Hum Reprod Update. 2026 Oct 9:dmag032. doi: 10.1093/humupd/dmag032. Online ahead of print.
ABSTRACT
BACKGROUND: Women with polycystic ovary syndrome (PCOS) face an elevated long-term risk of type 2 diabetes and cardiovascular disease, largely driven by clustered cardiometabolic abnormalities. Although lifestyle interventions primarily involving exercise and dietary interventions have shown benefits, the optimal strategy remains uncertain.
OBJECTIVE AND RATIONALE: We aimed to systematically evaluate and compare multiple lifestyle interventions for cardiometabolic outcomes in women with PCOS, and explore potential dose-response relationships between intervention effects and key covariates such as exercise dose. Outcomes were grouped into five domains: anthropometry and body composition, glucose and insulin regulation, blood lipid profiles, inflammation, and blood pressure, together comprising 16 outcomes. This study was designed to address the limited comparative evidence across lifestyle interventions and the lack of quantitative evidence on factors that may modify intervention effects, with the aim of informing individualized lifestyle prescriptions and cardiometabolic risk management in women with PCOS.
SEARCH METHODS: A systematic search across five databases retrieved randomized controlled trials of lifestyle interventions in PCOS, spanning records from inception to 9 June 2026, including lifestyle interventions targeting cardiometabolic outcomes in women with PCOS. Studies involving pharmacological interventions or with an intervention duration of <2 weeks were excluded. An arm-based multilevel Bayesian random-effects network meta-analysis was used to synthesize arm-level change scores as mean differences (MD), with effects reported as MDs and 95% credible intervals (CrIs); treatments were ranked using the surface under the cumulative ranking curve (SUCRA). The risk of bias was assessed using the Cochrane Risk of Bias tool 2, and the confidence in each network estimate was evaluated using the Confidence in Network Meta-Analysis (CINeMA) framework.
OUTCOMES: After exclusions, 43 studies with 2433 participants were included. Compared with the respective control condition, within the dietary network, the ketogenic diet reduced waist circumference (No (number of participants) = 86, k (number of study arms) = 2; MD = -5.74 cm, 95% CrIs -8.48 to -3.04) and body mass index (No = 86, k = 2; MD = -2.61 kg/m², -4.53 to -0.77). Meanwhile, within the exercise network, aerobic exercise ranked best for reducing waist to hip ratio (No = 141, k = 7; MD = -0.03, -0.04 to - 0.02), mind-body exercise ranked best for reducing insulin resistance assessed by the homeostatic model assessment of insulin resistance (HOMA-IR) (No = 72, k = 3; MD = -0.63, -1.06 to -0.24), fasting plasma glucose (No = 37, k = 2; MD = - 0.16 mmol/l, -0.31 to -0.03), triglycerides (No = 59, k = 2; MD = -0.28 mmol/l, -0.55 to -0.01), fasting insulin (No = 72, k = 3; MD = -6.26 pmol/l, -12.14 to -0.50) and low-density lipoprotein cholesterol (No = 59, k = 2; MD = -0.26 mmol/l, -0.42 to -0.14), and high-intensity interval training (HIIT) ranked best for improving high-density lipoprotein cholesterol (No = 117, k = 6; MD = 0.05 mmol/l, 0.01 to 0.10). Aerobic exercise also ranked best for reducing systolic blood pressure (No = 90, k = 4; MD = -3.49 mmHg; 95% CrI, -6.87 to -0.09). Overall, the certainty of evidence was generally rated as low or very low. An exercise dose of approximately 480-1012 metabolic equivalents of task (METs)·min/week was associated with the greatest overlap in predicted benefits across multiple cardiometabolic outcomes, and moderate-to-vigorous intensity exercise of 7.2-9.4 METs, 60 min/session, 3 sessions/week for 13-17 weeks was mainly associated with predicted favourable changes in lipid outcomes.
WIDER IMPLICATIONS: Collective evidence suggests that the ketogenic diet may reduce waist circumference and body mass index; HIIT and mind-body exercise appeared beneficial across multiple cardiometabolic outcomes. An exercise dose of approximately 480-1012 METs·min/week was associated with the greatest overlap in predicted benefits across outcomes. However, these findings should be interpreted cautiously because the certainty of evidence was generally low or very low, and some estimates were based on a limited number of studies. Nevertheless, this study provides a more targeted basis for exercise prescription and future research in PCOS.
REGISTRATION NUMBER: CRD420251079107.
PMID:42854088 | DOI:10.1093/humupd/dmag032