Predicting the risk of ovarian hyperstimulation syndrome in women undergoing assisted reproductive technology treatments: a systematic review and quality assessment of prediction models

Abstract

Importance

Ovarian hyperstimulation syndrome (OHSS) is a common iatrogenic complication of controlled ovarian stimulation (COS) in assisted conception. OHSS can be life-threatening and associated with significant morbidity. Several measures could help prevent OHSS; however, accurate risk prediction remains a challenge to enable early prevention.

Objective

To review available prediction models for OHSS in women undergoing assisted conception and to identify the best-performing models for their accuracy, generalizability, and applicability.

Evidence review

We searched electronic databases (MEDLINE, EMBASE, and CENTRAL) until October 2023. We included studies reporting on the development or evaluation of models predicting the risk of OHSS outcomes before or during COS among women undergoing assisted conception. We reported on models’ discrimination, calibration, type of validation, and any implementation tools for clinical practice.

Findings

We screened 5,699 citations and included 14 observational cohort studies reporting on 14 prediction models. The median sample size was 782 participants (range 105–256,381), and the majority of models were developed using logistic regression (13/14, 92.9%). The commonest predictor was maternal age (7/14, 50.0%), followed by number of antral and mature follicles (6/14, 42.9%). Six models were internally validated (6/14, 42.9%), and none were externally validated. Only one model had an implementation platform as a smartphone-based application (1/14, 7.1%). Most of the included studies had an unclear risk of bias (7/14, 50.0%), and only three studies were at low risk (3/13, 21.4%).

Conclusion and relevance

There are no clinically appropriate and validated prediction models for OHSS among women undergoing controlled ovarian stimulation. More research is needed to improve their generalizability and applicability into clinical practice.

PROSPERO

CRD42024509423