Background: The blood coagulation system is dysregulated in patients with cancer. While the risk of thrombosis is well established, bleeding represents an important yet understudied complication. The specific risk factors contributing to the risk of bleeding in cancer remain poorly characterized, limiting individual risk stratification. Use of antidepressant agents is common in patients with cancer, and these agents have been associated with bleeding in non-cancer populations. However, data on bleeding risk associated with antidepressant use in patients with cancer are lacking. Therefore, we aimed to determine the risk of bleeding in patients with cancer using antidepressant agents.

Methods: This analysis was performed within the framework of the Vienna Cancer, Thrombosis and Bleeding (CAT-BLED) study, a prospective, single-center, cohort study of patients with cancer undergoing systemic anti-cancer treatment at the Medical University of Vienna, Austria. At inclusion, electronic health records were reviewed for current medication use. For this analysis, patients were screened for antidepressant use, while those receiving anticoagulation were excluded. Patients were followed for the occurrence of bleeding events, defined as major bleeding (MB), clinically relevant bleeding (CRB), and CRB from the gastrointestinal tract (GI-CRB), which were independently adjudicated by an expert adjudication committee. Twelve-month cumulative incidences were calculated considering death as a competing event. Fine and Gray models adjusted for potential confounders were used to analyze the association of antidepressant use and bleeding events.

Results: Overall, 1,161 patients were analyzed (median age 61 years [IQR: 53-69]; 612 females [52.7%]). The most common cancer sites were lung (243 patients, 20.9%), breast (175, 15.1%), and pancreas (119, 10.3%). At study inclusion, 651 patients (56.1%) had metastatic disease and 714 (61.5%) had newly diagnosed cancer. The most common cancer-specific therapies after inclusion were chemotherapy (570 patients, 49.1%), chemotherapy with immune-checkpoint inhibitors (206 patients, 17.7%), and chemotherapy with targeted agents (155 patients, 13.4%). Antidepressants were used in 218 patients (18.8%) at inclusion, with selective-serotonin re-uptake inhibitors (SSRIs) in 91 (7.8%), serotonin antagonist-re-uptake inhibitors (SARIs) in 75 (6.5%), other agents (mirtazapine, amitriptyline or bupropion) in 68 (5.9%), and serotonin-noradrenaline re-uptake inhibitors (SNRIs) in 13 (1.1%) patients. One-hundred-ninety-three (16.6%) patients were on antiplatelet therapy, of whom 41 were concurrently using antidepressants. The median follow-up duration was 258 days (IQR: 126-506). Over the duration of follow-up, 113 patients had a CRB, 60 a MB, and 50 a GI-CRB, with corresponding 12-month cumulative incidences of 10.4% (95% CI: 8.3-12.4), 6.5% (95% CI: 4.9-8.1) and 5.0% (95% CI: 3.6-6.5), respectively.

The 12-month cumulative incidence of bleeding was higher in antidepressant users vs non-users (CRB: 12.3% [95% CI: 7.5-17.2] vs 9.8% [95% CI: 7.6-12.0]; MB: 11.7% [95% CI: 3.7-18.9] vs 6.0% [95% CI: 0.8-7.7]; GI-CRB 13.2% [95% CI: 5.5-21.0] vs 4.3% [95% CI: 2.8-5.7]). This was largely driven by SSRI users which had the highest incidence of bleeding when compared to non-antidepressant users (CRB: 18.6% [95% CI: 9.7-27.4] vs 9.8% [95% CI: 7.5-12.0]; MB: 11.7% [95% CI: 4.4-18.9] vs 5.9% [95% CI: 0.9-7.6]; GI-CRB 13.2% [95% CI: 5.5-21.0] vs 4.3% [95% CI: 0.8-5.9]).

In multivariable regression (adjusted for age, sex, cancer type, stage, new or recurrent disease, and antiplatelet therapy), SSRI use was associated with CRB (SHR: 2.11 [95% CI: 1.15-3.85]), MB (SHR: 2.53 [95% CI: 1.24-5.54]), and GI-CRB (SHR: 3.81 [95% CI: 1.78-8.17]) when compared to non-antidepressant users. No significant associations were observed for other antidepressant classes. The concomitant use of antiplatelet agents and antidepressants was associated with an increased risk of CRB (SHR: 4.73 [95% CI: 1.72–12.97]), compared to non-antidepressant and non-antiplatelet users.

Conclusions: Patients with cancer using antidepressant agents have a high risk of bleeding, particularly driven by GI-bleeding. Among different types of antidepressants, the highest bleeding-risk was observed among SSRI users. Concomitant use of antidepressants and antiplatelet agents was associated with a particularly high bleeding risk, warranting further investigations.

This content is only available as a PDF.
2025
Sign in via your Institution