Patients with normal platelet counts who have a GI bleed while on antiplatelets were almost six times more likely to die in the hospital if they had a platelet transfusion in a retrospective cohort study from the Yale University in New Haven, Conn.
Ten of the 14 deaths in the 204 transfused patients – versus none of the 3 deaths in the 204 nontransfused patients - were due to bleeding, so it’s possible that the mortality difference was simply because patients with worse bleeding were more likely to get transfused. “On the other hand, the adjusted [odds ratios] for mortality (4.5-6.8 with different sensitivity analyses) [were] large, increasing the likelihood of a cause-and-effect relationship,” said investigators led by gastroenterologist Liam Zakko, MD, now at the Mayo Clinic in Rochester, Minn. ().
Current guidelines suggest platelet transfusions are an option for antiplatelet patients with serious GI bleeds, but the Yale team found that they did not reduce rebleeding. “The observation of increased mortality without documentation of clinical benefit suggests a very cautious approach to the use of platelet transfusion. ... We do not support the use of platelet transfusions in patients with GI [bleeds] who are taking antiplatelet agents,” the investigators wrote.
Subjects in the two groups were matched for sex, age, and GI bleed location, and all had platelet counts above 100 × 109/L. Almost everyone was on aspirin for cardiovascular protection, and 30% were on also on clopidogrel.
Just over half in both groups had upper GI bleeds, and about 40% in each group had colonic bleeds. Transfused patients had more-severe bleeding, with overall lower blood pressure and lower hemoglobin; a larger proportion was admitted to the ICU.
On univariate analyses, platelet patients had more cardiovascular events (23% vs. 13%) while in the hospital. They were also more likely to stay in the hospital for more than 4 days (47% vs. 33%) and more likely to die while there (7% vs. 1%). On multivariable analysis, only the greater risk for death during admission remained statistically significant (odds ratio, 5.57; 95% confidence interval, 1.52-27.1). The adjusted odds ratio for recurrent bleeding was not significant.
Four patients in the platelet group died from cardiovascular causes. One patient in the control group had a fatal cardiovascular event.
Although counterintuitive, the authors said that it’s possible that platelet transfusions might actually increase the risk of severe and fatal GI bleeding. “Mechanisms by which platelet transfusion would increase mortality or [GI bleeding]–related mortality are not clear,” but “platelet transfusions are reported to be proinflammatory and alter recipient immunity,” they said.
At least for now, “the most prudent way to manage patients on antiplatelet agents with [GI bleeding] is to follow current evidence-based recommendations,” including early endoscopy, endoscopic hemostatic therapy for high-risk lesions, and intensive proton pump inhibitor therapy in patients with ulcers and high-risk endoscopic features.
“Although not based on high-quality evidence, we believe that hemostatic techniques that do not cause significant tissue damage (e.g., clips rather than thermal devices or sclerosants) should be used in patients on antiplatelet agents, especially if patients are expected to remain on these agents in the future,” they said.
The mean age in the study was 74 years, and about two-thirds of the subjects were men.
The authors had no disclosures.