The hypothesis that the restriction of dietary protein during lactation has different impacts on sow metabolic status and milk production according to body weight was evaluated. From 5-months of age until farrowing, the gilts were fed to achieve body weights of 180 or 240 kg at farrowing. At this time, 38 sows were assigned to one of three groups: "180 kg" sows not restricted in dietary protein during lactation (180CP); "180 kg" restricted in protein (180LP), or "240 kg" sows restricted in protein (240LP). Catheters were fitted in the jugular vein of 24 sows and serial blood samples were collected 1 d before and 1 d after weaning. Amongst the protein-restricted animals, heavy sows (240LP) had a smaller appetite than light sows in early lactation, resulting in lower energy and protein intakes in the 240LP than in the 180LP sows. Body protein losses were 8, 11 and 13.5% of calculated body protein mass at farrowing in the 180CP, 180LP and 240LP sows, respectively. At the end of lactation, IGF-I concentrations were lower in the 180LP than in the sows from the other groups, probably because of the uncoupling between GH and IGF-I secretions. Low IGF-I concentrations likely promote lean tissue mobilization. Glucose and insulin profiles suggested an insulin resistance state in the 240LP sows compared with the 180LP sows, which may explain, at least in part, the lower feed intake and body reserve mobilization in these sows. Plasma pre- and post-prandial concentrations of amino acids in late lactation differed among the three treatment groups. Throughout lactation, litters from the 180LP and 240LP sows had a slower growth rate than litters from sows which were not restricted, suggesting that endogenous protein mobilization throughout lactation does not completely compensate for a low protein intake.