Comparison of free range and graded exercise testing

Abstract

Previous studies with athletes have demonstrated greater physiologic responses during free range (FR) compared with graded (GXT) exercise testing. Since the sensitivity of clinical exercise testing depends upon the magnitude of physiologic responses, we sought to determine whether FR might provoke greater responses than GXT in nonathletic individuals and patients. Healthy, physically active nonathletes and clinically stable CHD patients (N= 12) performed GXT on cycle ergometer (15 W + 15 W·min^-1) and FR (minimal time for 75 kJ task) on a cycle ergometer. A starting power output was recommended for FR, but the patients were free to pedal at their own rates. During FR, ˙VO_2max (36.5 ± 10.1 vs 34.1 ± 9.4 mL·min^-1·kg^-1), HR_max (156 ± 25 vs 144 ± 27 beats·min^-1), double product (31.4 ± 4.9 vs 29.1 ± 5.9) and ˙VE_max (111 ± 26 vs 94 ± 17 L·min^-1) were all significantly greater than during cycle GXT. The mean peak power output during GXT (180 ± 45 W) was not significantly different than the mean power output during FR (204 ± 45 W). During FR, successive “0.5 mile laps” (≈12.5 kJ) were accomplished at power outputs of 217 ± 45, 217 ± 52, 192± 60, 194 ± 65, 199 ± 63, and 207 ± 63 W. No patient experienced angina or ECG changes during either FR or GXT. The patients uniformly reported that FR felt like “hurrying” in the real world. Some patients had to make large reductions in their power output in mid ride to allow recovery from a too aggressive start, much as they would in the real world. We conclude that FR exercise provides a clinically useful method of exercise testing that is not only more like real world exercise patterns but also provokes greater physiologic responses than are achievable during conventional GXT.

The sensitivity of exercise testing in the diagnosis of cardiovascular and respiratory disease depends in part on the magnitude of the physiologic stress achieved. Previous studies with athletes from our laboratory have demonstrated greater physiologic responses during exercise testing (˙VO₂, HR, VE, HLa) when the subjects were free to choose their own pattern of exercise(i.e., ride 5 km as fast as possible) (Free Range = FR) compared with incremental exercise testing (GXT) ⁽¹¹⁾.

At the present time the clinical use of FR exercise is primarily submaximal exercise in patients with very low exercise capacity (pulmonary disease, heart failure)^{(5,6,14-16,22,23,30,36)} in fitness testing ^(7,8,21) and in some applications with athletes^{(4,11,12,17,18,26)}. Extension of this ergometric technique to conventional diagnostic exercise testing has not been accomplished. Accordingly, the purpose of the present study was to compare the physiologic responses during FR exercise on the cycle ergometer to incremental GXT in nonathletic controls and clinically stable patients.

METHODS

Subjects. The subjects for this study were 12 volunteers. All provided informed consent before participation. Five were patients with documented CHD (participants in the maintenance phase of our cardiopulmonary rehabilitation program) and seven were moderately trained nonpatients. None of the subjects demonstrated electrocardiographic evidence of ischemia during screening exercise tests. On the average (± SD) the subjects were 47.3± 12.8 yr of age, 175 ± 5 cm in height, and 78.3 ± 10.2 kg in weight. Three patients had prior documented myocardial infarction, two had prior CABG surgery, and one had prior PTCA. Three patients were taking beta blockers and one an ACE inhibitor. All patients remained on prescribed medications throughout the study.

Protocol. Following prior screening/familiarization exercise tests on the treadmill, each subject performed two randomly ordered maximal exercise tests. One test (Cycle GXT) was performed on an electrically braked cycle ergometer. The initial power output was 15 W and was incremented by 15 W·min^-1 until volitional fatigue. The other test (Free Range) was performed on an air braked ergometer (Randall Windracer). Following a 5-min incremental warm-up to a workload of 50% of the estimated maximal power output and a 1 min recovery, the subject was required to complete a distance of “3 miles” on the ergometer video display which we calculate to equal ≈75 kJ. The subject was instructed to complete the distance as quickly as possible, as if they were hurrying to go somewhere on their own bicycle. A starting power output was recommended (equal the age predicted maximum). However, the subjects was free to choose their own patterns for completing the distance, including slowing down as necessary to deal with fatigue. The time required to complete each “0.5 mile” or 12.5 kJ“lap” was recorded as well as the time required to complete the entire task.

Throughout the tests the ECG was monitored and recorded at appropriate intervals. BP was measured by auscultation at the conclusion of each minute of exercise during the incremental cycle ergometer and treadmill protocols and at the end of each 12.5 kJ “lap” during the FR protocol. Pulmonary oxygen uptake was measured throughout the test by open circuit spirometry(Quinton Q-Plex, Seattle, WA). The comparative results (GXT vs FR) for power output, ˙VO₂, HR, BP, HR·BP, and ˙VE were compared using ANOVA. Post hoc comparisons were made using the Scheffe test.

RESULTS

The subjects universally reported that FR felt more like their pattern of activity in the “real world” compared with either cycle or treadmill GXT. The observed serial changes in power output HR and˙VO₂ during FR were similar to previous observations during time trials (simulated competition) in athletes (Fig. 1). Successive “0.5 mile laps” (≈12.5 kJ) were accomplished at mean power outputs of 217 ± 45, 217 ± 52, 192 ± 60, 194± 65, 199 ± 63, and 207 ± 63 W. In many cases the subjects had to change power output significantly over short periods of time to deal with fatigue (Fig. 2). The overall average power output during FR (204 ± 52 W) was not significantly different than the maximal power output attained during GXT (180 ± 45 W). The correlation between average power output during FR and peak power output during GXT was high (r = 0.969) (Fig. 3).

The observed values for ˙VO_2max (36.5 ± 10.1 vs 34.1± 9.4 mL·min^-1·kg^-1), HR_max (156± 25 vs 144 ± 27 beats·min^-1), HR·SBPmax(31.4 ± 4.9 vs 29.1 ± 5.9), and maximal ˙VE (111 ± 26 vs 94 ± 17 L·min^-1) were all significantly (P< 0.05) greater during FR than during GXT. Although the SBPmax was not significantly different (200 ± 22 vs 196 ± 23 mm Hg), every individual had an equal or higher SBP during FR compared with GXT(Fig. 4).

The relationship between peak power output during GXT and˙VO_2max and between mean power output during FR and˙VO_2max was similar (Fig. 5). However, there was a poor relationship between peak power output during FR and˙VO_2max.

DISCUSSION

One of the primary purposes of exercise testing is to allow provocation and evaluation of a patient's symptoms under controlled conditions in the presence of the technological resources to document the nature of the pathophysiology contributing to the symptoms. In our experience (unpublished), it is often comparatively difficult to reproduce symptoms of angina or dyspnea during GXT even in patients whose presentation is symptomatically consistent with angina. This problem is reflected in the known variability in the exercise duration and HR at the anginal threshold in control tests administered one week apart⁽³²⁾ Although the warm-up phenomenon is often observed with multiple repetitions of an exercise test^(19,34), it is possible that the considerable duration of submaximal exercise during the early moments of GXT may effectively warm patients up and serve to mask ischemic changes that might occur with more abrupt exercise. This phenomenon is well documented during more prolonged exercise as the “walk through” effect⁽²⁵⁾. Previous studies from our laboratory have demonstrated left ventricular dysfunction in response to abrupt increases in muscular power output ^(9,10). Using a similar protocol, Barnard had demonstrated electrocardiographic and myocardial perfusion changes consistent with subendocardial ischemia^(2,3). Thus, it may be that the nature of many GXT protocols serve to mask exertional ischemia. This suggestion is supported by observations that more rapidly incremented treadmill tests are more sensitive in detecting residual ischemic abnormalities 6 wk following myocardial infarction ⁽³¹⁾. Abrupt presentation of near maximal loads has been discouraged over at least the last generation on the basis that it might present safety problems ^(1,29). However, it is probably fair to say that a high percentage of exercise in the real world displays rapid on-off transitions. Without documentation of increased risk during more abrupt loading patterns, perhaps the loss of diagnostic sensitivity with conventional GXT overbalances the theoretical increase in risk. The data of Stein et al. ⁽³³⁾ where patients with stable cardiovascular disease (and taking beta adrenergic blocking medications) performed sudden strenuous exercise similar to our^(9,10) previous studies suggested no particular increase in test related complications. Additionally, the substantial accumulated experience with the 6-min walk test (effectively a free range test) in patients with heart failure and/or pulmonary disease does not indicate a particularly high risk of complications. Lastly, the brief submaximal stages that were performed at the beginning of our free range protocol should identify patients with strongly abnormal exercise responses before the onset of near maximal FR exercise and indicate which patients should not perform FR exercise, as well as providing some measure of physiologic warm-up.

Particularly during cycle ergometry but also during uphill treadmill walking, it is common for patients to complain about local muscular fatigue. This is most likely related to accumulation of muscle lactate. The sensation of effort in the leg has been shown to be related to both muscle and blood lactate accumulation ^(24,28). Given our knowledge of the interplay between muscle and blood lactate, it is easy to conceptualize that during rapidly incremented protocols muscle lactate accumulates to some critical concentration which requires that exercise be stopped, even without particularly high values of blood lactate ⁽²⁰⁾. In a FR type of protocol, the ability to make brief reductions in power output may allow lactate to diffuse from muscle to blood and delay the moment when a critical muscle lactate concentration is achieved. Direct experimental evidence for this is lacking. However, the higher post exercise blood lactate concentrations during free range exercise ⁽¹¹⁾, as well as our understanding of diffusion kinetics of muscle to blood lactate⁽¹³⁾, would support this concept. In this study all subjects were comparatively familiar with cycle exercise, although none trained extensively using the cycle. In view of the well described differences between treadmill and cycle exercise testing, it might be expected that greater levels of physiological response would have been achieved during treadmill exercise. Preliminary studies in our laboratory (unpublished) have suggested that FR cycle exercise yields about the same ˙VO_2max as observed during incremental treadmill exercise. Future studies await our ability to appropriately adjust treadmill exercise to the FR concept so that systematic comparisons may be made.

Previous studies in patients with heart failure comparing performance during a 9-min self-powered treadmill test to conventional incremental treadmill exercise demonstrated somewhat lower values of ˙VO_2max during the self powered treadmill test ⁽³⁶⁾. However, there are significant differences between exercising as hard as one can for a certain duration compared with attempting to complete a fixed task in minimal time (i.e., hurrying to finish). Studies in which patients have been“encouraged” versus not encouraged during free range walking have demonstrated a significant difference in performance and reinforces the importance of external motivation on the physiologic responses during exercise⁽¹⁴⁾. Given the format of the current ergometer which displays distance completed in an easy to follow format, an approach to using external motivation by asking the patient to “hurry up and finish” is feasible. The current protocol used cycle ergometry rather than treadmill testing which is the more common method of clinical exercise evaluation. This was done because it was simpler to control the FR protocol with a commercially available cycle ergometer. Ordinary motor driven treadmills present the problem of how to change workloads, an appropriate “distance” or total amount of work for the trial, and the confounding attributable to handrail support. Should an extension of the current cycle ergometer protocol prove to be clinically useful, future studies should explore the feasibility of converting the protocol to treadmill exercise.

In this study pulmonary gas exchange was measured in an attempt to directly quantitate the physiologic response to the differences in protocol. However, the similarity of the relationship between mean power output during FR and˙VO_2max during FR on one hand and between peak power output and˙VO_2max during GXT on the other suggest that predicting functional capacity may be nearly as accurate with the FR protocol as with conventional GXT. We did not systematically note the presence of the ventilatory threshold during FR. For convenience, we chose to use only 5 min of submaximal exercise to a power output of about 50% of the anticipated maximal. However, if direct gas exchange studies were to be made, it would be only a small modification of the protocol to continue the submaximal portion of FR until a clear ventilatory threshold was observed, give the patient a minute of easy pedaling to recover, then conduct the FR protocol. In this way the important information regarding both peak (˙VO_2max) and sustainable(ventilatory threshold) exercise capacity could be acquired as easily during FR as during conventional GXT.

The results of this study demonstrate that greater physiologic responses can be achieved during FR exercise than during conventional GXT in nonathletic individuals and patients with stable CHD. As such they extend our previous observations and support the concept that FR may be a superior method of clinical evaluation of patients with cardiovascular or respiratory disease. Also of significance relative to understanding the present data is the nature of the subjects. Even though the subjects in the present study were not systematically trained athletes and some had significant cardiovascular disease, all were relatively active. Future studies need to focus on sedentary individuals and patients referred for diagnostic studies to determine whether the greater physiologic responses achievable with FR allow for improved symptom provocation and increased diagnostic sensitivity. Further, given our knowledge of the association between unaccustomed heavy exertion and the onset of acute MI ^(27,35), future studies will need to document the safety of FR relative to incremental GXT.

Figure 1-Serial changes in power output, heart rate, and˙VO2 during GXT (•) and FR (○) in a representative patient. In this particular example, the duration of both procedures was identical. Often the FR protocol is completed somewhat more quickly than the GXT protocol: .

Figure 2-Serial changes in power output in individual subjects (thin lines) and for the group average (thick line) for the subjects during the FR protocol: .

Figure 3-Relationship between the peak power output during conventional GXT and average power output during the FR exercise test. The heavier line represents the line of identity, the thinner line the best fit relationship between observed GXT vs FR values for power output: .

Figure 4-Individual (thin lines) and mean (square with thick line) values for ˙VO2, HR, and pulmonary ventilation during GXT and FR cycle ergometer exercise: .

Figure 5-Relationship between peak power output during GXT and˙VO2max during GXT (closed circles), between mean power output and˙VO2max during FR (○), and between peak power output and˙VO2max during FR (□): .