Key Takeaways
Extended Takeaways
- In this 52-week, 15-center cohort, lithium exposure was sustained for a mean of 338 days (SD = 79), which strengthens the clinical relevance of the observed associations beyond brief or intermittent use.
- Lithium’s signal was strongest for earlier suicidal cognition: participants had lower rates of wish to be dead (50.0% vs 61.7%, P = .006) and nonspecific active suicidal thoughts (41.0% vs 52.5%, P = .007), supporting routine assessment of ideation severity rather than only attempts.
- Lithium was not associated with less nonsuicidal self-injury, despite reducing suicidality, so clinicians should not assume the same intervention will address both self-harm phenotypes in bipolar disorder.
- Lower hostility scores with lithium (mean [SD]: 21.42 [7.27] vs 23.03 [7.73], P = .01) suggest that changes in aggression-related traits may be one clinically relevant pathway to monitor alongside mood symptoms and suicidality.
- The association with suicide attempts did not reach significance in time-to-event analysis (adjusted HR = 0.74, 95% CI = 0.53–1.04, P = .08), so clinicians should interpret lithium’s benefit for attempts as supportive but less definitive than the ideation findings.
- Exploratory subgroup results suggest lithium’s association with lower suicidal ideation may be less robust in patients with comorbid anxiety, rapid cycling, or predominant depressive episode type, which may help prioritize closer follow-up in these bipolar disorder subgroups.
