Effect of exercise training on insulin sensitivity and glucose metabolism in lean, obese, and diabetic men KAREN R. SEGAL, MARIA B. TOMAS,
ALBERT EDANO, AMY ABALOS, AND F. XAVIER PI-SUNYER
JEANINE
ALBU,
LORNA
BLANDO,
Division of Pediatric Cardiology, Department of Pediatrics, Mount Sinai School of Medicine, Nezu York 10029; and Obesity Research Center, St. Luke’s=Roosevelt Hospital Center, New York, Nezu York 10025
SEGALJCAREN R., ALBERT EDANQAMY ABALOS, JEANINE ALBU, LORNA BLANDO, MARIA B. Tcmw, AND F. XAVIER PISUNYER. Effect of exercise training on insulin sensitivity and glucose metabolism in. lean, obese, and diabetic men. J. Appl. Physiol. 71(6): 2402-2411, 1991.-To clarify the impact of vigorous physical training on in vivo insulin action and glucose metabolism independent of the intervening effects of concomitant changes in body weight and composition and residual effects of an acute exercise session, 10 lean, 10 obese, and 6 dietcontrolled type II diabetic men trained for 12 wk on .a cycle ergometer 4 hfwk at -70% of maximal 0, uptake (VO, m,) while body composition and weight were maintained by refeeding the energy expended in each training session. Before and 4-5 days after the last training session, euglycemic hyperinsulinemic (40 mU. m2. min-l) clamps were performed at a plasma glucose of 90 mgldl, combined with indirect calorimetry. Total insulin-stimulated glucose disposal (M) was corrected for residual hepatic glucose output. Body weight, fat, and fat-free mass (FFM) did not change with training, but cardiorespiratory fitness increased by 27% in all groups. Before and after training, M was lower for the obese (5.33 t, 0.39 mg . kg FFM-1 min-’ pretraining; 5.33 t 0.46 posttraining) than for the lean men (9.07 t 0.49 and 8.91 -+ 0.60 mg . kg FFM-1 . min-l for pretraining and posttraining, respectively) and lower for the diabetic (3.86 t 0.44 and 3.49 t 0.21) than for the obese men (P < 0.001). Insulin sensitivity was not significantly altered by training in any group, but basal hepatic glucose production was reduced by 22% in the diabetic men. Thus, when intervening effects of the last exercise bout or body composition changes were controlled, exercise training per se leading to increased cardiorespiratory fitness had no independent impact on insulin action and did not improve the insulin resistance in obese or diabetic men. l
obesity; diabetes mellitus; euglycemic hyperinsulinemic indirect calorimetry; body composition
REDUCED insulin-mediated
clamp;
glucose uptake due to insulin resistance is a characteristic of both obesity and non-insulin-dependent diabetes mellitus (17). Physical exercise is prescribed widely as an adjunct in the treatment of both obesity and diabetes, as a modality for the induction of a negative energy balance, and also for the impact of exercise on in vivo insulin action. Although a single bout of physical exercise improves glucose metabolism acutely (4, 5), the impact of a regimen of physical training on metabolism is unclear. Cross-sectional comparisons indicate that basal and stimulated plasma insulin levels are lower and insulin sensitivity is greater in physically 2402
trained than in sedentary individuals (20). However, physically trained individuals are usually leaner than sedentary people, and it is likely that the differences can be attributed to the lower body fat rather than the level of cardiorespiratory fitness itself. Longitudinal studies involving controlled physical training are potentially difficult to interpret because of concomitant changes in other factors, such as the state of energy balance and changes in body composition, as well as the residual effects of the last exercise bout, which independent of physical training can exert strong metabolic effects for as long as 36 h postexercise. An acute bout of exercise enhances glucose uptake (5,9), and weight loss itself is also associated with improved insulin action (29). Furthermore, when physical training is accompanied by weight loss, insulin resistance is lessened (3). However, the specific changes in glucose metabolism induced by chronic exercise training leading to enhanced cardiorespiratory fitness are difficult to determine, because the effect of long-term training can be confounded by residual effects of the last isolated exercise bout, and the energy deficit induced by long-term rigorous physical training may lead to weight loss and/or changes in body composition, which in themselves are likely to alter glucose metabolism and insulin sensitivity. It has not been established whether exercise training without even a small reduction in body fat content and body weight can alter glucose metabolism. There is considerable evidence that a defect in energy expenditure, specifically in the component of total daily energy expenditure related to the stimulation of metabolic rate after the ingestion or infusion of nutrients (postprandial or dietary-induced thermogenesis), may have a role in the pathogenesis and persistence of the obese state (33,37). Several investigations have demonstrated a link between reduced insulin-stimulated glucose uptake and blunted thermogenesis (13,32). In fact, insulin resistance leading to impaired insulin-mediated glucose disposal may be the mechanism fur the blunted thermic responses to ingested or infused nutrients in obese and non-insulin-dependent diabetic individuals (13). Although it is clear that exercise itself profoundly increases energy expenditure in proportion to the intensity of exercise, the impact of physical training on energy expenditure (thermogenesis) remains unresolved (6,22). If impaired insulin-mediated glucose disposal is associated with the blunted thermogenic responses of obese and diabetic subjects, as previous research has suggested,
0161-7567/91 $1.50 Copyright 0 1991 the American Physiological Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.108.009.184) on October 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
EXERCISE
TRAINING
AND
GLUCOSE
METABOLISM
2403
(Koladex, Custom Laboratories, Baltimore, MD) was consumed within 3 min, and venous blood samples were drawn at 30-min intervals for 2 h and analyzed for plasma glucose and insulin. The subjects voided just before the glucose load, and urine samples were collected for the duration of the test for determination of urinary glucose loss. The integrated areas under the glucose and insulin curves were calculated. All the diabetic men met the following criteria for inclusion in the study: 1) fasting plasma glucose >140 mg/dl on more than one occasion, 2) 120-min OGTT plasma glucose >200 mg/dl, and 3) 309, 600, or 900min OGTT plasma glucose >200 mg/dl. The OGTT was administered twice before and twice 24 days after the last exercise session of the physical training program, and the average of the values obtained from the two tests before and the two tests after training was used. Body composition. Body fat content and FFM were deMETHODS termined by densitometry. Body density was determined Subjects. Ten lean males with normal glucose tolerby hydrostatic weighing by the method described by ance and no personal or family history of obesity or dia- Akers and Buskirk (1) and was corrected for residual betes, 10 obese males with normal glucose tolerance and lung volume, which was estimated by means of the no family history of diabetes mellitus, and six obese men closed-circuit 0, dilution method of Wilmore (42). Perwith non-insulin-dependent diabetes mellitus between cent body fat and FFM were derived from body density the ages of 25 and 40 yr participated in this study. The by use of the Siri equation (38). Body composition detergroups were matched with respect to age, height, and minations were made before and after the physical trainfat-free mass (FFM, determined densitometrically, see ing program. below). The lean men were ~18% body fat, and the obese Aerobic fitness test. Aerobic fitness was determined by and diabetic men were >28% body fat, determined by a maximal continuous multistage exercise test on an hydrostatic weighing. All subjects were healthy nonsmokelectromagnetically braked cycle ergometer (Robert ers with no personal history of cardiovascular disease. Bosch, Berlin, FRG). Before the test, the subjects were The men were weight stable at the time of the study, with familiarized with cycling on an ergometer at a constant no more than a 2-kg weight loss or gain over the 6 mo pedaling rate while breathing through the apparatus before the study. used for metabolic measurements. The subjects began An oral glucose tolerance test (OGTT) was adminiscycling at a rate of 60 rpm with zero external resistance tered to confirm that the lean and obese men had normal (unloaded cycling). The work rate was increased in 30-W glucose tolerance and to confirm diabetes mellitus in the increments every 2 min until volitional exhaustion was reached and the subject refused to continue despite vocal diabetic group, according to the criteria of the National Diabetes Data Group (27). The duration of diabetes in encouragement, or until he was unable to maintain the the diabetic subjects ranged from 0 (newly diagnosed) to pedaling rate. Ventilatory measurements were made continuously by open-circuit respirometry with the use of a 24 mo; the subjects managed their therapy by diet alone and were not taking any oral hypoglycemic agents, to Horizon metabolic measurement cart (Sensormedics, avoid any confounding metabolic effects of these drugs. Anaheim, CA), which includes a turbine volume transducer, a Beckman OM-11 polarographic 0, analyzer, and Furthermore, to obtain a homogeneous group of diabetic subjects, only non-insulin-dependent diabetic men with a Beckman LB-2 nondispersive infrared CO, analyzer. a high insulin response to the OGTT were included in the The subjects breathed through a nonrebreathing valve study. (Hans Rudolph, Kansas City, MO) and used a mouthpiece and noseclips. The gas analyzers were calibrated All subjects were sedentary, defined as no involvement in regular exercise (23 times/wk) for >2 yr before the before and after each aerobic fitness test with 100% N2, room air, and a gas mixture containing 4% CO, and 16% study and a current physical activity level of 1.15; 2) a leveling off of heart tained, and the protocol was approved by the Institurate, with no more than a 3-beat/min incrementbetween tional Review Board of the Mount Sinai School of Medicine. the final two work loads; and 3) a leveling off of Vo,, with OGTZ’. The subjects were tested in the morning in the less than a 3-ml. kg-’ amin-l increment between the final two work loads. If two or more of these criteria were postabsorptive state after an overnight (12-h) fast. After not met, the test was repeated on another day. a fasting blood sample was drawn, a 75-g glucose solution
and if exercise training improves insulin sensitivity, the capacity for thermogenesis might also be enhanced by means of improved insulin-stimulated glucose disposal. The effect of physical training on insulin-stimulated glucose disposal and energy metabolism has never been compared in lean, obese, and non-insulin-dependent diabetic subjects with rigorous control for changes in body composition. In addition, there is no information about the relationship between training-induced changes in insulin sensitivity and thermogenesis. The present study was designed to determine the independent impact on in vivo insulin action and glucose metabolism of a 12wk program of vigorous physical training during which body weight and body fat content were held constant by refeeding the energy expended in each training session in lean and obese men with normal glucose tolerance and obese men with non-insulin-dependent diabetes.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.108.009.184) on October 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
2404
EXERCISE
TRAINING
AND GLUCOSEMETABOLISM
Submaximal aerobic fitness was determined by estimation of the ventilatory threshold from the test data. The ventilatory threshold is the highest work rate or VU, before VE increases out of proportion to 00, (41) because of stimulation of ventilation by nonmetabolically produced CO,, which derives from the buffering of lactic acid. The graded exercise test was routinely performed before and after the physical training program. In a subset of men, the graded exercise test was repeated 7-10 days after the last training bout to ensure that the observed maximal 60, (iTo,,,) values observed after training were maintained during the period of posttraining metabolic testing and that no detraining effect had yet occurred. The fitness test data were used I) to provide descriptive information about the aerobic fitness level of the subjects before and after physical training and 2) for the individualized prescription of the appropriate training work rate. Physical training. The physical training program consisted of 12 wk of cycle ergometer exercise performed for 1 h and 10 min, four times per week. Each session consisted of 5 min of warm-up at a rate of 25 W, 1 h of cycling at the appropriate intensity, and 5 min of cool-down at a rate of 25 W. Initially, each subject trained at the work rate associated with his ventilatory threshold, determined from the graded aerobic fitness test. Thus the appropriate work intensity was prescribed on an individual basis, based on each subject’s submaximal aerobic fitness, The ventilatory threshold is a physiological parameter, determined from gas exchange parameters, which is made on the basis of the fact that ventilation is stimulated out of proportion to increases in 60, during graded exercise by nonmetabolically produced CO, that derives from the buffering of lactic acid (41). Throughout the training program, the work rate was increased so that it provided a similar stimulus as the subjects increased their level of aerobic fitness. The heart rate associated with the initially prescribed work intensity was recorded, and the work rate was adjusted weekly to maintain the same heart rate during exercise. The ventilatory breakpoint was reassessed periodically during training by means of a submaximal exercise test to ensure that the training intensity was appropriate. 90, and associated variables during the exercise period were measured biweekly throughout the training program to provide information about the energy expended during exercise. Diet and maintenunce of body weight during physical training. Because the objective of this project was to study the effect of physical training on in vivo insulin action and glucose metabolism independent of changes in body weight and body composition, the subjects were required to maintain their initial body weight throughout the duration of the study. Financial incentive was provided to ensure that the subjects complied with the requirement of weight maintenance. Body weight was measured on every test day, before and after training, and at each training session. The energy expended during a training session was measured biweekly for each subject. After each bout of exercise, the calories expended during exercise were refed in the form of Sustacal (Mead Johnson, Evansville,
IN; composition: 24% protein, 21% fat, 55% carbohydrate). None of the subjects had any difficulty maintaining weight during the training program, and none was dropped from the study because of weight gain or weight loss tl kg. Euglycemic hyperinsulinemic insulin clump plus indirect calorimetv. The subjects were tested at 7 A.M. after an overnight 12-h fast, having abstained from any exercise for >4-5 days before the test. After the training program, the clamp studies were performed >3 days after the last exercise bout to distinguish between the effects of acute exercise and chronic exercise training. The euglycemic hyperinsulinemic clamps were performed according to the technique described by DeFronzo et al. (8). When the euglycemic hyperinsulinemic clamp is combined with simultaneous indirect calorimetry, the rates of glucose oxidation and glucose storage can be calculated and the thermic effect of infused glucose and insulin, which is the increase in energy expenditure during the clamp, can be determined. An intravenous catheter was inserted in retrograde fashion in a hand vein, and the hand was placed in a heated box (>7O”C) for the duration of the entire test to provide arterialized venous blood samples (24). An indwelling Teflon catheter was placed in the antecubital vein of the other arm for infusion of glucose and insulin. A primed dose of 25 &i of [3-3H]glucose (New England Nuclear) was administered, followed by continuous infusion of 0.25 &i/min of [3-3H]glucose for 2 h to estimate hepatic glucose production in the basal state, and for the subsequent 2 h to determine hepatic glucose production during the clamp. The [3-3H]glucose was purified by high-performance liquid chromatography. Arterialized blood samples were obtained from the hand vein at 150 min intervals for the first 90 min and every 5 min during the last 30 min of this control period to measure the plateau steady-state plasma glucose specific activity. Steady-state enrichment of [ 3-3H]glucose was achieved in all three groups of subjects. This plateau value was used to calculate basal hepatic glucose production. After the 2-h control period, a primed continuous infusion of regular human insulin (40 mU rnA2 min-l, Humulin R, Eli Lilly, Indianapolis, IN) was administered for 2 h to raise and maintain the insulin concentration at -100 pU/ml. The plasma glucose level was maintained at 90 mg/dl by determining the plasma glucose level every 5 min and adjusting the infusion rate of a 20% dextrose-in-water solution with a servocontrol negative feedback principle (8). In the diabetic men, plasma glucose levels were allowed to decline to -90 mg/dl. This decline occurred within 60 min in all the subjects. To avoid rapid changes in glucose concentrations that might precipitate secretion of hormones counterregulatory to insulin, the rate of decline in plasma glucose was not >l.O mg +dl-l min? Blood was drawn at -30, -15, and 0 min during the baseline period and every 15 min during the clamp for determination of plasma insulin. Plasma [33H]glucose specific activity was measured at 150min intervals for the first 90 min of the insulin infusion and at 5-min intervals during the last 30 min of the study. Indirect calorimetry was applied for the last 30 min of the baseline period and again during the last 30 min of l
l
l
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.108.009.184) on October 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
EXERCISE
TRAINING
AND
GLUCOSE
METABOLISM
2405
the clamp to obtain measurements of energy expenditure Similar analyses were applied to the glucose oxidation and substrate utilization. The Horizon metabolic mea- and storage data obtained during the clamp. surement cart, described above, was used. Comparisons among groups of body composition, aerCalculation. The rate of appearance (Ra) of glucose in obic fitness, resting metabolic rate, physiological rethe plasma was calculated from the [3-3H]glucose spe- sponses to the acute exercise bout, and fasting glucose cific activity by use of the equations of Steele in the and insulin levels before and after training were also steady-state form (39) for the last 30 min of the basal made by application of 3 X 2 two-way ANOVA with reperiod and in the non-steady-state form, as validated by peated measures with group and training as the factors. Radziuk et al. (31), during the last 30 min of the clamp. The relationship between the thermic effect of glucose During the basal period, hepatic glucose production and the rate of glucose storage, before and after training, equals the Ra. The last 30 min of the euglycemic clamp was determined by regression analysis. Before the data study was used to calculate the rate of glucose uptake from the three groups were pooled, tests of the equality (M) by the entire body. The glucose distribution volume of the slopes of these regressions among the groups were of 40 ml/kg body wt was assumed. The rate of total body carried out. glucose disposal during the clamp was considered to be For all statistical analyses the 0.05 level of significance the sum of the rates of glucose infusion, corrected for the was used. glucose space, and the residual hepatic glucose production, if any. Correction was made for urinary glucose loss, RESULTS if any. The lean, obese, and diabetic men were matched with The subjects voided before the start of the test, and urine was collected at the end of the test for measure- respect to age and FFM. Thus FFM was similar among all three groups, whereas percent fat and total body ment of urinary urea nitrogen. The nonprotein respiratory quotient was calculated from the respiratory ex- weight were similar for the obese and diabetic groups change data and urinary urea nitrogen production rates. (Table 1). Body weight, FFM, and body fat did not The rates of carbohydrate and lipid oxidation were cal- change significantly with physical training in any of the culated according to the procedure of Lusk (23), which is groups (Table 1). The ratio of waist to hip circumference, based on a nonprotein respiratory quotient of 0.707 for an index of body fat patterning, was significantly lower in 100% fat oxidation and 1.000 for 100% carbohydrate oxi- the lean men (0.90 t 0.04) than in the obese (0.97 t 0.03) dation. Urinary urea nitrogen excretion was used as an and diabetic men (0.97 t 0.05) before training. The index of protein oxidation, on the basis of the assump- waist-to-hip ratio was not altered by training (0.90 t 0.04,0.97 t 0.04, and 0.96 in the lean, obese, and diabetic tion that 1 g nitrogen = 6.25 g protein. Nonoxidative glucose disposal, which includes storage men, respectively). The initial maximum fitness level was slightly lower for the diabetic men than for the other as glycogen, conversion to C, compounds, and conversion to lipid, was calculated as the difference between the rate two groups (see Table 1). Maximal work load and VO, of carbohydrate oxidation and the total uptake of glucose increased significantly in all three groups by ~27%. Subby the body. The thermic effect of glucose, which is the maximal aerobic fitness (Table 1) increased significantly change in energy expenditure during the clamp study, after training in all groups. Although the work intensity was prescribed on the was estimated as the difference between the energy expenditure during the last 30 min of the baseline period basis of the ventilatory breakpoint determined from the graded exercise test, the actual measured Vo2 during the and the last 30 min of the insulin/glucose infusion. 1 h of exercise training (Table 2) was higher than that Assay procedures. Plasma and urinary glucose levels were measured with a Beckman glucose analyzer II measured at the same work intensity during the 3-min (Beckman Instruments, Fullerton, CA). Insulin determigraded exercise test. Thus, although the pretraining vennations were made by radioimmunoassay with charcoal tilatory breakpoint during the graded exercise test ocabsorption with the use of a human insulin standard (16). curred at 50-55% BOB,,,, the actual training intensity Plasma [ 3-3H] glucose specific activity was determined was 70% V02max (Table 2). The energy expended during by the method of Katz and Dunn (19). Urinary urea ni- the exercise training sessions (Table 2), which indicates trogen was determined by the Kjeldahl method (14). the calories that were refed, increased significantly in all three groups but was slightly lower for the diabetic subPlasma C-peptide levels were determined by radioimmunoassay (21). jects because their fitness level was somewhat lower. Design and analysis of data. Comparison of the effect of ANOVA with repeated measures applied to the relative training on the thermic effect of infused glucose in the training intensity (Table 2) yielded a significant effect three groups was made by application of a 3 X 2 two-way for training but not for group or group-by-training interanalysis of variance (ANOVA) with repeated measures action, indicating that the relative training intensity was (44) to the basal and clamp data with group (lean, obese, similar for the three groups and was higher after training or obese diabetic) and training (pre- or posttraining) as than before, which reflects the increased submaximal as the factors. Parameters of glucose metabolism were ex- well as maximal cardiorespiratory fitness. The extra caloric intake that was refed to compensate for the energy pressed in absolute form (mg/min) and relative to FFM and total body weight. Significant F ratios from the AN- expended during exercise ranged from -2,400 kcal/wk OVA were followed by post hoc comparisons among cell initially to 3,400 kcal/wk by the end of the training promeans following the Newman-Keuls procedure, with use gram. The OGTT data are shown in Table 3. Fasting plasma of the appropriate error terms from the ANOVA (44). Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.108.009.184) on October 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
2406 TABLE
EXERCISE
TRAINING
AND
GLUCOSE
METABOLISM
1. Subject characteristics P Obese
Lean
Age, Yr
Height, cm Weight, kg Pre Post
Percent fat Pre Post Fat-free mass, kg Pre Post Maximal aerobic fitness Work load, W Pre Post
iTO,, ml/min Pre Post Percent increase in VO,, Submaximal aerobic fitness: ventilatory breakpoint Work load, W Pre Post VO,, mUmin Pre Post %Voz ma Pre Post
kcaUsession Pretraining Posttraining ire,, ml/min Pretraining Posttraining %Vo, *ax Pretraining Posttraining
Source
31t2 174t2
3622 17O-t3
NS NS
77.7t3.1
101.7*4.0
L
ALBERT EDANO, AMY ABALOS, AND F. XAVIER PI-SUNYER
JEANINE
ALBU,
LORNA
BLANDO,
Division of Pediatric Cardiology, Department of Pediatrics, Mount Sinai School of Medicine, Nezu York 10029; and Obesity Research Center, St. Luke’s=Roosevelt Hospital Center, New York, Nezu York 10025
SEGALJCAREN R., ALBERT EDANQAMY ABALOS, JEANINE ALBU, LORNA BLANDO, MARIA B. Tcmw, AND F. XAVIER PISUNYER. Effect of exercise training on insulin sensitivity and glucose metabolism in. lean, obese, and diabetic men. J. Appl. Physiol. 71(6): 2402-2411, 1991.-To clarify the impact of vigorous physical training on in vivo insulin action and glucose metabolism independent of the intervening effects of concomitant changes in body weight and composition and residual effects of an acute exercise session, 10 lean, 10 obese, and 6 dietcontrolled type II diabetic men trained for 12 wk on .a cycle ergometer 4 hfwk at -70% of maximal 0, uptake (VO, m,) while body composition and weight were maintained by refeeding the energy expended in each training session. Before and 4-5 days after the last training session, euglycemic hyperinsulinemic (40 mU. m2. min-l) clamps were performed at a plasma glucose of 90 mgldl, combined with indirect calorimetry. Total insulin-stimulated glucose disposal (M) was corrected for residual hepatic glucose output. Body weight, fat, and fat-free mass (FFM) did not change with training, but cardiorespiratory fitness increased by 27% in all groups. Before and after training, M was lower for the obese (5.33 t, 0.39 mg . kg FFM-1 min-’ pretraining; 5.33 t 0.46 posttraining) than for the lean men (9.07 t 0.49 and 8.91 -+ 0.60 mg . kg FFM-1 . min-l for pretraining and posttraining, respectively) and lower for the diabetic (3.86 t 0.44 and 3.49 t 0.21) than for the obese men (P < 0.001). Insulin sensitivity was not significantly altered by training in any group, but basal hepatic glucose production was reduced by 22% in the diabetic men. Thus, when intervening effects of the last exercise bout or body composition changes were controlled, exercise training per se leading to increased cardiorespiratory fitness had no independent impact on insulin action and did not improve the insulin resistance in obese or diabetic men. l
obesity; diabetes mellitus; euglycemic hyperinsulinemic indirect calorimetry; body composition
REDUCED insulin-mediated
clamp;
glucose uptake due to insulin resistance is a characteristic of both obesity and non-insulin-dependent diabetes mellitus (17). Physical exercise is prescribed widely as an adjunct in the treatment of both obesity and diabetes, as a modality for the induction of a negative energy balance, and also for the impact of exercise on in vivo insulin action. Although a single bout of physical exercise improves glucose metabolism acutely (4, 5), the impact of a regimen of physical training on metabolism is unclear. Cross-sectional comparisons indicate that basal and stimulated plasma insulin levels are lower and insulin sensitivity is greater in physically 2402
trained than in sedentary individuals (20). However, physically trained individuals are usually leaner than sedentary people, and it is likely that the differences can be attributed to the lower body fat rather than the level of cardiorespiratory fitness itself. Longitudinal studies involving controlled physical training are potentially difficult to interpret because of concomitant changes in other factors, such as the state of energy balance and changes in body composition, as well as the residual effects of the last exercise bout, which independent of physical training can exert strong metabolic effects for as long as 36 h postexercise. An acute bout of exercise enhances glucose uptake (5,9), and weight loss itself is also associated with improved insulin action (29). Furthermore, when physical training is accompanied by weight loss, insulin resistance is lessened (3). However, the specific changes in glucose metabolism induced by chronic exercise training leading to enhanced cardiorespiratory fitness are difficult to determine, because the effect of long-term training can be confounded by residual effects of the last isolated exercise bout, and the energy deficit induced by long-term rigorous physical training may lead to weight loss and/or changes in body composition, which in themselves are likely to alter glucose metabolism and insulin sensitivity. It has not been established whether exercise training without even a small reduction in body fat content and body weight can alter glucose metabolism. There is considerable evidence that a defect in energy expenditure, specifically in the component of total daily energy expenditure related to the stimulation of metabolic rate after the ingestion or infusion of nutrients (postprandial or dietary-induced thermogenesis), may have a role in the pathogenesis and persistence of the obese state (33,37). Several investigations have demonstrated a link between reduced insulin-stimulated glucose uptake and blunted thermogenesis (13,32). In fact, insulin resistance leading to impaired insulin-mediated glucose disposal may be the mechanism fur the blunted thermic responses to ingested or infused nutrients in obese and non-insulin-dependent diabetic individuals (13). Although it is clear that exercise itself profoundly increases energy expenditure in proportion to the intensity of exercise, the impact of physical training on energy expenditure (thermogenesis) remains unresolved (6,22). If impaired insulin-mediated glucose disposal is associated with the blunted thermogenic responses of obese and diabetic subjects, as previous research has suggested,
0161-7567/91 $1.50 Copyright 0 1991 the American Physiological Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.108.009.184) on October 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
EXERCISE
TRAINING
AND
GLUCOSE
METABOLISM
2403
(Koladex, Custom Laboratories, Baltimore, MD) was consumed within 3 min, and venous blood samples were drawn at 30-min intervals for 2 h and analyzed for plasma glucose and insulin. The subjects voided just before the glucose load, and urine samples were collected for the duration of the test for determination of urinary glucose loss. The integrated areas under the glucose and insulin curves were calculated. All the diabetic men met the following criteria for inclusion in the study: 1) fasting plasma glucose >140 mg/dl on more than one occasion, 2) 120-min OGTT plasma glucose >200 mg/dl, and 3) 309, 600, or 900min OGTT plasma glucose >200 mg/dl. The OGTT was administered twice before and twice 24 days after the last exercise session of the physical training program, and the average of the values obtained from the two tests before and the two tests after training was used. Body composition. Body fat content and FFM were deMETHODS termined by densitometry. Body density was determined Subjects. Ten lean males with normal glucose tolerby hydrostatic weighing by the method described by ance and no personal or family history of obesity or dia- Akers and Buskirk (1) and was corrected for residual betes, 10 obese males with normal glucose tolerance and lung volume, which was estimated by means of the no family history of diabetes mellitus, and six obese men closed-circuit 0, dilution method of Wilmore (42). Perwith non-insulin-dependent diabetes mellitus between cent body fat and FFM were derived from body density the ages of 25 and 40 yr participated in this study. The by use of the Siri equation (38). Body composition detergroups were matched with respect to age, height, and minations were made before and after the physical trainfat-free mass (FFM, determined densitometrically, see ing program. below). The lean men were ~18% body fat, and the obese Aerobic fitness test. Aerobic fitness was determined by and diabetic men were >28% body fat, determined by a maximal continuous multistage exercise test on an hydrostatic weighing. All subjects were healthy nonsmokelectromagnetically braked cycle ergometer (Robert ers with no personal history of cardiovascular disease. Bosch, Berlin, FRG). Before the test, the subjects were The men were weight stable at the time of the study, with familiarized with cycling on an ergometer at a constant no more than a 2-kg weight loss or gain over the 6 mo pedaling rate while breathing through the apparatus before the study. used for metabolic measurements. The subjects began An oral glucose tolerance test (OGTT) was adminiscycling at a rate of 60 rpm with zero external resistance tered to confirm that the lean and obese men had normal (unloaded cycling). The work rate was increased in 30-W glucose tolerance and to confirm diabetes mellitus in the increments every 2 min until volitional exhaustion was reached and the subject refused to continue despite vocal diabetic group, according to the criteria of the National Diabetes Data Group (27). The duration of diabetes in encouragement, or until he was unable to maintain the the diabetic subjects ranged from 0 (newly diagnosed) to pedaling rate. Ventilatory measurements were made continuously by open-circuit respirometry with the use of a 24 mo; the subjects managed their therapy by diet alone and were not taking any oral hypoglycemic agents, to Horizon metabolic measurement cart (Sensormedics, avoid any confounding metabolic effects of these drugs. Anaheim, CA), which includes a turbine volume transducer, a Beckman OM-11 polarographic 0, analyzer, and Furthermore, to obtain a homogeneous group of diabetic subjects, only non-insulin-dependent diabetic men with a Beckman LB-2 nondispersive infrared CO, analyzer. a high insulin response to the OGTT were included in the The subjects breathed through a nonrebreathing valve study. (Hans Rudolph, Kansas City, MO) and used a mouthpiece and noseclips. The gas analyzers were calibrated All subjects were sedentary, defined as no involvement in regular exercise (23 times/wk) for >2 yr before the before and after each aerobic fitness test with 100% N2, room air, and a gas mixture containing 4% CO, and 16% study and a current physical activity level of 1.15; 2) a leveling off of heart tained, and the protocol was approved by the Institurate, with no more than a 3-beat/min incrementbetween tional Review Board of the Mount Sinai School of Medicine. the final two work loads; and 3) a leveling off of Vo,, with OGTZ’. The subjects were tested in the morning in the less than a 3-ml. kg-’ amin-l increment between the final two work loads. If two or more of these criteria were postabsorptive state after an overnight (12-h) fast. After not met, the test was repeated on another day. a fasting blood sample was drawn, a 75-g glucose solution
and if exercise training improves insulin sensitivity, the capacity for thermogenesis might also be enhanced by means of improved insulin-stimulated glucose disposal. The effect of physical training on insulin-stimulated glucose disposal and energy metabolism has never been compared in lean, obese, and non-insulin-dependent diabetic subjects with rigorous control for changes in body composition. In addition, there is no information about the relationship between training-induced changes in insulin sensitivity and thermogenesis. The present study was designed to determine the independent impact on in vivo insulin action and glucose metabolism of a 12wk program of vigorous physical training during which body weight and body fat content were held constant by refeeding the energy expended in each training session in lean and obese men with normal glucose tolerance and obese men with non-insulin-dependent diabetes.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.108.009.184) on October 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
2404
EXERCISE
TRAINING
AND GLUCOSEMETABOLISM
Submaximal aerobic fitness was determined by estimation of the ventilatory threshold from the test data. The ventilatory threshold is the highest work rate or VU, before VE increases out of proportion to 00, (41) because of stimulation of ventilation by nonmetabolically produced CO,, which derives from the buffering of lactic acid. The graded exercise test was routinely performed before and after the physical training program. In a subset of men, the graded exercise test was repeated 7-10 days after the last training bout to ensure that the observed maximal 60, (iTo,,,) values observed after training were maintained during the period of posttraining metabolic testing and that no detraining effect had yet occurred. The fitness test data were used I) to provide descriptive information about the aerobic fitness level of the subjects before and after physical training and 2) for the individualized prescription of the appropriate training work rate. Physical training. The physical training program consisted of 12 wk of cycle ergometer exercise performed for 1 h and 10 min, four times per week. Each session consisted of 5 min of warm-up at a rate of 25 W, 1 h of cycling at the appropriate intensity, and 5 min of cool-down at a rate of 25 W. Initially, each subject trained at the work rate associated with his ventilatory threshold, determined from the graded aerobic fitness test. Thus the appropriate work intensity was prescribed on an individual basis, based on each subject’s submaximal aerobic fitness, The ventilatory threshold is a physiological parameter, determined from gas exchange parameters, which is made on the basis of the fact that ventilation is stimulated out of proportion to increases in 60, during graded exercise by nonmetabolically produced CO, that derives from the buffering of lactic acid (41). Throughout the training program, the work rate was increased so that it provided a similar stimulus as the subjects increased their level of aerobic fitness. The heart rate associated with the initially prescribed work intensity was recorded, and the work rate was adjusted weekly to maintain the same heart rate during exercise. The ventilatory breakpoint was reassessed periodically during training by means of a submaximal exercise test to ensure that the training intensity was appropriate. 90, and associated variables during the exercise period were measured biweekly throughout the training program to provide information about the energy expended during exercise. Diet and maintenunce of body weight during physical training. Because the objective of this project was to study the effect of physical training on in vivo insulin action and glucose metabolism independent of changes in body weight and body composition, the subjects were required to maintain their initial body weight throughout the duration of the study. Financial incentive was provided to ensure that the subjects complied with the requirement of weight maintenance. Body weight was measured on every test day, before and after training, and at each training session. The energy expended during a training session was measured biweekly for each subject. After each bout of exercise, the calories expended during exercise were refed in the form of Sustacal (Mead Johnson, Evansville,
IN; composition: 24% protein, 21% fat, 55% carbohydrate). None of the subjects had any difficulty maintaining weight during the training program, and none was dropped from the study because of weight gain or weight loss tl kg. Euglycemic hyperinsulinemic insulin clump plus indirect calorimetv. The subjects were tested at 7 A.M. after an overnight 12-h fast, having abstained from any exercise for >4-5 days before the test. After the training program, the clamp studies were performed >3 days after the last exercise bout to distinguish between the effects of acute exercise and chronic exercise training. The euglycemic hyperinsulinemic clamps were performed according to the technique described by DeFronzo et al. (8). When the euglycemic hyperinsulinemic clamp is combined with simultaneous indirect calorimetry, the rates of glucose oxidation and glucose storage can be calculated and the thermic effect of infused glucose and insulin, which is the increase in energy expenditure during the clamp, can be determined. An intravenous catheter was inserted in retrograde fashion in a hand vein, and the hand was placed in a heated box (>7O”C) for the duration of the entire test to provide arterialized venous blood samples (24). An indwelling Teflon catheter was placed in the antecubital vein of the other arm for infusion of glucose and insulin. A primed dose of 25 &i of [3-3H]glucose (New England Nuclear) was administered, followed by continuous infusion of 0.25 &i/min of [3-3H]glucose for 2 h to estimate hepatic glucose production in the basal state, and for the subsequent 2 h to determine hepatic glucose production during the clamp. The [3-3H]glucose was purified by high-performance liquid chromatography. Arterialized blood samples were obtained from the hand vein at 150 min intervals for the first 90 min and every 5 min during the last 30 min of this control period to measure the plateau steady-state plasma glucose specific activity. Steady-state enrichment of [ 3-3H]glucose was achieved in all three groups of subjects. This plateau value was used to calculate basal hepatic glucose production. After the 2-h control period, a primed continuous infusion of regular human insulin (40 mU rnA2 min-l, Humulin R, Eli Lilly, Indianapolis, IN) was administered for 2 h to raise and maintain the insulin concentration at -100 pU/ml. The plasma glucose level was maintained at 90 mg/dl by determining the plasma glucose level every 5 min and adjusting the infusion rate of a 20% dextrose-in-water solution with a servocontrol negative feedback principle (8). In the diabetic men, plasma glucose levels were allowed to decline to -90 mg/dl. This decline occurred within 60 min in all the subjects. To avoid rapid changes in glucose concentrations that might precipitate secretion of hormones counterregulatory to insulin, the rate of decline in plasma glucose was not >l.O mg +dl-l min? Blood was drawn at -30, -15, and 0 min during the baseline period and every 15 min during the clamp for determination of plasma insulin. Plasma [33H]glucose specific activity was measured at 150min intervals for the first 90 min of the insulin infusion and at 5-min intervals during the last 30 min of the study. Indirect calorimetry was applied for the last 30 min of the baseline period and again during the last 30 min of l
l
l
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.108.009.184) on October 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
EXERCISE
TRAINING
AND
GLUCOSE
METABOLISM
2405
the clamp to obtain measurements of energy expenditure Similar analyses were applied to the glucose oxidation and substrate utilization. The Horizon metabolic mea- and storage data obtained during the clamp. surement cart, described above, was used. Comparisons among groups of body composition, aerCalculation. The rate of appearance (Ra) of glucose in obic fitness, resting metabolic rate, physiological rethe plasma was calculated from the [3-3H]glucose spe- sponses to the acute exercise bout, and fasting glucose cific activity by use of the equations of Steele in the and insulin levels before and after training were also steady-state form (39) for the last 30 min of the basal made by application of 3 X 2 two-way ANOVA with reperiod and in the non-steady-state form, as validated by peated measures with group and training as the factors. Radziuk et al. (31), during the last 30 min of the clamp. The relationship between the thermic effect of glucose During the basal period, hepatic glucose production and the rate of glucose storage, before and after training, equals the Ra. The last 30 min of the euglycemic clamp was determined by regression analysis. Before the data study was used to calculate the rate of glucose uptake from the three groups were pooled, tests of the equality (M) by the entire body. The glucose distribution volume of the slopes of these regressions among the groups were of 40 ml/kg body wt was assumed. The rate of total body carried out. glucose disposal during the clamp was considered to be For all statistical analyses the 0.05 level of significance the sum of the rates of glucose infusion, corrected for the was used. glucose space, and the residual hepatic glucose production, if any. Correction was made for urinary glucose loss, RESULTS if any. The lean, obese, and diabetic men were matched with The subjects voided before the start of the test, and urine was collected at the end of the test for measure- respect to age and FFM. Thus FFM was similar among all three groups, whereas percent fat and total body ment of urinary urea nitrogen. The nonprotein respiratory quotient was calculated from the respiratory ex- weight were similar for the obese and diabetic groups change data and urinary urea nitrogen production rates. (Table 1). Body weight, FFM, and body fat did not The rates of carbohydrate and lipid oxidation were cal- change significantly with physical training in any of the culated according to the procedure of Lusk (23), which is groups (Table 1). The ratio of waist to hip circumference, based on a nonprotein respiratory quotient of 0.707 for an index of body fat patterning, was significantly lower in 100% fat oxidation and 1.000 for 100% carbohydrate oxi- the lean men (0.90 t 0.04) than in the obese (0.97 t 0.03) dation. Urinary urea nitrogen excretion was used as an and diabetic men (0.97 t 0.05) before training. The index of protein oxidation, on the basis of the assump- waist-to-hip ratio was not altered by training (0.90 t 0.04,0.97 t 0.04, and 0.96 in the lean, obese, and diabetic tion that 1 g nitrogen = 6.25 g protein. Nonoxidative glucose disposal, which includes storage men, respectively). The initial maximum fitness level was slightly lower for the diabetic men than for the other as glycogen, conversion to C, compounds, and conversion to lipid, was calculated as the difference between the rate two groups (see Table 1). Maximal work load and VO, of carbohydrate oxidation and the total uptake of glucose increased significantly in all three groups by ~27%. Subby the body. The thermic effect of glucose, which is the maximal aerobic fitness (Table 1) increased significantly change in energy expenditure during the clamp study, after training in all groups. Although the work intensity was prescribed on the was estimated as the difference between the energy expenditure during the last 30 min of the baseline period basis of the ventilatory breakpoint determined from the graded exercise test, the actual measured Vo2 during the and the last 30 min of the insulin/glucose infusion. 1 h of exercise training (Table 2) was higher than that Assay procedures. Plasma and urinary glucose levels were measured with a Beckman glucose analyzer II measured at the same work intensity during the 3-min (Beckman Instruments, Fullerton, CA). Insulin determigraded exercise test. Thus, although the pretraining vennations were made by radioimmunoassay with charcoal tilatory breakpoint during the graded exercise test ocabsorption with the use of a human insulin standard (16). curred at 50-55% BOB,,,, the actual training intensity Plasma [ 3-3H] glucose specific activity was determined was 70% V02max (Table 2). The energy expended during by the method of Katz and Dunn (19). Urinary urea ni- the exercise training sessions (Table 2), which indicates trogen was determined by the Kjeldahl method (14). the calories that were refed, increased significantly in all three groups but was slightly lower for the diabetic subPlasma C-peptide levels were determined by radioimmunoassay (21). jects because their fitness level was somewhat lower. Design and analysis of data. Comparison of the effect of ANOVA with repeated measures applied to the relative training on the thermic effect of infused glucose in the training intensity (Table 2) yielded a significant effect three groups was made by application of a 3 X 2 two-way for training but not for group or group-by-training interanalysis of variance (ANOVA) with repeated measures action, indicating that the relative training intensity was (44) to the basal and clamp data with group (lean, obese, similar for the three groups and was higher after training or obese diabetic) and training (pre- or posttraining) as than before, which reflects the increased submaximal as the factors. Parameters of glucose metabolism were ex- well as maximal cardiorespiratory fitness. The extra caloric intake that was refed to compensate for the energy pressed in absolute form (mg/min) and relative to FFM and total body weight. Significant F ratios from the AN- expended during exercise ranged from -2,400 kcal/wk OVA were followed by post hoc comparisons among cell initially to 3,400 kcal/wk by the end of the training promeans following the Newman-Keuls procedure, with use gram. The OGTT data are shown in Table 3. Fasting plasma of the appropriate error terms from the ANOVA (44). Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.108.009.184) on October 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
2406 TABLE
EXERCISE
TRAINING
AND
GLUCOSE
METABOLISM
1. Subject characteristics P Obese
Lean
Age, Yr
Height, cm Weight, kg Pre Post
Percent fat Pre Post Fat-free mass, kg Pre Post Maximal aerobic fitness Work load, W Pre Post
iTO,, ml/min Pre Post Percent increase in VO,, Submaximal aerobic fitness: ventilatory breakpoint Work load, W Pre Post VO,, mUmin Pre Post %Voz ma Pre Post
kcaUsession Pretraining Posttraining ire,, ml/min Pretraining Posttraining %Vo, *ax Pretraining Posttraining
Source
31t2 174t2
3622 17O-t3
NS NS
77.7t3.1
101.7*4.0
L
