Background:  Metformin has failed to gain wide acceptance as a first‐line treatment option for women with anovulatory infertility related to polycystic ovary syndrome. This study aimed to ascertain factors that predict fertility success with treatment that included metformin compared to standard (nonmetformin) treatment.
Methods:  Randomised trial data analysis by logistic regression of factors likely to have a differential influence on the likelihood of success of metformin versus nonmetformin treatment amongst women with ovulation dysfunction related to polycystic ovary syndrome.
Results:  For women within a BMI > 32 kg/m2 subpopulation, BMI had a significantly greater impact on the chance of pregnancy amongst women receiving metformin versus those receiving placebo and those with lower BMI who received metformin were more likely to become pregnant than their lower BMI counterparts who received placebo (P = 0.039). The subpopulation of women with BMI ≤ 32 kg/m2 had no factors showing a significantly different impact on the chance of pregnancy for women treated with metformin versus those receiving clomiphene treatment or combination metformin/clomiphene treatment versus clomiphene treatment. There were no significantly different effects of free testosterone, fasting insulin, duration of infertility or ultrasound appearance of polycystic ovaries in any treatment groups.
Conclusion:  This study provides preliminary evidence that BMI may be an important prognostic factor in response to metformin for women with ovulation dysfunction related to polycystic ovary syndrome, suggesting that women with a lower BMI may respond better to metformin treatment versus placebo amongst women with BMI > 32 kg/m2. Individual patient data meta‐analysis of existing randomised trials would clarify this further and would assess whether other factors might predict better response to metformin versus standard treatments.