Type 1 diabetes (T1D) and maternal obesity are independent risk factors for large-for-gestational-age (LGA) neonates. Continuous glucose monitoring (CGM) metrics have shown promise as potential predictors. However, the combined performance of CGM metrics with body-mass-index (BMI) remains underexplored. This study aimed to determine whether combining CGM metrics with BMI improves prediction of LGA compared to CGM metrics alone.
Pregnant women with T1D at Royal North Shore Hospital using CGM were recruited (2019-2024). Anthropometrics were recorded in trimester 1, and CGM metrics were downloaded concurrently. Women were stratified based on incidence of LGA, with statistical analysis completed in GraphPad. Normally distributed data was presented as mean ± SD and non-normal data as median (IQR). Area under the receiver operator characteristic (AUROC) curve was used to evaluate predictive potential.
In total, 20 women were included in this study, of which 45% delivered LGA neonates. These women had higher BMI (29.7±5.5kg/m2 vs. 22.7±3.3kg/m2, p=0.0024) and significantly higher rates of automated-insulin-delivery device use (88.9% vs. 18.2%, p=0.012). In trimester 1, coefficient of variance and time-below-range were lower in women who delivered LGA neonates (32.9±5.4% vs 38.6±5.0%, p=0.024, and 1.0(0.8-3.3)% vs 7.0(4.0-9.5)%, p=0.019 respectively). The combination of these significant CGM metrics with BMI outperformed these same CGM metrics alone (AUROC 0.91 vs. 0.83).
CGM data should be considered in the context of BMI when discussing relative risk of LGA in women with T1D. Future work should validate these findings in larger clinical cohorts and explore whether targeting these parameters reduces the incidence of LGA.