Testosterone concentrations, using different assays, in different types of ovarian insufficiency: a systematic review and meta-analysis.

BACKGROUND Increasing age and post-menopausal status are associated with decreasing androgen concentrations in females. Women with premature loss of ovarian function, such as primary ovarian insufficiency (POI) or iatrogenic menopause may be at increased risk for diminished testosterone levels at a relatively young age. Differentiation between a hypoandrogenic or normoandrogenic state in women with premature loss of ovarian function is problematic due to trueness and precision problems using various testosterone assays. The current meta-analysis was conducted to evaluate current literature reporting serum total testosterone concentrations under these conditions, including stratification for various testosterone assays. METHODS A systematic review and meta-analysis of controlled observational studies were performed. The electronic databases of Pubmed, Embase and the Cochrane Library were systematically searched until October 2011 for comparative studies on total testosterone concentrations in women with spontaneous POI or iatrogenic menopause compared with controls. The literature search, data extraction and critical appraisal, using the Newcastle-Ottawa Scale, were performed by two independent investigators. The effect measure was the weighted mean difference (WMD) with 95% confidence interval (95% CI) in a random effects model. RESULTS A total of 206 articles for spontaneous POI and 1358 for iatrogenic menopause were reviewed, of which 9 and 17 papers, respectively, were selected for final analysis. Both groups demonstrated significantly lower total testosterone concentrations compared with controls [WMD (95% CI) -0.38 (-0.55 to -0.22) nmol/l, and -0.29 (-0.39 to -0.18) nmol/l, respectively], but with substantial between-study heterogeneity. Subgroup analysis for assay type was statistically significant for spontaneous POI only. Sensitivity analyses of high-quality studies did not change the results, and resulted in a substantial decrease in heterogeneity in spontaneous POI studies. CONCLUSIONS The current meta-analysis demonstrates that total testosterone concentrations are decreased in women with spontaneous POI or iatrogenic menopause. The potential implications of hypoandrogenism in these women remain to be elucidated.