To determine the relationship between lipid profiles and the type of weight training and to assess the effects of anabolic-androgenic steroids on lipids, bodybuilders and powerlifters of similar age, body fat, and testosterone levels were studied before and after androgen use. Before androgen administration powerlifters had lower levels of plasma high-density-lipoprotein cholesterol (HDL-C) and HDL2-C (38 +/- 2; 6 +/- 1 mg/dL; means +/- SE, n = 8) than bodybuilders (55 +/- 2; 12 +/- 1 mg/dL; n = 8) and runners of comparable age and body fat (47 +/- 2; 14 +/- 2 mg/dL; n = 8), while levels of low-density-lipoprotein cholesterol (LDL-C) were higher in powerlifters (138 +/- 10 mg/dL) than in bodybuilders (104 +/- 7 mg/dL) and runners (110 +/- 6 mg/dL). Therefore, powerlifters had higher LDL-C/HDL-C ratios (3.7 +/- 0.3) than bodybuilders (2.0 +/- 0.2) and runners (2.4 +/- 0.2). Androgen use by eight bodybuilders and four powerlifters lowered values of both HDL-C and HDL2-C by 55% and raised values of LDL-C (61% +/- 10%) and LDL-C/HDL-C ratios (280% +/- 40%). Therefore, the training regimen of bodybuilders is associated with a more favorable lipid profile than the training used by powerlifters. Androgen use by strength-trained athletes may increase their risk for coronary heart disease.