Adverse live-born pregnancy outcomes among pregnant people with anorexia nervosa.

backgroundPrevious findings related to the association of adverse pregnancy outcomes with anorexia nervosa are mixed.

objectiveThis study aimed to investigate the association of adverse live-born pregnancy outcomes with anorexia nervosa using adjustment modeling accounting for confounding factors, and a mediation analysis addressing the contribution of underweight prepregnancy body mass index and gestational weight gain to those outcomes. STUDY

designThe sample included California live-born singletons with births between 2007 and 2021. The administrative data set contained birth certificates linked to hospital discharge records. Anorexia nervosa diagnosis during pregnancy was obtained from International Classification of Diseases codes on hospital discharge records. Adverse pregnancy outcomes examined included gestational diabetes, gestational hypertension, preeclampsia, anemia, antepartum hemorrhage, premature rupture of membranes, premature labor, cesarean delivery, oligohydramnios, placenta previa, chorioamnionitis, placental abruption, severe maternal morbidity, small for gestational age, large for gestational age, low birthweight, and preterm birth (by timing and indication). Risk of each adverse outcome was calculated using Poisson regression models. Unadjusted risk of each adverse outcome was calculated, and then the risks were adjusted for demographic factors. The final adjusted model included demographic factors, anxiety, depression, substance use, and smoking. A mediation analysis was performed to estimate the excess risk of adverse outcomes mediated by underweight prepregnancy body mass index and gestational weight gain below the American College of Obstetricians and Gynecologists recommendation.

resultsThe sample included 241 pregnant people with a diagnosis of anorexia nervosa and 6,418,236 pregnant people without an eating disorder diagnosis. An anorexia nervosa diagnosis during pregnancy was associated with many adverse pregnancy outcomes in unadjusted models (relative risks ranged from 1.65 [preeclampsia] to 3.56 [antepartum hemorrhage]) in comparison with people without an eating disorder diagnosis. In the final adjusted models, birthing people with an anorexia nervosa diagnosis were more likely to have anemia, preterm labor, oligohydramnios, severe maternal morbidity, a small for gestational age or low-birthweight infant, and preterm birth between 32 and 36 weeks with spontaneous preterm labor (adjusted relative risks ranged from 1.43 to 2.55). Underweight prepregnancy body mass index mediated 7.78% of the excess in preterm births and 18.00% of the excess in small for gestational age infants. Gestational weight gain below the recommendation mediated 38.89% of the excess in preterm births and 40.44% of the excess in low-birthweight infants.

conclusionAnorexia nervosa diagnosis during pregnancy was associated with a number of clinically important adverse pregnancy outcomes in comparison with people without an eating disorder diagnosis. Adjusting for anxiety, depression, substance use, and smoking during pregnancy decreased this risk. A substantial percentage of the excess risk of adverse outcomes was mediated by an underweight prepregnancy body mass index, and an even larger proportion of excess risk was mediated by gestational weight gain below the recommendation. This information is important for clinicians to consider when caring for patients with anorexia nervosa. Considering and treating anorexia nervosa and comorbid conditions and counseling patients about mediating factors such as preconception weight and gestational weight gain may improve live-born pregnancy outcomes among people with anorexia nervosa.