A Systematic Review of the Effects of Bronchodilators on Exercise Capacity in Patients With COPD^* FREE TO VIEW

COPD is a major health problem worldwide, and both morbidity and mortality are rising.¹There is currently no cure for COPD, and much attention has been paid to smoking cessation as the sole beneficial measure for both development and prognosis of COPD. Because this is only effective in 20% of patients,²symptomatic treatment with bronchodilators is the mainstream of therapy.^3–5 One of the main goals of bronchodilator therapy is to decrease airflow limitation in the airways and, as a consequence, improve dyspnea and exercise tolerance. The focus of this systematic review is to assess the effects of bronchodilator treatment on exercise capacity. Based on a literature search in MEDLINE, the following questions will be addressed: (1) Does bronchodilator therapy improve exercise capacity in patients with COPD? (2) Is the effect different for three categories of bronchodilators, ie, anticholinergics, βべーた₂-mimetics, and xanthine derivates? (3) Does a combination of bronchodilators have an additional effect compared to bronchodilator monotherapy? (4) Is the effect of a specific bronchodilator category comparable in both steady-state and incremental exercise tests? To this aim, we will first describe the most frequently used exercise tests, classified as incremental and steady-state tests, as well as our literature search strategy.

To assess exercise capacity, different protocols are used depending on the aim of the study. The tests can be classified as steady-state tests and incremental tests, both measuring a different aspect of exercise capacity (Table 1 ). The term steady state is employed to indicate a more or less constant work rate during the test. Incremental exercise tests assess maximal exercise capacity in terms of peak exercise level, whereas steady-state tests explore the maximal capacity that can be endured over a longer time period. The incremental tests are mostly performed on a cycle ergometer, but sometimes a treadmill or shuttle-walking test is used. The 6-min walking distance (6-MWD) and 12-min walking distance (12-MWD) tests are the most frequently used steady-state tests. Cycle endurance tests are also employed for the same goal.

Incremental exercise testing provides the possibility to study the integrated systemic responses to gradually increasing workloads, either using a treadmill or a bicycle.⁶ Both tests use a protocol of fixed increments in demanded exercise.^6–16 The workload is gradually intensified by increasing the resistance on the cycle ergometer, or the slope and/or speed on the treadmill ergometer. Patients are encouraged to continue the exercise until exhaustion. The test, therefore, is a symptom-limited exercise test. The most frequently used parameters are maximal workload (Wmax) and maximal oxygen consumption (V̇o₂max), but time to exhaustion (TTE), distance walked, maximal ventilation, maximal heart rate, maximal carbon dioxide production (V̇co₂max), and maximal Borg score (BSmax) are also used.

In the shuttle-walking test, subjects repeatedly walk a fixed distance of 10 m between two cones.^7,17 The increasing speed is dictated by audio signals, between which the fixed distance has to be covered. The time available for each of the following 10-m distances decreases after each completed 10-m distance. The outcome parameter is the distance walked until the patient stops because of dyspnea or other complaints, when the patient walks too slowly to cover the distance in time, or when a heart rate of 85% of the maximal heart rate for age is attained.

The 12-MWD and 6-MWD tests are self-paced tests that measure steady-state exercise capacity.^18–21 Subjects are instructed to walk as far as possible in 12 min (12-MWD) and 6 min (6-MWD).¹⁹ McGavin et al¹⁹ demonstrated the usefulness of the 12-MWD test, and Butland et al¹⁸ subsequently demonstrated that the 6-MWD test yielded similar results. Guyatt and coworkers²² demonstrated that the results are to a certain extent dependent on the patient’s motivation and encouragement. Therefore, they advocated not to encourage patients during the test. Patients need time to learn their optimal exercise level and strategy. To eliminate the learning effects from the result, it is recommended to perform two test sessions.

The endurance cycle test is performed with a constant workload. Moderate-intensity workloads result in steady-state responses that are comparable to many activities in daily life. High intensities of a constant workload will result in maximal values of exercise parameters. TTE, V̇co₂max, V̇o₂max, maximal ventilation, and BSmax are frequently used as outcome measures.

A systematic review of the literature searched with MEDLINE was performed including articles up to September 1999. Only randomized, controlled, double-blind trials written in English, were selected. Firstly, a database including all articles about obstructive lung diseases (key words: COPD, lung diseases, obstructive, plus all subheadings) was constructed. From this database, all articles that investigated patients with asthma were excluded (key-word strategy: database minus [asthma minus (asthma and COPD)]). A second database was defined including all articles about bronchodilators (key words:“ exploding” adrenergic-βべーた-agonist plus all subheadings,βべーた ₂-agonists, salmeterol, formoterol, adrenergic, long- and short-acting, “exploding” cholinergic-antagonist plus all subheadings, anticholinergics, ipratropium bromide, Oxivent, oxitropium, theophyllines, aminophyllines, sustained release, slow release). A third database was created with all articles describing exercise tests (key words: exercise, ergometry, treadmill, walking, cycling). Articles appearing in all three databases were selected. These articles were checked to see if they actually contained original, randomized, controlled, double-blind data on the effects of bronchodilator therapy on exercise capacity in patients with COPD.

Seventeen studies examining the effects of anticholinergics on exercise in patients with COPD were identified (Table 2 ). Ten studies applied ipratropium, 6 studies applied oxitropium, and 1 study applied atropine. At present, no tiotropium study is available. Twelve of 17 studies used a single-dose protocol; 16 of these studies primarily focusing on exercise testing showed a significant effect on FEV₁(Table 2).

We identified seven studies^23–29 with anticholinergics employing steady-state exercise tests. Three of four studies with single-dose anticholinergics showed significant improvements in walking distance.^23–24,27 In the maintenance-dose studies with ipratropium, significant improvements in the walking distance were found after 1 week of active treatment in one study,²⁵ and on only one of five time points in the other study.^28– O’Donnell et al²⁹ found a significant improvement of 2.8 min in symptom-limited exercise endurance time at 50 to 60% of Wmax with nebulized ipratropium, 500 μみゅーg. There were no obvious differences between studies using the 12-MWD and 6-MWD tests, nor between studies using oxitropium and ipratropium.

Three studies^30–32 of five studies^30–34 showed a significant effect of ipratropium bromide in an incremental cycle ergometer (ICE) test. There is a suggestion of a better effect on exercise capacity of higher doses of ipratropium compared to lower doses.^30–32 A single dose of oxitropium had a good effect on exercise tolerance in four of five studies.^{24–25,35–37} The effect of maintenance treatment was investigated in two studies,^32,38 one study with ipratropium for 1 week and one study with oxitropium for 1 year; both studies found significant improvements in exercise capacity.

Tables 3, 4 summarize the studies focusing on the effects of treatment withβべーた ₂-agonists on exercise capacity in patients with COPD. We identified 14 studies with short-actingβべーた ₂-agonists and 4 studies with long-actingβべーた ₂-agonists.

Salbutamol, irrespective of method of administration, dose, or duration of therapy, improved steady-state exercise results significantly in six of seven studies.^{23–26,39–41} Terbutaline, which was used in three studies, had no significant effect on exercise capacity,^42–43 and also when 5 mg bid was nebulized.⁴⁴Metaproterenol (orciprenaline) was tested in two studies.^45–46 The single-dose study^45– found a significant improvement, but the maintenance-dose study (1 week)⁴⁶ found no significant differences compared to placebo treatment. Three double-blind studies^28,47–48 using salmeterol, 50 to 100 μみゅーg bid, did not find a significant effect on walking distance (after 4, 16, and 12 weeks, respectively), despite a significant improvement in FEV₁ in the first two studies.

Two studies^45–46 could not detect a significant difference in V̇o₂max between an incremental test after treatment with metaprotenerol and placebo. One study^,49 with terbutaline yielded no difference compared to placebo treatment in an ICE test, which is similar to the lack of effect in the steady-state tests. Unfortunately, there is no study available assessing the effects of salbutamol in an ICE test. Therefore, we are unable to compare the outcomes of salbutamol directly to the results of steady-state exercise tests. Salmeterol was tested for its effect on exercise capacity by an ICE test in one study.⁴⁷ This group used the concept of physiologic strain according to Spiro et al,^50–51 and, like in the walking test, found no significant difference between both single-dose and maintenance treatment compared with placebo treatment. Liesker et al³² tested for the effects of treatments with formoterol on exercise with an ICE test. In this study,³² three different doses of formoterol were administered for 1 week (delivered doses, 4.5 μみゅーg bid, 6 μみゅーg bid, and 12 μみゅーg bid). Significant differences in TTE for all doses compared to placebo treatment were found, with a significant but small negative dose-response relationship for formoterol, which is as yet unexplained.³²

We identified three single-dose studies and five maintenance-dose studies with theophyllines (Table 5 ).

Five studies used a walking test to investigate the effects of theophyllines on exercise capacity in patients with COPD. Three of these studies^39,45,52– investigated maintenance therapy using theophylline for 1 to 4 weeks with checks on effective blood levels. None of these three studies found a significant effect on exercise capacity assessed by a 12-MWD test. Two studies used single-dose treatment, of which the clinical relevance is unclear. Leitch et al⁵³found a significant improvement in walking distance after a single dose of aminophylline, 450 mg. Evans⁵⁴ administered up to 800 mg of theophylline in single dose, yet they could not detect a significant improvement in exercise tolerance.

We identified six studies^{31,33,39,45,52,55} investigating the effect of theophylline by incremental exercise test (Table 5). One single-dose study,^,33 in which the medication was received orally, showed no effect in a treadmill test. Patients were treated with different doses of oral theophylline or aminophylline in five studies for 3 days, 7 days, 7 days, 28 days, and 1 month, respectively.^{31,39,45,52,55} These five studies checked for effective serum levels; if necessary, medication was adapted to achieve predefined effective blood levels.

The 3-day and 28-day studies found a significant improvement of exercise in terms of both Wmax and V̇o₂max.^31,55 Both 1-week studies and the 1-month study could not detect any significant improvement in exercise capacity.^39,45,52

Eight studies^{25–26,31–34,45,53} performed a head-to-head comparison of two bronchodilators of a different class, all in relatively small numbers of patients (range, 10 to 24 patients). None of these studies showed a significant difference between two bronchodilators in their bronchodilating capacity nor in the steady-state exercise test (Table 6 ). In seven studies, no significant differences were found in exercise tolerance while comparing theophyllines to anticholinergics,^,27 theophyllines toβべーた ₂-agonists,^,45,53 and anticholinergics toβべーた ₂-agonists.^{,25–26,31,34} One maintenance study³² in which subjects were treated for 7 days showed a longer cycle time in an ICE test with the use of ipratropium, 80 μみゅーg tid, than with formoterol, 18 μみゅーg bid.

In daily practice, many patients with moderate-to-severe COPD use more than one bronchodilator. The rationale for combining bronchodilators is a possible additive effect.

One study²⁶ showed a significant advantage of the combination of single-dose ipratropium with salbutamol vs placebo treatment in the 12-MWD test. Although better than placebo treatment, combination therapy was not better than either ipratropium or salbutamol alone.²⁶ Dullinger et al⁴⁵ and Leitch et al⁵³ investigated the combination of ipratropium plus theophylline using the 12-MWD test. Combination therapy was not better than either monotherapy, but was better than placebo treatment.

Three studies tested a combination of bronchodilators (fenoterol plus ipratropium,³⁴ metaproterenol plus theophylline,⁴⁵ and ipratropium plus theophylline³¹) vs the respective monotherapies and vs placebo treatment. Two of these studies did not find a difference in V̇o₂max between the combination therapy and monotherapy, nor between the monotherapies. Tsukino et al^,31 found a significant improvement in Wmax and FEV₁ by treatment with ipratropium, 160 μみゅーg, plus salbutamol, 200 μみゅーg, compared to placebo treatment and both monotherapies. Although Tobin et al^,34 found a significant improvement of the combination therapy (fenoterol, 400 μみゅーg, plus ipratropium, 40 μみゅーg) compared to monotherapy with fenoterol, there was, strangely enough, no improvement compared to placebo treatment.

Many authors (and patients) have claimed improvement in dyspnea scores in patients with COPD even in the absence of improvements in level of airflow limitation or exercise capacity. Therefore, we compared not only objective but also subjective improvements in exercise capacity by bronchodilators. It proved very difficult to aggregate the dyspnea ratings during exercise from the studies discussed in this article. Problems encountered were differences in exercise tests (steady-state and incremental tests); differences in dyspnea rating (BSmax, visual analog scale, breathlessness rating); differences in expression of results (eg, BSmax, end-exercise Borg score, Borg score slope); and, last but not least, difficulties in interpretation of the results. The most important interpretation problem in the dyspnea evaluation arises when a significant improvement in exercise is seen without an improvement in dyspnea. In this case, it is very likely that the dyspnea score would have shown an improvement if it had been measured at the same exercise level before and after the intervention, instead of at different end-exercise levels. This is illustrated nicely in at least in one study,²⁹ which showed no difference in the end-exercise Borg score, but significant lower Borg scores before compared to after the intervention, when measured at the same exercise duration. Therefore, we also scored this latter situation as an improved dyspnea score. Only studies assessing both exercise capacity and dyspnea related to exercise were included in this part of the analyses.

All four studies^23–25,29 traced in the literature assessing dyspnea during exercise with anticholinergics showed positive results. Four of five studies^{25,28,47–48} with βべーた₂-agonists showed an improved dyspnea rating. Only one study^,31 with a dyspnea rating was found for theophyllines, the results of which were negative.

Studies ofβべーた ₂-agonists^32,47 and theophyllines^31,39 contained data on exercise and on dyspnea; findings of two of these studies^31–32 were positive. With anticholinergics, all eight studies^{24,30–32,35–38} containing information about exercise and dyspnea showed positive effects on dyspnea and/or exercise.

We identified 33 double-blind, randomized, controlled studies investigating the effects of bronchodilators on exercise capacity in patients with COPD. Only a few studies have been published with a head-to-head comparison between different types of bronchodilators. Bronchodilators are frequently prescribed to patients with COPD in order to improve exercise intolerance. Nevertheless, our systematic review shows that approximately half of the studies with bronchodilators do not show a significant improvement in exercise capacity. Why is this?

Doctors prescribe a bronchodilator to patients with COPD in the hope of improving symptoms, among others, dyspnea during exercise and exercise capacity. This carries the implicit assumption that patients with COPD are primarily limited in their exercise by dyspnea due to a limited ventilatory capacity.⁵⁶Additionally, it implicates that this limitation can be (partly) improved by bronchodilator therapy. First of all, it is unclear whether the patients in the studies reported in this review were actually limited by their ventilatory capacity. This situation occurs if the demand for ventilation during exercise rises above approximately 80% of the maximal minute ventilation (37.5 × FEV₁).⁵⁷ None of the studies state if they checked at baseline (before bronchodilation) whether their patients were ventilatory limited in the exercise protocol used (type and time of exercise) by measuring maximal ventilation. If, for instance, patients with mild COPD do not reach their ventilatory limitation in a steady-state exercise protocol, it is doubtful whether bronchodilator therapy will provide any benefit. The same is true if patients are limited by nonventilatory reasons, such as cardiovascular limitation (maximal heart frequency), muscle performance (diminished lower-extremity force by atrophy or change of muscle type in COPD), motivation,¹² or diminished diffusion capacity.

The second reason for finding nonsignificant results can be the selection of the study population. In the case that patients with COPD are indeed limited by their ventilatory capacity, and thus by their airway obstruction, and additionally show a decrease in their airway obstruction by a bronchodilator, then these patients can be expected to improve their exercise capacity by bronchodilator therapy. However, some studies^{36–39,41,43,45–46,48–49,52,54,58} assessing the effects of bronchodilators on exercise exclude patients with reversibility to avoid inclusion of asthmatics in the study. Other studies of COPD did not use an exclusion criterion of reversibility of airflow limitation to select patients. Thus, the chance of finding positive effects on the measured exercise parameters is enhanced.^{23–28,30–31,33,35,39,42,44,47,53,55,59} In Figure 1 , we show that a positive correlation exists between the mean level of reversibility in different studies and the respective mean improvement in 12-MWD. All together, when looking at Tables 23456, the majority of studies did find a significant effect of bronchodilator therapy on FEV₁. However, in general, the improvements are small, which makes a significant effect on exercise capacity less likely.

The third possible reason for a lack of effect of bronchodilator therapy in some exercise studies is the dosing of the bronchodilator. In order to achieve an increase in exercise capacity in patients with COPD, it is essential to administer the patient an adequate dose of bronchodilators. For instance, it is unclear whether the therapeutic blood levels of theophyllines are reached with a single oral dosage, and thus whether the observed negative results in single-dose studies^33,53–54 are meaningful. Studies^26,28,30,34 using low doses of ipratropium (daily dose ≤ 80 μみゅーg) did not show a significant effect. Studies^{25,27,30–32} using a higher dose of ipratropium (daily dose > 80 μみゅーg) generally showed a significant improvement in exercise capacity. In the study of Ikeda et al,³⁰ only the higher doses of ipratropium provided a significant increase in exercise capacity. Thus the data suggest that the higher doses of ipratropium are needed to improve exercise capacity.

The fourth possibility for the lack of effect by bronchodilators with respect to exercise capacity may be potential negative effects of bronchodilators, like an increase in ventilation/perfusion mismatch,⁶⁰or diminution of peripheral muscle function, eg, by theβべーた ₂-agonist.⁶²

The fifth point of concern is that the tests can be sensitive to a learning effect. This learning effect has been described for the walking tests, but as far as we are aware, it has not been investigated whether this is also the case in other tests of exercise capacity. Two training sessions are necessary to eliminate a learning effect in walking tests.^18–19 However, some of the studies^27,40 failed to incorporate two practice sessions in their protocols. In uncontrolled studies, the increase in exercise capacity, as a consequence of learning, can be interpreted as an effect of the bronchodilator (false-positive effect). By contrast, in randomized, controlled studies, this phenomenon normally increases the variability in the results, thereby increasing the chance of false-negative results.

The last reason for lack of effects in exercise studies with bronchodilators in patients with COPD as evaluated in the current review is that the number of included subject in most of the studies was rather small (10 to 60 subjects). Thus, a lack of power may simply explain the lack of effect. On the other hand, no studies showed deleterious effects of bronchodilators on exercise capacity in patients with COPD.

Conclusions in randomized, controlled trials are based on statistical significance, yet a statistical significance does not always reflect clinical significance. For the 6-MWD test, Redelmeier and coworkers⁶³ showed that subjects have to improve their walking distance by 54 m in order to appreciate this increase as a beneficial effect. For the 12-MWD test, the minimal clinically relevant distance is unknown. None of the studies^{23–24,27,40,42–44,48,54,58} performed with a 6-MWD test that did find statistically significant results reached the minimal clinical significance limit of Redelmeier and coworkers.⁶³ Thus the relevance of the observed effect to the patient remains debatable. As far as we are aware, the minimal clinical significance levels in the other tests of exercise capacity have not been defined.

The BSmax was used in 11 ICE studies.^{24–25,30–32,35–38,43,47} In only two of these studies,^24,36 a significant difference was found in end-exercise BSmax. These results are not surprising. Since maximal exercise tests are limited by dyspnea, the individual BSmax should be expected not to change even if the exercise improves with the intervention. Therefore, we advocate not to use the BSmax score in incremental exercise tests, but other parameters or other dimensions of the Borg score, eg, the Borg scale slope (ΔでるたBorg score/ΔでるたV̇o₂max). This parameter indicates the rise in dyspnea when patients increase their oxygen consumption during exercise. Four of five studies^{,29,31,36–38} using this parameter found a significant improvement in the Borg scale slope, while no difference in BSmax score was found.^31,36–38 This signifies that patients experienced less dyspnea at the same level of oxygen consumption during exercise with the investigated treatment than with placebo treatment.

A point of interest to us was whether the effects varied for different classes of bronchodilators. Eight studies^{25–26,31–34,45,53} evaluated two bronchodilators of different classes in a head-to-head comparison. Most of these studies could not find a difference between the effects of the bronchodilator classes. However, Tables 23456 show more positive study findings with anticholinergics and βべーた₂-agonists than when theophyllines are being used.

Whether the effect of a specific bronchodilator category is similar in both steady-state and incremental exercise tests cannot be answered, since only six studies^{24,39,45–47,52} assessed both tests. Five of these studies showed a concordant result, ie, results of both tests were not significant, or results of both tests were significant.^{,24,39,45,47,52} Although it has been often suggested that steady-state exercise tests have more relevance to daily life activities in patients with COPD, no empirical preference can be deduced for a certain type of exercise protocol from the available studies in this systematic review.

In summary, this review shows that the effects of bronchodilators on exercise capacity in general are limited. Anticholinergic agents have significant beneficial effects in the majority of studies, especially when measured by steady-state exercise protocols. There is a trend toward a better effect of high-dose compared to low-dose anticholinergics. Short-acting βべーた₂-mimetics have favorable effects on exercise capacity in more than two third of the studies, but surprisingly, the situation is less clear for long-actingβべーた ₂-agents. The majority of findings of the published reports with theophyllines and their effects on exercise are negative. Direct comparisons of different classes of bronchodilators have not been made in a sufficient number of studies for a rational preference. The addition of a second bronchodilator has no proven advantage for improving exercise test results, but this has not been studied extensively and not in sufficiently large studies. The majority of studies reporting a measure of dyspnea found improvements, even in the absence of improvement in exercise capacity.

Until recently, only two large-scale studies have been performed to investigate the effect of bronchodilators (salmeterol⁴⁷ and ipratropium^47–48) on exercise capacity in patients with COPD. Studies with a small number of patients found variable effects of bronchodilators. Additionally, no large-scale study used the endurance cycle by O’Donnell et al,²⁹ which seems to be the most responsive test used until now. For a better understanding of the effects of bronchodilators on exercise capacity in patients with COPD, it would be of clinical and pathophysiologic interest to perform larger-scale studies with TTE on an endurance cycle test as the primary outcome, and additionally lung function parameters of the small airways as secondary outcomes. Also, the relationship of measurements of exercise capacity to measurements of daily life activities, for instance with an actometer, which documents body activity 24 h/d, could be a valuable additional research topic.