Insulin resistance and associated factors in female adolescents from two capital cities in the north and south of Brazil

Abstract

Background: It has been described that physiological changes in glucose metabolism, represented by insulin resistance (IR), are predicted during pubertal evolution, and obesity may be associated with its persistence even at the end of puberty. The aim of this study was to investigate the prevalence of IR in female adolescents with possible associated factors and evaluate the relationship of time since menarche (< 2 vs. ≥ 2 years) in the occurrence of IR in two Brazilian capital cities: Porto Velho (RO) and Porto Alegre (RS). Methods: This is a cross-sectional school-based study, using information from the Study of Cardiovascular Risks (ERICA) database for adolescents aged 12–17 years, enrolled in public and private schools, in municipalities with more than 100,000 inhabitants in Brazil, between 2013 and 2014. The present study included 889 adolescents, 382 in Porto Velho (PVh) and 507 in Porto Alegre (PoA). The homeostasis model assessment for insulin resistance (HOMA-IR) ≥ 3.16 and fasting insulin ≥ 15 mU/L was used to determine the outcome variable of IR. Estimates of crude and adjusted prevalence ratios with confidence intervals of 95% were calculated using Poisson regression with robust variance. Sociodemographic, behavioral, reproductive and nutritional characteristics were considered as potential confounding factors in multivariable models based on a conceptual framework of IR determination. Results: In the total sample, the prevalence of IR was 22.03% (95% CI 17.84–26.89). After adjusting the models, age 15–17 years and time since menarche ≥ 2 years were found to act as protective factors for IR; in contrast, the highest probability of IR was observed in black adolescents, with increased waist circumference (WC) and overweight/obesity (Ow/Ob). The protective effect of two or more years since menarche (post-menarche) was observed for both higher HOMA-IR and fasting insulin in PVh; in PoA, such protection was maintained only for fasting insulin ≥ 15 mU/L after adjustments in the multivariate models.