Nutritional and Related Interventions for Metabolic Markers in Polycystic Ovary Syndrome: A Systematic Review and Exploratory Meta-Analysis

Background/Objectives: Nutritional preparations are used by women with polycystic ovary syndrome (PCOS), but their short-term metabolic effects are uncertain. Methods: MEDLINE, Embase and the Cochrane Library were searched up to 4 December 2025 (PROSPERO: CRD42023424883). Between-group mean differences (MDs) in change from baseline were pooled separately for each intervention–comparator pair (restricted maximum likelihood, modified Hartung–Knapp 95% confidence intervals [CIs]). Risk of bias was assessed with the Cochrane risk-of-bias tool for randomized trials, version 2 (RoB 2) and certainty with GRADE (Grading of Recommendations Assessment, Development and Evaluation). Several analyses were post hoc. Results: Of 40 reports of 38 trials, 28 reports (26 trials) entered the primary analyses. Of 62 pooled estimates, two excluded no difference: 25-hydroxyvitamin D (25(OH)D) increased with vitamin D3 (five trials, 298 participants, MD +15.19 ng/mL, 95% CI 8.36 to 22.02, very low certainty) and homeostatic model assessment of insulin resistance (HOMA-IR) decreased with cinnamon (three trials, 185 participants, MD −1.03, 95% CI −1.97 to −0.09, low certainty), but this was not robust to omitting any single trial. Vitamin D3 did not clearly change fasting insulin or glucose. D-chiro-inositol estimates were highly imprecise. Of 176 graded estimates, 67 were of low and 109 of very low certainty. Conclusions: Apart from the expected rise in 25(OH)D, nutritional preparations showed no robust short-term metabolic effect in PCOS. The evidence does not support ranking preparations or treatment recommendations.