Anticoagulation patients at risk for upper gastrointestinal (GI) bleeding saw their risk reduced by more than a third when a proton pump inhibitor (PPI) was added to oral anticoagulant therapy, according to a retrospective analysis of Medicare claims.

In addition, the choice of anticoagulant can be important for outcomes, reported Wayne Ray, PhD, of Vanderbilt University Medical Center in Nashville, and colleagues. During 754,389 person-years of therapy, they found that the highest risk of upper GI hemorrhage was associated with rivaroxaban (Xarelto) but not apixaban (Eliquis), dabigatran (Pradaxa), or warfarin (Coumadin).

Ray noted in a Vanderbilt new release that an estimated 1% to 1.5% of patients on oral anticoagulants will experience upper GI bleeds every year, and anticoagulant choice and PPI co-therapy could affect the risk of upper GI tract bleeding, a frequent and potentially serious complication of oral anticoagulant treatment. The study, published online in JAMA, found that adding a PPI reduced overall bleeding across all anticoagulants, for an incidence rate ratio of 0.66 and a risk reduction of 34%.

“For patients starting oral anticoagulant treatment, both PPI co-therapy and anticoagulant choice can materially affect risk of upper gastrointestinal bleeding hospitalization,” Ray told MedPage Today. “These factors are particularly important for patients with elevated gastrointestinal risk and argue for a GI evaluation prior to initiating oral anticoagulant treatment.”

Ray also noted that without an added PPI the incidence of hospitalization for upper GI bleeding was 4% per year, while adding a PPI reduced the rate to 2.8% per year.

The retrospective analysis looked at hospitalizations for upper GI bleeds among Medicare beneficiaries in the period 2011-2015. The researchers identified 1,643,123 patients with 1,713,183 new episodes of oral anticoagulant treatment for conditions such as atrial fibrillation.

The mean age in the cohort was 76.4, and atrial fibrillation accounted for 870,330 person-years (74.9%) of follow-up. Women accounted for 56.1% of person-years of follow-up.

Warfarin was by far the most commonly used anticoagulant, Ray and co-authors found. Overall, 1,058,807 anticoagulation patients had no PPI co-therapy vs 239,672 with co-therapy, for a follow-up of 754,389 person-years without co-therapy.

During no PPI co-therapy, the adjusted incidence of hospitalization for upper GI bleeding (n=119) was 115 per 10,000 person-years (95% CI 112-118).

Comparing PPI co-therapy (264,447 person-years) with no co-therapy, the researchers found the overall risk of upper GI bleeding hospitalizations (n=2,245) was lower with co-therapy, for an incidence rate ratio (IRR) of 0.66 (95% CI 0.62-0.69).

Compared with the results of other anticoagulants without PPI co-therapy, hospitalization per 10,000 person-years was highest for rivaroxaban (n=1278) at 144 (IRR 1.97) and lowest for apixaban (n=279) at 73. “Because rivaroxaban is given as a single daily dose intended to maintain 24-hour therapeutic levels, the relative peak plasma concentrations are higher than those for other oral anticoagulants,” Ray and associates explained.

The corresponding hospitalization incidence for dabigatran (n=629) was 120 per 10,000 person-years and for warfarin (n=4,933), 113 per 10,000 person-years.

The association of anticoagulant choice and PPI co-therapy with hospitalization for upper GI bleeding varied according to patients’ underlying disease, with absolute differences driven by the upper quartile of risk.

“For these patients, the difference in the annual incidence of hospitalization for upper gastrointestinal tract bleeding between the treatment strategies with the lowest and the highest gastrointestinal safety (rivaroxaban treatment without PPI and apixaban treatment with PPI, respectively) was 2.1 hospitalizations per 100 person-years,” the authors wrote. “These findings indicate the potential benefits of a gastrointestinal bleeding risk assessment before initiating anticoagulant treatment.”

Ray noted that oral anticoagulants can be extremely valuable in preventing an ischemic stroke — “which is one thing you really do not want to have happen, and we can’t lose sight of that.”

“On the other hand,” he added, “they have potentially very dangerous side effects. What we’ve done with this study is show that clinicians can focus on a high-risk population and significantly improve care for those patients with the addition of a PPI.”

Important Because of Large Patient Numbers

Asked for his perspective, Suneal K. Agarawal, MD, of Baylor College of Medicine in Houston, who was not involved in the research, said: “This a very important study because it reinforces with large numbers of patients what gastroenterologists have assumed for some time, that PPIs reduce GI bleeding. The COGENT trial showed that a PPI reduced GI bleeding in patients on antiplatelet medication, and now we have data on the role of PPIs in anticoagulant as well as antithrombotic therapy.”

The Medicare-based analysis had several limitations, the authors said, including the potential misclassification of anticoagulant treatment, PPI co-therapy, and non-steroidal anti-inflammatory drug use, since these variables were determined from filled prescriptions, and Medicare restricts reimbursement for many over-the-counter drugs. In addition, confounding could also have been present by unmeasured factors, such as aspirin exposure and Helicobacter pylori infection.

Furthermore, bleeding risk was measured with a disease risk score, an internal measure suitable for risk stratification within the Medicare cohort and not studied in other populations, Ray and co-authors noted. In terms of generalizability, the cohort excluded patients previously hospitalized for GI bleeding or who switched to a different anticoagulant during the study period. In addition, its Medicare population had a greater prevalence of anticoagulant treatment and GI bleeding risk compared with younger populations.