We assessed the effects of twice weekly strength training on several proposed risk factors for breast and colon cancer: body fat, waist circumference, fasting insulin, fasting glucose, insulin-like growth factor I (IGF-I), and several IGF-binding proteins. Fifty-four healthy women, 30-50 years old, were randomized to no-contact control or treatment: 15 weeks of supervised strength training followed by 6 months of unsupervised training. Fifteen-week changes included reductions in percentage of body fat, fasting insulin, fasting glucose, and IGF-I that were larger in the treatment than control participants (treatment versus control mean +/- SE: % body fat -1.97 +/- 0.42 versus -0.43 +/- 0.40, P = 0.01; insulin (uU/ml) -0.29 +/- 0.35 versus 0.81 +/- 0.38, P = 0.055; glucose (mg/dl) -1.92 +/- 1.27 versus 1.21 +/- 1.36, P = 0.13; and IGF-I (ng/ml) -30.47 +/- 9.75 versus 5.86 +/- 10.44, P = 0.02). There was no treatment effect on IGF-binding proteins 1 and 3 or either of two surrogate measures of free IGF-I. By 39 weeks changes in percentages of body fat were largely maintained; IGF-I returned to baseline levels in the treatment group but remained 15% lower in treatment compared with control participants. Strength training produced favorable changes in several proposed cancer risk factors. The importance of these changes to long-term cancer prognosis, diagnosis, and/or recurrence remains to be determined.