Diabetes Can Impair Exercise Capacity: Abnormal oxygen uptake kinetic responses in women with type II diabetes mellitus
J Appl Physiol 85: 310-317, 1998
"To summarize, the results of the present study demonstrate that premenopausal women with uncomplicated diabetes have impaired VO2 responses to maximal and submaximal exercise...In the present study, we found that women with type II DM had impaired maximal and submaximal cardiopulmonary responses to exercise, even though they had no evidence of clinical cardiovascular disease or diabetic complications. We had previously observed a lower VO2 response to submaximal workloads during submaximal graded exercise testing in persons with type II DM compared with controls"
Judith G. Regensteiner1,2,3, Timothy A. Bauer1, Jane E. B. Reusch4, Suzanne L. Brandenburg1, Jeffrey M. Sippel5, Andria M. Vogelsong1, Susan Smith1, Eugene E. Wolfel3, Robert H. Eckel4, and William R. Hiatt1,6
1 Section of Vascular Medicine, Divisions of 2 Internal Medicine, 3 Cardiology, 4 Endocrinology and 6 Geriatrics, Department of Medicine, University of Colorado Health Sciences Center, Denver, Colorado 80262; and 5 Division of Pulmonary Medicine, University of Oregon Health Sciences Center, Portland, Oregon 97201
Persons with type II diabetes mellitus (DM), even without cardiovascular complications have a decreased maximal oxygen consumption (VO2 max) and submaximal oxygen consumption (VO2) during graded exercise compared with healthy controls. We evaluated the hypothesis that change in the rate of VO2 in response to the onset of constant-load exercise (measured by VO2-uptake kinetics) was slowed in persons with type II DM. Ten premenopausal women with uncomplicated type II DM, 10 overweight, nondiabetic women, and 10 lean, nondiabetic women had a VO2 max test. On two separate occasions, subjects performed 7-min bouts of constant-load bicycle exercise at workloads below and above the lactate threshold to enable measurements of VO2 kinetics and heart rate kinetics (measuring rate of heart rate rise). VO2 max was reduced in subjects with type II DM compared with both lean and overweight controls (P < 0.05). Subjects with type II DM had slower VO2 and heart rate kinetics than did controls at constant workloads below the lactate threshold. The data suggest a notable abnormality in the cardiopulmonary response at the onset of exercise in people with type II DM. The findings may reflect impaired cardiac responses to exercise, although an additional defect in skeletal muscle oxygen diffusion or mitochondrial oxygen utilization is also possible.
IT HAS PREVIOUSLY BEEN OBSERVED that persons with type II diabetes mellitus (DM), even in the absence of clinical cardiovascular disease, have a reduced maximal oxygen consumption (VO2 max) compared with nondiabetic persons (20). In addition, the rate of increase in oxygen consumption (VO2) during graded treadmill exercise is attenuated in persons with type II DM compared with nondiabetic individuals (20). These data suggest that the reduced VO2 max in type II DM may not simply be due to the early termination of graded exercise, or deconditioning, but to a qualitative difference in the rate of rise in VO2 with graded exercise. However, observations made during incremental exercise do not represent steady-state phenomena and may be influenced by a variety of factors, including the rate of progression in workload and the ability of the cardiovascular system to respond to the increased work demand. To control for these variables, constant-load, steady-state exercise can be utilized. Under constant-load conditions, VO2 at steady state is in direct proportion to the external work demand. However, during the transition from rest to constant-load exercise, the rate of rise of VO2 reflects the dynamic response of the cardiovascular system and skeletal muscle oxygen uptake. A rapid increase of VO2 to steady state is seen in healthy, physically trained individuals (3, 31). In contrast, a delayed rise in VO2 to steady state is observed in patients with decreased cardiac function or in chronic obstructive pulmonary disease (5, 25, 26).
To evaluate the effects of type II DM on VO2 kinetics, we had 10 sedentary women with uncomplicated type II DM perform bouts of constant-load exercise. In addition, we measured heart rate kinetics, which evaluate the rate of rise in heart rate at the beginning of exercise. The effects of obesity (commonly observed in type II DM) were controlled for by including two control groups (both sedentary). One group was moderately overweight, similar to the group with type II DM, and one group was composed of lean women.
In the present study, we found that women with type II DM had impaired maximal and submaximal cardiopulmonary responses to exercise, even though they had no evidence of clinical cardiovascular disease or diabetic complications. We had previously observed a lower VO2 response to submaximal workloads during submaximal graded exercise testing in persons with type II DM compared with controls (20). We reasoned that this finding might be due to slowed VO2 kinetic responses in subjects with type II DM. In the present study, constant-load testing was used to confirm that the VO2/workload relationship was impaired and that VO2 kinetic responses were in fact slowed in type II DM. In addition, the presence of slowed heart rate kinetics in the women with type II DM suggested that a cardiac component may be partially responsible for the abnormalities observed.
In the present study, to confirm our finding of slowed VO2 kinetics, we used multiple constant workloads, with each workload repeated several times over 2 days. We documented the consistent finding of slowed kinetics in persons with type II DM at the two workloads below the lactate threshold. However, at the one workload above the lactate threshold, persons with type II DM only tended to have slower kinetics than did controls. This may have been due in part to limitations in monoexponential modeling techniques at this workload.
Importantly, the presence of greater than ideal body weight could not account for the lower VO2 max or slowed VO2 and heart rate kinetic responses observed in the study because the overweight and lean groups had similar responses. Overweight controls were not different in terms of weight and fat-free mass from subjects with type II DM. In addition, VO2 max was lower in persons with type II DM than in controls, whether presented in milliliters per minute or milliliters per kilogram per minute. Differences in habitual physical activity level could also not account for the exercise differences observed between persons with type II DM and control subjects. The use of the LOPAR questionnaire revealed that physical activity levels did not significantly differ between groups.
The present study was performed in women only. The reason for studying women was that we observed that women with type II DM had a lower VO2 max relative to their nondiabetic counterparts than did men with type II DM compared with nondiabetic men (unpublished observations). The reason for studying premenopausal women only was for greater homogeneity of the sample in terms of age. The finding of cardiopulmonary exercise abnormalities during maximal and submaximal exercise was especially interesting given that the women studied had only had the clinical diagnosis of diabetes for a relatively short time.
Whereas, in healthy individuals, VO2 kinetic measurements are thought to closely reflect the time course of changing VO2 of exercising muscles, persons with specific cardiovascular or cardiopulmonary diseases have rate-limiting defects in the oxygen delivery and utilization process (5, 25, 26, 31). In the healthy individual, where oxygen delivery is not rate limiting during submaximal exercise, VO2 kinetics reflect the oxidative resynthesis rate of phosphocreatine (i.e., primarily reflect muscle bioenergetics and oxygen diffusion at the tissue level) (31) and therefore the utilization aspects of VO2 during exercise. However, in disease states in which oxygen delivery is compromised, for example, by a limited cardiac output response, VO2 kinetics also reflect the ability of the cardiovascular system to deliver oxygen to working muscle and therefore may reflect impaired oxygen delivery (28). Consistent with this thinking is the finding that the tau of phase II of VO2 kinetics (rise to steady state) is prolonged in patient groups with abnormal cardiovascular responses to exercise, such as pulmonary vascular disease and cyanotic congenital heart disease (25, 26). Further evidence for a relationship between impaired oxygen delivery and slowed kinetics is the observation that patients with pulmonary vascular disease who underwent surgical procedures that improved pulmonary hemodynamics had faster VO2 kinetics after the procedure (25).
There is evidence to support the idea that both central (cardiac) and peripheral factors may be related to the exercise abnormalities associated with type II DM. Studies have not thoroughly assessed the ability of the person with diabetes to utilize oxygen during exercise. Allenberg et al. (1) and Lithell et al. (16) reported that citrate synthase activity in skeletal muscle increased markedly after exercise training in type II DM, thereby showing the normal response. In contrast, Simoneau and Kelley (27) recently reported a higher than normal ratio between glycolytic and oxidative enzyme activities that was explained not only by an increased activity for glycolytic enzymes but also by decreased maximum velocities for citrate synthase and cytochrome-c oxidase enzymes in the subjects with type II DM compared with lean and nondiabetic obese subjects. Another recent study investigated whether older persons with impaired glucose tolerance or type II DM had an increased frequency of mitochondrial DNA deletions in skeletal muscle compared with an age-matched nondiabetic control group (15). The authors found that one particular deletion (4,977 bp) as well as other deletions were significantly increased in the muscle tissue of subjects with type II DM or impaired glucose tolerance compared with nondiabetic individuals. Future studies should further explore the effects of type II DM on skeletal muscle metabolism.
There is also evidence suggesting that impaired myocardial function (and subsequently oxygen delivery) in persons with type II DM may play a critical role in the abnormal exercise performance observed in persons with type II DM compared with controls (13, 24). In the present study, the finding of slowed heart rate kinetics supports the likelihood of a cardiac factor as a component of the exercise abnormalities observed. In other studies, a reduced cardiac output during exercise has been reported in persons with type II DM vs. controls (13, 24). One study used right heart catheterization to show the presence of a reduced cardiac output in men with diabetes during submaximal workloads of supine bicycling exercise (13). Another study, which used noninvasive methods (24), also measured cardiac output during exercise in persons with diabetes and reported similar results. Methodological issues limit interpretation in both studies. For instance, subjects were included who were taking insulin and oral agents. Persons with both type II DM and type I DM were studied, although evidence suggests that these groups may show differing hemodynamic responses to exercise (8). Also, subjects were not separated according to physical activity levels or carefully matched for age, factors which can strongly affect exercise performance. Finally, the presence of autonomic dysfunction was not evaluated in the first study. However, the suggestion from the literature is that some degree of LV dysfunction may occur in persons with diabetes.
To summarize, the results of the present study demonstrate that premenopausal women with uncomplicated diabetes have impaired VO2 responses to maximal and submaximal exercise. Further studies will be necessary to evaluate whether cardiac output, arteriovenous oxygen difference, and/or aspects of skeletal muscle metabolism are involved in causing the abnormalities observed. Understanding the magnitude and causes of the exercise impairments observed in this relatively healthy group of women with type II DM is important to potentially target appropriate interventions to improve exercise performance and thereby perhaps prevent increasing disability.
Ten subjects were enrolled in each group (Table 1). There were no significant differences among the three groups in age. The subjects with type II DM had been diagnosed with the disease an average of 3 yr. The group with type II DM and the overweight control group were not different with regard to weight, body mass index, or fat-free mass. However, these two groups differed from the lean control group in terms of the above measurements (Table 1). Fasting insulin, fasting glucose, and Hb AIc levels did not differ between the lean and overweight control groups. However, the group with type II DM had higher insulin, glucose, and Hb AIc levels than did the other two groups (all P < 0.05). Analysis of the physical activity recall revealed that there was no significant difference among the three groups in terms of habitual physical activity level measured by LOPAR. (Table 1).
Graded Maximal Exercise Test
VO2 max was lower in the group with type II DM than in the other two groups whether expressed in milliliters per minute or normalized to kilograms or kilograms of fat-free mass (Table 2). In addition, the maximal respiratory exchange ratio did not differ among groups and suggested a maximal effort in all three groups (all values were over 1.10). Maximal heart rate also was not different among groups.
To evaluate the VO2/work rate relationship, we compared the slopes of the Delta VO2/Delta work rate measured during the graded test between the three groups (Fig. 1). Overweight and lean groups exhibited nearly identical slopes (Fig. 1). Data from the overweight group were shifted upward, reflecting the effect of obesity on the absolute VO2/work rate. In contrast, data from the group with type II DM, which similarly shifted upward, revealed a decreased slope (P < 0.05).
Kinetic Responses to Constant-Load Exercise
VO2 kinetics were slower in persons with type II DM than in the lean and overweight control groups at the 20- and 30-W workloads (both P < 0.05, comparison between persons with type II DM and the other two groups) and tended to be slower at 80 W as well (P = 0.09; Figs. 2 and 3, Table 3). The heart rate kinetics were slower in persons with type II DM than in the other two groups at all three workloads (both P < 0.05, comparison between persons with type II DM and the other 2 groups). The values of the overweight and lean control groups were not different from each other in terms of the VO2 and heart rate kinetic measurements.
Steady-State Measurements During Constant-Load Exercise
Respiratory exchange ratios did not differ between groups during constant-load exercise workloads (Table 4). However, lactate concentrations were higher at 30 and 80 W (both P < 0.05) in the group with type II DM than in the other two groups. Lactate concentration was higher in the group with type II DM than in the lean group at 20 W and tended to be higher in persons with diabetes than in the overweight control group. Steady-state VO2 did not differ between groups at any workload. However, steady-state VO2 as a percentage of VO2 max was higher for the group with diabetes at all three constant-load workloads (Table 4).
Correlations Between Maximal and Steady-State Values
There was an inverse correlation across the three groups between the VO2 tau and VO2 max such that the shorter the tau , the greater the VO2 max (r = -0.36, r = -0.38, and r = -0.38, all P < 0.05 for 20, 30, and 80 W, respectively). There was also an inverse correlation between heart rate kinetics and VO2 max such that the shorter the tau for heart rate kinetics, the greater the VO2 max (r = -0.59, r = -0.45, and r = -0.60, all P < 0.05, for 20, 30, and 80 W, respectively).
Ten moderately overweight, sedentary, premenopausal women with type II DM and no complications or comorbid conditions, 10 moderately overweight but otherwise healthy women of similar age and activity levels, and 10 healthy lean women of similar age and physical activity levels were studied. Only women were enrolled because preliminary data suggested that exercise performance may be more affected by type II DM in women than in men (unpublished observations). No subject was >140% of ideal body weight. All women displayed sedentary behavior defined as not participating in a regular exercise program (>1 bout of exercise per week). Furthermore, use of a low-level physical activity recall (LOPAR) questionnaire ensured that physical activity levels were similar among subjects (20, 21). Presence of type II DM was documented by chart review, which confirmed the diagnosis as well as the presence of drug treatment for diabetes. Persons with type II DM were included in the study if their diabetes was treated by diet or oral agents, but not by insulin, because subjects treated with insulin tend to have more advanced disease. Two patients with type II DM were treated with diet only, six with glyburide, and two with glipizide. Duration of diabetes (from diagnosis) was noted. Other than oral agents, subjects with type II DM were taking no other medicines. Women with type II DM were accepted for study only if they had glycosylated hemoglobin levels (Hb AIc) levels <9% (adequate control) on therapy.
Women who were current smokers were not accepted for study because smoking can impair cardiovascular exercise performance. Former smokers must have been abstinent for the past 2 yr.
Premenopausal women between the ages of 30 and 50 yr were included in the study. Premenopausal status was evaluated in all women by history of regular menstrual cycles and by measurements of serum follicle-stimulating hormone (FSH) levels. For the purpose of uniformity and to rule out effects of widely differing levels of female hormones on exercise as well as to minimize potential effects of progesterone on ventilation, women were tested in the midfollicular phase (days 6-10) of the menstrual cycle (14, 22).
Absence of comorbid conditions was confirmed by history, physical examination, and laboratory testing. Distal symmetrical neuropathy was evaluated by symptoms (numbness, paresthesia) and signs. Persons who had clinically evident distal symmetrical neuropathy were excluded from further study because of possible effects on exercise performance (10). Three subjects were excluded by using these criteria.
Through the use of resting echocardiographic criteria, persons were excluded who had the presence of global or regional contractile abnormalities (12). Exclusions occurred if 1) regional wall motion abnormalities suggested coronary disease, 2) left ventricular (LV) wall thickness was >1.3 cm (suggesting moderate LV hypertrophy), or 3) there was decreased contractility, i.e., fractional shortening <35%. Subjects were also excluded if they had evidence of ischemic heart disease by history or abnormal resting or exercise electrocardiogram (ECG) (>= 1-mm flat or downsloping S-T segment depression). Persons with angina or any other cardiac or pulmonary symptoms potentially limiting exercise performance were excluded as well. Presence of systolic blood pressure >130 mmHg at rest or >190 mmHg with exercise or diastolic pressure >90 mmHg at rest or >100 mmHg with exercise was also grounds for exclusion. Persons with autonomic insufficiency, assessed by measuring variation in R-R intervals with cycled breathing and by the presence of a >20-mm fall in upright blood pressure without a change in heart rate, were excluded because of possible effects on exercise performance (9). Subjects with proteinuria (urine protein >200 mg/dl) or a creatinine >= 2.0 mg/dl, suggestive of renal disease, were excluded. Renal disease was grounds for exclusion because it can alter exercise performance (2).
Control subjects were screened identically to persons with type II DM. These subjects were taking no medications, had a normal Hb AIc, and had no history of any active medical problems.
Design (Study Protocol)
Subjects were evaluated over the course of six visits on separate days. Subjects made an initial visit to the General Clinical Research Center (GCRC) to have a history and physical examination, blood draws, and questionnaire administrations. A resting ECG was obtained, and a familiarization bicycle test was performed. During the two subsequent visits, a diet interview was administered and underwater weighing was performed. Three days before the fourth visit (to control for the effects of diet on exercise performance), subjects began receiving all meals from the GCRC. On the fourth visit, subjects performed a graded maximal bicycle exercise test to determine the lactate threshold and VO2 max. On the fifth and sixth visits, four 7-min bouts of constant-load exercise were performed each day (with 15-min rest periods between bouts) so that three bouts at 20 W, three bouts at 30 W, and two bouts at 80 W were performed in total over the 2-day period. Performance of multiple bouts enabled averaging of VO2 kinetic data within a workload to reduce variability of results.
Graded Maximal Exercise Test
After subjects fasted overnight, VO2 max and lactate threshold were determined during a graded bicycle protocol to exhaustion. Each test began with the subject seated at rest on the cycle ergometer (Cardio-2, Medical Graphics, Minneapolis, MN) breathing into a mouthpiece connected to a metabolic cart (CPX-D, Medical Graphics). All testing was done with the subject in the upright seated position. Three minutes of resting data were collected to obtain baseline measurements before exercise. The rest period was prolonged at the discretion of the investigator if additional time was required for adjustment to the mouthpiece and stabilization of physiological variables. To obviate the need to overcome inertia of the ergometer flywheel at the start of exercise, the flywheel was driven at 60 rpm during rest by an electric motor, which was turned off synchronous with the start of exercise. At the start of exercise, the work rate was increased in 10 W/min increments, and the incremental portion of the test was 12-15 min in duration. VO2 max was defined as VO2 remaining unchanged or increasing <1 ml · kg-1 · min-1 for 30 s or more despite an increment in workload (29).
VO2 and carbon dioxide production (VCO2) were measured, breath by breath, at rest and during exercise. Peak VO2 data were averaged over 30-s intervals. Arm blood pressure (by auscultation) and heart rate (by 12-lead ECG) were obtained every minute during exercise. Cardiac status was monitored throughout the test by 12-lead ECG. The respiratory exchange ratio was calculated as VCO2/VO2. VO2 was normalized on a per kilogram basis and per lean body mass as well as presented as milliliters per minute. Normalization by lean body mass was done to avoid confounding, which could result from differences in the fat mass between lean and overweight (type II DM and overweight control) subjects.
The slope of the increase in VO2 per increase in work rate (Delta VO2/Delta work rate) was analyzed by least squares linear regression excluding the first 2 min and last minute of graded exercise data as previously described (11).
Blood was drawn every minute during the first VO2 max test to enable determination of the lactate threshold. The lactate threshold was defined as the point at which a net increase in venous lactate accumulation was observed during incremental maximal exercise. Venous lactate (mmol) vs. VO2 was plotted for determination of the lactate threshold for each patient. The VO2 at the lactate threshold was determined and recorded from each plot. Confirmation of the lactate threshold was performed by using ventilatory data and the V-slope technique. VCO2 was plotted against VO2, and the ventilatory threshold was labeled where the slope of VCO2 vs. VO2 exceeded 1.0. In all cases, V-slope analyses confirmed lactate threshold measurements.
Constant-Load Exercise Tests
An overnight fast preceded each test day. Each test began with 3 min of resting baseline measurements as described in Graded Maximal Exercise Test. After this period, with the flywheel driven by the motor as described Graded Maximal Exercise Test, the preselected workload (20, 30, or 80 W) was then imposed and the subject maintained pedaling at 60 rpm for 7 min. This protocol allowed all subjects to reach steady-state VO2 at 20 and 30 W but not at 80 W, which was above the lactate threshold.
Kinetic Measurements During Constant-Load Exercise
VO2 kinetic measurements. Three phases to the response of VO2 from rest to moderate constant-load exercise were proposed by Whipp and Mahler (30). At the onset of exercise, VO2 from the lungs normally increases abruptly for the first 15 s (phase I) as pulmonary blood flow increases. Next, the VO2 increases exponentially in phase II with a time constant (tau ) of ∼30-45 s representing further increases in blood flow and decreased venous O2 content. Phase II ends as gas exchange approaches a steady state. Phase III is steady-state VO2 below the lactate threshold, but above the lactate threshold; phase III VO2 kinetic responses are not steady-state (phase III drift), and modeling is altered accordingly (30).
VO2 was measured breath by breath. After collection, the VO2 data were transferred to an ASCII file and filtered, and then a VO2 value was assigned to each second by extrapolation between breaths by using a program developed and validated at our laboratory. To dampen noise and enhance resolution of data, the data from repetitions within a workload were temporally aligned to a time at the start of exercise and superimposed to yield a single second-by-second averaged record of the tests for each subject at a given workload. The tau and the actual change in VO2 from rest to steady-state VO2 were then calculated by using a statistical program as previously described (6). With use of this program, a single exponential curve was fit to the data from the onset of exercise to the end of the sixth minute of steady-state exercise in 20- and 30-W transitions. Because we expected that 80 W would constitute high-intensity exercise (i.e., above the lactate threshold) in the majority of patients (given the sedentary profile of the patients), we modified the modeling procedures for this workload to enable comparability between those for whom 80 W was above the lactate threshold and those still below the lactate threshold at 80 W. Thus, to minimize the impact of a phase III drift in VO2 (non-steady-state VO2 associated with constant-load exercise VO2 above the lactate threshold), 80-W transitions were analyzed by using a single exponential curve fit from the onset of exercise to the end of the third minute of exercise.
Heart rate kinetic measurements. Heart rate was monitored beat by beat from the R-R interval of the ECG signal (Quinton Q-plex). These data underwent analog-to-digital conversion and were subjected to kinetic analyses as described in Kinetic Measurements During Constant-Load Exercise.
Echocardiographic measurements. Two-dimensional and Doppler echocardiography were performed by using standard methods (12) to exclude the presence of significant valvular pathology, LV global dysfunction and segmental wall motion abnormalities (Sonos 2500, Hewlett-Packard, Andover, MA). Chamber sizes, LV end-systolic and diastolic chamber dimensions and wall thickness, fractional shortening, and the area-length method for measurement of cardiac volume (to measure ejection fraction) were quantitated by standard techniques for all individuals. All readings were done by one of the authors (E. E. Wolfel), a cardiologist who is skilled in these readings and who was blinded to the diagnostic status of the patient. Measurements of diastolic filling were assessed by analyzing mitral valve and pulmonary venous flow patterns by using Doppler techniques.
Dietary control. Three days before the fourth visit, subjects began diet regulation. The diet was administered until the sixth visit was completed. With the use of a diet interview administered during the third visit, subjects were given a diet composed of the percentage of each macronutrient that they customarily ate for all meals. Customary macronutrient pattern was used because a change in diet may affect the respiratory exchange ratio. Overnight fasting (from 10:00 PM the preceding night) was observed before the underwater weighing test day and each exercise test day.
Body composition and hydrodensitometry. Body composition and hydrodensitometry measurements were performed according to standard methods and were used to derive fat-free mass. Percent body fat was estimated from body density by using the revised equation of Brozek et al. (4). Body fat distribution was determined by using the waist-to-hip ratio, where the waist circumference was measured at one-half the distance from the xiphoid process to the navel and the hip circumference was measured at the level of the greater trochanter.
Tests of autonomic insufficiency. To evaluate autonomic insufficiency, we measured variation in R-R intervals with cycled breathing (7, 9). The method for obtaining R-R variability was as follows. The patient, while resting supine, breathed five times per minute, coordinating breaths with a visual electronic signal. This was repeated for 5 min. To obtain data, maximum inspiratory heart rate was subtracted from the minimum expiratory heart rate. Variations of >30 beats/min were considered normal, and values <20 beats/min were considered abnormal (7, 9). In addition, autonomic insufficiency was evaluated by measuring, in lying and standing subjects, heart rates and blood pressures (>20-mm fall in upright systolic blood pressure without a change in heart rate). Subjects who failed to meet these criteria were excluded from study. Three potential subjects were excluded in this way.
Blood collection and preparation. Blood was drawn at baseline for the measurement of glucose, insulin, and plasma FSH levels and of Hb AIc. Blood lactate concentrations were measured every minute during the graded exercise test to determine the lactate threshold in all subjects. In addition, lactate was measured at rest and at peak exercise during the constant-load tests. For the measurement of blood lactate concentration, a 20-gauge intravenous catheter was placed in a forearm vein, with a three-way stopcock to facilitate blood drawing, and patency was maintained with heparinized saline. For each sample, 50 Ál of blood were withdrawn and immediately deproteinized in 3% perchloric acid and stored at room temperature.
Assay methods. Lactate concentration was assayed by a lactate dehydrogenase method (23). The lactate threshold was determined as the point at which blood lactate concentration began to progressively increase in the blood. Hb AIc was measured by glyc-affin GHB columns (Isolab). Serum insulin concentrations were measured by radioimmunoassay (18, 32). Serum glucose concentrations were measured by the glucose oxidase method (17). Plasma FSH levels were measured by a chemiluminescence assay (19).
LOPAR questionnaire. This questionnaire has been modified for use in persons with type II DM and peripheral arterial disease as well as in sedentary controls (20, 21). The subjects were asked a series of questions to itemize their time (reporting specific activities) into work, leisure, and housework categories for the previous week. Questionnaire results were expressed in metabolic equivalents (METs) where 1 MET equals resting VO2 (3.5 ml · kg-1 · min-1). Scores are reported in MET hours per week, derived by multiplying the amount of time spent performing an activity by the MET value of the activity. This questionnaire was primarily used in the present study to quantify the activity level of all participants.
Data analysis. The three groups were compared by using a between-subjects ANOVA. The Student-Newman-Keuls test was used for post hoc analysis. Where data were nonparametric, the Kruskal-Wallis test was used to make between-group comparisons. Correlations were done by using Pearson's product-moment correlation.