Abstract
OBJECTIVE: The prevalence of depression increases dramatically during puberty in girls. Earlier work in this sample reported that the sex steroids estradiol and testosterone were associated with increased depression in girls. Using three additional data waves (983 new observations), we retest the relative contributions of pubertal timing, pubertal status, and sex hormones on the increases in female depression. METHOD: Eight waves of data from the prospective, representative Great Smoky Mountains Study were used covering female participants in the community who were 9 to 16 years of age (3,005 assessments of 630 girls; 1993-2000). Structured interviews assessed depressive disorders. Youth rated their pubertal status using Tanner stage drawings, and sex steroids were assayed from dried blood spots. RESULTS: Risk for depression during puberty was associated with both age and Tanner stage in univariate models. In adjusted models accounting for pubertal timing and sex steroids, the apparent effects of age and Tanner stage were attenuated both in terms of statistical significance and effect size. The only significant predictors of change in depression status during puberty were early pubertal timing (odds ratio = 5.8, 95% CI = 1.9-17.9, p = .002 after age 12 years) and higher testosterone levels (odds ratio = 2.0, 95% CI = 1.1-3.8, p = .03 for quartile-split variable). CONCLUSION: The added observations have modified the original conclusions, implicating the following: testosterone only, but not estradiol; and early pubertal timing, but not age or pubertal status per se. These findings argue for multiple pubertal determinants of depression risk, including factors that are socially and biologically mediated.