An ovine-specific RIA, shown to be reliable for bovine leptin determination, was used to study the effects of breed, body fatness, feeding level, and meal intake on plasma leptin level in adult cattle. Eighteen fat Charolais, fat Holstein, and lean Holstein adult cows were either well-fed (130% of maintenance energy requirements [MER]) or underfed (60% of MER) for 3 wk. The breed tended to have a small effect on plasma leptin level, which was decreased by 70% (P < 0.05) in lean compared to fat Holstein cows. A strong curvilinear relationship was found between mean adipocyte volume and plasma leptin concentrations in well-fed (r = +0.95) and underfed (r = +0.91) cows. Underfeeding caused a significant decrease in plasma leptin levels from 8.0+/-3.1 to 6.1+/-2.3 ng/mL (P < 0.01). Nine adult Holstein cows initially fed at 130% of MER (control) were underfed to 21% of MER for 7 d, and five of them were refed to 237% of MER for 21 d. Plasma leptin measured 1 h before meal distribution was decreased from 5.9+/-0.4 to 3.8+/-0.2 ng/mL (P < 0.01) by underfeeding and increased to reach 8.8+/-1.0 ng/mL (P < 0.01) after refeeding. It was positively related to plasma glucose (r = +0.52, P < 0.01) and negatively related to plasma NEFA (r = -0.67, P < 0.001). Plasma leptin measured 4 h after meal distribution was positively related to feeding level and to plasma 3-OH-butyrate (r = +0.61, P < 0.005) and negatively related to plasma NEFA (r = -0.56, P < 0.01). Differences between pre- and postprandial leptin concentrations showed a decrease after meal intake in control and well-fed cows (-7 and -19%, P < 0.01, respectively) and an increase in underfed cows (+12%, P < 0.01). Leptin response to meal intake was positively related to glucose response (r = +0.66, P < 0.001) and negatively related to 3-OH-butyrate response (r = -0.78, P < 0.001). By using the "multispecies" commercial RIA, leptin concentrations were lower and we observed similar physiological responses, although less related to other hormones or metabolites. These data provide evidence, first, that a specific RIA for ruminant leptin determination is necessary to better understand leptin regulation, and second, that plasma leptin is strongly related to adipose cell size and positively related to feeding level in adult cattle, and that an effect of meal intake could be mediated by glucose and(or) ketone bodies.