Metformin vs thiazolidinediones for treatment of clinical, hormonal and metabolic characteristics of polycystic ovary syndrome: a meta‐analysis

Summary Objectives  Insulin-sensitizing drugs (ISDs) have been advocated for the long-term treatment of polycystic ovary syndrome (PCOS). It is therefore important to compare the efficacy and safety of ISDs such as metformin and thiazolidinediones (TZDs) for the treatment of this syndrome. Methods  A meta-analysis to assess the effectiveness and safety of metformin vs TZDs (including pioglitazone and rosiglitazone) in the treatment of PCOS was conducted, using MEDLINE (1966–May 2010) and EMBASE (1988–May 2010) to select randomized controlled trials comparing clinical, hormonal and metabolic results. Results  Ten trials were included. TZDs were superior to metformin in reducing serum levels of free testosterone (P = 0·03) and dehydroepiandrosterone sulfate (DHEA) (P = 0·002) after 3 months treatment. Decreases in triglyceride levels were more pronounced with metformin after 6 months (P < 0·0001). Decreases in body mass index (BMI) were greater with metformin treatment as assessed at 3 and 6 months (P < 0·00001). There were no significant between-group differences concerning improvements in ovulation, pregnancy rate, menstrual patterns or insulin sensitivity, or changes in serum levels of androstenedione, luteinizing hormone, follicle-stimulating hormone, total cholesterol, low-density lipoprotein C or insulin. Metformin caused a significantly higher incidence of side effects such as nausea, diarrhoea and abdominal cramping (P < 0·00001). Significant between-study heterogeneity was detected for several variables assessed. Conclusions  The findings from this meta-analysis do not indicate that metformin is superior to TZD’s for the treatment of PCOS or vice versa. Between studies, heterogeneity was a major confounder. A large scale, well-designed, randomized, controlled trial is needed to further address this issue.