Gastrointestinal Bleeding Not Affected by Resumption of Anticoagulation Therapy

When a patient requiring long-term anticoagulant therapy experiences gastrointestinal (GI) bleeding, requiring interruption of therapy, clinicians must decide whether or not to resume anticoagulation following that bleed. Interruption of warfarin following a GI bleed has also been associated with an increased risk of death and thromboembolic events, while resumption of warfarin may be associated with an increased risk of recurrent bleeding.

Chatree Chai-Adisaksopha, MD, from the Department of Medicine at McMaster University in Hamilton, Ontario, and colleagues conducted a literature review of phase III, randomized, controlled trials and cohort studies in patients with atrial fibrillation or venous thromboembolism who received oral anticoagulation to help answer the question of whether restarting warfarin after GI bleeding was safe in these patients.

Through a search of MEDLINE, EMBASE, and CENTRAL databases, Dr. Chai-Adisaksopha and co-authors identified three observational studies conducted between 1996 and July 2014 (TABLE).

Thromboembolic events occurred more often in patients who stayed off of warfarin, with the resumption of warfarin leading to a significant reduction in thromboembolic events (hazard ratio [HR] = 0.68 [95% CI 0.52-0.88]; p=0.004).

In addition, resuming warfarin therapy was associated with a significant reduction in patient mortality (HR=0.76; 95% CI 0.66-0.88; p<0.001).

The researchers also observed an increase in recurrent GI bleeding among patients who resumed warfarin compared with those who did not, but the finding was not statistically significant (HR=1.20; 95% CI 0.97-1.48]; p=0.10).

“[While] the percentage of patients experiencing recurrent GI bleeding was numerically higher in patients resuming warfarin therapy (10.1% and 5.5%, respectively),” they wrote, “when compared to the much larger relative reductions in thromboembolism and mortality, these findings provide some reassurance that resuming warfarin therapy would be a more appropriate strategy for many patients, especially when the GI bleeding source has been identified and definitively treated.” They added, though, that the non-significant increase “warrants further prospective studies.”

The “considerable” heterogeneity among the three studies was a limitation of the meta-analysis, Dr. Chai-Adisaksopha and co-authors noted, including that the duration of warfarin interruption varied between studies, ranging from a median of one day to 50 days. “We planned to perform an a priori subgroup analysis based on the hypothesis that the source of heterogeneity might be differences in participants and interventions; however, we were not able to retrieve sufficient clinical data from the original articles to confirm this hypothesis.”

The current review was also not able to draw conclusions about the optimal timing of warfarin resumption, which, Dr. Chai-Adisaksopha and co-authors wrote, should be the focus of future studies.



Chai-Adisaksopha C, Hillis C, Monreal M, et al. Thromboembolic events, recurrent bleeding and mortality after resuming anticoagulant following gastrointestinal bleeding. Thromb Haemost. 2015;114:819-25.

TABLE. Characteristics of the Included Studies
Study Design Number of patients Mean age, years Duration of withholding warfarin, days (median; IQR) Follow-up length, months
Resumed warfarin Did not resume warfarin Resumed warfarin Did not resume warfarin Resumed warfarin Did not resume warfarin
Qureshi et al. (2014) Retrospective cohort 653 676 74.8 75.3 50.0(21.0–78.0) N/A 24
Witt et al. (2012) Retrospective cohort 260 182 71.8 77.7 4.0(2.0–9.0) N/A 3
Nieto et al. (2008) Prospective registry 39 15 71.0 69.0 N/A 3 N/A