The aims of the present study were to determine whether available "fasting" and oral glucose tolerance test-derived insulin sensitivity indices could effectively discriminate between individuals with higher than normal insulin sensitivity, and whether they would all provide similar information in clinical practice. Sprint runners (n = 8), endurance runners (n = 8) and sedentary controls (n = 7) received a 75-g oral glucose tolerance test. All participants were healthy lean males, aged 21-29 years. Besides glucose and insulin responses, a total of nine such indices were computed. Fasting as well as post-load glucose concentrations were similar in the three groups, while basal plasma insulin and the insulinaemic response to glucose were both higher in untrained individuals (at P < 0.05 and P < 0.02, respectively). There were no differences between endurance and sprint runners. The results for insulin sensitivity, however, were quite variable: three indices showed that both groups of athletes were more insulin-sensitive than controls; three indicated that this was the case for endurance runners only; one indicated that this was the case for sprint runners only; and two showed that sprint runners were more insulin-sensitive than either sedentary individuals or endurance runners (all differences were significant at P < 0.05). Controlling for total body weight or lean mass did not effectively resolve this disagreement. Apparently, the various insulin sensitivity indices examined provided different quantitative and qualitative information, despite insulin action being greater in both groups of athletes relative to controls, as reflected by their similar glucose tolerance with lower insulin concentrations. We suggest, therefore, that the use and interpretation of such indices among physically active individuals be made with caution.