Determinants of Severe Late Ovarian Hyperstimulation Syndrome in Fresh Embryo Transfer Cycles Based on Integration of Decision Tree Classification and Conditional Logistic Regression

Background: To explore possible predictors of severe late ovarian hyperstimulation syndrome (SL-OHSS) in fresh embryo transfer cycles. Methods: We conducted a historical cohort study in a Chinese tertiary hospital from January 2017 to December 2019, with a total of 6931 women who had the first fresh embryo transfer included. SL-OHSS was defined as severe symptoms of OHSS occurring 12–17 days after ovulation triggering. Possible determinants of the occurrence of SL-OHSS were identified by a detection decision tree, effects of which were estimated by conditional logistic regression and restricted cubic spline. Results: Elevated estradiol (E2) on the day of ovulation triggering and elevated Anti-Mullerian hormone (AMH) were associated with an increased risk of SL-OHSS (p < 0.001). Women with an E2 on trigger day of ≥3320.2 pg/mL (odds ratio (OR): 2.20; 95% confidence interval (CI): 1.03–4.68), or with an E2 on trigger day of <3320.2 pg/mL and an AMH of ≥4.62 ng/mL (OR: 5.44; 95% CI: 2.29–12.90), had an increased risk of SL-OHSS compared to their counterparts. Women with E2 on trigger day of >3320.2 pg/mL and AMH of >4.62 ng/mL had the highest risk of SL-OHSS (OR: 13.20; 95% CI: 3.87–45.02) when compared with E2 on trigger day of ≤3320.2 pg/mL and AMH of ≤4.62 ng/mL. This association was not linear. Conclusions: Serum E2 concentration on trigger day and AMH levels at baseline may predict SL-OHSS occurrence in fresh embryo transfer cycles. These biomarkers may be valuable in determining the optimum transfer strategy to limit the occurrence of SL-OHSS.