1. We have assessed insulin secretion, insulin sensitivity, exercise capacity and muscle fibre composition in healthy men with a low insulin response to glucose (low insulin responders and endurance-trained subjects) and in healthy men with a normally high insulin response. Low insulin responders have been considered as prediabetic subjects. 2. During glucose infusion, low insulin responders and endurance-trained subjects had acute-phase (0-10 min) insulin responses that were 26% (P<0.001) and 31% (P<0.001), and C-peptide responses that were 35% (P<0.001) and 42% (P<0.01), respectively of the responses in high insulin responders. Also, the late-phase (10-60 min) insulin and C-peptide responses were lower in low insulin responders and endurance-trained subjects, Endurance-trained subjects, compared with high and Low insulin responders, had higher insulin sensitivity (blood glucose concentration during a somatostatin-insulin-glucose infusion test 3.6 +/- 0.4 versus 6.3 +/- 0.6 mmol/l in high insulin responders, P<0.01, and 5.5 +/- 0.4 mmol/l in low insulin responders, P<0.01), exercise capacity (4.2 +/- 0.2 versus 3.0 +/- 0.2 W/kg body weight, P<0.001, and 2.8 +/- 0.1 W/kg body weight, P<0.001) and a higher percentage of slow-twitch fibres (58.2 +/- 5.2 versus 38.5 +/- 3.5%, P<0.01 and 40.9 +/- 3.0%, P<0.001). 3. In conclusion, we demonstrate that low insulin responders exhibited decreased insulin and C-peptide responses to glucose, despite normal insulin sensitivity and exercise capacity. Thus, the low insulin response in low insulin responders reflects a true impairment of insulin release and not a compensatory decrease due to increased insulin sensitivity. The insulin response to glucose in endurance-trained subjects was also decreased, probably due to increased insulin sensitivity.