4 key variables may best predict PCOS
A new risk prediction score for polycystic ovary syndrome suggests that four clinical variables may best predict the condition in women, including free androgen index, anti-Müllerian hormone levels, waist circumference and 17-hydroxyprogesterone levels, according to findings published in Clinical Endocrinology.
Biochemical hyperandrogenism is a common component of the three available diagnostic criteria for PCOS from the NIH, Rotterdam consensus criteria and the Androgen Excess and PCOS Society, and can be assessed in multiple ways, Harshal Deshmukh, MBBS, MPH, MRCP, PhD, National Institute for Health Research (NIHR) clinical lecturer in endocrinology and diabetes at Hull York Medical School (HYMS) and Hull Royal Infirmary in the United Kingdom, and colleagues wrote in the study background. However, the number of available androgen biomarkers and indices — combined with little available guidance on cutoffs indicative of androgen excess — contributes to diagnosis- and risk stratification-related uncertainties, they wrote.
“While the focus has been placed upon biochemical and clinical hyperandrogenemia for the diagnosis of PCOS, recent data by our group and others have shown that elevated levels of anti-Müllerian hormone (AMH), a surrogate measure of follicle count on ultrasound, can be an important supplement to the hormonal parameters used in the diagnosis of PCOS,” Dashmukh and colleagues wrote. “The aim of this study was to use relevant biochemical markers and quantifiable clinical features to derive a risk score that can capture the PCOS disease spectrum.”
Deshmukh and colleagues analyzed data from 111 women with PCOS who met Rotterdam diagnostic criteria and 67 women without PCOS. Researchers measured BMI, waist circumference, hip circumference, AMH, salivary testosterone, total testosterone, salivary androstenedione, serum androstenedione, sex hormone-binding globulin, free androgen index, follicle-stimulating hormone, luteinizing hormone, fasting glucose, 2-hour glucose and insulin in all women. The researchers used logistic regression analysis with Firth’s bias-adjusted estimates to account for a small sample size or predictors that are strongly associated with one of the possible outcomes and calculated a prognostic index and risk score for developing PCOS using independent variables from the regression model.
Women in the cohort with PCOS tended to be younger (P = .01) and had higher BMI (P < .0001) and larger waist circumference (P < .0001) vs. controls, as well as higher levels of 17-hydroxyprogesterone (P = .03) and AMH (P < .0001).
The researchers found that the Firth’s logistic regression model identified four independent predictors of PCOS, including free androgen index (beta = 0.3; P = .008), 17-hydroxyprogesterone (beta = 0.2; P = .026), AMH (beta = 0.04; P < .0001) and waist circumference (beta = 0.08; P < .0001).
Women with PCOS with a high risk score (quartiles 1, 2 and 3 vs. quartile 4) presented with a worse metabolic profile, the researchers noted, characterized by higher 2-hour glucose (P = .01), serum insulin (P = .0003), triglycerides (P = .0005) and C-reactive protein (P < .0001) and lower HDL cholesterol (P = .02) vs. women with a lower risk score for PCOS.
The model, they noted, showed good discrimination ability and good calibration, although it still requires external validation.
“This model will have to be externally validated in populations across different ethnicities before a widespread clinical application,” the researchers wrote. – by Regina Schaffer
Disclosures: The authors report no relevant financial disclosures.