Individuals with impaired glucose tolerance or type 2 diabetes mellitus underwent 12 to 16 weeks of moderate-intensity exercise training, 5 days per week for 60 minutes per day. Individual subject data points are plotted on both panels; the x-axis represents the pretraining variable, and the y-axis indicates the exercise-induced change, such that the data points above the axis indicate an exercise-induced increase and vice versa. Open circles represent impaired glucose–tolerant subjects, and open triangles represent subjects with type 2 diabetes mellitus. The solid line represents the regression curve, and the dotted line represents the 95% confidence interval. A, There was a nonlinear quadratic relationship between pretraining 2-hour oral glucose tolerance test (OGTT) level and the training-induced change in 2-hour OGTT level (y = 0.06x2 − 1.5x + 7.6 [r2 = 0.26; P = .06] [N = 105]). For every 1-mmol/L increase in pretraining 2-hour glucose level above 13.1 mmol/L (the inflection point of the curve), there was a 0.2-mmol/L loss of improvement in 2-hour glucose level following exercise (to convert glucose to milligrams per deciliter, divide by 0.0555). B, There was also a nonlinear quadratic relationship between pretraining hemoglobin A1c (HbA1c) level and the training-induced change in HbA1c level (y = 0.31x2 − 3.8x + 11.7 [r2 = 0.33; P = .02] [n = 52]). For every 1–percentage-point increase in pretraining HbA1c level above 6.2% (the inflection point of the curve), there was a 0.2–percentage-point loss of improvement in HbA1c level following exercise. C, An inverse linear relationship between pretraining HbA1c level and the training-induced change in aerobic fitness was found (y = −0.11x + 0.91 [r = −0.38; P = .006] [n = 52]). For every 1–percentage-point increase in pretraining HbA1c level, there was 0.11-L/min loss of improvement in maximal oxygen uptake (V̇o2max) following exercise training.
eTable. Subject Characteristics
Solomon TPJ, Malin SK, Karstoft K, Haus JM, Kirwan JP. The Influence of Hyperglycemia on the Therapeutic Effect of Exercise on Glycemic Control in Patients With Type 2 Diabetes Mellitus. JAMA Intern Med. 2013;173(19):1834-1836. doi:10.1001/jamainternmed.2013.7783
Randomized clinical trials show that aerobic exercise training improves glycemic control in patients with type 2 diabetes mellitus (T2DM).1 However, interindividual variability is large.2 This may be explained by genetic variability,3 but ambient hyperglycemia4 and pancreatic β-cell function5 may also contribute. We examined whether changes in glycemic control following a 12- to 16-week aerobic exercise training intervention were influenced by the pretraining glycemic state in 105 individuals with impaired glucose tolerance or T2DM.
This study was approved by our institutional review board, and subjects provided informed consent. Before and following a 12- to 16-week period of aerobic exercise training, body composition, aerobic fitness (maximal oxygen uptake [V̇o2max]), and glycemic control (hemoglobin A1c [HbA1c], fasting glucose, and oral glucose tolerance test [OGTT] levels) were determined in a total of 105 older (mean [SEM] age, 61  years), overweight or obese (mean [SEM] body mass index, 33  [calculated as weight in kilograms divided by height in meters squared]) individuals with impaired glucose tolerance (n = 56) or T2DM (n = 49; diagnosed a mean [SEM] 4.8 [0.9] years prior and not insulin treated). Relationships between preintervention variables and intervention-induced changes in variables were assessed by linear and nonlinear regression. See eMethods in the Supplement for full details of the study design.
Mean (SEM) change in body weight (−4.6 [0.5] kg), whole-body adiposity (−1.9% [0.3%]), V̇o2max (+0.23 [0.03] L/min), fasting plasma glucose (−0.35 [0.08] mmol/L), and 2-hour OGTT (−0.8 [0.2] mmol/L) (to convert glucose to milligrams per deciliter, divide by 0.0555) were significantly improved following exercise training (full data are given in the eTable in the Supplement). Pretraining fasting plasma glucose level did not influence exercise-induced changes in glycemic control. However, there was a nonlinear quadratic relationship between pretraining 2-hour OGTT and exercise-induced changes in the 2-hour glucose response (r2 = 0.26; P = .06) (Figure, A). Subjects with a pretraining 2-hour OGTT level of less than 13.1 mmol/L showed greater exercise-induced decreases in 2-hour glucose level (r = −0.44; P < .001), whereas subjects with a pretraining 2-hour OGTT level of greater than 13.1 mmol/L had smaller improvements in 2-hour glucose level (r = 0.29; P = .07). The same nonlinear quadratic relationship existed between pretraining HbA1c level and exercise-induced changes in HbA1c level (r2 = 0.33; P = .02) (Figure, B), where subjects with a pretraining HbA1c level less than 6.2% had an exercise-induced decrease in HbA1c level (r = −0.55; P = .005), whereas subjects with a pretraining HbA1c level greater than 6.2% had smaller exercise-induced improvements in HbA1c level (r = 0.38; P = .04). Furthermore, pretraining HbA1c level was linearly and inversely related to the exercise-induced change in V̇o2max, such that high pretraining HbA1c level predicted smaller exercise-induced increases in V̇o2max (r = −0.38; P = .006) (Figure, C).
These findings emphasize that exercise-induced improvements in glycemic control are dependent on the pretraining glycemic level. We demonstrate that although moderate-intensity aerobic exercise can improve glycemic control, individuals with ambient hyperglycemia are the most likely to be nonresponders. Our key observation is that pretraining hyperglycemia predicts exercise-induced improvements in glycemic control: for every 1-mmol/L rise in pretraining 2-hour OGTT glucose level above 13.1-mmol/L (the curve inflection point in Figure, A) we predict a 0.2-mmol/L loss of improvement in 2-hour OGTT glucose following exercise. Accordingly, for every 1–percentage-point increase in pretraining HbA1c level above 6.2% (the curve inflection point in Figure, B), we predict a 0.2% point loss of improvement in HbA1c level following exercise. Pretraining hyperglycemia also predicted the exercise-induced increment in aerobic fitness: for a 1–percentage-point increase in pretraining HbA1c level, we predict a 0.11 L/min loss of improvement in V̇o2max following exercise.
Prior work shows that diabetes remission following exercise and diet intervention is more likely in individuals with a shorter disease history and lower HbA1c level.6 We show that aerobic exercise-induced improvements in glycemic control are blunted by ambient hyperglycemia, particularly in subjects with T2DM. Mechanistic studies are required to help us understand this phenomenon, but underlying impairments in β-cell function are likely to be very important.5 That hyperglycemia blunted the cardiovascular adaptations to exercise (V̇o2max) is in agreement with some prior reports7 and may be explained by the causal association between chronic hyperglycemia and microvascular and macrovascular dysfunction.8
The clinical relevance of these new findings is paramount and highlights the need to understand the metabolic “nonresponder.” Because chronic hyperglycemia (>6.2% HbA1c level; >13.1-mmol/L glucose level) potentially predicts a poor therapeutic effect of aerobic exercise on glycemic control and fitness, using exercise to treat patients with poorly controlled T2DM may have limited chances of a successful outcome.
Corresponding Author: Thomas P. J. Solomon, PhD, Centre of Inflammation and Metabolism, Rigshospitalet M7641, Blegdamsvej 9, Copenhagen 2100, Denmark (firstname.lastname@example.org).
Published Online: July 1, 2013. doi:10.1001/jamainternmed.2013.7783.
Author Contributions: Dr Solomon had access to the data and takes responsibility for the integrity of the data and the accuracy of the analysis.
Study concept and design: Solomon, Kirwan.
Acquisition of data: Solomon, Karstoft, Haus, Kirwan.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Solomon, Malin, Kirwan.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Solomon, Malin.
Obtained funding: Solomon, Malin, Kirwan.
Administrative, technical, and material support: All authors.
Study supervision: Solomon, Kirwan.
Conflict of Interest Disclosures: None reported.
Funding/Support: The study was funded by a Paul Langerhans program grant from the European Foundation for the Study of Diabetes (Dr Solomon) and was supported by grant RO1 AG12834 from the National Institute of Health (Dr Kirwan) and Clinical and Translational Science Award UL1 RR024989. Dr Malin was supported by NIH grant T32 DK007319.
Role of the Sponsors: The funding sources played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.
Trial Registration: clinicaltrials.gov Identifier: NCT01234155
Additional Contributions: Lisbeth Andreasen, MSc (Rigshospitalet, Denmark), assisted with biochemical analyses and Julianne Filion, RN (Cleveland Clinic, Cleveland, Ohio), assisted with subject recruitment and patient screening. Marc Cook, MSc (Cleveland Clinic), Thomas Grøndahl, MSc (Rigshospitalet), and Kamilla Winding, MSc (Rigshospitalet), helped with exercise training.