Although lack of assertiveness (LOA) and hostility are established risk factors for depression, the mechanisms linking them to depression and their genetic or environmental contributions remain unclear. We used the Swedish Adoption/Twin Study of Aging (SATSA) to test whether social support (SS) deficits mediated these pathways.

Community-dwelling twins (N = 2,034; 1,017 pairs) were followed for up to 23 years (1984–2007). We examined serial (LOA → hostility → SS → depression) and parallel (LOA + hostility → SS → depression) mediation models. Within-pair difference analyses in monozygotic (n = 317) and dizygotic (n = 645) pairs isolated non-shared environmental pathways.

Within monozygotic pairs, where genetic confounding is removed by design, LOA did not predict hostility. Hence, the serial indirect effect was not significant. The hostility → SS → depression indirect effect was significant in the serial model (b = 0.020, 95% CI [0.005, 0.038], β = 0.036); the parallel-model estimate was smaller and in the same direction. LOA and hostility directly predicted depression in both models (β = 0.100 to 0.227, p ≤ .039). Within dizygotic twins, where only 50% of additive genetic variance is removed on average, LOA predicted hostility (β = 0.089, p = .003).

The null monozygotic association between LOA and hostility is consistent with genetic confounding, although limited power and measurement-error attenuation cannot be ruled out. However, the hostility → SS → depression pathway and these direct effects are consistent with environmentally mediated processes that may be actionable intervention targets beyond genetic susceptibility.