OBJECTIVETo examine the utility of commonly used insulin sensitivity indices in nondiabetic European Americans (EAs) and African Americans (AAs).RESEARCH DESIGN AND METHODSTwo-hundred forty nondiabetic participants were studied. Euglycemic-hyperinsulinemic clamp was the gold standard approach to assess glucose disposal rates (GDR) normalized by lean body mass. The homeostatic model assessment for insulin resistance (HOMA-IR) and the quantitative insulin sensitivity check index (QUICKI) were calculated from fasting plasma glucose and insulin (FIL). Oral glucose tolerance test (OGTT) was performed to determine Matsuda index, the simple index assessing insulin sensitivity (SIisOGTT), Avignon index, and Stomvoll index. Relationships among these indices with GDR were analyzed by multiple regression.RESULTSGDR values were similar in EA and AA subgroups; even so, AA exhibited higher FIL and were insulin-resistant compared with EA, as assessed by HOMA-IR, QUICKI, Matsuda index, SIisOGTT, Avignon index, and Stumvoll index. In the overall study population, GDR was significantly correlated with all studied insulin sensitivity indices (/r/ = 0.381–0.513); however, these indices were not superior to FIL in predicting GDR. Race and gender affected the strength of this relationship. In AA males, FIL and HOMA-IR were not correlated with GDR. In contrast, Matsuda index and SIisOGTT were significantly correlated with GDR in AA males, and Matsuda index was superior to HOMA-IR and QUICKI in AAs overall.CONCLUSIONSInsulin sensitivity indices based on glucose and insulin levels should be used cautiously as measures of peripheral insulin sensitivity when comparing mixed gender and mixed race populations. Matsuda index and SIisOGTT are reliable in studies that include AA males.

Previous studies have used indirect measures of insulin sensitivity to link circulating amino acids with insulin resistance and identify potential biomarkers of diabetes risk. Using direct measures (i.e., hyperinsulinemic-euglycemic clamps), we examined the relationships between the metabolomic amino acid profile and insulin action (i.e., glucose disposal rate [GDR]). Relationships between GDR and serum amino acids were determined among insulin-sensitive, insulin-resistant, and type 2 diabetic (T2DM) individuals. In all subjects, glycine (Gly) had the strongest correlation with GDR (positive association), followed by leucine/isoleucine (Leu/Ile) (negative association). These relationships were dramatically influenced by BMI, the resting respiratory quotient (RQ), T2DM, and sex. Gly had a strong positive correlation with GDR regardless of BMI, RQ, or sex but became nonsignificant in T2DM. In contrast, Leu/Ile was negatively associated with GDR in nonobese and T2DM subjects. Increased resting fat metabolism (i.e., low RQ) and obesity were observed to independently promote and negate the association between Leu/Ile and insulin resistance, respectively. Additionally, the relationship between Leu/Ile and GDR was magnified in T2DM males. Future studies are needed to determine whether Gly has a mechanistic role in glucose homeostasis and whether dietary Gly enrichment may be an effective intervention in diseases characterized by insulin resistance.

Insulin resistance has been associated with the accumulation of fat within skeletal muscle fibers as intramyocellular lipid (IMCL). Here, we have examined in a cross-sectional study the interrelationships among IMCL, insulin sensitivity, and adiposity in European Americans (EAs) and African Americans (AAs). In 43 EA and 43 AA subjects, we measured soleus IMCL content with proton-magnetic resonance spectroscopy, insulin sensitivity with hyperinsulinemic–euglycemic clamp, and body composition with dual-energy X-ray absorptiometry. The AA and EA subgroups had similar IMCL content, insulin sensitivity, and percent fat, but only in EA was IMCL correlated with insulin sensitivity (r = −0.47, P < 0.01), BMI (r = 0.56, P < 0.01), percent fat (r = 0.35, P < 0.05), trunk fat (r = 0.47, P < 0.01), leg fat (r = 0.40, P < 0.05), and waist and hip circumferences (r = 0.54 and 0.55, respectively, P < 0.01). In a multiple regression model including IMCL, race, and a race by IMCL interaction, the interaction was found to be a significant predictor (t = 1.69, DF = 1, P = 0.0422). IMCL is related to insulin sensitivity and adiposity in EA but not in AA, suggesting that IMCL may not function as a pathophysiological factor in individuals of African descent. These results highlight ethnic differences in the determinants of insulin sensitivity and in the pathogenesis of the metabolic syndrome trait cluster.

Our data show for the first time that skeletal muscle protein-HNE adducts are related to the severity of insulin resistance in sedentary adults. These results suggest that muscle lipid peroxidation could be involved in the development of insulin resistance.