A short course of metformin does not reduce OHSS in a GnRH antagonist cycle for women with PCOS undergoing IVF: a randomised placebo-controlled trial

STUDY QUESTION: Does ‘metformin’ reduce the incidence of ovarian hyperstimulation syndrome (OHSS) for women with polycystic ovary syndrome (PCOS) undergoing a GnRH antagonist assisted conception treatment cycle? SUMMARY ANSWER: A short course of metformin does not reduce the incidence of OHSS for women with PCOS undergoing a GnRH antagonist treatment cycle. WHAT IS KNOWN ALREADY: Metformin does reduce the incidence of OHSS in a GnRH-agonist treatment cycle. STUDY DESIGN, SIZE, DURATION: A randomised placebo-controlled trial (RCT) using metformin or placebo. Randomisation was blinded to both patient and investigator, using a random permuted blocks method with a 50:50 allocation ratio. The study was completed over 5 years (2009–2014) with 153 randomised patients. A sample size calculation based on the incidence of OHSS was completed prospectively suggesting a minimum of 146 recruits was required for the trial with a power of 80% and a type 1 error of 0.05. PARTICIPANTS/MATERIALS, SETTING, METHODS: All patients met the Rotterdam criteria for PCOS and were treated with a standard GnRH antagonist IVF/ICSI treatment cycle in a tertiary infertility clinic. The study medication was started prior to stimulation and continued to oocyte retrieval. Of the 153 patients, 77 received metformin and 76 placebo. MAIN RESULTS AND THE ROLE OF CHANCE: There was no reduction in the incidence of moderate–severe OHSS (Placebo (PLA) 12.2%, metformin (MET) = 16%, 95% CI −0.08–0.16, P = 0.66). There was no difference in total gonadotrophin dose (PLA = 1200, MET = 1200, 95% CI −118.67–118.67, P = 0.75), oocytes retrieved (PLA = 15, MET = 14, 95% CI −2.37–4.37, P = 0.66) or fertilisation rate (PLA = 60.7%, MET = 53.3%, 95% CI −0.96–14.94, P = 0.07). However, using metformin resulted in a reduced clinical pregnancy rate (CPR) per cycle started (PLA = 48.7%, MET = 28.6%, 95% CI 0.04–0.35, P = 0.02) and live birth rate (PLA = 51.6%, MET = 27.6%, 95% CI 0.05–0.40, P = 0.02). Furthermore, when ethnicity was taken into account there was a significant reduction in pregnancy outcome for the South Asian population irrespective of metformin or placebo use (CPR per cycle started, White Caucasian = 44.4%, South Asian = 19.4%; 95% CI 0.06–0.39, P = 0.01). LIMITATIONS, REASONS FOR CAUTION: This study was only undertaken on an infertility population with PCOS with a limited duration of study medication use. WIDER IMPLICATIONS OF THE FINDINGS: This is the first adequately powered RCT to assess the impact of metformin on OHSS in a high-risk group (women with PCOS) undergoing a GnRH antagonist cycle. It does not support the empirical prescribing of metformin as an adjunct to a GnRH antagonist treatment cycle. STUDY FUNDING/COMPETING INTEREST(S): None. TRIAL REGISTRATION NUMBER: EudraCT number 2009-010952-81. TRIAL REGISTRATION DATE: 21 September 2009. DATE OF FIRST PATIENT'S ENROLMENT: 30 October 2009