material as Appendix 2.            

3.5. Assessment of Risk of Bias in the Included Studies

Two reviewers independently assessed the risk of bias in the included studies using the Quality in Prognostic Studies (QUIPS) tool (Hayden 2013). The QUIPS instrument classifies the risk of bias based on six domains: participant selection, attrition, prognostic factor measurement, outcome measurement, study confounding, statistical analysis, and reporting. Each domain can be rated as low, moderate, or high risk of bias. For each study, we rated the overall risk of bias as either high or low. When five QUIPS domains or more were at low risk of bias or only two domains were at moderate risk of bias, we classified the study as at an overall low risk of bias. Otherwise, the study was considered to be at a high overall risk of bias. To facilitate the assessment of the risk of bias, we extracted the suggested information regarding the risk of bias suggested by the CHARMS‐PF checklist. We also reported the risk of bias for each study in the forest plots.

3.6. Measures of association

We extracted the magnitude of associations using odds ratios (ORs), hazard ratios (HRs), and relative risks (RRs) for the purpose of meta‐analysis. Other measures of association were also gathered for narrative review. We employed validated statistical methods that use each study event rate, the prevalence of prognostic factors, and adjusted effect estimates to calculate the baseline risk of the outcome (Absolute Risk Calculator 2020; Foroutan 2020b). Consequently, using the mentioned values, HRs and RRs were converted to ORs. The reason to undertake these transformations was to enable pooling from a wider range of studies including different methodological designs. 

3.7. Dealing with Missing Data

We extracted the available information about the missing prognostic factors and outcomes data from the included studies’ text. This information was incorporated into the assessment of the risk of bias. Due to the large volume of included studies, we did not ask the original study authors for further information regarding missing data. We also did not apply other methods such as imputations due to the scarcity of sufficient information on the missing data.

3.8. Assessment of Heterogeneity 

We explored statistical heterogeneity by visually inspecting forest plots, considering the consistency of point estimates and the extent of overlap in confidence intervals. Even though we calculated statistical measures of heterogeneity such as I² and the chi² test significance value, we mostly relied on visual inspection since these measures in meta‐analyses of observational studies are usually very large and not helpful. We did not use strict thresholds for the statistical indices. These values along with the visual inspection of the forest plots provided the basis for interpretation of the amount of heterogeneity in a minimally contextualised review setting (Iorio 2015).

3.9. Assessment of Reporting Deficiencies

In order to assess the risk of bias in reporting the individual studies, we considered the reporting and statistical domain of the QUIPS tool. As for the other domains, the detailed guiding questions directed the final domain judgement. Furthermore, whenever 10 or more studies informed a meta‐analysis, we built a funnel plot (Sterne 2011). The symmetry and distribution of the effect estimates informed our evaluation of publication bias. We did not rely on other tests for the risk of publication bias.

3.10. Data Synthesis

             3.10.1 Data Synthesis and Meta‐Analysis Approaches

We used Stata 2015 for all analyses. We conducted random‐effects model meta‐analyses using the Dersimonian and Laird method (DerSimonian 1986) that pooled the effect estimates from all eligible studies to obtain a summary estimate and confidence interval for each outcome. In addition, for the timing of reported outcomes (other than time‐to‐event analysis), we decided to consider the closest timing of the reported outcome to 30 days from symptom development. Relevant outcomes included hospitalisation, severe COVID, and pneumonia. Such studies reporting on multiple time points were scarce in this review.

             3.10.2. Subgroup Analysis and Investigation of Heterogeneity

We included three subgroup analyses in this review, where possible, which included statistical adjustment set (adjusting for a minimum set of covariates), risk of bias, and statistical reference group for obesity comparison. We specified the subgroup analyses before data extraction. During team discussion, the lack of a strong and unanimous rationale for subgroup effects compelled us to limit the number of subgroup analyses to minimise the chance of observing spurious subgroup effects. For each subgroup analysis, we used the Instrument for assessing the Credibility of Effect Modification Analyses (ICEMAN) tool to guide us to estimate the credibility of the observed subgroup effect (Schandelmaier 2020). This tool has been developed for the evaluation of the credibility of a subgroup effect observed in randomised trials or systematic reviews of randomised trials. Even though the tool has not been validated for the systematic reviews of observational studies, each item in the tool is relevant to our review. Furthermore, after discussions, we were unable to add any critical items that were missing from the tool. As we still needed some methods to evaluate the credibility of any observed subgroup effect, we decided to use the concepts and guidance from this tool in addition to the methodologic and clinical expertise of the authors. If moderate or high credibility in the observed subgroup effect emerged after duplicate evaluation, we would report the more appropriate subgroup effect estimate alongside the results from the overall analysis. Furthermore, low or moderate credibility in the subgroup effect was an indicator of possible inconsistency in certainty ratings. The decision to provide both the subgroup and overall effects relies on the fact that the ICEMAN tool has been developed for randomised trials and uncertainty around its performance with observational studies remains.

In the first analysis, we considered the adjustment subgroup. As obesity is closely associated with other comorbidities such as diabetes and cardiovascular diseases, we believe that it is crucial for original studies to at least adjust for some specific variables. Therefore, we decided to prespecify the minimum adjustment set of variables based on the knowledge of COVID‐19 disease. Our minimum adjustment set was defined as age, sex, diabetes, hypertension, and cardiovascular disease. This subgroup analysis was performed on all of the mentioned outcomes. Our hypothesis for the direction of this subgroup analysis was that not adjusting for the minimum set would overestimate the magnitude of the effect. It is noteworthy that although the issue of confounders is considered in the statistical analysis domain of the risk of bias evaluation tool, we feel that the importance of the issue in this setting is not sufficiently represented only by a domain of risk of bias. Current evidence strongly supports the role of comorbidities in COVID‐19 adverse outcomes (CDC 2020).

We hypothesised that if there is an association between BMI and COVID outcomes that extends even below the 30 kg/m 2 threshold, considering lower BMI reference ranges would mean that each obesity class magnitude of effect would be estimated larger.

Next, we conducted another subgroup analysis for the effect of the obesity reference group each original study adopted in their regression model. We hypothesised that, if there was an association between BMI and the risk of COVID‐19 outcomes that extended even below the 30 kg/m² threshold, considering lower BMI reference ranges would mean a larger estimated magnitude of association. The choice of this subgroup analysis was also based on considering a wide range for the comparator group. The results of the analysis could advise the appropriateness of pooling across studies with different reference groups. We considered the  group without obesity (BMI < 30 kg/m²) to be the more appropriate reference group. It is important to notice that the association that has been proposed between obesity and COVID adverse outcomes is J‐ shaped. This means that, potentially, both the very low and very high values of BMI can increase the risk of adverse outcomes from COVID. Therefore, comparing obesity classes to different ranges of baseline BMI (Bhaskaran 2018) might affect the magnitude of the effect differently. We believe conducting this subgroup analysis clearly presents how the choice of the reference group influences the magnitude of the effect estimates.

The last subgroup analysis of this review was the overall risk of bias as determined by the QUIPS tool. This subgroup analysis also provided required information for the decision about the overall risk of bias when rating the certainty in the evidence. 

             3.10.3. Sensitivity Analysis

We only used sensitivity analysis to evaluate the effects of excluding the studies that were believed to drive a considerable portion of overall heterogeneity by having either non‐overlapping confidence intervals or noticeably heterogeneous point estimates from the majority of the other studies. We only undertook these sensitivity analyses if the number of excluded studies was below three (Schandelmaier 2020). The consequent assessment of heterogeneity and the changes in the statistical measures of heterogeneity were considered when assessing the inconsistency of the effect estimates. 

We also encountered instances where included studies used BMI < 40 kg/m² as the reference group for the statistical analyses regarding the association between class III obesity and the outcomes. As this range does not completely overlap the non‐obese BMI range, we conducted sensitivity analyses to investigate the effects of removing these studies from the pooled estimates on the interpretation of our results. 

3.11 Conclusion and Summary of Findings

For each outcome, we used the Grading of Recommendations Assessment, Development and Evaluation (GRADE) guidance for prognostic factors to assess the certainty of the evidence (Foroutan 2020a). This methodology guides reviewers to systematically evaluate the certainty of the evidence they have gathered. The certainty levels defined by GRADE include high, moderate, low, and very low. For prognostic factor reviews, the evidence originating from observational studies starts at a high certainty of evidence. Subsequently, this certainty can be rated down or up based on specific grounds. Domains each can rate down the certainty by one or two levels and they include risk of bias in included studies, inconsistency across the effect estimates, imprecision of the pooled estimate, indirectness of the gathered evidence, or risk of publication bias. On the other hand, the situations that allow rating up include a large magnitude of observed effect size, dose‐response, and the nature of plausible confounders. In order to rate our certainty in the minimally contextualised framework, we converted the relative effects to absolute risks by using the relative measure of association (OR), the median prevalence of obesity class in patients with COVID‐19, and the median of the overall risk of the outcomes across the included studies (Caussy 2020; Foroutan 2020b). To assess imprecision, we determined an absolute risk increase of 5% to be our threshold for a clinically meaningful prognostic factor (or 50 in 1000 as indicated in the summary of findings tables). When the subgroup analysis for risk of bias showed a significant difference across the high and low risk of bias studies, we used the analysis and conclusions of the meta‐analysis of studies at low risk of bias. To prepare the evidence tables presented here we used the Magic online review and guideline management software (Magic 2022). In the summary of findings tables, mortality, mechanical ventilation, hospitalisation, ICU admission, pneumonia, and severe COVID were considered as outcomes.

Results

 4.1. Description of Studies

4.1.1. Results of the Search

Our searches yielded 10,514 unique records. After the first stage of screening, 495 studies proceeded to full‐text screening and 171 were included in the review (Figure 1). One hundred and forty‐nine studies with a total of 12,045,976 participants provided quantitative data for at least one of our meta‐analyses. Three manuscripts were in Chinese and were reviewed by a pair of review authors proficient in Chinese. More details of each included study including their set of adjusted covariates can be found in the Characteristics of included studies.

1.

Study flow diagram

4.1.2. Included studies

Among the 171 included studies, only three studies used a case‐control design while 35 and 133 studies were registry data and cohort studies, respectively. Amongst the cohorts, 101 studies were retrospective cohorts and 32 were prospective. More than half of the included studies were conducted in the United States or China (Table 6). One hundred and fifty‐one studies used BMI as an indicator of obesity. Amongst these, 133 used obesity categorisations according to BMI, and 22 used BMI as a continuous variable in their analyses. In terms of the outcomes, 111 studies reported on mortality, 48 on mechanical ventilation, 47 on ICU admission, 34 on hospitalisation, 32 on severe COVID, six on pneumonia, five on length of hospitalisation, two on length of ICU admission, and one on length of mechanical ventilation. As per our inclusion criteria, all these studies adjusted for at least one covariate in addition to obesity. A table of characteristics of included studies provides the full details of each included study including the set of adjusted covariates (Characteristics of included studies).

1. Overall characteristics of included studies.

DESIGN	Number of studies
Prospective cohort	32
Retrospective cohort	101
Registry data	35
Case‐control	3
Total	171
SETTING	Number of studies
Inpatient	120
Outpatient	4
Outpatient and inpatient	47
Unspecified	5
Total	176
OBESITY MEASUREMENT TIME	Number of studies
Before or right at presentation	84
After presentation	15
Unspecified	74
Total	173
COUNTRY	Number of studies
US	70
China	18
UK	14
Italy	11
Spain	10
France	9
Mexico	8
Brazil	3
Iran	3
International	3
Germany	2
Netherlands	2
Ireland	2
Turkey	2
South Korea	2
Thailand	2
Kuwait	1
Belgium	1
Oman	1
Bolivia	1
Congo	1
India	1
Israel	1
Morocco	1
Qatar	1
Sweden	1
Total	171

4.1.3. Excluded Studies

In total, 324 studies were excluded after reviewing full texts. Reasons for this include inappropriate study design (n = 184), inappropriate study population (n = 31), the inclusion of duplicate records (n = 19), inappropriate statistical analysis (n = 73, only univariate analysis), and not reporting on any of the prespecified outcomes (n = 17). We have provided examples of eight key studies that were excluded in Characteristics of excluded studies. 

4.1.4. Risk of Bias in Included Studies

We used the QUIPS tool to assess the risk of bias in the included studies. We assigned scores of 0, 1, and 2 to each domain at a low, medium, and high risk of bias respectively. The sum of domain scores of more than 2 indicated the studies were at a high risk of bias. We judged 54% of all reported outcomes at low risk of bias (Risk of Bias Assessments). Regarding the ICU admission, most of the studies had low risk of bias (n = 27, 57%). Most of the studies that measured mortality, hospitalisation and mechanical ventilation had low risk of bias (56%, 57%, and 52%, respectively). We found one study that measured the length of mechanical ventilation, which we judged to have a low risk of bias. For pneumonia and length of ICU stay, 50% of studies had a low risk of bias. However, regarding the length of hospitalisation and severe Covid‐19 outcomes, most of the studies had a high risk of bias (60% and 58%, respectively). Table 7, Table 8, Figure 2, and Figure 3 summarise the judgements about the overall risk of bias and risk of bias domains judgments for each outcome. We used funnel plots to investigate the risk of publication bias for comparisons that included more than 10 studies (Appendix 3).

2. Overall risk of bias per outcome.

Outcome	Overall risk of bias (%)
	Low		High
	Number of studies	%	Number of studies	%
ICU admission	27	57.45	20	42.55
Hospitalisation	19	57.58	14	42.42
Length of ICU admission	1	50	1	50
Length of hospitalisation	2	40	3	60
Length of mechanical ventilation	1	100	0	0
Mechanical ventilation	26	52	24	48
Mortality	64	56.64	49	43.36
Pneumonia	3	50	3	50
Severe COVID	13	41.94	18	58.06
Total	156	54.17	132	45.83

3. Domain‐specific risk of bias per outcome.

	QUIPS domains
	Selection bias			Attrition bias			PF measurement bias			Outcome measurement bias			Confounding bias			Statistical analysis bias
	Low	Moderate	High	Low	Moderate	High	Low	Moderate	High	Low	Moderate	High	Low	Moderate	High	Low	Moderate	High
ICU admission	30	14	3	20	20	7	26	12	9	41	6	0	28	11	8	36	6	5
Hospitalisation	22	8	3	15	13	5	15	7	11	30	3	0	22	8	3	23	7	3
Length of ICU admission	0	2	0	2	0	0	2	0	0	1	1	0	1	1	0	1	1	0
Length of hospitalisation	3	1	1	3	2	0	3	2	0	5	0	0	2	1	2	3	2	0
Length of mechanical ventilation	1	0	0	1	0	0	0	1	0	1	0	0	0	1	0	1	0	0
Mechanical ventilation	37	10	3	24	20	6	26	13	11	41	9	0	28	12	10	39	7	4
Mortality	76	26	11	66	35	12	67	22	24	108	3	2	69	26	18	75	23	15
Pneumonia	4	2	0	3	2	1	2	3	1	4	1	1	3	1	2	3	2	1
Severe COVID	19	12	0	16	11	4	20	3	8	22	6	3	17	8	6	21	5	5
Total	192	75	21	150	103	35	161	63	64	253	29	6	170	69	49	202	53	33

ICU: Intensive care unit
PF: Prognostic factor
QUIPS: Quality in Prognostic Studies tool

2.

3.

4.2. Results of the Synthesis

Below, we provide a summary of the results of all the outcomes of interest to the review. Detailed findings for the outcomes we considered to be of greater importance to the patients (mortality, mechanical ventilation, and ICU admission) are presented in this section, with brief interpretations of the findings of the remaining outcomes. Further details on the findings for the remaining outcomes are detailed in the Summary of findings tables and forest plots.

4.2.1. Mortality

In total, 105 studies reported effect sizes on the association between obesity and mortality. The majority of the studies reported at least one association (n = 61) between unclassified obesity and mortality due to COVID‐19. Further, we were able to extract 15, 12, and 23 reported effect estimates for obesity class I, II, and III meta‐analyses, respectively.

4.2.1.1. Obesity Class I

We included 15 studies to estimate a pooled effect estimate for the association between obesity class I (versus patients without obesity) and COVID‐19 mortality. These studies included 335,209 participants in total. Nine of the studies were judged to be at low overall risk of bias. Synthesis of the results of these studies revealed that patients within class I obesity did not have a higher odds of mortality due to COVID (pooled OR 1.04, 95% CI 0.94 to 1.16, I² = 76.1%, Figures S1-S3) compared to patients without obesity. We incorporated the baseline risk of death from the included studies to transform this relative effect into an absolute risk difference (baseline risk: 180/1000). We concluded that when comparing 1000 obesity class I COVID‐19 patients to the ones without obesity, six more patients die amongst every 1000 obesity class 1 patients with COVID‐19 compared to COVID‐19 patients without obesity (95% CI from 9 fewer to 23 more, GRADE certainty in the evidence: High, Table 1). 

Comparing this finding to the minimally contextualised threshold of 50 patients in every 1000, we believe obesity class I makes little to no difference to the risk of mortality from COVID‐19. We also conducted three subgroup analyses. First, we compared effect estimates from studies that adjusted for at least age, sex, DM (diabetes mellitis), HTN (hypertension), and cardiovascular diseases with those from studies that did not (Figure S1). Next, we compared effect estimates between studies that used different BMI ranges as their reference group (Figure S2) and between studies with high and low overall risk of bias (Figure S3). Our evaluation of these subgroup effects, as guided by the ICEMAN tool, did not suggest any credible difference.

4.2.1.2. Obesity Class II

Eleven studies with 317,925 participants in total contributed to the pooled effect size for the association between COVID‐19 death and obesity class II. Six of the studies were at a low overall risk of bias. Ten studies used the normal BMI range (18.5 to 24.9 kg/m²) as the reference regression analysis group, one study used the non‐obese range (BMI < 30 kg/m²), and the last study did not specify its reference range. Our meta‐analysis indicated that obesity class II based on BMI can introduce 1.16 times higher odds of mortality compared to having a BMI between 18.5 and 24.9 kg/m² (pooled OR 1.16, 95% CI 0.99 to 1.36, I² = 83.3%, Figures S4-S6), but since the 95% CI includes 1, we cannot infer that the patients with obesity class II have a higher risk of mortality compared to the normal BMI patients. After considering the baseline risk of 180/1000 for COVID‐19 mortality in the included studies, this result translates into 23 more patients (95% CI from 1 fewer to 50 more, GRADE certainty in the evidence: High, Table 2) dying amongst every 1000 obesity class II patients with COVID‐19 compared to COVID‐19 patients without obesity. 

Therefore, we believe that obesity class II makes little to no difference to the risk of mortality from COVID‐19 when considering a minimally important difference threshold of 50 more deaths in every 1000 patients. Similar to the previous class, our three subgroup analyses did not find any credible subgroup effects (Figures S4-S6).

 4.2.1.3. Obesity Class III

We pooled the effect sizes from 19 studies with a total of 354,967 participants. Fourteen studies were at a low overall risk of bias. Similarly, 15 studies adjusted their effect estimate for at least age, sex, DM, HTN, and cardiovascular diseases. Most of the studies used the normal or non‐obese BMI range as the reference group (n = 15). However, four studies did not specify what reference group they reported their effect estimate against and one study used any BMI less than 40 kg/m² as the comparator (Bennett 2021). The results of the meta‐analysis indicate that COVID‐19 patients with class III obesity, as compared to patients with BMI < 30 kg/m² (without obesity), have approximately 1.6 times higher odds of mortality (pooled OR 1.67, 95%CI 1.39 to 2.00, I² = 85.6%, Figures S7-S9). Applying this to the baseline risk of mortality (180/1000 from the included studies) translates to 88 more deaths (95% CI from 54 more to 125 more, GRADE certainty in the evidence: Low, Table 3) amongst every 1000 COVID‐19 patients with class III obesity compared to patients with BMI < 30 kg/m².

Although the confidence interval of the pooled effect size surpasses the minimally contextualised threshold of 50 in 1000, we decided to rate down the certainty in the evidence because of inconsistency and the risk of publication bias (Table 3). The judgement about possible publication bias was based on the asymmetry in the funnel plot (Funnel Plots). Moreover, heterogeneity and considerable visual inconsistency among reported effect estimates with various possible interpretations informed the decision about penalising the certainty of evidence in the inconsistency domain. Therefore, we believe that obesity class III may increase the risk of mortality from COVID‐19.

The evidence suggests that obesity class III may increase the risk of mortality among COVID‐19 patients compared to those with BMI < 30 kg/m². We conducted three subgroup analyses on adjustment covariates, regression reference group, and risk of bias (Figures S7-S9). Our evaluation of these analyses did not suggest any credible subgroup effect. Furthermore, as the regression analysis reference group in  Bennett 2021 included other obesity classes, we carried out a sensitivity analysis to investigate the effect of excluding this study from the meta‐analysis. The result of this analysis did not indicate any change in the interpretation of the results (pooled OR 1.61, 95% CI 1.34 to 1.92, 18 studies, 335,178 participants, I² = 84.4%).

4.2.1.4. Obesity (Unclassified)

We also pooled the effect estimates from the studies that reported on the association between mortality among COVID‐19 patients and obesity without further classifying it (BMI ≥ 30 kg/m²). This meta‐analysis included 60 effect estimates from 54 studies with a total of 1,307,516 participants. Thirty studies were at low overall risk of bias. Twenty‐two studies used BMI < 30 kg/m², nine studies BMI between 18.5 and 24.9 kg/m², and 23 studies unspecified range as the reference group. The results suggest that  patients with obesity with COVID‐19, compared to the patients without (BMI < 30 kg/m²), demonstrate a higher odds of mortality (pooled OR 1.35, 95% CI 1.28 to 1.42, I² = 70.7%, Figures S10-S12). Similar to the previous categories, we translated this relative measure to an absolute risk difference by incorporating the baseline risk of mortality in COVID‐19 patients (180/1000) from the included studies. We concluded that, in every 1000 unclassified COVID‐19 patients with obesity, compared to 1000 COVID‐19 patients without, on average 49 more patients die (95% CI from 39 more to 58 more, GRADE certainty in the evidence: Low, Table 4).

We decided to rate down the certainty in this evidence for inconsistency and imprecision. The pooled confidence cut through the prespecified minimally important threshold. In addition, the effect estimates were visually inconsistent across the threshold. This, in our belief, cannot justify the confidence interval crossing over the threshold. Therefore, we believe that unclassified obesity may make little to no difference to the risk of mortality. Notice that our point estimate of the effect size lies exactly below the 50 in 1000 threshold for an important difference. The results of subgroup analyses could not convince us of any credible differences among subgroups (Figures S10-S12).

4.2.1.5. Units Change in BMI

Ten studies reported effect estimates considering BMI as a continuous independent variable. These studies had 6,937,151 participants. However, 6,910,695 participants were from one study. Six of the studies were at low risk of bias and six adjusted the effect estimate for at least age, sex, DM, HTN, and cardiovascular diseases. We calculated the effect estimates from each study corresponding to every 5 units increase in BMI using appropriate logarithmic transformations. The analysis revealed that increasing BMI corresponds to higher odds for mortality among COVID‐19 patients (for every 5 units increase in BMI: pooled OR 1.16, 95% CI 1.08 to 1.24, I² = 94.4%, Figures S13-S14).

We calculated that considering the baseline mortality risk of 180/1000, on average 23 more (95% CI from 12 more to 34 more, GRADE certainty in the evidence: High, Table 5) out of every 1000 COVID patients die compared to 1000 with 5 units lower BMI. Since this interpretation of results is dealing with a continuous exposure variable, we avoided the use of the previous threshold (defined for binary interpretations of exposure) and considered a difference of zero as the threshold. We, therefore, concluded that every 5‐unit increase in BMI increases the risk of mortality among patients with COVID‐19 infection.

4.2.2. Mechanical Ventilation

We were able to extract 81 effect estimates from 46 unique studies that associated some measure of obesity with the incidence of mechanical ventilation after contracting the COVID‐19 virus. Similar to mortality, most studies (n = 22) used unclassified obesity (BMI ≥ 30 kg/m²). This was followed by obesity class III (n = 12), class I (n = 10), class II and III aggregate (n = 9), and class II (n = 6).

4.2.2.1. Obesity Class I

We included 10 studies in the meta‐analysis on the association between obesity class I and the need for mechanical ventilation with a total of 187,895 participants. The risk of bias assessment indicated that six of these studies have a low overall risk of bias. All studies compared this obesity class to normal BMI (BMI 18.5 to 24.9 kg/m²) except one that did not indicate its analysis reference group. In terms of adjustment, eight studies adjusted their effect estimate for at least age, sex, DM, HTN, and cardiovascular diseases. The meta‐analysis revealed that COVID‐19 patients with class I obesity are at higher odds of needing mechanical ventilation compared to those within the normal BMI range of 18.5 to 24.9 kg/m² (pooled OR 1.38, 95% CI 1.20 to 1.59, I² = 62.1%, Figures S15-S17).

The calculations estimated that the baseline risk of the need for mechanical ventilation is 198 in every 1000 COVID‐19 infections. We applied the pooled OR to this risk and found that in every 1000 obesity class I patients with COVID‐19, 56 more people (95% CI from 31 more to 84 more, GRADE certainty in the evidence: Moderate, Table 1) need mechanical ventilation compared to BMI between 18.5 to 24.9 kg/m². We rated down the certainty in this evidence since the pooled confidence interval crossed the prespecified minimal important threshold of 50 in 1000. Therefore, we concluded that obesity class I, compared to the normal BMI range, probably increases the risk of mechanical ventilation. Our subgroup analyses did not point to any credible difference between the subgroups (Figures S15-S17).

4.2.2.2. Obesity Class II

A total of 171,149 participants from six studies contributed to estimating the effects of class II obesity on mechanical ventilation. Half of the studies were at low overall risk of bias. Similar to the previous class, only one study did not specify its reference group while others used the normal BMI range of 18.5 to 24.9 kg/m². The synthesis of the results of the included studies revealed that obesity class II COVID‐19 patients have higher odds of needing mechanical ventilation (pooled OR 1.67, 95% CI 1.42 to 1.96, I² = 61.7%, Figures S18-S20). 

This relative effect means that considering the baseline risk (198/1000), 94 more (95% CI from 62 more to 128 more, GRADE certainty in the evidence: High, Table 2) patients among every 1000 with COVID‐19 need mechanical ventilation when categorised as having class II obesity compared to the normal BMI range. Therefore, obesity class II increases the risk of requiring mechanical ventilation among COVID‐19 patients. The subgroup analyses did not demonstrate any considerable differences across the groups (Figures S18-S20).

4.2.2.3. Obesity Class III

Twelve studies with a total of 174,520 participants reported on the prognostic effects of class II obesity on mechanical ventilation. Five studies were at a low overall risk of bias. Reference groups in these studies included patients without obesity, normal BMI range, BMI less than 40, and unspecified. The analysis suggested that obesity class III, compared to the patients without obesity (BMI < 30 kg/m²), increases the odds of requiring mechanical ventilation among COVID‐19 (pooled OR 2.17, 95% CI 1.59 to 2.97, I² = 95.0%, Figures S21-S23).

This means that with the baseline risk of 198/1000, in every 1000 COVID‐19 patients 151 more with obesity class III (95% CI from 84 more to 225 more, GRADE certainty in the evidence: High, Table 3) require mechanical ventilation compared to those without obesity. Therefore, we believe that obesity class III increases the risk of mechanical ventilation among COVID‐19 patients. We also conducted a sensitivity analysis to evaluate the effects of removing studies with a BMI less than 40 as the reference for our interpretation. This analysis showed that removing these studies does not alter our findings (pooled OR 2.12, 95% CI 1.50 to 2.98, 10 studies, 173,679 participants, I² = 95.9%). Further, the subgroup analyses did not yield any credible subgroup effects (Figures S21-S23).

4.2.2.4. Obesity (Unclassified)

We also conducted a meta‐analysis of effect estimates from studies that reported the association between mechanical ventilation and obesity without further classifying it (BMI ≥ 30 kg/m²). Pooling information from 21 studies with a total of 62,348 participants showed that COVID‐19 patients with obesity have higher odds of needing mechanical ventilation compared to those without obesity (BMI < 30 kg/m²) (pooled OR 1.69, 95% CI 1.44 to 1.99, I² = 93.9%, Figures S24-S26).

Applying this relative measure to the baseline risk indicates that, in every 1000 patients with obesity, relative to those  patients without, 96 more (95% CI from 64 more to 131 more, GRADE certainty in the evidence: Low) patients would need mechanical ventilation. We rated down the certainty in the evidence due to inconsistency and publication bias. Therefore, we can conclude that obesity may increase the risk of mechanical ventilation among COVID‐19 patients. The subgroup analyses did not show any credible subgroup effects (Figures S24-S26).

 4.2.2.5. Units Change in BMI

Two studies with 13,527 participants investigated BMI as a continuous prognostic factor for mechanical ventilation. One of these was at low overall risk of bias. COVID‐19 infected patients with higher BMIs, on average, had higher odds of requiring mechanical ventilation (for every 5 units increase in BMI: pooled OR 1.28, 95% CI 1.24 to 1.31, I² = 0.0%, Figure S27). This translates into 42 more per 1000 (95% CI from 36 more to 46 more, GRADE certainty in the evidence: Moderate, Table 5). We rated down the certainty in the evidence due to imprecision as the number of studies and participants in the analysis was limited. Therefore, we believe that increasing BMI probably increases the risk of requiring mechanical ventilation among COVID‐19 patients. The subgroup analyses did not show any subgroup effect.

4.2.3. ICU admission

We extracted 63 effect estimates on the association between obesity and ICU admission. These were reported by 45 unique studies. Unclassified obesity (BMI ≥ 30 kg/m²), class I obesity, class II obesity, class III obesity, and continuous BMI accounted for 21, 7, 4, 7, and 5 effect estimates, respectively. 

4.2.3.1. Obesity Class I

For obesity class I, seven studies were included. In the analysis of the effects of obesity class I on ICU admission, seven studies with a total of 162,741 participants provided information. Just more than half of the studies were at low overall risk of bias (n = 4), and studies at a high risk of bias accounted for more than 40% of the total pooling weight and 90% of participants. Both patients without obesity (BMI < 30 kg/m², n = 2) and normal BMI ranges (18.5 ≤ BMI < 25 kg/m², n = 5) served as the reference group in the included studies. On average, COVID‐19 patients with class I obesity had higher odds of being admitted to ICU compared to those patients without obesity (pooled OR 1.36, 95% CI 1.06 to 1.75, I² = 82.4%, Figures S28-S30). 

We calculated the baseline risk of ICU admission in the included studies (208/1000). Combining this with the pooled OR revealed that in every 1000 COVID‐19 patients with class I obesity compared to patients without obesity, 55 more (95% CI from 10 more to 107 more, GRADE certainty in the evidence: Moderate, Table 1) patients were admitted to ICU. We rated down (by one level) the certainty in the evidence for imprecision and risk of bias combined as the confidence interval crossed the minimally contextualised threshold of 50 in 1000 and the high risk of bias studies accounted for nearly half of the total weight of the pooled analysis. Therefore, obesity class I probably increases the risk of ICU admission. The subgroup analyses for this class of obesity did not demonstrate a credible subgroup effect (Figures S28-S30). It is noteworthy that, however, the pooled effect size from studies with a reference group with normal range of BMI was smaller than that of studies with a reference group with a non‐obese range of BMI .

4.2.3.2. Obesity Class II

For obesity class II, four studies were included. We pooled the effect estimates from four studies with 157,665 patients that reported on the effects of obesity class II on ICU admission. Only one of the studies, contributing 30% of the total weight of the pooled analysis, was at low risk of bias. All of the studies used the normal BMI range (18.5 ≤ BMI < 25 kg/m²) as their statistical analysis reference group. Patients with class II obesity, compared to the normal range of BMI, had almost similar odds of ICU admission (pooled OR 1.02, 95% CI 0.90 to 1.15, I² = 35.9%, Figures S31-S33).

After applying the baseline risk (208/1000), we found that three more (95% CI from 17 fewer to 24 more, GRADE certainty in the evidence: Low, Table 2) out of 1000 COVID‐19 patients with obesity class II, compared to normal BMI range, were admitted to ICU out of every 1000 patients. Rating down the certainty in the evidence was due to the very serious risk of bias in the included studies. Therefore, we concluded from this result that obesity class II may have little to no difference in the risk of ICU admission.

4.2.3.3. Obesity Class III

Seven studies reported the difference between patients with a normal BMI and patients with class III obesity. For the effects of this class of obesity on ICU admission, we found a subgroup difference when comparing studies that adjusted the effect estimate for at least age, sex, DM, HTN, and cardiovascular diseases (Figure S34). As mentioned in the Methods section, we considered the guidance by the ICEMAN tool as a general framework for the evaluation of the subgroup‐effect credibility. Because the subgroup effect was credible, we conducted a meta‐analysis that only included studies that adjusted for the minimum set of DM, HTN, cardiovascular disease, age, and sex. For this reason, we present the results of both the overall meta‐analysis and the sensitivity analysis including only the studies adjusting for a minimum set of covariates in the following text and Table 3. 

For the overall group of studies, seven studies contributed to the meta‐analysis with a total of 159,691 participants. The four studies that were at a high risk of bias in this analysis contributed about 56% of the weight of the total pooled analysis. While two studies considered a BMI less than 40 kg/m² as the reference, four considered the normal BMI range (18.5 to 24.9 kg/m²), and one did not specify their reference group. The analysis suggests that  patients with morbid obesity have a higher odds of admission to the ICU (pooled OR 2.26, 95% CI 1.41 to 3.63, 7 studies, 159,691 participants, I² = 94.3%, Figures S35-S36) compared to the normal BMI range.

This finding can be expressed as an absolute risk difference after incorporating the baseline risk of 208/1000. When comparing 1000 COVID‐19 patients with class III obesity to 1000 who are not suffering from morbid obesity, 164 more patients are admitted to ICU (95% CI from 62 more to 280 more, GRADE certainty in the evidence: Low, Table 3). As mentioned earlier, these results should be cautiously interpreted alongside the results from the minimally adjusted subgroup.

In the sensitivity analysis including only the studies adjusting for a minimum set of covariates, we included three studies with 155,405 participants. Only one study with a weight of about 40% of the total pooled effect was at low risk of bias. Also, one study used BMI less than 40 as the reference group (Suleyman 2020), while others used the normal BMI range (18.5 ≤ BMI < 25 kg/m²). We found that, on average, COVID‐19 patients with obesity class III had an OR of 1.20 (95% CI 0.97 to 1.49, I² = 74.7%, Figures S37-S38) of requiring ICU admission compared to normal range BMI.

We translated this finding to an absolute risk difference by incorporating the baseline risk of ICU admission from included studies (208/1000). In every 1000 COVID‐19 patients with obesity class III, 32 more (95% CI from 5 fewer to 73 more, GRADE certainty in the evidence: Moderate, Table 3) patients would get admitted to ICU compared to those having a BMI range between 18.5 and 24.9 kg/m². Therefore, obesity class III probably makes little to no difference to the risk of ICU admission when adjusted for DM, HTN, cardiovascular diseases, age, and sex. A sensitivity analysis to investigate the effects of removing the study using a BMI less than 40 as the reference group demonstrated no change in the interpretation of the results or the certainty in the evidence (pooled OR 1.06, 95% CI 1.01 to 1.11, 2 studies, 155,050 participants, I² = 0.0%).

4.2.3.4. Obesity (Unclassified)

Twenty one‐studies investigated the association between obesity and ICU admission without specification of obesity class (BMI > 30 kg/m²). These studies included 69,147 participants and 12 of them were at a low overall risk of bias. The reference groups used in these 21 studies included normal BMI range (18.5 ≤ BMI < 25 kg/m², n = 3), non‐obese BMI range (BMI < 30 kg/m², n = 8), and unspecified (n = 10). On average,  COVID‐19 patients with obesity had an almost twice larger odds of ICU admission (pooled OR 1.86, 95% CI 1.56 to 2.21, I² = 88.4%, Figures S39-S41) compared to those patients without obesity.

By applying the baseline risk of 208/1000, this can be expressed as 123 more (95% CI from 84 more to 164 more, GRADE certainty in the evidence: Moderate, Table 4) people would get admitted to ICU amongst every 1000 patients with obesity compared to those without (BMI < 30 kg/m²). Therefore, after comparing this absolute risk difference to the threshold of 50 in every 1000, we conclude that obesity (without further classification) probably increases the risk of ICU admission.

 4.2.3.5. Units Change in BMI

No studies were found that looked at the impact of continuous obesity on ICU admission. 

4.2.4. Hospitalisation

4.2.4.1. Obesity Class I

Analyses provided moderate‐certainty evidence that patients with class I obesity probably have a similar risk of hospitalisation compared to patients without obesity (BMI < 30 kg/m²) with no evidence of credible subgroup effects (pooled OR 0.96, 95% CI 0.82 to 1.14, 5 studies, 515,115 participants, I² = 96.8%, Table 1; Figures S42-S44).

4.2.4.2. Obesity Class II

Analyses provided moderate‐certainty evidence that patients with class I and II obesity probably have a similar risk of hospitalisation compared to patients without obesity (BMI < 30 kg/m²) with no evidence of credible subgroup effects (pooled class II: OR =1.04, 95% CI 0.90 to 1.2, 3 studies, 293,707 participants, I² = 94.9%, Table 2; Figures S45-S46).

4.2.4.3. Obesity Class III

On the contrary to classes I and II obesity, patients with class III obesity (low certainty of evidence, Table 3) may have an increased risk of hospitalisation, when considering studies that adjusted for at least age, sex, DM, HTN, and cardiovascular diseases (pooled class III: OR = 1.35, 95% CI 1.19 to 1.52, 3 studies, 293,004 participants, I² = 91.4%, Table 3; Figures S47-S51). The analyses that considered all the studies suggested that obesity class III may be an important factor increasing the risk of hospitalisation (pooled class III: OR = 1.77, 95% CI 1.40 to 2.26, 7 studies, 747,176 participants, I² = 95.4%, low certainty‐evidence, Table 3; Figures S47-S51).

4.2.4.4. Obesity (Unclassified)

Patients with unclassified obesity (moderate certainty of evidence, Table 4) probably have an increased risk of hospitalisation, when considering studies that adjusted for at least age, sex, DM, HTN, and cardiovascular diseases (pooled obesity (unclassified): OR 1.31, 95% CI 1.20 to 1.44, 14 studies, 510, 405 participants, I² = 91.4%, Figures S52-S56). The analyses that considered all the studies suggested that unclassified obesity may be an important factor increasing the risk of hospitalisation (pooled obesity (unclassified): OR 1.49, 95% CI 1.34 to 1.64, 20 studies, 515,517 participants, I² = 82.1%, low certainty‐evidence, Table 4; Figures S52-S56). 

 4.2.4.5. Units Change in BMI

Also, low‐certainty evidence suggests that every 5 units increase in BMI may also increase this risk (pooled OR, 1.17, 95% CI 1.05 to 1.31, 3 studies, 6,911,600 participants, I² = 62.9%, Table 5; Figures S57-S58).

4.2.5. Severe COVID disease

4.2.5.1. Obesity Class I

Analyses provided low‐certainty evidence that patients with class I obesity may have an increased risk of severe COVID compared to patients without (BMI < 30 kg/m²), with no evidence of credible subgroup effects (pooled class I: OR = 1.48, 95% CI 1.16 to 1.87, 3 studies, 1040 participants, I² = 0.0%, Table 1; Figures S59-S61).

4.2.5.2. Obesity Class II

No studies were found that looked at the impact of obesity class II on severe COVID‐19 disease. 

4.2.5.3.  Obesity Class III

No studies were found that looked at the impact of obesity class III on severe COVID‐19 disease. 

4.2.5.4. Obesity (Unclassified)

As to the association between unclassified obesity and this outcome, high‐certainty evidence suggests that COVID‐19 patients with obesity, compared to patients without (BMI < 30 kg/m²), are at an increased risk of developing severe disease (pooled OR =1.94, 95% CI 1.62 to 2.32, 19 studies, 878,804 participants, I² = 69.5%, Table 4; Figures S62-S64).

 4.2.5.5. Units Change in BMI

Analyses provided low‐certainty evidence that patients with higher BMI (continuous) may have an increased risk of severe COVID‐19 compared to patients without (BMI < 30 kg/m²), with no evidence of credible subgroup effects (pooled BMI per 5 kg/m² increase: OR 1.94, 95% CI 1.39 to 2.7, 5 studies, 1041 participants, I² = 35.1%, Table 5; Figures S65-S66).

4.2.6. Pneumonia

4.2.6.1. Obesity Class I

No studies were found that looked at the impact of obesity class I on pneumonia due to COVID‐19 disease. 

4.2.6.2. Obesity Class II

No studies were found that looked at the impact of obesity class II on pneumonia due to COVID‐19 disease. 

4.2.6.4. Obesity (Unclassified)

Analysis of moderate‐certainty evidence suggests that patients with obesity (unclassified), compared to those with non‐obesity (BMI < 30 kg/m²) probably have an increased risk of pneumonia due to COVID‐19 (pooled OR = 1.44, 95% CI 1.21 to 1.72, 5 studies, 35,924 participants, I² = 81.4%, Table 4; Figure S67). 

 4.2.6.5. Units Change in BMI

No studies were found that looked at the impact of the increase in BMI on pneumonia due to COVID‐19 disease. 

4.3. Other Outcomes

We also identified other studies that reported on length of hospitalisation, length of ICU admission, and length of mechanical ventilation. All but one of these studies reported adjusted hazard ratios by considering hospital discharge, ICU discharge, or extubation as time‐to‐event outcomes. Biscarini 2020, in contrast, considered the length of hospitalisation a continuous outcome and reported a beta coefficient (slope) for the adjusted effects of obesity.

Hur 2020 included 564 patients hospitalised with COVID‐19 and analysed the data for a final cohort of 486 patients of whom a total of 138 patients were intubated during the course of the study. The study considered the time from intubation to extubation with censoring patients at death or those intubated at the time of the final follow‐up. This adjusted analysis for age, sex, race, HTN, and smoking indicated an HR of 0.53 (95% CI 0.32 to 0.90) for obesity class I and II combined (BMI 30 to 40 kg/m²) and an HR of 0.40 (95% CI 0.19 to 0.83) for obesity class III. These were both compared to the reference group of the patients without obesity. These findings suggest that COVID‐19 patients without obesity have a higher instantaneous chance of extubation at any time during their period of intubation.

Another study by Parikh 2020 reported on time from ICU admission to ICU discharge among 160 COVID‐19 patients who were admitted to ICU.  After adjusting for age, sex, and a history of asthma, the study reported an adjusted HR of 0.9 (95% CI 0.5 to 1.7) for ICU discharge and an adjusted HR of 1.2 (95% CI  0.7 to 2.2) for hospital discharge when comparing  patients with obesity to those without. While the point estimates may indicate that obesity lowers the instantaneous chance of ICU discharge but increases the same chance for hospital discharge after ICU admission, the wide confidence intervals include the no‐effect line.

Regarding the time from hospital admission to discharge, Hu 2020 and Van Zelst 2020 conducted Cox proportional hazard regression analysis, compared to the linear regression by Biscarini 2020. This latter study reported that the analysis of 427 patients suggests that patients with obesity, compared to those without, have an average increase of 1.19 days (95% CI  ‐1.88 to 4.26) in the length of hospital stay when adjusted for some covariates like age, sex, and some comorbidities. The study by Huh 2020 included 58 admitted patients with mild COVID‐19 from whom those with obesity had an adjusted HR of 0.83 (P value for trend = 0.001) for hospital discharge. The other study with a similar analysis, by Van Zelst 2020, analysed the data for 79 COVID‐19‐positive patients, of whom 74 were admitted to the hospital. After adjustment for age, sex, and presence of metabolic syndrome, this study reported a hospital discharge adjusted HR of 0.97 (95% CI 0.92 to 1.01) when comparing patients with obesity to those without.

Discussion

5.1. Summary of main results 

Our review investigated the prognostic effects of obesity on adverse COVID‐19 outcomes. We located and evaluated available evidence about patient‐important outcomes such as mortality, mechanical ventilation, and ICU admission. We also investigated the effects of obesity on hospitalisations, severe disease, and pneumonia due to COVID‐19.

High‐certainty evidence suggested that obesity class I and II does not meet a minimal threshold of 5% absolute risk increase for mortality. On the other hand, our analysis for class III obesity showed that the effect of this class exceeds this threshold, although this is low‐certainty evidence. A closer look at these observed effect sizes, interestingly, reveals that a dose‐response relationship exists. The absolute mortality risk difference (relative to patients without obesity) for COVID‐19 patients with obesity indicates a rise from 0.06% more to 8.8%, from class I to class III, respectively. This suggests that obesity possibly plays a role in increasing the risk of mortality, even though the milder two classes impose less than a 5% risk increase, which we considered the threshold of an important prognostic factor. This finding is also in line with the high‐certainty of evidence that exhibited an increased mortality risk for every five units of BMI increase. Our best estimate is that every five units of BMI increase inflates mortality risk by more than 2%. 

Coherently for all obesity classes, the evidence points toward an increased risk of mechanical ventilation among patients with obesity. The certainty of the evidence for classes II and III is high while for class I is moderate. Similar to mortality, we observed a dose‐response relationship which further supports the conclusion that obesity is associated with an increased risk for this outcome. The amount of this risk increase ranges from more than 5% to 15% for different obesity classes. Findings regarding every 5 units increase in BMI and unclassified obesity was also aligned with this elevated risk, however, with moderate and low certainty of evidence. We interpret these findings to indicate that all obesity classes are important prognostic factors for the need for mechanical ventilation among COVID‐19 patients.

In contrast to mortality and mechanical ventilation, the observed effects on ICU admission were not congruent across obesity classes. We observed the highest risk of ICU admission with an absolute risk increase of 5.5% from moderate‐quality evidence amongst participants with class I obesity. This suggests that this obesity class is probably an important prognostic factor. However, for obesity class II and class III, the absolute increase in ICU admission risk was only 0.03% and 2.3% from low‐quality and moderate‐quality evidence, respectively. Further adding to this incongruency, the low‐quality evidence from studies investigating unclassified obesity suggests a 12.3% increase in ICU admission risk. Therefore, we believe that interpretation of our findings for ICU admission should be accompanied by utmost caution. Even though unclassified obesity seems like an important prognostic factor for ICU admission, findings regarding obesity classes do not support this conclusion. One possible explanation for these contradictory results could be that the decision about ICU admission is made by the treating physician. This subjective decision can be informed by many different factors that are known to physicians, such as obesity. The resulting differential misclassification bias can introduce this unexpected difference. To a lesser degree, this bias can influence our results about mechanical ventilation too.

In terms of hospitalisation, we found moderate‐quality evidence that classes I and II of obesity alter the risk of hospitalisation negligibly (‐0.05% and 0.07%, respectively). On the other hand, the 3.2% and 5.5% increased risk of hospitalisation for class III and unclassified obesity arise from low certainty in the evidence. Therefore, although obesity might be a prognostic factor for hospitalisation, we do not think it is an important prognostic factor.

The available data for severe COVID‐19 disease only showed class I and unclassified obesity associations with this outcome. High‐quality evidence suggests that unclassified obesity is an important prognostic factor for severe disease with an increased risk of 12.3%. Similarly, we only found moderate‐quality evidence that unclassified obesity is an important prognostic factor for COVID‐19 pneumonia with an 8.2% increase in absolute risk.

5.2. Overall completeness and applicability of evidence

In this review, we sought to determine if obesity is an independent prognostic factor for adverse outcomes among adult patients with confirmed COVID‐19 infections. Obesity as a prognostic variable is complex and has been examined in different ways, which can complicate interpretation. Part of this complexity also includes using different comparators in analyses, such as those that include or exclude being overweight (BMI 25 to 29.9 kg/m²). Examining obesity as three separate classes, as continuous BMI, and also assessing the congruency across the various approaches, as we did, may address part of this complexity and provide some of the nuanced details required for clinical and policy decision‐making. 

Our findings demonstrate clear dose‐response relationships between obesity and in‐hospital mortality, and obesity and mechanical ventilation amongst adult COVID‐19 patients. Unclassified obesity is an important prognostic factor for severe COVID‐19 disease and is probably an important prognostic factor for COVID‐19 pneumonia. However, findings for ICU admission and hospitalisation had more variability and uncertainty. These findings could be used for risk stratification of COVID‐19 patients, and to inform customised clinical management and resource allocation when providing care. 

On the flip side of management, taking precautions to minimise the risk for infection with COVID‐19 is important for everyone, but even more so for certain groups of people, particularly regarding modifiable risk factors. Specifically, our findings indicate that adults living with obesity can be considered a potentially vulnerable group. These findings are, therefore, also relevant for COVID‐19 practice and policy decisions around prevention, and may also be informative for decisions around detection and testing. Importantly, the use of these findings for management and prevention decisions should be considered within the context of the roles of other prominent prognostic factors (e.g. age).

Based on the exclusion criteria of the included studies, the findings of this review are specifically most applicable to adults living with obesity who have been hospitalised with confirmed COVID‐19 infection. Included studies were undertaken in a wide range of countries and amongst people from a wide age range (approximately between 31 and 77 years old on average). Although present in a minority of studies, the following are examples of sub‐populations excluded from these studies: pregnant women, nursing home residents, and patients living with severe morbid obesity (e.g. > 300 kg, BMI > 100 kg/m²), organ transplants, active cancer, and terminal illnesses. Importantly, a large proportion of included studies did not report their exclusion criteria.

Numerous interdependent prognostic variables for adverse outcomes in COVID‐19 patients have been identified, including obesity. A better understanding of whether obesity per se is a key prognostic variable will help to identify vulnerable patients, inform treatment decisions and better allocate limited resources. For these reasons, we only included studies with multivariable analyses and then sought to examine if associations were different between studies using a minimum adjustment set (age, sex, diabetes, hypertension, and cardiovascular disease) compared to those that did not.  

5.3. Certainty of the evidence

As evident from the summary of findings tables, we have a variety of evidence quality levels for different comparisons and outcomes. We have started our certainty in the evidence as high for each comparison and subsequently rated the level down for pertinent issues. The most common reasons for downgrading the certainty in the evidence include imprecision and heterogeneity followed by risk of bias.  While we did not have problems with the directness of the evidence, suspected publication bias affected some of the ratings. You can refer to the individual summary of findings tables for each comparison and outcome reasons and explanations for the downgrading of certainty in the evidence.

5.4. Potential biases in the review process

Our review has several notable strengths. We systematically searched all bibliographic databases for studies published in any language. Therefore, our inferences on the prognostic value of obesity, in the setting of COVID‐19, are informed by the totality of peer‐reviewed publications on this topic. In addition to summarising the individual studies, we critically appraised their risk of bias with the use of QUIPS. Furthermore, for each of WHO’s obesity classes, we were able to conduct a separate meta‐analysis to obtain a pooled estimate and 95% CI. Beyond the WHO classification, our review also evaluated the prognostic utility of BMI as a continuous variable. Specifically, in evaluating the risk of bias, we determined a minimal set of key covariates (age, diabetes, cardiovascular disease) that authors of primary studies should have adjusted for in their multivariable regression modelling. We used subgroup analyses to evaluate whether this minimal set of key covariates had bearing on the association between obesity and outcomes. The combination of only including studies with multivariate analyses, the use of QUIPS to assess risk of bias (which explicitly incorporates the extent of covariates adjusted for) and the conducting of subgroup analyses based on the adequate adjustment of covariates allowed us to address the complex relationship between obesity and comorbidities in influencing mortality and other adverse COVID‐related outcomes. In summarising our findings, we translated the pooled relative effects to absolute risk difference to clinically contextualise the findings of our review. Our review also benefited from the GRADE framework, which informed our certainty in the prognostic value of obesity. 

Our review also had several notable limitations. We continuously strived to maintain an updated list of eligible articles. We did this by repeating our search and screening multiple times in the conduct of this systematic review. Despite our best efforts, however, we were unable to keep up with the rapid pace of publications on COVID‐19. Therefore, by the time this review is completed, several new articles may be published that are eligible for inclusion but not within our review. To date, our search has captured studies published up to April 2021. These additional studies may have variable impact on the outcomes reported in our review. We believe that additional studies may have minimal impact on the pooled estimates for mortality and mechanical ventilation. These outcomes had the highest number of patients and events, and thus the inclusion of additional studies may minimally impact on our findings. This, however, is not the case for other outcomes (such as ICU admission, hospitalisation, and severe COVID‐19), which were informed by much fewer studies. The inclusion of additional studies published after April 2021 may introduce an additional source of heterogeneity: inclusion of studies following cohorts of patients previously vaccinated for COVID‐19. Another point to mention is that, in the case of missing data, we did not contact the authors of the original works. However, this can also have minimal impact on the results based on the previously mentioned reasons. 

In our review, the included primary studies, in their multivariable regression modelling, adjusted for different types of covariates. Seldom did we observe two studies adjusting for the exact set of covariates, and some studies only adjusted for a single covariate. Although we prespecified a set of important covariates that authors should adjust for (as part of our risk of bias assessment, including age, sex, diabetes, hypertension, and cardiovascular disease), considerable heterogeneity exists in the covariates adjusted for. A notable limitation of our review is that we were unable to explore this methodologic heterogeneity beyond the exploration of subgroups. The most ideal method for dealing with this source of heterogeneity is to conduct a meta‐analysis of individual patient data. These differing covariates may be a contributing factor to the observed inconsistency, beyond random error, observed in meta‐analyses for mechanical ventilation, ICU admission, and hospitalisation. To account for and acknowledge this degree of heterogeneity, we ensured that we conducted all our meta‐analyses under the random‐effects framework. Furthermore, we acknowledge the doubt created by heterogeneity in assessing our certainty in the prognostic value of obesity.

Finally, the findings of our review are dependent on the underlying assumption that BMI is a meaningful and valid measure to capture obesity and its severity. The relationship between BMI and percentage of body fat is non‐linear, and the measure does not account for nonfat sources of mass (e.g. bone, muscle), the natural changes in body composition that accompany age, and the natural differences in body composition across different ethnicities (Hall 2006; Rothman 2008). Furthermore, there are differences between self‐reported and measured BMI (Maukonen 2018). However, despite these limitations, BMI remains the predominant measure of an individual's adiposity due to its ease of use and relatively low cost of collection (Burkhauser 2008). BMI was selected as the measure of obesity in our review due to its widespread use, however, we acknowledge the bias that the measure may bring when evaluating obesity's prognostic value.

5.5. Agreements and disagreements with other studies or reviews

Since the beginning of the COVID‐19 pandemic, multiple systematic reviews have strived to investigate the association of obesity with adverse outcomes of COVID‐19 disease. However, the methodologic quality, number of included studies, and reporting of these studies vary. 

Dessie 2021 included a total of 423,117 participants from a total of 42 studies in their review to locate mortality‐related risk factors of COVID‐19. Their quantitative analysis of 11 studies concluded with an OR of 1.34 (95% CI 1.17 to 1.52) and an HR of 1.5 (95% CI 1.26 to 1.75) for patients with obesity versus those without. These results are in line with our calculated pooled OR for obesity unclassified (OR 1.35, 95% CI 1.28 to 1.42). However, there are methodological concerns with this study. For example, the study did not mention any protocol registration or clear eligibility criteria.

Another study, Raeisi 2022, conducted a systematic search in June 2020 (published in July 2021). In contrast to other reviews, this study also evaluated the quality of evidence through the application of GRADE. This study investigated the association of obesity with mortality, mechanical ventilation, ICU admission, and hospitalisation amongst COVID‐19 patients. This study found low‐quality evidence for mortality, mechanical ventilation, and ICU admission and very low‐quality evidence for hospitalisation. Similar to other reviews, authors only compared patients with obesity versus those without further classification of obesity groups. The findings of this study agree with our findings in that they found ORs of 1.23 (95% CI 1.06 to 1.41), 2.24 (95% CI 1.70 to 2.94), 1.75 (95% CI 1.38 to 2.22), and 1.75 (95% CI 1.47 to 2.09) for mortality, mechanical ventilation, ICU admission, and hospitalisation. However, we reported lower ORs for mechanical ventilation and hospitalisation.

Other reviews also had systematic searches from May 2020 to July 2020 (Chu 2020; Földi 2020; Soeroto 2020). These reviews investigated the association between obesity and various outcomes such as mortality, mechanical ventilation, ICU admission, and poor outcomes. Most findings by these reviews were in line with our findings in terms of the association direction. However, Chu 2020 reported a statistically non‐significant lower odds for mortality among those with obesity. This study reported an OR of 0.89 (95% CI 0.32 to 2.51, 3856 participants from 3 studies) for patients with obesity versus those without obesity in COVID‐19 patients. In addition, another study investigated factors associated with mortality amongst COVID‐19 patients admitted to the ICU (Taylor 2021). The study included 21 studies to compare the mean BMI between the deceased and survival groups. What they found was a statistically non‐significant difference between the groups (SMD 0.05, 95%CI ‐0.06 to 0.16). It is noteworthy that an asymmetry in their forest plot suggests the possibility of publication bias.

Authors' conclusions

6.1. Implications for practice

Our findings are most applicable to adults living with obesity who have been hospitalised with confirmed COVID‐19 infection. In these patients, the evidence suggests clear dose‐response relationships between obesity and in‐hospital mortality, and obesity and mechanical ventilation. Unclassified obesity is an important prognostic factor for severe COVID‐19 disease and is probably an important prognostic factor for COVID‐19 pneumonia. However, findings for ICU admission and hospitalisation were less clear. These findings could be used to inform risk stratification of adult COVID‐19 patients for the provision of clinical care, allocation of resources, and planning prevention, detection, and testing strategies. 

6.2. Implications for research

As the findings of our review suggest, obesity is an important prognostic factor for at least some of the patient‐important COVID‐19 adverse outcomes. Therefore, it is important that future research acknowledge this fact and clearly report on their population in regard to obesity. Another important consideration is for researchers to ensure that patients with obesity should be sufficiently represented in trials on COVID‐19 treatments and vaccines. To ensure this, any systematic exclusion of this group should only take place with a strong rationale and in exceptional circumstances. Finally, our review demonstrates the variability in covariates adjusted for across studies exploring obesity as a prognostic factor amongst patients diagnosed with COVID‐19. Future individual patient data meta‐analyses are needed to explore the independent association of obesity and COVID‐19 outcomes under a controlled set of covariates.

What's new

Date	Event	Description
7 June 2023	Amended	Amendment to fix PDF display

History

Review first published: Issue 5, 2023

Appendices

Appendix 1. Search strategies

English Databases:

MEDLINE (PubMed)

#1 "COVID‐19"[Mesh] OR "SARS‐CoV‐2"[Mesh] OR "COVID‐19 Testing"[Mesh] OR "COVID‐19 Vaccines"[Mesh]

#2 

Search: ("2019 nCoV"[tiab] OR 2019nCoV[tiab] OR "2019 novel coronavirus"[tiab] OR ((coronavirus[tiab] OR "corona virus"[tiab]) AND (Huanan[tiab] OR Hubei[tiab] OR Wuhan[tiab])) OR "coronavirus‐19"[tiab] OR "coronavirus disease‐19"[tiab] OR "coronavirus disease‐2019"[tiab] OR "COVID 19"[tiab] OR COVID19[tiab] OR "nCov 2019"[tiab] OR "new coronavirus"[tiab] OR "new coronaviruses"[tiab] OR "novel coronavirus"[tiab] OR "novel coronaviruses"[tiab] OR "novel corona virus"[tiab] OR "SARS‐CoV2"[tiab] OR "SARS CoV‐2"[tiab] OR SARSCoV2[tiab] OR "SARSCoV‐2"[tiab] OR "SARS‐coronavirus‐2"[tiab] OR "SARS‐like coronavirus"[tiab] OR "Severe Acute Respiratory Syndrome Coronavirus‐2"[tiab] OR "COVID‐19"[nm] OR "COVID‐19 drug treatment"[nm] OR "COVID‐19 serotherapy"[nm] OR "LAMP assay"[nm] OR "severe acute respiratory syndrome coronavirus 2"[nm] OR "spike protein, SARS‐CoV‐2"[nm]) NOT ("animals"[mh] NOT "humans"[mh]) NOT (editorial[pt] OR newspaper article[pt])

#3 #1 OR #2

#4 obese OR obesity OR overweight 

#5 bmi or "body mass index" or "body mass" or "body weight" or "metabolic disorder" or "waist circumference"

#6 obesity [MeSH Terms] or "body mass index" [MeSH Terms]

#7 #4 OR #5 OR #6

#8 #3 AND #7 Filters: from 2019/12/1 ‐ 2021/4/23

Embase (Ovid)

#1 coronavirus.mp. or Coronavirinae/ or exp Coronavirinae/

#2 (coronavirus* or coronovirus* or coronavirinae* or Coronavirus* or Coronovirus* or "2019‐nCoV" or 2019nCoV or nCoV2019 or nCoV‐2019 or COVID‐19 or COVID19 or covid 19 or HCoV‐19 or HCoV19 or 2019 novel* or Ncov or n‐cov or SARS‐CoV‐2 or SARSCoV‐2 or SARSCoV2 or SARS‐CoV2 or SARSCov19 or SARS‐Cov19 or SARSCov‐19 or SARS‐Cov‐19 or Ncorona*).mp.

#3 SARS coronavirus/ or severe acute respiratory syndrome/ or severe acute respiratory syndrome*.mp.

#4 ((corona* or corono*) adj2 (virus* or viral* or virinae*)).mp.

#5 #1 or #2 or #3 or #4

#6 obesity/

#7 (obese or obesity or overweight or bmi or body mass index).mp.

#8 #6 or #7

#9 #5 and #8

#10 limit 9 to dd=20191201 ‐20210421

 

Cochrane COVID‐19 Study Register (https://covid‐19.cochrane.org/ )

Filtered by 

obese or obesity or overweight or BMI or “body mass index” or “body mass” or “body weight” or “metabolic disorder” or “waist circumference”

Date searched: 23 April 2021

 

WHO COVID‐19 database  (https://search.bvsalud.org/global-literature-on-novel-coronavirus-2019-ncov/)

(tw:(obese or obesity or overweight or BMI or body mass index or body mass or body weight or metabolic disorder or waist circumference))

Date searched: 23 April 2021

 

Chinese Databases:

China Network Knowledge Infrastructure (CNKI)

((SU = 肥胖+体重+超重+身体质量指数+体重指数+BMI+代谢失调+代谢异常+代谢障碍+代谢紊乱+腰围) or 

(TKA = 肥胖+体重+超重+身体质量指数+体重指数+BMI+代谢失调+代谢异常+代谢障碍+代谢紊乱+腰围)) and 

((SU = 2019冠状病毒+新型冠状病毒+新型冠状病毒肺炎+新冠肺炎+新冠肺炎疫情+冠状病毒+冠状病毒感染+新型冠状病毒感染+新型冠状病毒感染的肺炎+

"COVID‐19"+"COVID19"+"2019‐nCoV"+"2019nCoV"+"SARS‐CoV"+"SARSCoV"+"nCoV2019"+"nCoV‐2019"+"HCoV‐19"+"HCoV19"+

严重急性呼吸综合症+严重急性呼吸综合征+严重急性呼吸道症候群+急性呼吸综合征) or

(TKA = 2019冠状病毒+新型冠状病毒+新型冠状病毒肺炎+新冠肺炎+新冠肺炎疫情+冠状病毒+冠状病毒感染+新型冠状病毒感染+新型冠状病毒感染的肺炎+

"COVID‐19"+"COVID19"+"2019‐nCoV"+"2019nCoV"+"SARS‐CoV"+"SARSCoV"+"nCoV2019"+"nCoV‐2019"+"HCoV‐19"+"HCoV19"+

严重急性呼吸综合症+严重急性呼吸综合征+严重急性呼吸道症候群+急性呼吸综合征))

 

Chinese Scientific Journals Database (VIP)

(U = 肥胖+体重+超重+身体质量指数+体重指数+BMI+代谢失调+代谢异常+代谢障碍+代谢紊乱+腰围) and 

(U = 2019冠状病毒+新型冠状病毒+新型冠状病毒肺炎+新冠肺炎+新冠肺炎疫情+冠状病毒+冠状病毒感染+新型冠状病毒感染+新型冠状病毒感染的肺炎+

"COVID‐19"+"COVID19"+"2019‐nCoV"+"2019nCoV"+"SARS‐CoV"+"SARSCoV"+"nCoV2019"+"nCoV‐2019"+"HCoV‐19"+"HCoV19"+

严重急性呼吸综合症+严重急性呼吸综合征+严重急性呼吸道症候群+急性呼吸综合征)

 

Wanfang data 

((主题: "肥胖" or "体重" or "超重" or "身体质量指数" or "体重指数" or "BMI" or "代谢失调" or "代谢异常" or "代谢障碍" or "代谢紊乱" or "腰围") or 

(题名或关键词: "肥胖" or "体重" or "超重" or "身体质量指数" or "体重指数" or "BMI" or "代谢失调" or "代谢异常" or "代谢障碍" or "代谢紊乱" or "腰围")) and 

((主题: "2019冠状病毒" or "新型冠状病毒" or "新型冠状病毒肺炎" or "新冠肺炎" or "新冠肺炎疫情" or "冠状病毒" or "冠状病毒感染" or "新型冠状病毒感染" or "新型冠状病毒感染的肺炎" or 

"COVID‐19" or "COVID19" or "2019‐nCoV" or "2019nCoV" or "SARS‐CoV" or "SARSCoV" or "nCoV2019" or "nCoV‐2019" or "HCoV‐19" or "HCoV19" or 

"严重急性呼吸综合症" or "严重急性呼吸综合征" or "严重急性呼吸道症候群" or "急性呼吸综合征") or 

(题名或关键词: "2019冠状病毒" or "新型冠状病毒" or "新型冠状病毒肺炎" or "新冠肺炎" or "新冠肺炎疫情" or "冠状病毒" or "冠状病毒感染" or "新型冠状病毒感染" or "新型冠状病毒感染的肺炎" or 

"COVID‐19" or "COVID19" or "2019‐nCoV" or "2019nCoV" or "SARS‐CoV" or "SARSCoV" or "nCoV2019" or "nCoV‐2019" or "HCoV‐19" or "HCoV19" or 

"严重急性呼吸综合症" or "严重急性呼吸综合征" or "严重急性呼吸道症候群" or "急性呼吸综合征"))

 

SinoMed

        1)    "肥胖"[常用字段:智能] OR "体重"[常用字段:智能] OR "超重"[常用字段:智能] OR "身体质量指数"[常用字段:智能] OR "体重指数"[常用字段:智能] OR "BMI"[常用字段:智能] OR "代谢失调"[常用字段:智能] OR "代谢异常"[常用字段:智能] OR "代谢障碍"[常用字段:智能] OR "代谢紊乱"[常用字段:智能] OR "腰围"[常用字段:智能]

        2)    "2019冠状病毒"[常用字段:智能] OR "新型冠状病毒"[常用字段:智能] OR "新型冠状病毒肺炎"[常用字段:智能] OR "新冠肺炎"[常用字段:智能] OR "新冠肺炎疫情"[常用字段:智能] OR "冠状病毒"[常用字段:智能] OR "冠状病毒感染"[常用字段:智能] OR "新型冠状病毒感染"[常用字段:智能] OR "新型冠状病毒感染的肺炎"[常用字段:智能] OR " 'COVID‐19'"[常用字段:智能] OR "'COVID19'"[常用字段:智能] OR "'2019‐nCoV'"[常用字段:智能] OR "'2019nCoV'"[常用字段:智能] OR "'SARS‐CoV'"[常用字段:智能] OR "'SARSCoV'"[常用字段:智能] OR "'nCoV2019'"[常用字段:智能] OR "'nCoV‐2019'"[常用字段:智能] OR "'HCoV‐19'"[常用字段:智能] OR "'HCoV19'"[常用字段:智能] OR "严重急性呼吸综合症"[常用字段:智能] OR "严重急性呼吸综合征"[常用字段:智能] OR "严重急性呼吸道症候群"[常用字段:智能] OR "急性呼吸综合征"[常用字段:智能]

        3)    (#1) AND (#2)

Appendix 2. Blank Extraction Sheet

https://docs.google.com/spreadsheets/d/1y1xf8DpSPImBrPt36ROqHHwnIEMUS0TSDnyJd6cWOqY/edit?usp=sharing

Appendix 3. Funnel Plots

We presented only the funnel plots for the comparisons that included at least 10 effect estimates. Below you can find the links to each funnel plot and their interpretations: 

 

Mortality and obesity class I analysis funnel plot: Figure S68

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

 

Mortality and obesity class II analysis funnel plot: Figure S69

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

 

Mortality and obesity class III analysis funnel plot: Figure S70

Publication bias: serious. Asymmetrical around pooled estimate in the funnel plot.

 

Mortality and  unclassified obesity analysis funnel plot: Figure S71

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

 

Mortality and every 5 units increase in BMI analysis funnel plot: Figure S72

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

 

Mechanical ventilation and obesity class I analysis funnel plot: Figure S73

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

 

Mechanical ventilation and obesity class III analysis funnel plot: Figure S74

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

 

Mechanical ventilation and unclassified obesity analysis funnel plot: Figure S75

Publication bias: serious. Asymmetrical around pooled estimate in the funnel plot.

 

ICU admission and unclassified obesity analysis funnel plot: Figure S76

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

 

Hospitalisation and unclassified obesity analysis funnel plot: Figure S77

Publication bias: serious. Asymmetrical around pooled estimate in the funnel plot.

 

Hospitalisation and unclassified obesity analysis funnel plot (within minimum adjustment set subgroup): Figure S78

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

 

Severe disease and unclassified obesity analysis funnel plot: Figure S79

Publication bias: not serious. Symmetrical around pooled estimate in the funnel plot.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Abumayyaleh 2021.

Study characteristics
Notes	English title Does there exist an obesity paradox in COVID‐19? Insights of the international HOPE‐COVID‐19‐registry Study setting Start of study recruitment (MM/YYYY) NR End of study recruitment (MM/YYYY) 05/2020 Study design retrospective cohort Study centre(s) international Number of centres/clinics/areas 21 Study setting inpatient Number of participants recruited 3635 Sampling method consecutive participants Participants Female participants (absolute number), 1518 Age measure, value median (range), 63 (18, 99) Central tendency measure of age median Value 63 Dispersion measure of age range Value (report as single Value or as X1, X2) 18, 99 Inclusion criteria Hospitalised COVID‐19 patients were included. Exclusion criteria Lack of data about body mass index (BMI) and patients age < 18 Smoking (absolute number), 266 Diabetes (absolute number), 678 Hypertension (absolute number), 1808 Cardiovascular diseases (absolute number), 824 Please indicate if additional information is available  NR Asthma NR Chronic obstructive pulmonary disease NR Other pulmonary diseases (absolute number), 624 Please indicate if additional information is available  NR Immunosuppression (absolute number), 239 Please indicate if additional information is available Immunosuppressive therapy for psoriasis arthritis, lung transplantation, kidney transplantation or systemic lupus erythematosus; oncological disease such as mamma‐ca, prostate‐ca, myelodysplastic syndrome or gammopathy, glucocorticoid therapy caused by COPD, dialysis, HIV or hepatitis Chronic kidney disease (absolute number), 191 Cancer (absolute number), 401 Steroid administration NR Supplemental oxygen NR Differential values for various oxygenation methods (if indicated) 2552/3562 for O2 support at admission, 785/3515 for high‐flow nasal cannula Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Obesity is defined as a BMI ≥ 30 kg/m2 according to the recommended classification by the World Health Organization (WHO) The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  BMI ≥ 30 kg/m2 Measure of frequency absolute number Frequency value 1061 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (BMI ≥ 30 kg/m2) Outcome mortality Prognostic factor (category):    BMI ≥ 30 kg/m2 Follow‐up Number of patients followed completely for this outcome  3635 Number of obese patients followed completely for this outcome  1061 Number of non‐obese patients followed completely for this outcome  2574 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, BMI < 25 kg/m2, BMI > 30 kg/m2, connective tissue disease, elevated creatinine, GCS < 15, ICU (intensive care unit) admission, peripheral oxygen saturation (SpO2) < 92%  Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.15 (0.893,1.479)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Ahlstrom 2021.

Study characteristics
Notes	English title The Swedish Covid‐19 intensive care cohort: risk factors of ICU admission and ICU mortality Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): 05/2020 Study design: Case‐control Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Inpatient Number of participants recruited: 1981 Sampling method: Consecutive participants Participants Female participants (absolute number): 516 Age measure, value: Median (IQR), 61 (52‐69) Inclusion criteria: The study population was defined by at least one COVID‐19 registration in the SIRI until data acquisition on 27 May 2020. From RTB, four age‐ and sex‐matched controls per patient were drawn. Age matching was performed as close to ICU admission as possible, on the age at 31 January 2020. Cases could not become controls and controls could not be selected twice. Exclusion criteria: Exclusion criteria were aged < 18 years or the absence of a Swedish personal identification number (PIN) Smoking frequency: NR Diabetes frequency: 522 Hypertension frequency: 982 Cardiovascular disease frequency: 317 Asthma frequency: 133 Chronic obstructive pulmonary disease frequency: 75 Other pulmonary disease frequency: NR Immunosuppression frequency: 236 Chronic kidney disease frequency: 75 Cancer frequency: 94 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Obesity was defined based on ICD‐10 coding E66 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 123 Prognostic factor(s): BMI > 30 kg/m2 Outcome(s) Mortality Outcome (prognostic factor) Mortality (BMI > 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 1981 Number of obese patients followed completely for the outcome: 123 Number of non‐obese patients followed completely for the outcome: 1858 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, gender, simplified acute physiology score 3 (SAPS3) Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.87 (0.51, 1.48), 0.61
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Al‐Sabah 2020.

Study characteristics
Notes	English title COVID‐19: impact of obesity and diabetes on disease severity Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 1158 Sampling method: Consecutive participants Participants Female participants (absolute number): 213 Age measure, value: Median (IQR), 40.5 (31.5‐52.1) Inclusion criteria: The patients with positive results of COVID‐19 test were included in the study. Exclusion criteria: The patients with negative or equivocal results of COVID‐19 test were excluded. Smoking frequency: NR Diabetes frequency: 271 Hypertension frequency: 236 Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Normal weight (BMI of 18.5‐24.9 kg/m2), overweight (BMI of 25.0‐29.9 kg/m2) and obese (BMI ≥ 30 kg/m2). The subjects with obesity were further stratified into classes: class I obesity was defined as a BMI of 30‐34.9 kg/m2; class II obesity, by a BMI of 35‐39.9 kg/m2; and morbid obesity, by a BMI ≥ 40 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination: BMI Threshold used for definition: 30 Obesity frequency (Absolute number): 157 Prognostic factor(s): Class I obesity (BMI of 30‐34.9 kg/m2) Class II obesity (BMI of 35‐39.9 kg/m2) Morbid obesity (BMI ≥ 40 kg/m2) Outcome(s) ICU admission Outcome (prognostic factor) Mortality (Class I obesity (BMI of 30‐34.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 1158 Number of obese patients followed completely for the outcome: 157 Number of non‐obese patients followed completely for the outcome: 570 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.51 (1.60, 7.69), 0.002 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.70 (1.17, 6.20), 0.019 Outcome (prognostic factor) Mortality (Class II obesity (BMI of 35‐39.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 1158 Number of obese patients followed completely for the outcome: 157 Number of non‐obese patients followed completely for the outcome: 570 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.78 (0.93, 8.27), 0.066 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.61 (0.50, 5.15), 0.423 Outcome (prognostic factor) Mortality (Morbid obesity (BMI ≥ 40 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 1158 Number of obese patients followed completely for the outcome: 157 Number of non‐obese patients followed completely for the outcome: 570 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 5.18 (1.50, 17.85), 0.009 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.95 (1.00, 15.20), 0.046
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias ICU admission	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Al‐Salameh 2020.

Study characteristics
Notes	English title The association between body mass index class and coronavirus disease 2019 outcomes Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 329 Sampling method: Consecutive participants Participants Female participants (absolute number): 143 Age measure, value: Median (IQR), 72 (61‐84) Inclusion criteria: A confirmed diagnosis of COVID‐19 and inpatient admission to Amiens University Hospital Exclusion criteria: Opposition to data collection by the patient or his/her legal guardian and age under 18 Smoking frequency: 249 Diabetes frequency: 93 Hypertension frequency: 202 Cardiovascular disease frequency: 115 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 34 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 50 Cancer frequency: 53 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Underweight (BMI < 18.5), normal weight (BMI of 18.5‐24.9 kg/m2), overweight (BMI of 25.0‐29.9 kg/m2) and obese (BMI ≥ 30 kg/m2) The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 124 Prognostic factor(s): BMI ≥ 30 kg/m2 Outcome(s) ICU admission Mortality Mechanical ventilation Outcome (prognostic factor) ICU admission (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 329 Number of obese patients followed completely for the outcome: 124 Number of non‐obese patients followed completely for the outcome: 205 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 6 (3.08, 12.52), < 0.0001 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, ALAT > 40 U/I, ASAT > 40 U/I, cancer, cardiac disease, COPD, CRP on admission, CVD total, diabetes, sex (female) Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.05 (1.25, 7.82), 0.017 Outcome (prognostic factor) Mortality (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 329 Number of obese patients followed completely for the outcome: 124 Number of non‐obese patients followed completely for the outcome: 205 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.78 (0.41, 1.47), 0.44 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, ALAT > 40 U/I, ASAT > 40 U/I, cancer, cardiac disease, COPD, CRP on admission, CVD total, diabetes, sex (female) Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.35 (0.47, 3.96), 0.025 Outcome (prognostic factor) Mortality (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 329 Number of obese patients followed completely for the outcome: 124 Number of non‐obese patients followed completely for the outcome: 205 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 8.21 (3.21, 25.48), < 0.0001
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Confounding Bias Mechanical ventilation	No	Appendix 3
Confounding Bias ICU admission	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Alkhatib 2020.

Study characteristics
Notes	English title BMI is associated with coronavirus disease 2019 Intensive Care Unit admission in African Americans Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Registry data Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Outpatient and inpatient Number of participants recruited: 158 Sampling method: Consecutive participants Participants Female participants (absolute number): 97 Age measure, value: Mean (SD), 57 (15.1) Inclusion criteria: Self‐reported African American patients confirmed to have COVID‐19 who presented to a tertiary academic hospital Exclusion criteria: Patients with missing data or pending COVID‐19 confirmatory testing Smoking frequency: NR Diabetes frequency: 76 Hypertension frequency: 107 Cardiovascular disease frequency: 21 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 32 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 21 Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: World Health Organization (WHO) obesity class, defined as: class I obesity (30.0–34.9 kg/m2), class II obesity (35.0–39.9 kg/m2), and class III obesity (≥ 40.0 kg/m2) The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 96 Prognostic factor(s): BMI Outcome(s) ICU admission Outcome (prognostic factor) ICU admission (BMI) Follow‐up Number of patients followed completely for the outcome: 158 Number of obese patients followed completely for the outcome: 96 Number of non‐obese patients followed completely for the outcome: 62 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.063 (1.020, 1.108), 0.0044 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, CKD, congestive heart failure, DM, hypertension, obstructive lung disease (both chronic obstructive pulmonary disease and asthma), sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.115 (1.052, 1.182), 0.0002
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Anderson 2021.

Study characteristics
Notes	English title Body mass index and risk for intubation or death in SARS‐CoV‐2 infection: a retrospective cohort study Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 04/2020 Study design retrospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 2 Study setting inpatient Number of participants recruited 533 Sampling method consecutive participants Participants Female participants (percentage), 42 Age measure, value median (interquartile range), 67 (54, 78) Inclusion criteria COVID‐19 positive  Exclusion criteria NR Smoking (percentage), 12 Diabetes (percentage), 40 Hypertension (percentage), 52 Cardiovascular diseases NR Please indicate if additional information is available  NR Asthma NR Chronic obstructive pulmonary disease NR Other pulmonary diseases NR Please indicate if additional information is available  NR Immunosuppression NR Please indicate if additional information is available NR Chronic kidney disease (percentage), 18 Cancer (percentage), 13 Steroid administration NR Supplemental oxygen NR Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity The BMI categories were defined a priori by using the World Health Organization criteria: underweight (< 18.5 kg/m2), normal weight (18.5 to 24.9 kg/m2), overweight (25.0 to 29.9 kg/m2), class 1 obesity (30 to 34.9 kg/m2), class 2 obesity (35 to 39.9 kg/m2), and class 3 obesity (≥ 40 kg/m2).  The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  BMI ≥ 30 kg/m2 Measure of frequency absolute number Frequency value 189 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (class I obesity (30 to 34.9 kg/m2)) Outcome Mortality Prognostic factor (category):  class I obesity (30 to 34.9 kg/m2) Follow‐up Number of patients followed completely for this outcome  NR Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.2 (0.8, 1.7) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, asthma or chronic obstructive pulmonary disease, cancer, chronic kidney disease, diabetes, hypertension, pulmonary hypertension, race/ethnicity, sex, smoking  Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.4 (0.97, 2.0) Outcome (prognostic factor) mortality (class II obesity (35 to 39.9 kg/m2)) Outcome Mortality Prognostic factor (category):  class II obesity (35 to 39.9 kg/m2) Follow‐up Number of patients followed completely for this outcome  NR Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 0.7 (0.4, 1.1) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, asthma or chronic obstructive pulmonary disease, cancer, chronic kidney disease, diabetes, hypertension, pulmonary hypertension, race/ethnicity, sex, smoking  Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.0 (0.6, 1.7) Outcome (prognostic factor) mortality (class III obesity (≥ 40 kg/m2)) Outcome Mortality Prognostic factor (category):  class III obesity (≥ 40 kg/m2) Follow‐up Number of patients followed completely for this outcome  NR Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 0.9 (0.5, 1.5) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, asthma or chronic obstructive pulmonary disease, cancer, chronic kidney disease, diabetes, hypertension, pulmonary hypertension, race/ethnicity, sex, smoking  Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.6 (0.9, 2.7)
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Apea 2021.

Study characteristics
Notes	English title Ethnicity and outcomes in patients hospitalised with COVID‐19 infection in East London: an observational cohort study Study setting Start of study recruitment (MM/YYYY): 01/2020 End of study recruitment (MM/YYYY): 05/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 5 Study setting: Inpatient Number of participants recruited: 1996 Sampling method: Consecutive participants Participants Female participants (absolute number): 786 Age measure, value: Mean (SD), 63.40 (18.22) Inclusion criteria: NR Exclusion criteria: Those under 16 years were excluded. The first emergency admission encompassing the first positive SARS‐CoV‐2 test, or the first emergency admission within 2 weeks of positive outpatient testing was defined as the index admission; community diagnoses without an associated emergency hospital admission were excluded. Smoking frequency: 173 Diabetes frequency: 664 Hypertension frequency: 1009 Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 397 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 363 Cancer frequency: 144 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: World Health Organization (WHO) obesity class, defined as: class I obesity (30.0–34.9 kg/m2), class II obesity (35.0–39.9 kg/m2), and class III obesity (≥ 40.0 kg/m2) The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 409 Prognostic factor(s): BMI ≥ 30 kg/m2 Outcome(s) Mortality Outcome (prognostic factor) Mortality (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 1996 Number of obese patients followed completely for the outcome: 409 Number of non‐obese patients followed completely for the outcome: 839 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.42 (1.09, 1.85), 0.009
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Argenziano 2020.

Study characteristics
Notes	English title Characterization and clinical course of 1000 patients with coronavirus disease 2019 in New York: retrospective case series Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Outpatient and inpatient Number of participants recruited: 1000 Sampling method: Consecutive participants Participants Female participants (absolute number): 404 Age measure, value: Median (IQR), 63 (50‐75) Inclusion criteria: All patients with COVID‐19 who received emergency department or inpatient care at NYP/CUIMC Exclusion criteria: COVID‐19 patients who had performed their test in the outpatient setting or at another hospital Smoking frequency: 230 (including ex‐smokers) Diabetes frequency: 372 Hypertension frequency: 601 Cardiovascular disease frequency: NR Asthma frequency: 113 Chronic obstructive pulmonary disease frequency: 66 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: 67 Steroid administration frequency: 178 (only inpatient) Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 30 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 352 Prognostic factor(s): BMI Outcome(s) Mortality Mechanical ventilation Outcome (prognostic factor) Mortality (BMI) Follow‐up Number of patients followed completely for the outcome: 1000 Number of obese patients followed completely for the outcome: 352 Number of non‐obese patients followed completely for the outcome: 489 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.02 (1.00, 1.05), 0.025 Outcome (prognostic factor) Mechanical ventilation (BMI) Follow‐up Number of patients followed completely for the outcome: 1000 Number of obese patients followed completely for the outcome: 352 Number of non‐obese patients followed completely for the outcome: 489 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.02 (0.999, 1.04), 0.0682
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Confounding Bias Mechanical ventilation	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Awad 2021.

Study characteristics
Notes	English title Impact of hydroxychloroquine on disease progression and ICU admissions in patients with SARS‐CoV‐2 infection Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 04/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 336 Sampling method consecutive participants Participants Female participants (absolute number), 157 Age measure, value mean (standard deviation), 64.3 (17) Inclusion criteria Patients were included in the cohorts if they were admitted during the study time frame and tested positive for SARS‐CoV‐2. Exclusion criteria Patients were excluded if they did not test positive for SARS‐CoV‐2, if they required intubation within 24 hours of admission, or if they were not admitted to the hospital. Smoking NR Diabetes (absolute number), 94 Hypertension NR Cardiovascular diseases NR Please indicate if additional information is available  NR Asthma NR Chronic obstructive pulmonary disease NR Other pulmonary diseases NR Please indicate if additional information is available  NR Immunosuppression NR Please indicate if additional information is available NR Chronic kidney disease NR Cancer NR Steroid administration NR Supplemental oxygen NR Differential values for various oxygenation methods (if indicated) NR Other treatment hydroxychloroquine Dose if applicable loading dose of 400 mg every 12 hours on day 1 followed by a dose of 200 mg every 12 hours Duration if applicable 5‐day course of therapy  Percentage received this treatment 44.04 Prognostic factor(s) Study’s definition for obesity World Health Organization (WHO) obesity class, defined as: overweight (25.0‐ 29.90 kg/m2) class I obesity (30.0–34.9 kg/m2), class II obesity (35.0–39.9 kg/m2), and class III obesity (≥ 40.0 kg/m2) The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  BMI ≥ 25 kg/m2 Measure of frequency absolute number Frequency value 241 How many eligible outcomes reported? 2 How many eligible outcomes reported? 2 Outcome(s) ICU admission, mechanical ventilation Outcome (prognostic factor) ICU admission (BMI ≥ 25 kg/m2) Outcome ICU admission Prognostic factor (category):  BMI ≥ 25 kg/m2 Follow‐up Number of  patients followed completely for this outcome  336 Number of obese patients followed completely for this outcome  241 Number of non‐obese patients followed completely for this outcome  95 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.26 (0.69, 2.31) Outcome (prognostic factor) mechanical ventilation (BMI ≥ 25 kg/m2) Outcome mechanical ventilation Prognostic factor (category):  BMI ≥ 25 kg/m2 Follow‐up Number of patients followed completely for this outcome  336 Number of obese patients followed completely for this outcome  241 Number of non‐obese patients followed completely for this outcome  95 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.19 (0.64, 2.2) Outcome (prognostic factor) ICU admission (BMI ≥ 25 kg/m2) Outcome ICU admission Prognostic factor (category):  BMI ≥ 25 kg/m2 Follow‐up Number of patients followed completely for this outcome  122 Number of obese patients followed completely for this outcome  46 Number of non‐obese patients followed completely for this outcome  76 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 4.81 (0.94, 24.73) Outcome (prognostic factor) mechanical ventilation (BMI ≥ 25 kg/m2) Outcome mechanical ventilation Prognostic factor (category):  BMI ≥ 25 kg/m2 Follow‐up Number of patients followed completely for this outcome  122 Number of obese patients followed completely for this outcome  46 Number of non‐obese patients followed completely for this outcome  76 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 4.41 (1.01, 19.28) Outcome (prognostic factor) ICU admission (BMI ≥ 25 kg/m2) Outcome ICU admission Prognostic factor (category) BMI ≥ 25 kg/m2 Follow‐up Number of patients followed completely for this outcome  214 Number of obese patients followed completely for this outcome  73 Number of non‐obese patients followed completely for this outcome  141 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.03 (0.51, 2.06) Outcome (prognostic factor) mechanical ventilation (BMI ≥ 25 kg/m2) Outcome mechanical ventilation Prognostic factor (category):  BMI ≥ 25 kg/m2 Follow‐up Number of patients followed completely for this outcome  214 Number of obese patients followed completely for this outcome  73 Number of non‐obese patients followed completely for this outcome  141 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 0.95 (0.47, 1.94)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mechanical ventilation	No	Appendix 3
Confounding Bias ICU admission	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Baronio 2020.

Study characteristics
Notes	English title Italian SARS‐CoV‐2 patients in intensive care: towards an identikit for subjects at risk? Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Case series Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 191 (cohort 1), 157 (cohort 2) Sampling method: Consecutive participants Participants Female participants (absolute number): 42 (cohort 1), 33 (cohort 2) Age measure, value: Mean (SD), 64.5 (9.9) (cohort 1), median (IQR), 33 (59‐70) (cohort 2) Inclusion criteria: 157 critically ill patients from the Intensive Care Unit (ICU) and 34 stable patients from the Medical and Surgical Departments (hereafter referred to as 'controls'), who did not develop severe respiratory failure (cohort 1), 157 critically ill patients from the ICU (cohort 2) Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 34 (cohort 1), 25 (cohort 2) Hypertension frequency: NR Cardiovascular disease frequency: 114 (cohort 1), 94 (cohort 2) Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI values were classified in three categories: optimal (< 25 kg/m2), overweight (25‐30 kg/m2) and obese (≥ 30 kg/m2). The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 59 (cohort 1), 52 (cohort 2) Prognostic factor(s): BMI > 30 kg/m2 Outcome(s) ICU admission Mortality Outcome (prognostic factor) ICU admission (BMI > 30 kg/m2) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 191 Number of obese patients followed completely for the outcome: 59 Number of non‐obese patients followed completely for the outcome: 132 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Without adjustment Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 5.63 (1.73, 21.09), NR Outcome (prognostic factor) Mortality (BMI > 30 kg/m2) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 191 Number of obese patients followed completely for the outcome: 59 Number of non‐obese patients followed completely for the outcome: 132 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Without adjustment Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 4.66 (1.76, 13.15), NR Outcome (prognostic factor) Mortality (BMI > 30 kg/m2) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 157 Number of obese patients followed completely for the outcome: 52 Number of non‐obese patients followed completely for the outcome: 105 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, BMI, lymphocyte count, temperature, sex Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 2.23 (1.15, 4.35), 0.02
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Confounding Bias ICU admission	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Bartoletti 2020.

Study characteristics
Notes	English title Development and validation of a prediction model for severe respiratory failure in hospitalised patients with SARS‐CoV‐2 infection: a multicentre cohort study (PREDI‐CO study) Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 4 (cohort 1), 7 (cohort 2) Study setting: Inpatient Number of participants recruited: 644 (cohort 1), 469 (cohort 2) Sampling method: Consecutive participants Participants Female participants (absolute number): 268 (cohort 1), 141 (cohort 2) Age measure, value: Mean (SD), 63.7 (15.6) (cohort 1), 68.5 (14.1) (cohort 2) Inclusion criteria: All consecutive adults (18 years) diagnosed with SARS‐CoV‐2 infection during the study period Exclusion criteria: Hospital discharge within 24 hours of admission to Emergency Department and occurrence of SRF (severe respiratory failure) within 24 hours of hospitalisation Smoking frequency: NR Diabetes frequency: 37 (cohort 1), 23 (cohort 2) Hypertension frequency: 321 (cohort 1), 258 (cohort 2) Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 58 (cohort 1), 55 (cohort 2) Other pulmonary disease frequency: NR Immunosuppression frequency: 21 (cohort 1), 21 (cohort 2) Chronic kidney disease frequency: 61 (cohort 1), 54 (cohort 2) Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI > 30 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 122 (cohort 1), 74 (cohort 2) Prognostic factor(s): BMI > 30 kg/m2 Outcome(s) Severe respiratory failure Outcome (prognostic factor) Severe respiratory failure (BMI > 30 kg/m2) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 644 Number of obese patients followed completely for the outcome: 122 Number of non‐obese patients followed completely for the outcome: 522 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 6.09 (3.99, 9.30), NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 4.62 (7.70, 2.78), < 0.001 Outcome (prognostic factor) Severe respiratory failure (BMI > 30 kg/m2) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 469 Number of obese patients followed completely for the outcome: 74 Number of non‐obese patients followed completely for the outcome: 395 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.26 (1.37, 3.74), NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.07 (0.72, 1.60), 0.73
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Bellini 2021.

Study characteristics
Notes	English title Obesity as a risk factor for hospitalization in COronaVIrus Disease19 (COVID‐19) patients: analysis of the Tuscany regional database Study setting Start of study recruitment (MM/YYYY) 30/04/2020 End of study recruitment (MM/YYYY) 30/04/2020 Study design registry data Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting outpatient and inpatient Number of participants recruited 4481 Sampling method consecutive participants Participants Female participants (percentage), 49.5 Age measure, value NR Inclusion criteria NR Exclusion criteria patients with any missing data Smoking NR Diabetes NR Hypertension NR Cardiovascular diseases NR Please indicate if additional information is available  NR Asthma NR Chronic obstructive pulmonary disease NR Other pulmonary diseases NR Please indicate if additional information is available  NR Immunosuppression NR Please indicate if additional information is available NR Chronic kidney disease NR Cancer NR Steroid administration NR Supplemental oxygen NR Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Obesity was defined as BMI ≥ 30 kg/m2 The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  BMI ≥ 30 kg/m2 Measure of frequency absolute number Frequency value 157 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) hospitalisation Outcome (prognostic factor) Hospitalisation (BMI > 30 kg/m2 (obese)) Outcome Hospitalisation Prognostic factor (category):  BMI > 30 kg/m2 (obese) Follow‐up Number of patients followed completely for this outcome  4481 Number of obese patients followed completely for this outcome  157 Number of non‐obese patients followed completely for this outcome  4424 Univariable (unadjusted) analysis for obesity Effect measure for obesity relative risk Effect measure value (95% CI) 1.74 (1.56, 1.97) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, sex, all factors associated with hospitalisation at univariate analysis as possible confounders (factors were not reported) Effect measure for obesity odds ratio  Effect measure value (95% CI) 3 (2.16, 4.29) Outcome (prognostic factor) Hospitalisation (BMI > 30 kg/m2 (obese)) Outcome Hospitalisation Prognostic factor (category):  BMI > 30 kg/m2 (obese) Follow‐up Number of patients followed completely for this outcome  2307 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, sex, all factors associated with hospitalisation at univariate analysis as possible confounders (factors were not reported) Effect measure for obesity odds ratio  Effect measure value (95% CI) 4.08 (2.53, 6.77) Outcome (prognostic factor) Hospitalisation (BMI > 30 kg/m2 (obese)) Outcome Hospitalisation Prognostic factor (category):  BMI > 30 kg/m2 (obese) Follow‐up Number of patients followed completely for this outcome  2377 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, sex, all factors associated with hospitalisation at univariate analysis as possible confounders (factors were not reported) Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.91 (1.4, 2.65)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Hospitalisation	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Bello‐Chavolla 2021.

Study characteristics
Notes	English title Unequal impact of structural health determinants and comorbidity on COVID‐19 severity and lethality in older Mexican adults: considerations beyond chronological aging Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): 06/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 475 Study setting: Outpatient and inpatient Number of participants recruited: 101,238 Sampling method: Consecutive participants Participants Female participants (absolute number): 44,239 Age measure, value: NR Inclusion criteria: All SARS‐CoV‐2 PCR‐positive cases up to June 3, 2020, in individuals aged 60 and older. Exclusion criteria: NR Smoking frequency: 8333 Diabetes frequency: 17,489 Hypertension frequency: 20,955 Cardiovascular disease frequency: 2594 Asthma frequency: 2930 Chronic obstructive pulmonary disease frequency: 1990 Other pulmonary disease frequency: 26,925 Immunosuppression frequency: 1555 Chronic kidney disease frequency: 2339 Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: Unspecified Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): 20,599 Prognostic factor(s): Obesity Outcome(s) Pneumonia Hospitalisation ICU admission Mechanical ventilation Mortality Outcome (prognostic factor) Pneumonia (obesity) Follow‐up Number of patients followed completely for the outcome: 101,238 Number of obese patients followed completely for the outcome: 20,599 Number of non‐obese patients followed completely for the outcome: 80,639 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, male sex, indigenous, CVD, CKD, COPD, immunosuppression, smoking, diabetes, obesity, hypertension, social lag index Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.26 (1.11, 1.36), NR Outcome (prognostic factor) Hospitalisation (obesity) Follow‐up Number of patients followed completely for the outcome: 101,238 Number of obese patients followed completely for the outcome: 20,599 Number of non‐obese patients followed completely for the outcome: 80,639 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, male sex, indigenous, CVD, CKD, COPD, immunosuppression, smoking, diabetes, obesity, hypertension, social lag index Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.09 (1.01, 1.18), NR Outcome (prognostic factor) ICU admission (obesity) Follow‐up Number of patients followed completely for the outcome: 101,238 Number of obese patients followed completely for the outcome: 20,599 Number of non‐obese patients followed completely for the outcome: 80,639 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, male sex, indigenous, CVD, CKD, COPD, immunosuppression, smoking, diabetes, obesity, hypertension, social lag index Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.26 (1.09, 1.45), NR Outcome (prognostic factor) Mechanical ventilation (obesity) Follow‐up Number of patients followed completely for the outcome: 101,238 Number of obese patients followed completely for the outcome: 20,599 Number of non‐obese patients followed completely for the outcome: 80,639 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, male sex, indigenous, CVD, CKD, COPD, immunosuppression, smoking, diabetes, obesity, hypertension, social lag index Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.31 (1.15, 1.50), NR Outcome (prognostic factor) Mortality (obesity) Follow‐up Number of patients followed completely for the outcome: 101,238 Number of obese patients followed completely for the outcome: 20,599 Number of non‐obese patients followed completely for the outcome: 80,639 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, male sex, indigenous, CVD, CKD, COPD, immunosuppression, smoking, diabetes, obesity, hypertension, social lag index Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.19 (1.12, 1.27), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Study Attrition Pneumonia	Unclear	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Outcome Measurement Pneumonia	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Confounding Bias Pneumonia	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Bennett 2021.

Study characteristics
Notes	English title Underlying conditions and risk of hospitalisation, ICU admission and mortality among those with COVID‐19 in Ireland: a national surveillance study Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 07/2020 Study design retrospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 8 Study setting outpatient and inpatient Number of participants recruited 26,106 Sampling method consecutive participants Participants Female participants (absolute number), 11,153 Age measure, value NR Inclusion criteria NR Exclusion criteria NR Smoking NR Diabetes (absolute number), 1224 Hypertension NR Cardiovascular diseases (absolute number), 2700 Please indicate if additional information is available  CVD defined as chronic heart disease Asthma (absolute number), 467 Chronic obstructive pulmonary disease NR Other pulmonary diseases (absolute number), 2053 Please indicate if additional information is available  chronic respiratory disease Immunosuppression (absolute number), 402 Please indicate if additional information is available including HIV Chronic kidney disease (absolute number), 558 Cancer (absolute number), 747 Steroid administration NR Supplemental oxygen NR Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Detailed body mass index (BMI) information is not routinely captured by the ESF and instead recorded as the presence or absence of morbid obesity, defined as a BMI of ≥ 40 kg/m2 The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  BMI ≥ 40 kg/m2 (only morbidly obese patients) Measure of frequency absolute number Frequency value 298 How many eligible outcomes reported? 3 How many eligible outcomes reported? 3 Outcome(s) mortality, hospitalisation, ICU admission Outcome (prognostic factor) mortality (BMI ≥ 40 kg/m2 (morbid obesity)) Outcome mortality Prognostic factor (category):  BMI ≥ 40 kg/m2 (morbid obesity) Follow‐up Number of patients followed completely for this outcome  19,789 Number of obese patients followed completely for this outcome  298 Number of non‐obese patients followed completely for this outcome  19,491 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age (linear, quadratic, cubic), asthma (requiring meds), BMI > 40, cancer, diabetes, chronic heart disease, chronic kidney disease, chronic liver disease, chronic neurological disease, Community health office, chronic respiratory disease, immunodeficiency, other comorbidity, residential care facility, route of transmission, unknown comorbidity Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.89 (1.8, 4.64) Outcome (prognostic factor) hospitalisation (BMI ≥ 40 kg/m2 (morbid obesity)) Outcome hospitalisation Prognostic factor (category):  BMI ≥ 40 kg/m2 (morbid obesity) Follow‐up Number of  patients followed completely for this outcome  19,789 Number of obese patients followed completely for this outcome  298 Number of non‐obese patients followed completely for this outcome  19,491 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age (linear, quadratic, cubic), asthma (requiring meds), BMI > 40, cancer, diabetes, chronic heart disease, chronic kidney disease, chronic liver disease, chronic neurological disease, community health office, chronic respiratory disease, immunodeficiency, other comorbidity, residential care facility, route of transmission, unknown comorbidity Effect measure for obesity odds ratio  Effect measure value (95% CI) 4.29 (3.27, 5.65) Outcome (prognostic factor) ICU admission (BMI ≥ 40 kg/m2 (morbid obesity)) Outcome ICU admission Prognostic factor (category):  BMI ≥ 40 kg/m2 (morbid obesity) Follow‐up Number of  patients followed completely for this outcome  2811 Number of obese patients followed completely for this outcome  134 Number of non‐obese patients followed completely for this outcome  2677 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age (linear, quadratic, cubic), asthma (requiring meds), BMI > 40, cancer, diabetes, chronic heart disease, chronic kidney disease, chronic liver disease, chronic neurological disease, community health office, chronic respiratory disease, immunodeficiency, other comorbidity, residential care facility, route of transmission, unknown comorbidity Effect measure for obesity odds ratio  Effect measure value (95% CI) 7.53 (4.94, 11.48) Outcome (prognostic factor) mortality (BMI ≥ 40 kg/m2 (morbid obesity)) Outcome mortality Prognostic factor (category):  BMI ≥ 40 kg/m2 (morbid obesity) Follow‐up Number of patients followed completely for this outcome  2811 Number of obese patients followed completely for this outcome  134 Number of non‐obese patients followed completely for this outcome  2677 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age (linear, quadratic, cubic), asthma (requiring meds), BMI > 40, cancer, diabetes, chronic heart disease, chronic kidney disease, chronic liver disease, chronic neurological disease, community health office, chronic respiratory disease, immunodeficiency, other comorbidity, residential care facility, route of transmission, unknown comorbidity Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.19 (1.34, 3.56)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Study Attrition Hospitalisation	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Bhatt 2021.

Study characteristics
Notes	English title Clinical outcomes in patients with heart failure hospitalized with COVID‐19 Study setting Start of study recruitment (MM/YYYY): 04/2020 End of study recruitment (MM/YYYY): 09/2020 Study design: Prospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: More than 1041 centres Study setting: Inpatient Number of participants recruited: 8383 Sampling method: Consecutive participants Participants Female participants (absolute number): 4205 Age measure, value: Mean (SD), 72 (13.2) Inclusion criteria: Patients with at least 1 Heart Failure (HF) hospitalisation or 2 HF outpatient visits between January 1, 2019, and March 31, 2020, who were subsequently hospitalised between April and September 2020 with coronavirus disease‐2019 (COVID‐19) Exclusion criteria: NR Smoking frequency: 3665 Diabetes frequency: 5107 Hypertension frequency: 6997 Cardiovascular disease frequency: 4548 (arrhythmia), 1417 (valvular disease) Asthma frequency: 628 Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 3539 Immunosuppression frequency: NR Chronic kidney disease frequency: 5020 Cancer frequency: 290 Steroid administration frequency: NR Supplemental oxygen administration frequency: 17 Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: NR Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): 2461 Prognostic factor(s): Obesity Morbid obesity Outcome(s) In‐hospital mortality Outcome (prognostic factor) In‐hospital mortality (obesity) Follow‐up Number of patients followed completely for the outcome: 8383 Number of obese patients followed completely for the outcome: 2461 Number of non‐obese patients followed completely for the outcome: 5922 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, race, discharge month, region, LVEF, obesity, morbid obesity, diabetes mellitus, hypertension, kidney disease, pulmonary disease, smoking, malignancy Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.19 (1.01, 1.40), NR Outcome (prognostic factor) In‐hospital mortality (obesity) Follow‐up Number of patients followed completely for the outcome: 8383 Number of obese patients followed completely for the outcome: 2461 Number of non‐obese patients followed completely for the outcome: 5922 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, race, discharge month, region, LVEF, obesity, morbid obesity, diabetes mellitus, hypertension, kidney disease, pulmonary disease, smoking, malignancy Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.25 (1.07, 1.46), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Biscarini 2020.

Study characteristics
Notes	English title The obesity paradox: analysis from the SMAtteo COvid‐19 REgistry (SMACORE) cohort Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 03/2020 Study design: Retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 427 Sampling method: Consecutive participants Participants Female participants (absolute number): 136 Age measure, value: Mean (SD), 67 (21) Inclusion criteria: Patients with confirmed diagnosis of COVID‐19 hospitalised between 21st February and 31st March 2020 Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 66 Hypertension frequency: 174 Cardiovascular disease frequency: 98 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 212 Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: 22 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Obesity was defined as BMI > 30 kg/m2 The time when obesity has been measured: Some time after presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 80 Prognostic factor(s): BMI > 30 kg/m2 Outcome(s) Mortality ICU admission  Death in ICU Length of stay Outcome (prognostic factor) Mortality (BMI > 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 427 Number of obese patients followed completely for the outcome: 80 Number of non‐obese patients followed completely for the outcome: 252 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, antibiotic therapy, antiviral therapy, CRP > 10 mg/dL, diabetes, heart disease, hypertension, interstitial pneumonia, obesity, PF ratio < 260 (arterial partial pressure of oxygen (PaO2)/fractional inspired oxygen (fiO2)), respiratory frequency, sex, tumour Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.03 (0.65, 1.67), NR Outcome (prognostic factor) ICU admission (BMI > 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 427 Number of obese patients followed completely for the outcome: 80 Number of non‐obese patients followed completely for the outcome: 252 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, antibiotic therapy, antiviral therapy, CRP > 10 mg/dL, diabetes, heart disease, hypertension, interstitial pneumonia, obesity, PF ratio < 260 (arterial partial pressure of oxygen (PaO2)/fractional inspired oxygen (fiO2)), respiratory frequency, sex, tumour Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.96 (1.03, 3.75), NR Outcome (prognostic factor) Death in ICU (BMI > 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 427 Number of obese patients followed completely for the outcome: 80 Number of non‐obese patients followed completely for the outcome: 252 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, antibiotic therapy, antiviral therapy, CRP > 10 mg/dL, diabetes, heart disease, hypertension, interstitial pneumonia, obesity, PF ratio < 260 (arterial partial pressure of oxygen (PaO2)/fractional inspired oxygen (fiO2)), respiratory frequency, sex, tumour Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.65 (0.38, 7.15), NR Outcome (prognostic factor) Length of stay (BMI > 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 427 Number of obese patients followed completely for the outcome: 80 Number of non‐obese patients followed completely for the outcome: 252 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Linear regression The set of prognostic factors used for adjustment: Age, antibiotic therapy, antiviral therapy, CRP > 10 mg/dL, diabetes, heart disease, hypertension, interstitial pneumonia, obesity, PF ratio < 260 (arterial partial pressure of oxygen (PaO2)/fractional inspired oxygen (fiO2)), respiratory frequency, sex, tumour Effect measure for obesity: Slope (beta) Effect measure value (95% CI), P value: 1.19 (‐1.88, 4.26), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Bonifazi 2021.

Study characteristics
Notes	English title Predictors of worse prognosis in young and middle‐aged adults hospitalized with COVID‐19 pneumonia: a multi‐center Italian study (COVID‐UNDER50) Study setting Start of study recruitment (MM/YYYY) 02/2020 End of study recruitment (MM/YYYY) 05/2020 Study design retrospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 9 Study setting inpatient Number of participants recruited 263 Sampling method consecutive participants Participants Female participants (absolute number), 99 Age measure, value median (Interquartile range), 45.3 (40.4, 48.4) Inclusion criteria patients, aged 18–50 years, hospitalised for confirmed or probable diagnosis of SARS‐CoV2 infection Exclusion criteria NR Smoking NR Diabetes NR Hypertension NR Cardiovascular diseases NR Please indicate if additional information is available  NR Asthma NR Chronic obstructive pulmonary disease NR Other pulmonary diseases NR Please indicate if additional information is available  NR Immunosuppression NR Please indicate if additional information is available NR Chronic kidney disease NR Cancer NR Steroid administration NR Supplemental oxygen (absolute number), 88 Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity up to normal weight (BMI < 25 kg/m2), overweight (25 ≤ BMI < 30 kg/m2) and obese (BMI ≥ 30 kg/m2). The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  BMI ≥ 30 kg/m2 Measure of frequency absolute number Frequency value 51 How many eligible outcomes reported? 2 How many eligible outcomes reported? 2 Outcome(s) Mechanical ventilation, mortality Outcome (prognostic factor) Mechanical ventilation (BMI ≥ 30 kg/m2) Outcome Mechanical ventilation Prognostic factor (category):  BMI ≥ 30 kg/m2 Follow‐up Number of patients followed completely for this outcome  263 Number of obese patients followed completely for this outcome  51 Number of non‐obese patients followed completely for this outcome  146 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, comorbidities, smoking status, sex Effect measure for obesity odds ratio  Effect measure value (95% CI) 3.5 (1.44, 8.79) Outcome (prognostic factor) Mechanical ventilation (25 < BMI < 30 (overweight)) Outcome Mechanical ventilation Prognostic factor (category):  25 < BMI < 30 (overweight) Follow‐up Number of patients followed completely for this outcome  263 Number of obese patients followed completely for this outcome  51 Number of non‐obese patients followed completely for this outcome  146 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, comorbidities, smoking status, sex Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.43 (0.54, 3.81) Outcome (prognostic factor) mortality (BMI ≥ 30 kg/m2) Outcome mortality Prognostic factor (category):  BMI ≥ 30 kg/m2 Follow‐up Number of patients followed completely for this outcome  263 Number of obese patients followed completely for this outcome  51 Number of non‐obese patients followed completely for this outcome  146 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, comorbidities Effect measure for obesity odds ratio  Effect measure value (95% CI) 0.79 (0.27, 2.27) Outcome (prognostic factor) mortality (25 < BMI < 30 (overweight)) Outcome mortality Prognostic factor (category):  25 < BMI < 30 (overweight) Follow‐up Number of patients followed completely for this outcome  263 Number of obese patients followed completely for this outcome  51 Number of non‐obese patients followed completely for this outcome  146 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, comorbidities Effect measure for obesity odds ratio  Effect measure value (95% CI) 0.29 (0.05, 1.74)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Confounding Bias Mechanical ventilation	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Breland 2021.

Study characteristics
Notes	English title BMI and risk for severe COVID‐19 among Veterans Health Administration patients Study setting Start of study recruitment (MM/YYYY): 03/2021 End of study recruitment (MM/YYYY): 05/2021 Study design: Prospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Outpatient and inpatient Number of participants recruited: 9347 Sampling method: Consecutive participants Participants Female participants (absolute number): 833 Age measure, value: NR Inclusion criteria: The Veterans Health Administration (VHA) who tested positive for COVID‐19, who had a valid BMI measurement, and who were not VHA employees. Exclusion criteria: Weight < 75 or ≥ 700 lb and height < 48 or ≥ 84 inches Smoking frequency: NR Diabetes frequency: 3560 Hypertension frequency: 5820 Cardiovascular disease frequency: 3003 Asthma frequency: 554 Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 1881 Immunosuppression frequency: 710 Chronic kidney disease frequency: 290 Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: NR Obesity frequency (absolute number): NR Prognostic factor(s): BMI 23‐30 kg/m2 BMI 30‐39 kg/m2 Outcome(s) Mortality ICU admission Hospitalisation Outcome (prognostic factor) BMI 23‐30 kg/m2 (mortality) Follow‐up Number of patients followed completely for the outcome: 9347 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.96 (0.93, 0.98), NR Outcome (prognostic factor) BMI 23‐30 kg/m2 (ICU admission) Follow‐up Number of patients followed completely for the outcome: 9347 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.99 (0.97, 1.02), NR Outcome (prognostic factor) BMI 23‐30 kg/m2 (hospitalisation) Follow‐up Number of patients followed completely for the outcome: 9347 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.97 (0.95, 0.99), NR Outcome (prognostic factor) BMI 30‐39 kg/m2 (mortality) Follow‐up Number of patients followed completely for the outcome: 9347 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.02 (1.02, 1.04), NR Outcome (prognostic factor) BMI 30‐39 kg/m2 (ICU admission) Follow‐up Number of patients followed completely for the outcome: 9347 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.01 (1.00, 1.03), NR Outcome (prognostic factor) BMI 30‐39 kg/m2 (hospitalisation) Follow‐up Number of patients followed completely for the outcome: 9347 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.02 (1.01, 1.03), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Busetto 2020.

Study characteristics
Notes	English title Obesity and COVID‐19: an Italian snapshot Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 92 Sampling method: Consecutive participants Participants Female participants (absolute number): 35 Age measure, value: Mean (SD), 70.5 (13.3) Inclusion criteria: Being positive to an oropharyngeal swab used for real‐time reverse‐transcriptase polymerase chain reaction assays specific for SARS‐CoV‐2 Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 28 Hypertension frequency: 59 Cardiovascular disease frequency: 29 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 12 (chronic respiratory diseases) Immunosuppression frequency: NR Chronic kidney disease frequency: 5 Cancer frequency: 12 Steroid administration frequency: NR Supplemental oxygen administration frequency: 58 Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI > 25 The time when obesity has been measured: Some time after presentation Main variable used for determination of obesity: BMI Threshold used for definition: 25 Obesity frequency (absolute number): 60 Prognostic factor(s): Obesity (BMI > 25) Outcome(s) Mortality Outcome (prognostic factor) Mortality (obesity (BMI > 25)) Follow‐up Number of patients followed completely for the outcome: 92 Number of obese patients followed completely for the outcome: 32 Number of non‐obese patients followed completely for the outcome: 60 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.27 (0.03, 2.05), 0.204
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Cai 2020a.

Study characteristics
Notes	English title Association between obesity and clinical prognosis in patients infected with SARS‐CoV‐2 Study setting Start of study recruitment (MM/YYYY): 01/2020 End of study recruitment (MM/YYYY): 02/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 3 Study setting: Inpatient Number of participants recruited: 96 Sampling method: Consecutive participants Participants Female participants (absolute number): 42 Age measure, value: NR Inclusion criteria: All confirmed SARS‐CoV‐2 infection Exclusion criteria: NR Smoking frequency: 8 Diabetes frequency: NR Hypertension frequency: NR Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Obesity was defined as BMI > 24 kg/m2 The time when obesity has been measured: Some time after presentation Main variable used for determination of obesity: BMI Threshold used for definition: 24 Obesity frequency (absolute number): 37 Prognostic factor(s): BMI > 24 kg/m2 Outcome(s) ICU admission Outcome (prognostic factor) ICU admission (BMI > 24 kg/m2) Follow‐up Number of patients followed completely for the outcome: 92 Number of obese patients followed completely for the outcome: 37 Number of non‐obese patients followed completely for the outcome: 52 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 5.19 (2.11, 12.76), < 0.001 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, BMI Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.258 (1.07, 1.47), 0.005
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Cai 2020b.

Study characteristics
Notes	English title Obesity and COVID‐19 severity in a designated hospital in Shenzhen, China Study setting Start of study recruitment (MM/YYYY): 01/2020 End of study recruitment (MM/YYYY): 02/2020 Study design: Registry data Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 383 Sampling method: Consecutive participants Participants Female participants (absolute number): 200 Age measure, value: NR Inclusion criteria: hospitalised patients with COVID‐19 admitted, aged 18 years or above Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 22 Hypertension frequency: 58 Cardiovascular disease frequency: 35 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 32 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: 5 Steroid administration frequency: NR Supplemental oxygen administration frequency: 3 Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: Some time after presentation Main variable used for determination of obesity: BMI Threshold used for definition: NR Obesity frequency (absolute number): 164 Prognostic factor(s): Overweight Obesity Outcome(s) Severe COVID‐19 Outcome (prognostic factor) Severe COVID‐19 (overweight) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 383 Number of obese patients followed completely for the outcome: 164 Number of non‐obese patients followed completely for the outcome: 219 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: age, sex, epidemiological characteristics, days from disease onset to hospitalisation, disease history, and drugs used for treatment, compared with the normal weight group Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.84 (0.99, 3.43), 0.050 Outcome (prognostic factor) Severe COVID‐19 (obesity) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 383 Number of obese patients followed completely for the outcome: 164 Number of non‐obese patients followed completely for the outcome: 219 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: age, sex, epidemiological characteristics, days from disease onset to hospitalisation, disease history, and drugs used for treatment, compared with the normal weight group Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.40 (1.40, 8.26), 0.007 Outcome (prognostic factor) Severe COVID‐19 (overweight) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 183 Number of obese patients followed completely for the outcome: 73 Number of non‐obese patients followed completely for the outcome: 78 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: age, sex, epidemiological characteristics, days from disease onset to hospitalisation, disease history, and drugs used for treatment, compared with the normal weight group Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.98 (0.78, 5.00), 0.150 Outcome (prognostic factor) Severe COVID‐19 (obesity) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 183 Number of obese patients followed completely for the outcome: 32 Number of non‐obese patients followed completely for the outcome: 78 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: age, sex, epidemiological characteristics, days from disease onset to hospitalisation, disease history, and drugs used for treatment, compared with the normal weight group Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 5.66 (1.80, 17.57), 0.003 Outcome (prognostic factor) Severe COVID‐19 (overweight) (cohort 3) Follow‐up Number of patients followed completely for the outcome: 200 Number of obese patients followed completely for the outcome: 50 Number of non‐obese patients followed completely for the outcome: 141 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: age, sex, epidemiological characteristics, days from disease onset to hospitalisation, disease history, and drugs used for treatment, compared with the normal weight group Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.64 (0.63, 4.29), 0.310 Outcome (prognostic factor) Severe COVID‐19 (obesity) (cohort 3) Follow‐up Number of patients followed completely for the outcome: 200 Number of obese patients followed completely for the outcome: 9 Number of non‐obese patients followed completely for the outcome: 141 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: age, sex, epidemiological characteristics, days from disease onset to hospitalisation, disease history, and drugs used for treatment, compared with the normal weight group Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.70 (7.20, 0.07), 0.760
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Cai 2021.

Study characteristics
Notes	English title High body mass index is a significant risk factor for the progression and prognosis of imported COVID‐19: a multicenter, retrospective cohort study Study setting Start of study recruitment (MM/YYYY) NR End of study recruitment (MM/YYYY) 02/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 455 Sampling method consecutive participants Participants Female participants (absolute number), 239 Age measure, value mean (standard deviation), 44.53 (14.73) Inclusion criteria NR Exclusion criteria NR Smoking NR Diabetes (absolute number), 40 Hypertension (absolute number), 74 Cardiovascular diseases (absolute number), 10 Please indicate if additional information is available  Heart disease Asthma (unspecified), NR Chronic obstructive pulmonary disease (absolute number), 1 Other pulmonary diseases (unspecified), NR Please indicate if additional information is available  NR Immunosuppression (absolute number), 1 Please indicate if additional information is available NR Chronic kidney disease (absolute number), 3 Cancer (absolute number), 6 Steroid administration (absolute number), 89 Supplemental oxygen (absolute number), 8 Differential values for various oxygenation methods (if indicated) mechanical ventilation Other treatment Antibiotic therapy (n = 236) Use of corticosteroid (n = 89) Use of immunoglobulin (n = 87) Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity underweight is defined as BMI ≤ 18.5 kg/m2, overweight is defined as BMI ≥ 24 kg/m2, and obesity is defined as BMI ≥ 28 kg/m2. The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  24 Measure of frequency absolute number Frequency value 139 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) severe COVID Outcome (prognostic factor) severe COVID (BMI ≥ 24, < 28) Outcome severe COVID Prognostic factor (category):  BMI ≥ 24, < 28 Follow‐up Number of patients followed completely for this outcome  455 Number of obese patients followed completely for this outcome  187 Number of non‐obese patients followed completely for this outcome  268 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.83 (0.92, 3.64) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for age, sex, exposure to Wuhan, any coexisting medical condition, highest temperature, LDH, and C‐reactive protein Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.11 (0.47, 2.63) Outcome (prognostic factor) severe COVID (BMI ≥ 28) Outcome severe COVID Prognostic factor (category):  BMI ≥ 28 Follow‐up Number of patients followed completely for this outcome  455 Number of obese patients followed completely for this outcome  187 Number of non‐obese patients followed completely for this outcome  268 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 4.37 (1.96, 9.75) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for age, sex, exposure to Wuhan, any coexisting medical condition, highest temperature, LDH, and C‐reactive protein Effect measure for obesity odds ratio  Effect measure value (95% CI) 3.8 (1.32, 10.93)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Severe COVID	No	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Calmes 2021.

Study characteristics
Notes	English title Asthma and COPD are not risk factors for ICU stay and death in case of SARS‐CoV2 infection Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Registry data Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 596 Sampling method: Consecutive participants Participants Female participants (absolute number): 302 Age measure, value: NR Inclusion criteria: adult patients who were hospitalised in University Hospital of Liege between March 18 and April 17, 2020, for COVID‐19 Exclusion criteria: NR Smoking frequency: 50 Diabetes frequency: 124 Hypertension frequency: 246 Cardiovascular disease frequency: 116 Asthma frequency: 57 Chronic obstructive pulmonary disease frequency: 46 Other pulmonary disease frequency: emphysema (70), bronchiectasis (23) Immunosuppression frequency: 32 Chronic kidney disease frequency: 42 Cancer frequency: 70 Steroid administration frequency: inhaled corticosteroid (56), oral corticosteroid (23) Supplemental oxygen administration frequency: 41 Other treatments (frequency): hydroxychloroquine (596), doxycycline (596) Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: NR Obesity frequency (absolute number): 115 Prognostic factor(s): Obesity Outcome(s) ICU admission Death Outcome (prognostic factor) ICU admission (obesity) Follow‐up Number of patients followed completely for the outcome: 595 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 9 (4.5, 15), < 0.0001 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 8.5 (5.10, 14), < 0.0001 Outcome (prognostic factor) Death (obesity) Follow‐up Number of patients followed completely for the outcome: 595 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2 (1.2, 3.3), 0.0078 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Male gender, older age Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.80 (1.10, 3.20), 0.029
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	No	Appendix 3
Study Attrition ICU admission	No	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Cao 2021.

Study characteristics
Notes	English title Obesity and COVID‐19 in adult patients with diabetes Study setting Start of study recruitment (MM/YYYY) 02/2020 End of study recruitment (MM/YYYY) 03/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 1637 Sampling method consecutive participants Participants Female participants (absolute number), 823 Age measure, value median (interquartile range), 60 (50, 68) Inclusion criteria COVID patients aged > 18 years Exclusion criteria NR Smoking NR Diabetes (absolute number), 231 Hypertension (absolute number), 473 Cardiovascular diseases (absolute number), 106 Please indicate if additional information is available  Coronary artery disease: 90; congestive heart failure: 16 Asthma (unspecified), NR Chronic obstructive pulmonary disease (absolute number), 46 Other pulmonary diseases (absolute number), 355 Please indicate if additional information is available  Dyspnoea Immunosuppression (unspecified), NR Please indicate if additional information is available NR Chronic kidney disease (absolute number), 6 Cancer (absolute number), 29 Steroid administration (absolute number), 197 Supplemental oxygen (absolute number), 404 Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity BMI < 18.5 kg/m2; normal weight, 18.5–23.9 kg/m2; overweight, 24.0–27.9 kg/m2; and obesity > 28 kg/m2 The time when obesity has been measured some time after presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  28 Measure of frequency absolute number Frequency value 572 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) pneumonia Outcome (prognostic factor) pneumonia (other: BMI 24 to 27.9) Outcome pneumonia Prognostic factor (category):  other: BMI 24 to 27.9 Follow‐up Number of patients followed completely for this outcome  1637 Number of obese patients followed completely for this outcome  717 Number of non‐obese patients followed completely for this outcome  920 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.13 (1.32, 0.97) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age, sex, and comorbidities Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.14 (1.32, 0.98) Outcome (prognostic factor) pneumonia (other: BMI >= 28) Outcome pneumonia Prognostic factor (category):  other: BMI >= 28 Follow‐up Number of patients followed completely for this outcome  1637 Number of obese patients followed completely for this outcome  717 Number of non‐obese patients followed completely for this outcome  920 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.46 (1.89, 1.14) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age, sex, and comorbidities Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.47 (1.88, 1.15)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Pneumonia	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Pneumonia	Yes	Appendix 3
Confounding Bias Pneumonia	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Cariou 2020.

Study characteristics
Notes	English title Phenotypic characteristics and prognosis of inpatients with COVID‐19 and diabetes: the CORONADO study Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 53 Study setting: Inpatient Number of participants recruited: 1317 Sampling method: Consecutive participants Participants Female participants (absolute number): 462 Age measure, value: Mean (SD), 69.8 (13) Inclusion criteria: 1. Hospitalisation in a dedicated COVID‐19 unit with COVID‐19 diagnosis confirmed biologically (by SARS‐CoV‐2 PCR test) and/or clinically/radiologically (i.e. as ground‐glass opacity and/or crazy paving on chest computed tomography [CT] scan), 2. Personal history of diabetes or newly diagnosed diabetes on admission (i.e. HbA1c ≥ 48 mmol/mol [6.5%] during hospitalisation) Exclusion criteria: NR Smoking frequency: 57 (from 1029) Diabetes frequency: 1205 (including type 1 diabetes) Hypertension frequency: 1003 (from 1299) Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 133 (from 1278) Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 60 (from 831) Cancer frequency: 194 (active cancer, from 1282) Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 25 Obesity frequency (absolute number): 838 Prognostic factor(s): Obesity Outcome(s) Tracheal intubation and/or death within 7 days of admission Death Outcome (prognostic factor) Tracheal intubation and/or death within 7 days of admission (obesity) Follow‐up Number of patients followed completely for the outcome: 1317 Number of obese patients followed completely for the outcome: 838 Number of non‐obese patients followed completely for the outcome: 279 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.25 (1.09, 1.42), NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.24 (1.06, 1.44), 0.0064 Other measures of precision: “Prior to admission” model, stepwise selection with age and sex forced: OR (95% CI) = 1.28 (1.10, 1.47), P value = 0.0010 Outcome (prognostic factor) Death (obesity) Follow‐up Number of patients followed completely for the outcome: 1317 Number of obese patients followed completely for the outcome: 838 Number of non‐obese patients followed completely for the outcome: 279 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: NR Effect measure value (95% CI), P value: NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Unclear	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Cavallaro 2020.

Study characteristics
Notes	English title Contrasting factors associated with COVID‐19‐related ICU and death outcomes: interpretable multivariable analyses of the UK CHESS dataset Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): 06/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Inpatient Number of participants recruited: 13,954 Sampling method: Consecutive participants Participants Female participants (absolute number): 5661 Age measure, value: Median (IQR), 70 (56‐81) Inclusion criteria: Confirmed COVID‐19 Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 2219 Hypertension frequency: 3768 Cardiovascular disease frequency: 2247 Asthma frequency: 1172 Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 1521 (chronic respiratory disease) Immunosuppression frequency: 377 Chronic kidney disease frequency: 1172 Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Clinical obesity The time when obesity has been measured: NR Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): 1479 Prognostic factor(s): Obesity (clinical) Outcome(s) Mortality ICU admission Outcome (prognostic factor) Mortality (obesity (clinical)) Follow‐up Number of patients followed completely for the outcome: 13,954 Number of obese patients followed completely for the outcome: 1480 Number of non‐obese patients followed completely for the outcome: 12,474 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: 37 pre‐existing conditions (including immunosuppression due to disease, asthma requiring medication, immunosuppression due to treatment, neurological conditions, respiratory conditions, obesity, type‐1 and type‐2 diabetes, hypertension, heart conditions, renal disease, liver diseases, and other comorbidities) and demographic factors Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.63 (1.01, 1.33), NR Outcome (prognostic factor) ICU admission (obesity (clinical)) Follow‐up Number of patients followed completely for the outcome: 13,954 Number of obese patients followed completely for the outcome: 1480 Number of non‐obese patients followed completely for the outcome: 12,474 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: 37 pre‐existing conditions (including immunosuppression due to disease, asthma requiring medication, immunosuppression due to treatment, neurological conditions, respiratory conditions, obesity, type‐1 and type‐2 diabetes, hypertension, heart conditions, renal disease, liver diseases, and other comorbidities) and demographic factors Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.371 (2.900, 3.920), NR
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Cedano 2021.

Study characteristics
Notes	English title Characteristics and outcomes of patients with COVID‐19 in an intensive care unit of a community hospital; retrospective cohort study Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 04/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 132 Sampling method consecutive participants Participants Female participants (absolute number), 54 Age measure, value median (interquartile range), 63 (53, 71) Inclusion criteria adult patients, admitted to the ICU, with severe COVID‐19 infection, between 3 March 2020 and 22 April 2020, with positive PCR for SARS‐COV2 Exclusion criteria Patients that required cardiopulmonary resuscitation on the medical floors but did not survive to be transferred to the ICU were excluded from the study. Smoking NR Diabetes (absolute number), 60 Hypertension (absolute number), 78 Cardiovascular diseases (absolute number), 15 Please indicate if additional information is available  Coronary artery disease Asthma (absolute number), 7 Chronic obstructive pulmonary disease (absolute number), 11 Other pulmonary diseases (unspecified), NR Please indicate if additional information is available  NR Immunosuppression (unspecified), NR Please indicate if additional information is available NR Chronic kidney disease (absolute number), 25 Cancer (absolute number), 8 Steroid administration (absolute number), 96 Supplemental oxygen (absolute number), 104 Differential values for various oxygenation methods (if indicated) Mechanical ventilation Other treatment Azithromycin (n = 99; 79%), hydroxychloroquine (n = 109; 82%), tocilizumab (n = 17; 19%) Dose if applicable NR Duration if applicable NR Percentage received this treatment Azithromycin (79%), hydroxychloroquine (82%), tocilizumab (19%) Prognostic factor(s) Study’s definition for obesity BMI >= 30 The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency absolute number Frequency value 59 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (BMI ≥ 30) Outcome mortality Prognostic factor (category):  BMI ≥ 30 Follow‐up Number of patients followed completely for this outcome  132 Number of obese patients followed completely for this outcome  59 Number of non‐obese patients followed completely for this outcome  69 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.51 (1.06, 6.14) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age, sex, comorbidities Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.92 (1.07, 8.01)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Chand 2020.

Study characteristics
Notes	English title COVID‐19‐associated critical illness—report of the first 300 patients admitted to intensive care units at a New York City medical center Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Prospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 9 ICUs within 3 hospitals Study setting: Inpatient Number of participants recruited: 300 Sampling method: Consecutive participants Participants Female participants (absolute number): 118 Age measure, value: Mean (SD), 58.2 (12.6) Inclusion criteria: age > 18, ICU admitted Exclusion criteria: Exclusion criteria were aged < 18 years or the absence of a Swedish personal identification number (PIN) Smoking frequency: 67 Diabetes frequency: 134 Hypertension frequency: 200 Cardiovascular disease frequency: 65 Asthma frequency: 39 Chronic obstructive pulmonary disease frequency: 17 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 39 Cancer frequency: 18 Steroid administration frequency: 167 Supplemental oxygen administration frequency: 274 (mechanical ventilation), 2 (ECMO), 131 (NM blockade), 174 (prone positioning) Other treatments (frequency): 233 (any vasopressor support), 226 (norepinephrine), 89 (phenylephrine), 104 (vasopressin), 25 (epinephrine), 28 (chloroquine), 279 (hydroxychloroquine) Prognostic factor(s) Study’s definition for obesity: BMI ≥ 25 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 25 Obesity frequency (absolute number): 257 Prognostic factor(s): Obesity Outcome(s) Mortality Outcome (prognostic factor) Mortality (obesity) Follow‐up Number of patients followed completely for the outcome: 300 Number of obese patients followed completely for the outcome: 163 Number of non‐obese patients followed completely for the outcome: 137 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.01 (1.00, 1.03), 0.02 Comment: For different obesity categories: obesity class 1: 1.35 (0.88, 2.06), obesity class 2 (BMI > 35): 1.54 (0.98, 2.43) Multivariable analysis for obesity Modelling method: Linear regression The set of prognostic factors used for adjustment: Age, AKI status, Covid‐19 symptoms, comorbidities, laboratory values, race, sex, smoking, total number of comorbidities Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.020 (1.010, 1.040), 0.004
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Chris 2020.

Study characteristics
Notes	English title Risk factors associated with critical COVID‐19 requiring mechanical ventilation Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 05/2020 Study design case series Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 990 Sampling method consecutive participants Participants Female participants (absolute number), 479 Age measure, value median (interquartile range), 68 (55, 82) Inclusion criteria NR Exclusion criteria NR Smoking NR Diabetes (absolute number), 279 Hypertension (absolute number), 482 Cardiovascular diseases (absolute number), 253 Please indicate if additional information is available  coronary artery disease (n = 133), congestive heart failure (n = 120) Asthma (absolute number), 78 Chronic obstructive pulmonary disease (absolute number), 119 Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (absolute number), 55 Please indicate if additional information is available steroid in last month Chronic kidney disease (absolute number), 126 Cancer (absolute number), 14 Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity BMI >= 30 The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency absolute number Frequency value 352 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mechanical ventilation Outcome (prognostic factor)  mechanical ventilation (BMI >= 30) Outcome mechanical ventilation Prognostic factor (category):  BMI >= 30 Follow‐up Number of patients followed completely for this outcome  990 Number of obese patients followed completely for this outcome  352 Number of non‐obese patients followed completely for this outcome  638 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age, sex Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.035 (1.011, 1.06)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Chua 2021.

Study characteristics
Notes	English title Prognostication in COVID‐19: a prospectively derived and externally validated risk prediction score for in‐hospital death Study setting Start of study recruitment (MM/YYYY) NR End of study recruitment (MM/YYYY) NR Study design prospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 983 Sampling method consecutive participants Participants Female participants (unspecified), NR Age measure, value median (interquartile range), 70 (53, 83) Inclusion criteria NR Exclusion criteria NR Smoking NR Diabetes (unspecified), NR Hypertension (unspecified), NR Cardiovascular diseases (unspecified), NR Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity NR The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency unspecified Frequency value NR How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (BMI > 30) Outcome mortality Prognostic factor (category):  BMI > 30 Follow‐up Number of patients followed completely for this outcome  983 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age, sex Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.39 (1.88, 3.03)
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Mortality	No	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Coss‐Rovirosa 2020.

Study characteristics
Notes	English title Are overweight and obesity risk factors for invasive mechanical ventilation in severe coronavirus disease 2019 pneumonia? Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 07/2020 Study design registry data Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 355 Sampling method consecutive participants Participants Female participants (absolute number), 120 Age measure, value mean (standard deviation), 53.31 (15.29) Inclusion criteria We included patients 18 years old or older who had a documented diagnosis of COVID‐19 (defined as a positive PCR for SARS‐CoV2 or a chest CT scan showing characteristics of COVID‐19 pneumonia) Exclusion criteria excluded patients with missing values Smoking NR Diabetes (absolute number), 61 Hypertension (absolute number), 100 Cardiovascular diseases (percentage), 0.024 Please indicate if additional information is available  NR Asthma (unspecified), NR Chronic obstructive pulmonary disease (percentage), 0.02 Other pulmonary diseases (unspecified), NR Please indicate if additional information is available  NR Immunosuppression (unspecified), NR Please indicate if additional information is available NR Chronic kidney disease (percentage), 2 Cancer (percentage), 14 Steroid administration (absolute number), 24 Supplemental oxygen (absolute number), 121 Differential values for various oxygenation methods (if indicated) Required mechanical ventilation Other treatment Lopinavir/ritonavir; azithromycin; hydroxychloroquine; glucocorticoids; tocilizumab Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity A normal BMI is between 18.5 and 24.9 kg/m2, an overweight BMI ranges from 25‐29.9 kg/m2, and obesity BMI is > 30 kg/m2 The time when obesity has been measured some time after presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency absolute number Frequency value 160 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mechanical ventilation Outcome (prognostic factor) mechanical ventilation (overweight: BMI ranges from 25‐29.9 kg/m2) Outcome mechanical ventilation Prognostic factor (category):  overweight: BMI ranges from 25‐29.9 kg/m2 Follow‐up Number of patients followed completely for this outcome  355 Number of obese patients followed completely for this outcome  274 Number of non‐obese patients followed completely for this outcome  82 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.47 (0.82, 2.6) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age‐ and sex‐adjusted, also adjusted for possible confounders, such as  C‐reactive protein levels, oxygen‐saturation levels, and mean arterial pressure on admission. Effect measure for obesity odds ratio  Effect measure value (95% CI) 0.67 (0.29, 1.53) Outcome (prognostic factor) mechanical ventilation (obesity BMI is > 30 kg/m2) Outcome mechanical ventilation Prognostic factor (category):  obesity BMI is > 30 kg/m2 Follow‐up Number of patients followed completely for this outcome  355 Number of obese patients followed completely for this outcome  274 Number of non‐obese patients followed completely for this outcome  82 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.7 (0.9, 3.1) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age‐ and sex‐adjusted, also adjusted for possible confounders, such as C‐reactive protein levels, oxygen‐saturation levels, and mean arterial pressure on admission Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.82 (0.94, 3.53) Outcome (prognostic factor) mechanical ventilation (obesity as a BMI over 35 kg/m2) Outcome mechanical ventilation Prognostic factor (category):  obesity as a BMI over 35 kg/m2 Follow‐up Number of patients followed completely for this outcome  355 Number of obese patients followed completely for this outcome  37 Number of non‐obese patients followed completely for this outcome  318 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.55 (0.77, 3.08) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age‐ and sex‐adjusted, also adjusted for possible confounders, such as C‐reactive protein levels, oxygen‐saturation levels, and mean arterial pressure on admission Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.86 (1.09, 7.5)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Cummins 2021.

Study characteristics
Notes	English title Factors associated with COVID‐19 related hospitalisation, critical care admission and mortality using linked primary and secondary care data Study setting Start of study recruitment (MM/YYYY) 02/2020 End of study recruitment (MM/YYYY) 06/2020 Study design registry data Study centre(s) unspecified Number of centres/clinics/areas NR Study setting outpatient and inpatient Number of participants recruited 1781 Sampling method consecutive participants Participants Female participants (absolute number), 797 Age measure, value not reported Inclusion criteria aged 16 or older with confirmed COVID‐19 infection between 01/02/2020 and 30/06/2020 Exclusion criteria NR Smoking NR Diabetes (absolute number), 641 Hypertension (absolute number), 825 Cardiovascular diseases (absolute number), 107 Please indicate if additional information is available  Atrial fibrillation Asthma (absolute number), 244 Chronic obstructive pulmonary disease (absolute number), 145 Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (absolute number), 365 Cancer (absolute number), 148 Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity NR The time when obesity has been measured unspecified Main variable used for determination of obesity NR Threshold used for definition of obesity  NR Measure of frequency absolute number Frequency value 482 How many eligible outcomes reported? 3 How many eligible outcomes reported? 3 Outcome(s) hospitalisation, ICU admission, mortality Outcome (prognostic factor) Hospitalisation (obese) Outcome Hospitalisation Prognostic factor (category):  obese Follow‐up Number of patients followed completely for this outcome  1781 Number of obese patients followed completely for this outcome  482 Number of non‐obese patients followed completely for this outcome  1299 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment gender, age, stratified into 16‐49, 50‐69 and 70+ years of age, asthma; atrial fibrillation; cancer; chronic heart disease (CHD); chronic kidney disease (CKD); chronic obstructive pulmonary disease (COPD); dementia; depression; diabetes (type 1 and type 2 diabetes); epilepsy; heart failure; hypertension; learning disability; severe mental illness; peripheral arterial disease (PAD); and stroke. Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.64 (1.25, 2.15) Outcome (prognostic factor) ICU admission (obese) Outcome ICU admission Prognostic factor (category):  obese Follow‐up Number of patients followed completely for this outcome  1781 Number of obese patients followed completely for this outcome  482 Number of non‐obese patients followed completely for this outcome  1299 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment gender, age, stratified into 16‐49, 50‐69 and 70+ years of age, asthma; atrial fibrillation; cancer; chronic heart disease (CHD); chronic kidney disease(CKD); chronic obstructive pulmonary disease (COPD); dementia; depression; diabetes (Type 1 and Type 2 diabetes); epilepsy; heart failure; hypertension; learning disability; severe mental illness; peripheral arterial disease (PAD); and stroke. Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.74 (1.18, 2.56) Outcome (prognostic factor) mortality (obese) Outcome mortality Prognostic factor (category):  obese Follow‐up Number of patients followed completely for this outcome  1781 Number of obese patients followed completely for this outcome  482 Number of non‐obese patients followed completely for this outcome  1299 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment gender, age, stratified into 16‐49, 50‐69 and 70+ years of age, asthma; atrial fibrillation; cancer; chronic heart disease (CHD); chronic kidney disease(CKD); chronic obstructive pulmonary disease (COPD); dementia; depression; diabetes (Type 1 and Type 2 diabetes); epilepsy; heart failure; hypertension; learning disability; severe mental illness; peripheral arterial disease (PAD); and stroke. Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.15 (0.86, 1.55)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Czernichow 2020.

Study characteristics
Notes	English title Obesity doubles mortality in patients hospitalized for (SARS‐CoV‐2) in Paris hospitals, France: a cohort study on 5,795 patients Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Prospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 39 Study setting: Outpatient and inpatient Number of participants recruited: 5795 Sampling method: Consecutive participants Participants Female participants (absolute number): 2004 Age measure, value: Mean (SD), 59.7 (13.73) Inclusion criteria: Aged 18 to 79 years, hospitalised Exclusion criteria: Subjects who objected to the reuse of their data Smoking frequency: 786 Diabetes frequency: 2473 Hypertension frequency: 3142 Cardiovascular disease frequency: 264 (only heart failure) Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 543 Cancer frequency: 656 Steroid administration frequency: NR Supplemental oxygen administration frequency: 1984 Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI > 30 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 1264 Prognostic factor(s): Obesity class 1 (30 < BMI < 35) Obesity class 2 (35 < BMI < 40) Obesity class 3 (BMI >40) Outcome(s) Mortality Outcome (prognostic factor) Mortality (Obesity class 1 (30 < BMI < 35)) Follow‐up Number of patients followed completely for the outcome: 5795 Number of obese patients followed completely for the outcome: 1264 Number of non‐obese patients followed completely for the outcome: 2792 Comment: 1739 patients had missing data for BMI Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, comorbidities, sex, smoking status Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.89 (1.45, 2.47), NR Outcome (prognostic factor) Mortality (Obesity class 2 (35 < BMI < 40)) Follow‐up Number of patients followed completely for the outcome: 5795 Number of obese patients followed completely for the outcome: 1264 Number of non‐obese patients followed completely for the outcome: 2792 Comment: 1739 patients had missing data for BMI Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, comorbidities, sex, smoking status Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.79 (1.95, 3.97), NR Outcome (prognostic factor) Mortality (obesity class 3 (BMI > 40)) Follow‐up Number of patients followed completely for the outcome: 5795 Number of obese patients followed completely for the outcome: 1264 Number of non‐obese patients followed completely for the outcome: 2792 Comment: 1739 patients had missing data for BMI Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, comorbidities, sex, smoking status Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.55 (1.62, 3.95), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

De Souza 2021.

Study characteristics
Notes	English title On the analysis of mortality risk factors for hospitalized COVID‐19 patients: a data‐driven study using the major Brazilian database Study setting Start of study recruitment (MM/YYYY) 02/2020 End of study recruitment (MM/YYYY) 08/2020 Study design retrospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting inpatient Number of participants recruited 162,045 Sampling method non‐random sample Participants Female participants (absolute number), 202, 333 Age measure, value not reported Inclusion criteria The criteria for hospitalisation according to the Ministry of Health concerned the individual presenting gripal syndrome along with dyspnoea/respiratory distress or persistent pressure in the chest or blood oxygen saturation < 95% in room air or blue lips/face. The gripal syndrome concerned the individual with an acute respiratory condition, characterised by at least two of the following signs and symptoms: fever (even if referred), chills, sore throat, headache, cough, runny nose, olfactory disorders or taste disorders. We used in our study data from 162,045 patients who had closed outcomes (cure or death) in order to provide a profile overview of the patients and after, a 44,128 patient cohort with full symptom/comorbidity information aiming to analyse risk factors for mortality. Exclusion criteria NR Smoking NR Diabetes (absolute number), 17,573 Hypertension (unspecified) Cardiovascular diseases (absolute number), 22,957 Please indicate if additional information is available  NR Asthma (absolute number), 2118 Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (absolute number), 2788 Please indicate if additional information is available  Pneumopathy: 2788 Immunosuppression (absolute number), 2343 Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity NR The time when obesity has been measured unspecified Main variable used for determination of obesity other (please specify) Threshold used for definition of obesity  NR Measure of frequency absolute number Frequency value 3633 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) Mortality (obesity) Outcome Mortality Prognostic factor (category):  Obesity Follow‐up Number of patients followed completely for this outcome  44,128 Number of obese patients followed completely for this outcome  3633 Number of non‐obese patients followed completely for this outcome  40,495 Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 0.88 (0.83, 0.93) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Male sex, age (40‐60, 60‐80, 80+), fever, cough, dyspnoea, respiratory distress, SPO2, diarrhoea, other symptoms, cardiac disease, liver disease, asthma, diabetes, neuropathy, pneumopathy, immunodepression, kidney disease, other comorbidity, flu vaccine, flu antiviral, ICU admission, invasive mechanical ventilation, non‐invasive ventilation Effect measure for obesity hazard ratio Effect measure value (95% CI) NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Deng 2020.

Study characteristics
Notes	English title Obesity as a potential predictor of disease severity in young COVID‐19 patients: a retrospective study Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): 03/2020 Study design: Retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 65 Sampling method: Consecutive participants Participants Female participants (absolute number): 29 Age measure, value: Mean (SD), 33.6 (5.76) Inclusion criteria: Confirmed COVID‐19 based on a positive RNA test for SARS‐CoV‐2 in a respiratory sample, age between 18 and 40 years, chest computed tomography (CT) scan data available, and weight and height had been recorded Exclusion criteria: Pregnancy Smoking frequency: 1 Diabetes frequency: 2 Hypertension frequency: 3 Cardiovascular disease frequency: 0 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: 1 Steroid administration frequency: 12 Supplemental oxygen administration frequency: 23 Other treatments (frequency): 65 (antiviral), 52 (antibacterial), 12 (immunoglobulin), 3 (albumin) Prognostic factor(s) Study’s definition for obesity: No exact definition was given in the study. According to the categorisation in the table, the obesity was probably defined as a BMI ≥ 28 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 28 Obesity frequency (absolute number): 10 Prognostic factor(s): Obesity Outcome(s) Severe COVID Outcome (prognostic factor) Severe COVID (obesity) Follow‐up Number of patients followed completely for the outcome: 65 Number of obese patients followed completely for the outcome: 10 Number of non‐obese patients followed completely for the outcome: 55 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Comment: This study was adjusted for 6 models; no information is available regarding each model and adjusted covariates. Model 1: 115.89 (18.96,+∞), P = 0.001, Model 2: 73.77 (11.79, +∞), P < 0.001, Model 3: 88.76 (13.31,+∞), P < 0.001, Model 4: 6.46 (0.58,+∞), P = 0.1, Model 5: 86 (13.56,+∞), P < 0.001, Model 6: 90.03 (12.71,+∞), P < 0.001
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Severe COVID	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Severe COVID	No	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Dennison 2021.

Study characteristics
Notes	English title Circulating activated neutrophils in COVID‐19: an independent predictor for mechanical ventilation and death Study setting Start of study recruitment (MM/YYYY) 05/2020 End of study recruitment (MM/YYYY) 08/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 331 Sampling method consecutive participants Participants Female participants (absolute number), 107 Age measure, value median (interquartile range), 53 (41, 65) Inclusion criteria All adult patients presenting to the emergency department of SQUH with symptoms consistent with COVID‐19 and confirmed for SARS‐CoV‐2 by RT‐PCR from May to August 2020 were included in the study. Exclusion criteria Patients who had haemoglobinopathies, haematologic or solid malignancy on chemotherapy were excluded. Patients were also excluded if a CBC was not done at the time of admission Smoking NR Diabetes (absolute number), 116 Hypertension (absolute number), 118 Cardiovascular diseases (absolute number), 30 Please indicate if additional information is available  only CAD Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity a body mass index of >= 30 (patients with missing weight and height values were labelled obese if treating physicians labelled them as obese before the outcome) The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency absolute number Frequency value 31 How many eligible outcomes reported? 2 How many eligible outcomes reported? 2 Outcome(s) mechanical ventilation, mortality Outcome (prognostic factor) Mechanical ventilation (BMI >= 30) Outcome Mechanical ventilation Prognostic factor (category):  BMI >= 30 Follow‐up Number of patients followed completely for this outcome  300 Number of obese patients followed completely for this outcome  31 Number of non‐obese patients followed completely for this outcome  269 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, DM, HTN, IG (immature granulocytes), NEUT‐RI (neutrophil reactivity intensity), WBC Effect measure for obesity odds ratio  Effect measure value (95% CI) 6.55 (NR) Outcome (prognostic factor) Mortality (BMI >= 30) Outcome Mortality Prognostic factor (category):  BMI >= 30 Follow‐up Number of patients followed completely for this outcome  274 Number of obese patients followed completely for this outcome  unspecified Number of non‐obese patients followed completely for this outcome  unspecified Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, DM, HTN, NEUT‐RI (neutrophil reactivity intensity) Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.02 (NR)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Confounding Bias Mechanical ventilation	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Denova‐Gutiérrez 2020.

Study characteristics
Notes	English title The association of obesity, type 2 diabetes, and hypertension with severe coronavirus disease 2019 on admission among Mexican patients Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Outpatient and inpatient Number of participants recruited: 3844 Sampling method: Consecutive participants Participants Female participants (absolute number): 1614 Age measure, value: Mean (SD), 45.40 (15.8) Inclusion criteria: Laboratory‐confirmed cases with complete information Exclusion criteria: NR Smoking frequency: 365 Diabetes frequency: 669 Hypertension frequency: 557 Cardiovascular disease frequency: 727 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: 38 Chronic kidney disease frequency: 108 Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): antiviral (738) Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: NR Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): 668 Prognostic factor(s): Obesity Outcome(s) Severe COVID Outcome (prognostic factor) Severe COVID (obesity) Follow‐up Number of patients followed completely for the outcome: 3844 Number of obese patients followed completely for the outcome: 668 Number of non‐obese patients followed completely for the outcome: 3176 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, cardiovascular disease, CKD, drug treatment, immunosuppression, place of care, sex, smoking status, USMER (health units that monitor respiratory diseases) Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.43 (1.11, 1.83), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Severe COVID	Unclear	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Severe COVID	Unclear	Appendix 3
Confounding Bias Severe COVID	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Eastment 2021.

Study characteristics
Notes	English title BMI and outcomes of SARS‐CoV‐2 among US veterans Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 06/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Outpatient and inpatient Number of participants recruited: 276,564 Sampling method: Consecutive participants Participants Female participants (absolute number): 30,145 Age measure, value: Mean (SD), 61.70 (15.6) Inclusion criteria: All VA patients, who were tested for SARS‐CoV‐2 nucleic acid by polymerase chain reaction (PCR) in the inpatient or outpatient setting between February 28, 2020, and June 21, 2020 Exclusion criteria: VA employees, BMI < 12 kg/m2 or > 100 kg/m2 (n = 129) and those who were missing information on BMI (n = 5289) Smoking frequency: 57,525 (including ex‐smokers) Diabetes frequency: 94,861 Hypertension frequency: 173,129 Cardiovascular disease frequency: 64,715 Asthma or chronic obstructive pulmonary disease frequency: 75,502 Other pulmonary disease frequency: obstructive sleep apnoea (84,905), obesity hypoventilation (1936) Immunosuppression frequency: NR Chronic kidney disease frequency: 45,633 Cancer frequency: 75,225 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI 30 to 34.9 kg/m2 (class 1 obesity), 35 to 39.9 kg/m2 (class 2 obesity), and ≥ 40 kg/m2 (class 3 obesity) The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 119,417 Prognostic factor(s): Class I obesity (30 < BMI < 34.9) Class II obesity (35 < BMI < 39.9) Class III obesity (BMI > 40) Outcome(s) Hospitalisation ICU admission Mechanical ventilation Mortality Outcome (prognostic factor) Hospitalisation (Class I obesity (30 < BMI < 34.9)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.75 (0.70, 0.81), NR Outcome (prognostic factor) Hospitalisation (Class II obesity (35 < BMI < 39.9)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.83 (0.75, 0.91), NR Outcome (prognostic factor) Hospitalisation (Class III obesity (BMI > 40)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.96 (0.86, 1.07), NR Outcome (prognostic factor) ICU admission (Class I obesity (30 < BMI < 34.9)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.92 (0.81, 1.03), NR Outcome (prognostic factor) ICU admission (Class II obesity (35 < BMI < 39.9)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.94 (0.81, 1.10), NR Outcome (prognostic factor) ICU admission (Class III obesity (BMI > 40)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.15 (0.96, 1.36), NR Outcome (prognostic factor) Mechanical ventilation (Class I obesity (30 < BMI < 34.9)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.15 (0.94, 1.41), NR Outcome (prognostic factor) Mechanical ventilation (Class II obesity (35 < BMI < 39.9)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.35 (1.06, 1.72), NR Outcome (prognostic factor) Mechanical ventilation (Class III obesity (BMI > 40)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.77 (1.35, 2.32), NR Outcome (prognostic factor) Mortality (Class I obesity (30 < BMI < 34.9)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.89 (0.76, 1.03), NR Outcome (prognostic factor) Mortality (Class II obesity (35 < BMI < 39.9)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.95 (0.78, 1.16), NR Outcome (prognostic factor) Mortality (Class III obesity (BMI > 40)) Follow‐up Number of patients followed completely for the outcome: 25,925 Number of obese patients followed completely for the outcome: 12,672 Number of non‐obese patients followed completely for the outcome: 13,253 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age (continuous), alcohol dependence, asthma or chronic obstructive pulmonary disease, cancer, cerebrovascular disease, chronic kidney disease, cirrhosis, congestive heart failure, coronary artery disease, diabetes, dialysis, ethnicity, geographic region (COVID‐19 burden in each patient’s state or territory of residence as of August 19, 2020), hyperlipidaemia, hypertension, obstructive sleep apnoea, obesity hypoventilation syndrome, race (black, white, other), substance use dependence, sex, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.42 (1.12, 1.78), NR
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Ebinger 2020.

Study characteristics
Notes	English title Pre‐existing traits associated with Covid‐19 illness severity Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): NR Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Outpatient and inpatient Number of participants recruited: 442 Sampling method: Consecutive participants Participants Female participants (absolute number): 186 Age measure, value: Mean (SD), 57.72 (19.65) Inclusion criteria: NR Exclusion criteria: NR Smoking frequency: 16 Diabetes frequency: 84 Hypertension frequency: 161 Cardiovascular disease frequency: 49 Asthma or Chronic obstructive pulmonary disease frequency: 70 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: NR Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): NR Prognostic factor(s): Obesity Outcome(s) Hospitalisation ICU admission Mechanical ventilation Severe COVID Outcome (prognostic factor) Hospitalisation (obesity) Follow‐up Number of patients followed completely for the outcome: 442 Number of obese patients followed completely for the outcome: 71 Number of non‐obese patients followed completely for the outcome: 371 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, ACEI use, ARB use, asthma or COPD, DM, Elixhauser comorbidity score, ethnicity (Hispanic), HTN, myocardial infarction or HF, race (African‐American), sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.99 (0.97, 4.08), 0.059 Comment: In this study, age‐ and sex‐adjusted multivariate analysis is also given: 2.04 (1.14, 3.65), P = 0.016 Outcome (prognostic factor) ICU admission (obesity) Follow‐up Number of patients followed completely for the outcome: 442 Number of obese patients followed completely for the outcome: 71 Number of non‐obese patients followed completely for the outcome: 371 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.26 (0.62, 2.57), 0.520 Outcome (prognostic factor) Mechanical ventilation (obesity) Follow‐up Number of patients followed completely for the outcome: 442 Number of obese patients followed completely for the outcome: 71 Number of non‐obese patients followed completely for the outcome: 371 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.57 (0.72, 3.41), 0.260 Outcome (prognostic factor) Severe COVID (obesity) Follow‐up Number of patients followed completely for the outcome: 442 Number of obese patients followed completely for the outcome: 71 Number of non‐obese patients followed completely for the outcome: 371 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, ACEI use, ARB use, asthma or COPD, DM, Elixhauser comorbidity score, ethnicity (Hispanic), HTN, myocardial infarction or HF, race (African‐American), sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.95 (1.11, 3.42), 0.021 Comment: In this study, age‐ and sex‐adjusted multivariate analysis is also given: 1.96 (1.19, 3.24), P = 0.009. Moreover, after adding smoking to multivariate analysis (non‐missing data on smoking): 1.48 (0.76, 2.9), P = 0.25
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Study Attrition Hospitalisation	Yes	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mechanical ventilation	Unclear	Appendix 3
Outcome Measurement ICU admission	Unclear	Appendix 3
Outcome Measurement Hospitalisation	Unclear	Appendix 3
Outcome Measurement Severe COVID	Unclear	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Escalera 2020.

Study characteristics
Notes	English title Risk factors for mortality in patients with Coronavirus Disease 2019 (COVID‐19) in Bolivia: an analysis of the first 107 confirmed cases Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 03/2020 Study design retrospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas unspecified Study setting outpatient and inpatient Number of participants recruited 107 Sampling method consecutive participants Participants Female participants (absolute number), 52 Age measure, value median (standard deviation), 43.9 (17.6) Inclusion criteria unspecified Exclusion criteria unspecified Smoking NR Diabetes (absolute number), 5 Hypertension (absolute number), 10 Cardiovascular diseases (absolute number), 2 Please indicate if additional information is available  CHF Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified), Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment unspecified Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity unspecified The time when obesity has been measured unspecified Main variable used for determination of obesity other (please specify) Threshold used for definition of obesity  unspecified Measure of frequency absolute number Frequency value 6 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (obesity) Outcome mortality Prognostic factor (category):  obesity Follow‐up Number of patients followed completely for this outcome  107 Number of obese patients followed completely for this outcome  6 Number of non‐obese patients followed completely for this outcome  101 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 12.125 (1.69, 86.948) Multivariable (adjusted) analysis for obesity Modelling method other (please specify) The set of prognostic factors used for adjustment unclear Effect measure for obesity odds ratio  Effect measure value (95% CI) unspecified
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Statistical Analysis Bias	No	Appendix 3

FAI2R/SFR/SNFMI/SOFREMIP/CRI/IMIDIATE 2020.

Study characteristics
Notes	English title Severity of COVID‐19 and survival in patients with rheumatic and inflammatory diseases: data from the French RMD COVID‐19 cohort of 694 patients Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Prospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Outpatient and inpatient Number of participants recruited: 694 Sampling method: NR Participants Female participants (absolute number): 462 Age measure, value: Mean (SD), 56.1 (16.4) Inclusion criteria: Patients of all ages with confirmed iRMD (rheumatic and inflammatory diseases) and highly suspected/confirmed diagnosis of COVID‐19 Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 62 Hypertension frequency: 182 Cardiovascular disease frequency: 85 Asthma frequency: 52 Chronic obstructive pulmonary disease frequency: 28 Other pulmonary disease frequency: interstitial lung disease (26)  Immunosuppression frequency: NR Chronic kidney disease frequency: 42 Cancer frequency: 33 Steroid administration frequency: 215 Supplemental oxygen administration frequency: NR Other treatments (frequency): hydroxychloroquine (40), azithromycin (26), lopinavir/ritonavir (21), darunavir/ritonavir (10), remdesivir (2), tocilizumab (3), anakinra (1), HCQ + AZI (24), HCQ + AZI + anakinra (1) Prognostic factor(s) Study’s definition for obesity: BMI > 30 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 146 Prognostic factor(s): BMI 30‐39.9 BMI ≥ 40 Outcome(s) Severe COVID Mortality Outcome (prognostic factor) Severe COVID (BMI 30‐39.9) Follow‐up Number of patients followed completely for the outcome: 694 Number of obese patients followed completely for the outcome: 126 Number of non‐obese patients followed completely for the outcome: 459 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.25 (2.25, 0.69), 0.46 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.47 (0.76, 2.82), 0.25 Outcome (prognostic factor) Severe COVID (BMI ≥ 40) Follow‐up Number of patients followed completely for the outcome: 694 Number of obese patients followed completely for the outcome: 20 Number of non‐obese patients followed completely for the outcome: 459 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.43 (1.26, 9.32), 0.016 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 4.10 (1.28, 13.11), 0.017 Outcome (prognostic factor) Mortality (BMI 30‐39.9) Follow‐up Number of patients followed completely for the outcome: 675 Number of obese patients followed completely for the outcome: 121 Number of non‐obese patients followed completely for the outcome: 452 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.56 (0.78, 2.97), 0.19 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.95 (0.88, 4.18), 0.093 Outcome (prognostic factor) Mortality (BMI ≥ 40) Follow‐up Number of patients followed completely for the outcome: 675 Number of obese patients followed completely for the outcome: 19 Number of non‐obese patients followed completely for the outcome: 452 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.64 (1.07, 10.29), 0.026 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.77 (0.86, 15.09), 0.07
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Severe COVID	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Severe COVID	Unclear	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Farrell 2020.

Study characteristics
Notes	English title Sociodemographic variables as predictors of adverse outcome in SARS‐CoV‐2 infection: an Irish hospital experience Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 05/2020 Study design: Prospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 257 Sampling method: Consecutive participants Participants Female participants (absolute number): 104 Age measure, value: Mean (SD), 60.1 (18.4) Inclusion criteria: NR Exclusion criteria: NR Smoking frequency: 29 Diabetes frequency: NR Hypertension frequency: NR Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Overweight (BMI 25–30) or obese (BMI > 30) The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 25 Obesity frequency (absolute number): 166 Prognostic factor(s): Overweight or obese (BMI > 25) Outcome(s) Mortality ICU admission Outcome (prognostic factor) Mortality (overweight or obese (BMI > 25)) Follow‐up Number of patients followed completely for the outcome: 257 Number of obese patients followed completely for the outcome: 166 Number of non‐obese patients followed completely for the outcome: 91 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, Charlson Comorbidity Index Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 2.20 (0.88, 5.52), 0.093 Outcome (prognostic factor) ICU admission (overweight or obese (BMI > 25)) Follow‐up Number of patients followed completely for the outcome: 257 Number of obese patients followed completely for the outcome: 166 Number of non‐obese patients followed completely for the outcome: 91 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 2.37 (1.37, 6.83), 0.01
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Confounding Bias ICU admission	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Filardo 2020.

Study characteristics
Notes	English title Comorbidity and clinical factors associated with COVID‐19 critical illness and mortality at a large public hospital in New York City in the early phase of the pandemic (March‐April 2020) Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): NR Number of centres, clinics or areas: NR Study setting: Inpatient Number of participants recruited: 337 Sampling method: Consecutive participants Participants Female participants (absolute number): 108 Age measure, value: Mean (SD), 57.4 (12.64) Inclusion criteria: Patients aged 18 and older admitted to BHC with laboratory‐confirmed COVID‐19 between March 9th 2020 and April 8th 2020 Exclusion criteria: Incarcerated individuals Smoking frequency: 65 Diabetes frequency: 109 Hypertension or cardiovascular disease frequency: 175 Chronic pulmonary disease frequency: 43 Immunosuppression frequency: 14 Chronic kidney disease frequency: 27 Cancer frequency: 10 Steroid administration frequency: 102 Supplemental oxygen administration frequency: 272 Other treatments (frequency): lopinavir/ritonavir (31), HCQ (44), HCQ + azithromycin (200), tocilizumab (29), remdesivir study enrolment (receipt of remdesivir or placebo is unknown for these patients) (4), antimicrobials (206) Prognostic factor(s) Study’s definition for obesity: BMI ≥ 30 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 130 Prognostic factor(s): Obesity (BMI ≥ 30) Outcome(s) Mortality Outcome (prognostic factor) Mortality (obesity (BMI ≥ 30)) Follow‐up Number of patients followed completely for the outcome: 270 Number of obese patients followed completely for the outcome: 109 Number of non‐obese patients followed completely for the outcome: 161 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.19 (0.82, 1.74), < 0.05 Multivariable analysis for obesity Modelling method: Linear regression The set of prognostic factors used for adjustment: Age, cardiovascular comorbidity, dementia, diabetes, HIV, immunosuppression, malignancy, race, renal comorbidity, pulmonary comorbidity, sex Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.37 (1.07, 1.74), < 0.05
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Forest 2021.

Study characteristics
Notes	English title De novo renal failure and clinical outcomes of patients with critical coronavirus disease 2019 Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 03/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting inpatient Number of participants recruited 330 Sampling method consecutive participants Participants Female participants (absolute number), 130 Age measure, value not reported Inclusion criteria >= 18 years old, RT‐PCR+, endotracheal intubation and mechanical ventilation upon initial presentation or during inpatient hospitalisation, history of ESRD Exclusion criteria Patients who expired in the emergency department without planned hospital admission Smoking NR Diabetes (absolute number), 194 Hypertension (absolute number), 224 Cardiovascular diseases (absolute number), 55 Please indicate if additional information is available  CAD 55, CHF 55 Asthma (absolute number), 83 Chronic obstructive pulmonary disease (absolute number), 83 Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available HIV 10 Chronic kidney disease (absolute number), 81 Cancer (unspecified) Steroid administration (unspecified), Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity BMI greater than or equal to 30 The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency unspecified Frequency value NR How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (BMI >= 30) Outcome mortality Prognostic factor (category):  BMI >= 30 Follow‐up Number of patients followed completely for this outcome  330 Number of obese patients followed completely for this outcome  unspecified Number of non‐obese patients followed completely for this outcome  unspecified Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age (60 years old cut‐point), CVD, CKD, COPD or asthma, DM, HTN, race, renal replacement therapy, sex Effect measure for obesity NR Effect measure value (95% CI) 2.138 (1.039, 4.399)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Fresán 2021.

Study characteristics
Notes	English title Independent role of severe obesity as a risk factor for COVID‐19 hospitalization: a Spanish population‐based cohort study Study setting Start of study recruitment (MM/YYYY): 03/2020 (cohorts 1‐4), 02/2020 (cohorts 5‐8) End of study recruitment (MM/YYYY): 04/2020 Study design: Prospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Outpatient and inpatient Number of participants recruited: 216,054 (cohort 1), 132,126 (cohort 2), 85,815 (cohort 3), 433,995 (cohort 4), 650,000 (cohorts 5‐8) Sampling method: Consecutive participants Participants Female participants (absolute number): NR (cohorts 1‐3), 217,346 (cohort 4), 325,520 (cohorts 5‐8) Age measure, value: NR Inclusion criteria: Population aged 25 to 79 years and covered by the Health Service Exclusion criteria: Age under 25 or 80 and over, healthcare professionals, nursing home residents, not covered by Navarra health service, terminally ill patients Smoking frequency: NR (cohorts 1‐3), 1552 (cohort 4), 148,850 (cohorts 5‐8) Diabetes frequency: NR Hypertension frequency: NR (cohorts 1‐3), 71,888 (cohort 4), 107,640 (cohorts 5‐8) Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Class 3 obesity, defined as BMI ≥ 40 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 40 Obesity frequency (absolute number): NR (cohorts 1‐3), 7460 (cohort 4), 11,115 (cohorts 5‐8) Prognostic factor(s): BMI > 40 (obesity class 3) Outcome(s) Hospitalisation Severity Outcome (prognostic factor) Hospitalisation (BMI > 40 (obesity class 3)) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 216,054 Number of obese patients followed completely for the outcome: 2834 Number of non‐obese patients followed completely for the outcome: 213,220 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 7.01 (4.57, 10.47), < 0.001 Multivariable analysis for obesity Modelling method: Poisson regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 5.02 (3.19, 7.90), < 0.001 Outcome (prognostic factor) Hospitalisation (BMI > 40 (obesity class 3)) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 132,126 Number of obese patients followed completely for the outcome: 2661 Number of non‐obese patients followed completely for the outcome: 129,465 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 2.15 (1.30, 3.55), 0.003 Multivariable analysis for obesity Modelling method: Poisson regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.87 (1.12, 3.12), 0.017 Outcome (prognostic factor) Hospitalisation (BMI > 40 (obesity class 3)) (cohort 3) Follow‐up Number of patients followed completely for the outcome: 85,815 Number of obese patients followed completely for the outcome: 1965 Number of non‐obese patients followed completely for the outcome: 83,850 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.25 (0.72, 2.17), 0.429 Multivariable analysis for obesity Modelling method: Poisson regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.22 (0.70, 2.12), 0.488 Outcome (prognostic factor) Hospitalisation (BMI > 40 (obesity class 3)) (cohort 4) Follow‐up Number of patients followed completely for the outcome: 433,995 Number of obese patients followed completely for the outcome: 7360 Number of non‐obese patients followed completely for the outcome: 426,535 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 2.82 (2.14, 3.73), < 0.001 Multivariable analysis for obesity Modelling method: Poisson regression The set of prognostic factors used for adjustment: Health‐related characteristics: primary healthcare visits in prior 12 months, hospitalisation in prior 12 months, smoking status, hypertension, and major chronic conditions Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 2.20 (1.66, 2.93), < 0.001 Outcome (prognostic factor) Severity (BMI > 40 (obesity class 3)) (cohort 5) Follow‐up Number of patients followed completely for the outcome: 439,490 Number of obese patients followed completely for the outcome: 7460 Number of non‐obese patients followed completely for the outcome: 426,535 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 3.44 (1.82, 6.52), < 0.001 Multivariable analysis for obesity Modelling method: Poisson regression The set of prognostic factors used for adjustment: Sociodemographic characteristics (sex, age, country of origin, municipality size, and annual taxable income level), health‐related characteristics (primary healthcare visits in prior 12 months, hospitalisation in prior 12 months, smoking status, hypertension, and major chronic conditions) Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 2.30 (1.20, 4.40), < 0.001 Outcome (prognostic factor) Severity (BMI > 40 (obesity class 3)) (cohort 6) Follow‐up Number of patients followed completely for the outcome: 216,056 Number of obese patients followed completely for the outcome: 2834 Number of non‐obese patients followed completely for the outcome: 213,220 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 32.24 (8.34, 124.69), < 0.001 Multivariable analysis for obesity Modelling method: Poisson regression The set of prognostic factors used for adjustment: Sociodemographic characteristics (sex, age, country of origin, municipality size, and annual taxable income level), health‐related characteristics (primary healthcare visits in prior 12 months, hospitalisation in prior 12 months, smoking status, hypertension, and major chronic conditions) Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 13.80 (3.11,61.17), < 0.001 Outcome (prognostic factor) Severity (BMI > 40 (obesity class 3)) (cohort 7) Follow‐up Number of patients followed completely for the outcome: 132,126 Number of obese patients followed completely for the outcome: 2661 Number of non‐obese patients followed completely for the outcome: 129,465 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 3.04 (0.95, 9.76), 0.62 Multivariable analysis for obesity Modelling method: Poisson regression The set of prognostic factors used for adjustment: Sociodemographic characteristics (sex, age, country of origin, municipality size, and annual taxable income level), health‐related characteristics (primary healthcare visits in prior 12 months, hospitalisation in prior 12 months, smoking status, hypertension, and major chronic conditions) Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 2.07 (0.62, 6.85), < 0.001 Outcome (prognostic factor) Severity (BMI > 40 (obesity class 3)) (cohort 8) Follow‐up Number of patients followed completely for the outcome: 85,815 Number of obese patients followed completely for the outcome: 1965 Number of non‐obese patients followed completely for the outcome: 53,850 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.54 (0.57, 4.17), 0.398 Multivariable analysis for obesity Modelling method: Poisson regression The set of prognostic factors used for adjustment: Sociodemographic characteristics (sex, age, country of origin, municipality size, and annual taxable income level), health‐related characteristics (primary healthcare visits in prior 12 months, hospitalisation in prior 12 months, smoking status, hypertension, and major chronic conditions) Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.42 (0.52, 3.88), 0.496
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Hospitalisation	Yes	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Outcome Measurement Severe COVID	No	Appendix 3
Confounding Bias Hospitalisation	Unclear	Appendix 3
Confounding Bias Severe COVID	Unclear	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Fried 2021.

Study characteristics
Notes	English title Patient characteristics and outcomes of 11 721 patients with coronavirus disease 2019 (COVID‐19) hospitalized across the United States Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres/clinics/areas: 245 Study setting: Inpatient Number of participants recruited: 11,721 Sampling method: Consecutive participants Participants Female participants (absolute number): 5457 Age measure, value: NR Inclusion criteria: Patients aged ≥ 18 years indicating COVID‐19 with ICD‐10 code  Exclusion criteria: NR Smoking frequency: 1922 Diabetes frequency: 3254 Hypertension frequency: 5475 Cardiovascular disease frequency: 2182 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 1737 Immunosuppression frequency: NR Chronic kidney disease frequency: 1854 Cancer frequency: 2390 Steroid administration frequency: NR Supplemental oxygen administration frequency: 6896 Other treatments (frequency): Remdesivir (0.4%) Prognostic factor(s) Study’s definition for obesity: Obesity (BMI ≥ 30 kg/m2) The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition of obesity: 30 Obesity frequency (absolute number): 1891 Prognostic factor(s): Obesity (BMI ≥ 30 kg/m2) Outcome(s) Mechanical ventilation, mortality  Outcome (prognostic factor) Mechanical ventilation (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 11,721 Number of obese patients followed completely for the outcome: 1891 Number of non‐obese patients followed completely for the outcome: 9830 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, insurance status at admission, history of chronic kidney disease, stage 5 kidney disease, hypertension, diabetes, pulmonary disease, cardiovascular disease, liver disease, obesity, and smoking Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 1.47 (1.28, 1.69), NR Outcome (prognostic factor) Mortality (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 11,721 Number of obese patients followed completely for the outcome: 1891 Number of non‐obese patients followed completely for the outcome: 9830 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, insurance status at admission, history of chronic kidney disease, stage 5 kidney disease, hypertension, diabetes, pulmonary disease, cardiovascular disease, liver disease, obesity, and smoking Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 1.07 (0.93, 1.24), NR
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Gao 2020.

Study characteristics
Notes	English title Obesity is a risk factor for greater COVID‐19 severity Study setting Start of study recruitment (MM/YYYY): 01/2020 End of study recruitment (MM/YYYY): 02/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 3 Study setting: Inpatient Number of participants recruited: 150 Sampling method: Consecutive participants Participants Female participants (absolute number): 56 Age measure, value: Mean (SD), 48 (NR) Inclusion criteria: Adults with BMI > 25 Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 29 Hypertension frequency: NR Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 0 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: 0 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI > 25 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 25 Obesity frequency (absolute number): 75 Prognostic factor(s): Obesity Outcome(s) Severity Outcome (prognostic factor) Severity (obesity) Follow‐up Number of patients followed completely for the outcome: 150 Number of obese patients followed completely for the outcome: 75 Number of non‐obese patients followed completely for the outcome: 75 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.91 (1.31, 6.47), 0.007 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, smoking status, hypertension, diabetes, and dyslipidaemia, age, smoking status Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.00 (1.22, 7.38), NR Comment: Adjusted for each unit of BMI as a continuous outcome: 1.13 (1.01, 1.28)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Severe COVID	No	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Gao 2021.

Study characteristics
Notes	English title Associations between body‐mass index and COVID‐19 severity in 6·9 million people in England: a prospective, community‐based, cohort study Study setting Start of study recruitment (MM/YYYY) 01/2020 End of study recruitment (MM/YYYY) 04/2020 Study design prospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting outpatient Number of participants recruited 6,910,695 Sampling method unspecified Participants Female participants (percentage), 53.1 Age measure, value (reported in categories) 20 to 39 years: 2,384,223 (34.5%) 40 to 59 years: 2,444,011 (35.4%) 60 to 79 years: 1,652,615 (23.9%) >= 80 years: 429,846 (6.2%) Inclusion criteria aged 20–99 years who were registered at a general practice (GP) that contributes to the QResearch database and had available BMI data. Exclusion criteria Participants without at least one BMI measurement Smoking (percentage), 17.1 Diabetes (percentage), 8.4 Hypertension (percentage), 19.7 Cardiovascular diseases (percentage), 6 Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (percentage), 16.1 Please indicate if additional information is available  COPD and asthma Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (percentage), 4.7 Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Obesity l: BMI = 30‐34.9, obesity ll/lll = +35 The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  BMI ≥ 30 Measure of frequency absolute number Frequency value 1,681,112 How many eligible outcomes reported? 3 How many eligible outcomes reported? 3 Outcome(s) hospitalisation, ICU admission, mortality Outcome (prognostic factor) hospitalisation (BMI) Outcome hospitalisation Prognostic factor (category):  BMI Follow‐up Number of patients followed completely for this outcome  6,910,695 Number of obese patients followed completely for this outcome  1,681,112 Number of non‐obese patients followed completely for this outcome  5,229,583 Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.05 (1.05, 1.06) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, sex, ethnicity, economic status, geographical region, smoking status, non‐obesity‐related morbidity, including conditions related to severe COVID‐19 disease (namely, chronic obstructive pulmonary disease, asthma, autoimmune diseases [systemic lupus erythematosus, rheumatoid diseases], ulcerative colitis or Crohn’s disease, type 1 diabetes, chronic liver disease, chronic renal disease, chronic neurological disease, and cerebral palsy); obesity‐related morbidity, including hypertension, cardiovascular disease (including congestive heart failure and stroke), reflux disease or gastro‐oesophageal reflux disease, and sleep apnoea; and type 2 diabetes Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.04 (1.04, 1.05) Outcome (prognostic factor) ICU admission (BMI) Outcome ICU admission Prognostic factor (category):  BMI Follow‐up Number of  patients followed completely for this outcome  6,910,695 Number of obese patients followed completely for this outcome  1,681,112 Number of non‐obese patients followed completely for this outcome  5,229,583 Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.1 (1.09, 1.11) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, sex, ethnicity, economic status, geographical region, smoking status, non‐obesity‐related morbidity, including conditions related to severe COVID‐19 disease (namely, chronic obstructive pulmonary disease, asthma, autoimmune diseases [systemic lupus erythematosus, rheumatoid diseases], ulcerative colitis or Crohn’s disease, type 1 diabetes, chronic liver disease, chronic renal disease, chronic neurological disease, and cerebral palsy); obesity‐related morbidity, including hypertension, cardiovascular disease (including congestive heart failure and stroke), reflux disease or gastro‐oesophageal reflux disease, and sleep apnoea; and type 2 diabetes Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.09 (1.08, 1.1) Outcome (prognostic factor) mortality (BMI) Outcome mortality Prognostic factor (category):  BMI Follow‐up Number of patients followed completely for this outcome  6,910,695 Number of obese patients followed completely for this outcome  1,681,112 Number of non‐obese patients followed completely for this outcome  5,229,583 Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.03 (1.03, 1.02) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, sex, ethnicity, economic status, geographical region, smoking status, non‐obesity‐related morbidity, including conditions related to severe COVID‐19 disease (namely, chronic obstructive pulmonary disease, asthma, autoimmune diseases [systemic lupus erythematosus, rheumatoid diseases], ulcerative colitis or Crohn’s disease, type 1 diabetes, chronic liver disease, chronic renal disease, chronic neurological disease, and cerebral palsy); obesity‐related morbidity, including hypertension, cardiovascular disease (including congestive heart failure and stroke), reflux disease or gastro‐oesophageal reflux disease, and sleep apnoea; and type 2 diabetes Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.04 (1.04, 1.05)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Study Attrition Hospitalisation	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Garcia Moreno 2021.

Study characteristics
Notes	English title Analysis of factors related to the clinical course of COVID‐19 infection in patients with hypertension Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 03/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 571 Sampling method consecutive participants Participants Female participants (absolute number), 233 Age measure, value median (interquartile range), 76 (66, 83) Inclusion criteria diagnosis of hypertension, hospital admission for COVID‐19 between 1 March and 24 March 2020 Exclusion criteria unspecified Smoking NR Diabetes (unspecified) Hypertension (absolute number) 571 Cardiovascular diseases (unspecified) Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (percentage), 91.9 Differential values for various oxygenation methods (if indicated) oxygen therapy Other treatment unspecified Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity unspecified The time when obesity has been measured unspecified Main variable used for determination of obesity other (please specify) Threshold used for definition of obesity  unspecified Measure of frequency unspecified Frequency value unspecified How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (obesity) Outcome mortality Prognostic factor (category):  obesity Follow‐up Number of  patients followed completely for this outcome  571 Number of obese patients followed completely for this outcome  unspecified Number of non‐obese patients followed completely for this outcome  unspecified Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment unspecified Effect measure for obesity odds ratio  Effect measure value (95% CI) unspecified
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Unclear	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Garcia Olivares 2020.

Study characteristics
Notes	English title The age and comorbidities, are independent risk factors for mortality in critically ill COVID‐19 patients? Study setting Start of study recruitment (MM/YYYY) unspecified End of study recruitment (MM/YYYY) unspecified Study design prospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 150 Sampling method consecutive participants Participants Female participants (percentage), 29 Age measure, value mean (standard deviation), 58 (14) Inclusion criteria unspecified Exclusion criteria unspecified Smoking NR Diabetes (unspecified) Hypertension (percentage), 53 Cardiovascular diseases (unspecified) Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified), Supplemental oxygen (percentage), 88 Differential values for various oxygenation methods (if indicated) 88% mechanical ventilation, 74% prone position Other treatment percentage Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity unspecified The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency percentage Frequency value 53 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (obesity) Outcome mortality Prognostic factor (category):  obesity Follow‐up Number of patients followed completely for this outcome  150 Number of obese patients followed completely for this outcome  79 Number of non‐obese patients followed completely for this outcome  71 Univariable (unadjusted) analysis for obesity Effect measure for obesity relative risk Effect measure value (95% CI) 1.17 (0.6, 2.25) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment severity (APACHE II, SOFA and mechanical ventilation) Effect measure for obesity NR Effect measure value (95% CI) unspecified
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Gaur 2021.

Study characteristics
Notes	English title Macrolevel association of COVID‐19 with non‐communicable disease risk factors in India Study setting Start of study recruitment (MM/YYYY) 02/2020 End of study recruitment (MM/YYYY) 11/2020 Study design registry data Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas unspecified Study setting unspecified Number of participants recruited unspecified Sampling method other Participants Female participants (unspecified) Age measure, value not reported Inclusion criteria unspecified Exclusion criteria unspecified Smoking NR Diabetes (unspecified), the frequency of diabetes was reported for each province of India separately.  Hypertension (unspecified), the frequency of HTN was reported for each province of India separately.  Cardiovascular diseases (unspecified) Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity unspecified The time when obesity has been measured unspecified Main variable used for determination of obesity other (please specify) Threshold used for definition of obesity  unspecified Measure of frequency unspecified Frequency value the frequency of obesity was reported for each province of India separately.  How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (obesity not specified) Outcome mortality Prognostic factor (category):  Obesity not specified Follow‐up Number of patients followed completely for this outcome  unspecified Number of obese patients followed completely for this outcome  unspecified Number of non‐obese patients followed completely for this outcome  unspecified Univariable (unadjusted) analysis for obesity Effect measure for obesity slope (beta) Effect measure value (95% CI) 0.52 NR Multivariable (adjusted) analysis for obesity Modelling method other (please specify) The set of prognostic factors used for adjustment obesity, HTN, diabetes, literacy, smoking Effect measure for obesity slope (beta) Effect measure value (95% CI) ‐0.13 (unspecified, unspecified)
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	No	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Gayam 2020.

Study characteristics
Notes	English title Clinical characteristics and predictors of mortality in African‐Americans with COVID‐19 from an inner‐city community teaching hospital in New York Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 408 Sampling method: Consecutive participants Participants Female participants (absolute number): 177 Age measure, value: Median (IQR), 67 (56, 76) Inclusion criteria: African‐American inpatients with positive COVID‐19 PCR Exclusion criteria: NR Smoking frequency: 36 Diabetes frequency: 176 Hypertension frequency: 271 Cardiovascular disease frequency: 54 Asthma frequency: 54 Chronic obstructive pulmonary disease frequency: 43 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 69 Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI continuous  The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: NA Obesity frequency (absolute number): NA Prognostic factor(s): BMI continuous  Outcome(s) Mortality Outcome (prognostic factor) Mortality (BMI continuous) Follow‐up Number of patients followed completely for the outcome: 408 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR (NR), 0.002 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, BMI, C‐reactive protein, and D‐dimers, serum ferritin, other lab findings, shortness of breath, myalgia Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.07 (1.04, 1.11), < 0.001
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Giacomelli 2021.

Study characteristics
Notes	English title Impact of gender on patients hospitalized for SARS‐COV‐2 infection: a prospective observational study Study setting Start of study recruitment (MM/YYYY) 02/2020 End of study recruitment (MM/YYYY) 05/2020 Study design prospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 2 Study setting inpatient Number of participants recruited 520 Sampling method consecutive participants Participants Female participants (percentage), 33 Age measure, value median (interquartile range), 61(50, 72) Inclusion criteria COVID‐19 positive  Exclusion criteria NS Smoking NR Diabetes (absolute number), 61 Hypertension (unspecified) Cardiovascular diseases (absolute number), 254 Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (absolute number), 78 Please indicate if additional information is available  Respiratory disease  Immunosuppression (absolute number), 39 Please indicate if additional information is available Immune system disorders Chronic kidney disease (absolute number), 42 Cancer (absolute number), 50 Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity NR The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  NR Measure of frequency absolute number Frequency value 92 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) Mortality (obesity NS) Outcome Mortality  Prognostic factor (category):  Obesity NS  Follow‐up Number of patients followed completely for this outcome  520 Number of obese patients followed completely for this outcome  92 Number of non‐obese patients followed completely for this outcome  428 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.94 (1.14, 3.32) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment gender, age, obesity, CVD, cancer, flu vaccination, time from symptoms onset, critical disease at hospital admission, fever yes versus not, anaemia, INR > 1.3, D‐dimer ≥ 500 μg/L, CRP ≥ 50 mg/L, eGFR (MDRD) < 60 mL/min, LDH > 245 IU/L, CK > 185 IU/L Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.17 (1.1, 4.31)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Gil‐Rodrigo 2020.

Study characteristics
Notes	English title Analysis of clinical characteristics and outcomes in patients with COVID‐19 based on a series of 1000 patients treated in Spanish emergency departments Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 04/2020 Study design prospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 61 Study setting inpatient Number of participants recruited 1000 Sampling method random sample Participants Female participants (percentage), 43.8 Age measure, value median (range), 62.3 (17.8) Inclusion criteria confirmed or suspected COVID‐19 (PCR or clinical manifestations) Exclusion criteria unspecified Smoking NR Diabetes (percentage), 18.8 Hypertension (percentage), 44.6 Cardiovascular diseases (percentage), 7.7 Please indicate if additional information is available  for IHD Asthma (percentage), 7.3 Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (percentage), 4.9 Please indicate if additional information is available NR Chronic kidney disease (percentage), 7.3 Cancer (percentage), 9.7 Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity unspecified The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity other (please specify) Threshold used for definition of obesity  unspecified Measure of frequency percentage Frequency value 14.3 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) mortality (obesity unspecified) Outcome mortality Prognostic factor (category):  obesity unspecified Follow‐up Number of patients followed completely for this outcome  1000 Number of obese patients followed completely for this outcome  143 Number of non‐obese patients followed completely for this outcome  857 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.93 (1.32, 2.82) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, sex, HTN, DM, obesity, DLP, smoking, asthma, and other comorbidities Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.53 (1.47, 4.35) Outcome (prognostic factor) mortality (obesity class 1) Outcome mortality Prognostic factor (category):  obesity class 1 Follow‐up Number of patients followed completely for this outcome  2874 Number of obese patients followed completely for this outcome  812 Number of non‐obese patients followed completely for this outcome  2062 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment unspecified Effect measure for obesity NR Effect measure value (95% CI) 1.15 (0.93, 1.4) Outcome (prognostic factor) mortality (obesity class 2) Outcome mortality Prognostic factor (category):  obesity class 2 Follow‐up Number of patients followed completely for this outcome  2431 Number of obese patients followed completely for this outcome  369 Number of non‐obese patients followed completely for this outcome  2062 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment unspecified Effect measure for obesity NR Effect measure value (95% CI) 1.33 (1, 1.76) Outcome (prognostic factor) mortality (obesity class 3) Outcome mortality Prognostic factor (category):  obesity class 3 Follow‐up Number of patients followed completely for this outcome  2357 Number of obese patients followed completely for this outcome  295 Number of non‐obese patients followed completely for this outcome  2062 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment unspecified Effect measure for obesity NR Effect measure value (95% CI) 1.92 (1.4, 2.63)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Girardin 2021.

Study characteristics
Notes	English title Contribution of pulmonary diseases to COVID‐19 mortality in a diverse urban community of New York Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 05/2020 Study design registry data Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting inpatient Number of participants recruited 4446 Sampling method consecutive participants Participants Female participants (absolute number), 1766 Age measure, value mean (standard deviation), 62 (18.89) Inclusion criteria We included only patients with confirmed COVID‐19  who had been discharged alive or had died. Exclusion criteria omitting hospitalised patients with unknown state information Smoking NR Diabetes (absolute number), 1473 Hypertension (absolute number), 2390 Cardiovascular diseases (absolute number), 580 Please indicate if additional information is available  CAD Asthma (absolute number), 493 Chronic obstructive pulmonary disease (absolute number), 329 Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (absolute number), 472 Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment unspecified Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity NR The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  NR Measure of frequency absolute number Frequency value 1660 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) Mortality (obesity NS) Outcome Mortality  Prognostic factor (category):  Obesity NS  Follow‐up Number of patients followed completely for this outcome  4210 Number of obese patients followed completely for this outcome  1660 Number of non‐obese patients followed completely for this outcome  2550 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment Age, ethnicity, gender, income, smoking, obesity, COPD, asthma, sleep apnoea, HTN, hlp, DM, CAD, PAD, autoimmunity, cancer Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.19 (1.04, 1.37)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Goodman 2020.

Study characteristics
Notes	English title Impact of sex and metabolic comorbidities on COVID‐19 mortality risk across age groups: 66,646 inpatients across 613 U.S. hospitals Study setting Start of study recruitment (MM/YYYY): 04/2020 End of study recruitment (MM/YYYY): 07/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 613 Study setting: Inpatient Number of participants recruited: 66,646 Sampling method: Consecutive participants Participants Female participants (absolute number): 31,400 Age measure, value: Mean (SD), 62.8 (17.9) Inclusion criteria: Patients who their admission included an ICD‐10‐CM diagnosis code of 'COVID‐19' (U07.1) Exclusion criteria: Under 20‐year‐old patients Smoking frequency: NR Diabetes frequency: 25,611 Hypertension frequency: 42,813 Cardiovascular disease frequency: 9893 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: Chronic pulmonary disease (13,606) Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: 1795 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): 14,044 Prognostic factor(s): Obesity Outcome(s) Mortality Outcome (prognostic factor) Mortality (obesity) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 7371 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Log‐binomial models using the modified Poisson regression approach The set of prognostic factors used for adjustment: A list of underlying diseases, sex, race, admission month Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.92 (1.43, 2.57), < 0.001 Outcome (prognostic factor) Mortality (obesity) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 6947 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Log‐binomial models using the modified Poisson regression approach The set of prognostic factors used for adjustment: A list of underlying diseases, sex, race, admission month Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.57 (1.3, 2.9), < 0.001 Outcome (prognostic factor) Mortality (obesity) (cohort 3) Follow‐up Number of patients followed completely for the outcome: 11,138 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Log‐binomial models using the modified Poisson regression approach The set of prognostic factors used for adjustment: A list of underlying diseases, sex, race, admission month Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.33 (1.19, 1.49), < 0.001 Outcome (prognostic factor) Mortality (obesity) (cohort 4) Follow‐up Number of patients followed completely for the outcome: 14,343 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Log‐binomial models using the modified Poisson regression approach The set of prognostic factors used for adjustment: A list of underlying diseases, sex, race, admission month Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.26 (1.16, 1.36), < 0.001 Outcome (prognostic factor) Mortality (obesity) (cohort 5) Follow‐up Number of patients followed completely for the outcome: 12,855 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Log‐binomial models using the modified Poisson regression approach The set of prognostic factors used for adjustment: A list of underlying diseases, sex, race, admission month Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.16 (1.08, 1.25), < 0.001 Outcome (prognostic factor) Mortality (obesity) (cohort 6) Follow‐up Number of patients followed completely for the outcome: 13,472 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Log‐binomial models using the modified Poisson regression approach The set of prognostic factors used for adjustment: A list of underlying diseases, sex, race, admission month Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.11 (1.02, 1.22), < 0.001
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	No	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Goyal 2020.

Study characteristics
Notes	English title Obesity and COVID‐19 in New York City: a retrospective cohort study Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 05/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 2 Study setting: Inpatient Number of participants recruited: 1687 Sampling method: Consecutive participants Participants Female participants (absolute number): 683 Age measure, value: Median (IQR), 66.5 (53.7‐77.2) Inclusion criteria: NR Exclusion criteria: Patients who did not have height or weight data available to calculate body mass index (BMI) Smoking frequency: 81 Diabetes frequency: 526 Hypertension frequency: 956 Cardiovascular disease frequency: CAD (279), 121 (HF) Asthma frequency: 159 Chronic obstructive pulmonary disease frequency: 103 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 101 Cancer frequency: 121 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 30 The time when obesity has been measured: Some time after presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 525 Prognostic factor(s): Overweight (BMI 25–29.9 kg/m2) Mild‐to‐moderate obesity (BMI 30–39.9 kg/m2) Morbid obesity (BMI > 40 kg/m2) Outcome(s) Mortality Mechanical ventilation Outcome (prognostic factor) Mortality (overweight (BMI 25–29.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 1104 Number of obese patients followed completely for the outcome: 557 Number of non‐obese patients followed completely for the outcome: 547 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, race, smoking, diabetes, hypertension, chronic obstructive pulmonary disease, asthma, end‐stage renal disease, coronary artery disease, heart failure, and cancer Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.75 (0.56, 1), NR Outcome (prognostic factor) Mortality (mild‐to‐moderate obesity (BMI 30–39.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 981 Number of obese patients followed completely for the outcome: 434 Number of non‐obese patients followed completely for the outcome: 547 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, race, smoking, diabetes, hypertension, chronic obstructive pulmonary disease, asthma, end‐stage renal disease, coronary artery disease, heart failure, and cancer Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.98 (0.7, 1.36), NR Outcome (prognostic factor) Mortality (morbid obesity (BMI > 40 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 638 Number of obese patients followed completely for the outcome: 91 Number of non‐obese patients followed completely for the outcome: 547 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, race, smoking, diabetes, hypertension, chronic obstructive pulmonary disease, asthma, end‐stage renal disease, coronary artery disease, heart failure, and cancer Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.41 (0.74, 2.7), NR Outcome (prognostic factor) Mechanical ventilation (overweight (BMI 25–29.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 1104 Number of obese patients followed completely for the outcome: 557 Number of non‐obese patients followed completely for the outcome: 547 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, race, smoking, diabetes, hypertension, chronic obstructive pulmonary disease, asthma, end‐stage renal disease, coronary artery disease, heart failure, and cancer Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.32 (1.03, 1.69), NR Outcome (prognostic factor) Mechanical ventilation (mild‐to‐moderate obesity (BMI 30–39.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 981 Number of obese patients followed completely for the outcome: 434 Number of non‐obese patients followed completely for the outcome: 547 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, race, smoking, diabetes, hypertension, chronic obstructive pulmonary disease, asthma, end‐stage renal disease, coronary artery disease, heart failure, and cancer Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.8 (1.39, 2.35), NR Outcome (prognostic factor) Mortality (morbid obesity (BMI > 40 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 638 Number of obese patients followed completely for the outcome: 91 Number of non‐obese patients followed completely for the outcome: 547 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, race, smoking, diabetes, hypertension, chronic obstructive pulmonary disease, asthma, end‐stage renal disease, coronary artery disease, heart failure, and cancer Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.74 (1.08, 2.8), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	No	Appendix 3
Study Attrition Mechanical ventilation	No	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Gu 2020.

Study characteristics
Notes	English title Characteristics associated with racial/ethnic disparities in COVID‐19 outcomes in an academic health care system Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Registry data Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 5698 Sampling method: Consecutive participants Participants Female participants (absolute number): 3533 Age measure, value: Mean (SD), 47 (20.9) Inclusion criteria: Tested for Covid‐19 (either positive or negative) Exclusion criteria: Patients who were still staying at the hospital, age at BMI measurement was missing or below 18 years, height and/or weight were missing, height measurements were below 69 cm or above 234 cm, weight was above 400 kg, BMI deviated more than one unit from BMI calculated from height and weight. Outliers for multiple values per person were defined as values that exceeded the median BMI +/‐ 3 x the median absolute deviation (MAD) Smoking frequency: 484 (999 missing) Diabetes frequency: 1123 (1083 missing) Hypertension frequency: NR Cardiovascular disease frequency: 4205 (2291 missing) Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 78.9 (unspecified) (2291 missing)  Immunosuppression frequency: NR Chronic kidney disease frequency: 1117 (2291 missing) Cancer frequency: 1652 (2291 missing) Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI classified as 18.5, 18.5‐25, 25‐30 and more than 30 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 1943 (1090 missing) Prognostic factor(s): BMI > 30 25 < BMI < 30 BMI (continuous) Outcome(s) Hospitalisation ICU admission Mortality Outcome (prognostic factor) Hospitalisation (BMI > 30) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 688 Number of obese patients followed completely for the outcome: 525 Number of non‐obese patients followed completely for the outcome: 163 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 3.03 (1.63, 5.64), NR Outcome (prognostic factor) Hospitalisation (25 < BMI < 30) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 466 Number of obese patients followed completely for the outcome: 303 Number of non‐obese patients followed completely for the outcome: 163 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.26 (1.20, 4.25), NR Outcome (prognostic factor) Hospitalisation (BMI > 30) (cohort 3) Follow‐up Number of patients followed completely for the outcome: 371 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.89 (1.31, 6.35), NR Outcome (prognostic factor) Hospitalisation (25 < BMI < 30) (cohort 3) Follow‐up Number of patients followed completely for the outcome: 371 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.81 (0.81, 4.01), NR Outcome (prognostic factor) Hospitalisation (BMI > 30) (cohort 4) Follow‐up Number of patients followed completely for the outcome: 271 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 6.85 (0.90, 51.99), NR Outcome (prognostic factor) Hospitalisation (25 < BMI < 30) (cohort 4) Follow‐up Number of patients followed completely for the outcome: 271 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 71.6 (0.88, 57.92), NR Outcome (prognostic factor) ICU admission (BMI > 30) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 688 Number of obese patients followed completely for the outcome: 525 Number of non‐obese patients followed completely for the outcome: 163 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.71 (1.17, 6.29), NR Outcome (prognostic factor) ICU admission (25 < BMI < 30) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 466 Number of obese patients followed completely for the outcome: 303 Number of non‐obese patients followed completely for the outcome: 163 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.48 (0.62, 3.54), NR Outcome (prognostic factor) ICU admission (BMI > 30) (cohort 4) Follow‐up Number of patients followed completely for the outcome: 271 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 5.68 (0.27, 118.10), NR Outcome (prognostic factor) ICU admission (25 < BMI < 30) (cohort 4) Follow‐up Number of patients followed completely for the outcome: 271 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 4.79 (0.22, 105.00), NR Outcome (prognostic factor) ICU admission (BMI > 30) (cohort 5) Follow‐up Number of patients followed completely for the outcome: 398 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.72 (0.96, 7.71), NR Outcome (prognostic factor) ICU admission (25 < BMI < 30) (cohort 5) Follow‐up Number of patients followed completely for the outcome: 398 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.46 (0.50, 4.21), NR Outcome (prognostic factor) Mortality (BMI (continuous)) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 688 Number of obese patients followed completely for the outcome: 525 Number of non‐obese patients followed completely for the outcome: 163 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.22 (0.74, 2.01), NR Outcome (prognostic factor) Mortality (BMI (continuous)) (cohort 4) Follow‐up Number of patients followed completely for the outcome: 271 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.04 (0.52, 2.09), NR Outcome (prognostic factor) Mortality (BMI (continuous)) (cohort 5) Follow‐up Number of patients followed completely for the outcome: 398 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, alcohol consumption, BMI, comorbidities, ever‐smoker, race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.73 (0.28, 1.91), NR
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Mortality	No	Appendix 3
Study Attrition ICU admission	No	Appendix 3
Study Attrition Hospitalisation	No	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Guerson 2020.

Study characteristics
Notes	English title The impact of obesity among patients with COVID‐19 pneumonia Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 05/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 1 Study setting inpatient Number of participants recruited 3538 Sampling method consecutive participants Participants Female participants (percentage), 45 Age measure, value median (interquartile range), 65 (55, 75) Inclusion criteria COVID‐19 positive  Exclusion criteria NR Smoking NR Diabetes (absolute number), 954 Hypertension (absolute number), 1342 Cardiovascular diseases NR Please indicate if additional information is available  NR Asthma NR Chronic obstructive pulmonary disease NR Other pulmonary diseases NR Please indicate if additional information is available  NR Immunosuppression (absolute number), 354 Please indicate if additional information is available NR Chronic kidney disease (absolute number), 132 Cancer (absolute number), 3464 Steroid administration NR Supplemental oxygen NR Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Obesity was defined as per the CDC guidelines as body mass index (BMI) > 30 kg/m2. Obesity classes were defined as; Class 1 BMI 30 to 35 kg/m2, Class 2 BMI 35 to 40 kg/m2, and Class 3 BMI > 40 kg/m2. The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency percentage Frequency value 41.72 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) mortality Outcome (prognostic factor) Mortality (obesity class 1) Outcome Mortality Prognostic factor (category):  obesity class 1 Follow‐up Number of patients followed completely for this outcome  2874 Number of obese patients followed completely for this outcome  812 Number of non‐obese patients followed completely for this outcome  2062 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment unspecified Effect measure for obesity NR Effect measure value (95% CI) 1.15 (0.93, 1.4) Outcome (prognostic factor) Mortality (obesity class 2) Outcome Mortality Prognostic factor (category):  obesity class 2 Follow‐up Number of patients followed completely for this outcome  2431 Number of obese patients followed completely for this outcome  369 Number of non‐obese patients followed completely for this outcome  2062 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment unspecified Effect measure for obesity NR Effect measure value (95% CI) 1.33 (1.00, 1.76) Outcome (prognostic factor) Mortality (obesity class 3) Outcome Mortality Prognostic factor (category):  obesity class 3 Follow‐up Number of patients followed completely for this outcome  2357 Number of obese patients followed completely for this outcome  295 Number of non‐obese patients followed completely for this outcome  2062 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment unspecified Effect measure for obesity NR Effect measure value (95% CI) 1.92 (1.40, 2.63)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Gupta 2020.

Study characteristics
Notes	English title Factors associated with death in critically ill patients with coronavirus disease 2019 in the US Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 05/2020 Study design: Prospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 65 Study setting: Inpatient Number of participants recruited: 2215 Sampling method: Consecutive participants Participants Female participants (absolute number): 779 Age measure, value: Mean (SD), 60.5 (14.5) Inclusion criteria: Adults > 18 with positive COVID‐19 test Exclusion criteria: NR Smoking frequency: 656 Diabetes frequency: 861 Hypertension frequency: 1322 Cardiovascular disease frequency: 484 Asthma frequency: 258 Chronic obstructive pulmonary disease frequency: 173 Other pulmonary disease frequency: NR Immunosuppression frequency: 65 Chronic kidney disease frequency: 64 Cancer frequency: 112 Steroid administration frequency: 800 Supplemental oxygen administration frequency: 1496 Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): NR Prognostic factor(s): Overweight Obesity class I Obesity class II Obesity class III Outcome(s) Mortality Outcome (prognostic factor) Mortality (overweight) Follow‐up Number of patients followed completely for the outcome: 2215 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Active cancer, age, body mass index (calculated as weight in kilograms divided by height in meters squared), chronic obstructive pulmonary disease, congestive heart failure, coronary artery disease, covariates assessed at ICU admission (lymphocyte count, ratio of the PaO2 to the fraction of inspired oxygen [FIO2], shock, and the kidney, liver, and coagulation components of the Sequential Organ Failure Assessment score), current smoking status, diabetes, duration of symptoms before ICU admission, hypertension, sex, race  Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.01 (0.73, 1.39), NR Outcome (prognostic factor) Mortality (obesity class I) Follow‐up Number of patients followed completely for the outcome: 2215 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Active cancer, age, body mass index (calculated as weight in kilograms divided by height in meters squared), chronic obstructive pulmonary disease, congestive heart failure, coronary artery disease, covariates assessed at ICU admission (lymphocyte count, ratio of the PaO2 to the fraction of inspired oxygen [FIO2], shock, and the kidney, liver, and coagulation components of the Sequential Organ Failure Assessment score), current smoking status, diabetes, duration of symptoms before ICU admission, hypertension, sex, race  Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.97 (0.69, 1.37), NR Outcome (prognostic factor) Mortality (obesity class II) Follow‐up Number of patients followed completely for the outcome: 2215 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Active cancer, age, body mass index (calculated as weight in kilograms divided by height in meters squared), chronic obstructive pulmonary disease, congestive heart failure, coronary artery disease, covariates assessed at ICU admission (lymphocyte count, ratio of the PaO2 to the fraction of inspired oxygen [FIO2], shock, and the kidney, liver, and coagulation components of the Sequential Organ Failure Assessment score), current smoking status, diabetes, duration of symptoms before ICU admission, hypertension, sex, race  Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.24 (0.81, 1.89), NR Outcome (prognostic factor) Mortality (obesity class III) Follow‐up Number of patients followed completely for the outcome: 2215 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Active cancer, age, body mass index (calculated as weight in kilograms divided by height in meters squared), chronic obstructive pulmonary disease, congestive heart failure, coronary artery disease, covariates assessed at ICU admission (lymphocyte count, ratio of the PaO2 to the fraction of inspired oxygen [FIO2], shock, and the kidney, liver, and coagulation components of the Sequential Organ Failure Assessment score), current smoking status, diabetes, duration of symptoms before ICU admission, hypertension, sex, race  Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.51 (1.01, 2.25), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Hajifathalian 2020.

Study characteristics
Notes	English title Obesity is associated with worse outcomes in COVID‐19: analysis of early data from New York City Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 2 Study setting: Inpatient Number of participants recruited: 770 (cohort 1), 975 (cohort 2) Sampling method: Consecutive participants Participants Female participants (absolute number): 302 (cohort 1), 380 (cohort 2) Age measure, value: Mean (SD), 64 (16.7) Inclusion criteria: COVID‐19 positive patients admitted to either the Emergency Department or inpatient wards with complete BMI data (cohort 1), age > 18 and PCR‐positive for COVID‐19 (cohort 2) Exclusion criteria: NR (cohort 1), missing BMI data (cohort 2) Smoking frequency: NR Diabetes frequency: 238 (cohort 1), 301 (cohort 2) Hypertension frequency: 432 (cohort 1), 547 (cohort 2) Cardiovascular disease frequency: Obstructive sleep apnoea Asthma or chronic obstructive pulmonary disease frequency: 98 (cohort 1), 124 (cohort 2) Other pulmonary disease frequency: Obstructive sleep apnoea (36 (cohort 1), 46 (cohort 2)) Immunosuppression frequency: NR Chronic kidney disease frequency: 100 (cohort 1), 126.75 (cohort 2) Cancer frequency: 98 (cohort 1), 124 (cohort 2) Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 30 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 277 (cohort 1), NR (cohort 2) Prognostic factor(s): BMI > 30 Outcome(s) Mortality ICU admission Mechanical ventilation Outcome (prognostic factor) Mortality (BMI > 30) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 770 Number of obese patients followed completely for the outcome: 227 Number of non‐obese patients followed completely for the outcome: 493 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Generalised linear models (GLM; with maximum likelihood optimisation and robust standard error estimation) with a Poisson distribution for the dependent variable and a logarithmic (log) link function used to estimate risk ratios and their confidence intervals The set of prognostic factors used for adjustment: Age, race/ethnicity, and troponin I level Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.15 (0.62, 2.14), 0.663 Outcome (prognostic factor) ICU admission (BMI > 30) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 770 Number of obese patients followed completely for the outcome: 227 Number of non‐obese patients followed completely for the outcome: 493 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Generalised linear models (GLM; with maximum likelihood optimisation and robust standard error estimation) with a Poisson distribution for the dependent variable and a logarithmic (log) link function used to estimate risk ratios and their confidence intervals The set of prognostic factors used for adjustment: Age, race/ethnicity, and troponin I level Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.76 (1.24, 2.48), 0.001 Outcome (prognostic factor) Mechanical ventilation (BMI > 30) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 747 Number of obese patients followed completely for the outcome: 277 Number of non‐obese patients followed completely for the outcome: 465 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Generalised linear models (GLM; with maximum likelihood optimisation and robust standard error estimation) with a Poisson distribution for the dependent variable and a logarithmic (log) link function used to estimate risk ratios and their confidence intervals The set of prognostic factors used for adjustment: Age, race/ethnicity, and troponin I level Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.72 (1.22, 2.44), 0.002
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Study Attrition ICU admission	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Confounding Bias Mechanical ventilation	No	Appendix 3
Confounding Bias ICU admission	No	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Halasz 2020.

Study characteristics
Notes	English title Obesity, overweight and survival in critically ill patients with SARS‐CoV‐2 pneumonia: is there an obesity paradox? Preliminary results from Italy Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 242 Sampling method: Consecutive participants Participants Female participants (absolute number): 44 Age measure, value: Median (IQR), 64 (56‐71) Inclusion criteria: Patients with laboratory‐confirmed COVID‐19 infection treated with invasive ventilation and admitted to the ICU Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 37 Hypertension frequency: 110 Cardiovascular disease frequency: 35 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 21 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: WHO cut‐points The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 48 Prognostic factor(s): BMI 25‐29.9 BMI 30‐35 BMI 35‐40 BMI> 40 Outcome(s) Mortality Outcome (prognostic factor) Mortality (BMI 25‐29.9) Follow‐up Number of patients followed completely for the outcome: 142 Number of obese patients followed completely for the outcome: 104 Number of non‐obese patients followed completely for the outcome: 38 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, gender, comorbidities (hypertension, cardiovascular disease, COPD, diabetes) Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.04 (0.43, 2.5), 0.93 Outcome (prognostic factor) Mortality (BMI 30‐35) Follow‐up Number of patients followed completely for the outcome: 69 Number of obese patients followed completely for the outcome: 31 Number of non‐obese patients followed completely for the outcome: 38 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, gender, comorbidities (hypertension, cardiovascular disease, COPD, diabetes) Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.51 (0.51, 4.44), 0.45 Outcome (prognostic factor) Mortality (BMI 35‐40) Follow‐up Number of patients followed completely for the outcome: 45 Number of obese patients followed completely for the outcome: 7 Number of non‐obese patients followed completely for the outcome: 38 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, gender, comorbidities (hypertension, cardiovascular disease, COPD, diabetes) Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.74 (0.31, 9.69), 0.53 Outcome (prognostic factor) Mortality (BMI > 40) Follow‐up Number of patients followed completely for the outcome: 48 Number of obese patients followed completely for the outcome: 10 Number of non‐obese patients followed completely for the outcome: 38 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, gender, comorbidities (hypertension, cardiovascular disease, COPD, diabetes) Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 3.91 (NR), 0.09
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Hashemi 2020.

Study characteristics
Notes	English title Impact of chronic liver disease on outcomes of hospitalized patients with COVID‐19: a multicentre United States experience Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 9 Study setting: Inpatient Number of participants recruited: 363 Sampling method: Consecutive participants Participants Female participants (absolute number): 162 Age measure, value: Mean (SD), 63.4 (16.5) Inclusion criteria: All consecutive adult patients hospitalised with a positive SARS‐CoV‐2 infection via polymerase chain reaction (PCR) nasopharyngeal swab or tracheal aspirate from 11 March to 2 April 2020 Exclusion criteria: Liver transplant recipients Smoking frequency: 41 Diabetes frequency: 117 Hypertension frequency: 212 Cardiovascular disease frequency: 91 Asthma frequency: NR Pulmonary disease frequency: 76 Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: NR Obesity frequency (absolute number): NR Prognostic factor(s): Obesity Outcome(s) Mechanical ventilation ICU admission All‐cause in‐hospital mortality Outcome (prognostic factor) Mechanical ventilation (obesity) Follow‐up Number of patients followed completely for the outcome: NR Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.23 (0.77, 1.98), 0.39 Outcome (prognostic factor) ICU admission (obesity) Follow‐up Number of patients followed completely for the outcome: NR Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.26 (0.79, 1.98), 0.33 Outcome (prognostic factor) All‐cause in‐hospital mortality (obesity) Follow‐up Number of patients followed completely for the outcome: NR Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.03 (0.51, 2.09), 0.94
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Unclear	Appendix 3
Outcome Measurement ICU admission	Unclear	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Confounding Bias Mechanical ventilation	No	Appendix 3
Confounding Bias ICU admission	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Hendra 2021.

Study characteristics
Notes	English title Identifying prognostic risk factors for poor outcome following COVID‑19 disease among in‐centre haemodialysis patients: role of inflammation and frailty Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 05/2020 Study design retrospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas 4 Study setting outpatient and inpatient Number of participants recruited 148 Sampling method consecutive participants Participants Female participants (percentage), 43.2 Age measure, value mean (standard deviation), 64.13 (14.6) Inclusion criteria RT‐PCR positive patients Exclusion criteria NR Smoking NR Diabetes (absolute number), 78 Hypertension (absolute number), 122 Cardiovascular diseases (absolute number), 81 Please indicate if additional information is available  Chronic cardiac disease  Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (absolute number), 19 Please indicate if additional information is available  Chronic pulmonary disease  Immunosuppression (absolute number), 18 Please indicate if additional information is available Immunosuppressive treatment Chronic kidney disease (percentage), 100 Cancer (unspecified) Steroid administration (absolute number), 13 Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment ACEI/ARB Dose if applicable NR Duration if applicable NR Percentage received this treatment 15 Prognostic factor(s) Study’s definition for obesity BMI continuous  The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity NR Threshold used for definition of obesity  NR Measure of frequency NR Frequency value NR How many eligible outcomes reported? 2 How many eligible outcomes reported? 2 Outcome(s) mortality, hospitalisation Outcome (prognostic factor) Mortality (BMI continuous ) Outcome Mortality  Prognostic factor (category):  BMI continuous  Follow‐up Number of patients followed completely for this outcome  148 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age, gender, ethnicity, BMI, frailty score, deprivation index, type of vascular access, comorbidities (diabetes, hypertension, chronic cardiac and pulmonary disease), use of immunosuppression and biomarkers including CRP and NLR Effect measure for obesity odds ratio  Effect measure value (95% CI) 0.94 (0.87, 1.01) Outcome (prognostic factor) Hospitalisation (BMI continuous) Outcome Hospitalisation Prognostic factor (category):  BMI continuous Follow‐up Number of patients followed completely for this outcome  148 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment age, gender, ethnicity, BMI, frailty score, deprivation index, type of vascular access, comorbidities (diabetes, hypertension, chronic cardiac and pulmonary disease), use of immunosuppression and biomarkers including CRP and NLR Effect measure for obesity odds ratio  Effect measure value (95% CI) 0.96 (0.9, 1.03)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition Hospitalisation	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Hendren 2021.

Study characteristics
Notes	English title Association of body mass index and age with morbidity and mortality in patients hospitalized with COVID‐19: results from the American Heart Association COVID‐19 Cardiovascular Disease Registry Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): 07/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 88 Study setting: Inpatient Number of participants recruited: 7606 Sampling method: Consecutive participants Participants Female participants (absolute number): 3399 Age measure, value: Median (IQR), 63 (49‐75) Inclusion criteria: All adults (≥18 years old) with available BMI data and completed field entries for age, sex, admission date, discharge date, discharge disposition, and a medical history (selected as either none or as applicable conditions) Exclusion criteria: NR Smoking frequency: 510 Diabetes frequency: 2799 Hypertension frequency: 4525 Cardiovascular disease frequency: 1526 Asthma frequency: 741 Chronic obstructive pulmonary disease frequency: 630 Other pulmonary disease frequency: Pulmonary embolism (143) Immunosuppression frequency: NR Chronic kidney disease frequency: 972 Cancer frequency: 833 Steroid administration frequency: 1645 Supplemental oxygen administration frequency: NR Other treatments (frequency): Remdesivir (636) Prognostic factor(s) Study’s definition for obesity: World Health Organization (WHO) obesity classification, defined as: underweight (< 18.5 kg/m2), normal weight (18.5‐24.9 kg/m2), overweight (25.0‐29.9 kg/m2), class I obesity (30.0‐34.9 kg/m2), class II obesity (35.0–39.9 kg/m2), and class III obesity (≥ 40.0 kg/m2) The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 3311 Prognostic factor(s): Class I obesity (30.0‐34.9 kg/m2) Class II obesity (35.0‐39.9 kg/m2) Class III obesity (BMI ≥ 40) Outcome(s) Mortality Mechanical ventilation Outcome (prognostic factor) Mortality (Class I obesity (30.0‐34.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 7606 Number of obese patients followed completely for the outcome: 3311 Number of non‐obese patients followed completely for the outcome: 4295 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, chronic kidney disease, CVD (myocardial infarction, stroke, heart failure, or percutaneous coronary intervention), diabetes, hypertension, race/ethnicity, sex Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1 (0.83, 1.17), NR Outcome (prognostic factor) Mortality (Class II obesity (35.0‐39.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 7606 Number of obese patients followed completely for the outcome: 3311 Number of non‐obese patients followed completely for the outcome: 4295 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, chronic kidney disease, CVD (myocardial infarction, stroke, heart failure, or percutaneous coronary intervention), diabetes, hypertension, race/ethnicity, sex Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.15 (0.91, 1.44), NR Outcome (prognostic factor) Mortality (Class III obesity (BMI ≥ 40)) Follow‐up Number of patients followed completely for the outcome: 7606 Number of obese patients followed completely for the outcome: 3311 Number of non‐obese patients followed completely for the outcome: 4295 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, chronic kidney disease, CVD (myocardial infarction, stroke, heart failure, or percutaneous coronary intervention), diabetes, hypertension, race/ethnicity, sex Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.26 (1.00, 1.58), NR Outcome (prognostic factor) Mechanical ventilation (Class I obesity (30.0‐34.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 7606 Number of obese patients followed completely for the outcome: 3311 Number of non‐obese patients followed completely for the outcome: 4295 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, chronic kidney disease, CVD (myocardial infarction, stroke, heart failure, or percutaneous coronary intervention), diabetes, hypertension, race/ethnicity, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.54 (1.29, 1.84), NR Outcome (prognostic factor) Mechanical ventilation (Class II obesity (35.0‐39.9 kg/m2)) Follow‐up Number of patients followed completely for the outcome: 7606 Number of obese patients followed completely for the outcome: 3311 Number of non‐obese patients followed completely for the outcome: 4295 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, chronic kidney disease, CVD (myocardial infarction, stroke, heart failure, or percutaneous coronary intervention), diabetes, hypertension, race/ethnicity, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.88 (1.52, 2.32), NR Outcome (prognostic factor) Mechanical ventilation (Class III obesity (BMI ≥ 40)) Follow‐up Number of patients followed completely for the outcome: 7606 Number of obese patients followed completely for the outcome: 3311 Number of non‐obese patients followed completely for the outcome: 4295 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, chronic kidney disease, CVD (myocardial infarction, stroke, heart failure, or percutaneous coronary intervention), diabetes, hypertension, race/ethnicity, sex Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.08 (1.68, 2.58), NR
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Unclear	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Hernandez‐Galdamez 2020.

Study characteristics
Notes	English title Increased risk of hospitalization and death in patients with COVID‐19 and pre‐existing noncommunicable diseases and modifiable risk factors in Mexico Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 06/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 475 Study setting: Outpatient and inpatient Number of participants recruited: 211,003 Sampling method: NR Participants Female participants (absolute number): 95,561 Age measure, value: Mean (SD), 45.7 (16.3) Inclusion criteria: laboratory‐confirmed COVID‐19 cases were reported in the MoH database up to June 27. Exclusion criteria: missing or unknown comorbidity or condition Smoking frequency: 16,445 Diabetes frequency: 34,685 Hypertension frequency: 42,453 Cardiovascular disease frequency: 4949 Asthma frequency: 5854 Chronic obstructive pulmonary disease frequency: 3721 Other pulmonary disease frequency: NR Immunosuppression frequency: 2895 Chronic kidney disease frequency: 4581 Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: NR Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): 41,344 Prognostic factor(s): Obesity Outcome(s) Hospitalisation ICU admission Endotracheal Intubation (mechanical ventilation) Mortality Outcome (prognostic factor) Hospitalisation (obesity) Follow‐up Number of patients followed completely for the outcome: 21,103 Number of obese patients followed completely for the outcome: 41,344 Number of non‐obese patients followed completely for the outcome: 169,659 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, CKD, immunosuppression, diabetes, COPD, hypertension, CVD, asthma, obesity, smoking Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.29 (1.25, 1.32), < 0.001 Outcome (prognostic factor) ICU admission (obesity) Follow‐up Number of patients followed completely for the outcome: 21,103 Number of obese patients followed completely for the outcome: 41,344 Number of non‐obese patients followed completely for the outcome: 169,659 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, CKD, immunosuppression, diabetes, COPD, hypertension, CVD, asthma, obesity, smoking Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.59 (1.49, 1.69), < 0.001 Outcome (prognostic factor) Mechanical ventilation (obesity) Follow‐up Number of patients followed completely for the outcome: 21,103 Number of obese patients followed completely for the outcome: 41,344 Number of non‐obese patients followed completely for the outcome: 169,659 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, CKD, immunosuppression, diabetes, COPD, hypertension, CVD, asthma, obesity, smoking Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.62 (1.53, 1.71), < 0.001 Outcome (prognostic factor) Mortality (obesity) Follow‐up Number of patients followed completely for the outcome: 21,103 Number of obese patients followed completely for the outcome: 41,344 Number of non‐obese patients followed completely for the outcome: 169,659 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, CKD, immunosuppression, diabetes, COPD, hypertension, CVD, asthma, obesity, smoking Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.42 (1.37, 1.47), < 0.001
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Unclear	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Confounding Bias Mechanical ventilation	No	Appendix 3
Confounding Bias ICU admission	No	Appendix 3
Confounding Bias Hospitalisation	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Hosseinzadeh 2021.

Study characteristics
Notes	English title Should all patients with hypertension be worried about developing severe coronavirus disease 2019 (COVID‐19)? Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): NR Study design: retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 176 (cohort 1), 422 (cohort 2) Sampling method: NR Participants Female participants (absolute number): 115 (cohort 1), 289 (cohort 2) Age measure, value: Mean (SD), 58.21 (22.5) (cohort 1), 56.15 (20.2) (cohort 2) Inclusion criteria: Cases with COVID‐19 pneumonia who were admitted to Baqiyatallah Hospital in Tehran – Iran Exclusion criteria: Incomplete medical profiles, patients who were not receiving any kind of anti‐hypertensive medications and patients who were receiving any corticosteroids Smoking frequency: 4 (cohort 1), 12 (cohort 2) Diabetes frequency: 72 (cohort 1), 76 (cohort 2) Hypertension frequency: 176 (cohort 1), 0 (cohort 2) Cardiovascular disease frequency: 63 (cohort 1),44 (cohort 2) Asthma frequency: 31 (cohort 1), 67 (cohort 2) Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 36 (cohort 1), 52 (cohort 2) Immunosuppression frequency: NR Chronic kidney disease frequency: 25 (cohort 1), 33 (cohort 2) Cancer frequency: 4 (cohort 1), 8 (cohort 2) Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Comment: Asthma consists of both asthma and allergy history Prognostic factor(s) Study’s definition for obesity: BMI > 25 kg/m2 The time when obesity has been measured: NR Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): NR Prognostic factor(s): Overweight/obesity Outcome(s) Severe COVID‐19 Outcome (prognostic factor) Severe COVID‐19 (overweight/obesity) (cohort1) Follow‐up Number of patients followed completely for the outcome: 176 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age > 60 years, BMI > 25 kg/m2, increased hospital stays, CVD, diabetes, kidney disease Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.80 (1.02, 2.42), 0.027 Outcome (prognostic factor) Severe COVID‐19 (overweight/obesity) (cohort2) Follow‐up Number of patients followed completely for the outcome: 422  Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age > 60 years, BMI > 25 kg/m2, increased hospital stays, CVD, diabetes, kidney disease Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.34 (0.88, 1.89), 0.21
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Hu 2020.

Study characteristics
Notes	English title Clinical epidemiological analyses of overweight/obesity and abnormal liver function contributing to prolonged hospitalization in patients infected with COVID‐19 Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 02/2020 Study design: Retrospective cohort Study centre(s): Single centre/clinic/area within a country Number of centres, clinics or areas: 1 Study setting: Inpatient Number of participants recruited: 58 Sampling method: NR Participants Female participants (absolute number): 22 Age measure, value: Mean (SD), 49.2 (13.1) Inclusion criteria: Mild COVID‐19 patients Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: NR Hypertension frequency: NR Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 24 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 24 Obesity frequency (absolute number): 32 Prognostic factor(s): BMI ≥ 24 kg/m2 Outcome(s) Prolonged hospitalisation Outcome (prognostic factor) Prolonged hospitalisation (BMI ≥ 24 kg/m2) Follow‐up Number of patients followed completely for the outcome: 52 Number of obese patients followed completely for the outcome: 29 Number of non‐obese patients followed completely for the outcome: 23 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.83 (0.74, 0.92), 0.001 Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, SBP, DBP, peripheral absolute, neutrophil count, monocyte count, lymphocyte count, FPG, albumin, creatinine, BUN, CRP Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.75 (0.63, 0.90), 0.002
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Hospitalisation	No	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Huang 2020.

Study characteristics
Notes	English title Clinical findings of patients with coronavirus disease 2019 in Jiangsu province, China: a retrospective, multicentre study Study setting Start of study recruitment (MM/YYYY): 01/2020 End of study recruitment (MM/YYYY): 02/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 8 Study setting: Inpatient Number of participants recruited: 202 Sampling method: NR Participants Female participants (absolute number): 86 Age measure, value: Median (IQR), 44 (33‐54) Inclusion criteria: COVID‐19 patients from 8 designated hospitals in 8 cities of Jiangsu province, China Exclusion criteria: NR Smoking frequency: 16 Diabetes frequency: 19 Hypertension frequency: 29 Cardiovascular disease frequency: 5 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 7 Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: 2 Steroid administration frequency: 64 Supplemental oxygen administration frequency: 109 Other treatments (frequency): Antiviral therapy (196), atomised inhalation of interferon α‐2b (121), lopinavir/ritonavir (180), Arbidol (59), oseltamivir (32), antibiotic therapy (149), use of gamma globulin (31) Prognostic factor(s) Study’s definition for obesity: BMI > 28 kg/m2 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 28 Obesity frequency (absolute number): 24 Prognostic factor(s): BMI > 28 kg/m2 Outcome(s) Severe COVID‐19 Outcome (prognostic factor) Severe COVID‐19 (BMI > 28 kg/m2) Follow‐up Number of patients followed completely for the outcome: 202 Number of obese patients followed completely for the outcome: 24 Number of non‐obese patients followed completely for the outcome: 148 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 6.9 (2.381, 19.997), < 0.001 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, gender, BMI, hypertension, DM, smoking, WBC, neutrophils, lymphocyte, Hb, PLT, ALT, LDH, Tbil, ALB, CR, CRP, PT Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 9219 (2.731, 31.126), < 0.001
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	No	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Huh 2020.

Study characteristics
Notes	English title Impact of obesity, fasting plasma glucose level, blood pressure, and renal function on the severity of COVID‐19: a matter of sexual dimorphism? Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): NR Study design: Case‐control Study centre(s): Multiple centres/clinics/areas within a country  Number of centres, clinics or areas: NR Study setting: Outpatient and inpatient Number of participants recruited: 2231 Sampling method: NR Participants Female participants (absolute number): 1360 Age measure, value: Mean (NR), 53.7 (NR) Inclusion criteria: NR Exclusion criteria: Patients with lack of checkup data Smoking frequency: NR Diabetes frequency: 756 Hypertension frequency: 729 Cardiovascular disease frequency: 562 Asthma frequency: 574 Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 1093 Immunosuppression frequency: NR Chronic kidney disease frequency: 185 Cancer frequency: 182 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 25 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 25 Obesity frequency (absolute number): 855 Prognostic factor(s): Obesity class 1 Obesity class 2 & 3 Outcome(s) Severe COVID‐19 or death Outcome (prognostic factor) Severe COVID‐19 or death (obesity class 1) Follow‐up Number of patients followed completely for the outcome: 307 Number of obese patients followed completely for the outcome: 151 Number of non‐obese patients followed completely for the outcome: 156 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, NIHS expanded coverage for low household income, Charlson comorbidity index, diabetes, hypertension, chronic heart disease, chronic lung disease, asthma, chronic liver disease, chronic kidney disease, cancer, rheumatologic disease, chronic neurologic disease Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.69 (1.20, 2.38), 0.002 Outcome (prognostic factor) Severe COVID‐19 or death (obesity class 2 & 3) Follow‐up Number of patients followed completely for the outcome: 307 Number of obese patients followed completely for the outcome: 151 Number of non‐obese patients followed completely for the outcome: 156 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, NIHS expanded coverage for low household income, Charlson comorbidity index, diabetes, hypertension, chronic heart disease, chronic lung disease, asthma, chronic liver disease, chronic kidney disease, cancer, rheumatologic disease, chronic neurologic disease Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.23 (0.58, 2.60), 0.591
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Severe COVID	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Unclear	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Hur 2020.

Study characteristics
Notes	English title Factors associated with intubation and prolonged intubation in hospitalized patients with COVID‐19 Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 10 Study setting: Inpatient Number of participants recruited: 486 Sampling method: NR Participants Female participants (absolute number): 215 Age measure, value: Median (IQR), 59 (47‐69) Inclusion criteria: Age more than 18 years and were admitted to any of the 10 hospitals in the Northwestern Memorial HealthCare system spread across the Chicago metropolitan area between March 1 and April 8, 2020 Exclusion criteria: Hospitalised patients with documented ‘‘do not resuscitate and do not intubate’’ (DNR/DNI) orders and those who left the hospital against medical advice, and patients who had missing data on investigated predictor variables and did not reach a clinical endpoint of intubation or discharge from the hospital Smoking frequency: 163 Diabetes frequency: 160 Hypertension frequency: 267 Cardiovascular disease frequency: 111 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 78 Immunosuppression frequency: 45 Chronic kidney disease frequency: 42 Cancer frequency: 60 Steroid administration frequency: NR Supplemental oxygen administration frequency: 326 Other treatments (frequency): Antibiotics (329), hydroxychloroquine (268), IL‐6R inhibitor (33), remdesivir (9) Prognostic factor(s) Study’s definition for obesity: BMI ≥ 30 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 259 Prognostic factor(s): BMI 30‐39.99 BMI > 40 Outcome(s) Mechanical ventilation Time to extubation Outcome (prognostic factor) Mechanical ventilation (BMI 30‐39.99) Follow‐up Number of patients followed completely for the outcome: 468 Number of obese patients followed completely for the outcome: 259 Number of non‐obese patients followed completely for the outcome: 227 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, race and ethnicity, hospital, body mass index, respiratory rate temperature, O2 sat, pulse, DM, shortness of breath Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.46 (0.87, 2.46), 0.151 Outcome (prognostic factor) Mechanical ventilation (BMI > 40) Follow‐up Number of patients followed completely for the outcome: 468 Number of obese patients followed completely for the outcome: 259 Number of non‐obese patients followed completely for the outcome: 227 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, race and ethnicity, hospital, body mass index, respiratory rate temperature, O2 sat, pulse, DM, shortness of breath Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.92 (1.92, 4.00), 0.080 Outcome (prognostic factor) Time to extubation (BMI 30‐39.99) Follow‐up Number of patients followed completely for the outcome: 468 Number of obese patients followed completely for the outcome: 259 Number of non‐obese patients followed completely for the outcome: 227 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, race and ethnicity, hospital, body mass index, respiratory rate temperature, O2 sat, pulse, DM, shortness of breath Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.53 (0.32, 0.90), 0.018 Outcome (prognostic factor) Time to extubation (BMI > 40) Follow‐up Number of patients followed completely for the outcome: 468 Number of obese patients followed completely for the outcome: 259 Number of non‐obese patients followed completely for the outcome: 227 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, race and ethnicity, hospital, body mass index, respiratory rate temperature, O2 sat, pulse, DM, shortness of breath Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.40 (0.19, 0.82), 0.012
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mechanical ventilation	Yes	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mechanical ventilation	Unclear	Appendix 3
Confounding Bias Mechanical ventilation	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Iaccarino 2020.

Study characteristics
Notes	English title Gender differences in predictors of intensive care units admission among COVID‐19 patients: the results of the SARS‐RAS study of the Italian Society of Hypertension Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres/clinics/areas: 26 hospitals and centres Study setting: Inpatient Number of participants recruited: 2378 Sampling method: NR Participants Female participants (absolute number): 889 Age measure, value: Mean (SD) 68.21 (0.38) Inclusion criteria: patients aged 18 to 101 years with confirmed COVID‐19 according to World Health Organization interim guidance Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: 18.2% Hypertension frequency: 58.5% Cardiovascular disease frequency: coronary artery disease (14.3%), heart failure (12.1%) Asthma frequency: NR Chronic obstructive pulmonary disease frequency: 8.5% Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 6.1% Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Body mass index ≥ 30 kg/m2 according to the Center for Disease Control and Prevention The time when obesity has been measured:  Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition of obesity: 30 Obesity frequency (absolute number): 157 Prognostic factor(s): Obesity Outcome(s) ICU admission Outcome (prognostic factor) ICU admission (obesity) Follow‐up Number of patients followed completely for the outcome: 2378 Number of obese patients followed completely for the outcome: 157 Number of non‐obese patients followed completely for the outcome: 2221 Univariable unadjusted analysis for obesity Effect measure for obesity: Pearson Effect measure value (95% CI), P value: 0.103 (NR), 0.001 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, gender, hypertension, diabetes, CKD, heart failure, CAD, obesity Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 2.476 (1.724, 3.555), 0.0005
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Imam 2020.

Study characteristics
Notes	English title Independent correlates of hospitalization in 2040 patients with COVID‐19 at a large hospital system in Michigan, United States Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 8 Study setting: Outpatient and inpatient Number of participants recruited: 2040 Sampling method: NR Participants Female participants (absolute number): NR Age measure, value: NR Inclusion criteria: NR Exclusion criteria: NR Smoking frequency: NR Diabetes frequency: NR Hypertension frequency: NR Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: NR Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI > 30 kg/m2 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number):  Prognostic factor(s): BMI > 30 kg/m2 Outcome(s) Hospitalisation Outcome (prognostic factor) Hospitalisation (BMI > 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 2040 Number of obese patients followed completely for the outcome: NR Number of non‐obese patients followed completely for the outcome: NR Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.5 (1.2, 1.9), 0.002 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, CCI, race, ACEI/ARB use, BMI > 30, tachycardia (HR > 100), tachypnoea (RR < 20), hypoxia (sPO2 < 90%) Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.8 (1.4, 2.4), < 0.0005
Item	Authors' judgement	Support for judgement
Study Participation	No	Appendix 3
Study Attrition Hospitalisation	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Ioannou 2020.

Study characteristics
Notes	English title Risk factors for hospitalization, mechanical ventilation, or death among 10 131 US veterans with SARS‐CoV‐2 infection Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 05/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: NR Study setting: Outpatient and inpatient Number of participants recruited: 10,131 Sampling method: Consecutive participants Participants Female participants (absolute number): 912 Age measure, value: Mean (SD), 63.6 (16.2) Inclusion criteria: All VA enrollees who had nasopharyngeal swabs tested for SARS‐CoV‐2 nucleic acid by polymerase chain reaction in inpatient or outpatient VA facilities (including VA nursing homes) between February 28 and May 14, 2020 Exclusion criteria: VA employees Smoking frequency: 5207 (including ex‐smokers) Diabetes frequency: 3860 Hypertension frequency: 6291 Cardiovascular disease frequency: 3323 Asthma frequency: 750 Chronic obstructive pulmonary disease frequency: 1905 Other pulmonary disease frequency: Obstructive sleep apnoea (2715) Immunosuppression frequency: NR Chronic kidney disease frequency: 1864 Cancer frequency: 2300 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 30 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 4539 Prognostic factor(s): Obesity I Obesity II or III Outcome(s) Hospitalisation Mechanical ventilation Death Outcome (prognostic factor) Hospitalisation (obesity I) Follow‐up Number of patients followed completely for the outcome: 10,131 Number of obese patients followed completely for the outcome: 4542 Number of non‐obese patients followed completely for the outcome: 5337 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.81 (0.74, 0.90), NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Sex, age, race, ethnicity, urban vs rural, BMI, DM, cancer, HTN, CVD, CKD, cirrhosis, asthma, COPD, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.8 (0.72, 0.89), NR Outcome (prognostic factor) Hospitalisation (obesity II or III) Follow‐up Number of patients followed completely for the outcome: 10,131 Number of obese patients followed completely for the outcome: 4542 Number of non‐obese patients followed completely for the outcome: 5337 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.94 (0.84, 1.05), NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Sex, age, race, ethnicity, urban vs rural, BMI, DM, cancer, HTN, CVD, CKD, cirrhosis, asthma, COPD, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.87 (0.77, 0.98), NR Outcome (prognostic factor) Mechanical ventilation (obesity I) Follow‐up Number of patients followed completely for the outcome: 10,131 Number of obese patients followed completely for the outcome: 4542 Number of non‐obese patients followed completely for the outcome: 5337 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.23 (0.97, 1.57), NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Sex, age, race, ethnicity, urban vs rural, BMI, DM, cancer, HTN, CVD, CKD, cirrhosis, asthma, COPD, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.03 (0.80, 1.33), NR Outcome (prognostic factor) Mechanical ventilation (obesity II or III) Follow‐up Number of patients followed completely for the outcome: 10,131 Number of obese patients followed completely for the outcome: 4542 Number of non‐obese patients followed completely for the outcome: 5337 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.71 (1.33, 2.2), NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Sex, age, race, ethnicity, urban vs rural, BMI, DM, cancer, HTN, CVD, CKD, cirrhosis, asthma, COPD, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.22 (0.93, 1.61), NR Outcome (prognostic factor) Death (obesity I) Follow‐up Number of patients followed completely for the outcome: 10,131 Number of obese patients followed completely for the outcome: 4542 Number of non‐obese patients followed completely for the outcome: 5337 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.86 (0.71, 1.03), NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Sex, age, race, ethnicity, urban vs rural, BMI, DM, cancer, HTN, CVD, CKD, cirrhosis, asthma, COPD, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.84 (0.69, 1.01), NR Outcome (prognostic factor) Death (obesity II or III) Follow‐up Number of patients followed completely for the outcome: 10,131 Number of obese patients followed completely for the outcome: 4542 Number of non‐obese patients followed completely for the outcome: 5337 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.12 (0.91, 1.37), NR Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Sex, age, race, ethnicity, urban vs rural, BMI, DM, cancer, HTN, CVD, CKD, cirrhosis, asthma, COPD, smoking Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 0.97 (0.77, 1.21), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Jayanama 2021.

Study characteristics
Notes	English title The association between body mass index and severity of Coronavirus Disease 2019 (COVID‐19): a cohort study Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 04/2020 Study design prospective cohort Study centre(s) single centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting inpatient Number of participants recruited 147 Sampling method unspecified Participants Female participants (absolute number), 86 Age measure, value mean (standard deviation), 39.1 (13) Inclusion criteria confirmed COVID‐19, aged 15 years and older, and admitted to Chakri Naruebodindra Medical Institute between March 12 and April 30th, 2020 Exclusion criteria NR Smoking NR Diabetes (absolute number), 14 Hypertension (absolute number), 14 Cardiovascular diseases (unspecified) Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Obesity, defined as excessive accumulation of body fat, is generally determined by body mass index (BMI), calculated by body weight (kg) divided by height squared (m2) The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  25 Measure of frequency absolute number Frequency value 46 How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) severe COVID Outcome (prognostic factor) Severe pneumonia (severe COVID) (obesity (BMI > 25)) Outcome Severe pneumonia (severe COVID) Prognostic factor (category):  Obesity (BMI > 25) Follow‐up Number of  patients followed completely for this outcome  147 Number of obese patients followed completely for this outcome  46 Number of non‐obese patients followed completely for this outcome  101 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 6.41 (17.92, 2.29) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, sex, DM, HTN, dyslipidaemia Effect measure for obesity odds ratio  Effect measure value (95% CI) 4.73 (1.5, 14.94)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Severe COVID	Yes	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Severe COVID	Unclear	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Jirapinyo 2020.

Study characteristics
Notes	English title Effect of obesity and metabolic disease on severity of SARS‐CoV‐2 infection Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 04/2020 Study design prospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting unspecified Number of participants recruited 1680 Sampling method consecutive participants Participants Female participants (unspecified) Age measure, value mean (standard deviation), 51 (18) Inclusion criteria Confirmed COVID‐19 between March 1, 2020 and April 2, 2020 who were admitted to NR Centers. Exclusion criteria NR Smoking NR Diabetes (unspecified) Hypertension (unspecified) Cardiovascular diseases (unspecified) Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment unspecified Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Only defined morbid obesity as BMI > 35  The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  NR Measure of frequency percentage Frequency value 73.4 How many eligible outcomes reported? 3 How many eligible outcomes reported? 3 Outcome(s) hospitalisation, ICU admission, mechanical ventilation Outcome (prognostic factor) Hospitalisation (morbid obesity (BMI > 35)) Outcome Hospitalisation Prognostic factor (category):  Morbid obesity (BMI > 35) Follow‐up Number of patients followed completely for this outcome  1680 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.2 (1.6, 3.2) Outcome (prognostic factor) ICU admission (morbid obesity (BMI > 35)) Outcome ICU admission Prognostic factor (category):  Morbid obesity (BMI > 35) Follow‐up Number of patients followed completely for this outcome  1680 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 3.2 (1.9, 5.4) Outcome (prognostic factor) Intubation (morbid obesity (BMI > 35)) Outcome Intubation Prognostic factor (category):  Morbid obesity (BMI > 35) Follow‐up Number of patients followed completely for this outcome  1680 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity odds ratio  Effect measure value (95% CI) 3.4 (1.9, 5.9)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mechanical ventilation	No	Appendix 3
Study Attrition ICU admission	No	Appendix 3
Study Attrition Hospitalisation	No	Appendix 3
Prognostic Factor Measurement	Unclear	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Unclear	Appendix 3
Confounding Bias ICU admission	Unclear	Appendix 3
Confounding Bias Hospitalisation	Unclear	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Kaeuffer 2020.

Study characteristics
Notes	English title Clinical characteristics and risk factors associated with severe COVID‐19: prospective analysis of 1,045 hospitalised cases in North‐Eastern France, March 2020 Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 03/2020 Study design prospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 2 Study setting inpatient Number of participants recruited 1045 Sampling method consecutive participants Participants Female participants (percentage), 41.4 Age measure, value mean (standard deviation) 66.3 (16.0) Inclusion criteria COVID‐19 positive  Exclusion criteria NR Smoking (absolute number), 34 Diabetes (absolute number), 264 Hypertension (absolute number), 548 Cardiovascular diseases NR Please indicate if additional information is available  NR Asthma NR Chronic obstructive pulmonary disease NR Other pulmonary diseases NR Please indicate if additional information is available  NR Immunosuppression (absolute number), 48 Please indicate if additional information is available NR Chronic kidney disease (absolute number), 117 Cancer (absolute number), 109 Steroid administration NR Supplemental oxygen NR Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity patients had BMI ≥ 30 kg/m2 The time when obesity has been measured some time after presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency absolute number Frequency value 351 How many eligible outcomes reported? 2 How many eligible outcomes reported? 2 Outcome(s) mortality Outcome (prognostic factor) Mortality (BMI ≥ 30 kg/m2) Outcome Mortality Prognostic factor (category):  BMI ≥ 30 kg/m2 Follow‐up Number of patients followed completely for this outcome  1045 Number of obese patients followed completely for this outcome  351 Number of non‐obese patients followed completely for this outcome  236 Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 0.8 (0.6, 1.2) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, DM, HTN, sex, BMI, chronic lung disease, immunosuppression, chronic kidney disease, fever (≥ 38°C), dyspnoea, headache, lymphocytes count < 1000, neutrophil count ≥ 8000, CRP, AST Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.4 (0.7, 2.5) Outcome (prognostic factor) Severe COVID (BMI ≥ 30 kg/m2) Outcome Severe COVID Prognostic factor (category):  BMI ≥ 30 kg/m2 Follow‐up Number of patients followed completely for this outcome  1045 Number of obese patients followed completely for this outcome  351 Number of non‐obese patients followed completely for this outcome  236 Univariable (unadjusted) analysis for obesity Effect measure for obesity hazard ratio Effect measure value (95% CI) 1.6 (1.2, 2.0) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, DM, HTN,  sex, BMI, chronic lung disease, fever (≥ 38°C), dyspnoea, headache, lymphocytes count < 1000, neutrophil count ≥ 8000, CRP, AST Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.2 (1.5, 3.3)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Severe COVID	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Confounding Bias Severe COVID	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Kalligeros 2020.

Study characteristics
Notes	English title Association of obesity with disease severity among patients with coronavirus disease 2019 Study setting Start of study recruitment (MM/YYYY): 02/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres/clinics/areas: 2 Study setting: Inpatient Number of participants recruited: 103 Sampling method: Consecutive participants Participants Female participants (absolute number): 40 Age measure, value: Median (IQR), 60 (50‐72) Inclusion criteria: All consecutive adult patients (≥ 18 years old) who had a laboratory‐confirmed (using a reverse transcriptase–polymerase chain reaction assay) SARS‐CoV‐2 infection and who were admitted to Rhode Island Hospital, the Miriam Hospital, or Newport Hospital in Rhode Island between February 17 and April 5, 2020 Exclusion criteria: NR Smoking frequency: 48 Diabetes frequency: 38 Hypertension frequency: 66 Cardiovascular disease frequency: 25 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 20 Immunosuppression frequency: NR Chronic kidney disease frequency: 11 Cancer frequency: 9 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 30 kg/m2 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition of obesity: 30 Obesity frequency (absolute number): 49 Prognostic factor(s): BMI 30‐34.9 kg/m2, BMI ≥ 35 kg/m2 Outcome(s) ICU admission, mechanical ventilation Outcome (prognostic factor) ICU admission (BMI 30‐34.9 kg/m2) Follow‐up Number of patients followed completely for the outcome: 103 Number of obese patients followed completely for the outcome: 49 Number of non‐obese patients followed completely for the outcome: 54 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 2.8 (0.75, 10.48), 0.126 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, race, gender, BMI, diabetes, hypertension, heart disease, and chronic lung disease Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 2.65 (0.64, 10.95), 0.178 Outcome (prognostic factor) ICU admission (BMI ≥ 35 kg/m2) Follow‐up Number of patients followed completely for the outcome: 103 Number of obese patients followed completely for the outcome: 49 Number of non‐obese patients followed completely for the outcome: 54 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 3.02 (0.85, 10.74), 0.088 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, race, gender, BMI, diabetes, hypertension, heart disease, and chronic lung disease Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 5.39 (1.13, 25.64), 0.034 Outcome (prognostic factor) Mechanical ventilation (BMI 30‐34.9 kg/m2) Follow‐up Number of patients followed completely for the outcome: 103 Number of obese patients followed completely for the outcome: 49 Number of non‐obese patients followed completely for the outcome: 54 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 4.86 (0.88, 26.68), 0.069 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, race, gender, BMI, diabetes, hypertension, heart disease, and chronic lung disease Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 6.85 (1.05, 44.82), 0.045 Outcome (prognostic factor) Mechanical ventilation (BMI ≥ 35 kg/m2) Follow‐up Number of patients followed completely for the outcome: 103 Number of obese patients followed completely for the outcome: 49 Number of non‐obese patients followed completely for the outcome: 54 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 5.84 (1.12, 30.55), 0.036 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, race, gender, BMI, diabetes, hypertension, heart disease, and chronic lung disease Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 9.99 (1.39, 71.69), 0.022
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Kammar‐García 2020.

Study characteristics
Notes	English title Impact of comorbidities in Mexican SARS‐CoV‐2‐positive patients: a retrospective analysis in a national cohort Study setting Start of study recruitment (MM/YYYY) 01/2020 End of study recruitment (MM/YYYY) 04/2020 Study design registry data Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting outpatient and inpatient Number of participants recruited 13,842 Sampling method consecutive participants Participants Female participants (percentage), 42.3 Age measure, value mean (standard deviation), 46.6 (15.6) Inclusion criteria COVID‐19 positive  Exclusion criteria NR Smoking NR Diabetes NR Hypertension NR Cardiovascular diseases NR Please indicate if additional information is available  NR Asthma NR Chronic obstructive pulmonary disease NR Other pulmonary diseases NR Please indicate if additional information is available  NR Immunosuppression NR Please indicate if additional information is available NR Chronic kidney disease NR Cancer NR Steroid administration NR Supplemental oxygen NR Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity NR The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity NR Threshold used for definition of obesity  NR Measure of frequency absolute number Frequency value 2793 How many eligible outcomes reported? 5 How many eligible outcomes reported? 5 Outcome(s) mortality Outcome (prognostic factor) mortality (obesity) Outcome Mortality Prognostic factor (category):  obesity Follow‐up Number of patients followed completely for this outcome  13,842 Number of obese patients followed completely for this outcome  2793 Number of non‐obese patients followed completely for this outcome  11,049 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.8 (1.6, 2.01) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment sex, age, smoking status, and time from onset of symptoms to initial care Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.8 (1.6, 2.1) Outcome (prognostic factor) ICU admission (obesity) Outcome ICU admission Prognostic factor (category):  obesity Follow‐up Number of patients followed completely for this outcome  13,842 Number of obese patients followed completely for this outcome  2793 Number of non‐obese patients followed completely for this outcome  11,049 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.7 (1.4, 1.9) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment sex, age, smoking status, and time from onset of symptoms to initial care Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.7 (1.4, 2.01) Outcome (prognostic factor) pneumonia (obesity) Outcome pneumonia Prognostic factor (category):  obesity Follow‐up Number of patients followed completely for this outcome  13,842 Number of obese patients followed completely for this outcome  2793 Number of non‐obese patients followed completely for this outcome  11049 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.6 (1.4, 1.7) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment sex, age, smoking status, and time from onset of symptoms to initial care Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.6 (1.4, 1.7) Outcome (prognostic factor) hospitalisation (obesity) Outcome hospitalisation Prognostic factor (category):  obesity Follow‐up Number of patients followed completely for this outcome  13,842 Number of obese patients followed completely for this outcome  2793 Number of non‐obese patients followed completely for this outcome  11,049 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.6 (1.5, 1.7) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment sex, age, smoking status, and time from onset of symptoms to initial care Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.6 (1.4, 1.7) Outcome (prognostic factor) mechanical ventilation (obesity) Outcome mechanical ventilation Prognostic factor (category) obesity Follow‐up Number of patients followed completely for this outcome  13,842 Number of obese patients followed completely for this outcome  2793 Number of non‐obese patients followed completely for this outcome  11,049 Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.6 (1.3, 1.9) Multivariable (adjusted) analysis for obesity Modelling method Cox regression The set of prognostic factors used for adjustment sex, age, smoking status, and time from onset of symptoms to initial care Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.7 (1.4, 2.01)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	No	Appendix 3
Study Attrition Mechanical ventilation	No	Appendix 3
Study Attrition ICU admission	No	Appendix 3
Study Attrition Hospitalisation	No	Appendix 3
Study Attrition Pneumonia	No	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Unclear	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Outcome Measurement Pneumonia	No	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Confounding Bias Mechanical ventilation	Unclear	Appendix 3
Confounding Bias ICU admission	Unclear	Appendix 3
Confounding Bias Hospitalisation	Unclear	Appendix 3
Confounding Bias Pneumonia	Unclear	Appendix 3
Statistical Analysis Bias	No	Appendix 3

Khawaja 2020.

Study characteristics
Notes	English title Associations with COVID‐19 hospitalisation amongst 406,793 adults: the UK Biobank prospective cohort study Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 04/2020 Study design prospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas NR Study setting outpatient Number of participants recruited 406,793 Sampling method unspecified Participants Female participants (percentage), 55 Age measure, value mean (standard deviation), 68 (8) Inclusion criteria Individuals resident in England and alive in 2020 from UK Biobank Exclusion criteria We excluded participants that were tested but without a positive COVID‐19 test in case a proportion were false negatives and given the abundance of controls already available. Participants who died before 2020 or did not attend an assessment centre in England were excluded given they could not become cases. Smoking (absolute number), 40,181 Diabetes (absolute number), 19,897 Hypertension (absolute number), 135,604 Cardiovascular diseases (absolute number), 32,831 Please indicate if additional information is available  Ischaemic heart disease Asthma (absolute number), 55,127 Chronic obstructive pulmonary disease (absolute number), 13,805 Other pulmonary diseases (absolute number), 5377 Please indicate if additional information is available  Obstructive sleep apnoea Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Not specified The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  Not specified Measure of frequency unspecified Frequency value NR How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) hospitalisation Outcome (prognostic factor) Hospitalisation (BMI ≥ 25, < 30) Outcome Hospitalisation Prognostic factor (category):  BMI ≥ 25, < 30 Follow‐up Number of patients followed completely for this outcome  406,793 Number of obese patients followed completely for this outcome  94,690 Number of non‐obese patients followed completely for this outcome  312,103 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, sex, ethnicity, education level, Townsend deprivation index, BMI, DBP, alcohol intake frequency, smoking, loop diuretics use, HTN, IHD, stroke, COPD Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.26 (1.01, 1.56) Outcome (prognostic factor) Hospitalisation (BMI ≥ 30, < 35) Outcome Hospitalisation Prognostic factor (category):  BMI ≥ 30, < 35 Follow‐up Number of patients followed completely for this outcome  406,793 Number of obese patients followed completely for this outcome  94,690 Number of non‐obese patients followed completely for this outcome  312,103 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, sex, ethnicity, education level, Townsend deprivation index, BMI, DBP, alcohol intake frequency, smoking, loop diuretics use, HTN, IHD, stroke, COPD Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.37 (1.06, 1.76) Outcome (prognostic factor) Hospitalisation (BMI > 35) Outcome Hospitalisation Prognostic factor (category):  BMI > 35 Follow‐up Number of patients followed completely for this outcome  406,793 Number of obese patients followed completely for this outcome  94,690 Number of non‐obese patients followed completely for this outcome  312,103 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, sex, ethnicity, education level, Townsend deprivation index, BMI, DBP, alcohol intake frequency, smoking, loop diuretics use, HTN, IHD, stroke, COPD Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.04 (1.5, 2.77)
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Killerby 2020.

Study characteristics
Notes	English title Characteristics associated with hospitalization among patients with COVID‐19 ‐ metropolitan Atlanta, Georgia, March‐April 2020  Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres/clinics/areas: Six acute care hospitals and associated outpatient clinics  Study setting: Outpatient and inpatient Number of participants recruited: 531 Sampling method: Consecutive participants Participants Female participants (absolute number): 303 Age measure, value: NR Inclusion criteria: Hospitalised patients aged ≥ 18 years who were hospitalised with laboratory‐confirmed COVID‐19 (defined as a positive real‐time reverse transcription–polymerase chain reaction [RT‐PCR] test result for SARS‐CoV‐2) during March 1–30 and non‐hospitalised patients aged ≥ 18 years with laboratory‐confirmed COVID‐19 during March 1–April 7 Exclusion criteria: Persons lacking a healthcare visit during which a medical history could be recorded were excluded from analyses and if they stayed for observation or died in an ED. And persons lacking a healthcare visit during which a medical history could be recorded were excluded from analyses. Smoking frequency: 91 Diabetes frequency: 111 Hypertension frequency: 243 Cardiovascular disease frequency: 20 Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: 101 Immunosuppression frequency: 15 Chronic kidney disease frequency: 45 Cancer frequency: 34 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 30 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition of obesity: 30 Obesity frequency (absolute number): 227 Prognostic factor(s): BMI ≥ 30 kg/m2 Outcome(s) Hospitalisation Outcome (prognostic factor) Hospitalisation (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 531 Number of obese patients followed completely for the outcome: 227 Number of non‐obese patients followed completely for the outcome: 209 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: 1.82 (1.2, 2.57), NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: NR Effect measure for obesity: Odds ratio  Effect measure value (95% CI), P value: 1.9 (1.1, 3.3), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Hospitalisation	Unclear	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Hospitalisation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Kim 2020a.

Study characteristics
Notes	English title Risk factors for intensive care unit admission and in‐hospital mortality among hospitalized adults identified through the US Coronavirus Disease 2019 (COVID‐19)‐associated Hospitalization Surveillance Network (COVID‐NET) Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 05/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres/clinics/areas: 154 acute‐care hospitals in 74 counties in 13 states. Study setting: Inpatient Number of participants recruited: 2491 Sampling method: Consecutive participants Participants Female participants (absolute number): 1165 Age measure, value: Median (IQR), 62 (50‐75) Inclusion criteria: Eligible COVID‐19–associated hospitalisations occurred amongst persons who (1) resided in a predefined surveillance catchment area; and (2) had a positive SARS‐CoV‐2 test within 14 days prior to or during hospitalisation; included adults hospitalised within 154 acute‐care hospitals in 74 counties in 13 states with an admission date during 1 March–2 May 2020 who had either been discharged from the hospital or died during hospitalisation and had complete medical chart abstractions Exclusion criteria: children < 18 years of age due to small counts (n = 101) and 1 surveillance site (Iowa) for which medical chart abstractions were not conducted, also excluded patients who were still hospitalised at the time of this analysis and all patients for whom medical chart abstractions had not yet been completed as of 2 May 2020 Smoking frequency: 792 Diabetes frequency: 819 Hypertension frequency: 1428 Cardiovascular disease frequency: 859 Asthma frequency: 314 Chronic obstructive pulmonary disease frequency: 266 Other pulmonary disease frequency: 747 Immunosuppression frequency: 263 Chronic kidney disease frequency: NR Cancer frequency: 101 Steroid administration frequency: 106 Supplemental oxygen administration frequency: 814 Other treatments (frequency): Hydroxychloroquine (1065), azithromycin (725), tocilizumab (103), atazanavir (94), remdesivir (53), lopinavir/ritonavir (27), convalescent plasma (9), chloroquine (7), sarilumab (6), zinc (6) Prognostic factor(s) Study’s definition for obesity: Body mass index ≥ 30 kg/m2 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition of obesity: 30 Obesity frequency (absolute number): 1154 Prognostic factor(s): BMI ≥ 30 kg/m2 Outcome(s) In‐hospital mortality, ICU admission Outcome (prognostic factor) In‐hospital mortality (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 2490 Number of obese patients followed completely for the outcome: 1154 Number of non‐obese patients followed completely for the outcome: 1178 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 0.73 (0.61, 0.86), 0.001 Multivariable analysis for obesity Modelling method: Log‐linked Poisson The set of prognostic factors used for adjustment: Age, sex, race/ethnicity, smoking status, hypertension, obesity, diabetes, chronic lung disease, cardiovascular disease, neurologic disease, renal disease, immunosuppression, and outpatient use of an angiotensin receptor blocker Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.09 (1.30, 0.92), NR Outcome (prognostic factor) ICU admission (BMI ≥ 30 kg/m2) Follow‐up Number of patients followed completely for the outcome: 2490 Number of obese patients followed completely for the outcome: 1154 Number of non‐obese patients followed completely for the outcome: 1178 Univariable unadjusted analysis for obesity Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.25 (1.14, 1.37), 0.0013 Multivariable analysis for obesity Modelling method: Log‐linked Poisson The set of prognostic factors used for adjustment: Age, sex, race/ethnicity, smoking status, hypertension, obesity, diabetes, chronic lung disease, cardiovascular disease, neurologic disease, renal disease, immunosuppression, and outpatient use of an angiotensin receptor blocker Effect measure for obesity: Relative risk Effect measure value (95% CI), P value: 1.31 (1.16, 1.47), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias ICU admission	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Kim 2020b.

Study characteristics
Notes	English title Analysis of mortality and morbidity in COVID‐19 patients with obesity using clinical epidemiological data from the Korean Center for Disease Control & Prevention Study setting Start of study recruitment (MM/YYYY): NR End of study recruitment (MM/YYYY): 06/2020 Study design: Registry data Study centre(s): NR Number of centres, clinics or areas: NR Study setting: NR  Number of participants recruited: 4027 Sampling method: Consecutive participants Participants Female participants (absolute number): 2334 Age measure, value: NR Inclusion criteria: Patients with confirmed cases of COVID‐19 was released from isolation after achieving a complete recovery. Asymptomatic patients were determined to be completely recovered if the PCR results were negative two consecutive times with at least a 24‐h interval between them at least seven days after a definitive diagnosis had been made. Symptomatic patients with confirmed cases were determined to be completely recovered if they had no fever without taking antipyretic drugs, the clinical manifestations were improved and the PCR results were negative two consecutive times with at least a 24‐h interval between them at least seven days after a definitive diagnosis had been made. Exclusion criteria: Participants who did not have records of symptoms or past medical histories were excluded. Participants who did not have recorded BMI values were excluded. Smoking frequency: NR Diabetes frequency: 492 Hypertension frequency: 829 Cardiovascular disease frequency: 172 Asthma frequency: 96 Chronic obstructive pulmonary disease frequency: 30 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 43 Cancer frequency: 107 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: BMI ≥ 25 The time when obesity has been measured: NR Main variable used for determination of obesity: BMI Threshold used for definition: 25 Obesity frequency (absolute number): 1159 Prognostic factor(s): BMI ≥ 25 kg/m2 Outcome(s) Mortality Severe COVID Outcome (prognostic factor) Mortality (BMI ≥ 25 kg/m2) Follow‐up Number of patients followed completely for the outcome: 4027 Number of obese patients followed completely for the outcome: 1159 Number of non‐obese patients followed completely for the outcome: 2868 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.36 (0.9, 2.05), 0.149 Multivariable analysis for obesity Modelling method: Cox regression The set of prognostic factors used for adjustment: Age, sex, obesity, systolic blood pressure, diastolic blood pressure, heart rate, temperature, diabetes, hypertension, heart failure, chronic heart disease, asthma, chronic obstructive pulmonary disease, chronic kidney disease, cancer, chronic liver disease, rheumatic or autoimmune disease, dementia Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.71 (1.1, 2.66), 0.017 Outcome (prognostic factor) Severe COVID (BMI ≥ 25 kg/m2) Follow‐up Number of patients followed completely for the outcome: 4027 Number of obese patients followed completely for the outcome: 1159 Number of non‐obese patients followed completely for the outcome: 2868 Univariable unadjusted analysis for obesity Effect measure for obesity: Hazard ratio Effect measure value (95% CI), P value: 1.72 (1.39, 2.12), < 0.001 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, obesity, systolic blood pressure, diastolic blood pressure, heart rate, temperature, diabetes, hypertension, heart failure, chronic heart disease, asthma, chronic obstructive pulmonary disease, chronic kidney disease, cancer, chronic liver disease, rheumatic or autoimmune disease, dementia Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.71 (1.32, 2.21), < 0.001
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	No	Appendix 3
Study Attrition Severe COVID	No	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Severe COVID	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Kim 2020c.

Study characteristics
Notes	English title BMI as a risk factor for clinical outcomes in patients hospitalized with COVID‐19 in New York Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: NR Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 12 Study setting: Inpatient Number of participants recruited: 10,861 Sampling method: Consecutive participants Participants Female participants (absolute number): 4393 Age measure, value: Median (IQR), 65 (54‐77) Inclusion criteria: All adult patients admitted to 12 Northwell Health system acute‐care hospitals in New York between March 1, 2020, and April 27, 2020, with a confirmed diagnosis of COVID‐19 by a polymerase chain reaction of nasopharyngeal swabs Exclusion criteria: NR Smoking frequency: 1797 (including ex‐smokers) Diabetes frequency: 3995 Hypertension frequency: 6555 Cardiovascular disease frequency: 2379 Asthma frequency: 903 Chronic obstructive pulmonary disease frequency: 677 Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 515 Cancer frequency: 937 Steroid administration frequency: NR Supplemental oxygen administration frequency: NR Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Obesity class I (30‐34.9 kg/m2), obesity class II (35‐39.9 kg/m2), and obesity class III (≥ 40 kg/m2) The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 4090 Prognostic factor(s): 30 < BMI < 35 35 < BMI < 40 BMI > 40 Outcome(s) Invasive mechanical ventilation Mortality Outcome (prognostic factor) Invasive mechanical ventilation (30 < BMI < 35) Follow‐up Number of patients followed completely for the outcome: 10,861 Number of obese patients followed completely for the outcome: 4090 Number of non‐obese patients followed completely for the outcome: 6771 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Multivariable regression models, adjusting for patient characteristics and also secondary analysis using a Cox proportional‐hazards model The set of prognostic factors used for adjustment: Age, sex, race, hypertension, coronary artery disease, diabetes mellitus, heart failure, chronic kidney disease, end‐stage renal disease, cancer, asthma, COPD, smoking status, hospital type Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.48 (1.27, 1.72), NR Outcome (prognostic factor) Invasive mechanical ventilation (35 < BMI < 40) Follow‐up Number of patients followed completely for the outcome: 10,861 Number of obese patients followed completely for the outcome: 4090 Number of non‐obese patients followed completely for the outcome: 6771 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Multivariable regression models, adjusting for patient characteristics and also secondary analysis using a Cox proportional‐hazards model The set of prognostic factors used for adjustment: Age, sex, race, hypertension, coronary artery disease, diabetes mellitus, heart failure, chronic kidney disease, end‐stage renal disease, cancer, asthma, COPD, smoking status, hospital type Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.89 (1.56, 2.28), NR Outcome (prognostic factor) Invasive mechanical ventilation (BMI > 40) Follow‐up Number of patients followed completely for the outcome: 10,861 Number of obese patients followed completely for the outcome: 4090 Number of non‐obese patients followed completely for the outcome: 6771 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Multivariable regression models, adjusting for patient characteristics and also secondary analysis using a Cox proportional‐hazards model The set of prognostic factors used for adjustment: Age, sex, race, hypertension, coronary artery disease, diabetes mellitus, heart failure, chronic kidney disease, end‐stage renal disease, cancer, asthma, COPD, smoking status, hospital type Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 2.31 (1.88, 2.85), NR Outcome (prognostic factor) Mortality (30 < BMI < 35) Follow‐up Number of patients followed completely for the outcome: 10,861 Number of obese patients followed completely for the outcome: 4090 Number of non‐obese patients followed completely for the outcome: 6771 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Multivariable regression models, adjusting for patient characteristics and also secondary analysis using a Cox proportional‐hazards model The set of prognostic factors used for adjustment: Age, sex, race, hypertension, coronary artery disease, diabetes mellitus, heart failure, chronic kidney disease, end‐stage renal disease, cancer, asthma, COPD, smoking status, hospital type Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.00 (0.87, 1.16), NR Outcome (prognostic factor) Mortality (35 < BMI < 40) Follow‐up Number of patients followed completely for the outcome: 10,861 Number of obese patients followed completely for the outcome: 4090 Number of non‐obese patients followed completely for the outcome: 6771 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Multivariable regression models, adjusting for patient characteristics and also secondary analysis using a Cox proportional‐hazards model The set of prognostic factors used for adjustment: Age, sex, race, hypertension, coronary artery disease, diabetes mellitus, heart failure, chronic kidney disease, end‐stage renal disease, cancer, asthma, COPD, smoking status, hospital type Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.25 (1.03, 1.52), NR Outcome (prognostic factor) Mortality (BMI > 40) Follow‐up Number of patients followed completely for the outcome: 10,861 Number of obese patients followed completely for the outcome: 4090 Number of non‐obese patients followed completely for the outcome: 6771 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Multivariable regression models, adjusting for patient characteristics and also secondary analysis using a Cox proportional‐hazards model The set of prognostic factors used for adjustment: Age, sex, race, hypertension, coronary artery disease, diabetes mellitus, heart failure, chronic kidney disease, end‐stage renal disease, cancer, asthma, COPD, smoking status, hospital type Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.61 (1.30, 2.00), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Mechanical ventilation	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Klang 2020.

Study characteristics
Notes	English title Severe obesity as an independent risk factor for COVID‐19 mortality in hospitalized patients younger than 50 Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 05/2020 Study design: Registry data Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 5 Study setting: Inpatient Number of participants recruited: 572 (cohort 1), 2834 (cohort 2) Sampling method: Consecutive participants Participants Female participants (absolute number): 175 (cohort 1), 1270 (cohort 2) Age measure, value: NR Inclusion criteria: All adult patients who were positive for COVID‐19 by nasopharyngeal swab polymerase chain reaction test and were admitted to the hospital. Patients who were discharged or had died during the study period were included. Exclusion criteria: Patients who were still hospitalised at the time of analysis and patients with missing BMI data Smoking frequency: 76 (cohort 1), 717 (cohort 2) Diabetes frequency: 153 (cohort 1), 1446 (cohort 2) Hypertension frequency: 175 (cohort 1), 2124 (cohort 2) Cardiovascular disease frequency: 67 (cohort 1), 1190 (cohort 2) Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: 70 (cohort 1), 597 (cohort 2) Cancer frequency: 39 (cohort 1), 491 (cohort 2) Steroid administration frequency: NR Supplemental oxygen administration frequency: 79 (cohort 1), 730 (cohort 2) Other treatments (frequency): NR Prognostic factor(s) Study’s definition for obesity: Obesity was defined as BMI ≥ 30 kg/m2; obesity groups included the following: BMI of 30 to < 40 and BMI ≥ 40 The time when obesity has been measured: Before disease or right at presentation Main variable used for determination of obesity: BMI Threshold used for definition: 30 Obesity frequency (absolute number): 275 (cohort 1), 956 (cohort 2) Prognostic factor(s): BMI of 30 to < 40 BMI ≥ 40 Outcome(s) Mortality Mechanical ventilation Outcome (prognostic factor) Mortality (BMI of 30 to < 40) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 572 Number of obese patients followed completely for the outcome: 275 Number of non‐obese patients followed completely for the outcome: 297 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR, 0.313 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age decile, male sex, CAD, CHF, HTN, DM, hyperlipidaemia, CKD, history of cancer, smoking (past or present), BMI 30‐40, BMI ≥ 40, and race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.1 (0.5, 2.3), 0.755 Outcome (prognostic factor) Mortality (BMI ≥ 40) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 572 Number of obese patients followed completely for the outcome: 275 Number of non‐obese patients followed completely for the outcome: 297 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR, < 0.001 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age decile, male sex, CAD, CHF, HTN, DM, hyperlipidaemia, CKD, history of cancer, smoking (past or present), BMI 30‐40, BMI ≥ 40, and race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 5.1 (2.3, 11.1), < 0.001 Outcome (prognostic factor) Mortality (BMI of 30 to < 40) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 2834 Number of obese patients followed completely for the outcome: 956 Number of non‐obese patients followed completely for the outcome: 1878 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR, 0.117 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age decile, male sex, CAD, CHF, HTN, DM, hyperlipidaemia, CKD, history of cancer, smoking (past or present), BMI 30‐40, BMI ≥ 40, and race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.1 (0.9, 1.3), 0.421 Outcome (prognostic factor) Mortality (BMI ≥ 40) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 2834 Number of obese patients followed completely for the outcome: 956 Number of non‐obese patients followed completely for the outcome: 1878 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR, 0.532 Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age decile, male sex, CAD, CHF, HTN, DM, hyperlipidaemia, CKD, history of cancer, smoking (past or present), BMI 30‐40, BMI ≥ 40, and race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.6 (1.2, 2.3), 0.004 Outcome (prognostic factor) Mechanical ventilation (BMI of 30 to < 40) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 572 Number of obese patients followed completely for the outcome: 275 Number of non‐obese patients followed completely for the outcome: 297 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age decile, male sex, CAD, CHF, HTN, DM, hyperlipidaemia, CKD, history of cancer, smoking (past or present), BMI 30‐40, BMI ≥ 40, and race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.5 (0.8, 2.7), 0.2 Outcome (prognostic factor) Mechanical ventilation (BMI ≥ 40) (cohort 1) Follow‐up Number of patients followed completely for the outcome: 572 Number of obese patients followed completely for the outcome: 275 Number of non‐obese patients followed completely for the outcome: 297 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age decile, male sex, CAD, CHF, HTN, DM, hyperlipidaemia, CKD, history of cancer, smoking (past or present), BMI 30‐40, BMI ≥ 40, and race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.5 (1.1, 2.1), 0.025 Outcome (prognostic factor) Mechanical ventilation (BMI of 30 to < 40) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 2834 Number of obese patients followed completely for the outcome: 956 Number of non‐obese patients followed completely for the outcome: 1878 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age decile, male sex, CAD, CHF, HTN, DM, hyperlipidaemia, CKD, history of cancer, smoking (past or present), BMI 30‐40, BMI ≥ 40, and race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.3 (1, 1.6), 0.016 Outcome (prognostic factor) Mechanical ventilation (BMI ≥ 40) (cohort 2) Follow‐up Number of patients followed completely for the outcome: 2834 Number of obese patients followed completely for the outcome: 956 Number of non‐obese patients followed completely for the outcome: 1878 Univariable unadjusted analysis for obesity Effect measure for obesity: NR Effect measure value (95% CI), P value: NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age decile, male sex, CAD, CHF, HTN, DM, hyperlipidaemia, CKD, history of cancer, smoking (past or present), BMI 30‐40, BMI ≥ 40, and race Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.5 (1.1, 2.1), 0.025
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	No	Appendix 3
Study Attrition Mechanical ventilation	No	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Confounding Bias Mortality	Yes	Appendix 3
Confounding Bias Mechanical ventilation	Yes	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Klineova 2020.

Study characteristics
Notes	English title Covid‐19 infection in patients with multiple sclerosis: an observational study by the New York COVID‐19 Neuroimmunology Consortium (NYCNIC) Study setting Start of study recruitment (MM/YYYY) NR End of study recruitment (MM/YYYY) NR Study design prospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 5 Study setting outpatient and inpatient Number of participants recruited 349 Sampling method unspecified Participants Female participants (percentage), 70.8 Age measure, value median (range), 45 (13, 76) Inclusion criteria Patients with MS or related disorders, who self‐identified as diagnosed with COVID‐19 by a healthcare provider (based on characteristic symptoms, radiographic findings and/or positive COVID‐19 PCR/serology when available) were included. Exclusion criteria NR Smoking NR Diabetes (unspecified) Hypertension (unspecified) Cardiovascular diseases (unspecified) Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity NR The time when obesity has been measured unspecified Main variable used for determination of obesity NR Threshold used for definition of obesity  NR Measure of frequency NR Frequency value NR How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) hospitalisation Outcome (prognostic factor) Hopitalisation (obesity) Outcome Hospitalisation Prognostic factor (category):  Obesity Follow‐up Number of patients followed completely for this outcome  349 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment Age, obesity, EDSS, race, ethnicity, comorbidities (cardiac, pulmonary, diabetes), smoking status, specific DMT Effect measure for obesity odds ratio  Effect measure value (95% CI) 2.4 (1.1, 4.9)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Hospitalisation	Yes	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Hospitalisation	Unclear	Appendix 3
Statistical Analysis Bias	Unclear	Appendix 3

Kompaniyets 2021.

Study characteristics
Notes	English title Body mass index and risk for COVID‐19–related hospitalization, intensive care unit admission, invasive mechanical ventilation, and death — United States, March–December 2020 Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 12/2020 Study design registry data Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 238 Study setting outpatient and inpatient Number of participants recruited 148,494 Sampling method unspecified Participants Female participants (absolute number), 79,624 Age measure, value median (interquartile range), 55 (38, 70) Inclusion criteria aged >= 18 years with measured height and weight and an ED or inpatient encounter with an International Classification of Diseases, Tenth Revision, Clinical Modification (ICD‐10‐CM) code of U07.1 (COVID‐19, virus identified) during April 1–December 31, 2020, or B97.29 (other coronavirus as the cause of diseases classified elsewhere; recommended before April 2020 release of U07.1) Exclusion criteria Heights and weights were excluded if they were substantially larger or smaller than expected (defined as height < 44 inches [112 cm] or > 90 inches [229 cm]; weight < 25 kg [55 lbs] or > 454 kg [1000 lbs]; and BMI < 12 kg/m2 or > 110 kg/m2 Smoking NR Diabetes (unspecified) Hypertension (unspecified) Cardiovascular diseases (unspecified) Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (unspecified) Differential values for various oxygenation methods (if indicated) NR Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity Obesity (body mass index ≥ 30 kg/m2) is frequently categorised into three categories: class 1 (30.0–34.9 kg/m2), class 2 (35.0–39.9 kg/m2), and class 3 (≥ 40 kg/m2). Class 3 obesity is sometimes referred to as “extreme” or “severe” obesity. The time when obesity has been measured before disease or right at presentation Main variable used for determination of obesity BMI Threshold used for definition of obesity  30 Measure of frequency absolute number Frequency value 75,498 How many eligible outcomes reported? 4 How many eligible outcomes reported? 4 Outcome(s) hospitalisation, ICU admission, mechanical ventilation, mortality Outcome (prognostic factor) Hospitalisation (BMI 30‐34.9) Outcome Hospitalisation Prognostic factor (category):  BMI 30‐34.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.03 (1.01, 1.05) Outcome (prognostic factor) Hospitalisation (BMI 35‐39.9) Outcome Hospitalisation Prognostic factor (category):  BMI 35‐39.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.06 (1.04, 1.08) Outcome (prognostic factor) Hospitalisation (BMI 40‐44.9) Outcome Hospitalisation Prognostic factor (category):  BMI 40‐44.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.11 (1.08, 1.13) Outcome (prognostic factor) Hospitalisation (BMI ≥ 45) Outcome Hospitalisation Prognostic factor (category):  BMI ≥ 45 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.2 (1.17, 1.23) Outcome (prognostic factor) ICU admission (BMI 30‐34.9) Outcome ICU admission Prognostic factor (category) BMI 30‐34.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 0.99 (0.96, 1.01) Outcome (prognostic factor) ICU admission (BMI 35‐39.9) Outcome ICU admission Prognostic factor (category):  BMI 35‐39.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.01 (1.04, 0.98) Outcome (prognostic factor) ICU admission (BMI 40‐44.9) Outcome ICU admission Prognostic factor (category):  BMI 40‐44.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.04 (1, 1.07) Outcome (prognostic factor) ICU admission (BMI ≥ 45) Outcome ICU admission Prognostic factor (category):  BMI ≥ 45 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.12 (1.08, 1.16) Outcome (prognostic factor) Mechanical ventilation (BMI 30‐34.9) Outcome Mechanical ventilation Prognostic factor (category):  BMI 30‐34.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.31 (1.22, 1.41) Outcome (prognostic factor) Mechanical ventilation (BMI 35‐39.9) Outcome Mechanical ventilation Prognostic factor (category):  BMI 35‐39.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.45 (1.33, 1.57) Outcome (prognostic factor) Mechanical ventilation (BMI 40‐44.9) Outcome Mechanical ventilation Prognostic factor (category):  BMI 40‐44.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.62 (1.48, 1.77) Outcome (prognostic factor) Mechanical ventilation (BMI ≥ 45) Outcome Mechanical ventilation Prognostic factor (category):  BMI ≥ 45 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.95 (1.77, 2.16) Outcome (prognostic factor) Mortality (BMI 30‐34.9) Outcome Mortality Prognostic factor (category):  BMI 30‐34.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.04 (0.98, 1.1) Outcome (prognostic factor) Mortality (BMI 35‐39.9) Outcome Mortality Prognostic factor (category):  BMI 35‐39.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.07 (1, 1.15) Outcome (prognostic factor) Mortality (BMI 40‐44.9) Outcome Mortality Prognostic factor (category):  BMI 40‐44.9 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.24 (1.13, 1.36) Outcome (prognostic factor) Mortality (BMI ≥ 45) Outcome Mortality Prognostic factor (category):  BMI ≥ 45 Follow‐up Number of patients followed completely for this outcome  148,494 Number of obese patients followed completely for this outcome  75,498 Number of non‐obese patients followed completely for this outcome  72,996 Univariable (unadjusted) analysis for obesity Effect measure for obesity NR Effect measure value (95% CI) NR (NR, NR) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment adjusted for BMI category, underlying medical conditions (hypertension, diabetes, chronic kidney disease, asthma, coronary atherosclerosis and other heart disease, chronic obstructive pulmonary disease and bronchiectasis, and cancer), age, sex, race/ethnicity, payer type, hospital urbanicity, hospital census region, and admission month as controls Effect measure for obesity relative risk Effect measure value (95% CI) 1.48 (1.35, 1.62)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition Mortality	No	Appendix 3
Study Attrition Mechanical ventilation	No	Appendix 3
Study Attrition ICU admission	No	Appendix 3
Study Attrition Hospitalisation	No	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Mechanical ventilation	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Outcome Measurement Hospitalisation	Yes	Appendix 3
Confounding Bias Mortality	Unclear	Appendix 3
Confounding Bias Mechanical ventilation	Unclear	Appendix 3
Confounding Bias ICU admission	Unclear	Appendix 3
Confounding Bias Hospitalisation	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Krieger 2021.

Study characteristics
Notes	English title Emergency department characteristics and associations with intensive care admission among patients with coronavirus disease 2019 Study setting Start of study recruitment (MM/YYYY) 03/2020 End of study recruitment (MM/YYYY) 06/2020 Study design retrospective cohort Study centre(s) multiple centres/clinics/areas within a country Number of centres/clinics/areas 3 Study setting inpatient Number of participants recruited 330 Sampling method unspecified Participants Female participants (absolute number), 118 Age measure, value median (interquartile range), 65 (53, 76) Inclusion criteria Patients were included in the study if they presented to an ED and had laboratory‐confirmed severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection during the study period. Exclusion criteria Missing data were excluded from the analyses. Smoking NR Diabetes (unspecified) Hypertension (unspecified) Cardiovascular diseases (unspecified) Please indicate if additional information is available  NR Asthma (unspecified) Chronic obstructive pulmonary disease (unspecified) Other pulmonary diseases (unspecified) Please indicate if additional information is available  NR Immunosuppression (unspecified) Please indicate if additional information is available NR Chronic kidney disease (unspecified) Cancer (unspecified) Steroid administration (unspecified) Supplemental oxygen (absolute number) 40 Differential values for various oxygenation methods (if indicated) Nasal cannula (24), non‐rebreather mask (9), non‐invasive positive pressure ventilation (1), invasive mechanical ventilation (6) Other treatment NR Dose if applicable NR Duration if applicable NR Percentage received this treatment NR Prognostic factor(s) Study’s definition for obesity BMI > 29 The time when obesity has been measured unspecified Main variable used for determination of obesity BMI Threshold used for definition of obesity  29 Measure of frequency unspecified Frequency value NR How many eligible outcomes reported? 1 How many eligible outcomes reported? 1 Outcome(s) ICU admission Outcome (prognostic factor) ICU admission (obesity) Outcome ICU admission Prognostic factor (category):  Obesity Follow‐up Number of patients followed completely for this outcome  330 Number of obese patients followed completely for this outcome  NR Number of non‐obese patients followed completely for this outcome  NR Univariable (unadjusted) analysis for obesity Effect measure for obesity odds ratio  Effect measure value (95% CI) 1.01 (0.98, 1.04) Multivariable (adjusted) analysis for obesity Modelling method logistic regression The set of prognostic factors used for adjustment NR Effect measure for obesity NR Effect measure value (95% CI) 1.65 (1.04, 2.62)
Item	Authors' judgement	Support for judgement
Study Participation	Unclear	Appendix 3
Study Attrition ICU admission	Unclear	Appendix 3
Prognostic Factor Measurement	Yes	Appendix 3
Outcome Measurement ICU admission	Yes	Appendix 3
Confounding Bias ICU admission	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Kuderer 2020.

Study characteristics
Notes	English title Clinical impact of COVID‐19 on patients with cancer (CCC19): a cohort study Study setting Start of study recruitment (MM/YYYY): 03/2020 End of study recruitment (MM/YYYY): 04/2020 Study design: Registry data Study centre(s): International Number of centres, clinics or areas: NR Study setting: Outpatient Number of participants recruited: 928 Sampling method: Consecutive participants Participants Female participants (absolute number): 459 Age measure, value: Median (IQR), 66 (57‐76) Inclusion criteria: Patients who had baseline data entered onto the database between March 17 and April 16, 2020 and had follow‐up data entered up until May 7, 2020. Patients eligible for inclusion were adults (aged 18 years or older), with a diagnosed invasive or haematological malignancy at any time, and a resident of the USA, Canada, or Spain. Exclusion criteria: Patients with presumptive COVID‐19 who did not have a laboratory confirmed SARS‐CoV‐2 infection were excluded and patients with non‐invasive cancers including non‐melanomatous skin cancer, in‐situ carcinoma, or precursor haematological neoplasms were excluded from this analysis. Smoking frequency: 369 Diabetes frequency: NR Hypertension frequency: NR Cardiovascular disease frequency: NR Asthma frequency: NR Chronic obstructive pulmonary disease frequency: NR Other pulmonary disease frequency: NR Immunosuppression frequency: NR Chronic kidney disease frequency: NR Cancer frequency: 928 (396 were active) Steroid administration frequency: NR Supplemental oxygen administration frequency: 405 Other treatments (frequency): Hydroxychloroquine alone 89 (10%), azithromycin alone 93 (10%), azithromycin plus hydroxychloroquine 181 (20%) Prognostic factor(s) Study’s definition for obesity: NR The time when obesity has been measured: Some time after presentation Main variable used for determination of obesity: NR Threshold used for definition: NR Obesity frequency (absolute number): 172 Prognostic factor(s): Obesity Outcome(s) Mortality Severe COVID Outcome (prognostic factor) Mortality (obesity) Follow‐up Number of patients followed completely for the outcome: 928 Number of obese patients followed completely for the outcome: 172 Number of non‐obese patients followed completely for the outcome: 720 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds Ratio Effect measure value (95% CI), P value: 0.84 (0.5, 1.41), NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, smoking status, obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 0.99 (0.58, 1.71), NR Outcome (prognostic factor) Severe COVID (obesity) Follow‐up Number of patients followed completely for the outcome: 928 Number of obese patients followed completely for the outcome: 172 Number of non‐obese patients followed completely for the outcome: 720 Univariable unadjusted analysis for obesity Effect measure for obesity: Odds Ratio Effect measure value (95% CI), P value: 1.12 (0.77, 1.62), NR Multivariable analysis for obesity Modelling method: Logistic regression The set of prognostic factors used for adjustment: Age, sex, smoking status, obesity Effect measure for obesity: Odds ratio Effect measure value (95% CI), P value: 1.26 (0.8, 1.97), NR
Item	Authors' judgement	Support for judgement
Study Participation	Yes	Appendix 3
Study Attrition Mortality	Unclear	Appendix 3
Study Attrition Severe COVID	Unclear	Appendix 3
Prognostic Factor Measurement	No	Appendix 3
Outcome Measurement Mortality	Yes	Appendix 3
Outcome Measurement Severe COVID	Yes	Appendix 3
Confounding Bias Mortality	No	Appendix 3
Confounding Bias Severe COVID	No	Appendix 3
Statistical Analysis Bias	Yes	Appendix 3

Li 2020.

Study characteristics

Notes

English title Nutritional risk and therapy for severe and critical COVID‐19 patients: a multicenter retrospective observational study Study setting Start of study recruitment (MM/YYYY): 01/2020 End of study recruitment (MM/YYYY): 02/2020 Study design: Retrospective cohort Study centre(s): Multiple centres/clinics/areas within a country Number of centres, clinics or areas: 4 Study setting: Inpatient Number of participants recruited: 523 Sampling method: Consecutive participants Participants Female participants (absolute number): 273 Age measure, value: Mean (SD), 54.2 (15.9) Inclusion criteria: Severely ill patients if they met any of the following criteria: 1. respiratory distress and respiratory rate was