JENKINS, NATHAN T.; HAGBERG, JAMES M.
Regular physical activity and adoption of other healthy lifestyle habits may prevent or delay the onset of type 2 diabetes mellitus (T2DM) (18,30). The prediabetic state is characterized by impaired fasting glucose (IFG) or impaired glucose tolerance (IGT) and is associated with a substantially elevated risk for development of T2DM (1). Therefore, prediabetes is a condition in which primary prevention efforts through lifestyle modification are particularly important.
Several large-scale prospective trials have examined the role of physical activity as part of more comprehensive lifestyle interventions in prediabetic individuals to prevent or delay the transition to overt T2DM. The Da Qing Diabetes Prevention trial (24) found that exercise training with or without dietary cointervention was more effective than dietary changes alone in preventing T2DM. Other trials (12,16,22,32) using multifactorial lifestyle interventions, with exercise combined with diet and/or weight loss programs, have found that T2DM can be effectively prevented or delayed in the majority of prediabetic individuals. Further, one of these trials demonstrated that a dietary and exercise intervention in IGT patients can reduce long-term (i.e., 12-yr) mortality compared with IGT patients receiving routine treatment for IGT (basic advice about diet and exercise) by ∼50% (13). Data from several trials indicate that lifestyle interventions are at least as effective as pharmacologic treatments for most prediabetic individuals in terms of prevention of T2DM (4,7,8,16). However, it is unclear whether exercise training can, independently of other lifestyle modifications, reverse abnormal glucose metabolism in prediabetic individuals such that they achieve normoglycemia. Whether prediabetic subjects respond more favorably to an endurance exercise intervention compared with age- and body mass index (BMI)–matched normoglycemic subjects is also not known.
Therefore, the purpose of this study was to determine whether training adaptations in the glucose and insulin responses to an OGTT differed between prediabetic and normoglycemic subjects. We also examined whether a 6-month endurance exercise training intervention, independent of changes in dietary habits or weight loss, could completely reverse the prediabetic state. We hypothesized that (i) prediabetic subjects would improve glucose and insulin responses to OGTT to a greater extent than normoglycemic subjects and (ii) exercise training would completely reverse the prediabetes subjects to a normoglycemic state (i.e., fasting plasma glucose of <100 mg·dL−1 and 2-h OGTT glucose <140 mg·dL−1).
The methods and design of this study have been described in detail previously (15,21,31). Volunteers for the University of Maryland Gene Exercise Research Study were enrolled to assess the genetic determinants of interindividual variation in endurance exercise training-induced changes in cardiovascular disease risk factor phenotypes. Data are presented here from subjects who complied with all requirements of the training program and completed all testing procedures before and after training. Classification of subjects as initially prediabetic or normoglycemic was performed for the specific purpose of this report. Subjects were considered normoglycemic if their fasting plasma glucose was <100 mg·dL−1 and their 2-h plasma glucose was <140 mg·dL−1. Subjects were considered prediabetic if they had IFG (fasting glucose ≥100 mg·dL−1 and <126 mg·dL−1) or IGT (2-h glucose ≥140 mg·dL−1 and <200 mg·dL−1) (1). The institutional review board at the University of Maryland College Park approved all procedures, and participants provided written informed consent.
Inclusion and exclusion criteria.
Subjects were sedentary (aerobic exercise two times or less per week and <20 min per session, sedentary occupation), normotensive or hypertension-controlled (blood pressure of <160/99 mm Hg), nondiabetic (fasting plasma glucose ≤126 mg·dL−1 and 2-h glucose of ≤200 mg·dL−1), nonsmoking men and women age 50–75 yr who had no prior history of cardiovascular disease, were not on lipid- or glucose-lowering medications, and had a BMI of <37 kg·m−2. Women were postmenopausal and maintained their hormone replacement therapy status for the duration of the study. Subjects had at least one National Cholesterol Education Program lipid abnormality (total cholesterol >200 mg·dL−1, HDL-C <40 mg·dL−1, triglyceride >200 but <400 mg·dL−1) and total cholesterol and LDL-C of <90th and HDL-C of >20th percentiles for their age and sex to be excluded for possible familial dyslipidemia. A physician-supervised, symptom-limited maximal treadmill exercise test was conducted with blood pressure and ECG monitoring to ensure participants were free of cardiovascular disease (2). This screening maximal exercise test was conducted separately from the baseline maximal oxygen uptake (V˙O2max) test described later.
An important goal of our study was to examine the effects of exercise training on glucose and insulin metabolism independent of changes in diet composition and diet-induced weight loss. Therefore, qualified subjects completed a 6-wk American Heart Association (AHA) Step I dietary program (3) before they were admitted to the baseline testing phase of the study. Adherence to the prescribed diet (<30% caloric intake from fat, ∼55% from CHO, and ∼15% from protein and <300 mg·d−1 cholesterol intake) was monitored with 7-d food records by a study dietician. Subjects were admitted to the baseline testing and training phases of the study only when they had been weight-stable for 3 wk. During the dietary stabilization period, the normoglycemic group lost 0.7 ± 0.4 kg (mean ± SE), and the prediabetic group lost 0.7 ± 0.5 kg (P > 0.05 for difference between groups) body weight.
Subjects had body composition assessed by dual-energy x-ray absorptiometry (DXA; Lunar Corp, Madison, WI) and intra-abdominal and subcutaneous fat were quantified at L4–L5 by computed tomography. V˙O2max was assessed using a modified Bruce treadmill protocol as previously described (9). V˙O2 was considered maximum using the plateau criteria (i.e., increase in V˙O2 of <150 mL O2·min−1 with increased work rate). Secondary criteria of maximal effort, including RER of >1.10, and a maximum HR within 10 beats of the age-predicted max, were used in the absence of a clear plateau in V˙O2. A 75-g oral glucose tolerance test (OGTT) was conducted in the morning after an overnight fast, and plasma samples were analyzed for glucose and insulin concentrations. Glucose was measured using a YSI 2300 Glucose Analyzer (Yellow Springs Instruments, Yellow Springs, OH). Insulin was measured by radioimmunoassay as described previously (21). Areas under the glucose and insulin curves (AUC) were calculated using the trapezoidal principle (20). The homeostasis model assessment of insulin resistance (HOMA) was calculated as described previously (19).
Participants performed multimodal (treadmill, cycle, and rowing ergometers) endurance exercise training at an initial volume of 3 d·wk−1 for 20 min·d−1 at an HR corresponding to 50% V˙O2max. Duration increased by 5 min per session per week for 4 wk until reaching 40 min per session. Intensity was then increased by 5% V˙O2max·wk−1 for 4 wk, until reaching 70% V˙O2max. Intensity was verified using HR monitors, and these sessions were performed under close supervision of study personnel. An unsupervised 45- to 60-min walk at home was incorporated into the program after 12 wk. Only subjects who completed >75% of the prescribed training sessions were included in the final analyses. Dietary logs were examined by a staff dietician on a weekly basis throughout the training program, and subjects were counseled to maintain their habitual caloric intake while on the AHA Step I diet. It was expected that some subjects would lose weight gradually as a result of the energy expenditure associated with the exercise training program. Subjects were weighed weekly and were instructed to correct their dietary habits if they gained weight or if they lost more weight than expected from the exercise energy expenditure. Subjects were excluded from all final analyses if they had a change in body mass greater than ±5% from baseline.
All baseline assessments were repeated after the 24-wk training program. Because regular exercise-associated effects on glucose tolerance are quickly lost upon cessation of training (27), blood sampling and OGTTs were conducted 24–36 h after the final exercise bout.
Data were analyzed using SPSS software (Chicago, IL). Assumptions of normality and homoscedasticity were tested for all variables. Repeated-measures ANOVA was used to test for interactive and main effects of training and prediabetes status on outcome variables, with training (baseline vs after training) as the within-subjects factor and prediabetes status (normoglyemic vs prediabetic) as a between-subjects factor. Time was also included as a repeated factor for the OGTT analysis (0, 30, 60, 90, 120, and 180 min). Age, sex, hormone replacement therapy status, and race were examined as potential confounding variables using ANCOVA. To determine the effects of training on metabolic adaptations independent of changes in body composition, change scores for body composition variables that differed between groups (i.e., BMI, % total fat, and % trunk fat) were examined as covariates in between-group comparisons of training-induced changes in glucose and insulin variables. Covariates were only included in the final analyses when the association between the potential covariate and the outcome variable reached a threshold of P ≤ 0.10. A criterion of P ≤ 0.05 was used for statistical significance of between-group and training-induced differences in outcome variables. Data are presented as means ± SE.
Subject demographics and baseline characteristics.
Analysis of fasting and OGTT plasma glucose data from the 264 participants who originally completed the screening, dietary stabilization, and baseline testing phases of the study identified 196 normoglycemic subjects and 68 prediabetic subjects. Within the normoglycemic group, 61 subjects withdrew from the study for personal or medical reasons, 10 subjects were excluded from the analysis for poor attendance (<75% of exercise training sessions completed), and 6 subjects were excluded for noncompliance with the dietary requirements. Within the prediabetic group, 16 withdrew for personal or medical reasons, 2 subjects were excluded for poor attendance, and 3 were excluded for noncompliance with the dietary requirements. Thus, a total of 116 normoglycemic subjects and 47 prediabetic subjects were included in the analysis for the present report. The percentage of subjects excluded because of noncompliance or withdrawal was not statistically different between groups (39% for normoglycemic group, 31% for prediabetic group; χ2 test, P = 0.28). Within the prediabetic group, 15 subjects had IFG, 22 had IGT, and 10 had both IFG and IGT. The racial makeup of the normoglycemic group was 75% white (n = 89), 18% black (n = 22), and 7% (n = 8) Hispanic, Asian/Pacific Islander, Native American/Alaskan native, or other races. The racial makeup of the prediabetic group was 60% white (n = 28), 30% black (n = 14), and 10% (n = 5) Hispanic, Asian/Pacific Islander, Native American/Alaskan native, or other races. ANCOVA indicated that the difference between groups in racial makeup statistically affected only the insulin data. Therefore, racial differences between the prediabetic and normoglycemic groups were statistically controlled for by including race as a covariate in the ANCOVA models examining the effects of prediabetes and exercise training on insulin and related variables (e.g., HOMA).
At baseline, age, sex, body weight, and BMI did not differ significantly between normoglycemic and prediabetic groups, and the two groups were generally similar in terms of cardiovascular disease risk factors (Table 1). Diastolic blood pressure was 5 mm Hg lower in prediabetic subjects than normoglycemic subjects (P = 0.005), and serum triglyceride and cholesterol levels tended to be higher in prediabetic subjects (P = 0.051 and P = 0.08, respectively). Normoglycemic subjects had higher V˙O2max levels (mL·kg−1·min−1 and mL·kg−1 fat-free mass (FFM) ·min−1) than prediabetic subjects (P < 0.05). Both groups experienced significant training-induced increases in V˙O2max, and the magnitude of the increase was identical between groups (normoglycemic = 3.72 mL·kg−1·min−1, P < 0.001; prediabetic = 3.71 mL·kg−1·min−1, P < 0.001; between-group difference, P > 0.05).
Prediabetic subjects had similar BMI, percent total body fat, and abdominal subcutaneous fat levels compared with normoglycemic subjects at baseline (P > 0.05; Table 2). Groups differed or tended to differ in baseline levels of percent trunk fat (P < 0.05) and intra-abdominal fat levels (P = 0.052). Prediabetic subjects experienced a greater training-induced change in BMI than normoglycemic subjects (P < 0.05), and their reduction in percent trunk fat tended to be greater than normoglycemic subjects (P = 0.09). The change in FFM tended to be greater in the normoglycemic group than in the prediabetic group (P = 0.09). There were no statistically significant differences between groups in training-induced changes in percent total fat, intra-abdominal fat, or abdominal subcutaneous fat.
OGTT responses before and after training.
Fasting plasma glucose levels were higher in prediabetic subjects both before and after training (P < 0.05). Plasma glucose levels of prediabetic subjects also were significantly higher than those of normoglycemic subjects at each OGTT time point before and after training (Fig. 1A; all P < 0.05). The training-induced changes in glucose levels at 0, 60, 90, and 120 min of the OGTT were greater in prediabetic compared with normoglycemic subjects (Fig. 1B; P < 0.05). There were significant between-group differences in glucose AUC before and after training (P < 0.05), with only the prediabetic group having a significant training-induced reduction in glucose AUC (Fig. 1C; P < 0.05). The change in glucose AUC with training was virtually zero in the normoglycemic group, whereas the prediabetic group experienced a ∼5% reduction in glucose AUC (Fig. 1D; P < 0.05 vs baseline and vs P < 0.05 normoglycemic group). Groups differed significantly in insulin levels at 60, 90, 120, and 180 min of the OGTT at baseline and after training (Fig. 2A; all P < 0.05). Both groups experienced significant training-induced reductions in insulin levels at each time point of the OGTT (Fig. 2B; P < 0.05); however, the change was approximately two to three times greater for prediabetic subjects than for normoglycemic subjects at 90 and 120 min (P < 0.05). Insulin AUC was higher in prediabetic subjects at baseline and after training (P < 0.05), and both groups had lower insulin AUCs after training than at baseline (Fig. 2C; P < 0.05). Changes in insulin AUC were significant for both groups (P < 0.05); however, the prediabetic group had a 75% larger change with training than the normoglycemic group (Fig. 2D; P < 0.05).
HOMA index values were ∼35% higher in prediabetic than in normoglycemic subjects at baseline and after training (Fig. 3; P < 0.05), and both groups had significant training-induced reductions in the HOMA index (P < 0.05). There was a tendency (P = 0.095) for the magnitude of the change be greater in the prediabetic group (−16%) than the normoglycemic group (−11%).
The 2-h glucose value for the prediabetic group was, on average, still >140 mg·dL−1 after training. However, 17 of the 47 prediabetic subjects achieved normoglycemia after training. The breakdown of these prediabetic “responders” and “nonresponders” was as follows: 4 of the 15 individuals with IFG only normalized their fasting glucose to <100 mg·dL−1 after training, 10 of the 22 individuals with IGT only normalized their 2-h glucose to <140 mg·dL−1 after training, and 1 of the 10 individuals with both IFG and IGT achieved normoglycemia after training. The “responder” individuals who normalized their glucose metabolism and nonresponders were not statistically different in any variable, including age, race, cardiovascular risk factors, or body composition (P > 0.05). In addition, responders and nonresponders did not differ in magnitudes of training-induced changes in any body composition–related variables (P > 0.05).
Independent effects of prediabetes and training.
Importantly, insulin and HOMA data (Figs. 2 and 3) are adjusted for race because race was the only covariate found to significantly influence the associations between prediabetes and insulin variables when included in ANCOVA statistical models. Other potential covariates (age, sex, BMI, or hormone replacement therapy status) did not significantly mediate the associations between prediabetes status and any outcome variables and were therefore not included in any statistical models.
We also used ANCOVA to determine whether the changes in BMI and percent trunk fat explained the observed between-group differences in the training-induced changes in glucose and insulin responses to the OGTT. The changes in BMI and percent trunk fat were chosen for examination as potential confounders because the changes in these variables differed or tended to differ between groups (Table 2; P < 0.05 and P < 0.10 for BMI and percent trunk fat, respectively). Results of the ANCOVA revealed that: (i) inclusion of the change in BMI as a covariate had no effect on any of the between-group differences in training-induced changes in glucose, insulin, or HOMA variables (data not shown; patterns and statistical significance were identical with data shown in Figs. 1–3); (ii) inclusion of the change in percent trunk fat as a covariate had no effect on the association between prediabetes and training-induced changes in insulin variables or the HOMA index (data not shown); and (iii) inclusion of the change in percent trunk fat as a covariate indicated that the between-group differences in glucose at different points of the OGTT (Fig. 4A) and total glucose AUC (Fig. 4B) were related to the change in percent trunk fat. Specifically, the association between prediabetes and the change in glucose AUC was no longer statistically significant after adjusting for the change in percent trunk fat. However, the percent trunk fat–adjusted between-group difference in fasting plasma glucose (time 0 of the OGTT) remained statistically significant and displayed the same pattern as without adjustment, with normoglycemic subjects experiencing a significant increase and prediabetic subjects experiencing a significant decrease relative to baseline (P < 0.05).
The main findings of this study are (i) that prediabetic individuals experience greater endurance exercise training–induced reductions in glucose and insulin OGTT responses than normoglycemic individuals, but (ii) these changes are not sufficient, on average, to completely normalize baseline differences between normoglycemic and prediabetic groups. However, prediabetes can be reversed by exercise training alone in some individuals. In addition, the effects of exercise training on glucose and insulin OGTT responses in the prediabetic state are largely dependent on exercise training directly because both groups experienced minimal changes in weight and maintained strict adherence to dietary habits throughout the course of the intervention, and statistically controlling for any between-group differences in changes in BMI and body fatness produced, in general, similar results compared with unadjusted analyses.
Prediabetic subjects have greater metabolic adaptations to endurance training.
The effects of exercise training on glucose and insulin OGTT responses have not been adequately compared between normoglycemic and prediabetic subjects. However, there is some precedence for the trends evident in the present study. For example, in one study showing a favorable effect of moderate-intensity aerobic training in older individuals, the effect was driven by a subset of subjects who had IGT at baseline (10). In addition, it is generally understood that healthy normoglycemic individuals do not experience large, if any, changes in glucose tolerance with exercise training, especially compared with individuals with T2DM (14,17). From our data, it seems that metabolic adaptations to exercise training are also greater in the prediabetic state. These metabolic adaptations in prediabetic group are clearly clinically important, as the 15% reduction in the HOMA index represents a partial restoration toward the normoglycemic state, and ∼50% of the initial difference in the HOMA index between the groups was eliminated by exercise training. In addition, while our study quantifies the extent to which exercise training may help these patients independent of other lifestyle factors, the prediabetes status does not seem to be completely reversed with exercise alone. Perhaps the addition of a small degree of weight loss may further help prediabetic individuals to normalize their glucose metabolism.
Metabolic improvements with exercise training are insufficient to completely reverse prediabetic state.
Whereas the finding of a relationship between prediabetes status and magnitude of training-induced adaptations in glucose tolerance is a novel result and a critical finding of our study, training did not completely normalize between-group differences in glucose and insulin responses to an OGTT (i.e., groups still differed substantially after training). Furthermore, on average, the prediabetic subjects still remained glucose intolerant after training (i.e., 2-h glucose >140 mg·dL−1), despite having a significant training-induced improvement in glucose tolerance, and although 30%–40% of the initially prediabetic subjects achieved normoglycemia after training. These findings must be considered in light of several unique aspects of our study. Body weight and composition remained generally stable throughout the intervention. In addition, subjects were weight-stable for 3 wk before initiating the training program, and importantly, prediabetic and normoglycemic groups lost similar amounts of weight during the dietary stabilization phase. Finally, dietary habits of all subjects were monitored by a dietician throughout the intervention, and subjects were excluded from the final analysis if they were not compliant with the prescribed AHA Step I diet. Therefore, we are confident that we have isolated the effects of exercise training per se independent of dietary factors or weight loss, and thus we are able to conclude that training per se promotes favorable metabolic adaptations in the prediabetic state, but training alone is not sufficient to completely normalize differences between the groups in glucose tolerance.
Role of diet/weight loss versus exercise interventions in metabolic changes in prediabetes.
The literature is mixed regarding possible relationships between changes in dietary habits and/or body composition and glucose metabolism in prediabetic individuals. Several large-scale trials have targeted patients with IGT with the goal of reducing the rate of progression to overt T2DM (16,24,25,30). These trials have generally concluded that the degree of weight loss with lifestyle intervention—a mixed intervention with goals of increasing energy expenditure through physical activity and reducing caloric intake—is related to reduced risk for progression to T2DM. However, one of these large-scale trials (the Da Qing study ) compared the effects of exercise alone, diet alone, and exercise + diet on the risk of development of T2DM and found that exercise with or without changes in dietary habits was more effective than diet alone in preventing T2DM. Thus, the roles of diet, training, and their combination in prevention of T2DM are not yet clear. It is important to mention that our data set differs substantially from the Da Qing trial and other aforementioned large-scale trials in that we have compared prediabetic individuals to normoglycemic individuals undergoing the same intervention, whereas other trials included only prediabetic subjects assigned to different treatment conditions with the goal of preventing the progression to T2DM. We must also mention that our study did not track the progression to T2DM, and we cannot rule out the possibility the effects we have observed are transient. In addition, to mirror the Da Qing trial, we would have needed to include an additional group of sedentary individuals. Overall, the information currently available suggests that it is likely that regular exercise can prevent or delay the progression from prediabetes to T2DM, but whether exercise training can reverse prediabetes to normoglycemia needs substantial further research.
We attempted to rule out confounding effects of diet and changes in body composition by using an initial 3-wk dietary education and stabilization program and having a study dietician monitor dietary records throughout the study to ensure that subjects maintained their eating habits and did not attempt to lose weight during the program. We further sought to gain insight into this issue in our cohort by statistically examining the role of training-induced body composition changes in glucose and insulin adaptations using ANCOVA. These analyses revealed that changes in glucose responses to the OGTT were significantly related to changes in trunk fat, as indicated by the elimination of between-group differences in changes in glucose concentrations at 60, 90, and 120 min of the OGTT with the inclusion of the change in trunk fat as a covariate. However, the between-group difference in the change in fasting glucose values and total glucose AUC remained statistically significant in the ANCOVA model, and there were no other relationships between changes in body composition and changes in glucose, insulin, or the HOMA index. Furthermore, although a substantial proportion of our prediabetic subjects did in fact achieve normoglycemia after training, no differences in baseline or training-induced changes in any body composition–related variables were observed between prediabetic responders and nonresponders. Therefore, it seems that in prediabetic individuals, exercise training–induced changes in glucose metabolism are only partly related to changes in body composition. This is in agreement with previous studies, indicating that changes in body composition are not required for exercise training–induced improvements in glucose metabolism (5,23,28). However, we cannot rule out the possibility that glucose tolerance improvements would have been more pronounced in the presence of further weight loss through caloric restriction in these prediabetic subjects.
Role of exercise intensity, duration, and frequency for optimal treatment of prediabetes.
It is also possible that exercise training alone can completely normalize the prediabetic condition, but our training program was not of sufficient intensity, duration, and/or frequency to do so for most individuals. Given recent data showing improvements in insulin sensitivity after a 2-wk high-intensity interval training program in normoglycemic healthy young individuals (26), we speculate that if the intensity of our program had been higher, then perhaps the baseline differences between groups would have diminished to a greater extent as a result of training. It will be interesting for future prospective trials to test the recently proposed hypothesis (11) that high-intensity interval training programs will have insulin-sensitizing effects of greater magnitude for prediabetic individuals compared to traditional moderate-intensity programs. However, whether high-intensity interval training will be practical for widespread use in this population is doubtful.
Our findings differ from those of an important previous study demonstrating complete normalization of glucose tolerance in some patients with T2DM or IGT. Holloszy et al. (14) found that after a 12-month high-intensity (up to 90% V˙O2max), long-duration (up to 60 min), high-frequency (up to five sessions per week) endurance training program, glucose tolerance was completely normalized in most of their subjects who had T2DM and all subjects who had IGT before training. It is interesting to compare the results of the IGT subjects with the prediabetic subjects of our study. Although subjects in the earlier study were not instructed to lose weight, the energy expenditure from their intensive exercise training program was sufficient for subjects to lose an average of ∼5 kg. In comparison with the intervention of Holloszy et al., our training program, designed to be consistent with current exercise recommendations for sedentary older adults, was lower in frequency (maximum of 4 d·wk−1 in our study vs 5 d·wk−1 in Holloszy et al. ), duration (maximum of 40 vs 60 min), intensity (maximum of 70% V˙O2max vs 90% V˙O2max), and length of program (6 vs 12 months). As a result, it is likely that high-intensity, high-volume exercise training with accompanying weight loss is more appropriate for reversal of glucose intolerance in diabetic and prediabetic individuals than standard aerobic exercise training programs such as that used in the present study.
We acknowledge that our study is limited by the lack of a control group who remained sedentary throughout the training intervention. Our retrospective analysis of prediabetes status in a cohort of individuals recruited and tested for a study with a different original purpose is also a limitation. Our sample size was sufficient to detect differences between the prediabetic and normoglycemic groups but was too small to perform appropriately powered post hoc analyses with groups stratified by race, sex, or hormone replacement therapy. Because race was identified as a potential confounder in our ANCOVA results, future work should further examine exercise training effects on glucose and insulin metabolism in prediabetes in a race-specific manner. Similarly, our sample size was too small to determine why some prediabetic subjects achieved normoglycemia and some did not. Among our responder prediabetic subjects, the IGT subgroup was most highly represented (10 of the 17 responders), whereas 4 of 15 IFG-only and just 1 subject with both IFG and IGT achieved normoglycemia. Thus, from our data, we can tentatively conclude that an IGT-only prediabetic individual is the most responsive to exercise training, but this will need to be the focus of further research with larger sample sizes of prediabetic subgroups. We did not assess pancreatic function, which could be important in light of recent evidence indicating that the lifestyle-related changes in glucose tolerance of obese T2DM patients may involve increased pancreatic β-cell secretory capacity in addition to changes in peripheral insulin sensitivity (29). Future studies should also address whether the same is true in the setting of prediabetes. The finding that fasting glucose levels increased with exercise training in the normoglycemic group was unexpected but was somewhat consistent with previous data from the HERITAGE study, indicating substantial interindividual variations in training-induced changes in intravenous glucose tolerance (6). Notable strengths of our study include the weight-stable protocol and monitoring of dietary habits throughout the intervention to ensure that subjects adhered to the prescribed AHA Step 1 diet. Therefore, we are confident that we have isolated the effects of exercise training per se on our primary outcomes independent of dietary modifications and substantial weight loss.
In conclusion, metabolic adaptations to endurance exercise training are greater in magnitude in prediabetic subjects than in individuals who were initially normoglycemic. Traditional endurance exercise training without weight loss does not seem to completely reverse impaired glucose and insulin metabolism in all individuals with prediabetes. Our findings indicate a need for future research to determine whether exercise training of greater intensities can “cure” the prediabetic condition. Holistic approaches to lifestyle modification, including increased physical activity in combination with improved dietary habits, weight loss, cessation of smoking, etc., should be emphasized in health promotion efforts for improvement of glucose metabolism and reduction of cardiovascular disease risk in individual prediabetic patients. Nevertheless, future studies should continue to attempt to delineate the precise contribution of each individual component of the lifestyle modification to improvements in health-related outcomes.
This study was funded by R01AG017474 and R01AG015389 (J.M.H.). N.T.J. was supported by a National Institutes of Health predoctoral training grant (T32AG000268 to J.M.H.).
The authors report no conflict of interest.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med. 1998; 15: 539–53.
2. American College of Sports Medicine. ACSM’s Guidelines for Exercise Testing and Prescription. Baltimore (MD): Lippincott, Williams, and Wilkins; 2000. p. 93–4.
3. American Heart Association. Dietary guidelines for healthy American adults: a statement for physicians and health professionals by the Nutrition Committee, American Heart Association. Circulation. 1988; 77: 721–4A.
4. Bertram MY, Lim SS, Barendregt JJ, Vos T. Assessing the cost-effectiveness of drug and lifestyle intervention following opportunistic screening for pre-diabetes in primary care. Diabetologia. 2010; 53: 875–81.
5. Bloem CJ, Chang AM. Short-term exercise improves β-cell function and insulin resistance in older people with impaired glucose tolerance. J Clin Endocrinol Metab. 2008; 93: 387–92.
6. Boule NG, Weisnagel SJ, Lakka TA, et al.. Effects of exercise training on glucose homeostasis. Diabetes Care. 2005; 28: 108–14.
7. Crandall JP, Knowler WC, Kahn SE, et al.. The prevention of type 2 diabetes. Nat Clin Pract Endocrinol Metab. 2008; 4: 382–93.
8. Dagogo-Jack S, Egbuonu N, Edeoga C. Principles and practice of nonpharmacological interventions to reduce cardiometabolic risk. Med Princ Pract. 2010; 19: 167–75.
9. Dengel DR, Hagberg JM, Coon PJ, Drinkwater DT, Goldberg AP. Effects of weight loss by diet alone or combined with aerobic exercise on body composition in older obese men. Metabolism. 1994; 43: 867–71.
10. DiPietro L, Seeman TE, Stachenfeld NS, Katz LD, Nadel ER. Moderate-intensity aerobic training improves glucose tolerance in aging independent of abdominal adiposity. J Am Geriatr Soc. 1998; 46: 875–9.
11. Earnest CP. Exercise interval training: an improved stimulus for improving the physiology of pre-diabetes. Med Hypotheses. 2008; 71: 752–61.
12. Eriksson KF, Lindgarde F. Prevention of type 2 (non–insulin-dependent) diabetes mellitus by diet and physical exercise. The 6-year Malmo feasibility study. Diabetologia. 1991; 34: 891–8.
13. Eriksson KF, Lindgarde F. No excess 12-year mortality in men with impaired glucose tolerance who participated in the Malmo Preventive Trial with diet and exercise. Diabetologia. 1998; 41: 1010–6.
14. Holloszy JO, Schutltz J, Kusnierkiewicz J, Hagberg JM, Ehsani AA. Effects of exercise on glucose tolerance and insulin resistance. Brief review and some preliminary results. Acta Med Scand Suppl. 1986; 711: 55–65.
15. Jenkins NT, McKenzie JA, Damcott CM, Witkowski S, Hagberg JM. Endurance exercise training effects on body fatness, V˙O2max
, HDL-C subfractions, and glucose tolerance are influenced by a PLIN haplotype in older Caucasians. J Appl Physiol. 2010; 108: 498–506.
16. Knowler WC, Barrett-Connor E, Fowler SE, et al.. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002; 346: 393–403.
17. Krotkiewski M, Lonnroth P, Mandroukas K, et al.. The effects of physical training on insulin secretion and effectiveness and on glucose metabolism in obesity and type 2 (non–insulin-dependent) diabetes mellitus. Diabetologia. 1985; 28: 881–90.
18. Manson JE, Nathan DM, Krolewski AS, Stampfer MJ, Willett WC, Hennekens CH. A prospective study of exercise and incidence of diabetes among US male physicians. JAMA. 1992; 268: 63–7.
19. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985; 28: 412–9.
20. Matthews JN, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements in medical research. BMJ. 1990; 300: 230–5.
21. McKenzie JA, Weiss EP, Ghiu IA, et al.. Influence of the interleukin-6 −174 G/C gene polymorphism on exercise training-induced changes in glucose tolerance indexes. J Appl Physiol. 2004; 97: 1338–42.
22. Norris SL, Zhang X, Avenell A, et al.. Long-term effectiveness of weight-loss interventions in adults with pre-diabetes: a review. Am J Prev Med. 2005; 28: 126–39.
23. Oberbach A, Tonjes A, Kloting N, et al.. Effect of a 4 week physical training program on plasma concentrations of inflammatory markers in patients with abnormal glucose tolerance. Eur J Endocrinol. 2006; 154: 577–85.
24. Pan XR, Li GW, Hu YH, et al.. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance: the Da Qing IGT and Diabetes Study. Diabetes Care. 1997; 20: 537–44.
25. Ramachandran A, Snehalatha C, Mary S, Mukesh B, Bhaskar AD, Vijay V. The Indian Diabetes Prevention Programme shows that lifestyle modification and metformin prevent type 2 diabetes in Asian Indian subjects with impaired glucose tolerance (IDPP-1). Diabetologia. 2006; 49: 289–97.
26. Richards JC, Johnson TK, Kuzma JN, et al.. Short-term sprint interval training increases insulin sensitivity in healthy adults but does not affect the thermogenic response to β-adrenergic stimulation. J Physiol. 2010; 588: 2961–72.
27. Rogers MA, King DS, Hagberg JM, Ehsani AA, Holloszy JO. Effect of 10 days of physical inactivity on glucose tolerance in master athletes. J Appl Physiol. 1990; 68: 1833–7.
28. Rogers MA, Yamamoto C, King DS, Hagberg JM, Ehsani AA, Holloszy JO. Improvement in glucose tolerance after 1 wk of exercise in patients with mild NIDDM. Diabetes Care. 1988; 11: 613–8.
29. Solomon TP, Haus JM, Kelly KR, Rocco M, Kashyap SR, Kirwan JP. Improved pancreatic β-cell function in type 2 diabetic patients after lifestyle-induced weight loss is related to glucose-dependent insulinotropic polypeptide. Diabetes Care. 2010; 33: 1561–6.
30. Tuomilehto J, Lindstrom J, Eriksson JG, et al.. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med. 2001; 344: 1343–50.
31. Wilund KR, Ferrell RE, Phares DA, Goldberg AP, Hagberg JM. Changes in high-density lipoprotein-cholesterol subfractions with exercise training may be dependent on cholesteryl ester transfer protein (CETP) genotype. Metabolism. 2002; 51: 774–8.
32. Yates T, Khunti K, Bull F, Gorely T, Davies MJ. The role of physical activity in the management of impaired glucose tolerance: a systematic review. Diabetologia. 2007; 50: 1116–26.