Abstract Background Metabolic syndrome (MetS) affects ∼10% of U.S. adolescents, especially those with obesity. Adolescent obesity is associated with adolescent hyperandrogenemia (HA), while adult HA is associated with increased risk of MetS in both asymptomatic women and women with polycystic ovary syndrome (Torchen et al, Obesity [Silver Spring] 2020). Yet, the relationship between HA and severity of MetS during adolescence remains unclear. We investigated whether free testosterone (T) concentrations independently predict MetS severity in adolescent girls. Methods We performed a weighted analysis using data from the National Health and Nutrition Examination Survey (NHANES) for 2013–2016, utilizing a cross-sectional study design. In particular, we analyzed data from a nationally-representative sample of 230 girls aged 12-19 years. We performed regression analysis to evaluate whether free T (independent variable) predicted MetS Severity Z-score (outcome). In a multiple regression model, we adjusted for total percent body fat, age (months), and menarche status (premenarcheal vs. postmenarcheal). Free T was calculated (Vermuelen equation) from total T and sex hormone binding globulin (SHBG). MetS severity was determined by a calculator validated for children (Gurka et al, Metabolic Syndrome Severity Calculator, doi: 10.5281/zeondo.2542213; Gurka et al, Cardiovascular Diabetology 2012) using reported fasting labs and anthropometric parameters. As secondary analyses, we (1) repeated the above analysis substituting SHBG for free T in the primary cohort, and (2) repeated the above analysis substituting total T for tree T in an expanded cohort (n=251). Results In the primary cohort, MetS z-score was -0.27 (-0.41 to -0.13) [mean (95% CI)]; free T 4.0 pg/ml (3.7–4.3); total T 28.7 ng/dL (27.1–30.3); SHBG 60.8 nmol/L (54.0–67.6); total percent body fat 34.5% (33.3–35.6). In simple regression analyses, free T (R=0.11) and percent body fat (R=0.57) each predicted MetS z-score (p<0.001 for each), but age (p=0.68) and menarche status (p=0.18) did not. Our adjusted regression model accounted for 60% of the variance in MetS z-score (R2=0.60). Higher free T and percent body fat were independently associated with higher MetS z-score, while younger age was independently associated with higher MetS z-score (p<0.001 for each). When substituting SHBG and total T for free T, model R2 values were 0.61 and 0.59, respectively, with only percent body fat and age being significant independent predictors of MetS z-score in each model (SHBG, p=0.06; total T, p=0.90). Conclusion We conclude that free T is an independent predictor of MetS severity in a nationally-representative sample of adolescent girls aged 12 to 19 years. Total T was not an independent predictor of MetS severity, suggesting that the relationship between free T and MetS z-score may partly reflect obesity-associated reductions in SHBG levels. These data suggest that HA may contribute to MetS severity in adolescent girls. Presentation: Saturday, June 11, 2022 1:06 p.m. - 1:11 p.m., Monday, June 13, 2022 12:30 p.m. - 2:30 p.m.
Read full abstract