Antidepressant use and worsening of non-suicidal self-injury and suicidality in bipolar disorder: A multicenter retrospective cohort study

AI Summary

Antidepressant use in bipolar disorder associated with increased odds of worsening NSSI (OR 1.90), suicidal ideation (OR 1.70), and suicidal behaviour (OR 1.80)
Point estimates larger in low income and non-depressive dominant polarity; association significant only in patients not receiving concurrent lithium
Subgroup differences lacked significant interaction tests and are hypothesis generating; findings indicate caution when prescribing antidepressants in bipolar disorder and need replication

AI Summary

The safety of antidepressants in bipolar disorder (BD) remains controversial, particularly regarding the risk of behavioral activation and worsening Non-Suicidal Self-Injury (NSSI).

This multi-center retrospective cohort study included 575 patients with BD from 15 medical centers in China to evaluate the association between antidepressant use and NSSI frequency, suicidal ideation (SI), and suicidal behavior (SB) over a one-year period.

In exploratory subgroup analyses, point estimates were larger in patients with low household income (OR=2.74, p = 0.038) and non-depressive dominant polarity (OR=7.05, p < 0.001), and the association reached significance only in patients not receiving concurrent lithium (OR=2.29, p = 0.032) but not in lithium-treated patients (OR=1.38, p = 0.462); however, none of the formal interaction tests was statistically significant (all interaction p > 0.05).

Basic summary

Psychiatry Res. 2026 Jun 19;364:117301. doi: 10.1016/j.psychres.2026.117301. Online ahead of print.

ABSTRACT

The safety of antidepressants in bipolar disorder (BD) remains controversial, particularly regarding the risk of behavioral activation and worsening Non-Suicidal Self-Injury (NSSI). This multi-center retrospective cohort study included 575 patients with BD from 15 medical centers in China to evaluate the association between antidepressant use and NSSI frequency, suicidal ideation (SI), and suicidal behavior (SB) over a one-year period. Multivariable logistic regression revealed that antidepressant use was associated with higher odds of the worsening of NSSI (OR=1.90, p = 0.030), as well as the presence of SI (OR=1.70, p = 0.016) and SB (OR=1.80, p = 0.010). In exploratory subgroup analyses, point estimates were larger in patients with low household income (OR=2.74, p = 0.038) and non-depressive dominant polarity (OR=7.05, p < 0.001), and the association reached significance only in patients not receiving concurrent lithium (OR=2.29, p = 0.032) but not in lithium-treated patients (OR=1.38, p = 0.462); however, none of the formal interaction tests was statistically significant (all interaction p > 0.05). These findings indicate that antidepressant use in BD is associated with higher odds of worsening NSSI and suicidality. The observed subgroup differences did not reach statistical significance on interaction testing and should be regarded as hypothesis-generating.

PMID:42330862 | DOI:10.1016/j.psychres.2026.117301

Document this CPD