We studied separately the influence of two methods for losing fat weight on the levels of plasma lipids and lipoproteins in overweight sedentary men--decreasing energy intake without increasing exercise (diet), and increasing energy expenditure without altering energy intake (exercise, primarily running)--in a one-year randomized controlled trial. As compared with controls (n = 42), dieters (n = 42) had significant loss of total body weight (-7.8 +/- 0.9 kg [mean +/- SE]), fat weight (-5.6 +/- 0.8 kg), and lean (non-fat) weight (-2.1 +/- 0.5 kg) (P less than 0.001 for each variable), and exercisers (n = 47) had significant loss of total body weight (-4.6 +/- 0.8 kg) and fat weight (-3.8 +/- 0.7 kg) (P less than 0.001 for both variables) but not lean weight (-0.7 +/- 0.4 kg). Fat-weight loss did not differ significantly between dieters and exercisers. All subjects were discouraged from altering their diet composition; however, dieters and exercisers had slight reductions in the percentage of kilojoules derived from fat. As compared with the control group, both weight-loss groups had significant increases (P less than 0.01) in plasma concentrations of high-density lipoprotein (HDL) cholesterol (diet vs. exercise, 0.13 +/- 0.03 vs. 0.12 +/- 0.03 mmol per liter), HDL2 cholesterol (0.07 +/- 0.02 vs. 0.07 +/- 0.02 mmol per liter), and HDL3 cholesterol (0.07 +/- 0.02 vs. 0.06 +/- 0.02 mmol per liter) and significant decreases (P less than 0.05) in triglyceride levels (diet vs. exercise, -0.35 +/- 0.14 vs. -0.24 +/- 0.12 mmol per liter). Levels of total and low-density lipoprotein cholesterol were not significantly changed, relative to values in controls. None of these changes were significantly different between dieters and exercisers. Thus, we conclude that fat loss through dieting or exercising produces comparable and favorable changes in plasma lipoprotein concentrations.