Correlation of Continuous Glucose Monitoring and Glucose Tolerance Testing With Pregnancy Outcomes

For each patient's CGM data, we calculated the total area under the curve (AUC) for values above the predefined cutoffs of 110, 120, 130, 140, and 180 mg/dL. Patients wore the CGM for different amounts of time; therefore, the total AUC for the entire duration of CGM use was divided by the number of 24-hour periods of data collection. We called these normalized values "AUC-110," "AUC-120," "AUC-130," "AUC-140," and "AUC-180," and they reflect both the magnitude and duration of hyperglycemic excursions above the predetermined thresholds in an average 24-hour period. Birth weight percentile was determined using birth weight data derived from 1999 and 2000 United States Natality datasets. The correlation coefficient (r) was calculated between birth weight percentiles and each of the following: AUC-110, AUC-120, AUC-130, AUC-140, AUC-180, and 1-hour GCT result.

For our secondary outcome analyses,we chose to focus on AUC-130 because 130 mg/dL is a common threshold used when treating gestational diabetics. In addition, 130 mg/dL was the threshold used in an earlier pilot study performed at our institution because it had the best correlation with birth weight percentile.

Secondary outcomes were compared between these two groups using the chi-square test. Data were analyzed using Stata 11.2. AUC-130 values were divided into "high" and "low" at a cutoff of 22,000, which was the 90th percentile of AUC-130 values.