American Journal of Respiratory and Critical Care Medicine

Rationale: Exercise-induced dynamic hyperinflation contributes to decreased exercise tolerance in chronic obstructive pulmonary disease (COPD). It is unknown whether respiratory retraining (ventilation–feedback [VF] training) can affect exercise-induced dynamic hyperinflation and increase exercise tolerance.

Objectives: To determine whether patients with COPD would achieve longer exercise duration if randomized to a combination of exercise training plus VF training than either form of training on its own.

Methods: A total of 64 patients randomized to 1 of 3 groups: VF plus exercise (n = 22), exercise alone (n = 20), and VF alone (n = 22).

Measurements and Main Results: Exercise duration before and after 36 training sessions and exercise-induced dynamic hyperinflation and respiratory pattern before and after training were measured. In the 49 patients who completed training, duration of constant work-rate exercise was 40.0 (± 20.4) minutes (mean ± SD) with VF plus exercise, 31.5 (± 17.3) minutes with exercise alone, and 16.1 (± 19.3) minutes with VF alone. Exercise duration was longer in VF plus exercise than in VF alone (P < 0.0001), but did not reach predetermined statistical significance when VF plus exercise was compared with exercise alone (P = 0.022) (because of multiple comparisons, P ⩽ 0.0167 was used for statistical significance). After training, exercise-induced dynamic hyperinflation, measured at isotime, in VF plus exercise was less than in exercise alone (P = 0.014 for between-group changes) and less than in VF alone (P = 0.019 for between-group changes). After training, expiratory time was longer in VF plus exercise training (P < 0.001), and it was not significantly changed in the other two groups.

Conclusions: The combination of VF plus exercise training decreases exercise-induced dynamic hyperinflation and increases exercise duration more than VF alone. An additive effect to exercise training from VF was not demonstrated by predetermined statistical criteria.

Clinical trial registered with www.clinicaltrials.gov (NCT 00037973).

Scientific Knowledge on the Subject

Exercise-induced dynamic hyperinflation is a major contributor to decreased exercise tolerance in chronic obstructive pulmonary disease.

What This Study Adds to the Field

Ventilation–feedback (VF) plus exercise decreases dynamic hyperinflation more than VF alone or exercise alone, and increases exercise duration more than VF alone. The added effect of exercise and VF on exercise duration was not significant using preset statistical criteria.

Exercise tolerance is decreased in patients with chronic obstructive pulmonary disease (COPD) (15). Decreased exercise tolerance causes significant disability, which, in turn, profoundly affects quality of life (6). Mechanisms responsible for decreased exercise tolerance include inability to increase delivery of oxygen to the peripheral muscles, peripheral muscle dysfunction (7, 8), pulmonary hypertension (9), and, possibly, psychological factors (10). In many patients, however, exercise-induced dynamic hyperinflation is a major contributor to decreased exercise tolerance (1113). Dynamic hyperinflation reduces the capacity of the respiratory muscles to generate pressure (14), decreases Vt, increases mechanical load (15), and increases dead space-to-Vt ratio (13, 16).

In patients with COPD, the critical mechanisms responsible for development of dynamic hyperinflation during exercise are expiratory flow limitation and increased relaxation volume (17). Accordingly, it has been reasoned that slowing of the respiratory rate may decrease exercise-induced dynamic hyperinflation and, consequently, improve exercise capacity (18).

Using a computerized ventilation–feedback (VF) system, we recently demonstrated that patients with COPD can be trained to modify their respiratory pattern during cycle exercise (19). After 6 weeks of cycle-ergometry training, patients randomized to VF training demonstrated a slower respiratory rate (during exercise testing)—even when the computer monitor (used for feedback training) was not displayed (19).

The primary objective of this study was to determine whether patients with COPD would achieve longer exercise duration on a treadmill set at a constant work rate if randomized to a combination of exercise training plus VF training than they would to either form of training on its own. The secondary objectives were to establish whether VF training would alter respiratory pattern, reduce exercise-induced dynamic hyperinflation, and reduce perceived dyspnea during treadmill exercise testing.

This is the first randomized, controlled trial assessing the effects of a computerized respiratory program on exercise endurance and dynamic hyperinflation. Some of the results have been previously reported in abstract form (20), and preliminary results focused on cycle-ergometry testing after 6 weeks of training were reported in the form of a paper (19).

See the online supplement for additional details. Patients with moderate to severe COPD were recruited into the study (Table 1). Inclusion criteria were: ⩾40 years of age; PaO2 ⩾ 56 mm Hg at rest; SpO2 ⩾ 85% at peak exercise; stable clinical condition. The study was approved by the local human studies subcommittee and written informed consent was obtained.

TABLE 1. BASELINE PATIENT CHARACTERISTICS


Characteristics

VF plus Exercise (n = 22)

Exercise Alone (n = 20)

VF Alone (n = 22)
Age, yr68 ± 965 ± 769 ± 7
Height, cm177 ± 5174 ± 7176 ± 6
Weight, kg93.7 ± 18.085.5 ± 21.086.4 ± 18.7
BMI29.7 ± 5.328.3 ± 6.828.0 ± 6.1
MMSE28.5 ± 1.428.7 ± 1.128.3 ± 1.4
Smoking, pack-years55.9 ± 43.662.7 ± 34.565.2 ± 56.2
Comorbidity score2.1 ± 1.51.8 ± 1.52.0 ± 1.6
FEV1, L (% pred)1.36 ± 0.67 (41 ± 18)1.34 ± 0.52 (43 ± 15)1.40 ± 0.54 (46 ± 16)
FEV1/FVC (% pred)49 ± 12 (62 ± 15)47 ± 12 (60 ± 16)51 ± 12 (66 ± 15)
MVV (% pred)47.6 ± 27.4 (41 ± 20)42.7 ± 16.8 (34 ± 16)50.6 ± 19.7 (40 ± 14)
TLC (% pred)6.92 ± 0.91 (110 ± 17)7.43 ± 1.7 (122 ± 18)7.25 ± 2.29 (119 ± 43)
RV/TLC (% pred)
55 ± 13 (160 ± 38)
56 ± 11 (161 ± 44)
55 ± 12 (151 ± 30)

Definition of abbreviations: BMI = body mass index (weight [kg] / height [m]2); MMSE = Mini-Mental Status Exam; MVV = maximal voluntary ventilation; RV = residual volume; TLC = total lung capacity; VF = ventilation–feedback.

No significant differences were identified between the groups. Values are mean ± SD.

Patients completed the Chronic Respiratory Disease Questionnaire (21) and underwent symptom-limited cycle-ergometry (22) and symptom-limited treadmill exercise testing (Table 2) followed by constant work-rate treadmill exercise test (19) (Figure 1). Thereafter, using permuted block randomization, patients were assigned to: VF plus exercise training, exercise training alone, or VF alone (for details on the development of VF training, see Appendix E1 in the online supplement). After randomization, 15 patients (23%) withdrew or were withdrawn from the study (Figure 2). Characteristics of patients who completed the study and patients who did not were not different.

TABLE 2. BASELINE SYMPTOM-LIMITED TREADMILL EXERCISE TESTING: METABOLIC AND HEMODYNAMIC MEASUREMENTS RECORDED AT PEAK EXERCISE



VF plus Exercise (1) (n = 22)

Exercise Alone (2) (n = 20)

VF Alone (3) (n = 22)

P Value
Measures



1 vs. 2
2 vs. 3
1 vs. 3
o2 peak, ml · kg−1 · min−117.8 ± 4.217.7 ± 3.815.3 ± 2.80.970.040.07
e, L · min−149.4 ± 16.443.2 ± 15.443.9 ± 14.40.870.970.75
Vt, L1.69 ± 0.451.43 ± 0.471.44 ± 0.450.520.980.45
RR, breaths · min−129 ± 532 ± 631 ± 60.230.980.67
HR, beats · min−1123 ± 18126 ± 13113 ± 200.760.140.41
% HRpeak
81 ± 12
81 ± 7
75 ± 14
0.97
0.18
0.26

Definition of abbreviations: HR = heart rate; % HRPeak = % predicted peak heart rate; RR = respiratory rate; VF = ventilation–feedback; V̇o2 peak = peak oxygen uptake.

P values for all pairwise comparisons are included: 1 vs. 2 = VF plus exercise vs. exercise alone; 2 vs. 3 = exercise alone vs. VF alone; 1 vs. 3 = VF plus exercise vs. VF alone. Values are mean ± SD.

Training involved 36 sessions: 18 of leg-cycle exercise followed by 18 of treadmill exercise (Figure 1). The purpose of using cycling first was to make it easier for patients assigned to VF plus exercise to learn the respiratory technique while exercising.

Patients trained in the laboratory three times weekly. Exercise intensity (interval exercise training) was set at a percentage of baseline peak oxygen uptake (V̇o2 peak)—starting exercise intensity was 60% V̇o2 peak, and was followed by progressive increase to 85% of V̇o2 peak (see online supplement). Patients randomized to VF alone practiced VF for 30–35 minutes. A 10-minute period of low-intensity activity (18 sessions of unloaded cycling followed by 18 sessions of treadmill at 1.8 mph at 0% grade) were included in the workout so patients would experience using VF while active. Patients assigned to the two exercising groups began training for 25 minutes and progressed as tolerated up to 45 minutes of exercise. Training sessions during the last week of training lasted 37 (±5) minutes (mean ± SD) with VF plus exercise, 37 (±2) minutes with exercise alone, and 32 (±4) minutes with VF alone.

The VF system consisted of a heated pneumotachometer interfaced to a computer (19) (Figure 3). Patient duration of inhalation and exhalation were displayed on a monitor as a bar that moved left during inhalation and right during exhalation. Expiratory time (Te) goals were shown as “targets” on the right of the screen (Figure 3). The minimal expiratory flow rate (L/s) that patients were instructed to maintain during feedback training was based on data recorded during our pilot study (see Appendix E1). Goals of respiratory rate and the exhalation-to-inhalation ratio during feedback training were based on the breathing pattern recorded during baseline exercise stress test. At the start of training, the goal for respiratory rate was one to two breaths less than the respiratory rate recorded at a similar workload during the baseline exercise stress test. As training progressed, the respiratory rate goal was decreased by 1 breath/minute and the Te was increased by 0.05 s for a given workload (see online supplement for specific details).

The symptom-limited treadmill exercise test was repeated after 18 training sessions (when exercise training was changed from cycling to walking). As with the baseline symptom-limited treadmill exercise test, the symptom-limited treadmill protocol began with walking at 0% grade and 1.8 mph. Then, the grade of the treadmill was increased by 0.5% every 30 seconds throughout the test. In addition, at Minute 6, treadmill speed was increased by 0.2 mph every 3 minutes throughout the test. This test was performed to determine the target percentage of V̇o2 peak for treadmill training during the last 18 training sessions (see online supplement for details). To determine the value of V̇o2 peak at the conclusion of training, the symptom-limited treadmill exercise test was repeated after the 36th training session.

The constant work-rate treadmill exercise test was repeated after 18 training sessions. The constant work-rate treadmill test always began with a 2-minute warm-up period (0% grade at 1.8 mph) followed by an increase in grade and speed to the corresponding values recorded when patients had reached 85% of the V̇o2 peak during the baseline symptom-limited treadmill exercise test. The results of this test were used in the intention-to-treat analysis of patients who trained for 19–35 training sessions. To determine exercise endurance, extent of exercise dynamic hyperinflation and breathing pattern at the conclusion of training, the constant work-rate treadmill exercise test was repeated after the 36th session. Constant work-rate tests were continued until volitional exhaustion or until 60 minutes had elapsed (see online supplement and Figure E1 for details).

Statistical Analysis

Data are presented as means (±SD). Analysis of covariance was used to compare duration of constant work-rate treadmill exercise over time between groups. Pairwise comparisons were evaluated with the least significant difference procedure. Because multiple comparisons were involved, a P ⩽ 0.0167 was required to declare significant difference in duration of the constant work-rate treadmill exercise test (the main outcome variable). Paired t tests were used to determine differences within each group from baseline through the 36th training session (see online supplement for details). Gas exchange, respiratory pattern and extent of exercise-induced dynamic hyperinflation recorded during the constant work-rate treadmill exercise test were compared with the corresponding isotime values after training. Isotime was the time elapsed from the beginning of baseline constant work-rate treadmill exercise to the last set of measurements before completion of the baseline test, with the exception of two patients—in those two patients (one randomized to exercise plus VF and one randomized to VF alone), the baseline test was not the test with the shortest duration. In these cases, isotime was defined as the time elapsed from the beginning of constant work-rate treadmill exercise to the last set of measurements before completion of the test that was of the shortest duration. Adherence to training was quantified by computing the time to complete the 36 training sessions.

Sample

A total of 91 patients with COPD (97% male) were enrolled in the study (see Figure 2), and 27 patients were not randomized for the following reasons: lack of time to participate in the study, moved, or family illness (n = 10); cardiac or other medical conditions (n = 8); and not eligible after screening (n = 9). The remaining 64 patients were randomized to 1 of 3 groups: VF plus exercise training (n = 22), exercise training alone (n = 20), and VF alone (n = 22) (Table 1).

Effects of Training on Duration of Constant Work-Rate Treadmill Exercise Testing

At baseline, exercise duration on the constant work-rate treadmill exercise test was 10.5 (±5.1) minutes in patients randomized to VF plus exercise, 10.1 (±5.7) minutes in exercise alone, and 9.0 (±3.6) minutes in VF alone. These values include a 2-minute warm-up completed at 1.8 mph, 0% grade. To determine the independent contributions of the training protocol on the duration of (constant work rate) treadmill exercise, analysis of covariance (ANCOVA) was conducted on all randomized patients (intent-to-treat, n = 64) and on those who completed the study (n = 49). In the intent-to-treat analysis, end-of-training exercise duration was the exercise duration recorded during the last constant work-rate exercise test: third constant work-rate exercise test (after 36 training sessions) in 49 patients; second constant work-rate exercise test (after 18 training sessions) in 3 patients; and first constant work-rate exercise test (baseline test) in 12 patients. Covariates in the model included baseline exercise duration, severity of airway obstruction (FEV1% predicted), comorbidity score (Charlson Comorbidity Score), and adherence to the study (time to complete training sessions). Using ANCOVA with intent-to-treat procedures (n = 64), 54% of the variance for exercise duration at the end of training between groups was explained by the model (R2 = 0.54; P < 0.0001).

In the entire group of randomized patients, comparisons revealed that improvement in exercise duration was greater with VF plus exercise training (n = 22) than with VF alone (n = 22; P < 0.0001). Similarly, improvement in exercise duration was greater with exercise alone (n = 20) than with VF alone (P = 0.006). The improvement in exercise duration tended to be greater with VF plus exercise training (n = 17) than with exercise alone (n = 16; P = 0.051).

When the analysis was restricted to patients who completed the study (n = 49), mean exercise times at the end of training was 40.0 (±20.4) minutes in the VF plus exercise group, 31.5 (±17.3) minutes in the exercise-only group, and 16.1 (±19.3) minutes in the VF-only group. Improvement in exercise duration was again greater with VF plus exercise training (n = 17) than with VF alone (n = 16; P < 0.0001), and it was also greater with exercise alone (n = 16) than with VF alone (P = 0.004). The improvement in exercise duration tended to be greater with VF plus exercise training (n = 17) than with exercise alone (n = 16; P = 0.022). Observed power for the overall ANCOVA model with α equal to 0.05 was 100%. Observed power for the individual adjusted mean pairwise comparisons using pooled variance was 65% for the comparison of exercise plus VF versus exercise alone, 100% for the comparison of exercise plus VF versus VF alone, and 85% for the comparison of exercise alone versus VF alone.

Effects of Training on Isotime Exercise during Constant Work-Rate Exercise Testing

Compared with baseline, respiratory pattern at isotime in the VF plus exercise training group and in the VF-alone group demonstrated a magnitude of change not observed in the exercise-alone group (Table 3). Te increased by 36% (P < 0.001) with VF plus exercise training and 34% (P = 0.01) with VF alone. The corresponding value in the exercise-alone group was 11% (P = 0.03). The percent increase in Te recorded with VF plus exercise training was greater than the corresponding value with exercise alone (P = 0.018). Respiratory rate decreased by 24% with VF plus exercise training (P < 0.001), by 13% with VF alone (P = 0.01), and by 7% (P = 0.05) with exercise training alone. The percent decrease in respiratory rate recorded with VF plus exercise training was greater than the corresponding value with exercise alone (P = 0.018). Vt was not affected by any of the three interventions (P values = 0.56–0.87). V̇e decreased by 23% with VF plus exercise training (P < 0.05) and 15% with VF alone (P < 0.01). V̇e did not significantly change with exercise training alone.

TABLE 3. CONSTANT WORK-RATE TREADMILL EXERCISE TEST: HEMODYNAMIC AND RESPIRATORY VARIABLES RECORDED AT BASELINE AND AFTER COMPLETION OF TRAINING (AT CORRESPONDING ISOTIME)



VF plus Exercise Training

Exercise Alone

VF Alone
(n = 17)
(n = 16)
(n = 16)

Baseline
Post-Training
Baseline
Post-Training
Baseline
Post-Training
HR, beats · min−1118 ± 19106 ± 19127 ± 13111 ± 13116 ± 19104 ± 16*
o2, ml · kg−1 · min−117.3 ± 3.514.3 ± 3.017.3 ± 3.815.6 ± 3.1*14.9 ± 2.413.2 ± 2.1*
e, L · min−144.8 ± 16.134.3 ± 9.044.2 ± 13.141.1 ± 12.743.9 ± 12.337.4 ± 10.0
Vt, L · min−11.5 ± 0.51.6 ± 0.41.5 ± 0.51.5 ± 0.41.5 ± 0.51.6 ± 0.5
RR, breaths · min−129 ± 422 ± 330 ± 728 ± 6*30 ± 726 ± 8
Ti, s0.70 ± 0.160.86 ± 0.190.75 ± 0.210.78 ± 0.220.75 ± 0.190.85 ± 0.23*
Te, s1.40 ± 0.291.91 ± 0.451.33 ± 0.271.48 ± 0.281.32 ± 0.281.75 ± 0.61
Dyspnea score5.6 ± 4.11.3 ± 1.46.2 ± 3.82.2 ± 2.45.7 ± 3.03.3 ± 2.8
IC, L§
2.0 ± 0.6
2.5 ± 0.8
2.1 ± 0.9
2.2 ± 0.9
2.0 ± 0.6
2.2 ± 0.6

Definition of abbreviations: HR = heart rate; IC = inspiratory capacity; RR = respiratory rate; Te = expiratory time; Ti = inspiratory time; VF = ventilation–feedback.

Values are mean ± SD.

* P < 0.05 within-group differences.

P < 0.01 within-group differences.

P < 0.001 within-group differences.

§ n = 12 patients in each group for IC measures (due to technical problems with our pulmonary function software, our initial cohort of patients did not have pre- and postmeasurement of exercising IC).

At isotime, inspiratory capacity increased (less exercise-induced dynamic hyperinflation than baseline) by 0.54 (±0.43) L with VF plus exercise training (P < 0.001; n = 12). The increase of inspiratory capacity at isotime in the remaining two groups (n = 12 in each group) was not significant (Table 3). The increase in inspiratory capacity at isotime recorded with VF plus exercise training was greater than the corresponding change in inspiratory capacity recorded with exercise alone (P = 0.014, between-group changes) and the corresponding change in inspiratory capacity recorded with VF alone (P = 0.019, between-group changes).

At 5 minutes after exercise, patients were asked, “What was the primary reason that made you stop exercising?” (Table 4). Of patients assigned to VF plus exercise training, exercise alone, and VF alone, 53, 50, and 63%, respectively, stopped exercising secondary to dyspnea at baseline. The corresponding values after training were 18% (P = 0.014), 25% (P = 0.046), and 44% (P = 0.083). Perceived dyspnea decreased by 77% (±65%) with VF plus exercise training (P < 0.001), by 64% (±37%) with exercise alone (P < 0.001), and by 42% (±93%) with VF alone (P = 0.02).

TABLE 4. PRIMARY REASONS TO STOP SUBMAXIMAL TREADMILL EXERCISE TEST


Group

Breathlessness n (%)

Fatigue/Other n (%)

60-min Time Limit* n (%)
VF plus exercise (n = 17)
 Baseline testing9 (52.9)8 (47.1)0 (0)
 Post-training3 (17.6)7 (41.2)7 (41.2)
Exercise alone (n = 16)
 Baseline testing8 (50.0)8 (50.0)0 (0)
 Post-training4 (25.0)9 (56.3)3 (18.8)
VF alone (n = 16)
 Baseline testing10 (62.5)6 (37.5)0 (0)
 Post-training
7 (43.8)
9 (56.3)
0 (0)

Definition of abbreviation: VF = ventilation–feedback.

* Comparison of VF plus exercise vs. exercise alone: χ2 = 7.28; P = 0.03. Parenthetic values refer to the percentage of patients who experienced a given symptom in each group.

At baseline, no patient stopped because he had reached the preset exercise time limit of 60 minutes. After training, however, 41% of patients randomized to VF plus exercise training, and 19% of those randomized to exercise alone, stopped because they had reached the preset exercise time limit of 60 minutes. After training, no patient randomized to VF alone stopped because he had reached the exercise time limit of 60 minutes.

Effects of Training on Oxygen Uptake

At baseline, V̇o2 peak was greater in the VF plus exercise training than in the VF-alone group (P < 0.03) (Table 2). In patients randomized to VF plus exercise, V̇o2 peak increased from 17.4 (±3.6) ml · kg−1 · minute−1 at baseline to 18.9 (±4.9) ml · kg−1 · minute−1 at the end of training (P = 0.04). The corresponding values in patients randomized to exercise alone were 17.4 (±3.0) and 18.7 (±3.2) ml · kg−1 · minute−1 (P = 0.05) and, in those randomized to VF alone, were 15.1 (± 3.1) and 16.0 (±2.9) ml · kg−1 · minute−1 (P = 0.15). These results are similar to the findings of other investigators, where the observed changes in V̇o2 peak were modest when compared with improvements in the duration of submaximal constant work-rate exercise (3, 2427).

Patient Phenotype and Response to Training

To assess whether the addition of VF to exercise training may have greater benefit in patients with more severe obstructive disease, we divided the sample according to FEV1 values: patients with an FEV1 ⩽ 50% (n = 29) and patients with an FEV1 > 50% (n = 20). After adjusting for adherence and comorbidities, the end-of-training exercise duration of patients with an FEV1 ⩽ 50% was 36.7 (±3.7) minutes for the exercise plus VF group, 24.1 (±3.9) minutes for the exercise-alone group, and 11.9 (±4.6) minutes for the VF-alone group. The difference in exercise duration at the end of training between exercise plus VF and exercise alone was 52% (P = 0.027). After adjusting for adherence and comorbidities, the end-of-training exercise endurance of patients with an FEV1 > 50% was 48.0 (±7.6) minutes for the exercise plus VF group, 41.8 (±7.2) minutes for the exercise alone group, and 20.4 (±6.4) minutes for the VF-alone group. The difference in exercise duration at the end of training between exercise plus VF and exercise alone was 15% (P = 0.56).

Effects of Training on Health-related Quality of Life

Perceived dyspnea during activities of daily living improved by 4 (±6) points in the VF plus exercise training group (P = 0.02), by 6 (±5) points in the exercise alone group (P < 0.03), and by 6 (±5) points in the VF-alone group (P < 0.001). Emotional function improved only in the VF-alone group (6 ± 9 points; P = 0.05), and mastery improved only in the VF plus exercise training group (3 ± 3 points; P = 0.005). After training, fatigue scores were not significantly changed in any of the three groups. No differences between the groups were identified.

This is the first controlled trial designed to determine whether patients with COPD could achieve longer exercise duration if randomized to a combination of exercise training plus VF training than to either form of training on its own. There are four major findings. First, exercise duration was longer in patients randomized to VF combined with exercise training than in patients randomized to VF alone. Second, the combination of exercise training plus VF training achieved a strong trend toward longer exercise duration than did exercise training alone. Third, exercise-induced dynamic hyperinflation was less in patients randomized to VF combined with exercise training than in patients randomized to exercise training alone or VF training alone. Fourth, VF training, either alone or in combination with exercise training, could modify respiratory pattern during exercise.

Training and Exercise Duration

As expected (23, 2736), exercise duration (on the constant work-rate treadmill test) improved with exercise training—administered with or without VF. Moreover, for patients who completed training, exercise duration tended to be longer after VF plus exercise training than after exercise training alone (P = 0.022). Several factors may account for lack of statistical significance.

First, by study design, constant work-rate tests were terminated when patients completed 60 minutes of exercise. Seven patients undergoing VF plus exercise training achieved this predetermined target. In contrast, only three patients undergoing exercise training alone reached 60 minutes (χ2 = 7.28; P = 0.03). Thus, differences between the groups may have been underestimated by the a priori decision to place a time constraint on duration of the test.

Second, the duration of the constant work-rate treadmill test improved by 200% after exercise training alone. This improvement was more than that reported by Casaburi and colleagues (18) and by Ortega and colleagues (24), and similar to that reported by Emtner and colleagues (23). Training at high intensity (goal, 85% of the V̇o2 peak recorded after the first 18 sessions of training) is a likely mechanism for the recorded greater-than-expected improvement in exercise duration in the exercise training–alone group. Such greater-than-expected improvement in exercise duration after exercise training alone contributed to the lack of statistical difference.

Third, it is possible that treadmill tests may be limited, in part, by the ability of patients to move their feet. If correct, a learning effect in how to coordinate foot movement during treadmill exercise (i.e., improved efficiency with treadmill walking) might translate into large improvements in exercise duration.

Finally, the similar exercise duration with VF plus exercise training and exercise training alone could result from a true lack of benefit in adding VF to exercise training. The available data, however, cannot be used to accept or refute the lack of additive effect of VF to exercise training. At the conclusion of the study, using the ANCOVA model with a three-group design, we calculated that 17 patients per group were required to reach a statistical significance of α = 0.0167. As a consequence of dropout after randomization, this target was achieved in the VF plus exercise group, but not in the exercise alone (n = 16) or VF alone (n = 16) groups.

At the conclusion of the trial, exercise duration in the VF-alone group improved substantially. One patient in this group had a 36-minute improvement in exercise time, which was much larger than that recorded in the other patients assigned to this group (4.82 ± 3.85 min). This greater increase could have resulted from the low-intensity physical activity incorporated into each training session, and intended as a means for patients to practice the respiratory technique while active. These short bouts of very light cycling (first 18 training sessions) and walking (last 18 training sessions) may have been sufficient to elicit an exercise training effect in the skeletal muscles of the participants who were physically unfit at baseline. We also cannot rule out that this improvement may be the result of a Hawthorne effect or improved efficiency with treadmill walking.

Patient Phenotype and Response to Training

Patients with more severe obstructive disease are at higher risk of developing dynamic hyperinflation during exercise (13). Therefore, to assess whether the addition of VF to exercise training may have greater benefit in patients with more severe obstructive disease, we divided the sample according to FEV1 values: patients with an FEV1 ⩽ 50%, and patients with an FEV1 > 50%. After adjusting for adherence and comorbidities, the difference in exercise duration at the end of training between exercise plus VF and exercise alone was 52% (P = 0.027) in patients with an FEV1 ⩽ 50% and 15% (P = 0.56) in patients with an FEV1 > 50%. This post hoc analysis suggests that patients with more severe airway obstruction may benefit the most from the combination of exercise training plus VF training compared with patients with less severe airway obstruction.

Training and Respiratory Pattern

During the constant work-rate test (performed at the end of training), the computer monitor used for VF was not displayed to the patients; nevertheless, the respiratory pattern recorded at isotime was affected by randomization. Isotime V̇e with VF plus exercise training and with VF alone decreased by 23 and 15% from the respective baseline values (Table 3). This reduction in V̇e resulted from a decrease in respiratory rate (24% decrease in VF plus exercise training and 13% decrease in VF alone) (Table 3). The decrease in isotime V̇e of the two VF groups is greater than the post-training decreases in V̇e reported in the literature (23, 25, 26). In the two VF training groups, the increase in time of exhalation was three times greater than the increase with exercise training alone. These results indicate that changes in respiratory pattern were specific to the conditions of training (the nonsignificant 7% decrease in isotime V̇e with exercise training alone was similar to the results of other investigators [25]).

Work of breathing was not measured in this investigation. However, the combined reduction in respiratory rate and dynamic hyperinflation suggests that, at isotime, the decrease in dynamic compliance that accompanies increases in respiratory rate was less in the VF plus exercise group than in the other two groups. More favorable pulmonary mechanics (smaller reduction in exercise-associated dynamic compliance) and more favorable length–tension relationships of the inspiratory muscles (less dynamic hyperinflation)—while maintaining similar Vt—could have decreased work of breathing and increased the efficiency of oxygen use by the respiratory muscles in patients randomized to VF plus exercise training. Such possible decreases in work of breathing and increased efficiency in oxygen consumption could have contributed to prolongation of exercise duration.

At isotime, exercise-induced dynamic hyperinflation decreased only in patients randomized to VF plus exercise training—25% increase in inspiratory capacity over that at baseline (Table 3). Mechanisms responsible for this decrease in dynamic hyperinflation at the isotime may include training-specific effects on peripheral muscle metabolism and respiratory pattern. To perform the same work (definition of isotime) after training, patients in the VF plus exercise group required less V̇e than did the other two groups. This finding suggests a more efficient use of oxygen delivered to the peripheral muscles. A more efficient use of oxygen could have resulted from a more pronounced training-associated metabolic shift of skeletal muscle fibers: from glycolytic to aerobic (37, 38). A second potential mechanism is the specific effect of VF on respiratory pattern. After training, patients in the VF plus exercise training demonstrated a significant increase in Te and the largest decrease in respiratory rate (Table 3). A slower respiratory rate could have improved the matching of ventilation and perfusion—and, thus, decreased deadspace ventilation—and could have reduced ventilatory needs. Whether the purported decrease in deadspace ventilation could have been sufficient to decrease exercise-induced alveolar hypoventilation (12) remains to be determined. That a similar respiratory response was not observed with VF alone may have resulted from higher metabolic demands (less oxidative shift in the skeletal muscles) than after VF plus exercise training. The additive effect of combining exercise training and VF training in terms of dynamic hyperinflation may be an essential consideration if VF training is to be employed in pulmonary rehabilitation.

Training and Perceived Dyspnea during Exercise

After training, isotime dyspnea scores decreased in all three groups. Contrary to our expectations, dyspnea scores after VF plus exercise training were not significantly different from those after exercise training alone. This unexpected result may have resulted from the high interindividual variability of dyspnea ratings—coefficient of variation ranged from 62 to 106% among individuals—or a type II error. Contrary to our expectations, dyspnea scores after VF plus exercise training also did not differ from those after VF alone. Potential mechanism responsible for these results include a reduction in dyspnea after VF training alone, high interindividual variability of dyspnea ratings, and training of the peripheral muscles by the light exercise that was part of the VF training alone. The reduction in dyspnea with respiratory retraining alone is consistent with results of other respiratory retraining programs (39). Although not tested, the observed reduction of dyspnea after VF training alone raises the possibility that VF training alone may be useful as a tool to reduce dyspnea in patients unable to participate in formal exercise training programs.

Quality of Life

At the conclusion of the study, all three groups of patients reported clinically significant decreases in perceived dyspnea during activities of daily living. Improvements in dyspnea scores at the conclusion of the two protocols that included exercise training (VF plus exercise training and exercise training alone) support the results of earlier investigations (4042). To our knowledge, this is the first study to demonstrate an association between respiratory retraining alone and a decrease in perceived dyspnea during activities of daily living. Two non–mutually exclusive mechanisms may be responsible for this positive finding. First, VF training alone could have been sufficient to modify respiratory pattern during activities of daily living, enough to decrease perceived dyspnea. This possibility is supported by the changes in respiratory pattern recorded at the conclusion of training in patients randomized to VF alone (Table 3). Second, and as already discussed, the low-intensity physical activity incorporated into each training session of patients in the VF alone group may have improved the physical fitness of these patients.

In contrast to improvements in dyspnea scores, most of the improvements in fatigue, emotional function, and mastery did not reach statistical significance. It may be that the lack of improvement in the domains pertaining to fatigue, emotional function, and mastery result from a true lack of benefit when VF is combined with exercise training. Alternatively, we may have underestimated the improvements in fatigue, emotional function, and mastery. Underestimation of the improvements may have resulted from the strategy chosen to administer the Chronic Respiratory Disease Questionnaire (see online supplement for details). Unlike the procedures recommended by Guyatt (43), participants in the current investigation were not informed of their previous responses. This strategy was chosen to avoid or limit any bias that might occur from the patients' desire to please the investigators with their responses. Another mechanism for this lack of statistical difference may be a type II error.

Adherence and Drop Out

Adherence was similar across the three groups, with most patients completing the program within the 12–14 weeks allotted (76.5% in exercise training with VF group and 75% in the exercise-alone and VF-alone groups). Therefore, that some patients took longer than expected to complete the protocol resulted from factors outside of the research intervention. Three patients took more than 20 weeks to finish the program: two of these had problems with transportation and work commitments; the third patient developed symptomatic gastroesophageal reflux between the cycling and treadmill phases of the study and was allowed to continue when the symptoms subsided. We cannot exclude the possibility that prolonging the training period may decrease the extent of training effect. The drop-out/withdrawal rate of patients in this study was similar to that reported by other investigators (42, 44, 45).

Limitations

About 50% of patients with severe COPD become rapidly intolerant of exercise secondary to early dyspnea or limitation in ventilation (8). These patients do not sufficiently activate their leg muscles and, at the end of exercise, do not develop peripheral muscle fatigue (8). Such an observation raises the possibility that patients who do not develop peripheral muscle fatigue could be the ideal candidates for the combination of exercise training plus VF training. Although absence of peripheral muscle fatigue after exercise was not one of our entry criteria, we were still able to record greater increases in exercise duration after VF plus exercise training than after VF alone, and a tendency toward longer exercise duration after VF plus exercise training than after exercise training alone. We cannot, however, exclude the possibility that superior results might be obtained with VF plus exercise training if the study was confined to patients who do not develop peripheral muscle fatigue after exercise.

In the first cohort of patients enrolled in the study (n = 13), inspiratory capacity data were lost because of malfunction of a computer hard drive. Despite this unforeseen decrease in the number of observations, we were able to demonstrate an increase in inspiratory capacity at isotime in the VF plus exercise training group. Finally, whether the endurance and strength of the inspiratory muscles were improved by VF training and thus contributed to improved exercise capacity, even in patients randomized to the VF-alone group, remains to be determined.

Conclusions

When performing a submaximal constant work-rate treadmill test, patients with COPD who completed a training program of VF plus exercise training exercised for longer than did patients who underwent VF training alone. After VF plus exercise training, the improvement in exercise duration over exercise training alone was not significant according to predetermined statistical criteria. Patients training with VF plus exercise training developed less exercise-induced dynamic hyperinflation at isotime than did patients after VF alone or after exercise training alone. Finally, at the end of training, patients randomized to VF training, with or without concurrent exercise training, modified their respiratory pattern. The potential improvement in exercise duration by combining exercise training and VF training identified in this study clearly merits further investigation in future, large, prospective trials.

1. Casaburi R. Exercise training in chronic obstructive lung disease. In: Casuburi R, Petty TL, editors. Principles and practice of pulmonary rehabilitation. Philadelphia: W.B. Saunders; 1993. pp. 204–224.
2. Norweg AM, Whiteson J, Malgady R, Mola A, Rey M. The effectiveness of different combinations of pulmonary rehabilitation program components. Chest 2005;128:663–672.
3. Casaburi R, Bhasin S, Cosentino L, Porszasz J, Somfay A, Lewis MI, Fournier M, Storer TW. Effects of testosterone and resistance training in men with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2004;170:870–878.
4. Oga T, Nishimura K, Tsukino M, Sato S, Hajiro T, Mishima M. Exercise capacity deterioration in patients with COPD: longitudinal evaluation over 5 years. Chest 2005;128:62–69.
5. Troosters T, Casaburi R, Gosselink R, Decramer M. Pulmonary rehabilitation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2005;172:19–38.
6. Calverley PM, Walker P. Chronic obstructive pulmonary disease. Lancet 2003;362:1053–1061.
7. Gosselink R, Trooster T, Decramer M. Peripheral muscle weakness contributes to exercise limitation in COPD. Am J Respir Crit Care Med 1996;153:976–980.
8. Saey D, Debigare R, LeBlanc P, Mador MJ, Cote CH, Jobin J, Maltais F. Contractile leg fatigue after cycle exercise:a factor limiting exercise in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2003;168:425–430.
9. Roger N, Barbera JA, Roca J, Rovira I, Gomez FP, Rodriguez-Roisin R. Nitric oxide inhalation during exercise in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1997;156:800–806.
10. Eiser N, West C, Evans S, Jeffers A, Quirk F. Effects of psychotherapy in moderately severe COPD: a pilot study. Eur Respir J 1997;10:1581–1584.
11. Hussain O, Langbein WE, Choi Y, Bammert C, Collins E, Tobin MJ, Laghi F. Is the mechanism responsible for the improvement in exercise duration with non-invasive ventilation and with a low-density gas mixture in patients with COPD the same? [abstract]. Am J Respir Crit Care Med 2003;167:A186.
12. O'Donnell DE, D'Arsigny C, Fitzpatrick M, Webb KA. Exercise hypercapnia in advanced chronic obstructive pulmonary disease: the role of lung hyperinflation. Am J Respir Crit Care Med 2002;166:663–668.
13. O'Donnell DE, Revill SM, Webb KA. Dynamic hyperinflation and exercise intolerance in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;164:770–777.
14. Sinderby C, Spahilja J, Beck J, Kaminski D, Yan S, Comtois N, Sliwinski P. Diaphragm activation during exercise in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163:1637–1641.
15. Sliwinski P, Kaminski D, Zielinski J, Yan S. Partitioning of the elastic work of inspiration in patients with COPD during exercise. Eur Respir J 1998;11:416–421.
16. O'Donnell DE, Lam M, Webb KA. Measurement of symptoms, lung hyperinflation, and endurance during exercise in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998;158:1557–1565.
17. Koulouris NG, Dimopoulou I, Valta P, Finkelstein R, Cosio MG, Milic-Emili J. Detection of expiratory flow limitation during exercise in COPD patients. J Appl Physiol 1997;82:723–731.
18. Casaburi R, Porszasz J, Burns MR, Carithers ER, Chang RSY, Cooper CB. Physiologic benefits of exercise training in rehabilitation of patients with severe chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1997;155:1541–1551.
19. Collins EG, Fehr L, Bammert C, O'Connell S, Laghi F, Hanson K, Hagarty E, Langbein WE. Effect of ventilation–feedback training on endurance and perceived breathlessness during constant work-rate leg-cycle exercise in patients with COPD. J Rehabil Res Dev 2003;40:35–44.
20. Collins E, Langbein W, Fehr L, Laghi F, Maloney C, O'Connell S, Edwards L, Hagarty E, Tobin MJ. ventilation–feedback effect on exercise capacity in COPD. Am J Respir Crit Care Med 2002;165:A737.
21. Guyatt GH, Berman LB, Townsend M, Pugsley SO, Chambers LW. A measure of quality of life for clinical trials in chronic lung disease. Thorax 1987;42:773–778.
22. Laghi F, Langbein W, Antonescu-Turcu A, Jubran A, Bammert C, Tobin M. Respiratory and skeletal muscles in hypogonadal men with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2005;171:598–605.
23. Emtner M, Porszasz J, Burns M, Somfay A, Casaburi R. Benefits of supplemental oxygen in exercise training in nonhypoxemic chronic obstructive pulmonary disease patients. Am J Respir Crit Care Med 2003;168:1034–1042.
24. Ortega F, Toral J, Cejudo P, Vallagomez R, Sanchez H, Castillo J, Montemayor T. Comparison of effects of strength and endurance training in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2002;166:669–674.
25. O'Donnell DE, McGuire M, Samis L, Webb KA. General exercise training improves ventilatory and peripheral muscle strength and endurance in chronic airflow limitation. Am J Respir Crit Care Med 1998;157:1489–1497.
26. Porszasz J, Emtner M, Goto S, Somfay A, Whipp BJ, Casaburi R. Exercise training decreases ventilatory requirements and exercise-induced hyperinflation at submaximal intensities in patients with COPD. Chest 2005;128:2025–2034.
27. Collins EG, Langbein WE, Fehr L, Maloney C. Breathing pattern retraining and exercise in persons with chronic obstructive pulmonary disease. AACN Clin Issues 2001;12:202–209.
28. Gigliotti F, Romangnoli I, Scano G. Breathing retraining and exercise conditioning in patients with chronic obstructive pulmonary disease (COPD): a physiological appropach. Respir Med 2003;97:197–204.
29. Collins E, Langbein W, Fehr L, Bammert C, O'Connell S, Hanson K, Edwards L, Hagarty E, Tobin MJ. Can ventilation–feedback training decrease exertional dyspnea and improve exercise tolerance in patients with COPD? Am J Respir Crit Care Med 2003;167:A314.
30. Vallet G, Ahmaide S, Serres I, Febre C, Bourgouin D, Desplan J, Varray A, Prefaut C. Comparison of two training programmes in chronic airway limitation patients: standardized versus individualized protocols. Eur Respir J 1997;10:114–122.
31. Neder JA, Jones PW, Nery LE, Whipp BJ. Determinants of the exercise endurance capacity in patients with chronic obstructive pulmonary disease: the power–duration relationship. Am J Respir Crit Care Med 2000;162:497–504.
32. Ries AL, Kaplan RM, Limberg TM, Prewitt LM. Effects of pulmonary rehabilitation on physiologic and psychosocial outcomes in patients with chronic obstructive pulmonary disease. Ann Intern Med 1995;122:823–832.
33. O'Donnell DE, D'Arsigny C, Webb KA. Effects of hyperoxia on ventilatory limitation during exercise in advanced chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163:892–898.
34. Bye PT, Esau SA, Levy RO, Shiner RJ, Macklem PT, Martin JG, Pardy RL. Ventilatory muscle function during exercise in air and oxygen in patients with chronic airflow limitation. Am Rev Respir Dis 1985;32:236–240.
35. Potter WA, Olafson S, Hyatt RE. Ventilatory mechanics and expiratory flow limitation during exercise in patients with obstructive lung disease. J Clin Invest 1971;50:910–919.
36. Eastwood P, Hillman D, Finucane D. Ventilatory responses to inspiratory threshold loading and role of muscle fatigue in task failure. J Appl Physiol 1994;76:185–195.
37. Mador MJ, Kufel TJ, Pineda LA, Steinwald A, Aggarwal A, Upadhyah AM, Khan MA. Effect of pulmonary rehabilitation on quadriceps fatiguability during exercise. Am J Respir Crit Care Med 2001;163:930–935.
38. Maltais F, Leblanc P, Simard C, Jobin J, Berube C, Bruneau J, Carrier L, Belleau R. Skeletal muscle adaptation to endurance training in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1996;154:442–447.
39. Carrieri-Kohlman V, Nguyen H, Donesky-Cuenco D, Demir-Deviren S, Neuhaus J, Stulbarg MS. Impact of brief or extended exercise training on the benefit of a dyspnea self-management program in COPD. J Cardiopulm Rehabil 2005;25:275–284.
40. Puhan MA, Guyatt GH, Goldstein R, Mador J, McKim D, Stahl E, Griffith L, Schunemann HJ. Relative responsiveness of the chronic respiratory questionnaire, St. Georges Respiratory Questionnaire and four other health-related quality of life instruments for patients with chronic lung disease. Respir Med 2007;101:308–316.
41. Green RH, Singh SJ, Williams J, Morgan MDL. A randomised controlled trial of four weeks versus seven weeks of pulmonary rehabilitation in chronic obstructive pulmonary disease. Thorax 2001;56:143–145.
42. Ferreira G, Feuerman M, Spiegler P. Results of an 8-week, outpatient pulmonary rehabilitation program on patients with and without chronic obstructive pulmonary disease. J Cardiopulm Rehabil 2006;26:54–60.
43. Guyatt GH. The chronic respiratory questionnaire: background information and interviewing suggestions. Hamilton, ON, Canada: McMaster University; 1986.
44. Ries AL, Make BJ, Lee SM, Krasna MJ, Bartels M, Crouch R, Fishman AP; for the National Emphysema Treatment Trial Research Group. The effects of the pulmonary rehabilitation in the national emphysema treatment trial. Chest 2005;128:3799–3809.
45. Johnson JE, Gavin DJ, Adams-Dramiga S. Effects of training with heliox and noninvasive positive pressure ventilation on exercise ability in patients with severe COPD. Chest 2002;122:464–472.
Correspondence and requests for reprints should be addressed to Eileen G. Collins, Ph.D. Research and Development (151), Edward Hines, Jr. VA Hospital, Hines, IL 60141. E-mail:

Related

No related items
American Journal of Respiratory and Critical Care Medicine
177
8

Click to see any corrections or updates and to confirm this is the authentic version of record