Current issues of ACP Journal Club are published in Annals of Internal Medicine


Regular exercise was associated with a reduced incidence of diabetes mellitus among U.S. male physicians

ACP J Club. 1993 Jan-Feb;118:26. doi:10.7326/ACPJC-1993-118-1-026

Source Citation

Manson JE, Nathan DM, Krolewski AS, et al. A prospective study of exercise and incidence of diabetes among U.S. male physicians. JAMA. 1992 Jul 1;268:63-7.



To determine the association between regular exercise and the development of non-insulin-dependent diabetes mellitus in U.S. male physicians.


5-year cohort study of participants in the Physicians' Health Study.


United States.


21 271 male physicians aged 40 to 84 years were enrolled in a trial of prevention of cancer and cardiovascular disease. Exclusion criteria were a history of myocardial infarction, stroke, transient cerebral ischemia, and, for this analysis, diabetes mellitus. Mean follow-up was 60 months (range, 46 to 77 mo), and 99.7% were followed.

Assessment of risk factors

Participants filled out questionnaires at baseline that collected past medical history, physical characteristics, frequency of vigorous exercise (enough to work up a sweat), cigarette smoking, and alcohol consumption. A family history of diabetes was not collected. Questionnaires completed at 6 months, and then yearly, collected data on new medical diagnoses including diabetes. New cases of diabetes were assumed to be non-insulin-dependent.

Main outcome measure

Incidence of non-insulin-dependent diabetes mellitus.

Main results

285 new cases of diabetes mellitus were reported. Adjusted for age and body mass index (BMI), the relative risk (RR) for developing diabetes among men who exercised at least once per week compared with sedentary men was 0.71 (95% CI 0 56 to 0.91, P = 0.006). The results were similar in a multivariate analysis adjusted for age, BMI, smoking, alcohol consumption, blood pressure, history of hypertension and high serum cholesterol levels, and parental history of myocardial infarction (RR 0.71, CI 0.54 to 0.94, P = 0.02). The incidence of diabetes decreased with increasing exercise. The RR for diabetes was 0.77 for those who exercised once per week, 0.62 for 2 to 4 times per week of exercise, and 0.58 for ≥ 5 periods of exercise per week. Obesity was associated with an increased RR for diabetes. For a BMI > 26.4 kg/m2, the RR for diabetes was 3.09 (CI 2.02 to 4.72, P < 0.001). A history of hypertension also increased the risk for diabetes (RR 2.03; CI 1.47 to 2.80, P < 0.001).


Regular exercise was associated with a decreased incidence of non-insulin-dependent diabetes mellitus in U.S. male physicians. The association remained after adjusting for body mass index, smoking, alcohol consumption, blood pressure, history of hypertension, high serum cholesterol levels, and parental history of myocardial infarction.

Source of funding: National Institutes of Health.

For article reprint: Dr. J.E. Manson, 900 Commonwealth Avenue East, Boston, MA 02215-1204, USA. FAX 617-734-1437.


The cause and prevention of non-insulin-dependent diabetes mellitus are poorly understood. The observation by Manson and colleagues that an inverse correlation exists between how physicians characterize their level of physical activity and their subsequent development of type II diabetes provides an important clinical clue to our thinking about the pathophysiology of the disease.

The 5-year follow-up of this study and its large number of physician-participants underscore the credibility of the study's conclusion. The study is methodologically sound but limited in the information it provides. We do not know whether family history of non-insulin-dependent diabetes was equally distributed among the exercise cohorts. From an etiologic standpoint, however, little reason exists to think that a family history of diabetes would influence exercise patterns. We are also left to speculate whether exercise patterns were maintained during the study and what influence changes in exercise may have had on the development of glucose intolerance. In addition, it would be useful to know dietary patterns and how they correlate with exercise levels and incidence of diabetes so we can learn more about the effect of diet in the development of carbohydrate intolerance.

From a practical standpoint, we know from this and other studies that a strong correlation exists between obesity and the development of non-insulin-dependent diabetes mellitus. Manson and colleagues have shown an inverse correlation between individual patterns of exercise and development of non-insulin-dependent diabetes mellitus. Moreover, a recent randomized trial in China showed that exercise prevents diabetes in people with impaired glucose tolerance (1). Taken together, these studies underscore the important role of balancing the intake and expenditure of caloric nutrients in preventing carbohydrate intolerance and the expression of the diabetic state.

Eugene C. Corbett, Jr., MD
University of VirginiaCharlottesville, Virginia, USA


1. Pan XR, Li GW, Hu YH, et al. Effects of diet and excerise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care. 1997;20:537-44.