The impact of vitamin D changes during pregnancy on the development of maternal adverse events: a random forest analysis

Abstract

Abstract Background Maternal vitamin D deficiency during pregnancy has been associated with various maternal adverse events (MAE). However, the evidence regarding the effect of vitamin D supplementation on these outcomes is still inconclusive. Methods This secondary analysis utilized a case–control design. 403 samples with MAE and 403 samples without any outcomes were selected from the Khuzestan Vitamin D Deficiency Screening Program in Pregnancy study. Random forest (RF) analysis was used to evaluate the effect of maternal vitamin D changes during pregnancy on MAE. Results The results showed that women who remained deficient (35.2%) or who worsened from sufficient to deficient (30.0%) had more MAE than women who improved (16.4%) or stayed sufficient (11.8%). The RF model had an AUC of 0.74, sensitivity of 72.6%, and specificity of 69%, which indicate a moderate to high performance for predicting MAE. The ranked variables revealed that systolic blood pressure is the most important variable for MAE, followed by diastolic blood pressure and vitamin D changes during pregnancy. Conclusion This study provides evidence that maternal vitamin D changes during pregnancy have a significant impact on MAE. Our findings suggest that monitoring and treatment of vitamin D deficiency during pregnancy may be a potential preventive strategy for reducing the risk of MAE. The presented RF model had a moderate to high performance for predicting MAE.

Keywords

• Maternal adverse events
• Vitamin D
• Random forest