NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Dryden DM, Spooner CH, Stickland MK, et al. Exercise-Induced Bronchoconstriction and Asthma. Rockville (MD): Agency for Healthcare Research and Quality (US); 2010 Jan. (Evidence Reports/Technology Assessments, No. 189.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Exercise-Induced Bronchoconstriction and Asthma

Exercise-Induced Bronchoconstriction and Asthma.

Show details

Executive Summary

Introduction

Vigorous physical exercise can be followed by transient clinical signs and symptoms similar to an asthma attack and are due to post-exercise bronchoconstriction (i.e., a narrowing of the airways). Clinical symptoms include coughing, wheezing, shortness of breath, excessive mucus production, chest tightness, chest pain, or an ‘itching or scratching sensation’ in the chest. Though it is more common in people with asthma, it also occurs in people without asthma.

For the purposes of this evidence report, we defined exercise-induced bronchoconstriction (EIB) as “the airway obstruction that occurs in association with exercise without regard to the presence of chronic asthma”1 and we defined exercise-induced asthma (EIA) as “the condition in which exercise induces symptoms of asthma in patients who have asthma.”1 Research is ongoing to determine if the pathogenesis of the phenomenon is the same in non-asthmatics and asthmatics. In this report EIB and EIA were analyzed and discussed separately when the populations and data were specifically identified by authors of the primary studies as one or the other. When it was not clear or studies included patients with and without asthma, we referred to them as EIB/EIA.

The acute bronchoconstriction associated with EIB/EIA peaks rapidly, 3 to 15 minutes after exercise stops, then remits spontaneously within 20 to 60 minutes.2 It does not cause a persistent deterioration in lung function. Following recovery, a refractory period of 40 minutes to 3 hours has been reported. During this time repeat exercise causes less bronchoconstriction.3

Two of the most common pulmonary function measures used to quantify the degree of bronchoconstriction are the forced expiratory volume in 1 second (FEV1) and the peak expiratory flow (PEF), with FEV1 considered the more reliable and valid. Both measures will decrease from baseline pre-exercise values in susceptible individuals. It is proposed that the increase in minute ventilation caused by vigorous exercise triggers the airway narrowing.3,4 Some episodes are severe enough that the person will require an inhaled bronchodilator agent to reverse the bronchoconstriction.

Approximately 20.5 million people in the United States (7 percent) have asthma.5 Between 60 and 90 percent of people with asthma experience EIA and consider exercise a major trigger of asthma symptoms.6,7 Prevalence of EIB is lower (6 to 13 percent) in populations with no history of asthma or allergy.8,9 Among elite athletes the prevalence is reported to range from 10 to 50 percent.7

The signs and symptoms of EIB/EIA often go unrecognized or are blamed on lack of conditioning. Some people will avoid exercise and some parents, teachers, and coaches of children with asthma may impose restrictions on which activities will be allowed.10 With a proper diagnosis and treatment, children and adults have successfully competed at all levels of physical activity.

Key Questions

The objective of this report was to synthesize the evidence for six key questions on diagnostic test accuracy for EIB/EIA and six key questions on therapy for EIB/EIA (five involving pharmaceutical interventions and one on a nonpharmacologic intervention).

D-1. In patients with suspected EIB/EIA, what are the diagnostic test characteristics of a self-reported history/symptoms diary for diagnosing EIB/EIA compared with a standardized exercise challenge?

D-2. In patients with suspected EIB/EIA, what are the diagnostic test characteristics of a methacholine challenge (MCH) for diagnosing EIB/EIA compared with a standardized exercise challenge?

D-3. In patients with suspected EIB/EIA, what are the diagnostic test characteristics of sport/venue specific exercise challenges for diagnosing EIB/EIA compared with a standardized exercise challenge?

D-4. In patients with suspected EIB/EIA, what are the diagnostic test characteristics of eucapnic voluntary hyperpnea (EVH) for diagnosing EIB/EIA compared with a standardized exercise challenge?

D-5. In patients with suspected EIB/EIA, what are the diagnostic test characteristics of a free running asthma screening test (FRAST) for diagnosing EIB/EIA compared with a standardized exercise challenge?

D-6. In patients with suspected EIB/EIA, what are the diagnostic test characteristics of mannitol for diagnosing EIB/EIA compared with a standardized exercise challenge?

T-1. In patients with confirmed EIB/EIA, do patients using short- or long-acting beta-agonists (SABA or LABA) therapy develop tachyphylaxis to treatment and, if so, at what frequency, compared with standard comparator treatments and/or placebo?

T-2. In patients with confirmed EIB/EIA, does pre-exercise treatment with leukotriene receptor antagonist therapy (LTRA) reduce symptoms and prevent a 10 percent or more drop in FEV1 compared with no treatment/placebo?

T-3. In patients with confirmed EIB/EIA, does pre-exercise treatment with inhaled corticosteroid therapy (ICS) reduce symptoms and prevent a 10 percent or more drop in FEV1 compared with no treatment/placebo?

T-4. In patients with confirmed EIB/EIA, does pre-exercise treatment with mast cell stabilizers (sodium cromoglycate or nedocromil; MCS) therapy reduce symptoms and prevent a 10 percent or more drop in FEV1 compared with no treatment/placebo?

T-5. In patients with confirmed EIB/EIA, does pre-exercise treatment with anticholinergic therapy (SAAC) reduce symptoms and prevent a 10 percent or more drop in FEV1 compared with no treatment/placebo?

T-6. In patients with confirmed EIB/EIA, does a refractory period (10 to 15 minute warmup and/or cooldown) reduce symptoms and prevent a 10 percent or more drop in FEV1 compared with no treatment/placebo?

Methods

For questions involving diagnostic test accuracy, we searched MEDLINE®, EMBASE, EBM Reviews - Cochrane Central Register of Controlled Trials, AMED, PsycINFO, PASCAL, CINAHL®, SPORTDiscus with Full Text, Academic Search Elite, Web of Science®, BIOSIS Previews®, PubMed, Scopus®, the Medion Database of Diagnostic Reviews - University of Maastricht, and Proquest Dissertations and Theses. For questions involving EIB/EIA therapy we searched the Cochrane Airways Register. This Register contains references to randomized controlled trials (RCTs) from the Cochrane Central Register of Controlled Trials, MEDLINE®, EMBASE, CINAHL®, AMED, PsycINFO, as well as results from handsearches of respiratory journals and meeting abstracts. For all questions, we handsearched conference proceedings from key scientific meetings and reference lists of included studies.

Two reviewers independently screened the search results (titles and abstracts) to determine if an article met broad inclusion criteria. The full-text of potentially relevant articles was retrieved and two reviewers independently assessed each study using a standard inclusion/exclusion form. Disagreements were resolved by discussion or through third party adjudication, as needed.

Two reviewers independently assessed the methodological quality of individual diagnostic test accuracy studies using the Quality Assessment of Studies of Diagnostic Accuracy (QUADAS) tool11 and the methodological quality of RCTs in the therapy reviews using the Jadad12 scale and Schulz’s criteria to assess allocation concealment.13,14

Two reviewers also independently assessed the strength of evidence for each key question. For the diagnostic test reviews, we used the GRADE system for rating the quality of evidence and strength of recommendations for diagnostic tests.15 We assessed the strength of the study designs, the quantity and quality of individual studies, the indirectness of evidence, and the consistency and precision of the results. For all studies, the outcomes were measures of test accuracy (i.e., true positive, true negative, false positive, false negative), which are surrogates for patient-important outcomes. For each key question the quality of evidence was graded as high, moderate, low or very low. For therapy reviews, the strength of evidence for the primary outcome (maximum percent fall in FEV1) was assessed using the EPC approach to grading the strength of a body of evidence.16 A grade of high, moderate, low, or insufficient was based on four domains: risk of bias, consistency, directness, and precision. Disagreements were resolved by discussion or through third party adjudication, as needed.

Data analysis

The primary outcome used to determine the presence or absence of EIB/EIA was the maximum percent fall in FEV1 post-exercise challenge calculated using the following formula:

FEV1/PEFpre-exercise-minimum FEV1/PEFpost-exerciseFEV1/PEFpost-exercise×100

The peak expiratory flow (PEF) was used if that was the only pulmonary function measure reported. In the case of missing data, some assumptions and imputations were necessary to transform reported data into the form required for analysis. Data from graphs were extracted using CorelDRAW® 9.0 (Corel Corp., Ottawa, Canada). Means were approximated by medians, and 95 percent empirical intervals were used to calculate approximate standard deviations. In the therapy reviews the majority of studies used a crossover design, therefore, standard errors of mean differences were either computed exactly using individual patient data, or imputed using an estimated within-patient correlation of 0.5. Meta-analyses were conducted using the random effects model where appropriate. The I2 statistic was used to assess heterogeneity.17 Planned subgroup analyses included age (children [less than 18 years old] versus adult [18 years and older]), severity of EIB/EIA as defined by the percent fall index on placebo (mild [less than 30 percent] versus moderate-severe [30 percent or more]), and patients with EIB versus EIA. All data pooling was performed using Review Manager 5.0 (The Cochrane Collaboration, Copenhagen, Denmark).

Therapy reviews

For each study the maximum percent fall in FEV1 on the placebo ECT was subtracted from the maximum percent fall on the treatment ECT to obtain the difference. For each comparison a mean difference (MD) and 95 percent confidence interval (CI) were calculated using all included studies weighted according to sample size. There were two secondary outcomes of interest: (1) the proportion of people who received complete protection from EIB/EIA with treatment compared with placebo, for which a risk ratio (RR) and 95percent CI were calculated; and (2) the degree of protection received over placebo, known as the “clinical protection index,” which was calculated using the following formula:

maximum %fall FEV1(placebo)-maximum %fall FEV1(treatment)maximum %fall FEV1(placebo)×100

Results - Diagnostic Test Accuracy Review

In addition to a history of symptoms suggesting EIB/EIA following exercise, objective diagnostic testing is necessary to confirm a diagnosis. A fall in the FEV1 of 10 percent or more after a challenge test is the recommended diagnostic threshold to objectively diagnose EIB/EIA.18 Testing can be done using either direct or indirect methods. Currently, there is no universally accepted gold standard test to confirm a diagnosis of EIB or EIA. In the absence of a gold standard, a standardized exercise challenge test (ECT) on a treadmill or bicycle ergometer conducted according to the American Thoracic Society guidelines18 is the best defined reference standard that is available and was used as the reference test in this report. The index tests that were assessed included a self reported history/symptoms diary, one direct challenge, and four indirect challenge tests. Methacholine (MCH) is considered a direct challenge as it acts directly on smooth muscle receptors to cause constriction independent of airway inflammation.19 The indirect challenges included two alternative exercise protocols (sport/venue-specific and free running asthma screening test [FRAST]), eucapnic voluntary hyperpnea (EVH), and inhaled mannitol. The indirect challenges are thought to cause inflammatory cells to release mediators such as leukotrienes, prostaglandin, and histamine which provoke smooth muscle constriction.20

Sensitivity and specificity are two measures of diagnostic test characteristics. Sensitivity is the probability of testing positive if EIB/EIA is really present. Specificity is the probability of testing negative if EIB/EIA is truly absent. As the sensitivity of the test under investigation (index test) increases, the number of persons with EIB/EIA who are missed (false negatives) will decrease. As the specificity of the index test increases, the number of persons without EIB/EIA who are incorrectly classified (false positives) will decrease. To avoid the risk of spectrum bias, diagnostic test accuracy studies should include a broad spectrum of participants who range from high to low likelihood of EIB/EIA.

Twenty-eight studies met the inclusion criteria for the questions relating to diagnostic test accuracy: self-reported history/symptoms diary21,22 (n=2), MCH23–37 (n=15), sport or venue specific challenges34,38–41 (n=5), EVH34,38,39,42–45 (n=7), FRAST36,40,46 (n=3), and mannitol23,47 (n=2).

Table 1 and the following summaries provide an overview of the results of the diagnostic test accuracy review. Sensitivity and specificity are reported for individual studies but data were not pooled due to substantial heterogeneity.

Table 1. Summary of sensitivity and specificity for the diagnostic test accuracy review.

Table 1

Summary of sensitivity and specificity for the diagnostic test accuracy review.

Key Question D-1. Self-Reported History or Symptoms Diary

Many people with potential EIB/EIA present with complaints of symptoms such as shortness of breath, cough, wheeze, and chest pain, associated with exercise18 and these symptoms have been used to diagnose EIB/EIA. There is a concern that self-report alone leads to an unacceptably high rate of false positive and false negative decisions.4

Results. Two studies21,22 met the inclusion criteria for the diagnostic test accuracy of self-reported history/symptoms diary compared with a standardized ECT.

Frobase et al.21 matched 20 teenage athletes who answered “yes” to the question “do you experience cough after exercise” to 20 who responded “no.” All participants performed a treadmill ECT. The authors used a fall in FEV1 of 10 percent or more to define a positive ECT. They reported a sensitivity of 89 percent (95 percent CI: 67, 99); a specificity of 86 percent (95 percent CI: 64, 97).

Rupp et al.,22 gave 166 teenage athletes a questionnaire with two questions relevant to EIB/EIA (“Do you have trouble breathing after running 1 mile and resting?” “Do you have to stop when running for ½ mile?”). Twenty-nine percent were identified as being at risk of EIB/EIA. All took a treadmill ECT. At a threshold of 15 percent or more, the authors reported a sensitivity of 36 percent (95 percent CI: 17, 59); a specificity of 85 percent (95 percent CI: 78, 90).

Methodological quality. Neither study included a representative spectrum of participants or defined their selection criteria and neither study blinded assessors to the results of the other challenge. It is unclear whether the ECTs met the ATS guideline recommendations.18 Furthermore, the number of “unclear” responses for both studies using the QUADAS tool raises questions regarding bias and generalizability.

Discussion. Despite extensive searching, only two studies addressed this question. The results of one study21 suggest that exercise-induced cough may be a predictor of EIB/EIA. Self-reported history and symptoms may be a starting point for further investigations but on their own are not sufficient to confirm a diagnosis of EIB/EIA.

Conclusion: There is insufficient evidence to conclude that self-report of symptoms is a reliable diagnostic tool for EIB/EIA. The quality of evidence is classified as “very low.” The initial grade based on study design was low and there are serious limitations related to study quality. The diagnostic thresholds differed and the sensitivity estimates are inconsistent and imprecise.

Key Question D-2. Methacholine Challenge

MCH is considered a direct challenge. Once inhaled into the lungs, it acts on smooth muscle acetylcholine receptors causing contraction and airway narrowing.23 It has been noted that airway responsiveness to pharmacological agents such as MCH is different from hyper-responsiveness to osmotic agents or to hyper-responsiveness following exercise.18

The threshold for a positive ECT was a maximum fall in FEV1 of 10 percent or more from a pre-exercise baseline.18 The threshold for a positive MCH challenge was a fall in FEV1 of 20 percent or more at a provocative concentration of less than 8 mg/ml (PC20).18 We also looked at a threshold PC20 of less than 16 mg/ml.

Results. Fifteen studies23–37 met the inclusion criteria for the diagnostic test accuracy review of a MCH challenge compared with a standardized ECT. The number of participants ranged from 12 to 375. Considerable heterogeneity was observed across the studies. Overall, 331 participants (57 percent) had a positive ECT and 416 (71 percent) had a positive MCH at the 8 mg/ml threshold. Both sensitivity and specificity ranged from 0 to 100 percent. Using the 16 mg/ml threshold, 420 participants (53 percent) had a positive ECT and 498 (63 percent) had a positive MCH. Sensitivity ranged from 55 to 100 percent and specificity 0 to 100 percent. To explore heterogeneity, we conducted a post hoc subgroup analysis on studies where all participants had mild to moderate stable asthma24–26,28,29,32,33,35,36 and studies27,30,34,37 that reported fewer than 50 percent of participants had confirmed asthma. The heterogeneity was reduced somewhat. Among studies in which all participants had asthma, the sensitivity ranged from 66 to 100, specificity 0 to 100 percent; in 4 studies reporting fewer than 50 percent had asthma sensitivity ranged from 0 to 50 percent and specificity 42 to 84 percent.

There was no reduction in heterogeneity when we explored the effects of age (child versus adult), MCH method (5 breath dosimeter versus 2 minute tidal breathing), or ECT parameters (air temperature and humidity, or treadmill versus bicycle ergometer).

Methodological quality. There are several methodological issues in this group of studies. Of concern is the risk of spectrum bias. Most studies did not report how participants were recruited into the study, nor did they describe the inclusion criteria; three studies recruited volunteers.26,30,34 Blinding of the ECT results to the results of the MCH challenge was not reported. It is unclear whether eight of the ECTs met the ATS guideline recommendations.18

Discussion. Across these 15 studies, there was considerable heterogeneity in estimates for sensitivity and specificity although sensitivity estimates varied less when examining patients with a confirmed diagnosis of asthma. Unlike ECTs, a positive airway response to MCH does not imply the presence of inflammatory cells or their mediators that are known to be present in EIB/EIA;20 therefore, it is not surprising that MCH shows variable sensitivity and specificity to detect EIB/EIA. It has been suggested18 that a positive MCH test should not be used to infer EIB/EIA and likewise, a negative MCH should not be used to exclude EIB/EIA.

Conclusion. Based on the available evidence we cannot conclude that the MCH challenge is a reliable test to diagnose EIB/EIA in those with confirmed asthma or in populations with mixed asthma prevalence. The quality of evidence is classified as “moderate.” The initial grading based on study design was high; however, there are serious limitations to quality of individual studies. Sensitivity estimates are reasonably consistent but are imprecise in the participants with EIA. Specificity estimates are inconsistent and imprecise.

Key Question D-3. Sport or Venue Specific Exercise Challenge

Sport/venue specific challenges are advantageous in that the athlete can exercise in the conditions where they become symptomatic and the tests are reasonably inexpensive. In comparing sport/venue specific challenges with an ECT, coordinating the timing of the two tests and ensuring good quality control on the sport challenges can be problematic. The environmental conditions, level of intensity, and increase in minute ventilation reached during both challenges may be hard to match.

The threshold for a positive ECT or a sport specific challenge was a maximum fall in FEV1 of 10 percent or more from a pre-exercise baseline.18

Results. Five studies34,38–41 met inclusion criteria for the diagnostic test accuracy review of sport or venue specific exercise challenges compared with a standardized ECT. Three focused on swimming challenges34,38,40 and two assessed cold weather sport challenges.39,41 Rundell et al.41 only included athletes who tested positive on a sport specific challenge. The sensitivity of the swimming challenge tests ranged from 0 to 50 percent; specificity from 83 to 100 percent. In total 11 (19 percent) participants were positive on the ECT and 5 (9 percent) were positive on the swim challenge. In the two studies involving winter athletes, Dickensen et al.39 had no true positives but a specificity of 79 percent, and Rundell et al.41 sensitivity of 100 percent but had no true negatives and a specificity of 0 percent.

Methodological quality. In these five studies there are some methodological weaknesses. None of the studies had a representative patient spectrum which suggests the possibility of spectrum bias. Blinding of ECT results to the results of the index test was not reported. It is unclear whether two of the ECTs met the ATS guideline recommendations.18

Discussion. These studies compared challenges in the specific sport the athletes practiced (e.g., biathlon) or at the venue of activity (e.g., skating arena, swimming pool) to a laboratory ECT. The estimates of sensitivity and specificity both, ranged from 0 to 100 percent. Sport specific challenges are one of the recommended challenges for diagnosing EIB/EIA in elite athletes;34 however, in total only seven athletes were positive in their own field of competition while 16 were positive on a standardized ECT. It is not clear why this occurred, although it is possible that the level of minute ventilation and differences in environmental conditions played a role.

Conclusion. Sport specific challenges may be useful in identifying athletes who do not have EIB/EIA but other testing such as a standardized ECT may also be required. The quality of evidence is classified as “low.” The initial grading based on study design was high; however, there are serious limitations related to study quality. The sensitivity estimates are inconsistent and imprecise. Specificity estimates are also inconsistent and imprecise.

Key Question D-4. Eucapnic Voluntary Hyperpnea

Eucapnic voluntary hyperpnea (EVH) is a test based on the premise that the increased minute ventilation with exercise is responsible for bronchoconstriction in susceptible subjects.48 The EVH challenge requires the person to increase their minute ventilation up to 80–85 percent of the maximum minute ventilation for 5–6 minutes, which approximates the minute ventilation obtained during vigorous exercise. EVH was developed as a surrogate for exercise to identify EIB49 and currently, it is the challenge recommended by the International Olympic Committee Medical Commission to identify EIB/EIA among Olympic athletes.19

The threshold for a positive ECT or EVH challenge was a maximum fall in FEV1 of 10 percent or more from a pre-exercise baseline.18

Results. Seven studies34,38,39,42–45 met inclusion criteria for the diagnostic test accuracy review of EVH compared with a standardized ECT. The number of participants ranged from 10 to 33. Substantial heterogeneity was identified. Forty-two (30 percent) participants had a positive ECT and 74 (54 percent) had a positive EVH. Sensitivity estimates ranged from 25 to 90 percent and specificity from 0 to 71. A post hoc subgroup analysis of studies in which participants did not have known EIB/EIA34,38,39,42 explained some of the heterogeneity. In this population, the sensitivity ranged from 25 to 75 percent and specificity from 29 to 71 percent. Among studies in which all participants had EIB/EIA,42–45 the sensitivity of EVH was high, ranging from 71 to 90 percent; however, one study had no true positive cases42 and three had no true negative cases.

Methodological quality. As a group, there are methodological weaknesses that limit the interpretation and generalizability of the results. Of greatest concern is the risk of spectrum bias in all studies. Either the participants recruited into the studies were volunteers,34,38,39,42,44 or the recruitment source and methods were not reported.43,45 In some studies all participants had a history of EIB/EIA43–45 and therefore are not representative of the spectrum of patients who might be tested for EIB/EIA. Blinding of the ECT results to the results of the EVH challenge was not reported. It is unclear whether the ECTs met the ATS guideline recommendations.18

Discussion. Overall, both the sensitivity and specificity of EVH compared to an ECT showed substantial heterogeneity and no trend was observed. It is unclear whether an EVH challenge is identifying the same people that experience a fall in FEV1 of 10 percent or more on an ECT, particularly in a population of athletes with an unclear history of asthma. The EVH challenge resulted in a higher proportion of false positives (FP) (i.e., negative on the ECT but positive on EVH) among participants with no or unknown history of EIB/EIA. The proportion of FP ranged from 25 to 71 percent. The participants included both elite athletes (n=63) and those with unknown activity levels (n=20). Based on the available evidence, it is unclear if the EVH challenge is more sensitive to EIB/EIA, or if the mechanism that triggers the bronchoconstriction is different from that for the ECT, or if level of minute ventilation achieved on an ECT according to ATS recommendations is less than the level achieved in an EVH challenge particularly in athletes. Among participants with a history of EIB/EIA, the proportion of FPs ranged from 0 to 5 percent in participants with unknown activity levels. Among athletes, the proportion with a FP result was 27 percent. Further research is needed to determine if the EVH challenge would be an appropriate add-on test to an ECT in athletes with symptoms of EIB/EIA.

Conclusion. Based on the available data, we cannot conclude that an EVH challenge is a reliable diagnostic test for EIB/EIA. The quality of evidence is classified as “low.” The initial grading based on study design was high; however, there are serious limitations related to study quality. Furthermore, the sensitivity and specificity estimates are inconsistent and imprecise.

Key Question D-5. Free Running Asthma Screening Test (FRAST)

Free running has often been used in screening large groups for EIB/EIA because the test is relatively easy to perform, requires limited equipment, and multiple subjects can be tested at the same time.48 The threshold for a positive ECT or FRAST was a maximum fall in FEV1 of 10 percent or more from a pre-exercise baseline.18

Results. Three studies36,40,46 met inclusion criteria for the diagnostic test accuracy review of FRAST compared with a standardized ECT. Participants were children under 15 years, and all free running tests took place indoors. In two studies36,40 the threshold for a positive FRAST was a fall in FEV1 of 10 percent or more. In these 34 participants, 13 (38 percent) had a positive ECT and 18 (41 percent) had a positive FRAST. The sensitivities were 60 and 67 percent; specificities were 47 and 67 percent. The remaining study46 used a threshold of a fall in FEV1 of 16.5 percent or more for the FRAST and 20 percent or more for the ECT. Based on these thresholds the sensitivity was 53 percent (95 percent CI: 34, 72); specificity was 100 percent (94 percent CI: 88, 100).

Methodological quality. Overall, there were concerns about the methodological quality of the studies. Generalizing the results to a target population of people with suspected EIB/EIA may be limited as none of the studies had representative spectrum of participants. Blinding of results of the ECT to the results of the FRAST was not reported. It is unclear whether two of the ECTs met the ATS guideline recommendations.18

Discussion. Sensitivity and specificity of FRAST were modest (60 to 67 percent and 47 to 67 percent, respectively). While the FRAST may be easy to perform, there is insufficient evidence to conclude that it accurately identifies people with EIB/EIA compared with a standardized ECT. Furthermore, in a FRAST challenge there is an overall lack of control in terms of exercise intensity and therefore stimulus (i.e., minute ventilation).

Conclusion. There is insufficient evidence to conclude that FRAST is a reliable diagnostic test for EIB/EIA. The quality of evidence is classified as “very low.” The initial grading based on study design was moderate; however, there are serious limitations related to study quality. The sensitivity and specificity estimates are consistent but imprecise.

Key Questions D–6. Mannitol Challenge

Recently, a mannitol provocation test has been developed to examine airway hyper-responsiveness. Like exercise, mannitol is thought to cause airway narrowing indirectly through increasing the osmolarity of the airway surface, leading to the release of endogenous mediators such as prostaglandin, leukotrienes and histamine, and resulting in smooth muscle contraction.50,51 Mannitol has the advantage that it can be performed with minimal equipment (i.e., a metered-dose inhaler and spirometer).

The threshold for a positive ECT was a maximum fall in FEV1 of 10 percent or more. The threshold for a positive mannitol challenge was a 15 percent drop in FEV1 at less than 635mg or a drop of 10 percent between consecutive doses.23

Results. Three studies23,47,52 met the inclusion criteria for the diagnostic test accuracy review of mannitol compared with a standardized ECT. All participants in Brannan et al.47 and Kersten et al.52 had diagnosed asthma. In these 58 participants, 39 (67 percent) had a positive ECT test, 35 (67 percent) had a positive mannitol challenge. The sensitivity was 69 and 96 percent, respectively. Specificity was 78 percent in the Kersten et al.52 study. Anderson et al.23 included those with suspected but not confirmed asthma and reported a sensitivity of 58 percent (95 percent CI: 50, 66) and a specificity of 65 percent (95 percent CI: 58, 72).

Methodological quality. Concern regarding the risk of spectrum bias in the studies by Brannan et al.47 and Kersten et al. limits the conclusions that can be drawn.52 Participants were volunteers who had a history of EIA and are not representative of the spectrum of patients who might be tested for EIB/EIA. Blinding of the results of the ECT and the index test was not reported. All three of the ECTs met the ATS guideline recommendations.18

Discussion. Both mannitol and ECT are considered indirect tests, acting through a similar mechanism of increasing osmolarity of the airway surface. The advantage of mannitol is that exercise is not a requirement, and the test can be conducted in a physician’s office with minimal equipment. Among adults and children with a history of EIA or suspected asthma, the diagnostic test characteristics for mannitol compared to an ECT hold promise. Its role as a screening or diagnostic tool requires further study.

Conclusion. Based on these data, it is difficult to conclude that the mannitol provocation test is a reliable diagnostic test for EIB/EIA. The quality of evidence is classified as “moderate.” The initial grading based on study design was high; however, there are concerns regarding the limitations related to study quality. The sensitivity and specificity estimates are reasonably consistent but are imprecise.

Results – Therapy Reviews

The primary objective of therapy is to prevent EIB/EIA from occurring or, short of that, to at least attenuate the degree of bronchoconstriction. Pharmacologic effect is assessed by taking a single dose of an inhaled agent prior to an ECT that meets intensity and ventilation standards. FEV1 or PEF is measured immediately following the challenge and at 5 minute intervals until lung function begins to improve. The maximum fall in FEV1 is expressed as a percent of the baseline measure taken immediately prior to the ECT. The greater the percent fall, the worse the EIB/EIA or, conversely, the less the percent fall the more effective the therapy. This report conducted systematic reviews of randomized controlled trials (RCTs) involving four categories of prophylactic agents compared with a placebo: inhaled corticosteroids (ICS), leukotriene receptor antagonists (LTRA), mast-cell stabilizers (MCS), and short-acting anticholinergics (SAAC). We conducted two additional reviews. The first was to determine if using short- or long-acting beta agonists (SABA or LABA) on a daily basis caused people to develop a tolerance, or tachyphylaxis, to these agents thus reducing the prophylactic effect on EIB/EIA. If the post-exercise drop in FEV1 increases after prolonged use of SABA or LABA, it is an indication of tachyphylaxis. The second review investigated whether engaging in a specific warmup routine prior to an ECT caused a refractory period during which time the EIB/EIA response was prevented or attenuated on a subsequent ECT.

The primary outcome for these six reviews was the mean difference (MD) between the mean maximum percent fall in FEV1 on drug or warmup and the mean maximum percent fall in FEV1 with placebo or no warmup. In this executive summary, we report the absolute MD in order to phrase the results using positive numbers. In the forest plots in the main report, a negative MD that falls to the left of the null line indicates the drug performed better than the placebo and, by extension, a larger difference indicates a greater drug effect. The 95 percent CIs relate to whether or not the result represents a statistically significant improvement.

Overall, 109 studies were included in the therapy reviews. The included studies addressed the following research questions: tachyphylaxis to SABA and LABA53–63 (n=11), LTRA44,64–71 (n=9), ICS72–75 (n=4), MCS (n=3 Cochrane reviews and publications76–78), SAAC79–96 (n=18), and a refractory period97–103 (n=7).

Table 2 and the following summaries provide an overview of the results. The pharmacological treatments were further divided into bronchodilating and anti-inflammatory agents.

Table 2. Prophylactic therapy for EIA: Pulmonary function results.

Table 2

Prophylactic therapy for EIA: Pulmonary function results.

Pharmacological Interventions: Bronchodilating Agents

T–1. Tachyphylaxis to SABA or LABA Therapy

β2-Receptors on the airway smooth muscle are responsible for bronchodilation; beta-agonist agents attach to these receptors and are effective bronchodilators and improve breathing. SABA agents have been the mainstay of EIB/EIA treatment for many years. With the development and availability of LABA agents, more research has been applied to the effectiveness of these agents in preventing EIB/EIA. It has been reported in the literature that some people develop a tolerance (tachyphylaxis) to long term regular use of beta-agonist drugs and they lose some of their effectiveness and protective effect.57,59

Seven studies met inclusion criteria to address the question of whether people with EIB/EIA develop tachyphylaxis to LABA agents53–61 and two met the inclusion criteria for SABA agents.62,63 Overall, the quality of the trials was high. All were randomized, double-blind trials that described the withdrawals and dropouts. Five described an adequate process to ensure double-blinding.

SABA. Data from the two SABA crossover studies could not be combined and are therefore described individually. Both studies involved asthmatics who took 1 week of regular (four times per day) SABA or placebo prior to exercise challenges and measured the baseline FEV1 at the end of the treatment weeks. In one study63 there was a statistically significant 230 ml difference (p=0.02) in the measure of FEV1 after the treatment weeks that favored placebo. In the second study,62 there was no significant difference between the mean FEV1 after the placebo week compared with the salbutamol week (2.8L versus 2.8L).

Only the study by Inman and O’Byrne63 assessed prophylaxis. When the 10 adults took a single dose of placebo prior to an ECT on day 8, the mean fall in FEV1 was positive for EIA in both groups; however, the fall was slightly greater after the SABA week (29.4 ± 4.7 percent versus 24.9 ± 4.4; p=0.12). On day 9 following pre-treatment with a SABA, all received total protection from EIA; however, the fall in FEV1 was greater after the SABA week (5.1 ± 2.0 percent versus 1.1 ± 0.8; p=0.05) and remained lower throughout the recovery period. Adverse events were not reported. Hancox62 randomized 8 women with EIA and following a week of study drug performed an ECT with no pretreatment but gave SABA 5, 10, and 15 minutes post ECT. The fall in FEV1 was greater in the SABA treated group (p=0.001).

LABA. Four crossover trials55,57–59 randomized groups to 3 to 4 weeks of salmeterol55,57 once a day,59 or every other day58 or to a matching placebo. The average fall in FEV1 after the first ECT on day 1 was 3.7 percent in the LABA arms compared with 26.9 percent in the placebo arms. The absolute MD was 25.1 percent (95 percent CI: 18.0, 32.3) In the individual studies the percent fall in FEV1 following an ECT at both 2 weeks and 4 weeks was of a greater magnitude in the LABA arms than at day 1, indicating a greater degree of EIA and less protection from salmeterol. The fall in FEV1 in the placebo arms remained unchanged or showed a small decrease. After 4 weeks of daily LABA use, the average fall in FEV1 after an ECT (3 studies) increased to 11.4 percent in the LABA arms compared with a decrease to 21.3 percent in the placebo arms. The pooled MD was 10.5 percent (95 percent CI: 14.7, 6.4. Three headaches were attributed to salmeterol; one exacerbation and one complaint of chest tightness were attributed to placebo.

The study by Garcia et al.54 compared daily formoterol with placebo and also concluded that twice daily LABA over 4 weeks caused a significant reduction in bronchoprotection against EIA. Though the LABA was still effective, tachyphylaxis was evident by day 14. Adverse events were not reported.

Two studies compared regular use of salmeterol versus a LTRA (montelukast) once in the evening for 8 weeks.53,60 Both drugs attenuated the EIA response after 3 days of treatment to a similar degree. The mean fall in FEV1 on LABA was 19.8 percent compared with 18.2 percent on LTRA (MD=1.01; 95 percent CI: −2.2, 4.2). Montelukast maintained its effectiveness at 4 and 8 weeks compared with a slight decrease in effect with regular use of salmeterol. At 8 weeks the mean fall in FEV1 on LABA increased to 23.0 percent compared with a slight decrease to 17.1 percent on LTRA. The pooled MD was 5.4 (95 percent CI: 2.2, 8.7). This evidence, although indirect, provides additional support for the concept of tachyphylaxis following regular use of LABA agents in patients with EIA. Adverse events were infrequent, mild and occurred more or less equally between therapies.

Discussion. Based on the results of one study,63 SABA agents were found to have a prophylactic effect on EIA. The day 1 data from the LABA tachyphylaxis review also show that LABA is effective for attenuating the drop in FEV1 post exercise. However, the evidence indicates tachyphylaxis is associated with the regular use of both SABA and LABA agents. The results of this review demonstrated that the percent fall in FEV1 following an ECT after 1 to 4 weeks of use was greater (i.e., drugs were becoming less effective) in the SABA and LABA arms than in the placebo arms. Notwithstanding the tachyphylaxis, a prophylactic effect for SABA and LABA was still observed.

Conclusion. Given the consistency of results, the magnitude of effect and concordance with clinical consensus, the evidence indicates that prophylactic use of SABA and LABA agents is safe and effective for episodic prophylaxis of EIA. The evidence suggests that tachyphylaxis develops if these agents are used daily. The body of evidence comparing SABA/LABA to placebo was graded as “moderate”; evidence comparing LABA to LTRA was graded as “low.”

Pharmacological Interventions: Anti-Inflammatory Agents

T–2. Leukotriene Receptor Antagonist Therapy

Leukotrienes are produced by the cysteinyl leukotriene pathway and are implicated in both bronchoconstriction and in the inflammatory cascade leading to worsening asthma. LTRAs are relatively new agents available for the management of chronic and acute asthma which block this pathway.

Nine crossover trials44,64–71 met the inclusion criteria for the single prophylactic use of LTRA compared with placebo in the pre-treatment of EIA. Overall, the methodological quality of the trials was good. All were randomized, eight were double-blind trials and six described an adequate process to ensure double-blinding. Seven described the withdrawals and dropouts; however, none described the method of randomization and only one had adequate allocation concealment. All trials recruited people with asthma and confirmed EIA and all recorded the mean fall in FEV1 up to 2 hours after drug administration. The pooled results showed that LTRA was more effective than placebo in preventing EIA. The average fall in FEV1 in the LTRA arms after the first ECT ranged from 7.6 to 13.3 percent (average 10.9) compared with 15.0 to 23.2 percent (average 18.5) on the placebo challenges. The absolute MD was 8.3 percent (95 percent CI: 6.9, 11.0; I2=65 percent) and represented clinical protection of approximately 45 percent over placebo. Three trials68,70,71 recorded the mean maximum fall in FEV1 24 hours after drug administration. The average fall in the placebo arms was 13.8 percent (range 10.7 to 16.9) compared with 8.7 percent (range 8.3 to 10) in the LTRA arms. The pooled results showed that LTRA remained more effective than placebo in preventing EIA (MD=4.9; 95 percent CI: 1.8 to 8.0; I2=76 percent). Adverse events were infrequent, mild and occurred more or less equally between LTRA and placebo groups.

Discussion. We identified nine trials of LTRA compared with placebo in the pre-treatment of EIA. The methodological quality of the studies was high and the pooled results demonstrated a consistent benefit associated with prophylactic use. LTRAs reduced the maximum percent fall in FEV1 at the earliest measurement by approximately nine percent compared with placebo (Table 2).

Conclusion. From the available evidence, it appears that LTRAs are effective and safe as pre-exercise treatment for patients with mild to moderate stable asthma and EIA. The strength of the body of evidence is “moderate.”

T–3. Inhaled Corticosteroid Therapy

The main inhaled anti-inflammatory agents used for chronic and acute asthma are ICS agents. The search and selection methods employed for this review identified trials involving single prophylactic ICS use compared with placebo in the pre-treatment of EIA.

Four crossover trials72–75 met the inclusion criteria. Overall the quality of the trials was good. All were randomized and double-blind; two described an adequate process to ensure double-blinding. All described the withdrawals and dropouts; however, none described the method of randomization and one had inadequate allocation concealment. All trials recruited people with stable asthma and confirmed EIA. The pooled results showed no significant difference between ICS and placebo (absolute MD=5.0 percent; 95 percent CI 0.0, 10.0; I2=0 percent). None of the studies reported adverse events.

Discussion. The pooled results failed to produce a statistically significant or clinically important reduction in the percent fall in FEV1 compared with placebo.

Conclusion. From the available evidence, we cannot concude that a pre-exercise single-dose of an ICS agent will attenuate EIA. The strength of the body of evidence is “moderate.”

T–4. Mast Cell Stabilizers

Three systematic reviews were identifed that synthesized the evidence for MCS. One review compared nedocromil sodium (NCS) versus placebo for preventing EIA76,104 and one compared NCS versus sodium cromoglycate (SCG).77 The third review78 compared the effects of a prophylactic dose of either NCS or SCG (collectively called MCS) to that of atropine, ipratropium or oxitropium bromide (collectively called SAAC agents) and to SABA. The review also compared the effects of a combination of a SABA plus a MCS to a SABA alone. Our search strategy located no additional RCTs that would add to these reviews. In the three reviews, all trials recruited people with stable asthma and confirmed EIA.

Twenty-one RCTs were included in the NCS versus placebo review.76,104 NCS provided a statistically and clinically significant attenuation of EIA. The average fall in FEV1/PEF was 15.2 percent (range 10.0 to 29.8) in the NCS arms compared with 31.5 percent (range 17.5 to 47.2) in the placebo arms (absolute MD=15.6 percent; 95 percent CI: 13.2, 18.1; I2=20 percent). This represents a clinical protection index of approximately 51 percent over placebo. In those that had more severe EIA (i.e., a percent fall index 30 percent or more), the effect of NCS appeared more pronounced (absolute MD=21.4 percent; 95 percent CI: 25.5, 17.2).

There were nine RCTs included in the NCS versus SCG review.77 No significant differences between NCS and SCG were identified (absolute MD=0.88; CI: −4.5, 2.7; I2=0 percent).

There were 24 RCTs included in the review comparing MCS to other agents. Overall, the maximum fall on MCS was reduced to 7.1 percent compared with 13.8 percent on SAAC agents (MD=6.7 percent; 95 percent CI: 3.3, 10.0; I2=0 percent). When compared with SABA, MCS were not as effective at preventing EIA. The mean percent fall in FEV1 using MCS was 11.2 percent compared with 4.3 percent on SABA (MD=6.8 percent; 95 percent CI: 4.5, 9.2; I2=0 percent). Combining MCS with a SABA did not produce significant advantages to pulmonary function over SABA alone (MD=1.8 percent; 95 percent CI: −1.1, 4.6; I2=0 percent). Few trials reported adverse events due to MCS; those that did reported bad taste, throat irritation and cough.

Discussion. MCS stabilize the mast cell membranes and prevent the release of inflammatory mediators; they are used as weak anti-inflammatory agents in mild to moderate chronic asthma. Examining existing Cochrane reviews, this report highlights the efficacy (approximately 51 percent improvement in the post-exercise FEV1 compared with placebo) for NCS. An analysis failed to demonstrate a difference between NCS and SCG and further comparative analyses suggest that MCS are more effective than SAAC agents but less effective than SABA agents (Table 2).

Conclusion. From the available evidence, it appears that MCS agents are effective and safe as pre-exercise treatment for patients with stable asthma and EIA. No significant differences were identified between NCS and SCG on pulmonary function or degree of protection afforded to patients. MCS agents were somewhat more effective than SAAC agents but not as effective as SABA agents. The combination of SABA and MCS agents did not provide significant advantages over a SABA alone. The strength of the body of evidence is “moderate.”

T–5. Anticholinergic Therapy

SAAC agents are used in respiratory conditions to decrease mucus production and as weak bronchodilators. The original anticholinergic agent used in asthma was atropine; however, ipratropium bromide (IB) is now the most commonly used agent in this class. Long-acting anticholinergic agents such as titropium bromide are also now available. The search and selection methods employed for this review trials involving single prophylactic SAAC use compared with placebo in the pre-treatment of EIA.

Eighteen crossover trials79–96 met the inclusion criteria. Overall, the methodological quality was low. Though all were randomized, seven were not double-blind trials and only one described an adequate process to ensure double-blinding. Seventeen trials described the withdrawals and dropouts; however, none described the method of randomization and all had unclear allocation concealment.

All trials recruited people with stable asthma and confirmed EIA. The mean fall in FEV1/PEF on the placebo challenges ranged from 14 to 41 percent (average 32). Seven of the placebo groups79,83,85,88,90,93,96 had a mean fall in FEV1/PEF greater than 30 percent indicating moderate to severe EIA. The average fall in FEV1/PEF in the SAAC arms ranged from 10 to 33 percent (average 21). The pooled results showed that IB was more effective than placebo in preventing EIA (absolute MD=8.8 percent; 95 percent CI: 5.0, 14.6; I2=76 percent). Four trials compared atropine to placebo.80,84,86,91 The pooled results showed that atropine was more effective than placebo in preventing EIA (absolute MD=16.0; 95 percent CI: 10.2, 21.7; I2=0 percent). Two trials compared oxitropium bromide to placebo.88,89 The pooled results showed that oxitropium bromide was more effective than placebo in preventing EIA (absolute MD=13.8; 95 percent CI: 6.0, 21.6; I2=0 percent). Few trials reported adverse events; those that did reported dry mouth or thirst, bitter taste, and slight tremor.

Discussion. The pooled results demonstrated a modest but consistent benefit associated with the pre-exercise use of IB in EIA. IB reduced airway narrowing at the earliest measurement by 34 percent over placebo; however, the results should be viewed cautiously due to the presence of heterogeneity. Complete protection was achieved 4.5 times more often with IB than placebo.

Conclusion. The evidence suggests that when used as a pre-treatment, SAAC agents are effective and safe for patients with EIA and can offer a clinically relevant protective effect to some people. The strength of the body of evidence is “moderate.”

Nonpharmacologic Interventions

T–6. Refractory Period (10 to 15 Minute Warmup and/or Cooldown)

As an alternative, or in addition to using medications to attenuate EIB/EIA, many athletes, trainers and researchers advocate specific warmup routines as a method to trigger a refractory period. A refractory period is the time after the warmup routine during which further vigorous exercise will evoke significantly less severe or no EIB/EIA.

Seven trials97–103 met the inclusion criteria. Overall, the methodological quality was low. All were randomized, but none were double-blind trials. Six described the withdrawals and dropouts; however, none described the method of randomization and all had unclear allocation concealment. Six of the seven trials reported participants had stable asthma and confirmed EIA.

The effect of a warmup routine on EIA prior to a standard exercise challenge was examined for the following subgroups.

Interval protocol. Four trials compared an interval warmup to no warmup97,99,100,103 prior to an ECT. The MD in the percent fall FEV1 on the ECT ranged from an improvement of 4.8 to 16.1 percent over the fall after no warmup. The pooled results showed that a series of short intense sprints attenuated the EIA response by a mean of 10.6 percent (95 percent CI: 6.5, 14.7; I2=15 percent).

Continuous low intensity protocol. Three trials compared a continuous low intensity warmup that ranged from 3 minutes101 to 30 minutes102 to no warmup prior to an ECT. The mean difference in the percent fall in FEV1 on the ECT ranged from an improvement of 0 to 20.6 percent over the fall after no warmup. The pooled results failed to demonstrate statistically significant evidence that the low intensity warmup attenuated the EIA response (absolute MD=12.6 percent; 95 percent CI: −1.5, 26.7; I2=90 percent).

Continuous high intensity protocol. Two trials compared a continuous high intensity warmup that was identical to the ECT with no warmup prior to the ECT.102,103 The mean difference in the maximum percent fall in FEV1 on the ECT ranged from an improvement of 1.0 to 17.6 percent over the fall after no warmup. The pooled results failed to demonstrate statistically significant evidence that a continuous high intensity warmup attenuated the EIA response. (absolute MD=9.8 percent; 95 percent CI: −6.4, 26.0; I2=89 percent).

Discussion. Seven RCTs compared either different warmup routines with each other or with no warmup prior to an ECT. The evidence suggests that compared with no warmup, pre‐treatment of EIA with interval warmup exercise offers a statistically significant and homogeneous attenuating benefit of approximately 11 percent improvement in the percent fall FEV1 index (Table 2). For both high and low intensity continuous warmup protocols, the evidence is less clear. Combinations of interval and continuous warmup protocols compared within one study identified no differences and all protocols provided similar protection against EIA in the 10 to 11 percent range.

Conclusion. From the available evidence, it appears that certain warmup protocols are effective in reducing the degree of airway obstruction associated with EIA. Combination warmups show promise; however, it is unclear whether continuous warmups are effective in this condition. The strength of the body of evidence is “low.”

Gaps in Evidence and Recommendations for Future Research

Methodological Limitations of the Reviews

This review has methodological and logistic limitations. Some of the important limitations are listed below.

Diagnostic test accuracy reviews. Since there is no universally accepted gold standard to diagnose EIB/EIA, we used the standardized ECT based on ATS guidelines as our reference standard. Not all studies performed the ECT according to the ATS guidelines, however, and the variations in ECT protocol among studies may have affected the pulmonary response to the challenge and underestimated the number of positive results. Specifically, there were variations in environmental conditions (percent humidity and temperature of inspired air); restrictions on nose breathing; speed, grade, intensity and duration of the challenge; and target heart rate and minute ventilation achieved during the ECT. All of these have the potential to influence the stimulus for triggering EIB/EIA. These variations may have contributed to the heterogeneity in sensitivity and specificity observed with all the index tests.

Another concern is that many studies suffered from a potential spectrum bias. Ideally, studies should have a sample of participants who are representative of the population being examined for EIB/EIA. That is, studies should have sufficient subjects that span the range from high to low likelihood of EIB/EIA. In this review, many studies purposely recruited volunteers who either had known EIB/EIA, or definitely did not have it. In several studies no information on baseline likelihood was reported. The lack of a representative patient spectrum limits the generalizability of the results to a target population of people with suspected EIB/EIA and makes estimates of sensitivity and specificity unreliable.

A further limitation is that not all potentially relevant studies reported sufficient data to generate a 2x2 table comparing the reference standard result with an index test result based on one or more diagnostic thresholds.

Finally, poor reporting of study methods, challenge criteria, or participant characteristics meant we were not able to perform important subgroup analyses to provide more useful evidence summaries.

Therapy reviews. A double-blind RCT is considered the highest quality study design for assessing drug therapy. A short-coming of many of the included trials was a lack of disclosure on the methods used to generate and conceal the randomization code from investigators and participants as well as the adequacy of the blinding process.

Concerns regarding crossover trials, even though randomized, center on three factors: drug carryover effects, period effects, and statistical issues. Ideally, results should be reported for each treatment period and the test sequence reported so analyses can confirm the presence or absence of a carryover or period effect. No study reported data in such a manner and all merged the results from the treatment and placebo periods as though it were a parallel study.

Studies did not always provide sufficient quantitative data (e.g., measures of variance, mean endpoint estimates, statistical test results) needed to pool the individual trial results and to judge the treatment-related between group outcome differences.

In studies assessing the effect of a refractory period, the warmup protocols were not standardized. We categorized the warmup interventions into three groups, but still within each group the warmup approaches varied from each other.

Studies varied in the presentation of FEV1 and PEF data. Our primary outcome was the maximum percent fall in FEV1 from baseline. When possible we transformed data to this measure. The ATS convention is to accept a 10 percent maximum fall in FEV1 following ECT as the diagnostic cutpoint for EIB/EIA and the majority of studies used this. However, other cutpoints were used and it was not always possible to obtain data at the 10 percent level. The review provided data on 15 and 20 percent cutpoints when available. While outcome reporting mainly included detailed pulmonary function measures, data on the proportion obtaining clinical and complete protection were not always reported nor was it standardized (Table 2).

Side effect profiles were also poorly reported. While side effects were likely uncommon, better reporting and the use of patient treatment preference may have provided more robust assessments of the effectiveness of various treatment options. Finally, most studies relied exclusively on pulmonary function outcomes to assess effectiveness; symptom control and exercise performance data were not available.

Recommendations for Future Research

Efforts are needed to improve the overall quality of reporting of primary studies of diagnostic test accuracy. The STARD checklist105 details 25 items that address the level of detail that should be specified within such studies including descriptions of participants, tests methods, statistical methods, and results. This could be considered as a guide for authors reporting studies that evaluate diagnostic tests and for journals that publish EIB/EIA-related research.

Studies designed to more carefully examine the methodology of the standard ATS ECT test are needed. Specifically, guidelines state that the inspired air temperature must be less than 25oC, and less than 50 percent relative humidity. Studies have used inspired air as cold as -18oC and many have used medical compressed air, which has a relative humidity of 0 percent as the air leaves the tank. Colder, drier air would result in the greatest increase in osmolarity of the airway surface and thus most likely cause EIB/EIA. Additional studies are needed to more thoroughly examine how inspired air affects EIB/EIA so that a more standardized ECT guideline can be developed.

Future purposely designed studies to compare the diagnostic characteristics of different diagnostic methods are needed. New tests such as mannitol are encouraging; however, currently there are insufficient data to allow for a strong recommendation of this test. Particular attention in future studies must be given to including a representative sample of participants with suspected EIB/EIA. Studies that prospectively recruit participants and blind the reference standard results to those who interpret the index test results are needed.

To determine if the response to diagnostic tests differs in those with EIB versus EIA, those with atopy or no atopy, or other potentially defining characteristics, appropriate populations need to be included, adequate baseline data reported, and comparative analyses by the characteristics of interest performed and reported.

With regard to the systematic review of therapy for EIA, there are several issues with respect to methodological quality, populations, interventions, outcome assessments, and controls that require discussion.

Efforts are needed to improve the quality of reporting primary studies (i.e., randomized controlled trials). The CONSORT Statement106 could be considered as a guide for authors reporting trials and journals that publish EIB/EIA-related research. Most trials in this review used a crossover design. Concerns regarding crossover trials center on three factors: drug carryover effects, period effects, and statistical issues. Data should be reported in a manner that allows analyses to confirm the presence or absence of a carryover effect. Future studies should focus on complete reporting of results by period and sequence to assure readers that these concerns have been accounted for.

Population: The populations involved in the therapy trials all had stable asthma and confirmed EIA. Athletic status was not reported. This finding, coupled with the small number of studies investigating some interventions, precluded more detailed subgroup analysis of the issue of EIB/EIA in elite athletes. Additional investigations of treatment effects in patient subgroups defined by asthma severity, age, and activity level are clearly indicated.

More trials in clinically homogenous groups of patients with EIB and patients with EIA are needed to better explore and characterize differences in the efficacy of interventions between these two conditions.

Intervention: The nonpharmacologic interventions were not standardized. Although the warmup interventions were divided into similar groups, each warmup approach varied from the others. Future research exploring different standardized warmup approaches is clearly indicated.

Control: Many of the patients in the pharmacological studies were administered placebo agents as the control treatment. These agents were often similar in appearance or delivered in similar appearing devices (inhaler agents) to the active treatment; however, it may have been possible for patients to detect differences among treatments, especially the bronchodilators such as SABA, LABA and SAAC agents. It is impossible to determine how frequently this occurred since most studies did not report the patient’s perception of treatment received. In addition, it was not possible to blind the participants to the warmup programs in the refractory studies. When the participant cannot be blinded, it is particularly important that the outcome assessor be blinded to all information that could bias the outcome measure or assessment (e.g., other test results, intervention given, challenge performed).

Outcome: The pulmonary function outcomes reported varied and studies used different diagnostic cutpoints ranging from a fall index of 10 to 20 percent. The ATS convention is to accept a 10 percent maximum fall in FEV1 While outcome reporting included detailed pulmonary function measures, the format was not standardized. Data on the proportion obtaining clinical and complete protection were often missing (Table 2).

Side effect profiles were also poorly reported. While there is evidence for the safety of many of these agents from the chronic asthma literature, it would be prudent for future research to capture adverse effects in the EIB/EIA population. Future trials of interventions to prevent or attenuate EIB/EIA should include clinically relevant secondary outcomes such as patient preferences, symptom scores, and sport performance effects (e.g., changes in athletic performance or endurance). More robust outcome reporting would further inform decision-making by athletes, physicians, and sporting bodies.

Conclusion

Despite exhaustive efforts to identify evidence regarding the diagnostic test characteristics of self-reported history, MCH, sport or venue specific challenges, EVH, FRAST, or mannitol to diagnose patients with EIB/EIA, few studies were found that compared these diagnostic tests to a standardized ECT. The studies that were identified suffered from spectrum bias and considerable variability in test methodology and reported data.

Based on the available data as summarized in this review, none of the six tests provide the diagnostic test characteristics to make them individually attractive as an exclusive alternative to a standardized ECT. There is no clear evidence to suggest that any are equivalent to, or better than, a standardized ECT to diagnose EIB/EIA. More important, given the difference in response by a single individual to multiple challenges, a single negative test should not be used to exclude EIB/EIA. A specificity ranging from 79 to 100 percent in sport specific challenges may indicate they are useful in identifying athletes who do not have EIB/EIA.

Despite exhaustive efforts to locate randomized trials that compared the therapeutic effectiveness of six different interventions for EIB/EIA against placebo, only a small number of studies were identified for each question. The studies that were included suffered from several potential biases; however, this summary represents the most comprehensive review of therapies for EIB/EIA ever reported.

On the basis of this review, we can conclude that the bronchodilator agents examined (SABA, LABA, LTRA, and SAAC) are effective to varying degrees at attenuating or eliminating the drop in FEV1 associated with EIA. There is evidence of the development of tachyphylaxis associated with 1 to 4 weeks of regular use of SABA and LABA agents but it was not enough to negate their usefulness altogether.

On the basis of this review, we can conclude that the anti-inflammatory agents examined provide mixed results. The evidence suggests that MCS agents are effective in attenuating or eliminating the FEV1 drop associated with EIA. There is some role for LTRA agents in the treatment of EIA; however, the attenuation appears less than with other bronchodilator agents. There is limited evidence that a single prophylactic dose of ICS is of clinical benefit in preventing EIA. The long-term benefit of ICS agents in asthma is well known, and it is possible that better control of chronic airway inflammation may benefit patients with EIA specifically.107

Finally, from the available evidence, it appears that certain warmup protocols are effective at reducing the degree of airway obstruction associated with EIA. Combination warmups show promise; however, it is unclear whether continuous warmups are effective for this condition.

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...