It has been demonstrated that healthy centenarians have more favorable anthropometric characteristics and insulin-mediated glucose uptake than aged subjects. The plasma insulin-like-growth factor I (IGF-I) concentration may account for such differences. Three groups of subjects were studied: 1) adults (< 50 yr; n = 30), 2) aged subjects (75-99 yr; n = 30), 3) centenarians (> 100 yr; n = 19). In all subjects, fasting plasma IGF-I, IGF-binding protein-3 (IGFBP-3), leptin, and lipid concentrations were determined; body composition was assessed by bioimpedance analysis; and insulin-mediated glucose up-take was evaluated by euglycemic hyperinsulinemic glucose clamp. IGF-I declined with advancing age, but no differences between aged subjects and centenarians were found. IGFBP-3 showed a trend similar to IGF-I, but lower values were present in centenarians than in aged subjects. Nevertheless, centenarians had a plasma IGF-I/IGFBP-3 molar ratio greater than that in aged subjects. Centenarians had also a whole body glucose disposal (WBGD) greater than that in aged subjects, but similar to that in adults. Mini Mental State Examination (27 +/- 2.1 vs. 18.3 +/- 3.1; P < 0.02) and Instrumental Activities Daily Living (26 +/- 2.6 vs. 8.4 +/- 4.1; P < 0.001) scores were significantly different in aged subjects and centenarians, respectively. In centenarians, the plasma IGF-I/IGFBP-3 molar ratio correlated with the body mass index (r = -0.55; P < 0.009); the amount of body fat (r = -0.62; P < 0.003); fat-free mass (r = 0.56; P < 0.008); fasting plasma leptin (r = -0.63; P < 0.004), triglycerides (r = -0.58; P < 0.01), free fatty acid (r = -0.64; P < 0.005), and low density lipoprotein cholesterol (r = -0.59; P < 0.009) concentrations; Mini Mental State Examination (r = 0.53; P < 0.0.03); and WBGD (r = 0.64; P < 0.005). All correlations were independent of daily fat and carbohydrate intake and WBGD (P < 0.05 for all). No significant correlations between the plasma IGF-I/IGFBP-3 molar ratio and plasma total (r = 0.31; P = NS) and high density lipoprotein cholesterol (r = 0.34; P = NS) concentrations were present. The correlation between the plasma IGF-I/IGFBP-3 molar ratio and WBGD persisted after adjustment for body fat, fasting plasma insulin concentration, daily carbohydrate and fat intake, and daily physical activity (r = 0.55; P < 0.009), but not after further adjustment for plasma free fatty acid concentration (r = 0.30; P = 0.17). In conclusion, healthy centenarians have plasma IGF-I/IGFBP-3 molar ratio greater than aged subjects. A more elevated plasma IGF-I/IGFBP-3 molar ratio might improve insulin action and plasma lipid concentration in centenarians.