1. Unaccustomed eccentric exercise decreases whole-body insulin action in humans. To study the effects of one-legged eccentric exercise on insulin action in muscle and systemically, the euglycaemic clamp technique combined with arterial and bilateral femoral venous catheterization was used. Seven subjects participated in two euglycaemic clamps, performed in random order. One clamp was preceded 2 days earlier by one-legged eccentric exercise (post-eccentric exercise clamp (PEC)) and one was without the prior exercise (control clamp (CC)). 2. During PEC the maximal insulin-stimulated glucose uptake over the eccentric thigh was marginally lower when compared with the control thigh, (11.9%, 64.6 +/- 10.3 vs. 73.3 +/- 10.2 mumol kg-1 min-1, P = 0.08), whereas no inter-thigh difference was observed at a submaximal insulin concentration. The glycogen concentration was lower in the eccentric thigh for all three clamp steps used (P < 0.05). The glucose transporter GLUT4 protein content was on average 39% lower (P < 0.05) in the eccentric thigh in the basal state, whereas the maximal activity of glycogen synthase was identical in the two thighs for all clamp steps. 3. The glucose infusion rate (GIR) necessary to maintain euglycaemia during maximal insulin stimulation was lower during PEC compared with CC (15.7%, 81.3 +/- 3.2 vs. 96.4 +/- 8.8 mumol kg-1 min-1, P < 0.05). 4. Our data show that 2 days after unaccustomed eccentric exercise, muscle and whole-body insulin action is impaired at maximal but not submaximal concentrations. The local effect cannot account for the whole-body effect, suggesting the release of a factor which decreases insulin responsiveness systemically.