2018 Research Forum

Can a fasting insulin level predict gestational diabetes? Presenter: Jamie Markus MD Principal Investigator & Faculty Sponsor: Kurt Finberg MD Kurt Finberg MD 1 , Jamie Markus MD 2 , Brian Jean MS 3 , Nicholas Del Mundo RA 4 , Billy Huynh MS 5 , Kareem Tabsh MD 6

1 Associate Faculty 2 Chief Resident R4 3 Kern Medical Research Biostatistician; Professor of Mathematics, Taft College 4 Research Assistant 5 Medical Student MS4 6 Fellow, Maternal-Fetal Medicine, University of Arizona School of Medicine

INTRODUCTION In the United States, screening for gestational diabetes mellitus (GDM) has typically been performed at 24-28 weeks with either the one-step or two-step approaches. Various other laboratory markers have shown promise in aiding in the screening and diagnosis of GDM. PURPOSE The aim of this study Is to determine if measurement of fasting insulin (Fl) could predict abnormal results on 2-hour or 3-hour oral glucose tolerant tests (OGTT). METHOD FromNovember 2016 t0 June 2017, two cohorts of consecutive secondandearly third trimester pregnant patients were evaluated prospectively with fasting insulin levels at the time of either one step testing for gestational diabetes (fasting and 2-flour post 75-gram glucose challenge) or three-hour OGTI (after abnormal glucose screening with SQ-gram glucose challenge). Gestational age, body mass Index, fasting Insulin levels and results of glucose tolerance testing were collected. Glucose levels for the 2-hour group were interpreted according to recommendations from the ADA, and glucose levels for the 3-hour group were interpreted according to the recommendations of Carpenter and Coustan. Also, the homeostatic model assessment of insulin resistance (HOMA-IR), a model used to predict insulin levels due to Insulin resistance, was calculated using the formula (Fl)(FG)/405. In testing the viability of fasting insulin and HOMA-IR to predict gestational diabetes as compared to a 3HOGTT, two criterions were used in the analysis. Previous studies and meta-analysis have suggested that one abnormal value in 3HOGTT carries the same potential for adverse pregnancy outcome as the standard definition of GDM (2 abnormal values). Criteria #1 was defined as an abnormal value in any one of these indicators, and Criteria #2 was defined as any two abnormal values. The approach to the 3-hour analysis was identical to that of the 2-hour analysis but was repeated for both Criteria #1 and Criteria #2 using fasting Insulin and HOMA-IR as individual predictors of gestational diabetes. RESULTS Data was collected for 104 patients who received “ 2-hour OGTT, and 115 patients who received 3-hour OGTT. Based on an ROC analysis. threshold values were determined for Fl and HOMA-IR to predict GDM based on the 2HR test and Criteria #l and #2 as outlined above for 3-HR. Those values were used to create indicator variables (1 if greater than or equal and 0 if less than the threshold value). All predictors were found to be statistically significant in a logistic regression model (all p-values <= 0.0001). CONCLUSIONS Given the association between positive 2-hour OGTT test results and the 2-HR FI and HOMA-IR results, and the 3-hour OGTT results for both Criteria #1 and Criteria #2 the FI shows promise as a potential supplementary test in the diagnosis of gestational and overt diabetes mellitus in pregnancy. Out results suggest that further research is warranted to determine if use of the FI can help predict poor outcomes in the fetuses of mothers with abnormal OGTT results

36