New insights into the relationship of antimüllerian hormone with polycystic ovary syndrome and its diagnostic accuracy: an updated and extended meta-analysis using a marginal beta-binomial model

Abstract

Objective

This study aimed to investigate the diagnostic role of antimüllerian hormone in polycystic ovary syndrome using an advanced marginal beta-binomial statistical model, and present the optimal cutoff by different age groups, geographical locations, body mass indexes, and other relevant factors.

Data Sources

A comprehensive and systematic literature search was conducted in Web of Science, PubMed/Medline, Scopus, Cochrane Library, Embase, and ProQuest until August 2024.

Study Eligibility Criteria

Epidemiologic studies that used the Androgen Excess and Polycystic Ovary Syndrome Society, National Institutes of Health, or Rotterdam diagnostic criteria for polycystic ovary syndrome were included in this meta-analysis. Studies were eligible for inclusion if they provided information on the sensitivity and specificity of antimüllerian hormone or related data that allowed for the calculation of these parameters, and/or data on odds ratios and means.

Methods

The diagnostic efficacy of antimüllerian hormone was assessed using the marginal beta-binomial statistical model and the summary receiver operating characteristic method in terms of pooled sensitivity, specificity, and diagnostic odds ratio with 95% confidence interval. Pooled weighted mean difference and pooled odds ratios with 95% confidence interval were estimated using a random effects model.

Results

A total of 202 observational studies were included in the pooled analysis, of which 106 studies (including 19,465 cases and 29,318 controls) were used for meta-analysis of sensitivity/specificity and 186 studies (including 30,656 cases and 34,360 controls) for meta-analysis of mean difference. The pooled sensitivity, specificity, and diagnostic odds ratio for antimüllerian hormone were 0.79 (95% confidence interval, 0.52–0.97), 0.82 (95% confidence interval, 0.64–0.99), and 17.12 (95% confidence interval, 14.37–20.32), respectively. The area under the curve based on the summary receiver operating characteristic model was 0.90 (95% confidence interval, 0.87–0.93). Antimüllerian hormone levels were significantly higher in women with polycystic ovary syndrome than in control women (weighted mean difference, 4.91; 95% confidence interval, 4.57–5.27). In addition, individuals with higher antimüllerian hormone levels were more likely to be affected by polycystic ovary syndrome (odds ratio, 23.17; 95% confidence interval, 18.74–28.66; I²=94%; P<.001). A serum antimüllerian hormone concentration of >5.39 ng/mL was associated with polycystic ovary syndrome (sensitivity, 88.6%; specificity, 92.75%; likelihood ratio for a positive test result, 12.21; likelihood ratio for a negative test result, 0.12).

Conclusion

According to the results of this meta-analysis, serum antimüllerian hormone concentration is a valuable biomarker for the diagnosis of polycystic ovary syndrome. The cutoff points suggested by the current meta-analysis need to be evaluated and validated by future studies before their implementation into clinical practice.