Prediction of significant OHSS by ovarian reserve and ovarian response - implications for elective freeze-all strategy.

Ovarian Hyperstimulation Syndrome (OHSS) remains a risk to women undergoing assisted conception despite available preventative measures, which are usually applied on the basis of ovarian response. We performed a retrospective cohort study with robust ascertainment of OHSS cases in women undergoing treatment using GnRH antagonist. FSH dose was based on Anti-Mullerian Hormone concentration. A total of 1492 cycles were carried out over 18 months. Moderate/severe OHSS occurred in 24 cycles (1.6%). AMH of 35 pmol/L and/or AFC of 20 or more identified 18/24 (76%) OHSS cases. The optimal thresholds for predicting OHSS were 22.5 pmol/L for AMH (sensitivity 87.5%, specificity 60.6%), 19.5 for AFC (sensitivity 70.8%, specificity 67%), and 9.5 for egg numbers (sensitivity 83.5%, specificity 62.7%). Peak oestradiol levels had no predictive value. The utility of egg number is limited as it is only known after the ovulatory trigger has been administered. Thus, ovarian reserve parameters are better than ovarian response at predicting the risk of significant OHSS in GnRH antagonist cycles in modern clinical practice. Patients with a high ovarian reserve are at risk of OHSS even if their ovarian response is not excessive. Decisions about preventative measures should be based on ovarian reserve rather than ovarian response.