The effect of age on the relationship between negative life events and current and new depressive episodes: results from the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil).

ABSTRACT

BACKGROUND: There is scarce data concerning the relationship between negative life events (NLEs)* and current and new depressive episodes by age. METHODS: Cross-sectional (baseline) and prospective analyses (4-year/8-year follow-ups) were performed in 15,105 civil servants in 6 cities in Brazil classified according to age strata at baseline 1st(35-44), 2nd(45-54), 3rd(55-64), and 4th(65-74) years. The independent variable was NLEs in the last year (robbery, hospitalization, death of a relative, financial hardship, and rupture of a love relationship) collected at baseline. The dependent variable was depressive episodes assessed at baseline (current), and 4-year/8-year follow-ups (new). We built logistic (Odds Ratio[OR];95 % Confidence Interval[CI]) in the cross-sectional analysis and Poisson regression models (Relative Risk[RR], [CI]) in the prospective analysis. RESULTS: Robbery, hospitalization, financial hardship, rupture of a relationship were associated with current depressive episodes concentrated in the 35-44 age range, while for new depressive episodes, the association of hospitalization, death of a relative, and financial hardship were concentrated in the 45-54 age stratum. Financial hardship was associated with current depressive episodes in all age-strata 1st OR, 2.77(CI, 1.83-4.19); 2nd OR, 1.71(CI, 1.26-2.34); 3rd OR, 1.68(CI,1.15-2.46); 4th, OR, 4.07(CI, 1.24-13.34), and new depressive episodes 1st RR, 1.45(CI, 1.09-1.93); 2nd RR, 1.46(CI 1.15-1.84); 3rd RR, 1.97(CI, 1.38-2.82). Rupture of a relationship was associated only with current depressive episode, while death of a relative only with new depressive episode. LIMITATION: We cannot rule out the possibility of recall bias, since NLEs` information was self-reported. CONCLUSION: The association between NLEs with depressive episodes was concentrated in youngers.