Does the use of inhaled corticosteroids in asthma benefit lung function in the long-term? A systematic review and meta-analysis

Daniel J Tan; Din S Bui; Xin Dai; Caroline J Lodge; Adrian J Lowe; Paul S Thomas; Deborah Jarvis; Michael J Abramson; E Haydn Walters; Jennifer L Perret; Shyamali C Dharmage

doi:10.1183/16000617.0185-2020

. 2021 Jan 20;30(159):200185. doi: 10.1183/16000617.0185-2020

Does the use of inhaled corticosteroids in asthma benefit lung function in the long-term? A systematic review and meta-analysis

Daniel J Tan ¹, Din S Bui ¹, Xin Dai ¹, Caroline J Lodge ¹, Adrian J Lowe ¹, Paul S Thomas ², Deborah Jarvis ³, Michael J Abramson ⁴, E Haydn Walters ^1,⁵, Jennifer L Perret ^1,^6,⁷, Shyamali C Dharmage ^1,^7,^✉

PMCID: PMC9488672 PMID: 33472957

Abstract

While asthma is known to be associated with an increased risk of progressive lung function impairments and fixed airflow obstruction, there is ongoing debate on whether inhaled corticosteroids (ICS) modify these long-term risks.

Searches were performed of the PubMed, Embase and CENTRAL databases up to 22 July 2019 for studies with follow-up ≥1 year that investigated the effects of maintenance ICS on changes in lung function in asthma.

Inclusion criteria were met by 13 randomised controlled trials (RCTs) (n=11 678) and 11 observational studies (n=3720). Median (interquartile range) follow-up was 1.0 (1–4) and 8.4 (3–28) years, respectively. In the RCTs, predominantly in individuals with mild asthma, ICS use was associated with improved pre-bronchodilator (BD) forced expiratory volume in 1 s (FEV₁) across all age groups (2.22% predicted (95% CI 1.32–3.12), n=8332), with similar estimates of strength in association for children and adults. Improvements in post-BD FEV₁ were observed in adults (1.54% (0.87–2.21), n=3970), but not in children (0.20% (−0.49–0.90), n=3924) (subgroup difference, p=0.006). Estimates were similar between smokers and nonsmokers. There were no RCT data on incidence of fixed airflow obstruction. In the observational studies, ICS use was associated with improved pre-BD FEV₁ in children and adults. There were limited observational data for post-BD outcomes.

In patients with mild asthma, maintenance ICS are associated with modest, age-dependent improvements in long-term lung function, representing an added benefit to the broader clinical actions of ICS in asthma. There is currently insufficient evidence to determine whether treatment reduces incidence of fixed airflow obstruction in later life.

Short abstract

In patients with mild asthma, maintenance inhaled corticosteroids are associated with modest age-dependent improvements in long-term lung function https://bit.ly/32K2ZTF

Introduction

Asthma is a common global disease, affecting over 350 million people, for which many patients are treated using inhaled corticosteroids (ICS). While the benefits of such treatment have been demonstrated in terms of asthma control, hospital admissions and mortality, [1–4], the long-term effects of ICS in asthma on changes in lung function remains unclear. This information is necessary to guide optimal asthma management, particularly for patients with early-onset persistent disease who are the greatest risk of progressive lung function impairments [5].

Recent studies have shown that several lung function trajectories lead to fixed airflow obstruction [6, 7]. These include impaired lung growth in the first two to three decades of life and accelerated lung function decline from the third decade onwards [6]. Depending on age of onset, asthma is now known to contribute to both pathways [8, 9]. Given the rising prevalence and global impact of obstructive airways diseases, strategies aimed at preserving lung function in asthma are urgently needed. There is also growing international interest in this area, including a European Respiratory Society clinical initiative, CADSET (Chronic Airway Diseases Early Stratification), which aims to explore the factors that influence lung function trajectories over the lifespan [10].

ICS are currently recommended as a first-line option in persistent asthma [11], and are associated with improved airway calibre through a reduction in airway inflammation, mucus hypersecretion and bronchial hyperresponsiveness [12]. To date, the available evidence on their long-term effects on lung function in asthma have not been adequately summarised. Updates to the Global Initiative for Asthma (GINA) recommendations in 2019 [13], which now strongly advocate for ICS-containing treatment in all patients with mild asthma, further increase the need to critically evaluate the evidence on this topic. Therefore, we aimed to systematically review all studies which have investigated the long-term effects of ICS on lung function in asthma and to undertake a meta-analysis where similar data were available.

Methods

Search strategy

We conducted systematic searches of the PubMed, EMBASE and CENTRAL databases from inception to 22 July 2019 for peer-reviewed studies. For full details refer to table e1. Reference lists of related review articles were manually screened for additional studies meeting our inclusion criteria.

Eligibility criteria

We included randomised controlled trials (RCTs) and observational studies with ≥1year follow-up in which the effects of maintenance ICS on change in lung function (growth or decline) and risk of fixed airflow obstruction was assessed. Studies involving either children or adults with current asthma were included. Studies were required to have an appropriate placebo or control comparison group. Inclusion was limited to English-language studies published in full text. Studies with concomitant use of maintenance systemic corticosteroids were excluded.

A range of definitions of asthma were accepted, including physician-diagnosed asthma, spirometrically defined asthma and survey-reported asthma. Current asthma was defined as: asthma symptoms or asthma-related healthcare utilisation within the last 12 months.

Outcomes of interest

There were two outcomes of interest. 1) Change in lung function from baseline defined as: change in forced expiratory volume in 1 s (FEV₁), FEV₁/forced vital capacity (FVC), forced expiratory flow at 25–75% of FVC (FEF_25–75%) or bronchial hyperresponsiveness (BHR). 2) Fixed airflow obstruction defined as post-bronchodilator (BD) FEV₁/FVC ratio <0.7 or less than the lower limit of normal.

Selection of studies

All studies identified in the search strategy were independently screened by two review authors (D.J. Tan and either D.S. Bui or X. Dai). Full texts of all studies considered eligible, potentially eligible or unclear were retrieved and assessed. Disagreements at this stage or further on were settled by consultation with a third author (D.S. Bui or X. Dai).

Data extraction

Using a standardised data extraction form, study characteristics and outcome data were extracted independently by two review authors (D.J. Tan and either D.S. Bui or X. Dai). Information extracted included the first author, date of publication, study design, study setting, date of study, number of participants, mean age, age range, sex, asthma definition, asthma severity, smoking history, ICS use (type, frequency, dose and duration), outcome definitions, confounders and interactions, effect estimates and 95% confidence intervals.

Quality assessment and risk of bias

Risk of bias was assessed independently by two review authors (D.J. Tan and either D.S. Bui or X. Dai) using the Cochrane Collaboration's Tool for individual RCTs, modified Newcastle–Ottowa scale (NOS) for individual observational studies [14], and GRADE (Grading of Recommendations Assessment, Development and Evaluation) guidelines for quality by outcome across a range of studies [15].

Statistical analysis

Where multiple studies reported data for a single outcome on different scales (e.g. % predicted or mL), standardised mean differences (SMDs) were calculated to enable pooling of the data. If studies presented data in both formats, data presented as % predicted were selected in preference to mL for conversion to SMD.

RCTs and observational studies were meta-analysed separately. Observational studies were required to have adjusted for asthma severity and smoking status in the statistical analysis or have accounted for these factors in their study design (e.g. inclusion criteria of only mild asthma) to be included in the meta-analysis.

Heterogeneity was assessed using the I² statistic. Fixed-effects models were used if I² was <25%. Random-effects models were used if I² was between 25% and 75%. Pooled estimates were presented but considered to be unreliable if I² was >75%. The selected model for each main outcome was then applied to any related subgroup meta-analyses. Data were pooled using the program Review Manager, version 5.3 [16].

Subgroup analyses

The following planned subgroup analyses were performed where possible: 1) age group: children (aged <18 years) versus adults (aged ≥18 years); 2) atopic status: atopic versus nonatopic; 3) blood or sputum eosinophil level: high versus low; threshold as specified by the individual studies; 4) smoking status: current versus never- or former smoker; 5) length of follow-up: RCTs: 1 year versus >1 year; observational studies: 10 years or <10 versus >10 years.

Sensitivity analyses

The follow sensitivity analyses were performed where possible: 1) risk of bias assessments: with versus without studies considered at high risk of bias; 2) meta-analysis methodology: fixed-effects models versus random-effects models.

Protocol registration

The protocol for this systematic review was registered prospectively in the PROSPERO database in September 2017 (registration number: CRD42017053543). FVC and BHR were added as outcomes of interest after protocol registration.

Results

Search results

6517 records were identified from the database searches and reference lists of related articles (figure 1). Of these, 365 records were selected for full-text screening and of these, 38 articles from 24 unique studies met the pre-specified inclusion criteria. Excluded records and their reasons for exclusion are listed in table e2. Characteristics of the included studies have been summarised in table 1 (RCTs) and table e3 (observational studies).

PRISMA flow diagram for the study selection process. ICS: inhaled corticosteroid.

TABLE 1.

Study characteristics of the included randomised controlled trials

First author [ref.]	Age group	Duration years	Asthma definition	Asthma severity	ICS intervention	Sample size	Mean age years	Female %	Pre-BD FEV₁
									L	% pred
Beasley [17]	Adults	1	Physician-diagnosed asthma; treatment with SABA only last 3 months	Mild	Budesonide 200 µg twice daily^#	225	34.9	57		90.3
Beasley [17]	Adults	1		Mild	No placebo (open-label)	223	35.8	51		89.2
Becker [18]	Children	1	Physician-diagnosed asthma; ≥6 months of symptoms, FEV₁ >75% pred	Mild	Beclomethasone 200µg twice daily^¶	119	7.6	32.8	1.39	91.3
Becker [18]	Children	1		Mild	Placebo, twice daily	121	7.7	34.7	1.42	92.0
Boulet [19]	Adults	1	Physician-diagnosed asthma; BHR and FEV₁ >70% pred	Mild	Fluticasone 100–250µg daily^#	35	27	54.2	3.55	97.2
Boulet [19]	Adults	1	Physician-diagnosed asthma; BHR and FEV₁ >70% pred	Mild	Placebo daily	34	26	69.7	3.33	98.5
den Otter [20]	Adults	2	Symptoms typical of asthma; BHR or BDR; FEV₁ decline ≥80 mL·year⁻¹	No restriction	Fluticasone 250 µg daily^#	23
den Otter [20]	Adults	2		No restriction	Placebo daily	22
Jonasson [21]	Children	2.3	Physician-diagnosed asthma; ≥1 exacerbation in last 12 months or ≥3 exacerbations ever	Mild	Budesonide 200 µg daily^#	32	10.0	46.9	2.23	102.1
Jonasson [21]	Children	2.3		Mild	Placebo daily	34	9.4	35.3	2.04	104.6
Juniper [22]	Adults	1	Physician-diagnosed asthma; BHR and FEV₁ >70% pred	No restriction	Budesonide 200 µg twice daily^#	16	42.4	62.5		89.9
Juniper [22]	Adults	1	Physician-diagnosed asthma; BHR and FEV₁ >70% pred	No restriction	Placebo twice daily	16	35.1	56.3		92.1
Merkus [23]	Children	2–3	Physician-diagnosed asthma; BHR and FEV₁ 55–90% pred or FEV₁/FVC 50–75% pred	Moderate-severe	Budesonide 100 µg three times daily^¶	34	11.4	62.9		72.2
Merkus [23]	Children	2–3		Moderate-severe	Placebo three times daily	20	10.9	76.7		73.5
O'Byrne [24]	Mostly adults	1	Physician-diagnosed asthma; BHR	Mild	Budesonide 200 µg twice daily^#	1282	39.0	62.2		84.2
O'Byrne [24]	Mostly adults	1	Physician-diagnosed asthma; BHR	Mild	Placebo, twice daily	1277	40.0	60.4		84.1
Osterman [25]	Adults	1	Physician-diagnosed asthma diagnosed within the last year; BHR	No restriction	Budesonide 200 µg twice daily^#	38	33.0	57.9	3.31	93.1
Osterman [25]	Adults	1		No restriction	Placebo twice daily	37	35.0	54.1	3.23	88.7
Pauwels [3]	Adults and children	3	Physician-diagnosed asthma; recent onset ≤2 years; variable airflow limitation	Mild	Budesonide 200–400 µg daily^#	3597	24.0	54.2		86.3
Pauwels [3]	Adults and children	3		Mild	Placebo daily	3568	24.0	54.0		86.6
Simons [26]	Children	1	Physician-diagnosed asthma; BHR and BDR and FEV₁>70% pred	Mild	Beclomethasone 200 µg twice daily^¶	81	9.6	41.0	1.89	92.0
Simons [26]	Children	1	Physician-diagnosed asthma; BHR and BDR and FEV₁>70% pred	Mild	Placebo twice daily	80	9.5	45.0	1.88	96.0
Tonascia [27]	Children	4.3	Physician-diagnosed asthma; BHR	Mild-moderate	Budesonide 200 µg twice daily^¶	311	9.0	41.8		93.6
Tonascia [27]	Children	4.3	Physician-diagnosed asthma; BHR	Mild-moderate	Placebo, twice daily	418	9.0	44.0		94.2
Ward [12]	Adults	1	Physician-diagnosed asthma; positive skin prick testing to ≥3 common aeroallergens	Mild-moderate	Fluticasone 750 µg twice daily⁺	17				96.0
Ward [12]	Adults	1		Mild-moderate	Placebo twice daily	18				94.0

Open in a new tab

ICS: inhaled corticosteroid; BD: bronchodilator; FEV₁: forced expiratory volume in 1 s; SABA: short-acting β₂-agonist; BHR: bronchial hyperresponsiveness; BDR: bronchodilator reversibility; FVC: forced vital capacity. ^#: low-dose ICS; ^¶: medium-dose ICS; ⁺: high-dose ICS.

RCTs

Included studies

13 RCTs with a total of 11 678 participants were included [12, 17–27]. Five RCTs were performed in children (n=1250) [18, 21, 23, 26, 27], seven in adults (n=3263) [12, 17, 19, 20, 22, 24, 25] and one in both children and adults (n=7165) [3]. Three studies contributed 90% of the participants (START: n=7165; SYGMA1: n=2559; CAMP: n=729) [3, 24, 27]. Asthma was physician-diagnosed in all studies. BHR or bronchodilator reversibility was an additional inclusion criterion in nine studies (n=10 889) [3, 19, 20, 22–27]. Seven studies were in individuals with mild asthma (n=10 708) [3, 17–19, 21, 24, 26], two were in individuals with mild-to-moderate asthma (n=764) [12, 27], one was in individuals with moderate-to-severe asthma (n=54) [23], and three did not limit inclusion by asthma severity (n=152) [20, 22, 25]. Eight studies used budesonide (n=11 128) [3, 17, 21–25, 27], three used fluticasone (n=149) [12, 19, 20] and two used beclomethasone (n=401) [18, 26]. Based on GINA criteria [11], ICS were categorised as low dose in eight studies (n=10 459) [3, 17, 19–22, 24, 25], medium dose in four studies (n=1184) [18, 23, 26, 27] and high dose in one study (n=35) [12]. Median (interquartile range) follow-up was 1.0 (1–4) years.

Risk of bias assessments

Methodological quality in the RCTs was good overall (figure 2). All RCTs were assessed to be at low risk of selection, performance and detection bias. Five studies were assessed to be at high risk of attrition bias, related to high dropout rates [19–21, 23, 25]. Reporting bias was unclear in eight mainly older studies for which prospective protocols were not available [18–23, 25, 26]. GRADE certainty of evidence across outcomes ranged from low to high as outlined in tables 2 and 3.

TABLE 2.

Meta-analysis of randomised controlled trials for pre-bronchodilator (BD) outcomes stratified by age group

Outcome or subgroup	Studies n	Participants n	Statistical method	Effect estimate	p-value	I²	GRADE
ΔPre-BD FEV₁ % pred	8	8332	MD (Random, 95% CI)	2.22 (1.32–3.12)	<0.0001	41%	Moderate^#
Adults	5	4181	MD (Random, 95% CI)	2.47 (1.64–3.29)	<0.0001	0%	High
Children	4	4151	MD (Random, 95% CI)	2.08 (0.71–3.44)	0.003	64%	Moderate^#
ΔPre-BD FEV₁ mL	4	3603	MD (Random, 95% CI)	74.14 (54.47–93.81)	<0.0001	87%	Moderate^#
Adults	2	2634	MD (Random, 95% CI)	108.82 (84.40–133.23)	<0.0001	0%	High
Children	2	969	MD (Random, 95% CI)	10.00 (−23.21–43.21)	0.56	0%	Moderate^#
ΔPre-BD FEV₁ SMD	10	11131	SMD (Random, 95% CI)	0.21 (0.12–0.30)	<0.0001	68%	Moderate^#
Adults	6	6740	SMD (Random, 95% CI)	0.28 (0.15–0.41)	<0.0001	63%	Moderate^#
Children	5	4391	SMD (Random, 95% CI)	0.16 (0.04–0.28)	0.008	60%	Moderate^#
ΔPre-BD FVC % pred	2	804	MD (Fixed, 95% CI)	0.10 (−1.23–1.43)	0.88	0%	Moderate^#
Adults	1	75	MD (Fixed, 95% CI)	2.30 (−2.29–6.89)	0.33		Low^#^{,^¶}
Children	1	729	MD (Fixed, 95% CI)	−0.10 (−1.49–1.29)	0.89		Low^#^{,^¶}
ΔPre-BD FVC mL	2	804	MD (Random, 95% CI)	−12.84 (−149.33–123.65)	0.85	53%	Moderate^#
Adults	1	75	MD (Random, 95% CI)	90.00 (−100.51–280.51)	0.35		Low^#^{,^¶}
Children	1	729	MD (Random, 95% CI)	−60.00 (−124.87–4.87)	0.07		Low^¶^{,^¶}
ΔPre-BD FVC SMD	2	804	SMD (Fixed, 95% CI)	0.01 (−0.13–0.15)	0.86	0%	Moderate^#
Adults	1	75	SMD (Fixed, 95% CI)	0.23 (−0.23–0.68)	0.33		Low^#^{,^¶}
Children	1	729	SMD (Fixed, 95% CI)	−0.01 (−0.16–0.14)	0.89		Low^#^{,^¶}
ΔPre-BD FEV₁/FVC % pred	0	0	MD (Fixed, 95% CI)
Adults	0	0	MD (Fixed, 95% CI)
Children	0	0	MD (Fixed, 95% CI)
ΔPre-BD FEV₁/FVC ratio	1	729	MD (Fixed, 95% CI)	1.60 (0.64–2.56)	0.001		Moderate^#
Adults	0	0	MD (Fixed, 95% CI)
Children	1	729	MD (Fixed, 95% CI)	1.60 (0.64–2.56)	0.001		Moderate^#
ΔPre-BD FEV₁/FVC SMD	1	729	SMD (Fixed, 95% CI)	0.25 (0.10–0.39)	0.001		Moderate^#
Adults	0	0	SMD (Fixed, 95% CI)
Children	1	729	SMD (Fixed, 95% CI)	0.25 (0.10–0.39)	0.001		Moderate^#

Open in a new tab

FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; MD: mean difference; SMD: standardised mean difference. ^#: GRADE score downgraded for heterogeneity or inconsistency of results between studies; ^¶: GRADE score downgraded for imprecision, 95% CI includes important benefit and potential harm.

TABLE 3.

Meta-analysis of randomised controlled trials for post-bronchodilator (BD) outcomes stratified by age group

Outcome or subgroup	Studies n	Participants n	Statistical method	Effect estimate	p-value	I²	GRADE
ΔPost-BD FEV₁ % pred	2	7894	MD (Random, 95% CI)	0.61 (−0.31–1.54)	0.19	71%	Moderate^#
Adults	1	3970	MD (Random, 95% CI)	1.54 (0.87–2.21)	<0.0001		Moderate^#
Children	2	3924	MD (Random, 95% CI)	0.20 (−0.49–0.90)	0.57	12%	High
ΔPost-BD FEV₁ mL	1	729	MD (Fixed, 95% CI)	−40.00 (−115.38–35.38)	0.30		Moderate^#
Adults	0	0	MD (Fixed, 95% CI)
Children	1	729	MD (Fixed, 95% CI)	−40.00 (−115.38–35.38)	0.30		Moderate^#
ΔPost-BD FEV₁ SMD	2	8894	SMD (Random, 95% CI)	0.06 (−0.03–0.15)	0.18	69%	Moderate^#
Adults	1	3970	SMD (Random, 95% CI)	0.14 (0.08–0.21)	<0.0001		Moderate^#
Children	2	4924	SMD (Random, 95% CI)	0.02 (−0.05–0.09)	0.57	15%	High
ΔPost-BD FVC % pred	1	729	MD (Fixed, 95% CI)	−0.20 (−1.40–1.00)	0.74		Low^#^{,^¶}
Adults	0	0	MD (Fixed, 95% CI)
Children	1	729	MD (Fixed, 95% CI)	−0.20 (−1.40–1.00)	0.74		Low^#^{,^¶}
ΔPost-BD FVC mL	1	729	MD (Fixed, 95% CI)	−60.00 (−119.99– −0.03)	0.05		Low^#^{,^¶}
Adults	0	0	MD (Fixed, 95% CI)
Children	1	729	MD (Fixed, 95% CI)	−60.00 (−119.99– −0.03)	0.05		Low^#^{,^¶}
ΔPost-BD FVC SMD	1	729	SMD (Fixed, 95% CI)	−0.02 (−0.17–0.13)	0.77		Low^#^{,^¶}
Adults	0	0	SMD (Fixed, 95% CI)
Children	1	729	SMD (Fixed, 95% CI)	−0.02 (−0.17–0.13)	0.77		Low^#^{,^¶}
ΔPost-BD FEV₁/FVC % pred	0	0	MD (Fixed, 95% CI)
Adults	0	0	MD (Fixed, 95% CI)
Children	0	0	MD. (Fixed, 95% CI)
ΔPost-BD FEV₁/FVC ratio	1	729	MD (Fixed, 95% CI)	0.70 (−0.08–1.48)	0.08		Low^#^{,^¶}
Adults	0	0	MD (Fixed, 95% CI)
Children	1	729	MD (Fixed, 95% CI)	0.70 (−0.08–1.48)	0.08		Low^#^{,^¶}
ΔPost-BD FEV₁/FVC SMD	1	729	SMD (Random, 95% CI)	0.13 (−0.01–0.28)	0.07		Low^#^{,^¶}
Adults	0	0	SMD (Random, 95% CI)
Children	1	729	SMD (Random, 95% CI)	0.13 (−0.01–0.28)	0.07		Low^#^{,^¶}

Open in a new tab

Studies excluded from the quantitative analysis

Two small RCTs were excluded from the quantitative analysis due to methodological factors or nonrepresentative study samples [20, 23]. One study compared budesonide to placebo in 47 children with moderate-to-severe asthma, but had significant and selective dropout in the placebo arm, which resulted in early discontinuation [23]. The budesonide arm was continued to study completion at 3 years and the authors did not account for differences in treatment duration in the comparisons. The second study was published as a brief report and compared fluticasone to placebo in 45 adults with asthma and accelerated lung function decline (>80 mL·year⁻¹) [20]. The study sample was considered to be highly selected due the declining lung function inclusion criteria and not representative of the general asthma population.

Pre-BD lung function

Pre-BD FEV₁

Eight studies measured changes in pre-BD FEV₁ (% pred) (table 2, figure 3) [3, 12, 19, 21, 22, 25–27]. Improvements were observed with treatment across all age groups (2.22% (95% CI 1.32–3.12), n=8332), with similar estimates of strength in association in children (2.08% (95% CI 0.71–3.44), n=4151) and adults (2.47% (95% CI 1.64–3.29), n=4181). Findings were similar in four studies which measured pre-BD FEV₁ in mL (74 mL (95% CI 54–94), n=3603) [18, 24, 25, 27] and when SMDs were calculated and pooled (0.21 SMD (95% CI 0.12–0.30), n=11 131).

Forest plot comparison (randomised controlled trials): change in pre-bronchodilator forced expiratory volume in 1 s (% predicted) stratified by age. ICS: inhaled corticosteroid.

Pre-BD FVC

Two studies measured changes in pre-BD FVC (% pred) and found no evidence of a treatment benefit (0.10% (95% CI −1.23–1.43), n=804) [25, 27]. Findings were similar in two studies which reported pre-BD FVC in mL (−13 mL (95% CI −149–124), n=804) [25, 27] and when SMD were calculated and pooled (0.01 SMD (95% CI −0.13–0.15), n=804)

Pre-BD FEV₁/FVC

No studies measured changes in pre-BD FEV₁/FVC (% pred). One study performed in children measured change in pre-BD FEV₁/FVC (ratio multiplied by 100) and found significant improvements with treatment (1.60 (95% CI 0.64–2.56), n=729) [27].

Pre-BD FEF_25–75

One study measured changes in pre-BD FEF_25–75%, but was excluded from the quantitative analysis due to concerns around methodology [23]. The study reported benefits in pre-BD FEV_25–75% with ICS use over the treatment period.

Post-BD lung function

Post-BD FEV₁

One study in adults [3] and two in children [3, 27] measured changes in post-BD FEV₁ (% pred) (table 3, figure 4). Significant improvements were observed with treatment in adults (1.54% (95% CI 0.87–2.21), n=3970), but not in children (0.20% (95% CI −0.49–0.90), n=3924) (p-subgroup differences0.006). One study reported change in post-BD FEV₁ (mL) in children and also did not demonstrate a treatment benefit (−40 mL (95% CI −115–35, n=729) [27]. Findings were similar when SMD were calculated and pooled, with benefits found in adults (0.14 SMD (95% CI 0.08–0.21), n=3970), but not in children (0.02 SMD (95% CI −0.05–0.09), n=4924).

Forest plot comparison (randomised controlled trials): change in post-bronchodilator forced expiratory volume in 1 s (% predicted) stratified by age. ICS: inhaled corticosteroid.

Post-BD FVC

One study performed in children measured changes in post-BD FVC (% pred) and found no evidence of treatment benefit (−0.20% (95% CI −1.40–1.00), n=729) [27]. The same study also measured changes in post-BD FVC (mL) and a borderline adverse effect of treatment would have been reported if this measure was used (−60 mL (95% CI −120– −0), n=729) [27].

Post-BD FEV₁/FVC

No studies measured changes in post-BD FEV₁/FVC (% pred). One study in children measured change in post-BD FEV1/FVC (ratio) and found an association in the direction of a benefit (0.70 (95% CI −0.08–1.48), n=729), although this did not reach statistical significance [27].

Post-BD FEF_25–75%

One study measured changes in post-BD FEF_25–75%, but was excluded from the quantitative analysis due to methodological factors [23]. The study reported benefits in post-BD FEF_25–75% with ICS use over the treatment period.

Fixed airflow obstruction

There were no RCTs that reported data on incidence of fixed airflow obstruction.

BHR

Three studies measured BHR as change in doubling concentrations of methacholine provocative concentration causing a 20% fall in FEV₁ (PC₂₀) [19, 21, 22]. Treatment increased methacholine PC₂₀ compared to placebo (1.07 doubling concentrations (95% CI 0.65–1.49), n=223). Findings were similar when BHR was expressed as change in methacholine PC₂₀ mg·mL⁻¹ (0.99 (95% CI 0.32–1.66), n=161) [26], change in methacholine PC₂₀ factor increase from baseline (1.10× (95% CI 0.61–1.59), n=729) [27] and change in histamine PC₂₀ factor increase from baseline (2.59× (95% CI 1.18–4.00), n=75) [25].