Metformin as adjuvant therapy to IVF in women with PCOS: when is intention-to-treat unintentional?

Abstract

More than 15 years have passed since metformin was hailed as a possible treatment for infertility in women with polycystic ovary syndrome (PCOS) (Velazquez et al., 1994). Though a recent Cochrane Systematic review examining metformin for ovulation induction concludes that ‘the use of metformin in improving reproductive outcomes in women with PCOS appears to be limited’ (Tang et al., 2010), the debate over metformin’s role in infertility continues passionately. A systematic review is only as good as the studies contributing to it, and there have been multiple critiques of previous randomized trials of metformin in women with PCOS, including an over-representation of obese women, a too short pretreatment period with metformin, possible differences between immediate release and extended release preparations of metformin, and failure to recognize its benefit as an adjuvant therapy for other infertility treatments, for example, IVF/ICSI. Another Cochrane Systematic review of metformin for IVF/ICSI in women with PCOS noted no benefit of metformin on pregnancy rates, but stated ‘The risk of ovarian hyperstimulation syndrome (OHSS) in women with PCOS and undergoing IVF or ICSI cycles was reduced with metformin’ and recommended further multi-center trials (Costello et al., 2006). In this issue of Human Reproduction Kjotrod et al. (2011) have answered the call to provide a multi-center double-blind randomized controlled trial of the use of metformin as an adjuvant to PCOS, as well as identify a population (thin) and a treatment regimen (immediate release metformin given 3 months prior to IVF) that may further increase live birth. This is a well-designed trial, and the best of its kind to date. This tantalizing trial suggests a benefit to metformin in this adjuvant role, but the design and analysis raise as many questions as the study answers. These questions include the use of intention to treat (ITT) analysis for a multi-tiered intervention, the role of IVF in treating anovulatory infertility and the feasibility of randomized trials for a widely popular but unproven adjuvant therapy. The study of Kjotrod et al. (2011) randomized thin or mildly overweight women (BMI , 28 kg/m 2 ) with PCOS to either metformin or placebo for 12 weeks before or during IVF/ICSI, and focused on clinical pregnancy rate as the primary outcome. The sample size was well justified based on a pilot trial which had showed a significant benefit to