Androgenetic alopecia (AGA), the most common form of male hair loss, is primarily driven by genetic predisposition and androgen activity. Emerging evidence suggests that AGA is associated with metabolic dysfunction and sleep disorders, but their contribution to its pathogenesis remains unclear. To examine the hormonal, metabolic, and sleep profiles of adult men with self-reported AGA and their potential relevance as risk factors for this condition. This cross-sectional study used data from the 4th Edition of the São Paulo Epidemiological Sleep Study (EPISONO 2018). Participants self-reported AGA, completed sleep questionnaires, underwent polysomnography, and provided blood samples for laboratory measures. A first analysis of all included men was performed, followed by a propensity score matching (PSM), which was applied to match AGA and non-AGA groups by age, body mass index, waist-to-hip ratio, diabetes, and metabolic syndrome. Of the 769 EPISONO participants, 317 were men and 307 answered the AGA question with 64 reporting AGA and 243 classified as non-AGA. Before PSM, the AGA group was older (p = 0.010) and presented higher triglycerides (p = 0.015), lower HDL cholesterol (p = 0.041), lower total (p = 0.025) and free testosterone (p = 0.011), and higher prevalence of diabetes (p = 0.043), metabolic syndrome (p = 0.012), and moderate-to-severe obstructive sleep apnea (OSA; p = 0.041). Subjective sleep measures did not differ between groups. After PSM, 64 matched pairs were analyzed. All anthropometric, hormonal, metabolic, and sleep measures became comparable between groups and the association with OSA no longer persisted. Across univariate and multivariable logistic regression models, before and after PSM, triglycerides remained the only factor associated with AGA (p = 0.039). This study reinforces the link between AGA and lipid metabolism, with elevated triglycerides as a predictor, highlighting the importance of metabolic screening in men with AGA while supporting further investigation of lipid-related mechanisms.
Xerfan et al. (Mon,) studied this question.