Correlation of Polycystic Ovarian Syndrome Phenotypes With Pregnancy and Neonatal Outcomes

OBJECTIVE: To compare pregnancy and neonatal outcomes in women with hyperandrogenic polycystic ovarian syndrome (PCOS) phenotypes compared with nonhyperandrogenic PCOS phenotypes. METHODS: We conducted a retrospective cohort study of participants in the PPCOS (Pregnancy in Polycystic Ovary Syndrome) I and II randomized controlled trials; all of the participants met the National Institutes of Health diagnostic criteria for PCOS and were then sorted into three of the four Rotterdam criteria categories based on medical interview, demographics, physical examination, and laboratory data. The two hyperandrogenic (A and B) Rotterdam categories were compared with the nonhyperandrogenic phenotype of PCOS (phenotype D). Our outcomes of interest were clinical pregnancy, pregnancy loss, live birth, obstetric complications (including preterm labor, preeclampsia, gestational diabetes, intrauterine growth restriction, and premature rupture of membranes), and neonatal outcomes (including jaundice, respiratory distress syndrome, neonatal hospitalization, and neonatal infection). RESULTS: Of the 1,376 participants included in the study, 1,249 (90.8%) had hyperandrogenic PCOS phenotypes compared with 127 (9.2%) nonhyperandrogenic PCOS (nonhyperandrogenic PCOS). Compared with participants with nonhyperandrogenic PCOS, those with hyperandrogenic PCOS had higher body mass index (BMI) (35.5±8.9 vs 31.9±9.3 kg/m 2 , P <.001), fasting insulin (21.6±27.7 vs 14.7±15.0 micro-international units/mL, P <.001), and homeostatic model assessment for insulin resistance score (5.01±9.1 vs 3.4±4.1, P =.0002). Age and race were similar between groups. Months attempting pregnancy were greater in participants with hyperandrogenic PCOS compared with nonhyperandrogenic PCOS (41.8±37.3 vs 33.9±32.0). The proportion of participants who achieved pregnancy (29.9% vs 40.2%, P =.02) and live birth rates (20.1% vs 33.1%, P =.001) were lower among those with hyperandrogenic PCOS compared with nonhyperandrogenic PCOS, although pregnancy loss rates did not differ significantly (23.9% vs 32.3%, P =.06). The hyperandrogenic PCOS group had lower odds of live birth compared with the nonhyperandrogenic PCOS group (odds ratio [OR] 0.51, CI, 0.34–0.76), even after adjusting for BMI (adjusted odds ratio [aOR] 0.59, CI, 0.40–0.89). The hyperandrogenic PCOS group also had lower odds of achieving pregnancy compared with the nonhyperandrogenic PCOS group (OR 0.63, CI, 0.44–0.92); however, this association was no longer significant after adjusting for BMI (aOR 0.74, CI, 0.50–1.10). The overall low prevalence of prenatal complications and neonatal outcomes precluded a meaningful comparison between the two groups. CONCLUSION: Participants with hyperandrogenic PCOS achieved lower rates of pregnancy and live birth compared with those with nonhyperandrogenic PCOS. Evaluating distinct PCOS phenotypes may allow for individualized guidance regarding the probability of pregnancy and live birth. CLINICAL TRIALS REGISTRATION: ClinicalTrials.gov, NCT00068861 and NCT00718186.