Introduction: In the US, diabetes affects 7% of pregnancies. 86% of those are secondary to gestational diabetes (GDM). The GDM incidence is increasing worldwide, and gestational diabetes is associated with complications including preeclampsia, cesarean delivery, and development of type 2 diabetes. Oral agents such as glyburide are commonly initiated as first-line treatment for GDM. Inadequate glycemic control on oral agents necessitates transition to treatment with insulin. Past literature has explored prediction of glyburide failure, attributing it to factors of tobacco use, Medicaid, prior GDM, and glucose lab challenge results but not patient age, race, chronic hypertension, and pre-pregnancy BMI Objective: This study aims to identify factors associated with transition to insulin for the treatment of gestational diabetes (GDM) after the failure of oral agents. Methods: Case-control study of 492 patients with GDM treated with oral agents at a single academic center with a comprehensive diabetes program, 2011-2018. Exclusions were pregestational diabetes, patients with gestational diabetes treated with diet alone, and patients not treated with oral agents. The cases were GDM pregnancies treated with any oral medication who transitioned to insulin. Controls were GDM pregnancies treated with any oral agent only. Multivariable logistic regression was performed to identify patient factors associated with failure of oral agents. Receiver-operator characteristic curve estimated the ability of these factors to predict failure of oral agents. Results: Of the 492 patients who were treated with oral agents initially, 35 (7.1%) had inadequate control and were changed to insulin. Those who failed initial treatment with oral agents were more likely to be diagnosed at an earlier gestational age, have a higher 1-Hour GCT lab value, have a 1-Hour OGTT value ≥ 180, initiate treatment earlier, have a longer treatment course, and admission for hyperglycemia. Patients managed on oral medications alone tended to speak a language other than English and had appropriate pregnancy weight gain. After examining the above factors, a receiver-operator characteristic curve derived from the multivariable logistic regression identified protective factors of non-English primary language, non-white race, insufficient gestational weight gain, and risk factors of a value of 180 or more on the first hour of the 3-hour oral glucose tolerance test and treatment initiation prior to 20 weeks gesation. The model has a predictive value of 77% (AUC 0.77). Conclusions: Factors associated with GDM oral agent failure include higher glucose challenge lab results and earlier diagnosis. Conversely, non-English speaking, Black race and Hispanic ethnicity were associated with oral agent success.