IGF-Binding Protein-1 Levels Are Related to Insulin-Mediated Glucose Disposal and Are a Potential Serum Marker of Insulin Resistance: Response to Maddux et al.

We were interested in the article of Maddux et al. (1) in which they found a correlation coefficient of 0.75 for the relationship of IGF-binding protein (IGFBP)-1 with insulin sensitivity. Often, the homeostasis model assessment of insulin resistance (HOMA-IR) or other assessments such as the Stumvoll Index (2) are used. Because hepatic production of IGFBP-1 is under direct inhibitory influence of insulin, plasma IGFBP-1 levels could potentially be of use for the assessment of insulin sensitivity. IGFBP-2 has also been associated with the metabolic syndrome and nutritional status. For 62 nondiabetic subjects who underwent a hyperinsulinemic-euglycemic clamp (3), we determined plasma IGFBP-1 and IGFBP-2 levels through established radioimmunoassays (4). We found a linear correlation of insulin sensitivity (clamp; μm · kg lean body mass⁻¹ · min⁻¹ per pmol/l) with HOMA-IR (r = 0.573, P < 0.0005). As expected, plasma IGFBP-1 inversely correlated with fasting plasma insulin levels and HOMA-IR (r = −0.448 and −0.446, respectively; both P < 0.0005), but the correlation of plasma IGFBP-1 with insulin sensitivity (clamp) had an appreciably lower correlation coefficient (r = 0.228, P = 0.074). Similarly, plasma IGFBP-2 correlated with plasma insulin levels and HOMA-IR (r = −0.271, P = 0.027 and r = −0.262, P = 0.032, respectively) but not with insulin sensitivity (r = 0.141, P = 0.28).To study whether plasma IGFBP-1 levels could be of further use in assessing insulin sensitivity from basal values, we assessed the relationship of insulin sensitivity with age, BMI, waist circumference, fasting plasma insulin, and fasting plasma glucose with multiple linear correlation with and without the inclusion of IGFBP-1. Inclusion of IGFBP-1 did not improve the regression (nor did IGFBP-2, waist circumference, or fasting plasma glucose): insulin sensitivity = 0.141 − 0.0005 × age − 0.00185 × BMI − 0.000773 × fasting insulin (overall R = 0.646, P < 0.0005; insulin in pmol/l).

We conclude that plasma IGFBP-1 has only a very modest relation with insulin sensitivity; in our experience, plasma IGFBP-1 (or IGFBP-2) does not improve the assessment of an index for insulin sensitivity as performed with our modification of the Stumvoll Index (2).