We appreciate the comments of Stefan et al. (1) in this issue of Diabetes Care regarding adiponectin in youth. We are excited to have received the opinion of investigators in internal medicine, since traditionally pediatric investigations seldom cross the boundaries to adult medicine. The changing face of childhood diabetes might be playing a role (2). We offer the following responses.
1) Our study demonstrated associations and correlations between adiponectin and proinsulin and the proinsulin-to-insulin ratio. We never implied that a direct effect of adiponectin on insulin secretory function exists. However, as recommended by Stefan et al., we investigated the relationship between the proinsulin-to-insulin ratio and adiponectin after adjusting for insulin sensitivity per metabolically active fat-free mass. The partial correlation coefficient of adiponectin to the proinsulin-to-insulin ratio after controlling for insulin sensitivity was r = −0.24, two-tailed P = 0.12, and one-tailed P = 0.06. Furthermore, in a multiple regression analysis with proinsulin-to-insulin ratio as the dependent variable and adiponectin and insulin sensitivity as the independent variables, adiponectin (P = 0.007) and not insulin sensitivity (P = 0.60) was the significant independent correlate of the proinsulin-to-insulin ratio (R2 = 0.168, P = 0.02 with adiponectin and insulin sensitivity in the formula; R2 = 0.162, P = 0.007 with only adiponectin in the formula). Also, in an effort to determine if adiponectin is related to an index of glucose homeostasis, we evaluated the correlation of adiponectin with the glucose disposition index (product of insulin sensitivity × first-phase insulin secretion), which revealed that r = 0.35 and P = 0.013.
2) The observation by Roder et al. (3) that the association between the proinsulin-to-insulin ratio and acute insulin response was only present in diabetic adults and not in nondiabetic subjects is contrary to our findings in healthy adolescents. Extrapolating observations from 60-year-old subjects to adolescents may not be justified, especially taking into consideration the uniqueness of puberty-related changes in insulin sensitivity and secretion (2,4). Analysis of our unpublished data demonstrates that first-phase insulin secretion during a hyperglycemic clamp correlates positively with proinsulin (r = 0.43, P = 0.04) and the proinsulin-to-insulin ratio (r = 0.50, P = 0.01) in normal-weight adolescents (n = 23) with no correlations in obese adolescents (n = 26). Thus, in normal adolescents, the puberty-related increase in insulin secretion may also be accompanied by increased proinsulin secretion, whereas in obese adolescents, this relationship may disappear due to variable degrees of β-cell compensation.
In summary, despite the wealth of data in adults with respect to insulin sensitivity and secretion, gradually accumulating data in pediatrics would suggest that developmental differences in these parameters are distinguishing features of youth. At the moment, our data in pediatrics remain supportive of an important relationship between adiponectin and measures of β-cell function.
This work was supported by U.S. Public Health Service Grants Ro1 HD27503, K24 HD01357, and MO1-RR00084, the General Clinical Research Center, and Eli-Lilly.