Exercise training is an effective means to improve glycemic control in individuals with type 2 diabetes mellitus (1–4). Other benefits of exercise in this population include modest weight loss, improved insulin sensitivity, and modification of cardiovascular risk factors, including blood pressure and triglycerides (5,6). Patients with type 2 diabetes who participate in regular exercise training have also been found to have improved quality of life (7). The American Diabetes Association recommends that people with type 2 diabetes engage in a minimum of 150 min·wk−1 of aerobic physical activity and at least two weekly sessions of resistance exercise training while minimizing sedentary time (8).
Although clinical practice guidelines recommend exercise, the majority of individuals with type 2 diabetes are not achieving the physical activity recommendations in these guidelines (9). In a study examining the physical activity level in individuals with type 2 diabetes over the age of 65 yr in the United States, Zhao et al. (10) found that only 25% met the American Diabetes Association 2007 guideline recommendations for total physical activity. Another study looking at physical activity levels in the United States reported that 41.1% of individuals with type 2 diabetes met the aerobic exercise recommendations compared with only 12.4% for resistance training (11). As many individuals with type 2 diabetes are not achieving the recommended levels of physical activity, an important question arises when prescribing exercise in this patient population—is adherence to exercise prescription important in terms of glycemic control in the management of type 2 diabetes?
The primary aim of this study was to determine whether a dose–response relationship existed between the level of adherence to prescribed exercise over a 6-month exercise intervention and glycemic control (specifically, change in hemoglobin A1c [HbA1c]) in patients with type 2 diabetes. Our secondary objective was to determine if this association was affected by any of the following factors: modality of exercise, age, sex, or glycemic control before participating in exercise training.
RESEARCH DESIGN AND METHODS
This study is a post hoc analysis of data from the Diabetes Aerobic and Resistance Exercise (DARE) trial, a single-center, randomized controlled trial designed to evaluate in type 2 diabetes the effect of aerobic, resistance, and combined aerobic and resistance exercise training compared with no exercise training on glycemic control over 26 wk. The DARE trial protocol and main results have been previously reported (2); aerobic and resistance training each reduced HbA1c significantly, whereas combined training reduced HbA1c more than either aerobic or resistance training alone. The Ottawa Hospital Research Ethics Board reviewed and approved the study protocol, and informed consent was obtained from all participants. Data collected during the DARE trial were used in this current analysis to determine if there was a dose–response relationship between level of adherence to prescribed exercise and glycemic control.
Setting and participants
Supervised exercise training occurred at eight community-based facilities in the Ottawa-Gatineau region in Ontario and Quebec, Canada. Inclusion and exclusion criteria have been described previously (2). Among the 251 participants in the trial, 185 were randomized to exercise groups.
Participant screening and randomization
Details regarding participant screening and randomization were reported previously (2). In brief, participants were eligible for randomization if they met all inclusion and exclusion criteria, had a normal exercise stress test or were cleared by cardiology, and completed a minimum of 10 of 12 prescribed exercise sessions during a 4-wk run-in phase. Central randomization with allocation concealment before randomization was used with block sizes varying randomly between four and eight. Randomization was stratified by both sex and age (<55 and ≥55 yr).
Participants completed individually supervised, progressive exercise training programs within the exercise training group to which they were assigned (aerobic, resistance, or combined). Exercise training took place at community-based facilities three times per week for 6 months. Participants received gym memberships for the duration of the study. Each training session began with a 5- to 10-min warm-up and ended with a 5- to 10-min cool-down. Participants in the aerobic exercise training group progressed gradually in intensity and duration to 45 min at 75% of maximum heart rate (determined from a treadmill test) per session. Participants in the resistance training group progressed to three sets of eight repetitions of seven exercises at an intensity of eight maximal repetitions. The combined exercise training group completed both the aerobic and resistance training components.
Participants assigned to the control group were encouraged to maintain their usual pretrial level of physical activity from baseline to the end of the intervention period. As incentive to complete the study, control participants were offered the exercise training program of their choice at the end of the intervention period as well as a 6-month membership to a community-based training facility.
All participants were encouraged to follow a diet recommended by the Canadian Diabetes Association that would provide 90% of the energy intake needed to maintain their weights throughout the study and were followed by a dietician at 3-month intervals. Changes to glucose-lowering therapy, antihypertensives, and lipid-altering agents were discouraged unless clearly medically indicated while participants were enrolled in the study. Glucose-lowering therapy was increased in a stepwise manner if HbA1c was 10.5% (91 mmol·mol−1) or greater at the midintervention assessment, and glucose-lowering therapy was reduced systematically in response to frequent hypoglycemia.
Outcomes and follow-up
The primary outcome of the DARE trial was the change in HbA1c between baseline and the end of the 6-month supervised exercise intervention. Participants were assessed at baseline, midintervention (3 months), and end-of-intervention (6 months).
Adherence to prescribed exercise was recorded in two ways. First, at each session, participants entered detailed, structured logs recording their exercises in real time as they completed them, and these logs were collected and reviewed regularly by the personal trainers. Second, personal trainers provided direct individual supervision twice weekly during the run-in period, weekly for the first 4 wk of the intervention, and biweekly for the remainder of the intervention. The personal trainers had completed a Canadian Society for Exercise Physiology certification, or were in the final year of a Bachelor of Science in Human Kinetics and working toward their Canadian Society for Exercise Physiology certification. Attendance was recorded at each supervised session.
Only participants randomized to exercise groups (not the control group) with at least one postbaseline HbA1c result were included in this analysis. Adherence to exercise was calculated by dividing the number of sessions participants attended by the number of prescribed sessions. Adherence was reported using medians and interquartile ranges (IQR) due to skewness of the data. HbA1c changes between baseline and the end of the 6-month intervention were reported as means with standard deviations. Change in HbA1c was reported for three prespecified categories of exercise adherence: 1) <70%, 2) 70 to 90%, and 3) >90%.
Simple linear regression was used to assess the significance of the dose–response relationship between adherence to prescribed exercise in each modality exercise group and change in HbA1c between baseline and the end of the 6-month intervention. To further explore baseline characteristics that might contribute to this dose–response relationship, participants were stratified within each modality exercise group by 1) age (<55 and ≥55 yr), 2) sex, and 3) baseline HbA1c (<7.5% [58 mmol·mol−1] and ≥7.5% [58 mmol·mol−1]). These analyses were limited to individuals for whom HbA1c data were available at baseline and 6 months. For participants who were missing HbA1c data at 6 months but for whom HbA1c data were available at 3 months, a last observation carried forward analysis was used: seven (10.9%) from the aerobic exercise training group, 10 (16.7%) from the resistance training group, and two (3.2%) from the combined aerobic and resistance exercise training group. Three individuals who were randomized to an exercise training group had no HbA1c data available at either 3 or 6 months and, therefore, were excluded from the analysis: one (1.6%) from the aerobic exercise group, one (1.7%) from the resistance group, and one (1.6%) from the combined aerobic and resistance exercise group.
Data analyses were conducted using Stata (Version 15.1; StataCorp, College Station, TX). The prespecified α for all analyses was 0.05.
There were 185 participants included in this analysis. Baseline characteristics are presented in Table 1. Overall, participants had a mean age of 54.1 yr (SD, 7.2 yr) and 36% were female. The mean baseline HbA1c was 7.69% (61 mmol·mol−1) (SD, 0.87% [10 mmol·mol−1]). Randomized participants that were excluded as a result of missing data (n = 3) had a mean age of 57.7 yr (SD, 6.2 yr), mean baseline HbA1c of 7.67% (60 mmol·mol−1) (SD, 0.78% [9 mmol·mol−1]), and 67% were female.
Adherence to prescribed exercise
Exercise adherence for the 6-month intervention is presented in Table 2. Participants randomized to the exercise training groups completed a median of 85.9% (IQR, 74.4%–93.6%) of the prescribed exercise sessions. Participants assigned to the combined aerobic and resistance exercise training group had a median adherence of 85.9% (IQR, 74.4%–93.6%) over the 6 months, compared with the resistance training group with a median adherence of 85.9% (IQR, 78.2%–92.3%) and the aerobic exercise training group (median adherence, 84.6%; IQR, 70.8%–96.2%). Twenty-two (37.3%) participants in the aerobic exercise group completed more than 90% of the prescribed exercise sessions compared with 18 (28.5%) in the resistance exercise group and 26 (41.3%) in the combined aerobic and resistance exercise group.
The dose–response relationship between adherence to prescribed exercise and change in HbA1c preintervention and postintervention is presented in Table 3. Overall, a significant dose–response relationship was identified (β = −0.0076, R = −0.170, P = 0.021). Specifically, for every 20% increase in adherence to three prescribed exercise sessions per week (e.g., from 60 to 80%, which is equivalent to two additional exercise sessions per month), HbA1c was reduced by 0.15% (2 mmol·mol−1; 95% CI, 0.02%–0.28%). When stratified analyses were conducted, the dose–response relationship was found to be significant for participants assigned to the aerobic exercise group (P = 0.016) and the combined aerobic and resistance exercise group (P = 0.041), but not for the resistance exercise group (P = 0.233).
The HbA1c from baseline to 6 months for participants in the aerobic exercise group who completed less than 70% of their three prescribed sessions per week did not change significantly (0.13% [1 mmol·mol−1]; 95% CI, −0.48 to 0.83; Table 2). The participants in the aerobic exercise group who completed more than 90% of prescribed exercise sessions had a mean reduction in HbA1c of 0.69% (8 mmol·mol−1) (95% CI, −1.11% to −0.26%). The same trend was observed with the combined aerobic and resistance exercise group with the change in HbA1c ranging from a mean change of −0.52% (6 mmol·mol−1) (95% CI, −1.15% to 0.12%) in those who completed less than 70% of prescribed exercise compared with −1.09% (12 mmol·mol−1; 95% CI, −1.57 to −0.60) in those who completed more than 90% of prescribed sessions.
Dose–response relationship stratified by age, sex, and baseline HbA1c
Dose–response relationships stratified by age, sex, and baseline HbA1c are presented in Table 3. The DARE trial was not originally designed to evaluate these relationships, and, therefore, these analyses are limited by sample size. When including participants from all training groups, there was a statistically significant dose–response relationship identified for participants younger than 55 yr (P = 0.005), male participants (P = 0.010), and participants with baseline HbA1c ≥7.5% (58 mmol·mol−1) (P = 0.011). For participants with a baseline HbA1c of 7.5% (58 mmol·mol−1) or greater, an absolute increase by 20% in adherence to three prescribed exercise sessions per week (e.g., an additional two exercise sessions per month) was associated with a decrease in HbA1c of 0.26% (3 mmol·mol−1) (95%CI, 0.06%–0.46%). Within the aerobic exercise training group, dose–response relationships were also found to be statistically significant for younger than 55 yr (P = 0.012), males (P = 0.032), and baseline HbA1c ≥7.5% (58 mmol·mol−1) (P = 0.006). A significant dose–response relationship was found for those assigned to the combined aerobic and resistance training modality for participants younger than 55 yr (P = 0.006) but not the resistance training group (P = 0.466). When the resistance and combined aerobic and resistance exercise groups were stratified by sex and baseline HbA1c, there was no significant dose–response relationship for males or those with higher baseline HbA1c (Table 3).
This study aimed to explore whether there is a dose–response relationship between exercise adherence over a 6-month exercise intervention and glycemic control in patients with type 2 diabetes. The dose–response relationship in the overall study population was found to be statistically significant with a higher exercise dose being associated with a greater HbA1c reduction. Specifically, we found that for every 20% absolute increase in adherence to exercise training prescribed thrice weekly (e.g., an additional two exercise sessions per month), there was a 0.15% (2 mmol·mol−1) reduction in HbA1c over 6 months. When analyzed by exercise modality, we found that there was a significant dose–response relationship within the aerobic and combined aerobic and resistance exercise training groups, but not in the resistance training group. Additionally, significant dose–response relationships were identified for individuals younger than 55 yr, men and those with a baseline HbA1c of 7.5% (58 mmol·mol−1) or greater.
We are unaware of previous studies exploring the relationship between adherence to prescribed exercise and change in glycemic control in patients with type 2 diabetes. Our findings support the Physical Activity/Exercise guidelines put forward by the American Diabetes Association (8). Specifically, the guidelines recommend a combination of aerobic and resistance exercise training, which we found results in the greatest reduction in HbA1c regardless of degree of adherence to prescribed exercise. For aerobic exercise training, our results support recommendations of three or more exercise sessions per week as more aerobic exercise sessions were associated with a greater reduction in HbA1c. This is consistent with previous studies that have shown a positive association between aerobic exercise training volume and improvements in glycemic control (12). Conversely, two sessions per week (corresponding to 67% adherence in the DARE trial) may be adequate for resistance exercise training as no further benefit in terms of HbA1c reduction was identified with further sessions per week. This finding is consistent with previous studies exploring the effect of various prescribed resistance exercise training regimens in patients with type 2 diabetes (12). Yang et al. (13) found no difference in glycemic control with variation in prescribed resistance exercise training volume and intensity for patients with type 2 diabetes completing combined aerobic and resistance exercise. Resistance exercise training has been shown to improve insulin sensitivity in individuals with type 2 diabetes (14,15). The difference we observed between aerobic and resistance exercise training could be explained by the variable effect of these exercise modalities on insulin sensitivity.
The results of our stratified analyses suggest that the dose–response relationship between exercise adherence and glycemic control may exist for some subgroups of patients with type 2 diabetes but not others. Specifically, we identified three subgroups where greater adherence to prescribed exercise was associated with further reductions in HbA1c: 1) participants with HbA1c values greater than 7.5% (58 mmol·mol−1) before starting the exercise intervention, 2) men and 3) individuals younger than 55 yr.
Baseline HbA1c appears to affect the relationship between exercise adherence and glycemic control, with an association being identified only for those with HbA1c levels greater than 7.5% (58 mmol·mol−1) in our study. This association has been shown by previous studies (6,12) and is biologically plausible, as those with HbA1c levels less than 7.5% (58 mmol·mol−1) at baseline already have fairly good glycemic control and, therefore, have less room to improve when initiating an exercise intervention. Of note, although the dose–response relationship was not found to be significant in this stratified analysis, exercise is beneficial in this population. This was shown in the DARE trial where those randomized to combined aerobic and resistance exercise training had greater HbA1c reductions compared with control and with aerobic or resistance training alone, even when they started with an HbA1c less than 7.5% (58 mmol·mol−1).
Our findings suggest that there may be a different dose–response relationship based on biologic sex with men experiencing a greater HbA1c reduction with an increase in number of exercise sessions completed, whereas women do not. Biologic sex differences in physiologic response to exercise have previously been demonstrated in patients with type 2 diabetes in terms of cardiovascular fitness, tissue blood flow, and endothelial function (16). We speculate that higher testosterone levels in men may facilitate changes in muscle that increase glucose disposal (17). Studies are needed to further delineate the effect of exercise adherence on HbA1c reduction in this population, as biologic sex specific exercise prescriptions may be warranted.
We also found a dose–response relationship between exercise adherence and HbA1c reduction for patients younger than 55 yr regardless of exercise type. This is consistent with previous studies which have examined the effect of aerobic exercise training on HbA1c. Mourier et al. (18) (1997) found aerobic exercise training had a significant impact on HbA1c in participants with type 2 diabetes who had a mean age of 45 yr, compared with Tessier et al. (2000) who found no impact of aerobic training on HbA1c in older patients (19). Further research is needed to explore the effect of age on this dose–response relationship, as this may affect exercise prescription guidelines if the glycemic control response is different to prescribed doses based on a patient’s age. It should be noted however that, in studies that included only older participants, Dunstan et al. (20) and Castaneda et al. (21), large reductions were achieved in HbA1c through resistance training.
This study’s notable strength is that it provides novel findings regarding the effect of adherence to prescribed exercise on glycemic control. However, there are several limitations that should be acknowledged. First, this is a post hoc analysis that explored associations for which the trial was not originally designed to assess. The statistical power to examine the associations identified may be inadequate, especially in smaller subgroups. Second, exercise programs completed in this trial included supervised sessions and these associations may not translate to unsupervised exercise programs. Third, although measures were taken to standardize diet and minimize changes in medications over the 6-month behavioral intervention period, it is not possible to determine if the glycemic changes observed in this study are results of exercise training alone. Fourth, there were relatively fewer female participants in this study than male participants, which limited our ability to analyze the effect of biologic sex on this dose–response relationship.
In conclusion, we found a dose–response relationship between exercise adherence and change in HbA1c for aerobic and combined exercise but not for resistance exercise training. These results suggest that an increased volume of aerobic or combined aerobic and resistance exercise is associated with greater improvement in glycemic control. Our findings support clinical practice guideline recommendations for aerobic and combined exercise prescriptions.
The authors thank Dr. Gordon H. Fick and Devon Boyne for useful discussions. The authors are indebted to the DARE study coordinators, research, supporters and participants and all those involved in the DARE study.
J. L. B., J. E. B., M. J. D., S. D., N. G. B., G. P. K., and R. J. S., conceived and designed the study. J. L. B., J. E. B., M. J. D., S. D., and R. J. S. performed data analysis. J. L. B. wrote the initial article draft. J. L. B., J. E. B., M. J. D., S. D., N. G. B., G. P. K., and R. J. S. all participated in the critical revision of the article and approved the final version. J. L. B. and R. J. S. are the guarantors of the work and take responsibility for the integrity of the data.
The DARE trial was supported by grants from Canadian Institutes of Health Research (MCT-44155) and the Canadian Diabetes Association (the Lillian Hollefriend Grant). Dr. Sigal was supported by salary awards from the Canadian Institutes of Health Research the Ottawa Hospital Research Institute, and Alberta Innovates-Health Solutions; Dr. Kenny was supported by a Career Scientist Award from the Ontario Ministry of Health and Long Term Care, and a University of Ottawa Research Chair.
The authors have no conflicts of interest to declare. The results of the present study do not constitute endorsement by ACSM. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.
1. Balducci S, Zanuso S, Massarini M, et al. The Italian diabetes and exercise study (IDES): Design and methods for a prospective Italian multicentre trial of intensive lifestyle intervention in people with type 2 diabetes and the metabolic syndrome. Nutr Metab Cardiovasc Dis
2. Sigal RJ, Kenny GP, Boulé NG, et al. Effects of aerobic training, resistance training, or both on glycemic control in type 2 diabetes: a randomized trial. Ann Intern Med
3. Church TS, Blair SN, Cocreham S, et al. Effects of aerobic and resistance training on hemoglobin A1c levels in patients with type 2 diabetes : a randomized controlled trial
4. Umpierre D, Ribeiro PAB, Kramer CK, et al. Physical activity advice only or structured exercise training and association with HbA1c levels in type 2 diabetes: a systematic review and meta-analysis. JAMA
5. Chen L, Pei J-H, Kuang J, et al. Effect of lifestyle intervention in patients with type 2 diabetes: a meta-analysis. Metabolism
6. Snowling NJ, Hopkins WG. Effects of different modes of exercise training on glucose control and risk factors for complications in type 2 diabetic patients: a meta-analysis. Diabetes Care
7. Myers VH, McVay MA, Brashear MM, et al. Exercise training and quality of life in individuals with type 2 diabetes: a randomized controlled trial
. Diabetes Care
8. Colberg SR, Sigal RJ, Yardley JE, et al. Physical activity/exercise and diabetes: a position statement of the American Diabetes Association. Diabetes Care
9. Coyle ME, Francis K, Chapman Y. Self-management activities in diabetes care: a systematic review. Aust Health Rev
10. Zhao G, Ford ES, Li C, Balluz LS. Physical activity in U.S. older adults with diabetes mellitus: prevalence and correlates of meeting physical activity recommendations. J Am Geriatr Soc
11. Mu L, Cohen AJ, Mukamal KJ. Resistance and aerobic exercise among adults with diabetes in the U.S. Diabetes Care
12. Umpierre D, Ribeiro PAB, Schaan BD, Ribeiro JP. Volume of supervised exercise training impacts glycaemic control in patients with type 2 diabetes: a systematic review with meta-regression analysis. Diabetologia
13. Yang P, Swardfager W, Fernandes D, et al. Finding the optimal volume and intensity of resistance training exercise for type 2 diabetes: the FORTE study, a randomized trial. Diabetes Res Clin Pract
14. Ibañez J, Izquierdo M, Argüelles I, et al. Twice-weekly progressive resistance training decreases abdominal fat and improves insulin sensitivity in older men with type 2 diabetes. Diabetes Care
15. Brooks N, Layne JE, Gordon PL, Roubenoff R, Nelson ME, Castaneda-Sceppa C. Strength training improves muscle quality and insulin sensitivity in Hispanic older adults with type 2 diabetes. Int J Med Sci
16. Regensteiner JG, Bauer TA, Huebschmann AG, et al. Sex differences in the effects of type 2 diabetes on exercise performance. Med Sci Sports Exerc
17. Pitteloud N, Mootha VK, Dwyer AA, et al. Relationship between testosterone levels, insulin sensitivity, and mitochondrial function in men. Diabetes Care
18. Mourier A, Gautier J-F, Kerviler ED, et al. Mobilization of visceral adipose tissue related to the improvement in insulin sensitivity in response to physical training in NIDDM: effects of branched-chain amino acid supplements. Diabetes Care
19. Tessier D, Ménard J, Fülöp T, et al. Effects of aerobic physical exercise in the elderly with type 2 diabetes mellitus. Arch Gerontol Geriatr
20. Dunstan DW, Daly RM, Owen N, et al. High-intensity resistance training improves glycemic control in older patients with type 2 diabetes. Diabetes Care
21. Castaneda C, Layne JE, Munoz-Orians L, et al. A randomized controlled trial
of resistance exercise training to improve glycemic control in older adults with type 2 diabetes. Diabetes Care