乌鸦传媒

The use of warfarin was associated with higher rates of upper gastrointestinal bleeding, but not lower GI bleeding, compared with direct oral anticoagulants (DOACs), a population-based cohort study in Iceland showed.

Among over 7,000 patients, rates of upper GI bleeding were higher in warfarin users compared with DOAC users (1.7 vs 0.8 events per 100 person-years, HR 2.12, 95% CI 1.26-3.59), while rates of lower GI bleeding were similar (1.0 vs 1.6 events per 100 person-years, HR 0.65, 95% CI 0.35-1.19), reported Einar Stefan Bjornsson, MD, PhD, of the University of Iceland in Reykjavik, and colleagues.

Specifically, warfarin was tied to higher rates of upper GI bleeding compared with apixaban (Eliquis; HR 2.63, 95% CI 1.35-5.13), dabigatran (Pradaxa; HR 5.47, 95% CI 1.87-16.05), and rivaroxaban (Xarelto; HR 1.74, 95% CI 1.00-3.05), they noted in .

Although warfarin and DOAC users mostly experienced similar rates of overall and major GI bleeding, warfarin was associated with higher rates of major GI bleeding versus apixaban (2.3 vs 1.5 events per 100-person-years, HR 1.79, 95% CI 1.06-3.05).

"This [study] suggests that DOACs may be preferable to warfarin in patients at high risk of upper GI bleeding. For example, patients with history of peptic ulcer disease or major upper GI bleeding," Bjornsson and team concluded.

While peptic ulcer disease contributed to fewer cases of upper GI bleeding for those on warfarin versus DOACs (18% vs 39%; OR 0.33, 95% CI 0.10-0.96), absolute propensity-weighted incidence rates of peptic ulcer-induced GI bleeding were similar between warfarin and DOAC users (0.3 events per 100 person-years for both; HR 1.18, 95% CI 0.39-3.52).

GI bleeding has been established among warfarin and DOAC users, but prior studies and guidelines have not addressed rates of upper and lower GI bleeding specifically.

"As the drugs spend proportionally much longer time in the lower than the upper GI tract, it is perhaps unsurprising that DOACs are proportionally more likely to cause lower GI bleeding," the authors wrote. "Meanwhile, warfarin's anticoagulative effect is driven by inhibiting synthesis of active vitamin-K dependent factors in the liver and therefore has no local anticoagulant effects."

For this study, Bjornsson and colleagues examined data from electronic medical records at the National University Hospital of Iceland in addition to four regional hospitals on 7,081 patients who were prescribed apixaban (n=2,098), dabigatran (n=474), rivaroxaban (n=3,106), or warfarin (n=1,403) from March 2014 through February 2019.

Those who experienced GI bleeding were older (mean age 75 vs 69); had more comorbidities, such as hypertension (75% vs 59%) and congestive heart failure (17% vs 7%); and were on more concomitant medications, including proton pump inhibitors (57% vs 39%) and statins (48% vs 40%).

Inverse probability weighting by sex, age, history of GI bleeding, Charlson Comorbidity Index, and concomitant drug use, among other factors, yielded balanced groups. Overall, mean age was 67-73, and 53-59% were men.

Risk of upper GI bleeding was higher in men on warfarin versus DOACs (1.8 vs 0.7 events per 100 person-years) and compared with women (1.4 vs 1.0 events per 100 person-years). Rates of lower GI bleeding were similar between men and women.

Over a mean follow-up of 1.1-1.7 years, 295 GI bleeding events occurred; 51% were lower events, 36% were upper events, and 14% had an unknown location. About 70% required hospitalization, and 63% included major bleeding. Five patients died, including three taking warfarin.

Bjornsson and team noted that data on baseline lab values, over-the-counter medication use, and lifestyle and socioeconomic factors were not available, though socioeconomic factors likely did not contribute to bias, since Iceland offers universal healthcare.

Comment