PSD follows a robust three-trajectory structure across independent cohorts, countries, and instruments, with early clinical factors including cognitive impairment, sex, and vascular risk factors predicting membership in trajectories characterized by persistent depressive symptoms.
Key Findings
Results
A consistent three-class trajectory solution for post-stroke depression emerged across six independent analyses.
Data were drawn from three cohorts of the Stroke and Cognition Consortium (STROKOG; n = 750).
Analyses were conducted within individual cohorts and in pooled GDS-15 and HAMD-17 datasets, totaling six analyses.
The three-class solution was replicated across independent cohorts, countries, and measurement instruments.
Latent class growth analysis (LCGA) was the statistical method used to identify trajectory classes.
Results
The majority of stroke survivors showed no clinically significant depressive symptoms throughout follow-up.
The 'No Depression' class comprised 58–83% of participants across analyses.
This was the largest class identified in every one of the six analyses conducted.
Depressive symptoms were measured using the Geriatric Depression Scale (GDS-15) and the Hamilton Depression Rating Scale (HAMD-17).
Results
A second trajectory class displayed mild depressive symptoms that generally remitted over time.
The 'Mild Remitting' class comprised 14–29% of participants across analyses.
Symptoms in this class were mild and showed a general pattern of remission over the follow-up period.
This class was identified consistently across all six analyses.
Results
A smaller subgroup exhibited moderate-severity depressive symptoms throughout follow-up, with trajectory direction varying across analyses.
The moderate-severity class comprised 4–13% of participants across analyses.
In pooled HAMD-17 analyses, this class was characterized by moderately severe symptoms that gradually resolved over time ('Moderate Improving').
In pooled GDS-15 analyses, this class was characterized by moderate symptoms that remained stable ('Moderate Stable').
The variation in trajectory direction across instruments suggests instrument-specific sensitivity to symptom change.
Results
Baseline global cognitive impairment predicted Mild Remitting class membership across both pooled analyses.
This finding was replicated in both pooled GDS-15 and pooled HAMD-17 analyses.
Cognitive impairment was assessed at baseline as a predictor of trajectory class membership.
This was the only predictor consistent across both pooled analyses.
Results
Female sex predicted membership in the Moderate Improving class in pooled HAMD-17 analyses.
The Moderate Improving class was characterized by moderately severe depressive symptoms that gradually resolved over time.
This association was identified specifically in pooled HAMD-17 analyses.
Female sex was not identified as a significant predictor in the GDS-15 pooled analyses.
Results
Older age, hypertension, and diabetes predicted membership in the Moderate Stable class in pooled GDS-15 analyses.
The Moderate Stable class was characterized by moderate depressive symptoms that remained relatively stable over follow-up.
These three vascular and demographic risk factors were identified specifically in pooled GDS-15 analyses.
These predictors were not replicated in the HAMD-17 pooled analyses, where a different moderate-trajectory class ('Moderate Improving') was identified.
What This Means
This research suggests that depression after stroke does not follow a single uniform pattern — instead, stroke survivors tend to fall into one of three distinct groups based on how their depressive symptoms change over time. Most people (roughly 58–83%) experience little to no depression after their stroke. A second group (14–29%) has mild depressive symptoms that tend to improve on their own. A smaller group (4–13%) experiences moderate levels of depression that either gradually improve or remain persistently elevated. Importantly, this three-group pattern was found consistently across three different patient groups from different countries and using two different depression measurement tools, suggesting it reflects a genuine and reliable feature of post-stroke depression rather than a statistical artifact.
The study also identified several factors measured at the time of the stroke that could help predict which group a person is likely to fall into. Cognitive impairment at baseline was linked to the mild-remitting depression group across multiple analyses. Female sex was associated with moderate-but-improving depression when measured by one scale, while older age, high blood pressure, and diabetes were associated with persistent moderate depression when measured by another scale. These findings suggest that clinicians might be able to identify at-risk stroke survivors early — particularly those with cognitive difficulties or vascular risk factors — and provide targeted mental health support before depressive symptoms become entrenched.
This research matters because post-stroke depression affects roughly one in three stroke survivors and is linked to worse recovery, greater disability, and higher risk of death. By moving beyond simply asking whether someone is depressed or not, and instead tracking how depression changes over time, this study provides a more nuanced picture that could inform personalized care. The replication of findings across multiple independent groups strengthens confidence that these trajectory patterns are meaningful and potentially applicable in clinical settings.
Thurston M, Mattingley J, Lo J, Sachdev P, Desmond D, Traykov L, et al.. (2026). Trajectory classes of post-stroke depression severity and their baseline predictors: A multi-cohort replication study.. Journal of affective disorders. https://doi.org/10.1016/j.jad.2026.122017