In animal models, conflicting results on the effect of cirrhosis on glucose metabolism have been reported. The use of various toxins as well as differences in experimental protocols may be responsible for these controversial data. However, differences may be also be explained by the fact that glucose metabolism has been evaluated following different time intervals after cessation of the toxic injury. Therefore, we have performed intravenous glucose tolerance tests, euglycemic hyperinsulinemic clamps (at 2,6, and 30 mU/kg/min insulin infusion rates), and determination of peripheral tissue glucose metabolic index (by [3H]2-deoxy-glucose injection) in rats treated for 10 weeks with carbon tetrachloride, either 3 days (acute group) or 2 weeks (delayed group) after the last CCl4 dose was administered. Cirrhosis was confirmed by liver histological analysis, and by a 22% (P <.05) decrease in 13C-aminopyrine demethylation. In the acute group, whole-body glucose disposal was decreased at the highest insulin infusion rate only (19.7 ± 1.2 vs. 23.4 ± 1.2 mg/kg/min in controls, P <.05). In contrast, results of the delayed group were not different from controls at any insulin infusion rate. Peripheral tissue glucose metabolic index was significantly decreased in all muscles tested in the acute group compared with controls. A significant decrease of glucose utilization was found in some but not all muscles in the delayed group but was less pronounced than in the acute group. In conclusion, this study showed than insulin sensitivity in cirrhotic rats is time-dependent with regard to the last CCl4 administration. These results must be taken into account when using this experimental model of liver cirrhosis.