ABSTRACT
This network meta-analysis was conducted to compare effects of different placebo-controlled insulin-sensitizing drugs, including metformin, pioglitazone, rosiglitazone, and troglitazone on hormonal parameters in polycystic ovary syndrome (PCOS) patients. We searched PubMed, EMBASE, and Cochrane Library databases from their inception to July 2017. Randomized controlled trials (RCTs) met our inclusion criteria were included. We combined direct and indirect evidences to evaluate weighted mean difference (WMD) value and draw surface under the cumulative ranking curve (SUCRA). Totally 28 eligible RCTs were enrolled. The network meta-analysis results indicated that: Compared with placebo, patients treated with pioglitazone had relatively higher sex-hormone-binding globulin (SHBG) (nmol/L) level (WMD = 6.65, 95%CI = 0.57-12.98), patients treated with metformin had comparatively lower total testosterone (TT) (ng/mL) level (WMD = -0.20, 95%CI = -0.39 to -0.02); Compared with rosiglitazone, patients treated with metformin had relatively higher estradiol (E2 ) (pg/mL) level (WMD = 47.91, 95%CI = 11.44-85.55). However, there were no statistical significance among the placebo-controlled insulin-sensitizing drugs in follicle stimulating hormone (FSH) (IU/L), luteinizing hormone (LH) (IU/L), dehydroepiandrostrone-sulphate (DHEAS) (µg/dL), free testosterone (FT) (pg/mL) and androstenedione (ng/mL). The results of cluster analysis showed that rosiglitazone may be the best drug for PCOS patients regarding to DHEAS, TT, FSH, and LH, metformin may be the best drug for PCOS patients as for E2 , FT, and androstenedione. Rosiglitazone had the best effect on PCOS patients in terms of DHEAS, TT, FSH, and LH, metformin had the best effect on PCOS patients for E2 , FT, and androstenedione.
