Data on interleukin-6 (IL-6) and tumour necrosis factor-a (TNF-a) release during acute exercise are not conclusive, and information is lacking about the impact of physical inactivity. Some studies have shown an increase, but others report no changes in IL-6 and TNF-a release during exercise. We have now studied the temporal relationship of leg IL-6 and TNF-a release before and during isolated two-legged exercise after 14 days of one-leg immobilization (IM) while the other leg served as the control (CON) leg. Fifteen healthy male subjects (mean ± SEM age, 23 ± 1 years; body mass index, 23.6 ± 0.7 kg m; and maximal oxygen uptake, 46.8 ± 1.4 ml kg min) performed 45 min of two-legged dynamic knee-extensor exercise at 19.6 ± 0.8 W. Arterial and femoral venous blood samples from the CON and the IM leg were collected every 15 min during exercise, and leg blood flow was measured with Doppler ultrasound. The arterial plasma IL-6 concentration increased (P <0.05) with exercise (rest, 1.3 ± 0.1 pg ml; 15 min, 1.9 ± 0.2 pg ml; 30 min, 2.4 ± 0.2 pg ml; and 45 min, 3.1 ± 0.3 pg ml). Interleukin-6 release occurred after 15 min of exercise, and the release from the IM leg was significantly greater compared with the CON leg after 45 min (1114 ± 152 versus 606 ± 14 pg min, respectively, P <0.05). Tumour necrosis factor-a release did not differ between the CON and the IM leg, and arterial concentrations remained unchanged during exercise (P > 0.05). In conclusion, prior immobilization enhances release of IL-6 from the leg during exercise at a moderate workload, and the release is already present in the early phase of exercise. Neither immobilization nor exercise had an effect on TNF-a release in the working legs.