Early and late pregnancy loss in women with polycystic ovary syndrome undergoing IVF/ICSI treatment: a retrospective cohort analysis of 21 820 pregnancies

Objective To examine early and late pregnancy loss in women with and without polycystic ovary syndrome (PCOS) undergoing IVF/ICSI transfers. Design Retrospective cohort study. Setting Reproductive medicine centre at a tertiary hospital. Population We studied women with a positive beta-human chorionic gonadotropin (beta-hCG) after in vitro fertilisation/intra-cytoplasmic sperm injection (IVF/ICSI) treatment from May 2014 to April 2019. Methods Odds ratios (OR) for early (<= 13 weeks) and late (>13 weeks) pregnancy loss were calculated among women with and without PCOS for plurality of the pregnancy with adjustment for confounding factors. Main outcome measures Early pregnancy loss (EPL) and late pregnancy loss (LPL). Results From 21 820 women identified with a positive beta-hCG, 2357 (10.8%) women had PCOS, and 19 463 (89.2%) women did not. EPL occurred in 16.6% (391) of women with PCOS versus 18.3% (3565) in women with non-PCOS (OR 0.89, 95% CI 0.79-0.99, P = 0.04). After adjustment for age and other confounders, the rate of EPL was not statistically significantly associated with PCOS status (adjusted OR [aOR] 0.91, 95% CI 0.80-1.05). Women with PCOS demonstrated a higher rate of LPL (6.4% in PCOS versus 3.6% in non-PCOS, OR 1.81, 95% CI 1.48-2.21, P < 0.001). In multivariable analysis, the potential impact of PCOS was less strong (aOR 1.38, 95% CI 0.96-1.98), with BMI and maternal comorbidities also associated with LPL (aOR 1.08, 95% CI 1.04-1.1 and aOR 2.07, 95% CI 1.43-3.00, respectively). Conclusions Polycystic ovary syndrome was not independently associated with EPL. There was an increased risk of LPL but this difference was not statistically significant.

Automated summary

Polycystic ovary syndrome was not independently associated with early pregnancy loss. There was an increased risk of late pregnancy loss but this difference was not statistically significant.