点击显示 收起
1 From the School of Medicine and Pharmacology (KLC, VB, LJB, and IBP) and the School of Human Movement and Exercise Science (ARM), University of Western Australia, Western Australian Institute for Medical Research and HeartSearch WA
2 Supported by the National Heart Foundation of Australia, the National Health and Medical Research Council of Australia, and the Royal Perth Medical Research Foundation. 3 Reprints not available. Address correspondence to KL Cox, University of Western Australia, School of Medicine and Pharmacology, Royal Perth Hospital, PO Box X2213, GPO Perth, Western Australia, 6847. E-mail: kaycox{at}cyllene.uwa.edu.au.
ABSTRACT
Background: Overweight and inactivity are associated with impaired glucose tolerance, reduced insulin sensitivity, and diabetes. Few controlled trials have assessed the independent and combined effects of energy restriction and exercise on the prevention of these conditions.
Objective: The objective was to evaluate the independent and additive effects of 16 wk of energy restriction and exercise on glucose and insulin concentrations.
Design: Sixty nonsmoking, overweight, sedentary men aged 20–50 y were randomly assigned to either maintain or restrict their energy intake (4186–6279 kJ/d). Within each of these arms, the subjects were further randomly assigned to either a light-intensity (control) or a vigorous-intensity exercise program for 30 min 3 times/wk.
Results: Fifty-one subjects completed the study. Maximal oxygen uptake increased (
Key Words: Energy restriction • exercise • glucose • insulin • overweight men
INTRODUCTION
Energy restriction and exercise are commonly advocated as lifestyle approaches for the modification of cardiovascular disease risk factors such as obesity, high cholesterol, impaired glucose tolerance, and insulin resistance. Obesity and a sedentary lifestyle have been associated with decreased insulin sensitivity, increased concentrations of blood insulin in the fasting state and after a glucose challenge (1-3). Weight loss has been shown to improve insulin sensitivity but not glucose tolerance, particularly when glucose tolerance is in the normal range (4, 5). Exercise training has been shown to improve insulin sensitivity (6, 7), but with little or no influence on glucose tolerance (1, 8-10). Furthermore, exercise-induced weight loss has been shown to improve glucose and insulin responses, whereas exercise without weight loss has not (11). Some have reported both independent and additive effects of weight loss and exercise on glucose and insulin responses (5, 12), whereas others have not (13).
Few well-controlled randomized studies have assessed the independent and additive effects of energy restriction and vigorous exercise training on glucose and insulin concentrations in persons without diabetes. We took the opportunity to investigate, in a controlled randomized intervention trial, the effects of energy restriction and vigorous exercise on blood pressure and body composition in free-living, sedentary, obese men (14, 15). We now report the effects of 16 wk of energy restriction alone, exercise alone, and energy restriction combined with exercise on glucose and insulin metabolism.
SUBJECTS AND METHODS
Subjects
Overweight, sedentary, nonsmoking men aged 20–50 y were recruited from the community via the media. Five hundred men responded and 260 who met the initial entry criteria attended 2 screening visits. Participants were included if their body weight was between 120% and 160% of ideal weight for height and they had no substantial weight loss (>10 kg) in the preceding 12 mo. Sedentary was defined as fewer than two 30-min sessions of vigorous exercise (>6 metabolic equivalents) per week in the previous 6 mo. If the subjects were engaged in moderate activities, they were excluded if their energy expenditure exceeded the equivalent of the criteria for vigorous exercise. Furthermore, the subjects had to have no musculoskeletal injury that would preclude them from exercising, could not be taking nonsteroidal anti-inflammatory drugs, and had to have no history of diabetes, asthma, or heart, renal, or hepatic disease. Because this study was part of a parent study investigating the response of blood pressure to energy restriction and vigorous exercise (14), the subjects also had to have a blood pressure of 130–160 mm Hg (systolic) and 80–110 mm Hg (diastolic), could not be taking any antihypertensive drugs, and had to have an alcohol consumption of <210 mL/wk. Sixty subjects met all the entry criteria. The study was approved by the University of Western Australia Committee for Human Rights, and all subjects gave written informed consent.
Study design
A 4-wk run-in period included familiarization with the fitness test, diet records, and other intervention procedures. During a 2-wk baseline period at the end of the run-in, assessments were made of fitness, usual dietary intake, and body weight; blood samples were collected; and an oral-glucose-tolerance test was completed. The subjects were then randomly assigned to 1 of 4 groups in a two-factor factorial parallel design study of 16 wk duration. Two groups were asked to maintain their usual dietary habits, whereas the subjects in the other 2 groups were given an individually tailored program that aimed to reduce their total energy intake by 4186–6279 kJ/d (1000–1500 kcal/d); the diet provided protein, fat, and carbohydrates at 15%, 30%, and 55% of energy, respectively. Within each of these 2 dietary arms, the subjects were further allocated to a light-intensity exercise control group or a vigorous-intensity exercise group for three 0.5-h sessions/wk. The light exercise group was used to minimize cointervention bias, a factor not usually accounted for in exercise training studies. The light exercise protocol consisted of a series of slow flexibility exercises once per week and stationary cycling (against zero resistance; Monark Crescent AB, Varberg, Sweden) twice a week. Every second week, subjects substituted one cycling session for a session where they walked slowly, at a rate of 2 km in 30 min. The purpose of the session was to provide variety and to maximize compliance. To further encourage compliance in both groups, the sessions were offered at several different times during the day; make-up sessions were provided at the end of the week for those who missed their scheduled session. The vigorous exercise program consisted of stationary cycling only. Subjects cycled for 30 min at 60–70% of their maximum workload, which was determined from their fitness assessment. This exercise protocol and level of intensity was selected because previous reports had shown a training effect on insulin sensitivity (7). Furthermore, we had previously shown this protocol to be effective in improving fitness and to be well tolerated (16). All exercise programs included a 5-min warm-up before the exercise session and a 5-min cool down after the exercise session and were supervised by a trained exercise leader.
Serum glucose and insulin
After the subjects had fasted overnight, blood samples were collected for the measurement of glycated hemoglobin. Blood samples for the measurement of glucose and insulin were collected with the needle indwelling for 20 min while the subjects rested supine. Subjects were then given a 200-mL drink of Glucotol (Orion Laboratories, Perth, Australia) that contained 75 g glucose. Additional blood samples were drawn 30, 60, 90, and 120 min after the drink for the measurement of serum glucose and insulin. An attempt was made to test the subjects at the same time of the day before and after intervention. Blood was sampled before and after intervention with 24 h between the last exercise session and venipuncture.
Serum glucose from the oral-glucose-tolerance test was assayed on a Beckman Glucose analyzer 2 (Beckman Instruments, Fullerton, CA) with the use of reagents from Boehringer Mannheim (Mannheim, Germany). The CV (assay precision) was 2.6% at a concentration of 4.7 mmol/L and was 3.2% at a concentration of 15.4 mmol/L. Serum insulin was measured by radioimmunoassay (17), and the CV was 9% in the range of 15–154 mU/L. Glycated hemoglobin was measured from red blood cells in blood collected into EDTA-coated tubes by using the colorimetric method (18). The assay had a within batch CV of 2.8% at 7% glycated hemoglobin and of 3.2% at 13% glycated hemoglobin. The between batch CV was 2.6% at 7% glycated hemoglobin and 4% at 13.6% glycated hemoglobin.
Diet records
Three-day diet records were completed at baseline and at the end of intervention. The same dietitian monitored dietary intake throughout the study. The subjects were given careful written and verbal instructions on how to keep detailed and accurate records of food weight and volume on a 3-d diet record every fortnight. The subjects were given a target weight loss of 7–8 kg over the 16-wk study. Weight loss was achieved by substituting low-fat foods for high-fat foods, increasing fruit and vegetable consumption, and substituting complex for refined carbohydrates. All subjects were asked to maintain their consumption of tea, coffee, salt, and alcohol and were advised to make no major changes to their sodium intake. We previously used this dietary regimen and found the target weight loss to be achievable and the dietary changes to be manageable and well tolerated by the participants (19). Those subjects allocated to continue their usual energy intake were also seen fortnightly by the dietitian, who determined (from both the 3-d diet records and through an interview) that no substantial alterations to their usual energy intakes were being made. The maintenance of usual dietary patterns was reinforced at each visit, and the subjects were offered a weight-loss program on completion of the study. The diet records were coded by using the NUTTAB89 database (20).
Alcohol intake was monitored each fortnight and recorded in 7-d retrospective diaries, which were subsequently coded by using standard industry tables of alcohol beverage content. The diary data were used to estimate alcohol consumption (in mL ethanol/wk) before and after intervention. In addition, -glutamyltranspeptidase was measured as a potential objective biomarker of any change in alcohol intake (21).
Physical fitness assessment
Fitness was assessed from a multistage exercise test conducted with an electrically braked cycle ergometer (Siemans-Elema AB; Medicinsk Teknik, Solna, Sweden); the test began at zero workload and increased by 20 W every minute with a pedal rate of 60 rpm until the subjects reached volitional fatigue. Oxygen consumption was measured throughout the assessment and was recorded each minute. Maximum oxygen consumption (
Body weight was measured (to the nearest 1/100th of a kilogram) with electronic balance scales (August Sauter West Germany, FSE Scientific, Perth Australia) while the subjects were wearing light clothing and no shoes. Body composition was assessed by using the underwater weighing technique, and body fat distribution was determined from the waist-to-hip ratio (WHR) (15).
Statistical analysis
SPSS statistical software (release 10.1.0; SPSS Inc, Chicago) was used to analyze the data. A two-factor analysis of variance model with interaction was used to compare the changes in glucose, insulin, glycated hemoglobin, and dietary components. To assess the effect of exercise independent of weight loss on serum glucose and insulin concentrations, the analysis of variance was repeated with the change in weight entered as a covariate. Glucose and insulin values from the glucose tolerance test were calculated by the integrated area under the curve (AUC) method (22). Within-group comparisons of differences between baseline and postintervention measurements were made by using the Bonferroni t test. One-factor analysis of variance and chi-square tests were used to test for any between-group differences in baseline characteristics. Pearson’s correlation coefficients were used to determine relations between variables. It was calculated a posteriori that the study had 80% power at = 0.05 to detect energy restriction or exercise main effects of 15% on glucose AUC and 37% on insulin AUC. Baseline results are expressed as means ± SDs, and all other results are expressed as means and 95% CIs.
RESULTS
Baseline characteristics
Of the 60 subjects who started the study, 8 dropped out because of work commitments, and 1 subject sustained an injury during his everyday activity. At baseline, because of an administrative error, only 46 subjects had fasting insulin concentrations measured, and only 40 subjects had insulin concentrations measured during the oral-glucose-tolerance test. The missing insulin values were spread evenly over the 4 study groups. Baseline fitness, body weight, and glucose and insulin concentrations are shown in Tables 1 and 2. The mean (± SD) fasting glucose concentration was 5.26 ± 0.65 mmol/L, and the fasting insulin concentration was 16.52 ± 20.04 mU/L. Only 1 subject had a fasting glucose concentration >7 mmol/L, and 7 subjects were classified as having impaired glucose tolerance (plasma glucose concentration of 7.8–11.1 mmol/L 2 h after a glucose challenge). The mean total energy intake at baseline was 9929 ± 2930 kJ, with 36.4 ± 7.4% derived from fat, 17.7 ± 3.5% from protein, and 41.1 ± 6.8% from carbohydrate. Alcohol accounted for a mean of 2.42 ± 3.66% of the total energy intake. Baseline alcohol consumption assessed from 7-d retrospective diaries was 98.2 ± 104.6 mL ethanol/wk, which was equivalent to 10 standard drinks/wk. No significant differences in food nutrient or alcohol intakes were observed between groups at baseline.
View this table:
TABLE 1. Baseline characteristics and changes from baseline in body mass, body composition, and fitness for subjects in each of the 4 study groups1
View this table:
TABLE 2. Baseline characteristics and mean changes from baseline in glucose, insulin, and glycated hemoglobin for subjects in each of the 4 study groups1
Changes in fitness and body composition
Details of the changes in fitness and body composition were reported previously (14, 15). All subjects completed the prescribed 48 exercise sessions. The mean training intensity for the vigorous exercise group was 76% of the HR reserve (HRres) compared with an HRres of 18% in the light exercise group. The mean changes in fitness in each of the 4 groups are shown in Table 1
Diet components
Daily energy intake and all dietary components, except protein, were reduced in the group that reduced energy intake relative to the group who maintained their usual diet (Table 3). There was no significant effect of vigorous exercise on intake or nutrients.
View this table:
TABLE 3. Baseline characteristics and mean changes from baseline in daily energy intake and dietary components for subjects in each of the 4 study groups1
In the group that reduced their intake, 21.9% (19.1%, 24.7%) of energy was derived from fat, 45.4% (42.6%, 48.1%) from carbohydrate, and 26.5% (24.4%, 28.5%) from protein. The percentage of energy from fat in the diet was significantly reduced by 14.45% (18.8%, 10.1%; n = 51; P < 0.001) with energy restriction, and the percentage of energy from protein was significantly increased by 8.8% (6.78%, 10.8%; n = 51; P < 0.001) with energy restriction and by 3.0% (1.0%, 5.1%; n = 51; P = 0.02) with vigorous exercise. However, there was no significant change in the percentage of carbohydrate in the diet. The percentage of energy derived from alcohol was unchanged with energy restriction or vigorous exercise, and there was no significant change in dietary fiber or water intake.
Alcohol consumption assessed from the 7-d retrospective diaries was significantly reduced with energy restriction. The difference relative to the usual energy intake group was –52.5 (–93.7, –11.2) mL/wk (n = 51; P = 0.04). This amount is equivalent to the consumption of less than one standard drink per day, and no effect on the biochemical marker -glutamyltranspeptidase was observed.
Serum glucose and insulin
Changes in fasting glucose and insulin from baseline to after the intervention in the 4 groups are shown in Table 2. There was a significant main effect of vigorous exercise compared with that of light exercise on fasting glucose; concentrations in the vigorous exercise group were 0.30 (0.06, 0.54) mmol/L lower than those in the light exercise group at the end of the study (n = 51; P = 0.02). There was no significant main effect of vigorous exercise on fasting insulin (n = 40; P = 0.09). There was no significant effect of energy restriction on fasting glucose or insulin concentrations. Changes in fasting insulin concentrations were available for 40 subjects, and changes in insulin AUC were available for 37 subjects.
Changes in glucose and insulin AUCs over 2 h are shown in Table 2 for the 4 groups of subjects. Relative to light exercise, vigorous exercise had a significant effect on reducing the glucose AUC by 133.5 (59.8, 207.2) mmol · L–1 · 120 min–1 (n = 51; P = 0.01) and the insulin AUC by 3303 (677, 5929) mU · L–1 · 120 min–1 (n = 37; P = 0.02). This represented decreases of 13% and 20% for the glucose and insulin AUCs, respectively. An 8% decrease in the glucose AUC with energy restriction was nearly significant (P = 0.06). Five of the 7 subjects who had impaired glucose at baseline had plasma glucose concentrations in the normal range after the intervention. Energy restriction reduced the insulin AUC by 3925 (1299, 6551) mU · L–1 · 120 min–1 (n = 37; P = 0.01), or 40%. Energy restriction and vigorous exercise had additive effects on the insulin AUC.
The general linear model was repeated with the change in weight entered as a covariate. The reductions in fasting glucose and in the AUCs for glucose and insulin after vigorous exercise remained significant, whereas the reduction in the insulin AUC after energy restriction was no longer significant. When the analyses were repeated with the changes in diet composition also considered as covariates, the reductions in the AUCs for glucose and insulin remained significant. Glycated hemoglobin did not change significantly with energy restriction or vigorous exercise (Table 2). There was no interaction of energy restriction or vigorous exercise on any of the variables. The change in fitness was not significantly correlated with the changes in any of the glucose or insulin variables.
DISCUSSION
The results of our study show an independent effect of vigorous exercise and the glucose and insulin responses to an oral-glucose-tolerance test in sedentary overweight men without diabetes. When vigorous exercise was combined with energy restriction, there was an additive effect. The reported effects of energy restriction, exercise, or both on fasting glucose and insulin have previously been inconsistent; some studies have shown no effects of either intervention (11, 16), whereas others have shown a reduction in fasting insulin and glucose with weight loss but not exercise alone (10, 12). In contrast, we observed a reduction in fasting glucose but not in fasting insulin with vigorous exercise only.
The finding that exercise training resulted in lower insulin concentrations after a glucose challenge is consistent with the findings reported by others (4, 6, 7, 23); however, glucose concentrations were reported to be unchanged (1, 9, 10). Our findings that vigorous exercise independently reduced the insulin AUC by 20%, that energy restriction independently reduced the insulin AUC by 40%, and that both interventions had an additive effect are somewhat novel. We are aware of only one other study that had an appropriate design to evaluate the independent and additive effects of energy restriction and exercise training on glucose and insulin metabolism in obese nondiabetic persons (12). Our findings are consistent with those of a study that used a similar training regimen for 9 mo in older men (12). Other uncontrolled studies have reported additive effects of energy restriction and aerobic and resistance exercise (5), whereas others have not (13).
The role of exercise training in improving glucose tolerance is controversial, particularly in persons with normal glucose responses. In our study, only vigorous exercise reduced the glucose AUC, whereas Dengel et al (12) reported a reduction in the glucose AUC with energy restriction, exercise, and a combination of the 2 interventions. Although 12% of the men in our study had glucose responses to the glucose challenge at baseline that indicated impairment, the group overall had normal glucose tolerance. We showed a 13% improvement in glucose tolerance with exercise training. This finding is similar to that of Dengel et al (12). Our finding of an 8% reduction in the glucose AUC with energy restriction was of borderline significance, whereas similar reductions in other studies have been shown to be significant (10). We were unable to show an additive effect of energy restriction and exercise. The observation that glucose tolerance normalized in 5 of the 7 subjects with impaired glucose tolerance after the intervention was of interest; however, this small number of subjects did not allow for statistical analysis.
The effect of a change in diet composition and weight loss on glucose and insulin metabolism with energy restriction and exercise training is unclear. A low-fat, high-complex-carbohydrate diet has independent effects on glucose and insulin (24), and weight reduction has been shown to independently improve insulin sensitivity (4). In the current study the changes in the insulin AUC with energy restriction and exercise were independent of changes in diet composition and alcohol intake as was the effect of exercise on the glucose AUC. When the change in the glucose AUC was adjusted for the change in alcohol intake, energy restriction resulted in a reduction in the glucose AUC. Our observation that the changes in insulin and glucose with vigorous exercise were independent of any change in weight or body composition in obese men is consistent with the finding of some (25, 26) but not other (11) studies. The difference between our study and that of Dengel et al (12) was that in our exercise only group, the subjects were told to maintain their usual diet as was our control group, whereas in Dengel et al’s study the exercise only group was asked to increase their energy intake to maintain weight. This increase in energy intake, even if controlled for diet composition, may have resulted in different glucose and insulin responses compared with a protocol that maintained usual intake. The strength of our study is that we were able to compare the effects of vigorous exercise compared with control with both groups maintaining their usual diet. This may account for the differences in findings between our study and other studies.
There is debate over whether the effects of exercise training on glucose and insulin metabolism are due to the acute effects of the exercise bout (27). It is unknown how long the acute effects of exercise last, but they may persist from 12 to 60 h after the last bout of exercise (28, 29). However, after a single bout of exercise no improvement in glucose tolerance was seen after 18 h (30). In the present study the mean time from the last exercise session until the oral-glucose-tolerance test was 31 h with a maximum of 72 h. Although unlikely, it is possible that the improved glucose tolerance and blunting of the insulin response to a glucose load may have been due in part to the last bout of exercise. This latter point may partially account for the difference in findings between our study and that of Ross et al (11), who failed to find an improvement in insulin sensitivity with exercise without weight loss when measurements were taken 4 d after the last exercise session.
Nevertheless, with respect to the effect of lifestyle changes on the prevention and management of impaired glucose metabolism, regular exercise training without weight loss may still bestow benefits to the individual. Even if the acute effect of exercise is transient, cumulatively repeated bouts may improve longer-term glucose tolerance and insulin sensitivity. Combined with energy restriction the effect may be of greater magnitude and perhaps longer lasting as seen by the additive effect in this study.
The practical application of these findings deserves comment. The results of the present study and of another (31) suggest that exercise must be regular and vigorous to improve glucose tolerance. A study in older men and women showed a 9% improvement in insulin responses with low-intensity exercise but a 23% improvement with higher intensity (23). A recent study reported improvements in insulin sensitivity with intensive dietary and exercise regimens but not with the current moderate activity recommendations (32). A limitation of our study was that the 16-wk intervention was relatively short term and was conducted in a controlled setting; the excellent adherence rates observed may not be maintainable in the long term or in a less highly supervised setting. Vigorous intensity exercise may not be desirable or acceptable to some groups such as the obese, the elderly, or the habitually inactive. However, even though our protocol involved vigorous exercise, the intensity level was at the lower end of the vigorous range. We also used a non-weight-bearing activity (stationary cycling), which alleviates some of the problems of the overweight and obese doing exercise. This may be the reason why the activity was well tolerated. The availability of stationary cycling in fitness centers and homes increases the practicability of such a protocol for use in the community. Notwithstanding this, moderate-intensity exercise is likely to be more conducive to long-term maintenance. Hence, well-controlled studies are needed to establish whether more-moderate-intensity exercise programs can confer the same metabolic improvements in predominantly normoglycemic subjects that we showed with vigorous exercise. Some studies have shown that moderate activity and walking with dietary changes can reduce the incidence of diabetes in subjects with impaired glucose tolerance (33, 34); others have found that, compared with a more intensive exercise and diet program, moderate changes in diet and exercise were not effective in improving insulin sensitivity (32). Although moderate activity may be more appealing from a public health perspective, these studies used a daily exercise regimen. Such an expectation in previously sedentary persons may also be unrealistic for long-term maintenance.
We showed that weight loss through energy restriction significantly improved the insulin response to a glucose challenge in this group of previously sedentary, middle-aged obese men. Vigorous exercise in the absence of weight loss had an independent effect on both glucose and insulin responses to an oral-glucose-tolerance test. The combination of energy restriction and vigorous exercise resulted in additive effects on the insulin response to an oral-glucose-tolerance test. The magnitude of the improvements in glucose and insulin we observed is similar to that observed in other studies (12, 23). These improvements are of clinical significance because strategies that involve the use of pharmacologic agents such as metformin report improvements in insulin sensitivity of 25% (35). Another potential limitation of our study was that the low number of insulin samples may have lessened the magnitude of the change in the insulin AUC. However, because this change was significant with both interventions, our conclusions remain unchanged. Furthermore, even though the vigorous exercise program attenuated the loss of lean body mass with energy restriction, such a loss has undesirable effects on glucose metabolism. This may have lessened the effect of the exercise on the glucose and insulin responses. Both of these lifestyle interventions have independent and sometimes additive influences on other diabetes risk factors and they are a potent strategy that should be an integral part of any program to reduce the risk of impaired glucose tolerance, insulin sensitivity, and diabetes in overweight and sedentary persons.
ACKNOWLEDGMENTS
We are indebted to Penny Rogers for her valuable nursing experience.
KLC was responsible for the data collection and analysis, the interpretation of the results, the writing of the manuscript, and the concept, design, and formulation of the study hypothesis. VB contributed to the analysis and interpretation of the results and edited the manuscript. ARM contributed to the study design and protocol and edited the manuscript. LJB contributed to the study design, formulation of the hypothesis, and interpretation of the results and edited the manuscript. IBP contributed to the study design and the formulation of the hypothesis, the data collection, the interpretation of the results, and the writing and editing of the manuscript. None of the authors had any conflict of interest with the organization supporting this research.
REFERENCES