Meta-analytic review of longitudinal depression severity in stroke survivors

Depression is a prevalent and debilitating condition among stroke survivors, yet it is potentially treatable and preventable. Previous meta-analyses have primarily focused on prevalence rates, and thus lacked the nuanced insights needed to optimize timing for interventions and prevention strategies. In this meta-analysis, we examined the progression of depression severity from the acute and sub-acute stages up to 36 months post-stroke.

We included only studies involving stroke survivors not undergoing treatment interventions. A multi-level meta-analysis was conducted across 41 eligible studies to assess natural changes in depression severity over time using standardized mean change. We analysed data across all timepoints and conducted sensitivity analyses on potential sources of bias, including publication bias, methodological quality, and low-powered timepoints.

In the full model including all eligible studies, depression severity did not significantly change over time, and time since stroke was not a significant moderator. However, sensitivity analyses excluding studies identified as outliers revealed a small but significant overall reduction in depression severity (95 % CI [-0.123, -0.030]; p = 0.002), and a small increase in depression severity at 12-months post-stroke (95 % CI [0.006, 0.374]; p = 0.043).

Although the overall model did not show significant longitudinal change, sensitivity analyses revealed a potential pattern of overall improvement in depression, with a small resurgence at 12-months post-stroke. The findings underscore the importance of evaluating depression in stroke survivors on a continuous scale, which demonstrates a pattern of post-stroke depression not previously observed using prevalence rates. Future studies should more closely examine the influence of continuous longitudinal post-stroke depression severity on stroke recovery.