# Traditional herbal medicine for obesity-related polycystic ovary syndrome: a meta-analysis and data mining study

## TL;DR

This study finds that traditional herbal medicine, when combined with standard treatment, improves outcomes for women with obesity-related PCOS.

## Contribution

The study identifies a core combination of herbal ingredients using data mining and meta-analysis of clinical trials.

## Key findings

- THM combined with conventional therapy significantly improved clinical efficacy and pregnancy rates in PCOS patients.
- Key metabolic markers like HOMA-IR, BMI, and testosterone levels improved with THM treatment.
- Poria cocos, Citrus reticulata, Atractylodes lancea, and Cyperus rotundus were identified as a core herbal combination.

## Abstract

To systematically evaluate the clinical efficacy of traditional herbal medicine (THM) as an adjunctive therapy for obesity-related polycystic ovary syndrome (PCOS) and to identify core botanical drug combinations using evidence synthesis and data mining approaches.

We searched six databases from inception to March 2025 for randomized controlled trials Meta-analyses were performed using Stata 15.1. Association rule mining was performed with the Apriori algorithm. The Cochrane Risk of Bias Assessment Tool (ROB 2.0) and the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework were used to evaluate risk of bias and evidence certainty, respectively.

Seventy-two RCTs involving 5,308 patients were included. Meta-analysis indicated that THM combined with conventional therapy significantly improved the clinical efficacy rate (OR = 3.73, 95% CI: 3.12 to 4.46, p < 0.001; low certainty) and clinical pregnancy rate (OR = 3.03, 95% CI: 2.05 to 4.48, p < 0.001; moderate certainty). Significant improvements were also observed in Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) (SMD = −0.81, 95% CI: −1.02 to −0.60, p < 0.001; very low certainty), body mass index (BMI) (SMD = −0.95, 95% CI: −1.09 to −0.81, p < 0.001; low certainty), total testosterone (TT) (SMD = −0.90, 95% CI: −1.10 to −0.69, p < 0.001; very low certainty), and the luteinizing hormone/follicle-stimulating hormone (LH/FSH) ratio (SMD = −0.88, 95% CI: −1.05 to −0.70, p < 0.001; low certainty). Sensitivity analyses confirmed robustness, but substantial heterogeneity (I2 up to 89.3%) and publication bias (Egger’s test p < 0.05 for several outcomes) were noted. Association rule mining identified Poria cocos, Citrus reticulata, Atractylodes lancea, and Cyperus rotundus as the most strongly associated core botanical drug combination.

THM demonstrated superiority over conventional treatment alone in improving key clinical, metabolic, and reproductive outcomes in obesity-related PCOS. Association rule analysis revealed a core botanical drug combination as promising candidates for future research. However, more rigorously designed, large-scale RCTs are required.

https://www.crd.york.ac.uk/PROSPERO/view/CRD420251111078, identifier CRD420251111078.