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.)

Cover of Exercise-Induced Bronchoconstriction and Asthma

Exercise-Induced Bronchoconstriction and Asthma.

Show details

3Results

Diagnostic Test Accuracy Review

Literature Search

The search strategy for the diagnosis of EIB/EIA identified 5,318 citations from electronic databases and six additional references by handsearching (Figure 1). Screening these titles and abstracts identified 434 potentially relevant references. Ten studies could not be retrieved (Appendix D). Most of these were abstracts that were only available in supplementary issues of journals; they were requested through interlibrary loan but did not arrive by the time this report was written. The search update identified an additional six studies for further evaluation. Overall 28 studies met the inclusion criteria for this report; 412 studies were excluded. The included studies addressed the following research questions: self-report (n=2),21,22 methacholine (MCH) challenge (n=15),23–37 sport/venue specific challenges (n=54),34,38–41 eucapnic voluntary hyperpnea (EVH; n=7),34,38,39,42–45 FRAST (n=3),36,40,46 and mannitol (n=3).23,47

Figure 1. Diagnosis of EIB/EIA: Flow diagram for study retrieval and selection.

Figure 1

Diagnosis of EIB/EIA: Flow diagram for study retrieval and selection. *Some studies address more than one research question

The main reasons for excluding studies were (1) the diagnostic technique did not assess EIB/EIA (n=158), (2) inappropriate study design (n=116), (3) there was no comparison group (n=46), (4) there were insufficient data to construct a 2x2 table (n=26), (5) the study assessed other diagnostic tests that were not included in this report (n=24), and (6) there was an inappropriate reference standard (n=21). Eleven studies were excluded for other reasons. The list of excluded studies and reasons for exclusion are identified in Appendix D.

Key Question D–1: Self-Reported History or Symptoms Diary

Description of Included Studies

Two studies21,22 met the inclusion criteria for the diagnostic test accuracy review of self-reported history/symptoms diary compared with a standardized exercise challenge test (ECT). Tables 6 and 7 summarize the study and participant characteristics.

Table 6. Description of studies in the diagnostic test accuracy review: Self-report vs. ECT.

Table 6

Description of studies in the diagnostic test accuracy review: Self-report vs. ECT.

Table 7. Description of participants in the diagnostic test accuracy review: Self-report vs. ECT.

Table 7

Description of participants in the diagnostic test accuracy review: Self-report vs. ECT.

In the study by Frobase et al.21 20 junior high school athletes (11 to 15 years) who responded “yes” to the question “do you experience cough after exercise?” were included. They were matched to 20 controls who responded “no” to the same question. All participants underwent a treadmill ECT. They ran for 6 minutes at an intensity of 85 to 90 percent of predicted maximum heart rate. No information about a history of asthma was provided.

In the study by Rupp et al.22 166 high school athletes (12 to 18 years) completed a self-report questionnaire that included two questions relevant to EIB/EIA (“do you have trouble breathing after running 1 mile and resting?” and “do you have to stop when running for ½ mile?”). Forty-eight students (29 percent) were identified as being at risk for EIB/EIA based on baseline spirometry or by clinical history. All participants underwent a treadmill ECT. They ran for 6 minutes after reaching 80 percent of their predicted maximum heart rate.

Methodological Quality of the Included Studies

Table 8 summarizes the methodological quality of the included studies. The number of “unclear” responses for both studies raises questions regarding bias and generalizability. Neither study had a representative patient spectrum which suggests the possibility of spectrum bias. Blinding to the results of the ECT and the index test was not reported. There were no details about uninterpretable or indeterminate index test results. One study22 did not report their source of funding; one21 was supported by government grants. The body of evidence is classified as “low.”

Table 8. Methodological quality of the studies in the diagnostic test accuracy review: Self-report vs. ECT.

Table 8

Methodological quality of the studies in the diagnostic test accuracy review: Self-report vs. ECT.

Quantitative Results

The two studies used different thresholds to define a positive result on the ECT. Using a threshold of a maximum percent fall in FEV1 of 10 percent or more, Frobase et al.21 reported a sensitivity of 89 percent (95 percent CI: 67, 99) and a specificity of 86 percent (95 percent CI: 64, 97). In the study group, 17 (85 percent) participants had a positive ECT; in the control group 2 (10 percent) had a positive ECT.

Using a threshold of 15 percent or more, Rupp et al.22 reported a sensitivity of 36 (95 percent CI: 17, 59) and specificity of 85 percent (95 percent CI: 78, 90). Twenty participants (13 percent) had a positive ECT; 30 (18 percent) were at risk of EIB/EIA based on the questionnaire.

Key Question D–2: Methacholine Challenge

Description of Included Studies

Fifteen studies23–37 met the inclusion criteria for the diagnostic test accuracy review of a methacholine (MCH) challenge compared with a standardized exercise challenge test (ECT). All studies were conducted prospectively. Twelve used a cross sectional study design,23–25,27–32,34–36 three26,33,37 used a case-control design. One was published as a meeting abstract;24 fourteen were published in peer reviewed journals. The median year of publication was 1994 and ranged from 1973 to 2009. Studies were conducted in Europe,27,29–31,34,35,37 the United States,23,26,28 the Middle East,24,25 Asia,32,33 and South America.36

The characteristics of the studies are summarized in Table 9. The standardized ECT was performed on a treadmill23–28,30,32–34 or a bicycle ergometer.29,31,35–37 A positive test was defined at five different cutpoints using the maximum percent fall in FEV1 of 8 percent or greater,25 10 percent or greater,23,24,30,34,36,37 15 percent or greater,27,31 18 percent or greater,28 and 20 percent or greater.26,29,32,33,35 Most studies reported the target workload as between 80 to 90 percent of maximum predicted heart rate. In one study the target workload was described as “to exhaustion”34 and in another “target ventilation output of more than 60 percent of predicted maximum voluntary ventilation.”37 For two studies, the target workload was not reported.24,25

Table 9. Description of studies in the diagnostic test accuracy review: Methacholine vs. ECT.

Table 9

Description of studies in the diagnostic test accuracy review: Methacholine vs. ECT.

MCH challenges were based on the 2 minute tidal breathing method25,27,29,30,35 or the five breath dosimeter method.23,26,28,32–34,36,37 For the remaining studies, the protocol was either specific to the particular study group or was not clearly described.24,31

The baseline characteristics of the participants are presented in Table 10. The number of participants in each study ranged from 12 to 375 (median=52; IQR 28, 59). Six studies involved adults,26,28,31,33,34,37 seven involved children,24,27,29,30,32,35,36 and two included both adults and children.23,25 One study focused on elite female swimmers34 and one on endurance athletes.37 The remaining studies did not report on the level of fitness or sports participation. Most studies included participants who had a history of asthma.23–26,28,29,31–33,35,36 Generally, asthma was mild to moderate and stable. For the studies that used a case-control design,26,33,37 the control group comprised healthy controls who had no history of asthma.

Table 10. Description of participants in the diagnostic test accuracy review: Methacholine vs. ECT.

Table 10

Description of participants in the diagnostic test accuracy review: Methacholine vs. ECT.

Methodological Quality of Included Studies

Table 11 summarizes the methodological quality of the included studies. 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 results of the ECT to the results of the MCH challenge was not reported. Seven studies did not report their sources of funding.24,25,27–29,35,36 Four studies received funding support from industry,23,30–32 and six were supported by government and/or institution grants.26,30,31,33,34,37 The body of evidence is classified as “moderate.”

Table 11. Methodological quality of studies in the diagnostic test accuracy review: Methacholine vs. ECT.

Table 11

Methodological quality of studies in the diagnostic test accuracy review: Methacholine vs. ECT.

Quantitative Results

We analyzed the data using two concentration thresholds for MCH: PC20 MCH less than 8 mg/ml and PC20 MCH less than 16 mg/ml. Regardless of the threshold, there was considerable heterogeneity across the studies. Using the PC20 MCH less than 8 mg/ml threshold 331 (57 percent) tested positive for EIB/EIA on the ECT; 416 (71 percent) were positive on the MCH challenge. Sensitivity and specificity both ranged from 0 to 100 percent. (Figure 2)

Figure 2. Forest plot of sensitivity and specificity: Methacholine challenge vs. ECT (PC20 MCH less than 8 mg/ml; FEV1 fall index 10 percent or more for ECT).

Figure 2

Forest plot of sensitivity and specificity: Methacholine challenge vs. ECT (PC20 MCH less than 8 mg/ml; FEV1 fall index 10 percent or more for ECT).

Using the PC20 MCH less than 16 mg/ml threshold 420 (53 percent) tested positive for EIB/EIA on the ECT; 498 (63 percent) were positive on the MCH challenge. Heterogeneity in sensitivity estimates was reduced somewhat (range 55 to 100); specificity still ranged from 0 to 100 percent (Figure 3).

Figure 3. Forest plot of sensitivity and specificity: Methacholine challenge vs. ECT (PC20 MCH less than 16 mg/ml; FEV1 fall index 10 percent or more for ECT).

Figure 3

Forest plot of sensitivity and specificity: Methacholine challenge vs. ECT (PC20 MCH less than 16 mg/ml; FEV1 fall index 10 percent or more for ECT).

The area under the curve (AUC) of the ROC curve was marginally more than 0.5 using a 10 percent fall in FEV1 for a positive ECT and a PC20 MCH of less than 8 mg/ml (Figure 4) for a positive MCH. There was no improvement in the AUCs for any of the FEV1 cutpoints (10, 15 or 20 percent) or the MCH threshold (data not shown).

Figure 4. ROC curve plotting sensitivity vs. specificity: Methacholine challenge vs. ECT (PC20 MCH less than 8 mg/ml; FEV1 fall index 10 percent or more for ECT).

Figure 4

ROC curve plotting sensitivity vs. specificity: Methacholine challenge vs. ECT (PC20 MCH less than 8 mg/ml; FEV1 fall index 10 percent or more for ECT).

To explore heterogeneity, we performed a subgroup analysis on studies in which all participants had a diagnosis of asthma.24–26,28,29,32,33,35,36 This analysis (Figure 5) reduced the heterogeneity in sensitivity somewhat (range from 66 to 100); however, specificities ranged from 0 to 100 percent. Among this group of participants, 277 (70 percent) tested positive for EIA on the ECT; 350 (88 percent) were positive to the MCH challenge.

Figure 5. Forest plot of sensitivity and specificity: Methacholine challenge vs. ECT in studies in which all patients had asthma (PC20 MCH less than 8 mg/ml; FEV1 fall index 10 percent or more for ECT).

Figure 5

Forest plot of sensitivity and specificity: Methacholine challenge vs. ECT in studies in which all patients had asthma (PC20 MCH less than 8 mg/ml; FEV1 fall index 10 percent or more for ECT).

In the subgroup of four studies in which fewer than 50 percent of participants had asthma27,30,34,37 heterogeneity remained (Figure 6). Sensitivity ranged from 0 to 50 percent and specificity from 42 to 84 percent. Among this group of participants, 54 (29 percent) tested positive for EIB/EIA on the ECT; 66 (35 percent) were positive to the MCH challenge. Of note, the study by Pedersen et al.34 focused on elite swimmers; when this study was excluded, sensitivity ranged from 38 to 50 percent and specificity from 42 to 76 percent.

Figure 6. Forest plot of sensitivity and specificity: Methacholine challenge vs. ECT in studies in which less than 50 percent of patients had asthma (PC20 MCH less than 8mg/ml; FEV1 fall index 10 percent or more for ECT).

Figure 6

Forest plot of sensitivity and specificity: Methacholine challenge vs. ECT in studies in which less than 50 percent of patients had asthma (PC20 MCH less than 8mg/ml; FEV1 fall index 10 percent or more for ECT).

We also explored the following possible sources of heterogeneity: adults versus children, treadmill versus bicycle ergometer ECT, the five breath dosimeter versus 2 minute tidal breathing method for the MCH challenge, and cold/dry air versus other conditions during the ECT. There was no improvement in heterogeneity for any of these subgroups.

Key Question D–3: Sport or Venue Specific Exercise Challenges

Description of Included Studies

Five studies34,38–41 met the inclusion criteria for the diagnostic test accuracy review of sport or venue specific exercise challenges compared with a standardized exercise challenge test (ECT). Three studies focused on swimming challenges,34,38,40 one assessed short track speed skating and biathlon field challenges,39 and one assessed five cold weather sport challenges.41 Tables 12 and 13 summarize the study and participant characteristics.

Table 12. Description of studies in the diagnostic test accuracy review: Sport specific vs. ECT.

Table 12

Description of studies in the diagnostic test accuracy review: Sport specific vs. ECT.

Table 13. Description of participants in the diagnostic test accuracy review: Sport specific vs. ECT.

Table 13

Description of participants in the diagnostic test accuracy review: Sport specific vs. ECT.

Dickinson et al.39 included elite winter athletes who were members of Great Britain’s national teams for short track speed skating (n=10) and biathlon (n=4). The sport specific challenge for speed skaters involved indoor skating for 6 minutes at an 11 to 12 second pace per 250 meter lap. The sport specific challenge for biathletes was a 20 minute simulated race in Finland. The standardized ECT was performed on a treadmill for 8 minutes; the intensity level was more than 90 percent of maximum heart rate for the final 4 minutes. As part of this study, participants also completed an EVH challenge test (see Key Question D-4 for details). Two (14 percent) participants had a previous diagnosis of asthma and were taking asthma medication. Rundell et al.41 included 23 elite winter athletes who competed in a race or race simulation at their regular sport (biathlon, short-track speed skating, Nordic combined, cross country skiing, or kayaking) and demonstrated a drop in FEV1 of 10 percent or more. Seven of the participants had been diagnosed with EIA as children but took no medications at the time of the study.

Castricum et al.38 included 33 elite swimmers who were members of Australia’s state, national or international teams. The swim challenge required participants to swim for 8 minutes at more than 85 percent of their predicted maximum heart rate in an indoor 50 meter pool. The standardized ECT was a bicycle ergometer test; participants cycled at more than 90 percent of their predicted maximum heart rate. As part of this study, participants also completed an EVH challenge test (see Key Question D-4 for details). Thirteen (39 percent) participants had a previous diagnosis of asthma; of these, 11 were taking asthma medication.

In the study by Pedersen et al.34 16 elite (national or international caliber) female swimmers from Denmark underwent four different challenge tests: swimming, MCH, EVH, and a standardized ECT performed on a treadmill (see also Key Questions D-2 and D-4). The swim challenge was carried out under race conditions at the National Danish Swimming Championships. The distances were 200 meters (n=11), 400 meters (n=3) and 800 meters (n=2). For the treadmill test, the participants were instructed to run until exhaustion. None of the swimmers had a previous diagnosis of asthma.

Reggiani et al.40 assessed nine competitive swimmers from Italy; all had been involved in competitive swimming for at least 5 years. The swim challenge required participants to swim for 8 minutes at 80 percent of their maximum heart rate in an indoor 25 meter pool. The standardized ECT was a bicycle ergometer test; participants cycled at 80 percent of their maximum heart rate for 8 minutes. As part of this study, participants also completed a FRAST (see Key Question D-5 for details). All participants had a history of atopic diseases; they were asymptomatic at the time of the challenge tests.

Methodological Quality of the Included Studies

Table 14 summarizes the methodological quality of the included studies. There are several methodological issues. None had a representative patient spectrum which suggests the possibility of spectrum bias. Blinding of assessment of the ECT to the results of the index test was not reported. Two studies38,40 did not report their source of funding; one reported industry funding;39 one was supported by government grants,34 and one by the United States Olympic Committee.41 The body of evidence is classified as “low.”

Table 14. Methodological quality of studies in the diagnostic test accuracy review: Sport specific vs. ECT.

Table 14

Methodological quality of studies in the diagnostic test accuracy review: Sport specific vs. ECT.

Quantitative Results

In the study by Dickinson et al.39 three (21 percent) athletes had a positive test on the sport specific challenge using a FEV1 fall index of 10 percent or more. None of the participants had a positive test on the ECT; therefore, sensitivity could not be calculated. Specificity of the sport specific challenge was 79 percent (95 percent CI: 49, 95) indicating an 80 percent probability that for participants who do not have EIB/EIA, the sport challenge would correctly identify them as not having EIB/EIA. All of the athletes in the study by Rundell et al.41 were positive on a sport specific challenge. Five were also positive on an ECT at 95 percent of peak heart rate for 8 minutes. Sensitivity was 100 percent (95 percent CI: 48, 100). Overall, 5 (14 percent) of the cold weather athletes were positive for EIB/EIA on the ECT; 21 (57 percent) were positive for EIB/EIA on cold air challenges. None of the participants had a negative test on the sport specific challenge.

The sensitivity of the swimming challenge tests ranged from 040 to 50 percent34 (Figure 7). The specificity ranged from 83 to 100 percent. Overall, 11 (19 percent) of the swimmers were positive for EIB/EIA on the ECT; 5 (9 percent) were positive for EIB/EIA on the swimming challenge.

Figure 7. Forest plot of sensitivity and specificity: Sport specific challenges vs. ECT (FEV1 fall index 10 percent or more).

Figure 7

Forest plot of sensitivity and specificity: Sport specific challenges vs. ECT (FEV1 fall index 10 percent or more).

Key Question D–4: Eucapnic Voluntary Hyperpnea

Description of Included Studies

Seven studies34,38,39,42–45 met the inclusion criteria for the diagnostic test accuracy review of eucapnic voluntary hyperpnea (EVH) compared with a standardized ECT. All seven studies were conducted prospectively. Six 34,38,39,43–45 used a cross sectional design, and one42 used a case-control design with healthy controls. They were all published in peer reviewed journals. The median year of publication was 2005 and ranged from 1984 to 2008. Studies were conducted in Europe,34,39,45 Australia,38 and the United States.42–44

The characteristics of the studies are summarized in Table 15. The standardized ECT was performed on a treadmill34,39,43,45 or a bicycle ergometer.38,42,44 A positive test was defined as a FEV1 fall index of 10 percent or more,34,38,39,44 or 15 percent or more.45 Two studies did not report their definitions of a positive ECT.42,43 The target workload was reported as 90 percent of maximum heart rate,38,39 pulse rate of greater than 170 after 2 minutes of running (equivalent to 80 percent of predicted maximum heart rate),45 “until exhaustion,”34 “until a symptom limited maximum exercise capacity,”42 and “highest intensity.”44 One study43 did not report the criteria for target workload.

Table 15. Description of studies in the diagnostic test accuracy review: Eucapnic voluntary hyperpnea vs. ECT.

Table 15

Description of studies in the diagnostic test accuracy review: Eucapnic voluntary hyperpnea vs. ECT.

Most studies used five percent CO2 air content for the EVH challenge at room temperature. One study42 also tested participants with cold air. The duration of hyperpnea was either 5 minutes42 or 6 minutes.34,38,39,44,45 One study43 did not report duration. During the EVH challenge, the target ventilation rate was set at 80 percent maximum voluntary ventilation (MVV),42 85 percent MVV,38,44 a minimum ventilation of 30 times FEV1 equivalent to MVV,34 or 40 or more liters per minute.43 Two studies39,45 did not specify the rate of target ventilation.

The baseline characteristics of the participants are presented in Table 16. The number of participants in each study ranged from 10 to 33. Three studies included adults only,39,42,44 one included children only,45 two included both adults and children,34,38 and one43 did not report the age of participants. Four studies34,38,39,44 focused on elite athletes; three42,43,45 did not report the level of fitness or sports participation. Fourteen39 to 100 percent of participants43,45 had a diagnosis of asthma in the five studies that reported this. In three studies43–45 100 percent of participants had a history of EIB/EIA.

Table 16. Description of participants in the diagnostic test accuracy review: Eucapnic voluntary hyperpnea vs. ECT.

Table 16

Description of participants in the diagnostic test accuracy review: Eucapnic voluntary hyperpnea vs. ECT.

Methodological Quality of Included Studies

Table 17 summarizes the methodological quality of the included studies. As a group, there are methodological issues that limit interpretation and generalizability of the results. Of 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 results of the ECT to the results of the EVH challenge was not reported. Four studies did not report their sources of funding.38,42,43,45 Two studies received funding support from industry,39,44 and one was supported by government and institution grants.34 The body of evidence is classified as “low.”

Table 17. Methodological quality of studies in the diagnostic test accuracy review: Eucapnic voluntary hyperpnea vs. ECT.

Table 17

Methodological quality of studies in the diagnostic test accuracy review: Eucapnic voluntary hyperpnea vs. ECT.

Quantitative Results

A total of 138 participants were studied; 42 (30 percent) tested positive for EIB/EIA on the ECT and 74 (54 percent) were positive using EVH. Overall, the sensitivity and specificity of the seven studies were heterogeneous with values ranging from 25 to 90 percent for sensitivity and 0 to 71 percent for specificity (data not shown).

To explore the heterogeneity, we conducted a post hoc subgroup analysis of participants who did not have a history of EIB/EIA. This included all participants from three studies34,38,39 and one group of participants from a fourth study.42 Among these 83 individuals, 10 (12 percent) tested positive on the ECT while 44 (53 percent) tested positive using EVH. Heterogeneity was somewhat reduced and the range of sensitivity and specificity was narrowed—25 to 75 percent and 29 to 67 percent, respectively (Figure 8). Of note, the studies by Castricum et al.,38 Dickinson et al.,39 and Pedersen et al.34 focused on elite athletes (proportion with asthma 39, 14 and 0 percent, respectively).

Figure 8. Forest plot of sensitivity and specificity: Eucapnic voluntary hyperpnea vs. ECT (participants with no history of EIB/EIA; FEV1 fall index 10 percent or more for ECT and EVH).

Figure 8

Forest plot of sensitivity and specificity: Eucapnic voluntary hyperpnea vs. ECT (participants with no history of EIB/EIA; FEV1 fall index 10 percent or more for ECT and EVH).

For the three studies43–45 and one group42 in which all participants had a history of EIB/EIA the sensitivity of EVH was higher, ranging from 71 to 90 percent43 (Figure 9). Among these 55 people, 32 (58 percent) tested positive for EIB/EIA using the ECT and 30 (55 percent) tested positive using EVH.

Figure 9. Forest plot of sensitivity and specificity: Eucapnic voluntary hyperpnea vs. ECT (participants with a history of EIB/EIA; FEV1 fall index 10 percent or more for ECT and EVH).

Figure 9

Forest plot of sensitivity and specificity: Eucapnic voluntary hyperpnea vs. ECT (participants with a history of EIB/EIA; FEV1 fall index 10 percent or more for ECT and EVH).

Subgroup and sensitivity analyses. Sensitivity analyses were performed using different thresholds for a positive EVH challenge. Using a threshold of a FEV1 fall index of 15 percent or more for a positive EVH challenge, the sensitivity improved; however, there was still considerable heterogeneity (Figure 10). The results of the analyses using a threshold of 20 percent or more were not substantially different (data not shown). In a subgroup analysis on studies enrolling athletes the sensitivity ranged 25 to 88 percent; specificity ranged 0 to 67 percent (data not shown).

Figure 10. Forest plot of sensitivity and specificity: Eucapnic voluntary hyperpnea vs. ECT (all participants; threshold for a positive test is a FEV1 fall index 10 percent or more for ECT and 15 percent or more for EVH).

Figure 10

Forest plot of sensitivity and specificity: Eucapnic voluntary hyperpnea vs. ECT (all participants; threshold for a positive test is a FEV1 fall index 10 percent or more for ECT and 15 percent or more for EVH).

Cold air versus warm air. In the study by Eliasson et al.42 participants performed EVH tests while inspiring air at room temperature or inspiring air cooled to minus 18°C to minus 26°C. Despite the increased conditioning of the air required during the cold air trial, there was no difference in post-EVH FEV1 response between the cold versus room temperature challenges (data not shown).

Key Question D–5: Free Running Asthma Screening Test

Description of Included Studies

Three studies36,40,46 met the inclusion criteria for the diagnostic test accuracy review of a free running asthma screening test (FRAST) compared with a standardized exercise challenge test (ECT). All studies were conducted prospectively; two used a cross sectional design36,40 and one used a case-control design with healthy controls.46 The studies were published in peer reviewed journals in 1988,40 200536 and 2008.46 The studies were conducted in Europe40,46 and South America.36

The standardized ECT was performed on a treadmill46 and a bicycle ergometer.36,40 All free running tests took place indoors. Garcia de la Rubia et al.46 defined a positive test as a FEV1 fall index of 20 percent or more on the treadmill and a 16.5 percent or more for the FRAST. Among children with asthma the maximum heart rate ranged from 70 to 90 percent in treadmill running and 82 to 93 percent in free running. Results were not reported for the control group. Reggiani et al.40 defined a positive test as a FEV1 fall index of 20 percent or greater on the bicycle ergometer and the FRAST. The target workload was 80 percent of maximum heart rate for both tests. As part of the study the authors also assessed the diagnostic characteristics of a sport specific (swimming) challenge (see Key Question D-3 for details). Souza et al.36 defined a positive test as a FEV1 fall index of 10 percent or greater for both the ECT and FRAST. The target workload was between 80 and 90 percent of maximum heart rate for both tests. The characteristics of the studies are summarized in Table 18.

Table 18. Description of studies in the diagnostic test accuracy review: FRAST vs. ECT.

Table 18

Description of studies in the diagnostic test accuracy review: FRAST vs. ECT.

The baseline characteristics of the participants are presented in Table 19. Sixty people were enrolled in the study by Garcia de la Rubia et al.,46 30 had extrinsic asthma but were asymptomatic at testing. The 30 healthy controls had never presented with symptoms of asthma and had no history of allergy. The nine competitive swimmers who participated in the study by Reggiani et al.40 all had a history of atopy but were asymptomatic at the time of testing. In the study by Souza et al.36 the 30 participants analyzed had intermittent asthma defined as at least two episodes of dyspnea and/or wheezing relieved with the use of bronchodilators.

Table 19. Description of participants in the diagnostic test accuracy review: FRAST vs. ECT.

Table 19

Description of participants in the diagnostic test accuracy review: FRAST vs. ECT.

Methodological Quality of the Included Studies

Table 20 summarizes the methodological quality of the included studies. Overall there were concerns with 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 assessment of the ECT to the results of the FRAST was not reported. None of the studies reported their source of funding. The body of evidence is classified as “very low.”

Table 20. Methodological quality of studies in the diagnostic test accuracy review: FRAST vs. ECT.

Table 20

Methodological quality of studies in the diagnostic test accuracy review: FRAST vs. ECT.

Quantitative Results

For two studies36,40 the sensitivity and specificity could be calculated using a FEV1 fall index of 10 percent or more on both challenges. The sensitivity of FRAST was 60 and 67 percent; specificity was 47 and 67 percent (Figure 11). Overall, 13 (38 percent) participants had a positive ECT; 18 (53 percent) had a positive FRAST.

Figure 11. Forest plot of sensitivity and specificity: FRAST vs. ECT (FEV1 fall index ≥10%).

Figure 11

Forest plot of sensitivity and specificity: FRAST vs. ECT (FEV1 fall index ≥10%).

Garcia de la Rubia et al.46 used a FEV1 fall index of 20 percent or more as the threshold for a positive test for the ECT and 16.5 percent or more for the FRAST. Based on these thresholds the sensitivity was 53 percent (95 percent CI: 34, 72) and the specificity was 100 percent (94 percent CI: 88, 100). Thirty (50 percent) participants had a positive ECT; 16 (27 percent) had a positive FRAST.

Key Question D–6: Mannitol Challenge

Description of Included Studies

Three studies23,47,52 met the inclusion criteria for the diagnostic test accuracy review of Mannitol compared with a standardized exercise challenge test (ECT). All studies used a cross sectional design and all were conducted prospectively. The characteristics of the studies and the participants are summarized in Tables 21 and 22.

Table 21. Description of studies in the diagnostic test accuracy review: Mannitol vs. ECT.

Table 21

Description of studies in the diagnostic test accuracy review: Mannitol vs. ECT.

Table 22. Description of participants in the diagnostic test accuracy review: Mannitol vs. ECT.

Table 22

Description of participants in the diagnostic test accuracy review: Mannitol vs. ECT.

In the study by Anderson et al.23 the standardized ECT was performed on a treadmill. The target workload was between 80 and 90 percent of maximum heart rate. A positive test was defined as a FEV1 fall index of 10 percent or greater. A positive test for the mannitol challenge was defined as a FEV1 fall index of 15 percent or more or a fall of 10 percent or more between consecutive doses. As part of the study the authors also compared the ECT to a MCH challenge (See Key Question D-2). In total, there were 510 participants enrolled, 375 underwent all three challenge tests and were included in the final analyses. Both children (26 percent) and adults (74 percent) were included in the study. Participants had symptoms suggestive of asthma but none had a confirmed diagnosis. Symptoms suggested mild, persistent asthma (i.e., symptoms two or more times per week; asymptomatic between exacerbations; exacerbations of only a few hours to a few days; night time symptoms two or more times per month).109 Seventy-eight percent of participants were atopic.

Brannan et al.47 used a standardized bicycle ergometer challenge. The target workload was between 90 and 100 percent of maximum heart rate. A positive test was defined as a FEV1 fall index of 10 percent or greater. A positive test for the mannitol challenge was defined as a FEV1 fall index of 15 percent or more. As part of the study the authors also compared the ECT to an EVH challenge (See Key Question D-4). In total 36 adults were enrolled in the study; however, only 23 (64 percent) completed both the mannitol challenge and the ECT. All participants had atopic asthma, were taking asthma medications, and had a self-reported history of EIA.

In the study by Kersten et al.52 the ECT was performed on a treadmill. The target workload was approximately 90 percent of predicted maximum heart rate for 6 minutes. A positive test for both the ECT and the mannitol challenge was defined as a FEV1 fall index of 15 percent or greater. In total, 33 children were enrolled; 25 underwent both challenge tests and were included in the final analyses. All children had a history of allergic asthma and EIA.

Methodological Quality of Included Studies

Table 23 summarizes the methodological quality of the three studies. Of concern is the risk of spectrum bias in the studies by Brannan et al.47 and Kersten et al.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. There were no details about uninterpretable or indeterminate index test results in two of the three trials. All studies reported their sources of funding: Anderson et al.23 received funding support from industry; Brannan et al.47 and Kersten et al.52 were supported by government grants. The body of evidence is classified as “moderate.”

Table 23. Methodological quality of studies in the diagnostic test accuracy review: Mannitol vs. ECT.

Table 23

Methodological quality of studies in the diagnostic test accuracy review: Mannitol vs. ECT.

Quantitative Results

Using the threshold of a FEV1 fall index of 10 percent or more for the ECT, 23 (100 percent) participants in the Brannan et al.47 study were diagnosed with EIA; 22 (96 percent) were positive on the mannitol challenge. The sensitivity was 96 percent (95 percent CI: 78, 100). Specificity could not be calculated as none of the participants tested negative on the ECT (Figure 12).

Figure 12. Forest plot of sensitivity and specificity: Mannitol vs. ECT (FEV1 fall index ≥10% for ECT and ≥15% for mannitol).

Figure 12

Forest plot of sensitivity and specificity: Mannitol vs. ECT (FEV1 fall index ≥10% for ECT and ≥15% for mannitol).

Anderson et al.23 reported a sensitivity of 58 percent (95 percent CI: 50, 66) and a specificity of 65 percent (95 percent CI: 58, 72) (Figure 12). The ECT diagnosed 163 (44 percent) participants with EIA; 168 (45 percent) were positive on the mannitol challenge.

In the study by Kersten et al.52 sensitivity was 69 percent (95 percent CI: 41, 89) and specificity was 78 percent (95 percent CI: 40, 97) (Figure 12). The ECT diagnosed 16 (64 percent) participants with EIA; 13 (52 percent) were positive on the mannitol challenge. Three participants had to terminate the mannitol challenge due to persistent cough and were excluded from analysis.

Therapy Review

Literature Search

The search for controlled trials of the interventions targeted in this report identified 1,634 citations from electronic databases and eight by handsearching (Figure 13). Through screening of titles and abstracts, 490 references were selected for further examination. The full manuscripts of 13 citations could not be retrieved (Appendix D). Most of these were abstracts only available in supplementary conference proceedings. They were requested through interlibrary loan but did not arrive by the time this report was written. The search update identified an additional 19 citations; none met the screening criteria. We retrieved and evaluated the full-text of 477 potentially relevant articles. The application of the selection criteria resulted in 109 RCTs and CCTs being included; 368 studies were excluded.

Figure 13. Therapy for EIB/EIA: Flow diagram for study retrieval and selection.

Figure 13

Therapy for EIB/EIA: Flow diagram for study retrieval and selection.

During preliminary searches of the literature on mast cell stabilizing agents (MCS) three systematic reviews were identified. All three were published in the Cochrane Library76–78 and two were also published in journal format.104,161 Our search strategy included terms to identify new or additional trials that would update the existing reviews but none were identified. In consultation with AHRQ and the TEP, we made the decision to present a summary of these existing reviews (Appendix E).

The included articles addressed the following research questions: tachyphylaxis (n=9),53–55,57‐60,62,63 leukotriene receptor agonists (n=9),44,64–71 inhaled corticosteroids (n=4),72–75 MCS (n=62),76–78,104,161 anticholinergics (n=18),79–96 and a refractory period (warmup/cooldown; n=7).97–103

The main reasons for excluding studies were (1) the intervention did not address one of the research questions (n=214), (2) the study was not an RCT or CCT (n=60), (3) the study was not published in English (n=38), (4) the study population did not have a confirmed a diagnosis of EIB/EIA or included only children under the age of 6 (n=32). There were 24 studies excluded for other reasons. The list of excluded studies and reasons for exclusion are identified in Appendix D.

Key Question T–1: Tachyphylaxis to Short-Acting Beta- Agonists or Long-Acting Beta-Agonists

Description of Included Studies

Nine RCTs met the inclusion criteria for the review on whether people with EIB/EIA develop tachyphylaxis to long-acting beta-agonists (LABA)53–55,57–60 or to short-acting beta-agonists (SABA).62,63 Tachyphylaxis is defined as a waning or a diminishing response to prophylactic doses of a beta-agonist to attenuate or prevent EIB/EIA when these agents are also used on a continuous basis. In the case of continuous SABA use, tachyphylaxis is also defined as a diminished ability to cause bronchodilation when needed to reverse bronchoconstriction. Six trials used a crossover design55,57–59,62,63 and three were parallel trials.53,54,60 The LABA agents studied were formoterol,54 and salmeterol.53,55,57–60 The duration of LABA use was every other day for 3 weeks,58 daily use for 4 weeks,54,55,57,59 or daily use for 8 weeks.53,60

The studies were published in peer reviewed journals between 1994 and 2009. Five were conducted in North America,53,55,59,62,63 three in Europe,54,57,58 and one in 12 different countries.60 EIA was defined using different cutpoints of the maximum FEV1 fall index: 15 percent or greater,54,55,62,63 18 percent or greater,60 and 20 percent or greater.57 The exercise challenge test (ECT) was performed on a treadmill53,57,59,60 or bicycle ergometer.54,55,62,63 The characteristics of the trials are summarized in Tables 24a and 24b.

Table 24a. Description of trials in the therapy review: Tachyphylaxis to LABA.

Table 24a

Description of trials in the therapy review: Tachyphylaxis to LABA.

Table 24b. Description of trials in the therapy review: Tachyphylaxis to SABA.

Table 24b

Description of trials in the therapy review: Tachyphylaxis to SABA.

The baseline characteristics of the participants are presented in Tables 25a and 25b. The number of patients enrolled ranged from 9 to 248. Two studies58,59 included children only, two62,63 included adults only, and five53–55,57,60 included both adults and children. All participants had confirmed asthma; for most, their asthma was mild and stable.

Table 25a. Baseline characteristics of patients in trials in the therapy review: Tachyphylaxis to LABA.

Table 25a

Baseline characteristics of patients in trials in the therapy review: Tachyphylaxis to LABA.

Table 25b. Baseline characteristics of patients in trials in the therapy review: Tachyphylaxis to SABA.

Table 25b

Baseline characteristics of patients in trials in the therapy review: Tachyphylaxis to SABA.

Methodological Quality of Included Studies

Table 26 summarizes the methodological quality of included studies. Overall the quality of individual trials was high with a median Jadad score of 4 (IQR: 3, 4). Concealment of allocation was unclear in six trials54,55,57–60 and adequately reported in one.53 Two trials did not report their source of funding,57,58 three had industry support,53,59,60 three had government support,54,62,63 and one had a combination of institution, government and private industry support. 55

Table 26. Quality assessment of trials in the therapy review: Tachyphylaxis to LABA and SABA.

Table 26

Quality assessment of trials in the therapy review: Tachyphylaxis to LABA and SABA.

Quantitative Results: SABA

Two crossover trials that involved 1 week of regular (four times daily) SABA or placebo use prior to ECTs looked for evidence of tachyphylaxis to SABA.62,63 The SABA agent studied was albuterol (salbutamol). Both studies compared baseline FEV1 at the end of 1 week of use. Inman and O’Byrne63 assessed the prophylactic effect on EIA of an additional prophylactic dose of SABA. Hancox et al.62 did not address prophylaxis; however, did assess the bronchodilator effect of SABA following the ECT. Methodological differences precluded combining the data and the two studies are described separately.

Inman and O’Byrne63 randomized 10 adults with asthma and EIA (maximum percent fall FEV1 on pre-study ECTs ranged 15 to 45 percent) to 1 week of SABA 200μg or placebo four times per day. The washout period ranged from 7 to 21 days. Additional SABA use was allowed as needed throughout study weeks and the washout period. Baseline SABA use was less than 200μg per day and no other asthma medications had been taken in the previous month. No ECT was performed on day one of treatment and no baseline lung function measures were reported.

On day 8 of each treatment week, participants performed an ECT following a dose of placebo; on day 9, all performed an ECT following 200μg SABA (participants only were blinded to the prophylactic treatment taken).

Baseline FEV1 after 1 week of treatment. The mean (±SE) pre-challenge FEV1 on day 8 of the placebo week was 3.58±0.18L compared with 3.36±0.19L on day 8 of the SABA week; the pre-challenge FEV1 on day 9 of the placebo week was 3.58±0.17L compared with 3.34±0.20L after the SABA week. The mean difference of 230 ml between treatment weeks was statistically significant (p=0.02).

Post bronchodilator FEV1 .On day 9, all participants r eceived pre-challenge SABA 200μg. There was no significant difference between the placebo and SABA groups on the post-SABA FEV1 prior to the ECT (3.94 ±0.18L versus 3.83 ±0.18L; p=0.06). No comparison on the degree of bronchodilation that occurred was reported.

Effect on EIA. On day 8 after pre-treatment with placebo, both groups experienced EIA following the ECT. The mean maximum percent fall in FEV1 after the placebo week was 24.9±4.36 percent compared with a slightly greater fall of 29.4±4.71 percent after the SABA week. The difference of 4.5 percent was not statistically significant (p=0.12).

On day 9 following pre-challenge SABA, both groups experienced total protection from EIA. The mean maximum percent fall in FEV1 in the placebo group was 1.1±0.76 percent compared with a slightly greater fall of 5.1±2.03 percent in the SABA group. The difference of 4 percent was statistically significant (p=0.05). The post-exercise FEV1 measures were significantly lower at all time points over the recovery period after 1 week of continuous SABA use on both the day 8 placebo challenge (p=0.02) and the day 9 SABA challenge (p=0.01).

Other outcomes. Self-reported use of additional SABA for rescue during the placebo week was lower than during the SABA week (p>0.05).

The authors concluded that 1 week of regular SABA use resulted in a decrease in baseline FEV1, a decreased prophylactic effect on EIA, and more pronounced EIA.

Hancox et al.62 randomized eight, stable asthmatic women with EIA (maximum percent fall FEV1 of 15 percent or more) to 6 to 10 days of SABA 200μg or placebo four times per day. There was no washout period. Baseline SABA use was not reported. All bronchodilating drugs were withheld for 8 to 36 hours pre-ECT and no prophylactic treatment was given prior to the ECT at the end of the treatment week. SABA 100 μg, 100 μg, and 200 μg were given at 5, 10, and 15 minutes post-ECT following FEV1 measurements.

Baseline FEV1 after 1 week of treatment. There was no significant difference between the mean pre-challenge FEV1 after the SABA week compared with the placebo week (2.76L versus 2.80L).

Effect on EIA with no pre-treatment. At 5 minutes post-ECT the fall in FEV1 was greater in the SABA arm and the group as a whole experienced EIA. The mean minimum absolute FEV1 post-ECT after the SABA week was 2.28L (95 percent CI: 2.21, 2.35) compared with 2.55L (95 percent CI: 2.48, 2.62) after the placebo week (p=0.001). The percent fall in FEV1 was not reported and could not be calculated. The authors reported that the magnitude of the bronchodilator response to SABA in both arms was similar (no data); however, the FEV1 remained significantly lower in the SABA arm over the 25 minute recovery period despite administration of SABA at 5, 10, and 15 minutes.

The authors concluded that 1 week of regular SABA use results in more pronounced EIA and a suboptimal response to rescue medication.

Conclusions. The results from these two studies are equivocal on whether regular SABA use affects the baseline FEV1 over time. The studies are consistent in presenting data that indicates that regular SABA use may lead to a greater degree of EIA. It is important to note that pre‐treatment with SABA still offered complete protection to both groups in the one study.63 There is some evidence to indicate tachyphylaxis develops to the bronchodilating effect of SABA as rescue medication.62 The body of evidence is classified as “low.”

Quantitative Results: LABA

Five studies compared regular LABA use to placebo,54,55,58,59 four to salmeterol,55,57–59 and one to formoterol.54 Four studies reported regular LABA use for 4 weeks,54,55,57,59 and one58 used LABA treatment every other day for 21 days. Four studies were crossover trials55,57–59 and one used a parallel group design.54

Formoterol versus placebo. Garcia et al.54 randomized 19 stable asthmatics with EIA (maximum percent fall FEV1 of 15 percent or more) to 28 days of formoterol 12μg twice daily (n=10) or to a matching placebo (n=9) in a parallel group design. In the month prior to the study, SABA use was less than two doses per week and other asthma treatment remained unchanged. The primary outcome was the change in the bronchoprotection index (BI) defined as the reduction in the maximum percent fall in FEV1 after formoterol compared with no pre-treatment on the three test days.

On days 1, 14 and 28, participants refrained from taking all asthma medications and study drugs for 12 hours and performed two bicycle ECTs, 3 hours apart. The first took place with no prophylactic treatment; the second 30 minutes after taking formoterol.

Bronchoprotection index. The BI on day 1 was similar in the formoterol and placebo groups (71.3±32.6 percent versus 69.3±30.1 percent). During the study there was a significant decrease in the degree of protection offered by formoterol as the BI dropped to 52.3±31.7 percent on day 14 (p=0.012) and then to (27.8±32.6 percent) on day 28 (p=0.06). In the placebo group, the BI increased to 88.6±19.5 percent and 84.8±21.2 percent over the test days. The authors reported that these differences were not significant.

Complete protection (maximum percent fall FEV1 less than 10 percent). On study days 1, 14 and 28, a prophylactic dose of formoterol conferred complete protection to 9 of 10, 5 of 8, and 5 of 10 patients in the formoterol group, respectively (p=0.28). On days 1 and 14, a prophylactic dose of formoterol conferred complete protection to 8 of 9 patients, and on day 28 to 7 of 9 patients in the placebo group.

The authors concluded that twice daily formoterol over 4 weeks caused a significant reduction in bronchoprotection against EIA. Tachyphylaxis was evident by day 14 but did not progress. A single pre-exercise LABA dose in non-regular users (i.e., placebo group) maintained its efficacy.

Salmeterol versus placebo. Four crossover studies55,57–59 randomized groups to 3 or 4 weeks of salmeterol (50μg twice a day,55,57 once a day,59 or every other day for 3 weeks58) or to a matching placebo. Throughout the studies, the ECT was performed at 30 minutes,55 1 hour,59 6 hours,57 and 9 hours58 following administration of a study drug. The average fall in FEV1 after the first ECT on day 1 in the four studies was 3.7 percent (range 5.0 to 11.9) in the LABA arms compared with 26.9 percent (range 20.5 to 36.6) in the placebo arms. The pooled difference favored the salmeterol (MD = 25.1 percent [95 percent CI: 32.3, 18.0) (Figure 14). The pooled results are presented as the MD between salmeterol and placebo at day 1, week 2 or 3, and week 4. The decreasing MD between salmeterol and placebo from day 1 to week 4 reflects the decreasing effect of the LABA as a prophylactic EIA agent; over the same time period the placebo effect remained constant.

Figure 14. Effectiveness of LABA vs. placebo: Change from day 1 to week 3–4 weeks after the first ECT (MD in the maximum percent fall FEV1).

Figure 14

Effectiveness of LABA vs. placebo: Change from day 1 to week 3–4 weeks after the first ECT (MD in the maximum percent fall FEV1).

Maximum percent fall in FEV1. The mean maximum fall in FEV1 following an ECT after 2 and 4 weeks of continuous salmeterol use plus a pre-exercise dose was greater at week 4 than at day 1 (Figure 15). In contrast, the placebo arms showed a small decrease in the degree of EIA over the same time period. At the 4 week ECT, 30 minutes post-study drug, the average fall in FEV1 was 11.4 percent (range 4.0 to 24.0) in the LABA arms compared with 21.3 percent (range 16.0 to 32.9) in the placebo arms; however, the pooled difference still favored the LABA (MD = 10.5 percent; 95 percent CI: 14.7, 6.4) (Figure 14). The timing of the ECT following administration of the study drug (30 minutes to 9 hours) was not a factor. Onset of the bronchodilator effect of salmeterol occurs in 10 to 20 minutes and lasts for at least 12 hours.162

Figure 15. Effects over time of LABA and placebo on the maximum percent fall in FEV1 for the first ECT.

Figure 15

Effects over time of LABA and placebo on the maximum percent fall in FEV1 for the first ECT.

Duration of action (maximum percent fall in FEV1). Three studies55,57,59 investigated the duration of action of salmeterol. Participants underwent a second ECT on the same day; the ECT was given 9 to 12 hours after administration of the study drug. The pooled results (Figure 16) show that salmeterol continued to have a greater protective effect than placebo at 4 weeks (MD = 4.4 percent; 95 percent CI: 7.6, 1.2). However, the protective effect had decreased from day 1 when the absolute MD was 11.5 percent (95 percent CI: 14.3, 8.7).

Figure 16. Effectiveness of LABA vs. placebo: Change from day 1 after the second ECT (MD in the maximum percent fall FEV1).

Figure 16

Effectiveness of LABA vs. placebo: Change from day 1 after the second ECT (MD in the maximum percent fall FEV1).

Conclusions. Three to 4 weeks of regular LABA use results in a decreased prophylactic effect on EIA and more pronounced EIA. In studies that included a second ECT given 9 to 12 hours after the first ECT, the duration of the beneficial effects of LABA waned. The body of evidence is classified as “moderate.”

LABA compared to other agents. Two studies compared regular use of salmeterol 50μg, two puffs twice daily versus a leukotriene receptor antagonist (LTRA) (montelukast 10mg) once in the evening for 8 weeks. The studies included 360 participants with stable asthma and confirmed EIA.53,60 Both studies were funded by industry partners. The ECTs were performed near the end of the dosing interval for each drug (Table 24A). The pooled baseline fall in FEV1 in both groups in both studies (34.9 percent for LABA and 33.9 percent for LTRA) indicated that the participants experienced moderate to severe EIA. Both agents attenuated the EIA response after 3 days of treatment to a similar degree. The mean fall in the LABA group was reduced to 19.8 percent compared with 18.2 percent in the LTRA group; this difference was not significant (MD = 1.0 percent; 95 percent CI: −2.2, 4.2) (Figure 17). Both agents continued to be effective; however, montelukast maintained the same level of 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 the LABA group was 23.0 percent compared with 17.1 percent in the LTRA group; this difference was significant (MD = 5.4 percent; 95 percent CI: 2.2, 8.7). The studies also reported that a combined 52 percent of LABA participants compared with 65 percent of LTRA participants experienced falls in FEV1 of less than 20 percent at 8 weeks.

Figure 17. Effectiveness of LABA vs. LTRA: Change from day 1 to week 8 (MD in the maximum percent fall FEV1).

Figure 17

Effectiveness of LABA vs. LTRA: Change from day 1 to week 8 (MD in the maximum percent fall FEV1).

Although indirect evidence, these results lend support to other results that indicate some degree of tachyphylaxis develops early on with regular daily use of LABA. Furthermore, regular daily use of LABA lessens its ability to attenuate EIA when prophylactic doses are taken before exercise.

Key Question T–2: Leukotriene Receptor Antagonist Therapy

Description of Included Studies

Nine randomized crossover trials44,64–71 met the inclusion criteria for the review comparing a single dose of leukotriene receptor antagonist (LTRA) with no treatment or placebo to prevent a 10 percent or greater drop in FEV1 compared with no treatment or placebo. The studies were published between 1999 and 2007. Six trials were conducted in North America,44,64,65,67,68,71 two in Europe,66,69 and one70 involved multiple sites in North and South America.

The drugs studied were montelukast,44,64–66,68–71 zafirlukast,65,67 and zileuton.65 They were administered between 2 and 24 hours prior to an ECT. The thresholds for EIB/EIA measured by the FEV1 fall index were 10 percent or greater,44,66 15 percent or greater,64,65,69 and 20 percent or greater.67,68,70,71 The ECT was performed on a treadmill,64,66–71 or bicycle ergometer.44,65 The characteristics of the trials are summarized in Table 27.

Table 27. Description of trials in the therapy review: LTRA.

Table 27

Description of trials in the therapy review: LTRA.

The baseline characteristics of the participants are presented in Table 28. The number of patients enrolled ranged from 10 to 62 (total 267). Three studies64,67,69 included children only, four65,66,68,70 included adults only, and two44,71 included both children and adults. In eight trials44,64–71 all participants had a confirmed diagnosis of asthma; in the remaining trial44 36 percent of the participants had confirmed asthma. Participants had stable asthma with lung function greater than 85 percent predicted.

Table 28. Baseline characteristics of patients in trials in the therapy review: LTRA.

Table 28

Baseline characteristics of patients in trials in the therapy review: LTRA.

Methodological Quality of Included Studies

Table 29 summarizes the methodological quality of the included studies. Overall, the quality of the nine trials was good; the median Jadad score was 3 (IQR: 3, 5). Concealment of allocation was unclear in eight trials,44,64–69,71 and adequately reported in one.70 Two trials66,69 did not report their source of funding, five44,67,68,70,71 reported pharmaceutical industry support, and two64,65 had government/institutional support. The body of evidence is classified as “moderate.”

Table 29. Quality assessment of trials in the therapy review: LTRA.

Table 29

Quality assessment of trials in the therapy review: LTRA.

Quantitative Results

Pulmonary function measures. All nine trials recorded the maximum percent fall in FEV1 on an ECT performed 2 hours or less following drug administration. The times at which later ECTs were performed varied across studies thus data were categorized into two groups: 8 to 12 hours and 24 hours.

The mean fall in FEV1 on the placebo challenges ranged from 15.0 to 23.2 percent (average 18.5) indicative of mild 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). The absolute MD = 8.3 percent (95 percent CI: 6.9, 11.0) and represents clinical protection of approximately 41 percent over placebo. Substantial heterogeneity (I2=65 percent) was identified (Figure 18).

Figure 18. LTRA vs. placebo in pre-exercise treatment of EIB/EIA: Maximum decrease in FEV1.

Figure 18

LTRA vs. placebo in pre-exercise treatment of EIB/EIA: Maximum decrease in FEV1.

Six of the trials64,65,68–71 recorded the mean maximum percent fall in FEV1 on an ECT 8 to 12 hours after drug administration. The difference between groups ranged from 1.8 to 13.3 percent and all favored LTRA. The absolute MD = 6.8 percent (95 percent CI: 3.9, 9.6). Substantial heterogeneity (I2=70 percent) was identified (Figure 18).

The difference in the mean fall in FEV1 on an ECT 24 hours following drug administration was reported in three trials 68,70,71 and ranged from 4 to 8.6 percent in favor of LTRA. 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 indicate that LTRA remained significantly more effective than placebo in attenuating EIA (MD = 4.9 percent; 95 percent CI: 8.0, 1.8). Substantial heterogeneity (I2=76 percent) was identified (Figure 18).

The source of the heterogeneity could not be explained by subgroup analyses based on age, dose of study medication, or funding source. Other subgroup analyses (e.g., ICS use, atopic status) could not be performed due to lack of data.

Complete protection. Four trials66–69 reported the proportion of participants who received complete protection as an outcome measure. Using less than a 10 percent fall in FEV1 as the cutpoint, the pooled estimate indicated that there was no significant difference in the number who achieved complete protection with LTRA (RR = 1.9; 95 percent CI: 0.8 to 4.6) compared with placebo. Substantial heterogeneity (I2=72 percent) was identified (Figure 19).

Figure 19. LTRA vs. placebo in pre-exercise treatment of EIB/EIA: Achievemen t of complete protection (maximum percent fall FEV1 is less than 10 percent).

Figure 19

LTRA vs. placebo in pre-exercise treatment of EIB/EIA: Achievemen t of complete protection (maximum percent fall FEV1 is less than 10 percent).

Adverse effects. One study69 reported that there were no adverse effects using a 5 mg dose of montelukast. Using 10 mg of montelukast, three trials reported non-serious adverse effects that included mild headache,68,70,71 nausea,68 nervousness,68 arthralgia,71 and myalgia.71 One study67 reported dizziness, pharyngitis, fatigue and headache at a dose ranging from 5 to 40 mg of zafirlukast. The occurrence of side effects, however, was not markedly different between the treatment and placebo groups (Table 27).

Key Question T–3: Inhaled Corticosteroid Therapy

Description of Included Studies

Four randomized crossover trials72–75 met the inclusion criteria for the review on the prophylactic use of a single dose of inhaled corticosteroid therapy (ICS) prior to an ECT. The studies were published in peer reviewed journals between 1974 and 2001 and all were conducted in Europe. The definitions of EIA as measured by the FEV1 percent fall index were 10 percent or greater,73 15 percent or greater,74,75 and 20 percent or greater.72 The ECT was performed on a treadmill,73,74 bicycle ergometer,75 or by having participants run or exercise hard.72 The characteristics of the trials are summarized in Table 30.

Table 30. Description of trials in the therapy review: Inhaled corticosteroids.

Table 30

Description of trials in the therapy review: Inhaled corticosteroids.

The drugs studied were betamethasone valerate,72,73 budesonide,75 and fluticasone propionate.74 The number of patients enrolled ranged from 9 to 20 (total=50). Two studies73,74 included children only, one included adults only,75 and one included both adults and children.72 All participants had confirmed asthma; for most, their asthma was mild to moderate and stable. The baseline characteristics of the participants are presented in Table 31.

Table 31. Baseline characteristics of patients in trials in the therapy review: Inhaled corticosteroids.

Table 31

Baseline characteristics of patients in trials in the therapy review: Inhaled corticosteroids.

Methodological Quality of Included Studies

Table 32 summarizes the methodological quality of the included studies. Overall, the quality of the four trials was good with a median Jadad score of 3.5 (IQR: 3, 4). Concealment of allocation was unclear in three trials72,74,75 and inadequately reported in one.73 One trial did not report their source of funding,73 two had private industry support,72,74 and one was supported by a government grant.75 The body of evidence is classified as “moderate.”

Table 32. Quality assessment of trials in the therapy review: Inhaled corticosteroids.

Table 32

Quality assessment of trials in the therapy review: Inhaled corticosteroids.

Quantitative Results

Pulmonary function measures. Two trials72,74 reported the maximum percent fall in FEV1 and two73,75 reported the maximum percent fall in peak expiratory flow (PEF) post-ECT. The mean fall in FEV1/PEF on the placebo challenges ranged from 19.2 to 36.2 percent (average 30.5). Three of the placebo groups72,73,75 had a mean fall in FEV1/PEF greater than 30 percent indicating moderate to severe EIA. The average fall in FEV1/PEF in the ICS arms ranged from 9.7 to 31.9 percent (average 25.5). The pooled MD was 5.0 percent (95 percent CI: 0.0, 9.9). The differences were not statistically significant and the results failed to identify important heterogeneity (I2=0) (Figure 20).

Figure 20. Inhaled corticosteroids vs. placebo in pre-exercise treatment of EIA: MD in maximum percent fall in FEV1 or peak expiratory flow.

Figure 20

Inhaled corticosteroids vs. placebo in pre-exercise treatment of EIA: MD in maximum percent fall in FEV1 or peak expiratory flow.

Clinical protection. The mean clinical protection over placebo for participants was either reported as an outcome measure or we were able to calculate it from the data. The protection index in children was 49 percent after a high dose of fluticasone,74 and 19.6 percent after betamethasone.73 A combined population of children and adults who received betamethasone72 did better on the placebo (clinical protection was –5.9 percent). The clinical protection index for adults who received budesonide75 was 27.5 percent.

Complete protection. Thio et al74 provided IPD from which we could determine that 6 of 9 participants using high dose fluticasone obtained complete protection (maximum percent fall FEV1 less than 10 percent) compared with 4 of 9 on placebo (RR=1.5; 95 percent CI: 0.6, 3.6). This outcome was not reported in any of the other trials.

Adverse effects. None of the studies reported on adverse events.

Conclusions. We cannot conclude that ICS taken prior to exercise provides a clinical benefit to people with stable asthma who experience EIA.

Key Question T–4: Mast Cell Stabilizers

Background

With better understanding of the possible role of inflammatory mediators and mast cell degranulation in the pathogenesis of EIB/EIA, a variety of mast cell degranulation drugs have been investigated for their potential to protect against bronchoconstriction caused by exercise. Sodium cromoglycate (SCG; also referred to as "cromolyn sodium") and nedocromil sodium (NCS) are mast cell stabilizer (MCS) agents introduced in the late 1960s and early 1980s, respectively. Both are reportedly effective on a single-use basis for inhibiting bronchoconstriction due to antigens, fog, cold air, sulphur dioxide, and exercise. If effective and safe in EIB/EIA, these properties would make them attractive therapeutic options for active people.

During preliminary searches of the literature on MCS, three systematic reviews were identified. All three were published in the Cochrane Library and two were subsequently published in peer-reviewed journals. The reviews addressed the following topics: NCS to prevent EIA,76,104 NCS versus SCG to prevent EIA,77 and MCS versus an anticholinergic agent, a SABA agent, or a combination of MCS and SABA agents to prevent EIA.78 Our search strategy included terms to locate any new or additional RCTs that would add to the existing reviews. None were identified (Appendix E).

SCG is no longer available in a dry-powdered inhaler in North America; the only remaining formulations for asthma are solutions for nebulization. The evidence related to these drugs in EIB/EIA are summarized from the existing reviews.

NCS versus Placebo Review

Design issues. All the studies included in this review were randomized crossover trials.76,104 The ECTs were conducted on separate days and the researchers adhered to the recommended washout time of 5 to 10 times the half-life of a drug between challenges.

Objective. The objective of the review was to compare the effects of a prophylactic dose of inhaled NCS with placebo in persons with confirmed EIA. The primary outcome was a measure of the change in the percent fall index pre- or post-exercise. Secondary outcomes were the mean percent fall in FEV1 at varying time points up to 1 hour post-ECT and the degree of clinical protection afforded by NCS over placebo. A mean protection index of 50 percent or more was considered clinically significant. The review also investigated subgroups based on dose, delivery device, timing of pre-treatment, severity of EIA, and age.

Populations. Recruitment procedures were not often described but generally seemed to be volunteers from asthma clinics or schools for asthmatic children. All had documented asthma and demonstrated reproducible EIA. All were described as having stable asthma at testing, were otherwise healthy, non pregnant, and had no respiratory comorbidity.

Exercise challenge tests. All ECTs were conducted indoors on an inclined treadmill (n=18), a bicycle ergometer (n=1), FRAST (n=1), or sport specific (n=1) (3 sprinters ran outdoors). Room temperatures ranged from 17 to 24 degrees Celsius and relative humidity ranged from 35 to 60 percent. One trial163 conducted a trial using cold air (−18 degrees Celsius). Regardless of the challenge format, the intensity and duration demanded was sufficient to induce EIA.

Results. Description of studies. There were 21 RCTs included in the review (20 peer reviewed articles and one unpublished thesis). The trials all used the crossover design but did not report data by period and only one reported the treatment sequence. The reviewers analyzed the data as though from parallel studies determining that this approach would provide a more conservative estimate of treatment effect. Data on continuous measures were pooled and reported as MD with 95 percent CI using a random effects model. In the subgroup analyses examining the effect of treatment based on baseline severity, mild EIA was defined a priori as a mean maximum percent fall in FEV1/PEF less than 30 percent on the placebo challenge. Moderate to severe EIA was defined as a mean maximum percent fall greater than 30 percent.

Methodological quality. Using the Jadad scale for RCTs, two studies were rated “strong” (5 points), nine “very good” (4 points), with the remaining 10 studies rated as “good” (3 points). All studies were double-blinded. Two studies were rated as “adequate” concealment of allocation; the remaining studies were rated “unclear.” The body of evidence is classified as “moderate.”

Pulmonary function results. Seventeen trials measured the FEV1 to determine the response to treatment following an ECT (total sample size=240; 11 children and 6 adult studies); seven measured the PEF (total sample size=115; 4 children and 3 adult studies). Three trials assessed the duration of effect over two or three challenges given 2 to 4 hours apart on the same day.

The maximum percent fall in either FEV1 or PEF improved significantly following NCS indicating a statistically and clinically significant attenuation of the EIA response over placebo (MD =15.6 percent; 95 percent CI: 13.2, 18.1) (Table 33). Subgroup analyses indicated no significant differences: age (children less than 18 years; n=11), dose (low less than 4 mg, n=3; standard 4 mg, n=16; high more than 4mg, n=1), delivery system (metered dose inhaler [MDI], n=12 versus MDI with spacer device, n=5), and timing of pre-treatment (less than 30 minutes, n=10 versus 30 minutes or more, n=7). The effect of NCS appeared more pronounced in those that experienced moderate to severe EIA (i.e., a fall index 30 percent or more). In this the MD increased to 21.36 percent (95 percent CI: 25.52, 17.20).

Table 33. Description of trials included in the Cochrane reviews of mast cell stabilizers.

Table 33

Description of trials included in the Cochrane reviews of mast cell stabilizers.

Other outcomes. Thirteen studies compared lung function up to 30 minutes post-challenge. At each of seven time points pre-treatment with NCS resulted in significant improvements in FEV1 and a return to within 10 percent of baseline function occurred in 10 minutes compared with 30 minutes with placebo.

In three trials that examined the duration of action on subsequent ECTs within 4.5 hours on the same day with no further medication, NCS appeared to offer benefit but of a lesser magnitude (MD=5.7 percent; 95 percent CI: 2.8, 14.2).

The trials provided no data on symptom scores, performance measures or participant preference. Subgroup comparisons by sex and corticosteroid use were not performed due to lack of data. Sensitivity analyses based on study quality, fixed versus random models, and publication status were all nonsignificant.

Adverse events. Twelve studies mentioned side effects. Seven stated that no adverse effects or symptoms attributable to NCS were observed. Minor effects reported were bad taste, throat irritation, and cough, and in one study a clinically insignificant increase in heart rate of four beats per minute.

Conclusions. In people with stable asthma that experience EIA, NCS taken 15 to 60 minutes prior to exercise reduces the severity and duration of EIA in both children and adults. The benefit appears more pronounced in those with more severe EIA.

NCS versus SCG Review

Design issues. Only crossover trials were included in this revew.77,161 To ensure drug clearance, the challenges within each study were conducted with a minimum 24 hour break.

Objective. The objective of the review was to compare the effects on post-exercise lung function following prophylactic doses of NCS and SCG by MDI, in persons with confirmed EIA. The primary outcome was the maximum percent fall index. The review also reported the odds ratio (OR) and 95 percent CIs of not obtaining complete protection and clinical protection over placebo. Complete protection was not obtained if the post-exercise fall in FEV1 exceeded either the 10 or 15 percent diagnostic cutpoints. Clinical protection was not obtained if the post‐exercise fall in FEV1 after NCS or SCG was less than 50 percent of the drop following placebo.

Population. Recruitment procedures were not well described but appeared to be healthy volunteers from either asthma clinics or schools for asthmatic children. Five studies involved children (6.5 to 15 years), three involved adults (17 to 38 years), and one study enrolled a mixed population (13 to 30 years). The majority recruited stable asthmatics with a reproducible post‐exercise fall in FEV1 of either 15 percent or 20 percent.

Exercise challenge tests. All ECTs were performed indoors on either an inclined treadmill (n=8) or a bicycle ergometer (n=1) for a duration of 6 to 8 minutes. Intensity required was reported as either a heart rate of 170 to 180 or 85 percent of predicted maximum for age. ECTs were conducted in controlled environments; however, room temperature and humidity were not reported. In all studies, the ECTs were conducted at the same time of day on different days.

Results. Description of studies. The review included nine RCTs (eight peer-reviewed articles and one conference abstract). As in the previous review, the data were analyzed as though they were from parallel studies. Data on continuous measures were pooled and reported as a MD with 95 percent CI, using a random effects model. Dichotomous variables for individual and pooled statistics were calculated as OR with 95 percent CI, using a random effects model. We have converted the OR of not obtaining complete or clinical protection to the RR of obtaining complete or clinical protection to allow comparisons to the key questions from the current review.

Subgroup analyses were performed on adults (18 years and older) versus children (under 18 years), baseline asthma severity (mild versus moderate-severe), different doses of NCS and SCG, and the timing of ECT post-treatment.

Methodological quality. Using the Jadad scale one study was rated “strong” (5 points), one “very good” (4 points), with the remaining studies rated as “good” (3 points). All studies except one were double-blinded. One study was rated as “adequate” concealment of allocation; the remaining studies were rated “unclear.” The body of evidence is classified as “moderate.”

Pulmonary function results. Seven trials reported the maximum percent fall in FEV1 (sample size=97; 5 children, 2 adult studies). No significant differences between NCS and SCG were identified (MD=−0.88; CI: −4.50, 2.74; I2=0 percent). Subgroup analyses based on dose (NCS 4 mg versus SCG 10 mg, n=5; NCS 4 mg versus SCG 20 mg, n=1; NCS 8 mg versus SCG 4 mg, n=1) and timing of the ECT (30 minutes or less, n=7; 120 to 140 minutes, n=2; 240 minutes, n=1) failed to identify significant differences (Table 33).

Other outcomes. Six trials reported complete protection using a fall in FEV1 of less than 10 percent and six used less than 15 percent; six also reported clinical protection over placebo. There were no significant differences between NCS and SCG by diagnostic cutpoint in providing complete or clinical protection (Table 33). The trials provided no data on symptom scores, performance measures or participant preference.

Adverse events. Three trials reported on treatment side effects (unpleasant taste N=3; sore throat N=2). No significant differences were found between NCS and SCG with respect to unpleasant taste (RR=3.77; 95 percent CI: 1.29, 10.97) or sore throat (RR=4.96; 95 percent CI: 20.1, 9.1). Overall, the trend in both cases was for fewer side effects in the SCG group.

Conclusions. In people with stable asthma who experience EIA, both NCS and SCG provide a significant protective effect against EIA. No significant differences were evident between the two drugs on pulmonary function; however, more people experienced a sore throat or an unpleasant taste from NCS.

MCS to Prevent EIA Review

Design issues. This review78 was restricted to full manuscripts regardless of publication status and three foreign language studies were included. All but two studies used a crossover design.

Objective. The objective of the review was to compare the effects of a prophylactic dose of either NCS or SCG (collectively called MCS) to that of atropine, ipratropium or oxitropium bromide (collectively called anticholinergics), and SABA. The review also compared the effects of a combination of SABA plus MCS to a SABA alone. The primary outcome was the percent fall in FEV1 post-exercise. The review also reported the OR and 95 percent CIs of obtaining complete protection and clinical protection. Complete protection was obtained if the post‐exercise fall in FEV1 was less than either the 10 or 15 percent diagnostic cutpoints. Clinical protection was obtained if the post-exercise fall in FEV1 after drug intervention was at least 50 percent less than that with placebo.

Population. Recruitment procedures were not well described but appeared to be healthy volunteers from either asthma clinics or schools for asthmatic children. Thirteen trials recruited children (6 to 13 years) and 11 recruited adults (18 years and older). Most studies documented recruiting healthy, stable, asthmatics and all of the trials required at least a 15 percent fall in FEV1 to diagnose EIA. The populations in seven trials were considered to have mild EIA defined as a percent fall index less than 30 percent on placebo; 16 were classified moderate to severe; one did not report baseline data.

Exercise challenge tests. The ECTs involved an inclined treadmill (n=19), FRAST (n=1), bicycle ergometer (n=3), and stair climbing (n=1). Most studies reported an exercise challenge time of between 5 to 8 minutes with a target heart rate ranging from 150 to 180 beats per minute.

Results. Description of studies. The 24 RCTs included in the review reported data on 518 participants. The studies compared the following drug combinations: MCS versus anticholinergics (8 studies), MCS versus SABA (20 studies), and SABA versus a combination of SCG plus SABA (10 studies). The method of delivery varied across studies (8 nebulizer, 5 MDI, 3 spinhaler, and 2 not described), with six studies comparing two or more devices. All studies required participants to abstain from taking any bronchodilator for at least 8 hours prior to a challenge; other medications were withheld for longer periods of time.

Methodological quality. Using the Jadad score, two studies were rated “strong” (5 points), five rated “very good” (4 points), nine rated “good” (3 points), and eight rated “poor” (2 points or less). Four studies were rated to have “adequate” concealment of allocation; the rest were rated “unclear.” The body of evidence is classified as “moderate.”

Pulmonary function results. Sixteen studies (9 children, 7 adult) reported the response to treatment using different FEV1 measures (maximum percent fall FEV1=14 studies; percent predicted=1 study; change=1 study); 8 studies (5 children, 3 adult) reported maximum percent fall PEF.

MCS versus anticholinergics: MCS provided a modest but significant advantage in post‐exercise lung function over anticholinergics. Pooled results indicate MCS attenuated the fall in FEV1 to 7.1 percent compared with 13.8 percent on anticholinergics (MD=6.7 percent; 95 percent CI: 3.3, 10.0; I2=0 percent). MCS provided more individuals with complete protection (OR=2.2; 95 percent CI: 1.3, 3.7) and clinical protection (OR=2.7; 95 percent CI: 1.1, 6.4). There were no subgroup differences based on age, severity, or study quality. No adverse effects were reported for either agent group.

MCS versus SABA: When compared with SABA, MCS were not as effective at preventing EIA. The mean maximum 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=20 percent).). MCS provided fewer individuals with complete protection (OR=0.3; 95 percent CI: 0.2, 0.5) or clinical protection (OR=0.4; 95 percent CI: 0.2, 0.8). There were no significant subgroup differences based on age, severity, drug, delivery, or study quality.

SABA versus combination of SCG plus SABA: Combining a MCS with a SABA did not indicate significant advantages to pulmonary function over SABA alone. On average the maximum fall in FEV1 on SABA alone was reduced to 5.3 percent compared with 3.5 percent using a combination (MD=1.3 percent; 95 percent CI: −6.3, 8.9). There were no significant differences in the number achieving complete protection (OR=0.5; 95 percent CI: 0.2, 1.4) or clinical protection (OR=0.4; 95 percent CI: 0.1 to 1.2). There were no subgroup differences. Overall, every comparison favored the combination; however, none reached statistical significance.

Other outcomes. One study measured the work effort involved during exercise. This analysis showed that SCG and SABA significantly decreased the energy cost of running, ventilation, oxygen consumption, and tidal volume which significantly increased running duration. The trials provided no data on symptom scores, performance measures or participant preference.

Adverse events. A nonsignificant difference in side effects was demonstrated with 11 percent of SABA patients experiencing side effects compared with 3 percent of those receiving MCS (OR=0.2; 95 percent CI: 0.0, 8.2). No adverse events were reported for MCS versus anticholinergics or SABA versus combinations of MCS and SABA.

Conclusions. In people with stable asthma who experience EIA, SABA, MCS, and anticholinergic agents can provide a significant protective effect against EIA with few adverse effects. On average, SABA agents were more effective than MCS, and MCS agents were more effective than anticholinergics. The combination of SABA and MCS agents did not provide significant advantages over a SABA alone; however, this approach may be appropriate for some people.

Table 34Baseline characteristics of patients in trials included in the Cochrane reviews of mast cell stabilizers

Author YearAge yr (mean±SD)
Males: N (%)
Asthma statusPulmonary function (% predicted FEV1 / mean±SD)Max % fall FEV1 (mean±SD)Smoking status: N (%)Atopic status: N studiesICS history N (%)
Spooner76 2002C: 11.3±2.2
A: 26.5±5.0
M ~ 64%
Not all reported male/female mix
All stable at time of ECTFEV1 >70% predicted; variability <10–15% between challengesMean range of EIA severity on control or placebo ECT
C: 28.7 to 50.0%
A: 15.00 to 40.6%
Mean fall <30% N=16
Mean fall ≥30% N=5
Nonsmokers N=3; NR N=17All atopic N=12; Mixed/NR N=8None on ICS N=7; mixed/NR=2 3
Kelly77 2002C:11.85±4.4;
A: 24±4.2
M ~ 54%
Not all reported male/female mix
All stable at time of ECTFEV1 >70% predicted; variability <10–15% between challengesMean range of EIA severity on control or placebo ECT
C: 27.4 to 35.9%
A: 15.00 to 33.2%
Mean fall <30% N=4
Mean fall ≥30% N=4
NR N=1
Nonsmokers N=1; NR N=8All atopic N=3; mixed/NR N=8None N=4; mixed/NR N=5
Spooner78 2003C:11.49±3.99
A: 26.98±11.15
M ~ 48%
Not all reported male/female mix
All stable at time of ECTFEV1 >70% predicted; variability <10–15% between challengesMean range of EIA severity on control or placebo ECT
C: 14.3 to 45.2%
A: 15 to 47%
Mean fall <30% N=7
Mean fall ≥30% N=16
NR N=1
Nonsmokers N=4; 3/18 smokers included N=1 ; NR N=192/49 atopic N=1; all atopic N=1; NR N=22Unclear N=6; mixed N=4; steroid use an exclusion criteria N=9

A = adult; C = children; ECT = exercise challenge test; EIA = exercise-induced asthma; FEV1 = forced expiratory volume in 1 second; ICS = inhaled corticosteroids; M = males; NR = not reported; SD = standard deviation

Key Question T–5: Anticholinergic Agents

Description of Included Studies

Eighteen randomized crossover trials79–96 met the inclusion criteria for the review of short-acting anticholinergic (SAAC) therapy. The studies were published between 1976 and 1989. All studies were conducted either in Europe79,82,85–90,92–94 or North America.80,81,83,84,91,95,96

The drugs studied were atropine,84,86,91 ipratropium bromide (IB),79–83,85,87,88,90,93–96 and oxitropium bromide (OB).88,89,92 The definitions of EIA as measured by the percent fall in FEV1 or PEF were 10 percent or greater,81,86,96 15 percent or greater,80,83,87–89,91,95 and 20 percent or greater.79,85,92,93 The ECT was performed on a treadmill,79,83–86,89,90,92–94,96 bicycle ergometer,80,81,87,88,91,95 and free running (FRAST).82 The characteristics of the trials are summarized in Table 35.

Table 35. Description of trials in the therapy review: Anticholinergic therapy.

Table 35

Description of trials in the therapy review: Anticholinergic therapy.

The baseline characteristics of the participants are presented in Table 36. The number of patients enrolled ranged from 6 to 20. Seven studies79,80,84,86,89,91,96 included only children, eight included only adults, 81–83,87,88,90,94,95 and three included both.85,92,93 All participants in these trials had a confirmed asthma status; for most, the asthma was mild and stable.

Table 36. Baseline characteristics of patients in trials in the therapy review: Anticholinergic therapy.

Table 36

Baseline characteristics of patients in trials in the therapy review: Anticholinergic therapy.

Methodological Quality of Included Studies

Table 37 summarizes the methodological quality of the included studies. Overall the quality of the 18 trials was low with a median Jadad score of 2.5 (IQR: 2, 3) (Table 37). Concealment of allocation was unclear in all the trials. Seven trials did not report their source of funding,79,82,85,86,91,92,94 eight received private industry support,80,83,87–90,93,95 and three were supported by government grants.81,84,96 The body of evidence is classified as “moderate.”

Table 37. Quality assessment of trials in the therapy review: Anticholinergic therapy.

Table 37

Quality assessment of trials in the therapy review: Anticholinergic therapy.

Quantitative Results

Pulmonary function measures. Eleven trials compared IB to placebo79,80,83,85,87,88,90,93–96 and reported the maximum percent fall in FEV1 as the outcome measure; one82 reported the maximum percent fall in peak expiratory flow (PEF). 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 they experienced moderate to severe EIA. The average fall in FEV1/PEF in the SAAC arms ranged from 10 to 33 percent (average 21). The pooled difference favored IB over placebo (MD = 9.8 percent; 95 percent CI: 5.0, 14.6) was statistically significant (Figure 21); however, substantial heterogeneity (I2=76 percent) was identified.

Figure 21. Ipratropium vs. placebo in pre-exercise treatment of EIA: Mean difference in maximum percent fall in FEV1 or peak expiratory flow.

Figure 21

Ipratropium vs. placebo in pre-exercise treatment of EIA: Mean difference in maximum percent fall in FEV1 or peak expiratory flow.

Four trials compared atropine to placebo80,84,86,91 and all reported the maximum percent fall in FEV1 as the outcome measure. The mean difference in the percent fall in FEV1 ranged from 11.9 to 22.7 percent. The pooled estimate indicated that atropine was significantly more effective than placebo in attenuating EIA (MD=16.0; 95 percent CI: 10.2, 21.7) (Figure 22). Negligible heterogeneity was identified (I2=0 percent).

Figure 22. Atropine vs. placebo in pre-exercise treatment of EIA: Mean difference in maximum percent fall in FEV1 or peak expiratory flow.

Figure 22

Atropine vs. placebo in pre-exercise treatment of EIA: Mean difference in maximum percent fall in FEV1 or peak expiratory flow.

Two trials compared OB to placebo88,89 and both reported the maximum percent fall in FEV1 as the outcome measure. The pooled results showed that OB was significantly more effective than placebo in attenuating EIA (MD=13.8; 95 percent CI: 6.0, 21.6) (Figure 23). Negligible heterogeneity was identified (I2=0 percent).

Figure 23. Oxitropium vs. placebo in pre-exercise treatment of EIA: Mean difference in maximum percent fall in FEV1 or peak expiratory flow.

Figure 23

Oxitropium vs. placebo in pre-exercise treatment of EIA: Mean difference in maximum percent fall in FEV1 or peak expiratory flow.

Complete protection. Seven trials involving IB reported complete protection as an outcome measure.79,83,85,87,88,94,96 The proportion of patients who achieved complete protection ranged from 0 to 13 percent. The pooled estimate showed that patients are 4.5 times more likely to achieve complete protection with IB compared with placebo (RR=4.5; 95 percent CI: 1.9, 10.9) (Figure 24). Negligible heterogeneity was identified (I2=0 percent).

Figure 24. Ipratropium vs. placebo in pre-exercise treatment of EIA: Achievement of complete protection (maximum fall FEV1 less than 10 percent).

Figure 24

Ipratropium vs. placebo in pre-exercise treatment of EIA: Achievement of complete protection (maximum fall FEV1 less than 10 percent).

Godfrey et al.86 found no significant difference in achieving complete protection in patients who received atropine versus placebo (RR=5.0; 95 percent CI: 0.3, 88.5). The remaining studies on atropine did not report this outcome.

The pooled estimate of RR for complete protection from the three studies88,89,92 comparing OB with placebo was nonsignificant (RR=4.7; 95 percent CI: 0.4, 60.3) (Figure 25). There was substantial heterogeneity (I2=67 percent).

Figure 25. Oxitropium bromide vs. placebo in pre-exercise treatment of EIA: Achievement of complete protection (maximum fall FEV1 is less than 10 percent).

Figure 25

Oxitropium bromide vs. placebo in pre-exercise treatment of EIA: Achievement of complete protection (maximum fall FEV1 is less than 10 percent).

Clinical protection. Six trials that examined IB reported clinical protection as an outcome measure.79,81,83,85,87,94 The pooled estimate showed that 60 percent (95 percent CI: 48, 72) of patients achieved clinical protection when defined as at least a 50 percent improvement in FEV1 over placebo (Figure 26). There was negligible heterogeneity (I2=0 percent).

Figure 26. Ipratropium vs. placebo in pre-exercise treatment of EIA: Achievement of clinical protection (50 percent or greater improvement over placebo in FEV1).

Figure 26

Ipratropium vs. placebo in pre-exercise treatment of EIA: Achievement of clinical protection (50 percent or greater improvement over placebo in FEV1).

In the study by Godfrey et al.86 29 percent (95 percent CI: 8, 64) of patients achieved clinical protection when taking atropine. The remaining studies did not report on this outcome or provide data for calculation.

Adverse effects. Overall, IB was well tolerated with seven studies reporting no adverse effects.79,82,87,88,93–95 Dry mouth or thirst was reported in three studies: 16 percent of patients in two studies in which the dose was 0.04 mg of IB,80,96 and 20 percent at a dose of 0.2 mg and 60 percent at a dose of 2 mg in the third study.90 Other side effects included bitter taste90 and slight tremor.96

Larsson et al. reported no side effects with oxitropium.88 The remaining two studies did not report on adverse effects. None of the studies that examined atropine reported on adverse effects.

Subgroup/sensitivity analyses. Subgroup analyses based on age suggest similar results for children and adults for maximum percent decrease in FEV1. Analyses based on asthma severity did not reduce heterogeneity. Sensitivity analysis on high versus low quality studies did not reduce the heterogeneity.

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

Description of Included Studies

Seven trials97–103 met the inclusion criteria for the review investigating whether a defined exercise warmup protocol induces a refractory period that attenuates or prevents the EIB/EIA phenomenon within 2 hours of a subsequent ECT. Six trials used a crossover design;97,99–103 for one abstract the specific design could not be determined.98 The studies were published between 1979 and 2007. Four trials were conducted in Europe,97,98,101,102 two in Australia101,103 and two in North America.99,100 The definition of EIB/EIA was measured by at least a 1098,100,101 or 1597,99,102 percent fall in FEV1.

In three trials97,98,102 it was not clear if the ECT following the warmup met the ATS criterion of a work rate equal to 80 to 90 percent of an individually calculated maximum.164 The time between warmup and ECT ranged from 1 to 49 minutes. The characteristics of the trials are summarized in Table 38.

Table 38. Description of trials in the therapy review: Refractory period.

Table 38

Description of trials in the therapy review: Refractory period.

The baseline characteristics of the participants are presented in Table 39. The number of patients enrolled ranged from 6 to 46. One study97 included children only, three included adults only,99,100,102 and three included both children and adults.98,101,103 All participants in these trials except for one98 had confirmed asthma; for most, their asthma was mild and stable.

Table 39. Baseline characteristics of patients in trials in the therapy review: Refractory period.

Table 39

Baseline characteristics of patients in trials in the therapy review: Refractory period.

Five of the seven studies investigated two or more warmup protocols. In order to capture all comparisons the protocols were categorized into three subgroups based on the intensity of the routine. Four routines involved several short sprints and are designated “interval” warmups;97,99,100,103 two studies involved two standardized challenges 45 minutes apart and are designated “continuous high intensity” warmups;102,103 three involved treadmill runs at work rates of 60 percent and are designated as “continuous low intensity” warmups.99,101,102 Study arms that involved drug therapy with or without a warmup are not included in this analysis.100 Finally, the study by Eck et al.,98 and one arm of the study by Schnall et al.103 that used a continuous routine followed by an interval routine are reported separately.

Methodological Quality of Included Studies

Table 40 summarizes the methodological quality of the included studies. Overall the seven trials received low Jadad scores ranging from 1 to 2. No study described the randomization method. None stated that the assessors were blind; it was not possible to blind the participants to the type of warmup. Concealment of allocation was unclear in all the trials. The body of evidence is classified as “low.”

Table 40. Quality assessment of trials in the therapy review: Refractory period.

Table 40

Quality assessment of trials in the therapy review: Refractory period.

Quantitative Results

In all studies the mean maximum percent fall in FEV1 or PEF was measured after a control ECT of 5 to 10 minutes and compared with the percent fall in an identical challenge after a designated warmup routine. The pooled results are presented as a MD in the maximum percent fall in FEV1 or PEF between the two challenges. Because studies are included in more than one group, only subgroup pooled estimates were calculated.

Interval protocol. Four trials compared an interval warmup to a control challenge with no warmup.97,99,100,103 The protocols involved repetitive sprints of 26 to 30 seconds at 100 percent maximal oxygen consumption (VO2max) or higher. The mean difference in the maximum percent fall in FEV1 ranged from an improvement of 4.8 to 16.1 percent over the control challenge. The pooled results showed that a series of short intense sprints attenuated the EIB/EIA response by a mean of 10.6 percent (95 percent CI: −14.7, −6.5; I2=15 percent) (Figure 27). One study reported that one of the 12 participants had falls in FEV1 of less than 15 percent99 and would be classified as having obtained complete protection from EIB/EIA following the interval warmup.

Figure 27. Warmup vs. no warmup in pre-exercise treatment of EIB/EIA: maximum percent decrease in FEV1 or peak expiratory flow.

Figure 27

Warmup vs. no warmup in pre-exercise treatment of EIB/EIA: maximum percent decrease in FEV1 or peak expiratory flow.

Continuous low intensity protocol. Three trials compared a continuous low intensity warmup that ranged from 3 minutes101 to 30 minutes102 to a control challenge with no warmup. Exercise intensity for the warmup was 60 percent of maximum heart rate,101 60 percent of VO2max,99 and reported as low intensity in the third study.102 The mean difference in the maximum percent fall in FEV1 ranged from no improvement to 20.6 percent over the control challenge. The pooled results showed that this type of warmup attenuated the EIB/EIA response by a mean of 12.6 percent (95 percent CI: −26.7, 1.5; I2=90 percent) (Figure 27). One study also reported that 6 of the 12 participants had falls in FEV1 of less than 15 percent and classified them as having obtained complete protection from EIB.99

Continuous high intensity protocol. Two trials compared a continuous high intensity warmup102,103 to a control challenge with no warmup. Exercise intensity for the warmup was a heart rate equal to180 beats per minute103 and 98±2 percent of predicted maximum.102 The mean difference in the maximum percent fall in FEV1 ranged from very little improvement (0.99) to 17.6 percent over the control challenge. The pooled results showed that this type of warmup attenuated the EIB/EIA response by a mean of 9.8 percent (95 percent CI: −26.0, 6.4) (Figure 27). Substantial heterogeneity was identified (I2=89 percent).

Combination protocols. Eck et al.98 reported that there was no significant difference among the three protocols they investigated—one each of continuous 10 minutes of low intensity running, 10 minutes of running in intervals, and 10 minutes of exercising with increasing intensity. No data on the intensity of each warmup were reported. When combined, the three protected 79 percent (36/46) of participants from EIB/EIA (cutpoint not reported). The combined mean maximum percent fall in FEV1 on the challenge following all warmup protocols was compared with no warmup and indicated a mean improvement of 11.0 percent (95 percent CI: 7.5, 14.6) (Figure 27).

One of three protocols in the Schnall and Landau103 trial involved a combination of continuous and interval segments—a 6 minute treadmill run (heart rate 180 beats per minute), a 10 minute rest followed by 7 × 30 second sprints (treadmill speed increased 120 to 130 percent over first run), then a 20 minute rest before the final challenge of the same intensity as the original 6 minute run. The mean difference in the maximum percent fall in FEV1 compared with no warmup was 10.4 percent (95 percent CI: 0.2, 21.1).

Other comparisons: One study reported no significant changes among three formats of warmup in measures of respiratory heat and water loss.99

PubReader format: click here to try

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...