Mild versus conventional ovarian stimulation for IVF in poor, normal and hyper-responders: a systematic review and meta-analysis

BACKGROUND: Mild ovarian stimulation has emerged as an alternative to conventional IVF with the advantages of being more patient-friendly and less expensive. Inadequate data on pregnancy outcomes and concerns about the cycle cancellation rate (CCR) have prevented mild, or low-dose, IVF from gaining wide acceptance. OBJECTIVE AND RATIONALE: To evaluate parallel-group randomised controlled trials (RCTs) on IVF where comparisons were made between a mild (≤150 IU daily dose) and conventional stimulation in terms of clinical outcomes and cost-effectiveness in patients described as poor, normal and non-polycystic ovary syndrome (PCOS) hyper-responders to IVF. SEARCH METHODS: Searches with no language restrictions were performed using Medline, Embase, Cochrane central, Pre-Medicine from January 1990 until April 2020, using pre-specified search terms. References of included studies were hand-searched as well as advance access articles to key journals. Only parallel-group RCTs that used ≤150 IU daily dose of gonadotrophin as mild-dose IVF (MD-IVF) and compared with a higher conventional dose (CD-IVF) were included. Studies were grouped under poor, normal or hyper-responders as described by the authors in their inclusion criteria. Women with PCOS were excluded in the hyper-responder group. The risk of bias was assessed as per Cochrane Handbook for the included studies. The quality of evidence (QoE) was assessed according to the GRADE system. PRISMA guidance was followed for review methodology. OUTCOMES: A total of 31 RCTs were included in the analysis: 15 in the poor, 14 in the normal and 2 in the hyper-responder group. Live birth rates (LBRs) per randomisation were similar following use of MD-IVF in poor (relative risk (RR) 0.91 (CI 0.68, 1.22)), normal (RR 0.88 (CI 0.69, 1.12)) and hyper-responders (RR 0.98 (CI 0.79, 1.22)) when compared to CD-IVF. QoE was moderate. Cumulative LBRs (5 RCTs, n = 2037) also were similar in all three patient types (RR 0.96 (CI 0.86 1.07) (moderate QoE). Risk of ovarian hyperstimulation syndrome was significantly less with MD-IVF than CD-IVF in both normal (RR 0.22 (CI 0.10, 0.50)) and hyper-responders (RR 0.47 (CI 0.31, 0.72)), with moderate QoE. The CCRs were comparable in poor (RR 1.33 (CI 0.96, 1.85)) and hyper-responders (RR 1.31 (CI 0.98, 1.77)) but increased with MD-IVF among normal responders (RR 2.08 (CI 1.38, 3.14)); all low to very low QoE. Although fewer oocytes were retrieved and fewer embryos created with MD-IVF, the proportion of high-grade embryos was similar in all three population types (low QoE). Compared to CD-IVF, MD-IVF was associated with less gonadotrophin use and lower cost. WIDER IMPLICATIONS: This updated review provides reassurance on using MD-IVF not only for the LBR per cycle but also for the cumulative LBR, with moderate QoE. With risks identified with 'freeze-all' strategies, it may be time to recommend mild-dose ovarian stimulation for IVF for all categories of women i.e. hyper, poor and normal responders to IVF.