Risk of Thromboembolism, Recurrent Hemorrhage, and Death After Warfarin Therapy Interruption for Gastrointestinal Tract Bleeding

We read the study by Dr. Witt and colleagues with great interest. In this retrospective cohort study, resumption of warfarin therapy after gastrointestinal bleeding was associated with 95% lower risk of thrombosis and 69% lower risk of death within 90 days after gastrointestinal bleeding. This occurred at the expense of a statistically non-significant increase in the risk of recurrent gastrointestinal bleeding. 

Interestingly, we have, independent of Dr Witt and colleagues, investigated the risk of death and acute cardiovascular events among aspirin users complicated by peptic ulcer bleeding. We performed a retrospective chart review of all patients who were hospitalized for peptic ulcer bleeding at Karolinska University Hospital in Sweden between 2007 and 2010. Among 118 patients identified, 47 (40%) discontinued aspirin therapy at discharge (1). We observed an almost 7-fold (hazard ratio, 6.9; 95% confidence interval 1.4 – 34.8) increased risk of death and acute cardiovascular events within 6 months of hospital discharge among patients who discontinued aspirin therapy compared to patients who resumed this therapy at discharge. This relation was only evident among patients with cardiovascular disease. Recurrent peptic ulcer bleeding was found among seven patients, four of who had discontinued aspirin therapy. Due to the small numbers, no statistical analyses could be performed to clarify the association between aspirin continuation and the risk of recurrent peptic ulcer bleeding. 

To conclude, there is now increasing evidence that antithrombotic therapy should be continued after an episode of peptic ulcer bleeding.

1.  Derogar M, Sandblom G, Lundell L, et al. Discontinuation of Low-Dose Aspirin Therapy After Peptic Ulcer Bleeding Increases Risk of Death and Acute Cardiovascular Events. Clin Gastroenterol Hepatol. Sep 10 2012.

The work by Witt and colleagues will be valuable to us in our decisions regarding whether and when to restart warfarin after a patient on warfarin suffers a GIB. However, we believe that the authors and the invited commentators (Brotmann and Jaffer) should modify their statements about one aspect of their presentation of results and discussion which plays an important part in clinicians' decisions about restarting. 

The authors write, "Compared with those who did not resume warfarin therapy, a numerically higher proportion of patients resuming warfarin therapy had recurrent GIB, but this difference was not statistically significant (10.0% [resumed warfarin therapy] vs 5.5% [did not resume warfarin therapy]; p = .09)." Commentators Brotman and Jaffer stated that there was, "only a modest and nonsignificant increased risk of recurrent bleeding than those in whom warfarin was witheld." With 260 patients resuming therapy and 182 patients not resuming therapy, and with the percentages of rebleed as given, the study had only the power to detect a real difference in rebleed rates between the two groups of 0.324 (with continuity correction) (1,2). Each group would have needed 600 subjects to reach the generally accepted power of 0.80 to detect a true difference or be able with reasonable confidence (80%) to state that that there was no difference in the rates of rebleed (1). Readers of the article and commentary who do not appreciate this major limitation behind the statement that the rebleed "difference was not significant" may make clinical decisions with the expectation that rebleed rates between the two groups have been shown to be essentially equal. Based on the small numbers in the study for this particular measure, such a conclusion is premature. Indeed, if the difference in rebleed rates persisted as found, it would become statistically significant with even a two-fold increase in study subjects (1). 

(1) Glantz SA. Primer of Biostatistics, 5th ed, 2002. McGraw-Hill, New York

(2) Hulley SB, Cummings SR, Browner WS, et. al. Designing Clinical Research, 3rd ed, 2007. Lippincott, Williams and Wilkins, Philadelphia