Background
Polycystic ovary syndrome (PCOS) affects up to one in five women of childbearing age. Observational studies assessing the association between maternal PCOS and adverse obstetric outcomes have reported varying results, depending on patient population, diagnostic criteria for PCOS and covariates accounted for in their analyses. We aimed to assess the risk of obstetric outcomes among a population-based representative cohort of women with PCOS compared to an age-matched cohort of women without PCOS.

Methods
A retrospective cohort study was conducted of pregnancies of women in England aged 15–49 years identified from the Clinical Practice Research Datalink (CPRD) GOLD pregnancy register and linked Hospital Episodes Statistic (HES) data between March 1997 and March 2020. Pregnancies from the register that had a linked HES delivery record were included. Linked CPRD primary care data was used to ascertain maternal PCOS exposure prior to pregnancy. To improve detection of PCOS, in addition to PCOS diagnostic codes, codes for (1) polycystic ovaries or (2) hyperandrogenism and anovulation together were also considered. Sensitivity analysis was limited to only pregnant women with a diagnostic code for PCOS.

Primary outcomes ascertained from linked HES data were (1) preterm delivery (gestation < 37 weeks), (2) mode of delivery, (3) high (> 4000 g) or low birthweight (< 2500 g) and (4) stillbirth. Secondary outcomes were (1) very preterm delivery (< 32 weeks), (2) extremely preterm delivery (< 28 weeks), (3) small and (4) large for gestational age.

Conditional logistic regression models were performed adjusting for age, ethnicity, deprivation, dysglycaemia, hypertension, thyroid disorders, number of babies born at index pregnancy, and pre-gravid BMI. Multiple imputation was performed for missing outcome data.

Results
27,586 deliveries with maternal PCOS were matched for age (± 1 year) to 110,344 deliveries without PCOS. In the fully adjusted models, maternal PCOS was associated with an increased risk of (1) preterm birth [aOR: 1.11 (95% CI 1.06–1.17)], and (2) emergency caesarean, elective caesarean and instrumental vaginal compared to spontaneous delivery [aOR: 1.10 (1.05–1.15), 1.07 (1.03–1.12) and 1.04 (1.00–1.09), respectively]. There was absence of association with low birthweight, high birthweight and stillbirth. In the sensitivity analysis, the association with preterm birth [aOR: 1.31 (95% CI 1.13–1.52)], emergency caesarean [aOR: 1.15 (95% CI 1.02–1.30)], and elective caesarean [aOR: 1.03 (95% CI 1.02–1.03)] remained.

While there was no significant association with any of the secondary outcomes in the primary analysis, in the sensitivity analysis maternal PCOS was associated with increased risk of extremely preterm delivery [aOR: 1.86 (95% CI 1.31–2.65)], and lower risk of small for gestational age babies [aOR: 0.74 (95% CI 0.59–0.94)].

Conclusions
Maternal PCOS was associated with increased risk of preterm and caesarean delivery. Association with low birthweight may be largely mediated by lower gestational age at birth.