Objective To examine the effect of beta2-agonists on aerobic performance in healthy, non-asthmatic study participants.
Design Systematic review and meta-analysis.
Eligibility criteria We searched four databases (PubMed, Embase, SPORTDiscus and Web of Science) for randomised controlled trials published until December 2019. Studies examining the effect of beta2-agonists on maximal physical performance lasting longer than 1 min were included in the meta-analysis. Data are presented as standardised difference in mean (SDM) with 95% CI.
Results The present meta-analysis includes 47 studies. The studies comprise 607 participants in cross-over trials, including 99 participants in three-way cross-over trials and 27 participants in a four-way cross-over trial. Seventy-three participants were included in parallel trials. Beta2-agonists did not affect aerobic performance compared with placebo (SDM 0.051, 95% CI −0.020 to 0.122). The SDM for the included studies was not heterogeneous (I2=0%, p=0.893), and the effect was not related to type of beta2-agonist, dose, administration route, duration of treatment or performance level of participants. Beta2-agonists had no effect on time trial performance, time to exhaustion or maximal oxygen consumption (p<0.218).
Conclusion/implication The present study shows that beta2-agonists do not affect aerobic performance in non-asthmatic subjects regardless of type, dose, administration route, duration of treatment or performance level of participants. The results of the present study should be of interest to WADA and to anyone who is interested in equal opportunities in competitive sports.
Systematic review registration PROSPERO CRD42018109223.
- elite performance
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Asthma is the most common chronic disease in elite athletes,1 and endurance athletes regularly performing heavily increased ventilation are at increased risk of developing asthma.2 Asthma is usually associated with airway hyper-responsiveness (AHR), and the recommended therapy for asthma is inhaled glucocorticoids with inhaled beta2-agonists pre-exercise and as a reliever of symptoms.3 However, ever since inhaled beta2-agonists became available just before the Olympic Games in 1972, antidoping authorities have regulated the use of beta2-agonists in athletes due to possible performance-enhancing effects.1 WADA annually updates the prohibited list, a list of substances and methods prohibited in elite sports. The prohibited list, effective from 1 January 2020, prohibits all use of beta2-agonists except inhaled salbutamol (maximum of 1600 μg over 24 hours in divided doses, not to exceed 800 μg over 12 hours starting from any dose), inhaled formoterol (maximum delivered dose of 54 μg over 24 hours) and inhaled salmeterol (maximum of 200 μg over 24 hours).4
Athletes with a medical history of asthma have consistently outperformed athletes without this condition during the Olympic Games.2 There has been some suspicion that non-asthmatic athletes use beta2-agonists with intention to improve their performance.5 Thus, the possible performance-enhancing effect of beta2-agonists has been examined in multiple studies. In 2007, Kindermann5 reviewed the effect of beta2-agonists and concluded that inhaled beta2-agonists do not enhance endurance performance, while oral beta2-agonists enhance endurance performance. The year after, the IOC consensus statement claimed that inhaled beta2-agonists do not enhance endurance performance,6 and a joint Task Force of European Respiratory Society and European Academy of Allergy and Clinical Immunology concluded that there was no evidence to suggest that asthma drugs can improve physical performance in healthy athletes.7 In 2011, Pluim et al 8 published the first systematic review and meta-analysis on the effect of beta2-agonists on physical performance in healthy athletes. They did not identify any effect of inhaled beta2-agonists on endurance, strength or sprint performance, but some weak evidence indicating a performance-enhancing effect of systemic beta2-agonists on anaerobic performance. Since August 2009, multiple studies have investigated the effect of beta2-agonists on aerobic performance, and controversy with regard to the use of beta2-agonists in sports continuously exists, which has been highlighted in recent beta2-agonist antidoping investigations involving world-class athletes.9 10 In a recent meta-analysis,11 we assessed the effect of beta2-agonists on anaerobic performance in non-asthmatic subjects. Therefore, the aim of this systematic review and meta-analysis was to assess the effect of beta2-agonists on aerobic performance in healthy, non-asthmatic subjects.
Search strategy and selection criteria
The study protocol for this systematic review and meta-analysis was registered at PROSPERO (International Prospective Register of Systematic Reviews) on 18 September 2018, with registration number CRD42018109223, and complied with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2009 guidelines.12
We systematically searched for published randomised controlled trials (RCT) that examined the effect of beta2-agonists on physical performance in healthy humans on 29 October 2018. Peer-reviewed articles published in English were identified from four electronic databases: PubMed (all fields), Embase (all fields), SPORTDiscus (text) and Web of Science (topic). The search strategy consisted of four blocks of terms: (healthy OR non-asthmatic OR athletes) AND (salbutamol OR formoterol OR salmeterol OR terbutaline OR albuterol) AND (exhaustion OR power OR endurance OR strength OR aerobe OR anaerobe OR exercise OR performance) AND (rct OR randomized controlled trial* OR randomized control trial* OR controlled trial*). The search identified 398 records (PubMed 100, Embase 105, Web of Science 36 and SPORTDiscus 157). After elimination of duplicates, 290 records remained. On 18 December 2019 we performed an updated search in the four databases, adding the term beta2-agonist and all beta2-agonists listed in the WADA prohibited list4 but were not included in the original search (beta2-agonist OR Fenoterol OR Higenamine OR Indacaterol OR Olodaterol OR Procaterol OR Reproterol OR Tretoquinol OR Tulobuterol OR Vilanterol). The updated search identified 57 records (PubMed 11, Embase 22, Web of Science 5 and SPORTDiscus 19). After elimination of duplicates, 44 records remained (figure 1). The searches were performed by the first author and a librarian.
Inclusion criteria and selection process
Two authors (AR and LBA) independently assessed the studies for eligibility, with subsequent consensus by discussion. We included RCTs involving healthy, non-asthmatic subjects examining the effect of beta2-agonists on maximal physical performance.
Studies investigating the effect of salbutamol/albuterol, salmeterol, formoterol and terbutaline alone or in various combinations, administered by inhalation, orally or by infusion, were included. There were no restrictions related to dose or duration of treatment.
We excluded studies examining physical performance with a duration of 60 s or less and non-performance variables such as neuromuscular function, oxygen kinetics and ventilation.
From the first search 45 studies were selected for full-text eligibility assessment after screening of titles and abstracts, while 22 other studies were included based on previous knowledge of the studies or screening of the reference lists of the included studies. From the updated search four studies were selected for full-text eligibility assessment after screening of titles and abstracts. In total 71 studies met the primary inclusion criteria. As the present study includes only performance outcomes lasting longer than 1 min, only 21 studies presenting data from performance outcomes with a duration of 1 min or less were excluded. Two studies presented results that were published in two other studies and were therefore excluded, and one study reported no performance outcome. Thus, 47 studies were included in the present meta-analysis (figure 1).
Study quality assessment
The included studies were assessed using the Cochrane Collaboration Risk of Bias Tool to evaluate seven bias domains.13 The domains were scored as low risk of bias, high risk of bias or unknown risk of bias according to the tool criteria. For the domain ‘blinding of participants and personnel’, the studies were scored as high risk of bias if the subject experienced side effects of the beta2-agonists even if the blinding procedure in the study was performed according to the criteria for low risk of bias. The domains ‘incomplete data’ and ‘selective reporting’ were set to ‘low risk of bias’ due to the nature of the studies. The seventh domain (other bias) was defined as ‘participants screened for AHR’. For a study to be classified as ‘low risk for bias’ on the seventh domain, an objective measure of AHR was required. Studies which screened participants using lung function measurement at rest, stethoscopy or a questionnaire on medical history or bronchial complaints were considered as ‘high risk of bias’. An eighth domain was added to the risk of bias tool reporting the days between the treatments in cross-over studies. Carryover effect and period effect were also extracted from the study as part of bias assessment. Two authors (AR and JS) independently assessed the included studies that were not assessed in the meta-analysis by Pluim et al.8 Any discrepancies in the assessments were resolved by discussion.
AR and TS conducted data extraction of study results separately and settled discrepancy by mutual agreement. The main outcome was aerobic performance, defined as maximal physical performance lasting more than 1 min. Other outcomes were maximal/peak oxygen consumption (VO2max/peak), measured by analysis of expiratory gas; and physical performance, measured as time to exhaustion, distance covered in a preset time, time to cover a distance/amount of work or contractions to task failure. If a study reported both VO2max/peak and physical performance, physical performance was included in the main analysis. If only VO2max/peak was reported, VO2max/peak was included as outcome in the main analysis. Performance tests were categorised into time to exhaustion/open-ended tests when the subjects performed running, walking, cycling or quadriceps contractions for as long as they could, and time trials/closed-ended tests when the subjects performed a given amount of work as fast as possible. The interventions were categorised in four different ways: (1) type of beta2-agonist: short-acting (salbutamol and terbutaline) and long-acting (formoterol and salmeterol); (2) administration route: inhaled, oral and infusion (infusion was only used in one comparison and was not included as a category in the meta-regression); (3) duration of treatment: acute treatment and multiple weeks of treatment; and (4) dose: approved and prohibited by WADA.4 The four interventions were treated as categorical variables in the meta-regression analysis. The included studies were classified as high risk of bias if they scored ‘high risk of bias’ in one domain or more, and low risk of bias if all domains in the risk of bias assessment tool were scored as ‘low risk of bias’ or ‘unclear risk of bias’. If the subject in a study performed endurance training for more than 10 hours per week or had VO2max >65 mL/kg/min (women >60 mL/kg/min), the subjects were classified as high-performance endurance athletes. Correlations between performance with active treatment and placebo were seldom reported in the included studies; thus, a correlation coefficient of 0.5 was imputed for all comparisons.
Extracted data from individual studies were collated and prepared for meta-analysis (computing SD when SEM and 95% CI were reported) in Excel (Microsoft) prior to transfer into Comprehensive Meta-Analysis (CMA) V.3 (Biostat, Englewood, New Jersey, USA). Further analyses were performed in CMA. The meta-analyses were performed with random effects models, and effect estimates are presented as standardised difference in mean (SDM) with 95% CI. Heterogeneity is presented as I2 and p values. Whether or not the effect size was related to type of beta2-agonist, administration route, duration of treatment, dose, publication bias or level of fitness was analysed by meta-regression (test of model). In addition, meta-regression was used to perform the goodness of fit to assess the presence of unexplained variance in the model. The proportion of total between-study variance explained by the covariate is expressed as R2 analogue. Potential publication bias was assessed by funnel plot and Begg and Mazumdar rank correlation test. Adequacy of sample size in each included study was assessed by calculation of the sample size required for the effect found in the respective study to obtain an alpha of 0.05 and a beta of 0.2.14 Skewness of outcomes was assessed as baseline mean/SD. Variables with a mean to SD ratio >2 were considered skewed.15 Significance level was set to p<0.05.
The present meta-analysis consists of 47 RCTs, including 43 studies with a cross-over design,16–58 of which 8 studies compared two interventions33 35 39 44 46 53 55 58 and 1 study compared three different interventions40 with placebo, and 4 studies with parallel design59–62 (one with both acute and multiple-week intervention62). The meta-analysis of aerobic performance includes 60 different randomised and placebo-controlled comparisons with beta2-agonists, comprising 607 participants in cross-over trials that include 99 participants in three-way cross-over trials comparing two different treatments with placebo and 27 participants in a four-way cross-over trial comparing three different beta2-agonist treatments with placebo. Seventy-three participants were included in parallel trials. The included studies are presented in table 1.
Risk of bias
Twenty-five (53%) studies had high risk of bias in one domain or more, and the washout period varied from overnight to 4 weeks between the studies (online supplementary appendix table 1). The effect of beta2-agonists was not related to risk of bias (table 2). One study52 reported period effect (p=0.12) and carryover effect (p=0.51) in addition to treatment effect (p=0.54). Examination of potential publication bias by assessing the funnel plot indicated no publication bias (online supplementary appendix figure 1). The Begg and Mazumdar rank correlation test found no publication bias, with a one-tailed p value of 0.466.
Effect of beta2-agonists
Beta2-agonists did not affect aerobic performance as compared with placebo (SDM 0.051, 95% CI −0.020 to 0.122). The SDM for the included studies was not heterogeneous (I2=0%, p=0.893) (table 3). Neither type of beta2-agonist, administration route, duration of treatment nor dose of beta2-agonist did influence the SDM (p>0.340) (table 2). In a stratified analysis beta2-agonists prohibited by WADA (SDM 0.032, 95% CI −0.082 to 0.146) and beta2-agonists approved by WADA (SDM 0.063, 95% CI −0.02 to 0.153) had no effect on aerobic performance.
The effect of beta2-agonists on physical performance was assessed in 53 comparisons. Beta2-agonists did not improve aerobic physical performance (SDM 0.047, 95% CI −0.028 to 0.121). The SDM for the included studies was not heterogeneous (I2=0%, p=0.778) (table 3, figure 2). Neither type of beta2-agonist, administration route, duration of treatment nor dose of beta2-agonist did influence the SDM (p>0.325) (table 2).
The SDM from 28 comparisons showed that beta2-agonists did not affect VO2max (p=0.809) (table 3, figure 3), and type of beta2-agonist, administration route, duration of treatment or dose of beta2-agonist did not influence SDM (p>0.249) (table 2).
The effect on time trial performance was assessed in 17 comparisons and the effect on performance to exhaustion was assessed in 36 comparisons. Beta2-agonists did not improve time to exhaustion (p=0.411) or time trial performance (p=0.345) (table 3).
Sample size and skewness
Two of the 47 included studies included adequate numbers of participants to obtain an alpha <0.05 and a beta <0.2 (online supplementary table 1). Baseline performance values were skewed in four comparisons, and sensitivity analysis excluding these four comparisons found no effect of beta2-agonists on aerobic performance (SDM 0.046, p=0.213).
Eleven studies included high-performance endurance athletes (table 3), and there was no difference in response to beta2-agonists in the high-performance endurance athletes compared with the less fit subjects included in the other studies (table 2; p=0.808).
In a sensitivity analysis excluding the 16 comparisons with 10 or less pairwise comparisons between beta2-agonists and placebo, we did not find any effect of beta2-agonists on aerobic performance (SDM 0.059, p=0.171).
This meta-analysis of RCTs that examined the effect of beta2-agonists on aerobic performance provides the most comprehensive quantitative summary of the evidence to date, including 47 RCTs with 60 placebo-controlled comparisons comprising 680 participants. Twenty-five studies had high risk of bias due to side effects, single blinding or inadequate screening for AHR. Our study extends previous reviews by including 21 studies not previously meta-analysed and with an indepth analysis of aerobic performance. The results from our analysis demonstrated that beta2-agonists had no effect on aerobic performance. The result was consistent and not heterogeneous.
To our knowledge, no other studies have pooled data and meta-analysed the effects of beta2-agonists on aerobic performance to this extent. Pluim et al 8 presented a meta-analysis stratified by administration route (oral or inhaled) and analysed test-specific outcomes separately and did not find any effect of inhaled or oral beta2-agonists on any aerobic performance outcome. In our study, we included comparisons with inhaled and oral beta2-agonists in the same analysis, as we hypothesised that inhalation and oral ingestion may provide the same physiological stimuli, which depends on the dose and systemic bioavailability, because the two administration routes may induce similar serum concentrations of beta2-agonists.63 This assumption was supported by the findings in the present study, as route of administration was not related to the effect of beta2-agonists on aerobic performance. We also investigated whether beta2-agonists prohibited by WADA had different effects from beta2-agonists (type and dose) approved by WADA and we found no difference. In addition, we meta-analysed different types of aerobic performance and stratified by VO2max, performance, time trials (closed-ended tests) performance, and performance until exhaustion (open-ended tests). We combined 59 comparisons as compared with 2–18 comparisons (depending on the number of studies measuring aerobic performance in the same way) in the study by Pluim et al.8 Thus, our results strengthen the findings from Pluim et al.8 Type of beta2-agonist, administration route, duration of treatment, dose, fitness level of participants or study quality did not affect the result.
Analysis of performance categories
Maximal oxygen consumption was unaffected by the use of beta2-agonists. The effect size from the 27 comparisons included was not heterogeneous, and type of beta2-agonist, administration route, duration of treatment or dose did not affect the result. The finding builds on the evidence from Pluim et al,8 who meta-analysed 18 studies with inhaled beta2-agonists.
Beta2-agonists did not improve aerobic physical performance measured by closed-ended or open-ended tests. Closed-ended tests are usually recommended over open-ended tests due to better reliability64 and possibly a better chance of detecting minor differences in performance. However, in the present study, neither type of protocol indicated an effect of beta2-agonists on aerobic performance.
The funnel plot and the Begg and Mazumdar rank correlation test were negative for publication bias, indicating that the results do not influence whether the studies are published or not.
Tachycardia and tremor are characteristic adverse side effects of beta2-agonists.65 These side effects may break the blinding if the participants are aware of whether they receive beta2-agonists or placebo, and possibly motivate them to perform differently when receiving beta2-agonists. Fifteen studies were classified as high risk of bias due to lack of blinding (single-blinded design or reported side effects of the beta2-agonists), and 10 additional studies did not screen the participants sufficiently for AHR. However, high risk of bias did not influence the SDM in any analysis.
Our study included 16 comparisons between placebo and beta2-agonists that comprised less than 10 pairs (less than 20 subjects in parallel studies and 10 subjects in cross-over studies), and only two of the included studies had a sample size providing an alpha <0.05 and a beta <0.2 for the measured effect. This low sample size in the individual studies may have introduced sparse data bias in the SDM.66 67 However, when we performed a sensitivity analysis excluding these 16 comparisons, the effect of beta2-agonists on aerobic performance was still not statistically significant. Normal distribution of data is an important assumption in the meta-analysis of continuous data and five comparisons had skewed baseline data; however, execution of these comparisons from the meta-analysis did not influence the result.
Strength and limitations
The present study is strengthened by the systematic search of the literature in multiple databases. It is therefore likely that all relevant studies were identified, and we included RCTs only. We consider all maximal performance tests lasting more than 1 min to be a measure of aerobic performance; thus, we used SDM as outcome in the meta-analysis.68 This resulted in a large sample size and a high statistical power. We also investigated the effect of beta2-agonists separately on VO2max and physical performance, which was further divided into open-ended and closed-ended tests to investigate any physiological (oxygen consumption) or performance test-specific effects. Further, we performed subgroup analyses on outcome categories and meta-regression to investigate the effect of the different types of intervention. For example, two recent studies69 70 reported a potential negative effect of multiple weeks of beta2-agonist treatment, while one study44 reported a positive effect of acute administration. If this was representative of the included studies, the opposite effects could even each other out in a meta-analysis but show difference in effects in a meta-regression categorised by duration of treatment.
Splitting up the interventions into subcategories can also be a limitation as the statistical power is reduced. Categories with few studies/participants (multiple weeks of treatment) usually have larger uncertainties in the effect estimates, and it is difficult to interpret if the lack of effect is due to no effect or to low statistical power to reveal the real effect. Another weakness in the present study is that all outcomes are assessed by laboratory tests, not identical to actual athletic competitions. Therefore, reliable, sensitive and valid test protocols are of importance. Closed-ended tests are recommended over open-ended tests due to better reliability, and the coefficients of variation for open-ended tests are reported to decrease with increased intensity or decreased duration.64 Thus, we performed separate meta-analyses on open-ended and closed-ended tests. However, both types of tests failed to demonstrate any effect of beta2-agonists. The fitness level of the participants included in our study varied from untrained to elite athletes, and fitness level has been suggested to confound the effect of beta2-agonists on physical performance.71 However, in the present study beta2-agonists did not affect aerobic performance differently in high-performance endurance athletes compared with less endurance trained participants. The meta-analysis assumes independence between the subjects included. In the present study, the same subjects are included twice in the analysis if they participated in a three-way cross-over study with two different interventions with beta2-agonists, or if the same subjects are assessed after acute treatment and subsequently after multiple weeks of beta2-agonist treatment or placebo. To investigate the effect of this potential bias, we performed a meta-analysis including comparisons with different people only, and the effect size was practically the same as when all relevant comparisons were included. There is also a possibility that the same subjects are included in different studies. Few studies reported correlation between trial results; thus, the correlation coefficient for pretest and post-test has been set to 0.5 for all studies. This is lower than the data made available from Dickinson et al 60 by request and similar to what Pluim et al 8 reported. RCTs are regarded as high-quality studies, but there is a possibility of bias especially related to the side effects from beta2-agonists breaking the blinding for the participants. In addition, the meta-analysis includes several single-blinded studies where the investigators knew when the subjects received beta2-agonists. This lack of blinding may allow the investigator to treat the subjects systematically different when knowing what the subjects have received. The possible difference in the way the investigator interacts with the subjects may lead to a systematic difference in performance. Many studies did not include objective tests for AHR and thus may have failed to exclude participants with AHR. However, risk of bias did not influence the effect size in the present meta-analysis, and Koch et al 51 found no difference in time trial performance between cyclists with and without AHR. In addition, a systematic review from 201472 concluded the current evidence is insufficient to prove a negative effect of AHR on physical performance.
Based on the previously mentioned limitations, the findings should be interpreted with caution, especially the results from subgroups with few studies/participants, but there is consistency in the results demonstrating that beta2-agonists do not affect aerobic performance in non-asthmatic subjects.
The present study, which summarises the best scientific evidence, shows that beta2-agonists do not affect aerobic performance in non-asthmatic subjects. Beta2-agonists had no effect on performance tests or VO2max. The results from the present study should be of interest to WADA when revising the antidoping regulations and planning antidoping sample analysis, and to anyone who is interested in equal opportunities in competitive sports.
What is already known
Asthma is the most common chronic disease in athletes.
The gold standard for asthma therapy is inhaled glucocorticoids with inhaled beta2-agonists pre-exercise and as a reliever of symptoms.
The use of beta2-agonists in sports is regulated by WADA due to possible performance-enhancing effects.
What are the new findings
Beta2-agonists do not affect physical aerobic performance or maximal oxygen consumption in healthy subjects.
Route of administration, type of beta2-agonist, duration of treatment or dose were not related to the effect of beta2-agonists in healthy subjects.
Study bias or performance level of subjects had no impact on the effect of beta2-agonists on aerobic performance in healthy subjects.
Contributors All authors reviewed the report. AR generated the hypotheses, did the literature search, analysed the data and wrote the first draft of the manuscript. AR, TS, JS and LBA revised the manuscript critically for important intellectual content. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. AR and TS extracted the data. JS and AR assessed bias. AR and LBA evaluated the studies for inclusion.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.