(852) What best predicts for need for hypoglycemic agents in patients with gestational diabetes?

This study aimed to identify antenatal characteristics that best predict the need for hypoglycemic agents in patients with gestational diabetes (GDM). 

   

Study Design:

This study was a single-center retrospective cohort study of patients with singleton pregnancies between 2018-2023, diagnosed with GDM between 24-34 weeks gestation by Carpenter-Coustan criteria on a 3hour OGTT. Hypoglycemic agents were added when optimal control was not achieved as determined by serial finger stick analysis (FBS < 95mg/dl, postprandial values at 1h 140mg/dl , 2h 120mg/dl).Logistic regression models were developed to determine the need for hypoglycemic agents and included commonly associated predictors of disease severity. Data collected included age, body mass index (BMI), race, gestational age (GA) at diagnosis, history of GDM, and the 1h, 2h, and 3h OGTT values.Model 1 utilized only the fasting OGTT value as the predictor.Model 2 included the aforementioned variables excluding the fasting OGTT value.Model 3 incorporated the fasting OGTT value along with the other predictors from Model 2. 

Results:

1,407 patients were diagnosed with GDM;567(40%) were prescribed hypoglycemic agents. 
A comparative analysis of the three models was performed to evaluate their predictive performance(Table 1, Figure 1).Model 1, using only the fasting OGTT value, had an area under the curve (AUC) of 0.77.Model 2, which excluded the fasting OGTT value, had an AUC of 0.71.Model 3, incorporating both the fasting OGTT value and the other predictors from Model 2, achieved an AUC of 0.76.This was significantly different from Model 2 (p  < 0.0001) and Model 1 (p=0.0021) (Table 1).  

Conclusion:

Our model suggests that the fasting value on the OGTT alone (Model 1) is a significant predictor of the need for hypoglycemic agents when compared to common predictors for disease severity (Model 2). Therefore, the isolated fasting value is a valuable clinical predictor in assessing the risk for hypoglycemic agents and should be used when counseling patients.