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


Resistance strength training decreased total cholesterol, HbA1c and glucose levels, while increasing strength and endurance in men with IDDM

ACP J Club. 1991 Jan-Feb;114:19. doi:10.7326/ACPJC-1991-114-1-019

Source Citation

Durak EP, Jovanovic-Peterson L, Peterson CM. Randomized crossover study of effect of resistance training on glycemic control, muscular strength, and cholesterol in type I diabetic men. Diabetes Care. 1990 Oct;13:1039-43.



To evaluate the effect of a program of resistance weight training on lipid, blood glucose, and HbA1c levels, and overall strength in insulin-dependent diabetes mellitus (IDDM).


Randomized, unblinded, crossover trial.


Training sessions took place in a physical therapy clinic in the United States.


8 men with IDDM (mean age 31 y, mean duration of diabetes 12 y) were recruited (the recruitment strategy was not reported). Patients had previously been successfully treated for diabetic retinopathy; otherwise participants had no evidence of microvascular or macrovascular complications or hypertension. Incoming fitness or activity levels were not reported.


Patients received resistance strength training of arms and legs. 1-hour sessions of 40 to 50 sets were held 3 d/wk for 10 weeks. Group A (n = 4) finished 10 weeks of training, then rested for 6 weeks. Group B (n = 4) rested for 6 weeks before 10 weeks of training. Blood samples were drawn 3 times: at baseline, after 10 weeks of training, and after 6 weeks of rest.

Main outcome measures

Lipid profile; plasma glucose and glycosylated hemoglobin; and strength.

Main results

Significant increases in strength (94% in squat, P < 0.001, and 58% in upper-body strength, P =0.002) were measured at the end of the training program. The bench press was the only endurance test to show a statistically significant increase from baseline (P = 0.02). All cholesterol levels decreased (except high-density lipoprotein cholesterol levels) during training but only the decrease in total cholesterol levels reached statistical significance (5.04 mM ± standard deviation [SD] 1.06 mM vs 4.63 mM ± SD 0.081 mM, P = 0.015). Glycosylated hemoglobin levels dropped a mean of 1.1% during the exercise program (P = 0.05). There were 10 occurrences (out of a total of 152 training sessions) of blood glucose levels of < 3.64 mM after exercise, none with reported clinical consequences. Only cholesterol, triglyceride, and insulin levels were reported after the rest period, and these values all increased towards pre-exercise values. No morbidity other than minor soreness occurred.


A supervised, heavy-resistance, strength-training program for men with insulin-dependent diabetes mellitus increased strength and endurance and decreased total cholesterol levels, glycosylated hemoglobin, and self-monitored glucose levels.

Source of funding: Valeo, Inc. and Vallhalla Scientific.

Address for article reprint: E. P. Durak, Sansum Medical Research Foundation, 221 Bath Street, Santa Barbara, CA 93105, USA.


Previous studies, cited by Durak and colleagues, have shown lower insulin levels and lower glucose responses to oral glucose tolerance tests and better lipid levels in nondiabetic persons participating in weight-training exercise. However, in the study by Durak and colleagues, analyses comparing baseline, post-training, and post-resting values for each lipid fraction show no significant lipid changes with training. The significant but small decrease in total cholesterol levels—comparing only the 2 points in time, baseline and after training—occurred to the greatest extent in those with the highest initial total cholesterol levels and thus may represent a regression to the mean rather than a true training effect.

Although HbA1c levels in these patients decreased by 1.1 %, an appropriate "repeated measures" analysis of the 3 periods was not done, weakening the crossover design and the strength of this finding. Although insulin levels and total carbohydrate intake remained the same, lowering of glucose levels in these fairly well-controlled patients could have been caused by the changes within the carbohydrate component of the diet rather than the weight-training effect.

Because the study showed no adverse effects in these patients with well-controlled diabetes and no complications, weight training should not be precluded in these patients. However, one should not advise this training in patients with diabetes who have proliferative retinopathy or uncontrolled hypertension where severe and sudden increases in blood pressure with resistive exercises may cause further damage. Also, patients with glucose values ≥ 13.9 mmol/L (250 mg/L) should not do such exercise until their glucose levels are lower, because at this elevated level of glucose, exercise can increase hyperglycemia.

Donald A. Smith, MD, MPH
Mount Sinai Medical CenterNew York, New York, USA