Oral anticoagulants (OACs) are widely prescribed for stroke prevention in patients with atrial fibrillation (AF), prevention and treatment of venous thromboembolism (VTE), and anticoagulant management in patients with heart valves (1–3). While vitamin K antagonists (VKAs), such as warfarin, have served as the cornerstone of anticoagulant therapy for decades, direct oral anticoagulants (DOACs)—have revolutionized clinical anticoagulant practice due to their superior safety profile and simplified medication management (1, 2). Notably, the coexistence of cardiovascular diseases and psychiatric disorders is increasingly prevalent in modern clinical practice, particularly among older adults (4–7), with 4.47–14.1% of patients on oral anticoagulants concomitantly receiving antidepressant therapy (8–10). Selective serotonin reuptake inhibitors (SSRIs), representing a major class of modern antidepressants, are the most widely prescribed in this context (9, 11–13). However, this combination therapy carries a significant risk of bleeding, posing a critical safety concern for clinicians and pharmacists (10, 14–16).

Multiple studies have indicated that antidepressants, especially SSRIs, may independently increase bleeding risk by impairing platelet function (17, 18). When combined with OACs, this risk may be substantially amplified (9, 15, 19). Such synergistic effects can lead to severe adverse events, including gastrointestinal and intracranial hemorrhages, which elevate hospitalization rates and mortality risk, thereby threatening patient safety and necessitating cautious pharmacotherapeutic management (10, 20–22), including careful selection of concomitant medications, dose adjustment, and close monitoring for hemorrhagic events. A 5-year cross-sectional analysis of patients with AF prescribed DOACs revealed that the most common drug classes involved in potential interactions were SSRIs and serotonin-norepinephrine reuptake inhibitors (SNRIs), followed by nonsteroidal anti-inflammatory drugs (NSAIDs), calcium channel blockers (CCBs), and amiodarone. While the overall rate of potential DOAC-related drug interactions slightly declined over the 5-year period, the interaction rate with SSRIs/SNRIs remained relatively stable (23). This underscores the need for heightened vigilance among prescribing clinicians, which should include a comprehensive bleeding risk assessment, review of unnecessary concomitant medications that may increase bleeding risk (such as NSAIDs), consideration of antidepressants with a lower bleeding risk profile when appropriate, use of gastroprotective agents to prevent gastrointestinal bleeding in high-risk patients, closer clinical monitoring for signs of bleeding, and patient education regarding warning symptoms (24–26). These measures can help mitigate the synergistic bleeding risk while maintaining the efficacy of both anticoagulant and antidepressant therapies.

Current evidence on the bleeding risk associated with combined antidepressant and anticoagulant use is limited and has notable constraints. First, well-designed clinical trials often select patients based on strict inclusion and exclusion criteria, which may not reflect real-world patient populations. Second, studies investigating the bleeding risk associated with antidepressants and anticoagulants have primarily focused on SSRIs and SNRIs, leaving uncertainty regarding the impact of other antidepressant classes. Additionally, most studies have employed limited follow-up periods, potentially missing critical events and underestimating the true incidence of adverse events (AEs). Therefore, data from larger-scale, multicenter studies are essential to provide valuable complementary evidence. To address this, we utilized the Food and Drug Administration Adverse Event Reporting System (FAERS) database, the world’s largest real-world AE reporting system for post-marketing surveillance, to aggregate data and identify bleeding signals associated with the co-administration of DOACs and various classes of antidepressants.

This study used real-world data from the FAERS database to explore the impact of antidepressants on DOAC-related bleeding risk. The results showed that compared with DOAC monotherapy, concurrent use of antidepressants significantly increased the bleeding risk, among which the combination of SSRIs and DOACs had the highest bleeding risk. Subgroup analysis of nervous system hemorrhage also indicated that the risk of the combined regimen was higher than that of DOAC monotherapy. This finding is consistent with the conclusions of multiple high-quality cohort studies and meta-analyses (9, 14, 16, 33). For example, a large cohort study in Stockholm, Sweden, found that the incidence of severe bleeding in the combined group (4.7 cases per 100 person-years) was significantly higher than that in the monotherapy group (2.7 cases per 100 person-years), with an adjusted Hazard Ratio (aHR) of 1.42 (9).

To minimize interference from other drug–drug interactions, we implemented rigorous inclusion and exclusion criteria, resulting in a relatively small sample size. Within our included cohort, 80.5% of the cases were individuals over 60 years of age. Several factors may account for this: first, cardiovascular diseases are prevalent among the elderly, making this population the primary recipients of anticoagulant therapy; furthermore, elderly patients with cardiovascular conditions are more likely to present with comorbid depression (6, 12, 34). Secondly, age-related declines in renal and hepatic functions, alongside changes in body composition, significantly alter the metabolism and clearance rates of therapeutic agents, which subsequently leads to an increased incidence of adverse drug reactions (35). This serves as a critical physiological basis for the elevated bleeding risk observed in elderly patients (36). Thirdly, a high proportion of the elderly population suffers from multiple chronic diseases, leading to polypharmacy (37, 38). Additionally, the risk of adverse outcomes increases accordingly in patients receiving multiple medications (39).

Between 2008 and 2012, dabigatran, rivaroxaban, and apixaban were successively launched in Europe and the United States (40–42). Edoxaban was approved in Japan in 2011 and in the United States in 2015 (43). At those times, DOACs were not yet approved in many countries or included in formularies, resulting in lower international post-marketing sales, which in turn delayed the documentation of relevant post-marketing data. Regarding the reporter types, compared with the DOAC monotherapy group, the interaction group had a higher proportion of physician reporters (34.5% vs. 27.5%) and a lower proportion of consumers (33.3% vs. 55.2%). This suggests the presence of reporting bias—physicians may be more inclined to report bleeding events when they are aware of potential drug–drug interactions; therefore, we adjusted for this factor in our subsequent analysis.

The core mechanism underlying the increased bleeding risk is a pharmacodynamic (PD) interaction, mainly resulting from the direct effect of antidepressants on the hemostatic process. Platelet aggregation relies on the 5-hydroxytryptamine (5-HT)-mediated signaling pathway. SSRIs and some SNRIs can deplete intraplatelet 5-HT reserves by inhibiting the serotonin transporter (5-HTT) on the platelet membrane, thereby impairing platelet aggregation function and prolonging bleeding time (17, 18, 21, 23). This antiplatelet effect, combined with the inhibition of coagulation factors by DOACs, forms a “double-hit” that significantly increases the probability of bleeding events (19, 44).

The intensity of bleeding risk is positively correlated with the drug’s binding affinity for 5-HTT (45, 46): paroxetine, sertraline, and fluoxetine have high affinity for 5-HTT; citalopram, escitalopram, etc., have moderate affinity; trazodone, bupropion, and mirtazapine have very low affinity. The lower the affinity, the weaker the inhibition of 5-HT reuptake (47). Shao et al. pointed out that in atrial fibrillation patients receiving oral anticoagulants, initiating SSRIs is associated with a higher bleeding risk than starting other types of antidepressants (8). In the elderly population, the absolute difference in upper gastrointestinal bleeding between users of high-affinity antidepressants and those of low-affinity antidepressants increases with age and the severity of underlying diseases (48). Consistent with this, our study found that the combination of SSRIs and DOACs had the highest bleeding risk (ROR 1.78, 95%CI 1.54–2.04), while no significant bleeding signals were detected for SNRIs and other classes of antidepressants (SNRIs: ROR 1.12, 95%CI 0.88–1.44; other classes: ROR 0.67, 95%CI 0.44–1.03).

Mirtazapine (a noradrenergic and specific serotonergic antidepressant) and bupropion (a norepinephrine-dopamine reuptake inhibitor) are often recommended as alternative medications for patients at high bleeding risk due to their lack of 5-HT reuptake inhibition (49, 50). However, a meta-analysis found no difference in bleeding risk between mirtazapine and SSRIs (51). A potential mechanism is mirtazapine’s strong antagonistic effect on the 5-HT2A receptor, as 5-HT-mediated enhancement of platelet activation in whole blood is mediated by this receptor (52, 53). Our study also confirmed that mirtazapine can increase the bleeding risk of apixaban, and the risk of nervous system hemorrhage associated with their combination is significantly increased, suggesting that it is premature to recommend mirtazapine as an alternative for patients at high bleeding risk.

Intracranial hemorrhage (ICH) is the most dangerous complication in atrial fibrillation patients treated with DOACs, with an annual incidence rate of 0.23–0.80% (54–57). DOAC-related ICH is more severe and has a higher mortality rate than spontaneous ICH (58). Currently, studies on ICH caused by the combination of DOACs and antidepressants are limited, mostly focusing on the combination of SSRIs and anticoagulants (10, 44, 59). Subgroup analysis of nervous system hemorrhage in this study found that SSRIs can increase the risk of DOAC-related cerebral hemorrhage, which is consistent with previously reported research results (60). Specific combinations with strong bleeding signals included edoxaban + sertraline, rivaroxaban + sertraline, and rivaroxaban + escitalopram. In addition, strong bleeding signals were observed for the combination of apixaban with mirtazapine/duloxetine. Therefore, we recommend maintaining high vigilance for nervous system hemorrhage associated with the combined use of anticoagulants and antidepressants (especially SSRIs).

This study has several limitations. First, as a spontaneous reporting system, the FAERS possesses inherent defects, including suboptimal data quality, missing values, and duplicate reports. A significant proportion of submissions by non-healthcare professionals may lead to data inconsistencies or misreporting, thereby introducing potential bias. Second, the data primarily originate from the United States, which may be influenced by factors such as ethnicity, geography, and variations in clinical practice, limiting the generalizability of the findings. Third, bleeding risk is dose-dependent; however, the lack of data on drug dosages and renal function makes it difficult to determine whether the observed risk arises from the drug class itself, individual patient factors, or medication misuse. Fourth, a limited number of reports for certain combinations may inflate signal scores in frequency-based statistical methods, potentially leading to instability in the results. Fifth, due to missing information or small case counts, this study did not adjust for key confounding factors such as DOAC indications (AF vs. VTE), renal function, bleeding history, alcohol consumption, or frailty status. Consequently, these findings should be interpreted with caution in real-world clinical applications.

This study revealed that the co-administration of antidepressants and DOACs significantly increases the risk of bleeding, particularly the risk of nervous system hemorrhage. Among the different classes of antidepressants, SSRIs combined with DOACs showed the highest bleeding risk. However, the bleeding risk associated with mirtazapine, which lacks 5 − HT reuptake inhibition, when combined with DOACs should also not be overlooked. This underscores the need for clinicians to cautiously evaluate the bleeding risk during co-medication. Future research should further explore the underlying mechanisms and optimize therapeutic strategies.