Longitudinal assessment of insulin sensitivity in pregnancy. Validation of the homeostasis model assessment

OBJECTIVE: To validate the use of the homeostasis model assessment (HOMA) for measurement of insulin sensitivity in obese women during gestation and the postpartum period. DESIGN: Three consecutive measurements of insulin resistance (IR) were performed during and after pregnancy to compare the homeostasis model assessment insulin sensitivity index (HOMA-IR) to glucose utilization rates obtained during hyperinsulinaemic euglycaemic clamps (G(Rd)). PATIENTS: Six obese women (mean second trimester BMI = 30.4 kg/m(2)) with normal glucose tolerance were studied during the second and third trimesters of pregnancy and once in the postpartum period. Thus, there were a total of 18 measurements for analysis. RESULTS: Correlations between the rate of glucose disappearance (G(Rd)) (the gold standard) and the HOMA-derived metabolic parameters of insulin sensitivity were significant, with a multiple R(2) of 43.5% (P = 0.003). However, when controlling for variations between patients using dummy variables, we observed that one patient differed from the other five in the relationship between G(Rd) and HOMA-IR. Applying this regression we obtained a R(2) of 72.6% (P < 0.001). When the regression constant was omitted, we observed that the individual trends during pregnancy and postpartum in two patients differed statistically from the other patients between the two assessments, and we obtained a multiple R(2) of 97.3% (< 0.001). CONCLUSIONS: HOMA estimation of insulin resistance is appropriate for use during both the second and third trimesters of pregnancy and postpartum in obese women with normal glucose tolerance. It lacks sensitivity for the evaluation of individuals, where more precise measures of insulin sensitivity should be utilized.