Evaluating the effect of maternal health conditions on severe maternal morbidity adjusting for emergent cesarean delivery: A mediation analysis approach

Abstract

The prevalence of pre-existing and gestational maternal health conditions is rising; affected women are at increased risk for bothcesarean delivery (CD) and severe maternal morbidity (SMM). We estimated the increased risk of SMM associated with keymaternal health conditions, and determined the extent to which it was directly associated with these conditions versus indirectlyassociated because of an increased risk of CD. Nulliparous, term, singleton, vertex (NTSV) gestations were identified amonglaboring women in California 2008-2010 hospital discharge datasets. Logistic regression models (stratified by younger vs. olderwomen [35 years]) were constructed for each condition: chronic/gestational diabetes mellitus (DM), chronic/gestationalhypertension, obesity, heart disease, and mental health diagnoses. Using mediation analysis, the potential effect associated witheach condition (exposure) on SMM (outcome) consisted of a direct effect and an indirect effect caused by modification of therisk of an emergent CD (mediator). Of 502,654 deliveries, 24.3% had an emergent CD; 1.7% had SMM, which was present in1.6% of younger vs. 2.2% of older mothers. The modeled odds ratios (ORs) for the direct effect on SMM were 2.00-2.67 for anyhypertension, 0.78 for obesity in younger women, 6.57-7.97 for heart disease, and 1.43-1.58 for mental health diagnoses. Nodirect effect was identified for DM or obesity in older women. The ORs for the indirect effect were low, ranging from 0.94-1.11 for all models. For women laboring with NTSV gestations where key maternal health conditions were present, the risk of SMMmediated by an emergent CD was low, relative to the direct risk.