PCOS during the menopausal transition and after menopause: a systematic review and meta-analysis

Abstract BACKGROUND Current knowledge about the consequences of PCOS during the late reproductive years and after menopause is limited. OBJECTIVE AND RATIONALE We performed a systematic review and meta-analysis of data on the pathophysiology, clinical manifestations, diagnosis, prognosis, and treatment of women ≥45 years of age—peri- or postmenopausal—with PCOS. SEARCH METHODS Studies published up to 15 April 2023, identified by Entrez-PubMed, EMBASE, and Scopus online facilities, were considered. We included cross-sectional or prospective studies that reported data from peri- or postmenopausal patients with PCOS and control women with a mean age ≥45 years. Three independent researchers performed data extraction. Meta-analyses of quantitative data used random-effects models because of the heterogeneity derived from differences in study design and criteria used to define PCOS, among other confounding factors. Sensitivity analyses restricted the meta-analyses to population-based studies, to studies including only patients diagnosed using the most widely accepted definitions of PCOS, only menopausal women or only women not submitted to ovarian surgery, and studies in which patients and controls presented with similar indexes of weight excess. Quality of evidence was assessed using the GRADE system. OUTCOMES The initial search identified 1400 articles, and another six were included from the reference lists of included articles; 476 duplicates were deleted. We excluded 868 articles for different reasons, leaving 37 valid studies for the qualitative synthesis, of which 28 studies—published in 41 articles—were considered for the quantitative synthesis and meta-analyses. Another nine studies were included only in the qualitative analyses. Compared with controls, peri- and postmenopausal patients with PCOS presented increased circulating total testosterone (standardized mean difference, SMD 0.78 (0.35, 1.22)), free androgen index (SMD 1.29 (0.89, 1.68)), and androstenedione (SMD 0.58 (0.23, 0.94)), whereas their sex hormone-binding globulin was reduced (SMD −0.60 (−0.76, −0.44)). Women with PCOS showed increased BMI (SMD 0.57 (0.32, 0.75)), waist circumference (SMD 0.64 (0.42, 0.86)), and waist-to-hip ratio (SMD 0.38 (0.14, 0.61)) together with increased homeostasis model assessment of insulin resistance (SMD 0.56 (0.27, 0.84)), fasting insulin (SMD 0.61 (0.38, 0.83)), fasting glucose (SMD 0.48 (0.29, 0.68)), and odds ratios (OR, 95% CI) for diabetes (OR 3.01 (1.91, 4.73)) compared to controls. Women with PCOS versus controls showed decreased HDL concentrations (SMD −0.32 (−0.46, −0.19)) and increased triglycerides (SMD 0.31 (0.16, 0.46)), even though total cholesterol and LDL concentrations, as well as the OR for dyslipidaemia, were similar to those of controls. The OR for having hypertension was increased in women with PCOS compared with controls (OR 1.79 (1.36, 2.36)). Albeit myocardial infarction (OR 2.51 (1.08, 5.81)) and stroke (OR 1.75 (1.03, 2.99)) were more prevalent in women with PCOS than controls, the ORs for cardiovascular disease as a whole, coronary artery disease as a whole, breast cancer and age at menopause, were similar in patients and controls. When restricting meta-analysis to studies in which women with PCOS and controls had a similar mean BMI, the only difference that retained statistical significance was a decrease in HDL-cholesterol concentration in the former and, in the two studies in which postmenopausal women with PCOS and controls had similar BMI, patients presented with increased serum androgen concentrations, suggesting that hyperandrogenism persists after menopause, regardless of obesity. WIDER IMPLICATIONS Hyperandrogenism appeared to persist during the late-reproductive years and after menopause in women with PCOS. Most cardiometabolic comorbidities were driven by the frequent coexistence of weight excess and PCOS, highlighting the importance of targeting obesity in this population. However, the significant heterogeneity among included studies, and the overall low quality of the evidence gathered here, precludes reaching definite conclusions on the issue. Hence, guidelines derived from adequately powered prospective studies are definitely needed for appropriate management of these women.