Depressive Disorder Subtypes as Predictors of Incident Obesity in US Adults: Moderation by Race/Ethnicity

Abstract

We compared the relative importance of atypical major depressive disorder (MDD), nonatypical MDD, and dysthymic disorder in predicting 3-year obesity incidence and change in body mass index and determined whether race/ethnicity moderated these relationships. We examined data from 17,787 initially nonobese adults in the National Epidemiologic Survey on Alcohol and Related Conditions waves 1 (2001-2002) and 2 (2004-2005) who were representative of the US population. Lifetime subtypes of depressive disorders were determined using a structured interview, and obesity outcomes were computed from self-reported height and weight. Atypical MDD (odds ratio (OR) = 1.68, 95% confidence interval (CI): 1.43, 1.97; P < 0.001) and dysthymic disorder (OR = 1.66, 95% CI: 1.29, 2.12; P < 0.001) were stronger predictors of incident obesity than were nonatypical MDD (OR = 1.11, 95% CI: 1.01, 1.22; P = 0.027) and no history of depressive disorder. Atypical MDD (B = 0.41 (standard error, 0.15); P = 0.007) was a stronger predictor of increases in body mass index than were dysthymic disorder (B = -0.31 (standard error, 0.21); P = 0.142), nonatypical MDD (B = 0.007 (standard error, 0.06); P = 0.911), and no history of depressive disorder. Race/ethnicity was a moderator; atypical MDD was a stronger predictor of incident obesity in Hispanics/Latinos (OR = 1.97, 95% CI: 1.73, 2.24; P < 0.001) than in non-Hispanic whites (OR = 1.54, 95% CI: 1.25, 1.91; P < 0.001) and blacks (OR = 1.72, 95% CI: 1.31, 2.26; P < 0.001). US adults with atypical MDD are at particularly high risk of weight gain and obesity, and Hispanics/Latinos may be especially vulnerable to the obesogenic consequences of depressions.

Keywords: body mass index; depressive disorders; ethnicity; obesity; prospective study; race.

Figure 1.

Logistic regression models examining major depressive disorder (MDD) subtypes in wave 1 (2001–2002) as predictors of 3-year incidence of obesity in wave 2 (2004–2005), stratified by racial/ethnic group (n = 16,724), National Epidemiologic Survey on Alcohol and Related Conditions. The reduced sample size was due to the removal of persons in the “other” race/ethnicity category and in the dysthymic disorder only category. Models were adjusted for age, sex, educational level, wave-1 body mass index, lifetime alcohol use disorders, lifetime tobacco use, lifetime antidepressant use, cardiovascular disease, liver disease, arthritis, and study sampling design. CI, confidence interval; OR, odds ratio.