Does Depression in Adolescence Cause Later Growth in BMI? a Co-Twin Control Analysis

Introduction:

Previous research has found an association between MDD during adolescence and the later onset of obesity (e.g., Marmorstein et al., 2014). We used a longitudinal community-based sample of twins and a co-twin control (CTC) design to examine whether MDD during adolescence is causally related to change in body mass index (BMI) during early adulthood.

In brief, the co-twin difference design used in this study examines whether the experience of having MDD in adolescence increases risk for growth in BMI through early adulthood by modeling differences in BMI growth between twins who experienced MDD and their co-twins who did not.

Methods:

This study utilized data from the community-based, longitudinal Minnesota Twin Family Study (MTFS). The MTFS uses an accelerated longitudinal design with two cohorts, one first assessed at age 11 (n=1512) and one first assessed at age 17 (n=1252). All participants were re-assessed every 3-4 years through age 29. The sample is representative of the population of Minnesota during specified birth years.

MDD was assessed via structured interview and BMI was computed based on study assessments of height and weight. For this study, we combined the cohorts and began with a “depression by age 17” variable, based on the first assessment of the older cohort and a combination of the first 3 assessments of the younger cohort. This variable was used to predict BMI at ages 17, 20, 24/25, and 29.

We used mixed effects models followed by CTC analyses, in which the effect of MDD on BMI was decomposed into a within-twin pair effect and a between-twin pair effect. The within-pair effect yielded the difference on BMI for members of the twin pair who were discordant on their histories of MDD—i.e., an estimate of the effect of exposure to MDD on the intercept and slope (change over time) of BMI. Next, a zygosity-by-within pair interaction term was included in order to examine whether the effect differed for monozygotic (MZ) and dizygotic (DZ) twins.

Results:

MDD by age 17 was associated with smaller negative quadratic slopes of BMI from 17 to 29, indicating that there was less slowing of growth in BMI during early adulthood among individuals with histories of MDD in adolescence. CTC analyses indicated that affected (by MDD) twins had these BMI patterns while non-affected twins did not. Comparisons of MZ and DZ twins were consistent with a causal effect of MDD on future change in BMI.

Conclusions:

MDD during adolescence is associated with less slowing of growth in BMI during early adulthood. Using co-twins without MDD as controls, this study demonstrates that this effect is likely causal: the experience of having MDD in adolescence increases risk for a failure to show a normative moderation in growth in BMI during early adulthood.