Follicle-Stimulating Hormone and Diabetes in Postmenopausal Women: A Systematic Review and Meta-Analysis

Abstract

Abstract Context The co-occurrence of hormonal changes during menopause and the risk of cardio-metabolic disorders has been well studied. Objective We explored the association of circulating levels of follicle-stimulating hormone (FSH) with diabetes (DM) among postmenopausal women. Method In this systematic review and meta-analysis, the search was performed in PubMed, Scopus, Web of Sciences, Epistemonikos, and Cochrane Library up to September 2023. Risk of bias was assessed by Newcastle-Ottawa Quality Assessment Scale. Pooled estimates of mean differences in FSH levels were compared between postmenopausal women with and without DM. Correlations between FSH and fasting blood glucose (FBG)/insulin/homeostatic model assessment for insulin resistance (HOMA-IR) as well as pooled effect sizes with their 95% CIs for risk of DM were calculated. Results In this study, 14 articles, including 7878 postmenopausal women, met eligibility criteria. Most of the included studies had a low/moderate risk of bias. Women with DM had significantly lower FSH levels than those without DM (standardized mean difference [SMD] −0.751 [95% CI, −1.129 to −.372], I2 = 82.46%, n = 1416). The pooled effect size for diabetes was 0.861 (95% CI, 0.740-1.001; I2 = 80.11%). The pooled risk estimate for DM based on the categorical FSH levels (high vs low) was (HR = 0.550; 95% CI, 0.356-0.850, I2 = 0). The significant inverse correlation was found between FSH levels and glycemic parameters: FBG (r= −0.285 [95% CI −0.441 to −0.113]; n = 1229), HOMA-IR (r = −0.241[−0.378 to −0.0924]; n = 1229) and insulin (r = −0.337 [−0.434 to −0.232]; n = 959)]. There were no statistically significant differences between estradiol levels among diabetic and nondiabetic groups; however, the SMD for luteinizing hormone was similar to that reported for FSH. Conclusion The available data indicated an indirect association between FSH levels and glucose disturbances among postmenopausal women, notwithstanding heterogeneity among included studies, and the complexity of various influential factors needs to be considered. Further efforts should be made to clarify the underlying mechanisms.