A Prospective Twin Study of the Bidirectional Associations Between Depression and Stressful Life Events

Major Depression (MD) and stressful life events (SLEs) commonly occur in close temporal proximity to each other. Theoretical models of depression have proposed various hypotheses explaining this relationship including stress exposure (SLE -> MD), stress generation (MD -> SLE), transactional causation (SLE <-> MD), and common genetic or environmental factors increasing risk for both MD and SLEs. This study compared these theoretical hypotheses by applying multivariate twin modeling to a sample of adult female twins (N = 2,091). The occurrence of MD onsets and various SLEs in each month in the prior year was assessed via interview at up to four waves. MD onset data was extracted for one month per twin pair and SLE data was extracted for the preceding and subsequent six months. MD onset was phenotypically associated with increased risk for any SLE in the subsequent (but not preceding) six months. Twin modeling suggested this association may be explained by environmental factors increasing risk for both MD and a SLE. Additional analyses examined whether associations with MD onset varied across SLEs using incrementally more homogeneous operationalizations of SLEs: (a) any interpersonal SLE, (b) any non-interpersonal SLE, and (c) individual SLEs (e.g., divorce/separation, major financial problems). Common individual-specific environmental risk factors similarly explained MD onset’s associations with any interpersonal SLE and any non-interpersonal SLE. Analyses of individual SLEs suggested that the sources and magnitudes of associations with MD onset did not differ between two individual non-interpersonal SLEs (serious illness/injury and major financial problems), but may vary across individual interpersonal SLEs. The best-fitting model indicated that divorce/separation directly increased risk for MD onset, which in turn directly increased risk for serious marital problems. These two causal pathways were significantly stronger than pathways between MD onset and the other interpersonal SLEs, suggesting potential specificity. Broadly, this study demonstrates some of the promise and challenges associated with using twin modeling to compare competing theoretical hypotheses regarding the association between MD and SLEs.