Factors Determining Constant Work Rate Exercise Tolerance in COPD and their Role in Dictating the Minimal Clinically Important Difference in Response to Interventions

Abstract

Constant work rate exercise testing has been used to assess effectiveness of therapeutic interventions in chronic obstructive pulmonary disease. It has been noted to yield larger fractional increases in exercise tolerance than other measures. Reasons for this are rooted in physiological determinants of the time course of pulmonary ventilation. Following exercise onset, ventilation increases in three phases; the slowest phase is seen only at high exercise intensities. In chronic obstructive pulmonary disease exercise proceeds until ventilation reaches a limiting value. Because both ventilatory requirement for exercise and limiting ventilation can be manipulated by several interventions (exercise training, oxygen inhalation, bronchodilator administration), constant work rate tolerance responds to these interventions. The power-duration relationship predicts the effect of the work rate imposed on tolerated duration of a constant work rate test. Arguments are presented that a pre-intervention constant work rate duration of 4–7 minutes is desirable. At present, the recommended strategy to achieve this target duration is to choose a work rate equal to 85% of peak work rate achieved in a constant work rate test. There is clearly insufficient information to reliably determine a minimal clinically important difference for this test. A lower bound of 1.75 minutes is suggested.

• COPD
• Exercise
• Ventilation
• Critical power
• Minimal clinically important difference

Constant Work Rate Exercise Testing as a Measure of Exercise Tolerance Improvement

Constant work rate exercise testing is one of a number of measures of exercise tolerance that can be used to assess the effectiveness of an intervention. It belongs to one of the three classes of exercise tolerance measures that have been used in clinical research: questionnaires [[1]], activity monitoring [[2]] and exercise testing. Questionnaires are generally easy to administer and are capable of assessing the perceptions of the patient regarding the effect of an intervention on symptoms of exercise intolerance; they are considered in-depth in other presentations in this symposium. Activity monitors are able to discern whether interventions actually yield increased levels of ambulation, surely a key goal of these interventions. However, activity monitoring has not yet achieved a greatdegree of standardization. Exercise testing modalities include timed walking tests (such as the 6-minute walk and the shuttle walk test), incremental cardiopulmonary exercise testing, and constant work rate exercise testing. These tests have advantages and disadvantages that have been commented on previously [3-5]. However, in the current context, it is important to point out that the fractional improvement in exercise tolerance for a given intervention is substantially greater for constant work rate testing than for the other two modalities. This was pointed out by Oba et al. [[6]], who determined the exercise tolerance of 38 COPD patients before and after bronchodilator administration utilizing three exercise modalities. Average 6-minute walk increase was 2%, peak work rate increase in the incremental test averaged 4%, while the increase in constant work rate duration averaged 20%. To understand why such large increases occur and to determine the advantages of constant work rate exercise testing in assessing responses to interventions, it is necessary to appreciate this test's physiologic determinants.

Physiologic Responses to Onset of Constant Work Rate Exercise

When exercise begins, a coordinated set of physiologic adaptations occurs that respond to the increased metabolic demands of exercise [[7]]. Muscle oxygen uptake and carbon dioxide output increase. Muscle blood flow and total cardiac output increase. Pulmonary vasodilation occurs, allowing increased perfusion of pulmonary capillaries. Pulmonary ventilation increases, allowing for greater gas exchange in the lungs. These adaptations do not reach their full steady-state requirement instantaneously; all rise with individual time courses. These individual time courses dictate the time courses of variables that are easily observed, e.g., oxygen uptake, CO₂ output, heart rate, and pulmonary ventilation. These rates of adaptation vary among individuals and, in particular, are slower in older individuals [[8]] and in individuals with chronic disease [9&10].

Importantly, the adaptations to exercise onset occur with different kinetic time courses depending on the intensity of the work rate imposed [[11]]. Three phases of kinetic response of ventilation and gas exchange have been defined. Phase I occurs within the first 20 seconds or so of exercise and is dictated, at least in part, by the abrupt change in cardiac output that occurs following exercise onset [though the mechanism(s) of the ventilatory response remain controversial [[12]]]. Phase II is a slower, exponential phase lasting roughly 2–4 minutes; the gas exchange responses are dictated by the kinetics of response of the metabolic products of exercise at the lung (i.e., deoxygenated blood with high carbon dioxide content). Phase II ventilatory response is closely linked to (and lags slightly behind) the kinetics of CO₂ output [[13]]. Phase III is seen only at heavy work rates and reflects the consequences of lactic acid production in the exercising muscles: a slow upward drift in oxygen uptake, extra CO₂ output from lactic acid buffering, and an acidification of the arterial blood. It constitutes a slowly rising phase generally commencing about 2 minutes after exercise onset [[14]]. The phase III component is generally steeper, the greater the lactic acidosis engendered by the exercise [15&16]. Importantly, the work rate at which lactic acidosis becomes prominent varies greatly among subjects and is lower in older and more sedentary individuals and can be markedly low in patients with chronic disease [[17]].

Limiting Factors to Constant Work Rate Exercise in Healthy Subjects and COPD Patients

The mechanisms leading to muscular fatigue in healthy subjects are a subject of controversy [[18]]. However, exercise does not continue much after a limiting oxygen uptake has been reached. The limiting oxygen uptake in constant work rate exercise is approximately the same as that observed in incremental tests [[19]]. The time to reach the limiting oxygen uptake during constant work rate exercise is, therefore, predominantly dictated by the kinetic response of oxygen uptake. For very high work rates, exercise limitation may occur while the subject is in phase II of the oxygen uptake response. However, for all but the highest work rates, exercise limitation occurs when the phase III oxygen uptake response brings the subject to the limiting oxygen uptake. Note that, while the phase II oxygen uptake response to a given work varies little among individuals (and is roughly 10 ml/min/Watt) [[7]], the work rate at which a phase III response is engendered and the steepness of the phase III response will vary markedly among individuals. These factors are therefore prime determinants of the tolerated duration of a given constant work rate.

COPD patients are conventionally believed to be limited in their exercise tolerance by the level of ventilation they can sustain [[20]], i.e., that ventilatory muscle fatigue rather than limb muscle fatigue dictates exercise limitation. This scenario implies that the kinetics of pulmonary ventilation rather than the kinetics of oxygen uptake is the key consideration dictating exercise tolerance to a constant work rate task. Indeed, studies have shown that the limiting time in constant work rate testing generally coincides with the point at which ventilation rises to equal the peak ventilation achieved in incremental exercise testing and/or the maximum voluntary ventilation achieved in studies performed at rest [[19]]. The factors dictating the ventilatory requirement for a given work rate are strongly determined by the inefficiency of pulmonary gas exchange, quantified as the ratio of dead space to tidal volume (V_D/V_T). Therefore the phase II steady-state response will be higher (and therefore closer to the limiting ventilation) the higher is the V_D/V_T. If the work rate engenders lactic acidosis, a superimposed phase III ventilatory response will also occur. The magnitude of the V_D/V_T and of the phase III response will dictate the point at which the limiting ventilation is reached and, therefore, the tolerated duration of a given work rate.

Research in recent years has shown that the limiting ventilatory response in COPD patients is inextricably linked to the occurrence of progressive hyperinflation as exercise proceeds [21&22]. Expiratory airflow obstruction limits the ability to exhale before the next breath must be initiated and full exhalation does not occur. As respiratory rate is driven upward and, therefore, expiratory time shortens, dynamic hyperinflation occurs. When the difference between the end-inspiratory lung volume and the volume at total lung capacity (the inspiratory reserve volume) reaches a critically low value, exercise limitation is reached [[23]]. This seems to be related to the amount of elastic work that is necessary to achieve lung volumes so close to total lung capacity.

Complicating these considerations is the recent finding that, even in patients with severe COPD, exercise tolerance may be limited primarily by fatigue of the muscles of ambulation rather than by ventilatory limitation [[24]]. Initial evidence for this concept is drawn from the subjective complaints of patients when performing symptom limited exercise; leg fatigue rather than dyspnea is often the limiting symptom [[25]]. In a recent study [[26]], half of a group of COPD subjects exhibited objective evidence of leg fatigue during constant work rate exercise. In these subjects, improving lung mechanics (via an inhaled bronchodilator) failed to improve exercise tolerance, suggesting that leg fatigue was, indeed, the primary factor limiting exercise tolerance.

The Power-Duration Relationship

If a subject performs a series of constant work rate tests covering a range of work rates, it has been found that the relationship between the work rate and the tolerated duration of the test has the form of a rectangular hyperbola [[27]]:where W is the work rate, t is the tolerated duration, Θ_F is the asymptotic work rate known as the “critical power,” the highest work rate that can be sustained indefinitely. W′ is the curvature constant of the hyperbola and represents the amount of work that can be performed above the critical power. Θ_F and W′ can be determined graphically if 1/t is plotted on the abscissa and W is plotted on the ordinate; the slope of this plot is W′ and the y-intercept is the critical power. In healthy subjects, the critical power is a work rate somewhat above the lactic acidosis threshold, and approximates the highest work rate at which blood lactate levels can achieve a steady (though elevated) level [[28]].

In COPD subjects, the critical power generally occurs at a lower work rate than in healthy subjects [[19]]. However, as a fraction of the peak work rate achieved in an incremental test, the critical power is higher than in healthy subjects [[19]]. This is consistent with the observation that COPD patients can sustain high fractions of the peak work rate observed in an incremental exercise test and that they, therefore, tolerate high relative exercise training intensities [[29]].

Utilizing Constant Work Rate Duration as an Outcome Measure in Interventional Studies in COPD

Strategies to improve the exercise tolerance of patients with COPD are actively sought. Constant work rate cycle ergometry has been found to be useful in assessing the effectiveness of such strategies. If a given work rate can be sustained for a longer period after an intervention, this is taken as evidence of improved exercise tolerance. Further, improvement in the physiologic ability to perform exercise is accompanied by a constellation of changes in physiologic response variables (e.g., reduced ventilation, CO₂ output, blood lactate) for a given exercise task (a given work rate performed for a specified period of time) [[30]]; these constitute effort independent evidence of improvement in exercise tolerance.

From the foregoing discussion, it is apparent that selection of the work rate for initial testing must be individualized to the patient's exercise tolerance. Considering the power-duration relationship for a given subject, if the work rate is too high, the work rate will be tolerated only for a short duration and improvements in the physiologic ability to exercise will result in only small improvements in exercise time. If the work rate is too low (below the critical power), then exercise can be continued indefinitely and will not be a helpful discriminant. Even for work rates mildly above the critical power, exercise durations will be long and any small improvement in the physiologic ability to exercise will yield essentially indefinite exercise durations; this severely limits the ability to discriminate large from small effect sizes. This principle is illustrated in Figure 1 based on simulated data yielding the power-duration curve for a healthy subject undergoing exercise training.

Figure 1. Effect of the choice of work rate in a constant work rate exercise test on the increase in exercise duration after an exercise training program. Responses are stimulated data, but are adapted from those presented by Poole et al. [[28]], who studied the power-duration relationship in healthy subjects before and after training. Curved lines in each panel are power-duration curves constructed from a series of constant work rate tests performed at a variety of work rates. Dashed curved line: before an exercise training program. Solid curved line: after a training intervention. Upper, middle, and lower panels show, respectively, the results of choosing a work rate of 270, 248, and 226 watts as the measure of constant work rate duration is short (about 3.3 minutes) and the increase in exercise tolerance engendered by training is only about 1.5 minutes. For 226 watts (lower panel), the initial exercise duration is long (about 10.5 minutes); after training this work rate is below critical power and can be tolerated indefinitely. In contrast, a work rate of 248 watts (middle panel) yields an initial duration of about 5.3 minutes and the increase in exercise tolerance is approximately 3.3 minutes.

We have asserted that exercise durations in the range of 4–7 minutes are desirable baselines for interventional studies [[31]]. The lower limit (4 minutes) is roughly the time for the phase II response to be complete and the phase III response (if any) to be apparent. Durations greater than the upper limit imply that the work rate is not much above the critical power; a modestly effective intervention may result in an indefinitely long work rate duration. Establishing the work rate that will achieve the goal of a 4–7-minute exercise duration is therefore crucial if constant work rate testing is to be an effective tool in interventional studies. Having the subject perform a series of work rates to establish the power-duration relationship is laborious and, for most purposes, impractical. We recently summarized a large experience in our laboratory that examined, in a COPD population, the relation between the fraction of the peak work rate achieved in an incremental exercise test used in a constant work rate and the tolerated duration of that constant work rate. Based on 251 pairs of studies, choosing a work rate of approximately 85% of the peak work rate in the incremental test yielded the best chance of achieving a test duration in the 4–7-minute range [[31]].

A number of studies in COPD patients can be cited that have used constant work rate testing to define the responseto an intervention posited to increase exercise tolerance. For example:

• O'Donnell et al. [[21]] have evaluated the effect of chronic tiotropium (a long-acting anticholinergic bronchodilator) therapy on constant work rate exercise tolerance in a group of 187 COPD patients in a double-blinded placebo-controlled trial. Baseline constant work rate exercise tolerance was approximately 8 minutes. Exercise tolerance was improved by an average of about 1.6 minutes more in the tiotropium group than in the placebo group.

• Somfay et al. [[22]] investigated the effect of breathing gas mixtures containing elevated fractions of oxygen on constant work rate exercise tolerance in COPD patients without clinically significant hypoxemia in a single-blinded randomized study. Average exercise tolerance breathing air was 4.2 minutes. Breathing 30% oxygen increased the average exercise duration to 7.8 minutes and 50% oxygen yielded an average duration of 10.3 minutes.

• Emtner et al. [[32]] determined the effect, separately and combined, of oxygen inhalation and high intensity exercise training in a group of 29 COPD patients. Half underwent training while respiring 3 L/min of supplemental oxygen by nasal cannula and half while receiving supplemental air (in a double-blinded fashion). At baseline (before training), while breathing air, constant work rate duration averaged 5 minutes. Breathing 30% oxygen increased exercise duration by 6 minutes. After training, constant work rate duration while breathing air averaged 20 minutes and while breathing 30% oxygen averaged 25 minutes. These improvements were somewhat underestimated as 30% of the air-breathing tests and 47% of the oxygen-breathing tests were terminated by the laboratory staff at 30 minutes. For these latter tests, the work rate imposed apparently no longer exceeded the critical power.

These studies illustrate the utility of constant work rate testing in assessing the relative magnitude of benefit of interventions designed to improve exercise tolerance. From these (and other) studies, the effectiveness of these interventions in increasing constant work rate duration can be ranked as: high intensity exercise training > supplemental oxygen > bronchodilator. Of course, the “costs” of these benefits must be considered in evaluating the desirability of an intervention. It might be argued that the difficulty in instituting these interventions is in precisely the opposite order as is their effectiveness.

Considerations for Specifying a Minimal Clinically Important Difference for Constant Work Rate Testing in COPD

Powering Studies Using a Constant Work Rate Test Duration Outcome

Knowledge of the standard deviation of change in a given outcome measure for a given intervention is essential foraccurate sample size estimates. From the preceding discussion, it seems quite likely that the standard deviation of the change in constant work rate test duration will increase the longer the initial test duration and the larger the effect size. This likely makes this measure of exercise tolerance more difficult to incorporate into clinical trials than some other laboratory exercise tests. However, other considerations, principally the large fractional change in test duration and the opportunity to observe non-effort dependent outcomes may, nonetheless, make constant work rate testing an attractive measure. However, for tests with baseline durations in the desired range, it should be possible to estimate variance of the test. For the study of O'Donnell et al. [[21]], where baseline exercise duration was approximately 8 minutes, the standard deviation of change in exercise duration among subjects receiving placebo was approximately 3.3 minutes. In a study of testosterone and/or strength training in 47 COPD subjects [[33]], baseline constant work rate exercise duration averaged 6.0 minutes. Neither strength training nor testosterone was associated with appreciable increase in constant work rate duration; the standard deviation of change in exercise duration among subjects was 2.9 minutes (unpublished data). Therefore, as a preliminary estimate, assuming a standard deviation of change on the order of 3 minutes might be appropriate for constant work rate testing when the initial duration of the test was in the desirable range and the intervention did not increase exercise duration by a very large amount.

Toward a MCID

As a first step, it is worthwhile considering the magnitude of constant work rate increase seen in COPD interventions. Since exercise intolerance is often a chief compliant of patients with COPD, it would reasonably be expected that increasing exercise tolerance by an amount that was clinically significant would yield clinically significant increases in other outcomes. In particular, health-related quality of life measures might be considered a reasonable yardstick. In this disease, chronic bronchodilator administration yields QOL increases that are generally near the MCID [34&35]. Chronic oxygen therapy generally yields QOL increases above the MCID [[36]]. Rehabilitative exercise training generally yields QOL increases clearly above the MCID [37&38]. The preliminary evidence cited above suggests that a lower bound for the MCID might be considered as 1.75 minutes (that associated in one trial [[21]] with bronchodilator administration). However, since bronchodilator administration is associated with improvements in quality of life measures that only intermittently exceed the MCID, it is possible that an appreciably greater improvement in constant work rate exercise tolerance might be identified as the minimum clinically important difference as further evidence accumulates (Table 1).

Table 1.  MCID Data Table.

Name of outcome measure	Constant work rate duration
Range	?
SEM	N/A
Reliability	?
Effect size	?
Distribution-based MCID	½ SD approx. 1.5 minutes
Anchor-based MCID	Lower limit approx. 1.75 minutes
Expert physician-based MCID	N/A
Patient-based MCID	N/A
Your suggested MCID	Lower limit approx 1.75 minutes

It might be noted that the “half standard deviation method” of estimating the MCID [[39]] yields a value on the order of 1.5 minutes, only mildly below our current best estimate of the lower bound of 1.75 minutes.