A comparison of the homeostasis model assessment (HOMA) with direct measurements of insulin resistance in pregnancy with and without gestational diabetes (GDM)

Abstract

MEASUREMENTS OF INSULIN RESISTANCE IN PREGNANCY WITH AND WITHOUT GESTATIONAL DIABETES (GDM) MARTIN MONTORO, SIRI KJOS, ANNY XIANG, THOMAS A. BUCHANAN, University of Southern California, Women’s and Children’s Hospital, Dept. of OB/GYN, Los Angeles, California, HarborUCLA Med. Ctr., Dept. of OB/GYN (LA BioMed), Torrance, California, University of Southern California, Preventive Medicine, Los Angeles, California OBJECTIVE: To correlate third trimester HOMA [fasting serum glucose (mmol/L) ! fasting serum insulin (uU/mL)]/22.5 with insulin resistance and b-cell function in control and GDM women. STUDY DESIGN: 150 GDM and 25 non-diabetic women underwent (a) oral glucose tolerance test (OGTT), (b) intravenous glucose tolerance test (IVGTT) with minimal model analysis and (c) euglycemic clamps with measures of basal and steady state glucose turnover. Exclusion criteria included history of hypertension, previous insulin use and positive anti-pancreatic islet cell antibodies. Linear regression analysis was used to evaluate relationships between HOMA (dependent variable) and four direct measures of insulin sensitivity: (a) MINIMOD SI, (b) clamps steady state (SS) glucose requirements (GINF), (c) clamp ratio of SS/GINF: SS plasma insulin (GINF: ins), and (d) clamp SS glucose clearance rate. RESULTS: During pregnancy, the HOMA calculations were significantly but weakly correlated with MINIMOD SI (R = 0.34, R 2 = 0.12); clamp GINF (R = 0.454, R =0.21), clamp SS GINF: insulin (R = 0.377, R = 0.142) and clamp SS glucose clearance (R = 0.404, R = 0.163). Raw data demonstrate non-linear relationship between HOMA and direct measures of insulin sensitivity in pregnant women. CONCLUSION: The HOMA, an empirical calculation based on fasting insulin has a non-linear relationship to insulin sensitivity. During pregnancy the range of insulin sensitivity is truncated compared to non-pregnant individuals thus limiting the applicability of HOMA as a measure of insulin sensitivity in pregnancy.