Association between cognition and suicidal ideation in patients with major depressive disorder: A longitudinal study

•This is the first study to explore the association between cognition and SI in Chinese depressed patients.•Depressed patients with SI showed worse performance on verbal learning than those without SI.•Improvement in speed of processing was negatively associated with reduction in suicidal scores across treatment. Suicidal ideation (SI) is common in patients with major depressive disorder (MDD) and often related to cognitive deficits. Limited longitudinal study has shown that cognitive improvement is associated with reduced SI. However, the comparatively study in Chinese depressed patients is still absent. The objective of this study was to explore the specific cognitive deficits in Chinese MDD with SI and investigate the relationship between changes in cognition and change in SI across antidepressant treatment. Three hundred and five patients with MDD received four weeks of antidepressant treatment. The 17-item Hamilton Depression Rating Scale (HAMD-17) and four domains of the MATRICS Consensus Cognitive Battery (MCCB), including speed of processing, working memory, visual learning and verbal learning were measured at baseline and four-week follow-up. One hundred and thirty patients (42.6%) expressed suicidal ideation. Suicidal patients performed worse on verbal learning than non-suicidal patients. Change in speed of processing domain was negatively associated with change in suicidal scores over time. Logistic regression analysis showed that reduction of SI was associated with improvement of speed of processing. The major limitation was that there was no healthy control group in the current study, which might limit the interpretation of cognitive deficits in depressed patients with SI. Our findings suggest that suicidal patients performed worse on verbal learning which can potentially serve as a cognitive biomarker of suicide risk in MDD. Moreover, reduced suicidal ideation was associated with improved speed of processing.