Effectiveness and tolerability of dual and triple combination inhaler therapies compared with each other and varying doses of inhaled corticosteroids in adolescents and adults with asthma: a systematic review and network meta‐analysis

Yuji Oba; Sumayya Anwer; Tinashe Maduke; Tarang Patel; Sofia Dias

doi:10.1002/14651858.CD013799.pub2

. 2022 Dec 6;2022(12):CD013799. doi: 10.1002/14651858.CD013799.pub2

Effectiveness and tolerability of dual and triple combination inhaler therapies compared with each other and varying doses of inhaled corticosteroids in adolescents and adults with asthma: a systematic review and network meta‐analysis

Yuji Oba ^1,^✉, Sumayya Anwer ², Tinashe Maduke ¹, Tarang Patel ¹, Sofia Dias ²

Editor: Cochrane Airways Group

PMCID: PMC9723963 PMID: 36472162

Abstract

Background

Current guidelines recommend a higher‐dose inhaled corticosteroids (ICS) or adding a long‐acting muscarinic antagonist (LAMA) when asthma is not controlled with medium‐dose (MD) ICS/long‐acting beta2‐agonist (LABA) combination therapy.

Objectives

To assess the effectiveness and safety of dual (ICS/LABA) and triple therapies (ICS/LABA/LAMA) compared with each other and with varying doses of ICS in adolescents and adults with uncontrolled asthma.

Search methods

We searched multiple databases for pre‐registered randomised controlled trials (RCTs) of at least 12 weeks of study duration from 2008 to 18 February 2022.

Selection criteria

We searched studies, including adolescents and adults with uncontrolled asthma who had been treated with, or were eligible for, MD‐ICS/LABA, comparing dual and triple therapies. We excluded cluster‐ and cross‐over RCTs.

Data collection and analysis

We conducted a systematic review and network meta‐analysis according to the previously published protocol. We used Cochrane’s Screen4ME workflow to assess search results and Grading of Recommendations Assessment, Development and Evaluation (GRADE) to assess the certainty of evidence. The primary outcome was steroid‐requiring asthma exacerbations and asthma‐related hospitalisations (moderate to severe and severe exacerbations).

Main results

We included 17,161 patients with uncontrolled asthma from 17 studies (median duration 26 weeks; mean age 49.1 years; male 40%; white 81%; mean forced expiratory volume in 1 second (MEF 1)1.9 litres and 61% predicted). The quality of included studies was generally good except for some outcomes in a few studies due to high attrition rates.

Medium‐dose (MD) and high‐dose (HD) triple therapies reduce steroid‐requiring asthma exacerbations (hazard ratio (HR) 0.84 [95% credible interval (CrI) 0.71 to 0.99] and 0.69 [0.58 to 0.82], respectively) (high‐certainty evidence), but not asthma‐related hospitalisations, compared to MD‐ICS/LABA.

High‐dose triple therapy likely reduces steroid‐requiring asthma exacerbations compared to MD triple therapy (HR 0.83 [95% CrI 0.69 to 0.996], [moderate certainty]). Subgroup analyses suggest the reduction in steroid‐requiring exacerbations associated with triple therapies may be only for those with a history of asthma exacerbations in the previous year but not for those without.

High‐dose triple therapy, but not MD triple, results in a reduction in all‐cause adverse events (AEs) and likely reduces dropouts due to AEs compared to MD‐ICS/LABA (odds ratio (OR) 0.79 [95% CrI 0.69 to 0.90], [high certainty] and 0.50 [95% CrI 0.30 to 0.84], [moderate certainty], respectively). Triple therapy results in little to no difference in all‐cause or asthma‐related serious adverse events (SAEs) compared to dual therapy (high certainty).

The evidence suggests triple therapy results in little or no clinically important difference in symptoms or quality of life compared to dual therapy considering the minimal clinically important differences (MCIDs) and HD‐ICS/LABA is unlikely to result in any significant benefit or harm compared to MD‐ICS/LABA.

Authors' conclusions

Medium‐dose and HD triple therapies reduce steroid‐requiring asthma exacerbations, but not asthma‐related hospitalisations, compared to MD‐ICS/LABA especially in those with a history of asthma exacerbations in the previous year. High‐dose triple therapy is likely superior to MD triple therapy in reducing steroid‐requiring asthma exacerbations.

Triple therapy is unlikely to result in clinically meaningful improvement in symptoms or quality of life compared to dual therapy considering the MCIDs.

High‐dose triple therapy, but not MD triple, results in a reduction in all‐cause AEs and likely reduces dropouts due to AEs compared to MD‐ICS/LABA. Triple therapy results in little to no difference in all‐cause or asthma‐related SAEs compared to dual therapy.

HD‐ICS/LABA is unlikely to result in any significant benefit or harm compared to MD‐ICS/LABA, although long‐term safety of higher rather than MD‐

ICS remains to be demonstrated given the median duration of included studies was six months.

The above findings may assist deciding on a treatment option when asthma is not controlled with MD‐ICS/LABA.

Plain language summary

What is triple inhaled therapy, when is it used, and what does it do in asthma?

How are inhalers used for the management of asthma?

Management of asthma involves a series of stepwise therapies depending on the severity of the disease. Initial therapy typically starts with as needed short‐acting inhaler therapy (step 1), and a daily low‐ to medium‐dose inhaled steroids is added for better asthma control when needed (step 2). Subsequently, a bronchodilator known as long‐acting beta₂‐agonist (LABA), which causes the passages of the airways to expand and relax so that breathing difficulty is reduced, is typically added to inhaled steroids if needed (steps 3 and 4).

What are the options when asthma is not controlled with a combination of inhaled steroids and LABA?

Current guidelines recommend a higher‐dose of inhaled steroids or adding another bronchodilator known as long‐acting muscarinic antagonist (LAMA), (i.e. triple inhaled therapy) (step 5), when asthma is not controlled with medium‐dose inhaled steroids and LABA dual inhaled therapy.

How did we answer the question?

We collected and analysed data from 17 studies, including a total of 17,161 adolescents and adults with uncontrolled asthma, using a special method called a network meta‐analysis, which enabled us to simultaneously compare multiple inhaler groups.

What did we find?

Triple inhaled therapy (i.e, inhaled steroids + LABA + LAMA) reduces asthma flare‐ups, but not asthma‐related hospitalisations. High‐dose triple therapy, not medium‐dose triple, is likely to be better tolerated due to less side effects compared to dual inhaled therapy (i.e. inhaled steroids + LABA).

Triple therapy may improve symptom and quality of life scores compared to dual therapy but not enough to be perceived by those being on it.

Higher than medium‐dose inhaled steroids in dual inhaled therapy are unlikely to result in any additional benefit or harm.

Conclusions

Triple inhaled therapy, especially high‐dose formulations, reduces asthma flare‐ups and is likely to be better tolerated due to less side effects compared to dual therapy.

Triple inhaled therapy may or may not to improve symptoms or quality of life compared to dual therapy.

Increasing the strength of inhaled steroids from medium to high dose is likely beneficial in triple inhaled therapy but probably not in dual therapy.

Immuno modulators, which are injectable medications, or other options may be considered if asthma symptoms are not well controlled or for those requiring asthma‐related hospitalisations despite being on medium‐dose dual inhaled therapy.

Summary of findings

Summary of findings 1. NMA Summary of Findings for severe exacerbations (asthma‐related hospitalisations).

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Severe exacerbations Setting(s): Outpatient					Geometry of the Network in Figure 1*
Total studies: 8 RCTs Total Participants: 9983	Hazard ratio (95% CrI)**	Anticipated absolute effect at the end of 1 year(95% CrI)*		Certainty of the evidence	Ranking** (95% CrI)**	Interpretation of Findings
Total studies: 8 RCTs Total Participants: 9983	Hazard ratio (95% CrI)**	*With intervention* *(With MD‐ICS/LABA)*	Difference	Certainty of the evidence	Ranking** (95% CrI)**	Interpretation of Findings
HD‐ICS/LABA (Direct evidence; 7 RCTs; 7023 participants)	1.43 (0.76 to 2.77)	15 per 1000	5 per 1000 more (from 2 fewer to 18 more)	⊕⊕⊕◯ Moderate Due to substantial heterogeneity¹	3.0 (1.0 to 4.0)	Probably little or no difference
MD‐TRIPLE (Direct evidence; 2 RCTs; 1023 participants)	1.73 (0.90 to 3.32)	18 per 1000	8 per 1000 more (from 1 fewer to 24 more)	⊕⊕◯◯ Low Due to imprecision²	4.0 (1.0 to 4.0)	Suggest little or no difference
HD‐TRIPLE (Direct evidence; 2 RCTs; 1024 participants)	1.14 (0.54 to 2.41)	12 per 1000	2 per 1000 more (from 4 fewer to 15 more)	⊕⊕◯◯ Low Due to imprecision²	2.0 (1.0 to 4.0)	Suggest little or no difference
MD‐ICS/LABA	Reference Comparator	(10 per 1000)³	Reference Comparator	Reference Comparator	1.0 (1.0 to 3.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted. Network Meta‐Analysis estimates are reported as hazard ratio. Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Anticipated absolute effect (exacerbation rate at 1 year). Anticipated absolute effect compares two rates by calculating the difference between the rates of the intervention group with the rate of MD‐ICS/LABA group. **** Median and credible intervals are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ Substantial heterogeneity I²>= 50% to 90% in the direct pairwise comparison. ² Very serious imprecision. Due to wide confidence intervals and suboptimal sample sizes in the direct and/or indirect estimate(s). ³ Based on the average rate in patients treated with MD‐ICS/LABA in the included studies.

Outcome № of participants (studies)	Relative effect (95% CI)	Anticipated absolute effects (95% CI)			Certainty of the evidence	What happens
Outcome № of participants (studies)	Relative effect (95% CI)	With active control	With experimental comparator	Difference	Certainty of the evidence	What happens
Severe exacerbations ‐ HD‐ICA/LABA vs MD‐ICS LABA № of participants: 4492 (5 RCTs) Follow up: 3 to 12 months	RR 1.49 (0.74 to 3.01)	0.8%	1.1% (0.6 to 2.3)	0.4% more (0.2 fewer to 1.5 more)	⨁⨁⨁◯ Moderate^a	HD‐ICA/LABA likely results in little to no difference in severe exacerbations compared to MD‐ICS LABA.
Severe exacerbations ‐ MD TRIPLE vs MD‐ICS/LABA № of participants: 813 (1 RCT) Follow up: 12 months	RR 1.00 (0.35 to 2.83)	1.7%	1.7% (0.6 to 4.9)	0.0% fewer (1.1 fewer to 3.1 more)	⨁⨁◯◯ Low^{b, c}	The evidence suggests that MD TRIPLE results in little to no difference in severe exacerbations compared to MD‐ICS/LABA.
Severe exacerbations ‐ HD TRIPLE vs MD‐ICS/LABA № of participants: 815 (1 RCT) Follow up: 12 months	RR 0.57 (0.17 to 1.93)	1.7%	1.0% (0.3 to 3.3)	0.7% fewer (1.4 fewer to 1.6 more)	⨁⨁◯◯ Low^{b, c}	The evidence suggests that HD TRIPLE results in little to no difference in severe exacerbations compared to MD‐ICS/LABA.
Severe exacerbations ‐ MD TRIPLE vs HD‐ICS/LABA № of participants: 812 (1 RCT) Follow up: 12 months	RR 1.40 (0.45 to 4.37)	1.2%	1.7% (0.6 to 5.4)	0.5% more (0.7 fewer to 4.2 more)	⨁⨁◯◯ Low^{b, c}	The evidence suggests that MD TRIPLE results in little to no difference in severe exacerbations compared to HD‐ICS/LABA.
Severe exacerbations ‐ HD TRIPLE vs HD‐ICS/LABA № of participants: 1727 (2 RCTs) Follow up: 12 months	RR 0.80 (0.45 to 1.42)	2.9%	2.3% (1.3 to 4.1)	0.6% fewer (1.6 fewer to 1.2 more)	⨁⨁⨁◯ Moderate^b	HD TRIPLE likely results in little to no difference in severe exacerbations compared to HD‐ICS LABA.
Severe exacerbations ‐ HD TRIPLE vs MD TRIPLE № of participants: 814 (1 RCT) Follow up: 12 months	RR 0.57 (0.17 to 1.93)	1.7%	1.0% (0.3 to 3.3)	0.7% fewer (1.4 fewer to 1.6 more)	⨁⨁◯◯ Low^{b, c}	The evidence suggests that HD TRIPLE results in little to no difference in severe exacerbations compared to MD TRIPLE.
Severe exacerbations ‐ TRIPLE vs DUAL № of participants: 2540 (2 RCTs) Follow up: 12 months	RR 0.84 (0.51 to 1.40)	2.5%	2.1% (1.3 to 3.5)	0.4% fewer (1.2 fewer to 1 more)	⨁⨁⨁◯ Moderate^d	TRIPLE likely results in little to no difference in severe exacerbations compared to DUAL.
Moderate to severe exacerbations ‐ HD‐ICS/LABA vs MD‐ICS/LABA № of participants: 5452 (6 RCTs) Follow up: 3 to 12 months	RR 0.93 (0.82 to 1.05)	15.0%	14.0% (12.3 to 15.8)	1.1% fewer (2.7 fewer to 0.8 more)	⨁⨁⨁⨁ High	HD‐ICS/LABA results in little to no difference in moderate to severe exacerbations compared to MD‐ICS/LABA.
Moderate to severe exacerbations ‐ MD TRIPLE vs MD‐ICS/LABA № of participants: 3184 (3 RCTs) Follow up: 12 months	RR 0.86 (0.75 to 0.99)	22.8%	19.6% (17.1 to 22.6)	3.2% fewer (5.7 fewer to 0.2 fewer)	⨁⨁⨁◯ Moderate^b	MD TRIPLE likely reduces moderate to severe exacerbations compared to MD‐ICS/LABA.
Moderate to severe exacerbations ‐ HD TRIPLE vs MD‐ICS/LABA № of participants: 2037 (2 RCTs) Follow up: 12 months	RR 0.78 (0.66 to 0.92)	24.0%	18.7% (15.8 to 22)	5.3% fewer (8.1 fewer to 1.9 fewer)	⨁⨁⨁⨁ High	HD TRIPLE reduces moderate to severe exacerbations compared to MD‐ICS/LABA.
Moderate to severe exacerbations ‐ MD TRIPLE vs HD‐ICS/LABA № of participants: 2651 (2 RCTs) Follow up: 12 months	RR 1.05 (0.78 to 1.41)	23.4%	24.6% (18.2 to 33)	1.2% more (5.1 fewer to 9.6 more)	⨁⨁◯◯ Low^{a, d}	MD TRIPLE may result in little to no difference in moderate to severe exacerbations compared to HD‐ICS/LABA.
Moderate to severe exacerbations ‐ HD TRIPLE vs HD‐ICS/LABA № of participants: 4989 (4 RCTs) Follow up: 12 months	RR 0.83 (0.75 to 0.92)	25.2%	20.9% (18.9 to 23.2)	4.3% fewer (6.3 fewer to 2 fewer)	⨁⨁⨁⨁ High	HD TRIPLE reduces moderate to severe exacerbations compared to HD‐ICS/LABA.
Moderate to severe exacerbations ‐ HD TRIPLE vs MD TRIPLE № of participants: 3470 (3 RCTs) Follow up: 6 to 12 months	RR 0.85 (0.72 to 1.01)	15.2%	12.9% (10.9 to 15.3)	2.3% fewer (4.2 fewer to 0.2 more)	⨁⨁⨁◯ Moderate^e	HD TRIPLE likely results in a slight reduction in moderate to severe exacerbations compared to MD TRIPLE.
Moderate to severe exacerbations ‐ TRIPLE vs DUAL № of participants: 8173 (5 RCTs) Follow up: 12 months	RR 0.85 (0.78 to 0.92)	24.3%	20.6% (18.9 to 22.3)	3.6% fewer (5.3 fewer to 1.9 fewer)	⨁⨁⨁⨁ High	TRIPLE reduces moderate to severe exacerbations compared to DUAL.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
Explanations a. Substantial heterogeneity I2 > 50% to 90% b. Optimal information size is not met (Guyatt 2011b) c. Total size of less than 1000 participants may suggest small study effect (Dechartres 2013) d. Confidence interval includes a clinically important difference. e. Confidence interval includes the null effect.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Moderate to severe exacerbations Setting(s): Outpatient					Geometry of the Network in Figure 2*
Total studies: 10 RCTs Total Participants: 12407	Hazard ratio (95% CrI)**	Anticipated absolute effect at the end of 1 year(95% CrI)*		Certainty of the evidence	Ranking** (95% CrI)**	Interpretation of Findings
Total studies: 10 RCTs Total Participants: 12407	Hazard ratio (95% CrI)**	*With intervention*	Difference compared to MD‐ICS/LABA	Certainty of the evidence	Ranking** (95% CrI)**	Interpretation of Findings
HD‐ICS/LABA (Direct evidence; 6 RCTs; 5452 participants)	0.90 (0.77 to 1.04)	176 per 1000	20 per 1000 fewer (from 45 fewer to 8 more)	⊕⊕⊕⊕ High	3.0 (2.0 to 4.0)	Little or no difference
MD‐TRIPLE (Direct evidence; 3 RCTs; 3184 participants)	0.84 (0.71 to 0.99)	165 per 1000	31 per 1000 fewer (from 2 fewer to 57 fewer)	⊕⊕⊕◯ Moderate Due to imprecision¹	2.0 (2.0 to 3.0)	Probably superior
HD‐TRIPLE (Direct evidence; 2 RCTs; 2037 participants)	0.69 (0.58 to 0.82)	135 per 1000	61 per 1000 fewer (from 35 fewer to 82 fewer)	⊕⊕⊕⊕ High	1.0 (1.0 to 1.0)	Superior
MD‐ICS/LABA	Reference Comparator	196 per 1000²	Reference Comparator	Reference Comparator	4.0 (3.0 to 4.0)	Reference Comparator
HD Triple vs. MD Triple
HD‐TRIPLE (Direct evidence; 3 RCTs; 3470 participants)	0.83 (0.69 to 0.996)	162 per 1000	34 per 1000 fewer (from 1 fewer to 61 fewer)	⊕⊕⊕◯ Moderate Due to imprecision¹	NA	Probably superior
MD Triple	Reference Comparator	196 per 1000³	Reference Comparator	Reference Comparator	NA	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Network Meta‐Analysis estimates are reported as hazard ratio. Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Anticipated absolute effect (exacerbation rate at 1 year). Anticipated absolute effect compares two rates by calculating the difference between the rates of the intervention group with the rate of MD‐ICS/LABA group. **** Median and credible intervals are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ Serious imprecision. Due to suboptimal sample size(s) in the direct and/or indirect estimate(s). ² Based on the average rate in participants treated with MD‐ICS/LABA in the included studies. ³ Based on the average rate in participants treated with MD Triple in the included studies.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Change from baseline in ACQ scores at 3 months Setting(s): Outpatient						Geometry of the Network in Figure 3*
Total studies: 4 RCTs Total Participants: 4529	Relative effect (95% CrI)	Anticipated absolute effect(95% CrI)**		Certainty of the evidence	Ranking* (95% CrI)**	Interpretation of Findings
Total studies: 4 RCTs Total Participants: 4529	Relative effect (95% CrI)	With intervention	*Difference compared to MD‐ICS/LABA¹*	Certainty of the evidence	Ranking* (95% CrI)**	Interpretation of Findings
HD‐ICS/LABA (Direct evidence; 3 RCTs; 2450 participants)	0.01 (‐0.05 to 0.07)	0.72 (0.67 to 0.78)	Change from baseline in ACQ score was 0.01 lower (0.07 lower to 0.05 higher)	⨁⨁⨁◯ Moderate Due to imprecision²	4.0 (2.0 to 4.0)	Probably little or no clinically meaningful difference⁴
MD‐TRIPLE (Direct evidence; 1 RCT; 768 participants)	‐0.06 (‐0.14 to 0.03)	0.78 (0.70 to 0.87)	Change from baseline in ACQ score was 0.06 higher (0.03 lower to 0.14 higher)	⊕⊕◯◯ Low Due to imprecision³	2.0 (1.0 to 4.0)	Suggest little or no clinically meaningful difference⁴
HD‐TRIPLE (Direct evidence; 1 RCT; 764 participants)	‐0.09 (‐0.18 to ‐ 0.01)	0.82 (0.74 to 0.90)	Change from baseline in ACQ score was 0.09 higher (0.01 higher to 0.18 higher)	⊕⊕◯◯ Low Due to imprecision³	1.0 (1.0 to 2.0)	Suggest little or no clinically meaningful difference⁴
MD‐ICS/LABA	Reference Comparator¹	0.72	Reference Comparator	Reference Comparator	3.0 (2.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Estimates are reported as mean difference and credible interval (CrI). Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Ranking and confidence intervals for efficacy outcome are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ The mean change from baseline in ACQ scores was 0.72with MD‐ICS/LABA. ² Serious imprecision. Due to small sample sizes in the direct and/or indirect estimate(s). ³ Very serious imprecision. Due to very small sample sizes in the direct and/or indirect estimate(s). ⁴ Minimal clinically important difference is 0.5.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Change from baseline in ACQ scores at 6 months Setting(s): Outpatient					Geometry of the Network in Figure 4*
Total studies: 6 RCTs Total Participants: 7957	Relative effect (95% CrI)	Anticipated absolute effect(95% CrI)**		Certainty of the evidence	Ranking* (95% CrI)**	Interpretation of Findings
Total studies: 6 RCTs Total Participants: 7957	Relative effect (95% CrI)	With intervention	*Difference compared to MD‐ICS/LABA¹*	Certainty of the evidence	Ranking* (95% CrI)**	Interpretation of Findings
HD‐ICS/LABA (Direct evidence; 3 RCTs; 3762 participants)	‐0.03 (‐0.09 to 0.02)	0.90 (0.84 to 0.95)	Change from baseline in ACQ score was 0.03 higher (0.02 lower to 0.09 higher)	⊕⊕⊕◯ Moderate Due to imprecision²	3.0 (2.0 to 4.0)	Probably little or no clinically meaningful difference³
MD‐TRIPLE (Direct evidence; 2 RCTs; 1961 participants)	‐0.07 (‐0.13 to 0.00)	0.93 (0.86 to 1.00)	Change from baseline in ACQ score was 0.07 higher (0.00 lower to 0.13 higher)	⊕⊕⊕◯ Moderate Due to imprecision²	2.0 (1.0 to 3.0)	Probably little or no clinically meaningful difference³
HD‐TRIPLE (Direct evidence; 2 RCTs; 1952 participants)	‐0.10 (‐0.16 to ‐0.03)	0.96 (0.90 to 1.02)	Change from baseline in ACQ score was 0.1 higher (0.03 higher to 0.16 higher)	⊕⊕⊕◯ Moderate Due to imprecision²	1.0 (1.0 to 2.0)	Probably little or no clinically meaningful difference³
MD‐ICS/LABA	Reference Comparator¹	0.86	Reference Comparator	Reference Comparator	4.0 (3.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Estimates are reported as mean difference and credible interval (CrI). Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Ranking and confidence intervals for efficacy outcome are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ The mean change from baseline in ACQ scores was 0.86 with MD‐ICS/LABA. ² Serious imprecision due to small sample sizes in the direct and/or indirect estimate(s). ³ Minimal clinically important difference is 0.5.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Change from baseline in ACQ scores at 12 months Setting(s): Outpatient					Geometry of the Network in Figure 5*
Total studies: 5 RCTs Total Participants: 5440	Relative effect (95% CrI)	Anticipated absolute effect(95% CrI)**		Certainty of the evidence		Ranking* (95% CrI)**	Interpretation of Findings
		With intervention	*Difference compared to MD‐ICS/LABA¹*
HD‐ICS/LABA (Direct evidence; 3 RCTs; 3152 participants)	0.00 (‐0.06 to 0.06)	1.00 (0.94 to 1.06)	Change from baseline in ACQ score was 0.00 (0.06 lower to 0.06 higher)	⊕⊕⊕◯ Moderate Due to imprecision²		3.0 (2.0 to 4.0)	Probably little or no clinically meaningful difference³
MD‐TRIPLE (Direct evidence; 2 RCTs; 1366 participants)	0.02 (‐0.07 to 0.11)	0.98 (0.89 to 1.07)	Change from baseline in ACQ score was 0.08 higher (0.01 lower to 0.17 higher)	⊕⊕⊕◯ Moderate Due to imprecision²		4.0 (2.0 to 4.0)	Probably little or no clinically meaningful difference³
HD‐TRIPLE (Direct evidence; 2 RCTs; 1379 participants)	‐0.08 (‐0.16 to 0.00)	1.08 (1.00 to 1.16)	Change from baseline in ACQ score was 0.08 higher (0.00 lower to 0.16 higher)	⊕⊕⊕◯ Moderate Due to imprecision²		1.0 (1.0 to 2.0)	Probably little or no clinically meaningful difference³
MD‐ICS/LABA	Reference Comparator¹	1.00	Reference Comparator	Reference Comparator		3.0 (2.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Estimates are reported as mean difference and credible interval (CrI). Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Ranking and confidence intervals for efficacy outcome are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ The mean change from baseline in ACQ scores was 1.00with MD‐ICS/LABA. ² Serious imprecision due to small sample sizes in the direct and/or indirect estimate(s). ³ Minimal clinically important difference is 0.5.

Outcome № of participants (studies)	Relative effect (95% CI)	Anticipated absolute effects (95% CI)*		Certainty of the evidence	What happens†
Outcome № of participants (studies)	Relative effect (95% CI)	With active control	Difference	Certainty of the evidence	What happens†
CFB in ACQ at 3 months ‐ HD‐ICS/LABA vs MD‐ICS/LABA № of participants: 2450 (3 RCTs)	‐	‐0.72	MD 0.01 higher (0.05 lower to 0.07 higher)	⨁⨁⨁◯ Moderate^a	HD‐ICS/LABA likely results in little to no difference in CFB in ACQ at 3 months compared to MD‐ICS/LABA.
CFB in ACQ at 3 months ‐ MD TRIPLE vs MD‐ICS/LABA № of participants: 768 (1 RCT)	‐	‐0.58	MD 0.06 lower (0.16 lower to 0.04 higher)	⨁⨁◯◯ Low^{a, b}	The evidence suggests that MD TRIPLE results in little to no difference in CFB in ACQ at 3 months compared to MD‐ICS/LABA.
CFB in ACQ at 3 months ‐ HD TRIPLE vs MD‐ICS/LABA № of participants: 764 (1 RCT)	‐	‐0.58	MD 0.12 lower (0.22 lower to 0.02 lower)	⨁⨁◯◯ Low^{a, b}	The evidence suggests that HD TRIPLE results in little to no difference in CFB in ACQ at 3 months compared to MD‐ICS/LABA.
CFB in ACQ at 3 months ‐ MD TRIPLE vs HD‐ICS/LABA № of participants: 771 (1 RCT)	‐	‐0.61	MD 0.04 lower (0.14 lower to 0.06 higher)	⨁⨁◯◯ Low^{a, b}	The evidence suggests that MD TRIPLE results in little to no difference in CFB in ACQ at 3 months compared to HD‐ICS/LABA.
CFB in ACQ at 3 months ‐ HD TRIPLE vs HD‐ICS/LABA № of participants: 767 (1 RCT)	‐	‐0.61	MD 0.09 lower (0.19 lower to 0.01 higher)	⨁⨁◯◯ Low^{a, b}	The evidence suggests that HD TRIPLE results in little to no difference in CFB in ACQ at 3 months compared to HD‐ICS/LABA.
CFB in ACQ at 3 months ‐ HD TRIPLE vs MD TRIPLE № of participants: 2079 (2 RCTs)	‐	‐0.85	MD 0.04 lower (0.11 lower to 0.03 higher)	⨁⨁⨁◯ Moderate^a	HD TRIPLE likely results in little to no difference in CFB in ACQ at 3 months compared to MD TRIPLE.
CFB in ACQ at 3 months ‐ TRIPLE vs DUAL № of participants: 1535 (1 RCT)	‐	‐0.59	MD 0.08 lower (0.15 lower to 0.01 lower)	⨁⨁⨁◯ Moderate^a	TRIPLE likely results in little to no difference in CFB in ACQ at 3 months compared to DUAL.
CFB in ACQ at 6 months ‐ HD‐ICS/LABA vs MD‐ICS/LABA № of participants: 3762 (3 RCTs)	‐	‐0.86	MD 0.04 lower (0.12 lower to 0.04 higher)	⨁⨁⨁◯ Moderate^c	HD‐ICS/LABA likely results in little to no difference in CFB in ACQ at 6 months compared to MD‐ICS/LABA.
CFB in ACQ at 6 months ‐ MD TRIPLE vs MD‐ICS/LABA № of participants: 1961 (2 RCTs)	‐	‐0.79	MD 0.09 lower (0.17 lower to 0.02 lower)	⨁⨁⨁◯ Moderate^a	MD TRIPLE likely results in little to no difference in CFB in ACQ at 6 months compared to MD‐ICS/LABA.
CFB in ACQ at 6 months ‐ HD TRIPLE vs MD‐ICS/LABA № of participants: 1952 (2 RCTs)	‐	‐0.79	MD 0.11 lower (0.18 lower to 0.04 lower)	⨁⨁⨁◯ Moderate^a	HD TRIPLE likely results in little to no difference in CFB in ACQ at 6 months compared to MD‐ICS/LABA.
CFB in ACQ at 6 months ‐ MD TRIPLE vs HD‐ICS/LABA № of participants: 2561 (2 RCTs)	‐	‐0.91	MD 0.01 lower (0.08 lower to 0.06 higher)	⨁⨁⨁◯ Moderate^a	MD TRIPLE likely results in little to no difference in CFB in ACQ at 6 months compared to HD‐ICS/LABA.
CFB in ACQ at 6 months ‐ HD TRIPLE vs HD‐ICS/LABA № of participants: 3459 (3 RCTs)	‐	‐0.82	MD 0.06 lower (0.15 lower to 0.03 higher)	⨁⨁◯◯ Low^{a, c}	The evidence suggests that HD TRIPLE results in little to no difference in CFB in ACQ at 6 months compared to HD‐ICS/LABA.
CFB in ACQ at 6 months ‐ HD TRIPLE vs MD TRIPLE № of participants: 3288 (3 RCTs)	‐	‐0.94	MD 0.02 lower (0.08 lower to 0.04 higher)	⨁⨁⨁◯ Moderate^a	HD TRIPLE likely results in little to no difference in CFB in ACQ at 6 months compared to MD‐ICS/LABA.
CFB in ACQ at 6 months ‐ TRIPLE vs DUAL № of participants: 5408 (4 RCTs)	‐	‐0.81	MD 0.07 lower (0.14 lower to 0.01 lower)	⨁⨁⨁◯ Moderate^a	TRIPLE likely results in little to no difference in CFB in ACQ at 6 months compared to DUAL.
CFB in ACQ at 12 months ‐ HD‐ICS/LABA vs MD‐ICS/LABA № of participants: 3152 (3 RCTs)	‐	‐1.00	MD 0  (0.12 lower to 0.12 higher)	⨁⨁◯◯ Low^{a, c, d}	The evidence suggests that HD‐ICS/LABA results in little to no difference in CFB in ACQ at 12 months compared to MD‐ICS/LABA.
CFB in ACQ at 12 months ‐ MD TRIPLE vs MD‐ICS/LABA № of participants: 1366 (2 RCTs)	‐	‐0.93	MD 0.01 lower (0.11 lower to 0.08 higher)	⨁⨁⨁◯ Moderate^{a, d}	MD TRIPLE likely results in little to no difference in CFB in ACQ at 12 months compared to MD‐ICS/LABA.
CFB in ACQ at 12 months ‐ HD TRIPLE vs MD‐ICS/LABA № of participants: 1379 (2 RCTs)	‐	‐0.93	MD 0.09 lower (0.23 lower to 0.06 higher)	⨁⨁⨁◯ Moderate^{a, d}	HD TRIPLE likely results in little to no difference in CFB in ACQ at 12 months compared to MD‐ICS/LABA.
CFB in ACQ at 12 months ‐ MD TRIPLE vs HD‐ICS/LABA № of participants: 1967 (2 RCTs)	‐	‐1.03	MD 0.01 higher (0.2 lower to 0.21 higher)	⨁⨁◯◯ Low^{a, c, d}	The evidence suggests that MD TRIPLE results in little to no difference in CFB in ACQ at 12 months compared to HD‐ICS/LABA.
CFB in ACQ at 12 months ‐ HD TRIPLE vs HD‐ICS/LABA № of participants: 2887 (3 RCTs)	‐	‐0.89	MD 0.07 lower (0.15 lower to 0)	⨁⨁⨁◯ Moderate^{a, d}	HD TRIPLE likely results in little to no difference in CFB in ACQ at 12 months compared to HD‐ICS/LABA.
CFB in ACQ at 12 months ‐ HD TRIPLE vs MD TRIPLE № of participants: 1381 (2 RCTs)	‐	‐0.94	MD 0.07 lower (0.23 lower to 0.09 higher)	⨁⨁⨁◯ Moderate^{a, d}	HD TRIPLE likely results in little to no difference in CFB in ACQ at 12 months compared to MD TRIPLE.
CFB in ACQ at 12 months ‐ DUAL vs TRIPLE № of participants: 4253 (4 RCTs)	‐	‐0.91	MD 0.04 lower (0.1 lower to 0.02 higher)	⨁⨁⨁◯ Moderate^{a, d}	TRIPLE likely results in little to no difference in CFB in ACQ at 12 months compared to DUAL.
‡ ACQ scores range from 0 to 6 with lower scores indicating better asthma control. *The effect in the intervention group (and its 95% confidence interval) is based on the assumed effect in the comparison group and the relative effect of the intervention (and its 95% CI). † Minimal Clinically Important Difference is 0.5
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
Explanations a. Optimal information size is not met (Guyatt 2011b) b. Total size of less than 1000 participants may suggest small study effect (Dechartres 2013) c. Substantial heterogeneity I2 > 50% to 90% d. Lee 2020 had very high attrition rates and is considered at high risk of bias. However, excluding the study did not change the results.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Change from baseline in AQLQ score at 6 months Setting(s): Outpatient					Geometry of the Network in Figure 6*
Total studies: 4 RCTs Total Participants: 3454	Relative effect (95% CrI)	Anticipated absolute effect(95% CrI)**		Certainty of the evidence	Ranking* (95% CrI)**	Interpretation of Findings
Total studies: 4 RCTs Total Participants: 3454	Relative effect (95% CrI)	With intervention	*Difference compared to MD‐ICS/LABA¹*	Certainty of the evidence	Ranking* (95% CrI)**	Interpretation of Findings
HD‐ICS/LABA (Direct evidence; 1 RCT; 1223 participants)	‐0.06 (‐0.14 to 0.03)	0.71 (0.63 to 0.80)	Change from baseline in AQLQ score was 0.06 lower (0.14 lower to 0.03 higher)	⊕⊕⊕◯ Moderate Due to imprecision²	4.0 (2.0 to 4.0)	Probably little or no clinically meaningful difference⁴
MD‐TRIPLE (Direct evidence; 0 RCTs; 0 participants)	0.03 (‐0.23 to 0.29)	0.80 (0.54 to 1.06)	Change from baseline in AQLQ score was 0.03 higher (0.23 lower to 0.29 higher)	⊕⊕◯◯ Low Due to imprecision³	2.0 (1.0 to 4.0)	Suggest little or no clinically meaningful difference⁴
HD‐TRIPLE (Direct evidence; 0 RCTs; 0 participants)	0.11 (‐0.09 to 0.30)	0.88 (0.68 to 1.07)	Change from baseline in AQLQ score was 0.11 higher (0.09 lower to 0.30 higher)	⊕⊕◯◯ Low Due to imprecision³	1.0 (1.0 to 3.0)	Suggest little or no clinically meaningful difference⁴
MD‐ICS/LABA	Reference Comparator¹	0.77	Reference Comparator	Reference Comparator	3.0 (1.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted. Estimates are reported as mean difference and credible interval (CrI). Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Ranking and confidence intervals for efficacy outcome are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ The mean change from baseline in AQLQ scores was 0.77 with MD‐ICS/LABA. ² Serious imprecision due to small sample sizes in the direct and/or indirect estimate(s). ³ Very serious imprecision due to very small sample sizes in the indirect estimate. ⁴ Minimal clinically important difference is 0.5.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Change from baseline in AQLQ score at 12 months Setting(s): Outpatient					Geometry of the Network in Figure 7*
Total studies: 4 RCTs Total Participants: 4809	Relative effect (95% CrI)	Anticipated absolute effect(95% CrI)**		Certainty of the evidence		Ranking* (95% CrI)**	Interpretation of Findings
		With intervention	*Difference compared to MD‐ICS/LABA¹*
HD‐ICS/LABA (Direct evidence; 2 RCTs; 2815 participants)	‐0.02 (‐0.09 to 0.04)	0.81 (0.74 to 0.87)	Change from baseline in AQLQ score was 0.02 lower (0.09 lower to 0.04 higher)	⊕⊕⊕◯ Moderate Due to imprecision²		3.0 (2.0 to 4.0)	Probably little or no clinically meaningful difference³
MD‐TRIPLE (Direct evidence; 1 RCT; 1071 participants)	‐0.08 (‐0.17 to 0.02)	0.75 (0.66 to 0.85)	Change from baseline in AQLQ score was 0.08 lower (0.17 lower to 0.12 higher)	⊕⊕⊕◯ Moderate Due to imprecision²		4.0 (2.0 to 4.0)	Probably little or no clinically meaningful difference³
HD‐TRIPLE (Direct evidence; 1 RCT; 1088 participants)	0.05 (‐0.04 to 0.13)	0.88 (0.79 to 0.13)	Change from baseline in AQLQ score was 0.05 higher (0.04 lower to 0.13 higher)	⊕⊕⊕◯ Moderate Due to imprecision²		1.0 (1.0 to 3.0)	Probably little or no clinically meaningful difference³
MD‐ICS/LABA	Reference Comparator¹	0.83	Reference Comparator	Reference Comparator		2.0 (1.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Estimates are reported as mean difference and credible interval (CrI). Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Ranking and confidence intervals for efficacy outcome are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ The mean change from baseline in ACQ scores was 0.83 with MD‐ICS/LABA. ² Serious imprecision due to small sample sizes in the direct and/or indirect estimate(s). ³ Minimal clinically important difference is 0.5.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: ACQ responders at 6 months Setting(s): Outpatient					Geometry of the Network in Figure 8*
Total studies: 7 RCTs Total Participants: 10453	Risk ratio (95% CrI)**	Anticipated absolute effect(95% CrI)*		Certainty of the evidence		Ranking** (95% CrI)**	Interpretation of Findings
		*With intervention*	Difference compared to MD‐ICS/LABA
HD‐ICS/LABA (Direct evidence; 3 RCTs; 3700 participants)	1.05 (0.92 to 1.20)	632 per 1000	12 per 1000 more (from 19 fewer to 43 more)	⊕⊕⊕◯ Moderate Due to imprecision¹		3.0 (3.0 to 4.0)	Probably little or no difference
MD‐TRIPLE (Direct evidence; 3 RCTs; 3063 participants)	1.25 (1.09 to 1.44)	670 per 1000	50 per 1000 more (from 19 more to 81 more)	⊕⊕◯◯ Low Due to imprecision¹ and heterogeneity²		1.0 (1.0 to 2.0)	Possibly superior
HD‐TRIPLE (Direct evidence; 2 RCTs; 1916 participants)	1.25 (1.07 to 1.45)	670 per 1000	50 per 1000 more (19 more to 81 more)	⊕⊕◯◯ Low Due to imprecision¹ and heterogeneity²		2.0 (1.0 to 2.0)	Possibly superior
MD‐ICS/LABA	Reference Comparator	620 per 1000³	Reference Comparator	Reference Comparator		4.0 (3.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Network Meta‐Analysis estimates are reported as risk ratio. Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Anticipated absolute effect. Anticipated absolute effect compares two rates by calculating the difference between the rates of the intervention group with the rate of MD‐ICS/LABA group. **** Median and credible intervals are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ Serious imprecision due to suboptimal sample size in the direct and/or indirect estimate(s). ² Serious heterogeneity in the direct estimate. ³ Based on the average rate in participants treated with MD‐ICS/LABA in the included studies.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: ACQ responders at 12 months					Geometry of the Network in Figure 9*
Total studies: 5 RCTs Total Participants: 7391	Risk ratio (95% CrI)**	Anticipated absolute effect(95% CrI)*		Certainty of the evidence	Ranking** (95% CrI)**	Interpretation of Findings
Total studies: 5 RCTs Total Participants: 7391	Risk ratio (95% CrI)**	*With intervention*	Difference compared to MD‐ICS/LABA	Certainty of the evidence	Ranking** (95% CrI)**	Interpretation of Findings
HD‐ICS/LABA (Direct evidence; 2 RCTs; 2817 participants)	1.00 (0.94 to 1.05)	676 per 1000	0 per 1000 fewer (from 41 fewer to 30 more)	⊕⊕⊕◯ Moderate Due to imprecision¹	3.0 (2.0 to 4.0)	Probably little or no difference
MD‐TRIPLE (Direct evidence; 2 RCTs; 2237 participants)	0.99 (0.94 to 1.05)	669 per 1000	7 per 1000 more (from 41 fewer to 34 more)	⊕⊕⊕◯ Moderate Due to imprecision¹	3.0 (2.0 to 4.0)	Probably little or no difference
HD‐TRIPLE (Direct evidence; 1 RCT; 1088 participants)	1.08 (1.02 to 1.14)	730 per 1000	54 per 1000 more (14 more to 95 more)	⊕⊕⊕◯ Moderate Due to imprecision²	1.0 (1.0 to 1.0)	Probably superior
MD‐ICS/LABA	Reference Comparator	676 per 1000³	Reference Comparator	Reference Comparator	3.0 (2.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Network Meta‐Analysis estimates are reported as risk ratio. Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Anticipated absolute effect. Anticipated absolute effect compares two rates by calculating the difference between the rates of the intervention group with the rate of MD‐ICS/LABA group. **** Median and credible intervals are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ Serious imprecision due to suboptimal sample size in the direct and/or indirect estimate(s). ² Serious imprecision due 95% CI or CrI including the null effect in the direct and/or indirect estimate(s). ³ Based on the average rate in participants treated with MD‐ICS/LABA in the included studies.

Study, year	Arms included Dose in micrograms	Duration (weeks)	No. of participants included	Mean age	Male (%)	White (%)	Current smoker excluded/ maximum PYs allowed for ex‐smokers	Baseline FEV1 (L) prebronchodilator (% predicted)	History of at least one asthma exacerbation
Bernstein 2011	MF/FM 200/10 bid	12	371	44.8	87	87	Y/10	2.3 (74)	Not required
Bernstein 2011	FP/SAL 250/50 bid	12	351	45.1	86	86	Y/10	2.4 (74)	Not required
Bernstein 2015	FF/VI 100/25 qd	12	346	44.7	41	89	Y/10	2.0 (63)	Not required
Bernstein 2015	FF/VI 200/25 qd	12	346	45.9	35	87	Y/10	2.0 (62)	Not required
Bodzenta‐Lukaszyk 2012	FP/FM 250/10 bid	12	140	49.8	37	96	Y/10	NR (65)	Not required
Bodzenta‐Lukaszyk 2012	BUD/FM 400/12 bid	12	139	48.1	27	96	Y/10	NR (64)	Not required
Busse 2008	BUD/FM 320/9 bid	24	427	39.4	34	82	N/20	2.5 (79)	Not required
Busse 2008	FP/SAL 250/50 bid	24	406	38.8	43	84	N/20	2.6 (78)	Not required
Cukier 2013	FP/FM 250/12 bid	12	97	34.5	24	67	Y/10	2.5 (86)	Not required
Cukier 2013	BUD/FM 400/12 bid	12	99	35.6	27	72	Y/10	2.5 (85)	Not required
Gessner 2020	MF/GLY/IND 80/50/150 qd	24	474	51.9	35	85	N/20	NR (63)	Required
	MF/GLY/IND 160/50/150 qd		476	52.7	39	82		NR (62)
	FP/SAL 500/50 bid + Tio 5 qd		476	53.1	36	82		NR(63)
Kerstjens 2012a	HD‐ICS/LABA	48	237	52.9	38	84	Y/10	1.6 (55)	Required
Kerstjens 2012a	HD‐ICS/LABA+Tio 5 qd	48	222	53.9	36	84	Y/10	1.6 (55)	Required
Kerstjens 2012b	HD‐ICS/LABA	48	219	51.4	42	80	Y/10	1.7 (55)	Required
Kerstjens 2012b	HD‐ICS/LABA+Tio 5 qd	48	234	53.6	42	84	Y/10	1.6 (55)	Required
Kerstjens 2020	MF/GLY/IND 80/50/150 qd	52	620	52.4	42	74	Y/10	1.6 (54)	Required
	MF/GLY/IND 160/50/150 qd		619	52.1	38	74		1.6 (55)
	MF/IND 160/150 qd		617	51.8	39	73		1.6 (55)
	MF/IND 320/150 qd		618	52.0	39	73		1.6 (54)
	FP/SAL 500/50 bid		618	52.9	33	76		1.6 (55)
Lee 2020	FF/VI 100/25 qd	24‐52	407	53.3	38	80	Y/10	1.7 (58)	Not required
	FF/UMEC/VI 100/62.5/25 qd		406	52.9	39	83		1.8 (59)
	FF/VI 200/25 qd		406	53.9	38	78		1.7 (59)
	FF/UMEC/VI 200/62.5/25 qd		408	53.7	37	80		1.7 (59)
Mansfield 2017	FP/SAL 250/50 bid	26	41	45.9	51	78	Y/10	2.4 (NR)	Not required
	FP/SAL 200/12.5 bid		133	46.1	46	71		2.3 (NR)
	FP/SAL 500/50 bid		44	45.6	48	70		2.5 (NR)
Papi 2007	BDP/FM 200/12 bid	12	115	47.3	45	NR	Y/10	2.1 (68)	Not required
Papi 2007	FP/SAL 250/50 bid	12	113	49.7	43	NR	Y/10	2.0 (67)	Not required
Peters 2008	BUD/FM 640/18 bid	52	443	41.0	37	87	Y/20	2.4 (75)	Not required
Peters 2008	BUD/FM 320/9 bid	52	132	38.6	41	89	Y/20	2.4 (72)	Not required
Stempel 2016	FP/SAL 250/50 bid	26	580	43.4	34	75	Y/10	NR (PEF>=50%)	Required
Stempel 2016	FP/SAL 500/50 bid	26	982	43.4	34	75	Y/10	NR (PEF>=50%)	Required
van Zyl‐Smit 2020	MF/IND 320/150 qd	26‐52	445	47.1	41	70	Y/10	2.1 (67)	Not required
	MF/IND 160/150 qd		439	47.4	42	71		2.1 (67)
	FP/SAL 500/50 bid		446	48.9	43	68		2.1 (67)
Virchow 2019a	BDP/FM/G 200/12/20 bid	26‐52	576	52.6	38	100	Y/10	1.7 (55)	Required

Treatment arm	No. of participants included	Mean age	Male %	White %	Maximum pack years allowed for smokers	Baseline FEV1 % predicted	History asthma exacerbation (%)
MD‐ICS/LABA	3502	48.4	39	82	10	60.3	33
HD‐ICS/LABA	5377	51.2	40	83	10	63.8	60
MD TRIPLE	2652	52.5	39	88	10‐20	57.0	85
HD TRIPLE	4151	52.8	37	88	10‐20	55.9	90

Model	p	Mean LHR (95% CrI)
MD Triple vs. HD‐ICS/LABA
Direct	0.717	0.091 (‐0.928, 0.995)
Indirect		0.328 (‐0.785, 1.422)
Network		0.184 (‐0.586, 0.930)
HD Triple vs. MD‐ICS/LABA
Direct	0.492	‐0.189 (‐1.519, 0.912)
Indirect		0.250 (‐0.606, 1.124)
Network		0.131 (‐0.621, 0.880)
HD Triple vs. MD Triple
Direct	0.506	‐0.700 (‐2.004, 0.392)
Indirect		‐0.261 (‐1.234, 0.694)
Network		‐0.416 (‐1.235, 0.414)

	Median HR (95% CrI)
Comparison	Overall	High Risk	Low Risk
HD‐ICS/LABA vs MD‐ICS/LABA	1.43 (0.76, 2.77)	13.41 (2.04, 191.20)*	0.76 (0.33, 1.80)
MD Triple vs. MD‐ICS/LABA	1.73 (0.90, 3.32)	1.89 (0.80, 4.47)	1.03 (0.36, 2.78)
HD Triple vs MD‐ICS/LABA	1.14 (0.54, 2.41)	11.77 (1.61, 169.90)*	0.56 (0.14, 1.79)
MD Triple vs. HD‐ICS/LABA	1.20 (0.56, 2.53)	0.14 (0.01, 1.14)	1.34 (0.45, 3.87)
HD Triple vs HD‐ICS/LABA	0.79 (0.48, 1.29)	0.85 (0.49, 1.48)	0.73 (0.18, 2.45)
HD Triple vs MD Triple	0.66 (0.29, 1.51)	6.23 (0.70, 104.30)*	0.55 (0.14, 1.85)

Outcome № of participants (studies)	Relative effect (95% CI)	*Anticipated absolute effects (95% CI)**			Certainty of the evidence	What happens
Outcome № of participants (studies)	Relative effect (95% CI)	With active control	With experimental comparator	Difference	Certainty of the evidence	What happens
ACQ responders at 6 months ‐ HD‐ICS/LABA vs MD‐ICS/LABA № of participants: 3700 (3 RCTs)	RR 1.02 (0.96 to 1.08)	66.8%	68.1% (64.1 to 72.2)	1.3% more (2.7 fewer to 5.3 more)	⨁⨁⨁◯ Moderate^{a, b}	HD‐ICS/LABA likely results in little to no difference in ACQ responders at 6 months compared to MD‐ICS/LABA.
ACQ responders at 6 months ‐ MD TRIPLE vs MD‐ICS/LABA № of participants: 3063 (3 RCTs)	RR 1.09 (0.99 to 1.19)	58.3%	63.5% (57.7 to 69.3)	5.2% more (0.6 fewer to 11.1 more)	⨁⨁◯◯ Low^{c, d}	The evidence suggests MD TRIPLE increases ACQ responders at 6 months compared to MD‐ICS/LABA.
ACQ responders at 6 months ‐ HD TRIPLE vs MD‐ICS/LABA № of participants: 1916 (2 RCTs)	RR 1.11 (0.91 to 1.35)	62.8%	69.7% (57.2 to 84.8)	6.9% more (5.7 fewer to 22 more)	⨁◯◯◯ Very low^{e, f}	HD TRIPLE may increase ACQ responders at 6 months compared to MD‐ICS/LABA, but the evidence is very uncertain.
ACQ responders at 6 months ‐ MD TRIPLE vs HD‐ICS/LABA № of participants: 2480 (2 RCTs)	RR 1.02 (0.97 to 1.08)	67.5%	68.9% (65.5 to 72.9)	1.4% more (2 fewer to 5.4 more)	⨁⨁⨁◯ Moderate^b	MD TRIPLE likely results in little to no difference in ACQ responders at 6 months compared to HD‐ICS/LABA.
ACQ responders at 6 months ‐ HD TRIPLE vs HD‐ICS/LABA № of participants: 4818 (4 RCTs)	RR 1.07 (1.01 to 1.14)	61.2%	65.5% (61.8 to 69.8)	4.3% more (0.6 more to 8.6 more)	⨁⨁⨁◯ Moderate^b	HD TRIPLE likely results in little to no difference in ACQ responders at 6 months compared to HD‐ICS/LABA.
ACQ responders at 6 months ‐ HD TRIPLE vs MD TRIPLE № of participants: 2821 (3 RCTs)	RR 0.99 (0.95 to 1.03)	73.2%	72.5% (69.6 to 75.4)	0.7% fewer (3.7 fewer to 2.2 more)	⨁⨁⨁◯ Moderate^b	HD TRIPLE likely results in little to no difference in ACQ responders at 6 months compared to MD TRIPLE.
ACQ responders at 6 months ‐ TRIPLE vs DUAL № of participants: 7881 (5 RCTs)	RR 1.09 (1.02 to 1.15)	60.1%	65.5% (61.3 to 69.1)	5.4% more (1.2 more to 9 more)	⨁⨁◯◯ Low^{b, c}	The evidence suggests TRIPLE increases ACQ responders at 6 months compared to DUAL.
ACQ responders at 12 months ‐ HD‐ICS/LABA vs MD‐ICS/LABA № of participants: 2817 (2 RCTs)	RR 0.99 (0.90 to 1.07)	77.0%	76.2% (69.3 to 82.3)	0.8% fewer (7.7 fewer to 5.4 more)	⨁⨁◯◯ Low^{a, b, c}	HD‐ICS/LABA likely results in little to no difference in ACQ responders at 12 months compared to MD‐ICS/LABA.
ACQ responders at 12 months ‐ MD TRIPLE vs MD‐ICS/LABA № of participants: 2222 (2 RCTs)	RR 1.01 (0.95 to 1.07)	65.9%	66.6% (62.6 to 70.6)	0.7% more (3.3 fewer to 4.6 more)	⨁⨁⨁◯ Moderate^b	MD TRIPLE likely results in little to no difference in ACQ responders at 12 months compared to MD‐ICS/LABA.
ACQ responders at 12 months ‐ HD TRIPLE vs MD‐ICS/LABA № of participants: 1088 (1 RCT)	RR 1.08 (1.01 to 1.15)	73.1%	79.0% (73.9 to 84.1)	5.9% more (0.7 more to 11 more)	⨁⨁⨁◯ Moderate^b	HD TRIPLE likely results in an increase in ACQ responders at 12 months compared to MD‐ICS/LABA.
ACQ responders at 12 months ‐ MD TRIPLE vs HD‐ICS/LABA № of participants: 1631 (1 RCT)	RR 0.97 (0.91 to 1.03)	75.3%	73.1% (68.5 to 77.6)	2.3% fewer (6.8 fewer to 2.3 more)	⨁⨁⨁◯ Moderate^b	MD TRIPLE likely results in little to no difference in ACQ responders at 12 months compared to HD‐ICS/LABA.
ACQ responders at 12 months ‐ HD TRIPLE vs HD‐ICS/LABA № of participants: 3982 (3 RCTs)	RR 1.11 (0.99 to 1.23)	64.2%	71.3% (63.6 to 79)	7.1% more (0.6 fewer to 14.8 more)	⨁◯◯◯ Very low^{b, c, d}	HD TRIPLE may increase ACQ responders at 12 months compared to HD‐ICS/LABA, but the evidence is very uncertain.
ACQ responders at 12 months ‐ HD TRIPLE vs MD TRIPLE № of participants: 1089 (1 RCT)	RR 1.08 (1.01 to 1.16)	72.8%	78.6% (73.5 to 84.5)	5.8% more (0.7 more to 11.6 more)	⨁⨁⨁◯ Moderate^b	HD TRIPLE likely increases ACQ responders at 12 months compared to MD TRIPLE.
ACQ responders at 12 months ‐ TRIPLE vs DUAL № of participants: 6204 (4 RCTs)	RR 1.07 (0.99 to 1.17)	64.8%	69.4% (64.2 to 75.8)	4.5% more (0.6 fewer to 11 more)	⨁◯◯◯ Very low^{b, c, d}	TRIPLE may increase ACQ responders at 12 months compared to DUAL, but the evidence is very uncertain.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
Explanations a. van Zyl‐Smit 2020 had very high attrition rates and is considered at high risk of bias. However, excluding the study did not change the results. b. Optimal information size is not met (Guyatt 2011b) c. Substantial heterogeneity I2 > 50% to 90% d. Confidence interval includes the line of no effect e. Considerable heterogeneity. I2 >75% to 100% f. Confidence intervals include clinically important outcomes.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: All‐cause serious adverse events (SAEs) Setting(s): Outpatient					Geometry of the Network in Figure 10*
Total studies: 13 RCTs Total Participants: 144476	Risk ratio (95% CrI)**	Anticipated absolute effect(95% CrI)*		Certainty of the evidence		Ranking** (95% CrI)**	Interpretation of Findings
		*With intervention*	Difference compared to MD‐ICS/LABA
HD‐ICS/LABA (Direct evidence; 8 RCTs; 7511 participants)	1.06 (0.86 to 1.33)	54 per 1000	3 per 1000 more (from 7 fewer to 16 more)	⊕⊕⊕⊕ High		3.0 (1.0 to 4.0)	Little or no difference
MD‐TRIPLE (Direct evidence; 3 RCTs; 3187 participants)	1.10 (0.84 to 1.45)	56 per 1000	5 per 1000 more (from 8 fewer to 21 more)	⊕⊕⊕◯ Moderate Due to imprecision¹		3.0 (1.0 to 4.0)	Probably little or no difference
HD‐TRIPLE (Direct evidence; 2 RCT; 2039 participants)	1.05 (0.81 to 1.64)	54 per 1000	3 per 1000 more (from 10 fewer to 33 more)	⊕⊕⊕◯ Moderate Due to imprecision¹		2.0 (1.0 to 4.0)	Probably little or no difference
MD‐ICS/LABA	Reference Comparator	51 per 1000²	Reference Comparator	Reference Comparator		2.0 (1.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Network Meta‐Analysis estimates are reported as risk ratio. Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Anticipated absolute effect. Anticipated absolute effect compares two rates by calculating the difference between the rates of the intervention group with the rate of MD‐ICS/LABA group. **** Median and credible intervals are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ Serious imprecision due to wide confidence intervals in the direct and/or indirect estimate(s). ² Based on the average rate in participants treated with MD‐ICS/LABA in the included studies.

Patient or population: Adolescent sand adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Asthma‐related serious adverse events (SAEs) Setting(s): Outpatient						Geometry of the Network in Figure 11*
Total studies: 11 RCTs Total Participants: 13209	Relative risk (95% CrI)**	Anticipated absolute effect(95% CrI)*		Certainty of the evidence	Ranking** (95% CrI)**		Interpretation of Findings
		*With intervention*	Difference compared to MD‐ICS/LABA
HD‐ICS/LABA (Direct evidence; 6 RCTs; 6244 participants)	1.27 (0.79 to 2.05)	13 per 1000	3 per 1000 more (from 2 fewer to 12 more)	⊕⊕⊕⊕ High	3.0 (1.0 to 4.0)		Little or no difference
MD‐TRIPLE (Direct evidence; 3 RCTs; 3188 participants)	1.70 (0.99 to 1.8)	18 per 1000	8 per 1000 more (from 0 fewer to 9 more)	⊕⊕⊕◯ Moderate Due to imprecision¹	4.0 (2.0 to 4.0)		Probably little or no difference
HD‐TRIPLE (Direct evidence; 2 RCTs; 2039 participants)	1.05 (0.60 to 1.80)	11 per 1000	1 per 1000 more (from 4 fewer to 9 more)	⊕⊕⊕⊕ High	2.0 (1.0 to 3.0)		Little or no difference
MD‐ICS/LABA	Reference Comparator	10 per 1000²	Reference Comparator	Reference Comparator	1.0 (1.0 to 3.0)		Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Network Meta‐Analysis estimates are reported as risk ratio. Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Anticipated absolute effect. Anticipated absolute effect compares two risks by calculating the difference between the risk of the intervention group with the risk of the control group. **** Median and credible intervals are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ Serious imprecision due to wide confidence intervals in the direct and/or indirect estimate(s). ² Based on the average rate in participants treated with MD‐ICS/LABA in the included studies.

Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: All‐cause adverse events (AEs) Setting(s): Outpatient					Geometry of the Network in Figure 12*
Total studies: 12 RCTs Total Participants: 12915	Relative effect (95% CrI)**	Anticipated absolute effect(95% CrI)*		Certainty of the evidence		Ranking** (95% CrI)**	Interpretation of Findings
		*With intervention*	Difference compared to MD‐ICS/LABA
HD‐ICS/LABA (Direct evidence; 7 RCTs; 5949 participants)	1.00 (0.89 to 1.12)	508 per 1000	0 per 1000 more (from 29 fewer to 28 more)	⊕⊕⊕⊕ High		3.0 (2.0 to 4.0)	Little or no difference
MD‐TRIPLE (Direct evidence; 3 RCTs; 3188 participants)	0.89 (0.78 to 1.02)	479 per 1000	29 per 1000 fewer (from 62 fewer to 5 more)	⊕⊕⊕◯ Moderate Due to imprecision¹		2.0 (1.0 to 3.0)	Probably little or no difference
HD‐TRIPLE (Direct evidence; 2 RCTs; 2039 participants)	0.79 (0.69 to 0.90)	449 per 1000	59 per 1000 fewer (from 26 fewer to 92 fewer)	⊕⊕⊕⊕ High		1.0 (1.0 to 2.0)	Superior
MD‐ICS/LABA	Reference Comparator	508 per 1000²	Reference Comparator	Reference Comparator		3.0 (2.0 to 4.0)	Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Network Meta‐Analysis estimates are reported as risk ratio. Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Anticipated absolute effect. Anticipated absolute effect compares two rates by calculating the difference between the rates of the intervention group with the rate of MD‐ICS/LABA group. **** Median and credible intervals are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ Serious imprecision due to 95% CIs including the null effect. ² Based on the average rate in participants treated with MD‐ICS/LABA in the included studies.

‐Patient or population: Adolescents and adults with symptomatic asthma Interventions: HD‐ICS/LABA, MD‐TRIPLE, HD‐TRIPLE Comparator (reference): Medium‐Dose ICS/LABA (MD‐ICS/LABA) Outcome: Dropouts due to adverse events (AEs) Setting(s): Outpatient						Geometry of the Network in Figure 13*
Total studies: 12 RCTs Total Participants: 12915	Risk ratio** (95% CrI)	Anticipated absolute effect***(95% CrI)		Certainty of the evidence	Ranking**** (95% CrI)		Interpretation of Findings
		With intervention	Difference compared to MD‐ICS/LABA
HD‐ICS/LABA (Direct evidence; 7 RCTs; 5969 participants)	0.91 (0.64 to 1.36)	15 per 1000	1 per 1000 fewer (from 6 fewer to 5 more)	⊕⊕⊕⊕ High	3.0 (2.0 to 4.0)		Little or no difference
MD‐TRIPLE (Direct evidence; 3 RCTs; 3205 participants)	0.88 (0.53 to 1.43)	14 per 1000	2 per 1000 fewer (from 8 fewer to 7 more)	⊕⊕⊕◯ Moderate Due to imprecision¹	3.0 (2.0 to 4.0)		Probably little or no difference
HD‐TRIPLE (Direct evidence; 2 RCTs; 2051 participants)	0.50 (0.30 to 0.84)	8 per 1000	8 per 1000 fewer (from 11 fewer to 3 fewer)	⊕⊕⊕◯ Moderate Due to imprecision¹	1.0 (1.0 to 2.0)		Probably superior
MD‐ICS/LABA	Reference Comparator	16 per 1000²	Reference Comparator	Reference Comparator	4.0 (2.0 to 4.0)		Reference Comparator
NMA‐SoF table definitions * The size of the nodes and the thickness of edges depend on the number of people randomised and the number of trials conducted, respectively. Network Meta‐Analysis estimates are reported as risk ratio. Results are expressed in credible intervals as opposed to the confidence intervals since a Bayesian analysis has been conducted. * Anticipated absolute effect. Anticipated absolute effect compares two rates by calculating the difference between the rates of the intervention group with the rate of MD‐ICS/LABA group. **** Median and credible intervals are presented. Rank statistics is defined as the probabilities that a treatment out of n treatments in a network meta‐analysis is the best, the second, the third and so on until the least effective treatment.
GRADE Working Group grades of evidence (or certainty in the evidence) High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanatory Footnotes ¹ Serious imprecision due to wide confidence intervals in the direct and/or indirect estimate(s). ² Based on the average rate in participants treated with MD‐ICS/LABA in the included studies.

	Overall			High Risk Subgroup			Low Risk Subgroup
Treatments	Mean Rank	Median Rank	95% CrI	Mean Rank	Median Rank	95% CrI	Mean Rank	Median Rank	95% CrI
MD‐ICS/LABA	1.55	1.0	(1.0, 3.0)	1.077	1.0	(1.0, 2.0)	3.05	3.0	(1.0, 4.0)
HD Triple	1.97	2.0	(1.0, 4.0)	3.223	3.0	(2.0, 4.0)	1.64	1.0	(1.0, 4.0)
HD‐ICS/LABA	3.01	3.0	(1.0, 4.0)	3.679	4.0	(3.0, 4.0)	2.25	2.0	(1.0, 4.0)
MD Triple	3.47	4.0	(1.0, 4.0)	2.021	2.0	(1.0, 4.0)	3.06	3.0	(1.0, 4.0)

	Lower Threshold		Upper Threshold
Comparison	New Optimal Treatment	Change in lnHR	New Optimal Treatment	Change in lnHR
HD‐ICS/LABA vs. MD‐ICS/LABA	HD Triple	‐0.220	N/A	Inf
MD Triple vs. MD‐ICS/LABA	HD Triple	‐0.481	N/A	Inf
HD Triple vs. MD‐ICS/LABA	HD Triple	‐0.958	N/A	Inf
MD Triple vs. HD‐ICS/LABA	MD Triple	‐3.574	HD Triple	1.190
HD Triple vs. HD‐ICS/LABA	HD Triple	‐0.186	HD‐ICS/LABA	2.013
HD Triple vs. MD Triple	HD Triple	‐0.807	MD Triple	4.990

Comparison	Median HR (95% CrI)
BUD/FM 320 vs. FP/SAL 250	0.40 (0.01, 5.33)
FF/VI 100 vs. FP/SAL 250	0.38 (0.01, 4.50)
MF/IND 160 vs. FP/SAL 250	3.00 (0.06, 78.87)*
FP/FM 250 vs. FP/SAL 250	0.07 (0.0001, 5.64)
FP/SAL 200 vs. FP/SAL 250	9.34 (0.59, 464.10)*
FP/SAL 500 vs. FP/SAL 250	6.92 (1.61, 80.79)*
BUD/FM 640 vs. FP/SAL 250	0.11 (0.002, 3.51)
FF/VI 200 vs. FP/SAL 250	0.26 (0.01, 4.13)
MF/IND 320 vs. FP/SAL 250	11.70 (1.00, 237.80)*
MF/GLY/IND 80 vs. FP/SAL 250	5.77 (0.10, 170.20)*
FF/UMEC/VI 100 vs. FP/SAL 250	0.37 (0.010, 5.52)
MF/GLY/IND 160 vs. FP/SAL 250	7.10 (1.17, 93.56)*
FF/UMEC/VI 200 vs. FP/SAL 250	0.20 (0.005, 3.24)
FF/VI 100 vs. BUD/FM 320	0.94 (0.012, 62.44)*
MF/IND 160 vs. BUD/FM 320	7.87 (0.06, 812.80)*
FP/FM 250 vs. BUD/FM 320	0.20 (0.0004, 5.64)
FP/SAL 200 vs. BUD/FM 320	26.98 (0.53, 3628)*
FP/SAL 500 vs. BUD/FM 320	19.87 (0.88, 1034)*
BUD/FM 640 vs. BUD/FM 320	0.29 (0.03, 2.82)
FF/VI 200 vs. BUD/FM 320	0.64 (0.007, 50.68)*
MF/IND 320 vs. BUD/FM 320	32.17 (0.84, 2463)*
MF/GLY/IND 80 vs. BUD/FM 320	14.87 (0.11, 1687)*
FF/UMEC/VI 100 vs. BUD/FM 320	0.92 (0.01, 70.48)*
MF/GLY/IND 160 vs. BUD/FM 320	19.92 (0.74, 1124)*
FF/UMEC/VI 200 vs. BUD/FM 320	0.49 (0.006, 38.79)
MF/IND 160 vs. FF/VI 100	8.54 (0.08, 785.50)*
FP/FM 250 vs. FF/VI 100	0.18 (0.0002, 46.45)
FP/SAL 200 vs. FF/VI 100	28.91 (0.60, 3650)*
FP/SAL 500 vs. FF/VI 100	21.04 (0.93, 1036)*
BUD/FM 640 vs. FF/VI 100	0.32 (0.003, 39.12)
FF/VI 200 vs. FF/VI 100	0.69 (0.20, 2.21)
MF/IND 320 vs. FF/VI 100	34.58 (0.84, 2500)*
MF/GLY/IND 80 vs. FF/VI 100	16.25 (0.13, 1628)*
FF/UMEC/VI 100 vs. FF/VI 100	0.99 (0.34, 2.93)
MF/GLY/IND 160 vs. FF/VI 100	21.23 (0.79, 1173)*
FF/UMEC/VI 200 vs. FF/VI 100	0.54 (0.14,1.84)
FP/FM 250 vs. MF/IND 160	0.02 (0.00001, 8.51)
FP/SAL 200 vs. MF/IND 160	3.35 (0.09, 372.80)*
FP/SAL 500 vs. MF/IND 160	2.41 (0.20, 100.10)*
BUD/FM 640 vs. MF/IND 160	0.04 (0.0002, 8.01)
FF/VI 200 vs. MF/IND 160	0.08 (0.0007, 10.23)
MF/IND 320 vs. MF/IND 160	3.79 (0.41, 147.00)*
MF/GLY/IND 80 vs. MF/IND 160	1.89 (0.80, 4.47)
FF/UMEC/VI 100 vs. MF/IND 160	0.12 (0.001, 14.12)
MF/GLY/IND 160 vs. MF/IND 160	2.45 (0.17, 109.20)*
FF/UMEC/VI 200 vs. MF/IND 160	0.06 (0.0006, 8.22)
FP/SAL 200 vs. FP/FM 250	170.90 (0.74, 284800)*
FP/SAL 500 vs. FP/FM 250	121.20 (0.98, 127600)*
BUD/FM 640 vs. FP/FM 250	1.61 (0.03, 949.80)*
FF/VI 200 vs. FP/FM 250	3.81 (0.01, 5159)*
MF/IND 320 vs. FP/FM 250	200.60 (1.10, 272600)*
MF/GLY/IND 80 vs. FP/FM 250	91.20 (0.21, 145100)*
FF/UMEC/VI 100 vs. FP/FM 250	5.51 (0.02, 7370)*
MF/GLY/IND 160 vs. FP/FM 250	122.10 (0.86, 136600)*
FF/UMEC/VI 200 vs. FP/FM 250	2.96 (0.01, 4057)*
FP/SAL 500 vs. FP/SAL 200	0.82 (0.03, 8.69)
BUD/FM 640 vs. FP/SAL 200	0.01 (0.0001, 1.03)
FF/VI 200 vs. FP/SAL 200	0.02 (0.0002, 1.38)
MF/IND 320 vs. FP/SAL 200	1.25 (0.03, 30.15)
MF/GLY/IND 80 vs. FP/SAL 200	0.56 (0.005, 23.05)
FF/UMEC/VI 100 vs. FP/SAL 200	0.03 (0.0002, 1.91)
MF/GLY/IND 160 vs. FP/SAL 200	0.81 (0.03, 10.71)
FF/UMEC/VI 200 vs. FP/SAL 200	0.02 (0.0001, 1.13)
BUD/FM 640 vs. FP/SAL 500	0.01 (0.0002, 0.71)
FF/VI 200 vs. FP/SAL 500	0.03 (0.001, 0.93)
MF/IND 320 vs. FP/SAL 500	1.57 (0.24, 12.53)
MF/GLY/IND 80 vs. FP/SAL 500	0.77 (0.02, 10.92)
FF/UMEC/VI 100 vs. FP/SAL 500	0.05 (0.001, 1.26)
MF/GLY/IND 160 vs. FP/SAL 500	1.00 (0.37, 2.68)
FF/UMEC/VI 200 vs. FP/SAL 500	0.02 (0.0004, 0.73)
FF/VI 200 vs. BUD/FM 640	2.15 (0.01, 299.50)*
MF/IND 320 vs. BUD/FM 640	112.30 (1.463, 15230)*
MF/GLY/IND 80 vs. BUD/FM 640	51.28 (0.23, 9561)*
FF/UMEC/VI 100 vs. BUD/FM 640	3.09 (0.02, 410.20)*
MF/GLY/IND 160 vs. BUD/FM 640	69.57 (1.25, 6911)*
FF/UMEC/VI 200 vs. BUD/FM 640	1.66 (0.01, 228.10)*
MF/IND 320 vs. FF/VI 200	51.33 (1.03, 4278)*
MF/GLY/IND 80 vs. FF/VI 200	23.70 (0.17, 2801)*
FF/UMEC/VI 100 vs. FF/VI 200	1.43 (0.44,4.97)
MF/GLY/IND 160 vs. FF/VI 200	31.90 (0.94, 2048)*
FF/UMEC/VI 200 vs. FF/VI 200	0.78 (0.18, 3.10)
MF/GLY/IND 80 vs. MF/IND 320	0.49 (0.01, 5.52)
FF/UMEC/VI 100 vs. MF/IND 320	0.03 (0.0003, 1.39)
MF/GLY/IND 160 vs. MF/IND 320	0.64 (0.06, 5.34)
FF/UMEC/VI 200 vs. MF/IND 320	0.02 (0.0002, 0.80)
FF/UMEC/VI 100 vs. MF/GLY/IND 80	0.06 (0.0005, 8.15)
MF/GLY/IND 160 vs. MF/GLY/IND 80	1.31 (0.08, 62.71)*
FF/UMEC/VI 200 vs. MF/GLY/IND 80	0.03 (0.0003, 4.74)
MF/GLY/IND 160 vs. FF/UMEC/VI 100	21.94 (0.68, 1362)*
FF/UMEC/VI 200 vs. FF/UMEC/VI 100	0.55 (0.14, 1.89)
FF/UMEC/VI 200 vs. MF/GLY/IND/160	0.02 (0.0004, 0.85)

Comparison	Median Mean Difference (95% CrI)
HD‐ICS/LABA vs MD‐ICS/LABA	0.008 (‐0.053, 0.069)
MD Triple vs. MD‐ICS/LABA	‐0.056 (‐0.141, 0.029)
HD Triple vs MD‐ICS/LABA	‐0.094 (‐0.178,‐0.011)
MD Triple vs. HD‐ICS/LABA	‐0.064 (‐0.149, 0.022)
HD Triple vs HD‐ICS/LABA	‐0.103 (‐0.187, ‐0.018)
HD Triple vs MD Triple	‐0.039 (‐0.111, 0.034)

Model	p	Mean Difference (95% CrI)
MD Triple vs. MD‐ICS/LABA
Direct	0.472	‐0.092 (‐0.231, 0.047)
Indirect		0.003 (‐0.302, 0.283)
Network		‐0.071 (‐0.173, 0.026)
HD Triple vs. MD‐ICS/LABA
Direct	0.733	‐0.108 (‐0.229, 0.016)
Indirect		‐0.075 (‐0.295, 0.128)
Network		‐0.103 (‐0.204, ‐0.015)
MD Triple vs. HD‐ICS/LABA
Direct	0.266	‐0.007 (‐0.115, 0.101)
Indirect		‐0.121 (‐0.323, 0.077)
Network		‐0.038 (‐0.135, 0.050)

PERMALINK

Effectiveness and tolerability of dual and triple combination inhaler therapies compared with each other and varying doses of inhaled corticosteroids in adolescents and adults with asthma: a systematic review and network meta‐analysis

Yuji Oba

Sumayya Anwer

Tinashe Maduke

Tarang Patel

Sofia Dias

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings 1. NMA Summary of Findings for severe exacerbations (asthma‐related hospitalisations).

1.

Summary of findings 2. Asthma exacerbations ‐ pairwise comparisons.

Summary of findings 3. NMA Summary of Findings for moderate to severe (steroid‐requiring) exacerbations.

2.

Summary of findings 4. NMA Summary of Findings for change from baseline in ACQ scores at 3 months.

3.

Summary of findings 5. NMA Summary of Findings for change from baseline in ACQ scores at 6 months.

4.

Summary of findings 6. NMA Summary of Findings for change from baseline in ACQ scores at 12 months.

5.

Summary of findings 7. Asthma Control Questionnaire: change from baseline ‐ pairwise comparisons ‡.

Summary of findings 8. NMA Summary of Findings for change from baseline in AQLQ scores at 6 months.

6.

Summary of findings 9. NMA Summary of Findings for change from baseline in AQLQ scores at 12 months.

7.

Summary of findings 10. Asthma Quality of Life Questionnaire: change from baseline ‐ pairwise comparisons ‡.

Summary of findings 11. NMA Summary of Findings for ACQ responders at 6 months.

8.

Summary of findings 12. NMA Summary of Findings for ACQ responders at 12 months.

9.

Summary of findings 13. Asthma Control Questionnaire responders ‐ pairwise comparisons.

Summary of findings 14. NMA Summary of Findings for all‐cause SAEs.

10.

Summary of findings 15. Serious adverse events, adverse events, and dropouts due to adverse event ‐ pairwise comparisons.

Summary of findings 16. NMA Summary of Findings for asthma‐related SAEs.

11.

Summary of findings 17. NMA Summary of Findings for all‐cause AEs.

12.

Summary of findings 18. NMA summary of findings table for dropouts due to AEs.

13.

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Assessment of bias in conducting the systematic review

Network meta‐analysis

Direct pairwise meta‐analysis

Unit of analysis issues

Dealing with missing data

Network meta‐analysis

Direct pairwise meta‐analysis

Network meta‐analysis

Direct pairwise meta‐analysis

Network meta‐analysis (NMA)

Direct pairwise meta‐analysis

Summary of findings 7. Asthma Control Questionnaire: change from baseline ‐ pairwise comparisons ^‡.

Summary of findings 10. Asthma Quality of Life Questionnaire: change from baseline ‐ pairwise comparisons ^‡.

Model	p	LORs (95% CrI)
HD‐ICS/LABA vs. MD‐ICS/LABA
Direct	0.360	0.080 (‐0.142, 0.306)
Indirect		‐0.14 (‐0.664, 0.373)
Network		0.046 (‐0.150, 0.236)
MD Triple vs. MD‐ICS/LABA
Direct	0.402	0.207 (‐0.026, 0.441)
Indirect		0.472 (‐0.167, 1.117)
Network		0.228 (0.028, 0.432)
HD Triple vs. MD‐ICS/LABA
Direct	0.720	0.247 (‐0.114, 0.612)
Indirect		0.156 (‐0.327, 0.620)
Network		0.216 (0.005, 0.425)
MD Triple vs. HD‐ICS/LABA
Direct	0.267	0.089 (‐0.218, 0.408)
Indirect		0.343 (‐0.058, 0.781)
Network		0.183 (‐0.020, 0.394)
HD Triple vs. HD‐ICS/LABA
Direct	0.391	0.185 (‐0.006, 0.383)
Indirect		‐0.077 (‐0.718, 0.584)
Network		0.172 (‐0.001, 0.343)
HD Triple vs. MD Triple
Direct	0.359	‐0.061 (‐0.305, 0.171)
Indirect		0.158 (‐0.339, 0.672)
Network		‐0.011 (‐0.222, 0.188)

Comparison	Odds Ratio (95% CrI)
HD‐ICS/LABA vs MD‐ICS/LABA	1.052 (0.919, 1.203)
MD Triple vs. MD‐ICS/LABA	1.248 (1.086, 1.437)
HD Triple vs MD‐ICS/LABA	1.246 (1.073, 1.446)
MD Triple vs. HD‐ICS/LABA	1.187 (1.030, 1.370)
HD Triple vs HD‐ICS/LABA	1.184 (1.054, 1.331)
HD Triple vs MD Triple	0.998 (0.861, 1.155)

Treatments	Mean Rank	Median	95% CrI
MF/IND 320	1.78	1	(1.00, 4.00)
MF/GLY/IND 80	2.17	2	(1.00, 5.00)
MF/IND 160	3.44	3	(1.00, 6.00)
MF/GLY/IND 160	3.54	4	(1.00, 6.00)
FP/SAL 500	4.89	5	(3.00, 6.00)
FP/SAL 500 + Tio	5.17	6	(1.00, 6.00)

Model	p	LOR (95% CrI)
HD Triple vs. MD‐ICS/LABA
Direct	0.804	0.310 (‐0.404, 1.018)
Indirect		0.203 (‐0.566, 0.960)
Network		0.263 (‐0.172, 0.672)
MD Triple vs. HD‐ICS/LABA
Direct	0.412	‐0.129 (‐0.775, 0.527)
Indirect		0.228 (‐0.599, 1.077)
Network		0.014 (‐0.408, 0.469)
HD Triple vs. MD Triple
Direct	0.752	0.327 (‐0.384, 1.033)
Indirect		0.167 (‐0.822, 1.143)
Network		0.286 (‐0.192, 0.747)

Fixed Effects Model
Treatments	Mean Rank	Median Rank	95% CrI
HD Triple	1.01	1	(1.00, 1.00)
MD‐ICS/LABA	2.88	3	(2.00, 4.00)
HD‐ICS/LABA	2.99	3	(2.00, 4.00)
MD Triple	3.11	3	(2.00, 4.00)
Random Effects Model
Treatments	Mean Rank	Median Rank	95% CrI
HD Triple	1.09	1	(1.00, 2.00)
MD‐ICS/LABA	2.81	3	(1.00, 4.00)
MD Triple	3.04	3	(1.00, 4.00)
HD‐ICS/LABA	3.07	3	(2.00, 4.00)

Model	p	LORs (95% CrI)
HD‐ICS/LABA vs. MD‐ICS/LABA
Direct	0.410	‐0.005 (‐0.364, 0.298)
Indirect		0.388 (‐0.612, 1.359)
Network		0.044 (‐0.232, 0.298)
MD Triple vs. MD‐ICS/LABA
Direct	0.237	0.166 (‐0.258, 0.563)
Indirect		‐0.409 (‐1.401, 0.485)
Network		0.087 (‐0.250, 0.402)
HD Triple vs. MD‐ICS/LABA
Direct	0.961	0.035 (‐0.502, 0.555)
Indirect		0.012 (‐0.565, 0.543)
Network		0.039 (‐0.293, 0.349)
MD Triple vs. HD‐ICS/LABA
Direct	0.746	0.083 (‐0.424, 0.587)
Indirect		‐0.038 (‐0.656, 0.625)
Network		0.042 (‐0.276, 0.368)
HD Triple vs. HD‐ICS/LABA
Direct	0.248	‐0.051 (‐0.379, 0.267)
Indirect		0.511 (‐0.442, 1.523)
Network		‐0.005 (‐0.282, 0.261)
HD Triple vs. MD Triple
Direct	0.456	0.013 (‐0.390, 0.438)
Indirect		‐0.340 (‐1.315, 0.547)
Network		‐0.050 (‐0.381, 0.270)

Comparison	Odds Ratio (95% CrI)
FF/ VI 200 vs. FP/SAL 250	0.451 (0.019, 4.920)
MF/FM 200 vs. FP/SAL 250	0.836 (0.024, 22.820)
MF/IND 160 vs. FP/SAL 250	2.724 (0.634, 14.430)
FP/SAL 200 vs. FP/SAL 250	3.846 (0.808, 26.520)
FP/SAL 500 vs. FP/SAL 250	2.926 (0.924, 12.610)
FF/VI 200 vs. FP/SAL 250	0.376 (0.014, 5.295)
MF/FM 400 vs. FP/SAL 250	3.166 (0.028, 889.500)
MF/IND 320 vs. FP/SAL 250	4.102 (1.003, 21.370)
MF/GLY/IND 80 vs. FP/SAL 250	5.028 (1.228, 26.260)
FF/UMEC/VI 100 vs. FP/SAL 250	0.446 (0.017, 6.139)
MF/GLY/IND 160 vs. FP/SAL 250	3.105 (0.722, 16.610)
FF/UMEC/VI 200 vs. FP/SAL 250	0.239 (0.008, 3.558)
FP/SAL 500 + Tio vs. FP/SAL 250	2.025 (0.182, 18.300)
MF/FM 200 vs. FF/VI 100	1.932 (0.029, 169.300)*
MF/IND 160 vs. FF/VI 100	6.296 (0.358, 218.700)*
FP/SAL 200 vs. FF/VI 100	8.997 (0.478, 337.100)*
FP/SAL 500 vs. FF/VI 100	6.779 (0.440, 210.900)*
FF/VI 200 vs. FF/VI 100	0.852 (0.265, 2.621)
MF/FM 400 vs. FF/VI 100	7.663 (0.036, 4,274.000)*
MF/IND 320 vs. FF/VI 100	9.489 (0.544, 319.600)*
MF/GLY/IND 80 vs. FF/VI 100	11.630 (0.668, 391.800)*
FF/UMEC/VI 100 vs. FF/VI 100	1.005 (0.336, 2.986)
MF/GLY/IND 160 vs. FF/VI 100	7.182 (0.405, 247.500)*
FF/UMEC/VI 200 vs. FF/VI 100	0.548 (0.137, 1.868)
FP/SAL 500 + Tio vs. FF/VI 100	4.602 (0.150, 211.300)*
MF/IND 160 vs. MF/FM 200	3.361 (0.089, 152.900)*
FP/SAL 200 vs. MF/FM 200	4.809 (0.120, 246.900)*
FP/SAL 500 vs. MF/FM 200	3.637 (0.106, 153.300)*
FF/VI 200 vs. MF/FM 200	0.438 (0.004, 33.710)
MF/FM 400 vs. MF/FM 200	3.487 (0.141, 500.800)*
MF/IND 320 vs. MF/FM 200	5.069 (0.136, 227.900)*
MF/GLY/IND 80 vs. MF/FM 200	6.216 (0.168, 280.300)*
FF/UMEC/VI 100 vs. MF/FM 200	0.520 (0.005, 39.290)
MF/GLY/IND 160 vs. MF/FM 200	3.813 (0.101, 174.100)*
FF/UMEC/VI 200 vs. MF/FM 200	0.278 (0.003, 22.100)
FP/SAL 500 + Tio vs. MF/FM 200	2.418 (0.042, 145.300)*
FP/SAL 200 vs. MF/IND 160	1.404 (0.278, 8.683)
FP/SAL 500 vs. MF/IND 160	1.086 (0.452, 2.624)
FF/VI 200 vs. MF/IND 160	0.134 (0.003, 2.880)
MF/FM 400 vs. MF/IND 160	1.141 (0.008, 395.800)*
MF/IND 320 vs. MF/IND 160	1.502 (0.673, 3.506)
MF/GLY/IND 80 vs. MF/IND 160	1.837 (0.825, 4.300)
FF/UMEC/VI 100 vs. MF/IND 160	0.158 (0.004, 3.366)
MF/GLY/IND 160 vs. MF/IND 160	1.135 (0.466, 2.821)
FF/UMEC/VI 200 vs. MF/IND 160	0.085 (0.002, 1.946)
FP/SAL 500 + Tio vs. MF/IND 160	0.745 (0.089, 4.096)
FP/SAL 500 vs. FP/SAL 200	0.782 (0.154, 3.019)
FF/VI 200 vs. FP/SAL 200	0.093 (0.002, 2.142)
MF/FM 400 vs. FP/SAL 200	0.797 (0.005, 285.000)*
MF/IND 320 vs. FP/SAL 200	1.076 (0.181, 5.240)
MF/GLY/IND 80 vs. FP/SAL 200	1.313 (0.222, 6.467)
FF/UMEC/VI 100 vs. FP/SAL 200	0.111 (0.003, 2.519)
MF/GLY/IND 160 vs. FP/SAL 200	0.811 (0.132, 4.132)
FF/UMEC/VI 200 vs. FP/SAL 200	0.059 (0.001, 1.446)
FP/SAL 500 + Tio vs. FP/SAL 200	0.515 (0.038, 4.730)
FF/VI 200 vs. FP/SAL 500	0.123 (0.003, 2.364)
MF/FM 400 vs. FP/SAL 500	1.046 (0.007, 339.600)*
MF/IND 320 vs. FP/SAL 500	1.384 (0.633, 3.121)
MF/GLY/IND 80 vs. FP/SAL 500	1.690 (0.773, 3.868)
FF/UMEC/VI 100 vs. FP/SAL 500	0.146 (0.004, 2.754)
MF/GLY/IND 160 vs. FP/SAL 500	1.046 (0.437, 2.538)
FF/UMEC/VI 200 vs. FP/SAL 500	0.079 (0.002, 1.591)
FP/SAL 500 + Tio vs. FP/SAL 500	0.688 (0.081, 3.717)
MF/FM 400 vs. FF/VI 200	9.133 (0.039, 5,439.000)*
MF/IND 320 vs. FF/VI 200	11.330 (0.526, 441.100)*
MF/GLY/IND 80 vs. FF/VI 200	13.890 (0.642, 542.400)*
FF/UMEC/VI 100 vs. FF/VI 200	1.181 (0.381, 3.783)
MF/GLY/IND 160 vs. FF/VI 200	8.546 (0.392, 338.100)*
FF/UMEC/VI 200 vs. FF/VI 200	0.645 (0.158, 2.346)
FP/SAL 500 + Tio vs. FF/VI 200	5.441 (0.148, 286.500)*
MF/IND 320 vs. MF/FM 400	1.324 (0.004, 199.100)*
MF/GLY/IND 80 vs. MF/FM 400	1.622 (0.005, 241.900)*
FF/UMEC/VI 100 vs. MF/FM 400	0.130 (0.0002, 29.950)
MF/GLY/IND 160 vs. MF/FM 400	1.002 (0.003, 152.000)*
FF/UMEC/VI 200 vs. MF/FM 400	0.069 (0.0001, 16.970)
FP/SAL 500 + Tio vs. MF/FM 400	0.617 (0.001, 121.300)*
MF/GLY/IND 80 vs. MF/IND 320	1.223 (0.587, 2.576)
FF/UMEC/VI 100 vs. MF/IND 320	0.105 (0.003, 2.221)
MF/GLY/IND 160 vs. MF/IND 320	0.756 (0.328, 1.701)
FF/UMEC/VI 200 vs. MF/IND 320	0.056 (0.001, 1.273)
FP/SAL 500 + Tio vs. MF/IND 320	0.495 (0.060, 2.606)
FF/UMEC/VI 100 vs. MF/GLY/IND 80	0.085 (0.002, 1.815)
MF/GLY/IND 160 vs. MF/GLY/IND 80	0.620 (0.289, 1.274)
FF/UMEC/VI 200 vs. MF/GLY/IND 80	0.046 (0.001, 1.040)
FP/SAL 500 + Tio vs. MF/GLY/IND 80	0.409 (0.053, 1.874)
MF/GLY/IND 160 vs. FF/UMEC/VI 100	7.224 (0.334, 279.900)*
FF/UMEC/VI 200 vs. FF/UMEC/VI 100	0.546 (0.138, 1.869)
FP/SAL 500 + Tio vs. FF/UMEC/VI 100	4.605 (0.127, 240.700)*
FF/UMEC/VI 200 vs. MF/GLY/IND 160	0.075 (0.002, 1.725)
FP/SAL 500 + Tio vs. MF/GLY/IND 160	0.660 (0.084, 3.172)
FP/SAL 500 + Tio vs. FF/UMEC/VI 200	8.563 (0.221, 485.500)*

Comparison	Odds Ratio (95% CrI)
BUD/FM 320 vs. FP/SAL 250	1.912 (1.048, 3.607)
FF/VI 100 vs. FP/SAL 250	1.051 (0.770, 1.436)
MF/FM 200 vs. FP/SAL 250	1.058 (0.746, 1.503)
MF/IND 160 vs. FP/SAL 250	0.861 (0.357, 2.065)
FP/FM 250 vs. FP/SAL 250	2.177 (0.935, 5.196)
FP/SAL 200 vs. FP/SAL 250	0.899 (0.442, 1.819)
FP/SAL 500 vs. FP/SAL 250	0.944 (0.398, 2.227)
BUD/FM 640 vs. FP/SAL 250	2.885 (1.256, 6.663)
FF/VI 200 vs. FP/SAL 250	0.935 (0.631, 1.385)
MF/FM 400 vs. FP/SAL 250	1.498 (0.565, 4.177)
MF/IND 320 vs. FP/SAL 250	0.775 (0.321, 1.859)
MF/GLY/IND 80 vs. FP/SAL 250	0.773 (0.318, 1.864)
FF/UMEC/VI 100 vs. FP/SAL 250	1.063 (0.701, 1.618)
MF/GLY/IND 160 vs. FP/SAL 250	0.707 (0.291, 1.706)
FF/UMEC/VI 200 vs. FP/SAL 250	0.910 (0.597, 1.391)
FP/SAL 500 + Tio vs. FP/SAL 250	0.703 (0.283, 1.737)
FF/VI 100 vs. BUD/FM 320	0.550 (0.271, 1.083)
MF/FM 200 vs. BUD/FM 320	0.553 (0.269, 1.109)
MF/IND 160 vs. BUD/FM 320	0.449 (0.153, 1.306)
FP/FM 250 vs. BUD/FM 320	1.136 (0.629, 2.065)
FP/SAL 200 vs. BUD/FM 320	0.469 (0.183, 1.189)
FP/SAL 500 vs. BUD/FM 320	0.492 (0.170, 1.413)
BUD/FM 640 vs. BUD/FM 320	1.509 (0.851, 2.600)
FF/VI 200 vs. BUD/FM 320	0.488 (0.232, 1.002)
MF/FM 400 vs. BUD/FM 320	0.783 (0.245, 2.575)
MF/IND 320 vs. BUD/FM 320	0.404 (0.137, 1.177)
MF/GLY/IND 80 vs. BUD/FM 320	0.403 (0.136, 1.177)
FF/UMEC/VI 100 vs. BUD/FM 320	0.556 (0.261, 1.157)
MF/GLY/IND 160 vs. BUD/FM 320	0.369 (0.125, 1.078)
FF/UMEC/VI 200 vs. BUD/FM 320	0.476 (0.222, 0.994)
FP/SAL 500 +Tio vs. BUD/FM 320	0.366 (0.122, 1.090)
MF/FM 200 vs. FF/VI 100	1.006 (0.629, 1.609)
MF/IND 160 vs. FF/VI 100	0.818 (0.321, 2.073)
FP/FM 250 vs. FF/VI 100	2.071 (0.842, 5.219)
FP/SAL 200 vs. FF/VI 100	0.855 (0.393, 1.843)
FP/SAL 500 vs. FF/VI 100	0.897 (0.359, 2.230)
BUD/FM 640 vs. FF/VI 100	2.742 (1.129, 6.708)
FF/VI 200 vs. FF/VI 100	0.889 (0.698, 1.132)
MF/FM 400 vs. FF/VI 100	1.425 (0.509, 4.163)
MF/IND 320 vs. FF/VI 100	0.737 (0.289, 1.861)
MF/GLY/IND 80 vs. FF/VI 100	0.735 (0.287, 1.864)
FF/UMEC/VI 100 vs. FF/VI 100	1.012 (0.764, 1.338)
MF/GLY/IND 160 vs. FF/VI 100	0.672 (0.263, 1.705)
FF/UMEC/VI 200 vs. FF/VI 100	0.866 (0.651, 1.150)
FP/SAL 500 +Tio vs. FF/VI 100	0.668 (0.255, 1.733)
MF/IND 160 vs. MF/FM 200	0.814 (0.316, 2.085)
FP/FM 250 vs. MF/FM 200	2.057 (0.825, 5.235)
FP/SAL 200 vs. MF/FM 200	0.850 (0.385, 1.860)
FP/SAL 500 vs. MF/FM 200	0.892 (0.352, 2.248)
BUD/FM 640 vs. MF/FM 200	2.726 (1.105, 6.755)
FF/VI 200 vs. MF/FM 200	0.883 (0.522, 1.497)
MF/FM 400 vs. MF/FM 200	1.415 (0.570, 3.713)
MF/IND 320 vs. MF/FM 200	0.733 (0.284, 1.878)
MF/GLY/IND 80 vs. MF/FM 200	0.731 (0.282, 1.880)
FF/UMEC/VI 100 vs. MF/FM 200	1.005 (0.582, 1.737)
MF/GLY/IND 160 vs. MF/FM 200	0.668 (0.258, 1.723)
FF/UMEC/VI 200 vs. MF/FM 200	0.861 (0.497, 1.491)
FP/SAL 500 +Tio vs. MF/FM 200	0.665 (0.250, 1.751)
FP/FM 250 vs. MF/IND 160	2.534 (0.747, 8.695)
FP/SAL 200 vs. MF/IND 160	1.046 (0.512, 2.130)
FP/SAL 500 vs. MF/IND 160	1.097 (0.918, 1.310)
BUD/FM 640 vs. MF/IND 160	3.356 (1.002, 11.224)
FF/VI 200 vs. MF/IND 160	1.086 (0.417, 2.847)
MF/FM 400 vs. MF/IND 160	1.748 (0.470, 6.710)
MF/IND 320 vs. MF/IND 160	0.900 (0.756, 1.073)
MF/GLY/IND 80 vs. MF/IND 160	0.897 (0.735, 1.097)
FF/UMEC/VI 100 vs. MF/IND 160	1.237 (0.469, 3.277)
MF/GLY/IND 160 vs. MF/IND 160	0.822 (0.673, 1.003)
FF/UMEC/VI 200 vs. MF/IND 160	1.058 (0.401, 2.810)
FP/SAL 500 +Tio vs. MF/IND 160	0.817 (0.614, 1.086)
FP/SAL 200 vs. FP/FM 250	0.413 (0.135, 1.241)
FP/SAL 500 vs. FP/FM 250	0.433 (0.128, 1.447)
BUD/FM 640 vs. FP/FM 250	1.326 (0.583, 2.967)
FF/VI 200 vs. FP/FM 250	0.429 (0.165, 1.089)
MF/FM 400 vs. FP/FM 250	0.690 (0.186, 2.603)
MF/IND 320 vs. FP/FM 250	0.355 (0.104, 1.205)
MF/GLY/IND 80 vs. FP/FM 250	0.354 (0.103, 1.204)
FF/UMEC/VI 100 vs. FP/FM 250	0.489 (0.186, 1.251)
MF/GLY/IND 160 vs. FP/FM 250	0.324 (0.094, 1.102)
FF/UMEC/VI 200 vs. FP/FM 250	0.418 (0.159, 1.076)
FP/SAL 500 +Tio vs. FP/FM 250	0.322 (0.092, 1.113)
FP/SAL 500 vs. FP/SAL 200	1.049 (0.526, 2.096)
BUD/FM 640 vs. FP/SAL 200	3.212 (1.079, 9.581)
FF/VI 200 vs. FP/SAL 200	1.039 (0.465, 2.338)
MF/FM 400 vs. FP/SAL 200	1.671 (0.501, 5.785)
MF/IND 320 vs. FP/SAL 200	0.861 (0.423, 1.758)
MF/GLY/IND 80 vs. FP/SAL 200	0.858 (0.419, 1.768)
FF/UMEC/VI 100 vs. FP/SAL 200	1.184 (0.522, 2.701)
MF/GLY/IND 160 vs. FP/SAL 200	2.701 (0.383, 1.617)
FF/UMEC/VI 200 vs. FP/SAL 200	1.013 (0.446, 2.315)
FP/SAL 500 +Tio vs. FP/SAL 200	0.781 (0.370, 1.652)
BUD/FM 640 vs. FP/SAL 500	3.062 (0.927, 10.102)
FF/VI 200 vs. FP/SAL 500	0.991 (0.387, 2.554)
MF/FM 400 vs. FP/SAL 500	1.594 (0.434, 6.051)
MF/IND 320 vs. FP/SAL 500	0.821 (0.688, 0.980)
MF/GLY/IND 80 vs. FP/SAL 500	0.818 (0.669, 1.000)
FF/UMEC/VI 100 vs. FP/SAL 500	1.128 (0.435, 2.936)
MF/GLY/IND 160 vs. FP/SAL 500	0.750 (0.612, 0.916)
FF/UMEC/VI 200 vs. FP/SAL 500	0.965 (0.372, 2.522)
FP/SAL 500 +Tio vs. FP/SAL 500	0.745 (0.560, 0.991)
FF/VI 200 vs. BUD/FM 640	0.324 (0.129, 0.814)
MF/FM 400 vs. BUD/FM 640	0.521 (0.143, 1.947)
MF/IND 320 vs. BUD/FM 640	0.268 (0.080, 0.898)
MF/GLY/IND 80 vs. BUD/FM 640	0.267 (0.080, 0.900)
FF/UMEC/VI 100 vs. BUD/FM 640	0.368 (0.145, 0.937)
MF/GLY/IND 160 vs. BUD/FM 640	0.245 (0.073, 0.823)
FF/UMEC/VI 200 vs. BUD/FM 640	0.315 (0.124, 0.803)
FP/SAL 500 +Tio vs. BUD/FM 640	0.243 (0.071, 0.832)
MF/FM 400 vs. FF/VI 200	1.604 (0.558, 4.812)
MF/IND 320 vs. FF/VI 200	0.829 (0.316, 2.159)
MF/GLY/IND 80 vs. FF/VI 200	0.826 (0.314, 2.156)
FF/UMEC/VI 100 vs. FF/VI 200	1.138 (0.858, 1.511)
MF/GLY/IND 160 vs. FF/VI 200	0.756 (0.287, 1.975)
FF/UMEC/VI 200 vs. FF/VI 200	0.974 (0.731, 1.296)
FP/SAL 500 +Tio vs. FF/VI 200	0.752 (0.279, 2.004)
MF/IND 320 vs. MF/FM 400	0.515 (0.134, 1.916)
MF/GLY/IND 80 vs. MF/FM 400	0.514 (0.133, 1.922)
FF/UMEC/VI 100 vs. MF/FM 400	0.710 (0.234, 2.058)
MF/GLY/IND 160 vs. MF/FM 400	0.470 (0.122, 1.761)
FF/UMEC/VI 200 vs. MF/FM 400	0.607 (0.200, 1.767)
FP/SAL 500 +Tio vs. MF/FM 400	0.467 (0.120, 1.778)
MF/GLY/IND 80 vs. MF/IND 320	0.997 (0.817, 1.217)
FF/UMEC/VI 100 vs. MF/IND 320	1.373 (0.522, 3.644)
MF/GLY/IND 160 vs. MF/IND 320	0.912 (0.748, 1.113)
FF/UMEC/VI 200 vs. MF/IND 320	1.175 (0.446, 3.122)
FP/SAL 500 +Tio vs. MF/IND 320	0.907 (0.682, 1.205)
FF/UMEC/VI 100 vs. MF/GLY/IND 80	1.378 (0.521, 3.665)
MF/GLY/IND 160 vs. MF/GLY/IND 80	0.915 (0.771, 1.086)
FF/UMEC/VI 200 vs. MF/GLY/IND 80	1.179 (0.445, 3.146)
FP/SAL 500 +Tio vs. MF/GLY/IND 80	0.910 (0.718, 1.152)
MF/GLY/IND 160 vs. FF/UMEC/VI 100	0.664 (0.249, 1.758)
FF/UMEC/VI 200 vs. FF/UMEC/VI 100	0.856 (0.636, 1.150)
FP/SAL 500 +Tio vs. FF/UMEC/VI 100	0.660 (0.242, 1.784)
FF/UMEC/VI 200 vs. MF/GLY/IND 160	1.288 (0.486, 3.437)
FP/SAL 500 +Tio vs. MF/GLY/IND 160	0.994 (0.784, 1.258)
FP/SAL 500 +Tio vs. FF/UMEC/VI 200	0.772 (0.282, 2.088)

Model	p	LOR (95% CrI)
*HD‐ICS/LABA vs. MD‐ICS/LABA*
Direct	0.003	‐0.027 (‐0.646, 0.557)
Indirect		‐22.573 (‐71.914, ‐2.303)
Network		‐0.141 (‐0.797, 0.475)
MD Triple vs. MD‐ICS/LABA
Direct	0.405	‐0.652 (‐2.291, 0.280)
Indirect		0.301 (‐2.064, 2.681)
Network		‐0.350 (‐1.441, 0.355)
HD Triple vs. MD‐ICS/LABA
Direct	0.840	‐0.822 (‐2.143, 0.263)
Indirect		‐0.683 (‐2.060, 0.526)
Network		‐0.798 (‐1.688, ‐0.065)
MD Triple vs. HD‐ICS/LABA
Direct	0.117	0.210 (‐0.907, 1.247)
Indirect		‐1.158 (‐2.969, 0.269)
Network		‐0.204 (‐1.274, 0.487)
HD Triple vs. HD‐ICS/LABA
Direct	0.402	‐0.554 (‐1.457, 0.322)
Indirect		‐1.509 (‐4.439, 0.756)
Network		‐0.660 (‐1.410, ‐0.021)
HD Triple vs. MD Triple
Direct	0.002	‐0.588 (‐1.396, 0.300)
Indirect		23.163 (1.997, 74.056)
Network		‐0.446 (‐1.195, 0.521)

Database/search platform/date of last search	Search strategy	Results
Airways Register (via Cochrane Register of Studies) Date of most recent search: 1 December 2020	#1 MESH DESCRIPTOR Asthma EXPLODE ALL AND INSEGMENT #2 asthma:ti,ab AND INSEGMENT #3 #1 OR #2 #4 MESH DESCRIPTOR Formoterol Fumarate AND INSEGMENT #5 MESH DESCRIPTOR Salmeterol Xinafoate AND INSEGMENT #6 formoterol:ti,ab AND INSEGMENT #7 salmeterol:ti,ab AND INSEGMENT #8 indacaterol:ti,ab AND INSEGMENT #9 vilanterol:ti,ab AND INSEGMENT #10 #4 OR #5 OR #6 OR #7 OR #8 OR #9 #11 MESH DESCRIPTOR Budesonide AND INSEGMENT #12 MESH DESCRIPTOR Fluticasone AND INSEGMENT #13 MESH DESCRIPTOR Beclomethasone AND INSEGMENT #14 budesonide:ti,ab AND INSEGMENT #15 fluticasone:ti,ab AND INSEGMENT #16 mometasone:ti,ab AND INSEGMENT #17 beclomethasone:ti,ab AND INSEGMENT #18 ciclesonide:ti,ab AND INSEGMENT #19 (inhal NEAR3 (steroid* or corticosteroid* or glucocorticoid*)):ti,ab AND INSEGMENT #20 #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 #21 MESH DESCRIPTOR Budesonide, Formoterol Fumarate Drug Combination AND INSEGMENT #22 MESH DESCRIPTOR Mometasone Furoate, Formoterol Fumarate Drug Combination AND INSEGMENT #23 MESH DESCRIPTOR Fluticasone‐Salmeterol Drug Combination AND INSEGMENT #24 #21 OR #22 OR #23 #25 #3 AND #10 AND #20 #26 #3 AND #24 #27 #25 OR #26 #28 (2008 or 2009 or 2010 or 2011 or 2012 or 2013 or 2014 or 2015 or 2016 or 2017 or 2018 or 2019 or 2020):yr AND INSEGMENT #29 #27 AND #28 #30 INREGISTER #31 #29 AND #30	Dec 2020=915
CENTRAL (via Cochrane Register of Studies) Date of most recent search: 1 December 2020	#1 MESH DESCRIPTOR Asthma EXPLODE ALL AND CENTRAL:TARGET #2 asthma:ti,ab AND CENTRAL:TARGET #3 #1 OR #2 AND CENTRAL:TARGET #4 MESH DESCRIPTOR Formoterol Fumarate AND CENTRAL:TARGET #5 MESH DESCRIPTOR Salmeterol Xinafoate AND CENTRAL:TARGET #6 formoterol:ti,ab AND CENTRAL:TARGET #7 salmeterol:ti,ab AND CENTRAL:TARGET #8 indacaterol:ti,ab AND CENTRAL:TARGET #9 vilanterol:ti,ab AND CENTRAL:TARGET #10 #4 OR #5 OR #6 OR #7 OR #8 OR #9 AND CENTRAL:TARGET #11 MESH DESCRIPTOR Budesonide AND CENTRAL:TARGET #12 MESH DESCRIPTOR Fluticasone AND CENTRAL:TARGET #13 MESH DESCRIPTOR Beclomethasone AND CENTRAL:TARGET #14 budesonide:ti,ab AND CENTRAL:TARGET #15 fluticasone:ti,ab AND CENTRAL:TARGET #16 mometasone:ti,ab AND CENTRAL:TARGET #17 beclomethasone:ti,ab AND CENTRAL:TARGET #18 ciclesonide:ti,ab AND CENTRAL:TARGET #19 (inhal NEAR3 (steroid* or corticosteroid* or glucocorticoid*)):ti,ab AND CENTRAL:TARGET #20 #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 AND CENTRAL:TARGET #21 MESH DESCRIPTOR Budesonide, Formoterol Fumarate Drug Combination AND CENTRAL:TARGET #22 MESH DESCRIPTOR Mometasone Furoate, Formoterol Fumarate Drug Combination AND CENTRAL:TARGET #23 MESH DESCRIPTOR Fluticasone‐Salmeterol Drug Combination AND CENTRAL:TARGET #24 #21 OR #22 OR #23 AND CENTRAL:TARGET #25 #3 AND #10 AND #20 AND CENTRAL:TARGET #26 #3 AND #24 AND CENTRAL:TARGET #27 #25 OR #26 AND CENTRAL:TARGET #28 (2008 or 2009 or 2010 or 2011 or 2012 or 2013 or 2014 or 2015 or 2016 or 2017 or 2018 or 2019 or 2020):yr AND CENTRAL:TARGET #29 #27 AND #28 AND CENTRAL:TARGET	Dec 2020=1665
MEDLINE (Ovid) ALL Date of most recent search: 1 December 2020	1 exp Asthma/  2 asthma$.tw.  3 1 or 2  4 Formoterol Fumarate/  5 Salmeterol Xinafoate/  6 formoterol.tw.  7 salmeterol.tw.  8 indacaterol.mp.  9 vilanterol.mp.  10 or/4‐9  11 Budesonide/  12 Fluticasone/  13 Mometasone Furoate/  14 Beclomethasone/  15 budesonide.tw.  16 fluticasone.tw.  17 mometasone.tw.  18 beclomethasone.tw.  19 ciclesonide.mp.  20 (inhal$ adj3 (steroid$ or corticosteroid$ or glucocorticoid$)).tw.  21 or/11‐20  22 Budesonide, Formoterol Fumarate Drug Combination/  23 Mometasone Furoate, Formoterol Fumarate Drug Combination/  24 Fluticasone‐Salmeterol Drug Combination/  25 or/22‐24  26 3 and 10 and 21  27 3 and 25  28 26 or 27  29 (controlled clinical trial or randomized controlled trial).pt.  30 (randomized or randomised).ab,ti.  31 placebo.ab,ti.  32 dt.fs.  33 randomly.ab,ti.  34 trial.ab,ti.  35 groups.ab,ti.  36 or/29‐35  37 Animals/  38 Humans/  39 37 not (37 and 38)  40 36 not 39  41 28 and 40  42 limit 41 to yr="2008 ‐Current"	Dec 2020=993
Embase (Ovid) Date of most recent search: 1 December 2020	1 exp asthma/  2 asthma$.tw.  3 1 or 2  4 formoterol fumarate/  5 salmeterol xinafoate/  6 formoterol.tw.  7 salmeterol.tw.  8 indacaterol.mp.  9 vilanterol.mp.  10 or/4‐9  11 budesonide/  12 fluticasone/  13 mometasone furoate/  14 beclometasone/  15 budesonide.tw.  16 fluticasone.tw.  17 mometasone.tw.  18 beclomethasone.tw.  19 ciclesonide.mp.  20 (inhal$ adj3 (steroid$ or corticosteroid$ or glucocorticoid$)).tw.  21 or/11‐20  22 budesonide plus formoterol/  23 formoterol fumarate plus mometasone furoate/  24 exp fluticasone propionate plus salmeterol/  25 or/22‐24  26 3 and 10 and 21  27 3 and 25  28 26 or 27  29 Randomized Controlled Trial/  30 randomization/  31 controlled clinical trial/  32 Double Blind Procedure/  33 Single Blind Procedure/  34 Crossover Procedure/  35 (clinica$ adj3 trial$).tw.  36 ((singl$ or doubl$ or trebl$ or tripl$) adj3 (mask$ or blind$ or method$)).tw.  37 exp Placebo/  38 placebo$.ti,ab.  39 random$.ti,ab.  40 ((control$ or prospectiv$) adj3 (trial$ or method$ or stud$)).tw.  41 (crossover$ or cross‐over$).ti,ab.  42 or/29‐41  43 exp animals/ or exp invertebrate/ or animal experiment/ or animal model/ or animal tissue/ or animal cell/ or nonhuman/  44 human/ or normal human/ or human cell/  45 43 and 44  46 43 not 45  47 42 not 46  48 28 and 47  49 limit 48 to yr="2008 ‐Current"	Dec 2020=1758
Global Health (Ovid) Date of most recent search: 1 December 2020	1 exp asthma/  2 asthma$.tw.  3 1 or 2  4 formoterol.tw.  5 salmeterol.tw.  6 indacaterol.mp.  7 vilanterol.mp.  8 4 or 5 or 6 or 7  9 exp corticoids/  10 budesonide.tw.  11 fluticasone.tw.  12 mometasone.tw.  13 beclomethasone.tw.  14 ciclesonide.mp.  15 (inhal$ adj3 (steroid$ or corticosteroid$ or glucocorticoid$)).tw.  16 or/9‐15  17 3 and 8 and 16  18 randomized controlled trials/  19 (randomized or randomised).ab,ti.  20 placebo.ab,ti.  21 randomly.ab,ti.  22 trial.ab,ti.  23 or/18‐22  24 17 and 23  25 limit 24 to yr="2008 ‐Current"	Dec 2020=32
ClinicalTrials.gov Date of most recent search: 1 December 2020	Study type: Interventional Condition: asthma Intervention: (formoterol OR salmeterol OR indacaterol OR vilanterol) AND (budesonide OR fluticasone OR mometasone OR beclomethasone OR ciclesonide)	Dec 2020=270