Antiplatelet Therapy During Gastrointestinal Bleeding: Risk or Benefit?

Article Outline

• Comment
• Copyright

Sung JJY, Lau, JYW, Ching JYL, et al. (Institute of Digestive Disease, Chinese University of Hong Kong, Hong Kong, China). Continuation of low-dose aspirin therapy in peptic ulcer bleeding. Ann Intern Med 2010;152:1–20.

A large number of patients with cardiovascular diseases are on a regimen of low-dose aspirin or other inhibitors of platelet function. Even with low-dose aspirin, the risk of upper gastrointestinal (GI) bleeding increases by 2- to 3-fold (BMJ 1995;310:827–830; BMJ 2006;333:726; Aliment Pharmacol Ther 2006;24:897–908.). Comedication with proton pump inhibitors significantly reduces the risk of bleeding complications of aspirin (Am J Gastroenterol 2007;102:507–515) as well as the rebleeding rate after endoscopic therapy (N Engl J Med 2000;343:310–316). In fact, proton pump inhibitors (PPI) alone showed a similar therapeutic effect in upper GI bleeding as endoscopic treatment (Dig Dis Sci 2007;52:3371–3376). There is some uncertainty how to proceed with the aspirin medication in patients with a clinical need for this therapy when a GI bleed occurs. The paper by Sung et al (Ann Intern Med 2010;152:1–20) studied the impact of high-dose PPI treatment on the continuation of a low-dose aspirin therapy in patients with an acute peptic ulcer bleeding.

In a prospective, double-blind, placebo-controlled design, Sung et al investigated 156 patients in a tertiary center over a 3-year period (2003–2006). Patients had to be on low-dose aspirin treatment (LDASS) for cardiovascular or neurovascular diseases. Patients were only eligible if they (1) showed signs of an upper GI bleeding, (2) had a peptic ulcer showing active bleeding, visible blood vessels, or adherent clots that were successfully treated by endoscopic therapy, and (3) continued to require low-dose aspirin (≤325 mg/d) for prophylaxis or treatment of cardiovascular diseases. All patients received endoscopy within 24 hours using injection therapy and heater probe application. Seventy-eight patients received aspirin, 80 mg/d, and 78 received placebo for 8 weeks immediately after endoscopic therapy. All patients received a 72-hour infusion of pantoprazole (80 mg + 192 mg/d) followed by oral pantoprazole (40 mg/d) and all patients completed follow-up. All 156 patients were included in an intention-to-treat analysis. Three patients withdrew from the trial before finishing follow-up.

The primary end point of the study was recurrent ulcer bleeding within 30 days of the admission; repeat endoscopy with verification of the bleeding source and site was mandatory. Secondary end points of the study were all-cause and specific-cause mortality within 8 weeks of the initial bleeding. Recurrent ulcer bleeding within 30 days was 10.3% in the aspirin group and 5.4% in the placebo group. Patients who received aspirin had lower all-cause mortality rates than patients who received placebo (1.3% vs 12.9%). Patients in the aspirin group had lower mortality rates attributable to cardiovascular, cerebrovascular, or GI complications than patients in the placebo group (1.3% vs 10.3%).

The adjudication committee evaluated 22 cases of suspected recurrent upper GI bleeding. The committee identified 12 cases of confirmed recurrent bleeding: 8 in the aspirin group and 4 in the placebo group. The 30-day cumulative incidence of recurrent ulcer bleeding in the intention-to-treat population was 10.3% in the aspirin group and 5.4% in the placebo group (hazard ratio, 1.9; 95% confidence interval, 0.6–6.0). The site of recurrent bleeding was the same site as in the original bleeding event. Ten patients did not meet the prespecified criteria for recurrent upper GI bleeding (5 patients in the aspirin group and 5 in the placebo group). Among them, 8 patients had no endoscopic evidence of recurrent bleeding and 2 (in the placebo group) did not have endoscopy (1 patient had recurrent hematemesis and died before arriving at the hospital, and 1 patient developed recurrent melena but was too ill to undergo further endoscopic examination). Not explicitly stated and discussed in the manuscript, the Helicobacter pylori incidence in the rebleeding group was numerically lower (3/12; 25%) than in patients with no rebleeding (61/144; 42.4%).

Back to Article Outline

Comment 

With advancing age and more aggressive interventional angiographic therapy the percentage of patients with LDASS for primary or secondary prophylaxis of cardiovascular or neurovascular diseases is constantly increasing. Antiplatelet therapy increases the risk of upper GI bleeding by 2- to 3-fold (BMJ 1995;310:827–830; BMJ 2006;333:726; Aliment Pharmacol Ther 2006;24:897–908).

The primary intention of this study was to prove the concept that high-dose PPI therapy would be effective to prevent rebleeding after an upper GI bleeding event even with continuing LDASS therapy. High-dose PPI medication in patients not taking LDASS reduces the risk of rebleeding within 30 days of endoscopic therapy from 22.5% to 6.7% (N Engl J Med 2000;343:310–316). Patients with recurrent bleeding are usually subjected to repeat endoscopic therapy without increased mortality, when compared with immediate surgery. Sung et al showed that, even with high-dose PPI, the continuation of LDASS increased the risk of rebleeding by about 1.9-fold. Despite the increased rebleeding rate (10.3% vs 5.4%), there was no significant difference with respect to transfusion requirements, length of hospital stay, or bleeding-related mortality. The only proven fatal cases owing to GI bleeding occurred in the placebo group, and the only fatal case in the aspirin group was due to cardiac failure. In addition, 2 further suspected fatal GI bleeding events, which could not be documented by endoscopy occurred in the placebo group. If these 2 patients, who were not included, are added to the patients with recurrent bleeding the difference becomes even less. Another study found no cases of rebleeding in patients with aspirin-associated endoscopic ulcers, who were treated with a PPI and randomly assigned to restart anti-platelet therapy with aspirin or clopidogrel within 1 day of endoscopy (Aliment Pharmacol Ther 2004;19:359–365). However, in this study only patients with a low or medium rebleeding risk were included.

The patients in the study by Sung et al were exclusively on antiplatelet medication with aspirin and the combination therapy of aspirin and clopidogrel, which is a therapeutic standard for prevention of stent thrombosis after percutaneous transluminal coronary angiography with a bare metal stent and, especially, drug-eluting stents (N Engl J Med 2001;345:494–502; JAMA 2002;288:2411–2420; J Am Coll Cardiol 2007;50:e1–e157). With the combination therapy of aspirin plus clopidogrel, the GI bleeding risk is elevated, which is mainly attributed to the aspirin effect (N Engl J Med 2001;345:494–502). In contrast with the PPI effect on bleeding risk with aspirin the risk reduction for clopidogrel in combination with PPI is inconsistent (Clin Ther 2009;31:2038–2047; Am J Gastroenterol 2007;102:507–515).

However, if GI bleeding occurs in a patient requiring LDASS, all clinical aspects have to be considered and the risks of rebleeding and of new cardiovascular and neurovascular complications have to be considered and balanced (N Engl J Med 2005;353:2373–2383). When patients present with peptic ulcer bleeding, the usual protocol according to the international guidelines is to treat the active bleeding with an endoscopic device, offer antisecretory therapy with high-dose PPI, and discontinue aspirin or other antiplatelet agents. However, a restart/reinstitution of the antiplatelet therapy usually within 7 days is recommended (Ann Intern Med 2010;152:101–113).

The most impressive finding of the study by Sung et al is the impact of aspirin discontinuation on overall and cardiovascular mortality. Whereas in the aspirin group the overall and cardiovascular mortality was very low (1/78; 1.3%), there was a significantly increased mortality in the placebo group (10/78; 12.9%) where aspirin was discontinued. To fully understand and critically discuss the impact of this study, it is of major importance to look at the fatal cases in the placebo group. From the 10 fatal cases, only 5 patients died from a cardiovascular disease or complication and 4 of these 5 died within the first 30 days. Three patients died from GI complications (including the only 1 who died from bleeding) and 2 from pneumonia. Thus, half of the fatal cases cannot be directly linked to an effect of aspirin discontinuation. If the 2 patients from placebo group who died with signs of GI bleeding but without endoscopy are added, the mortality difference is increased. Both aspirin and clopidogrel cause an irreversible inhibition of platelet function by an irreversible inhibition of cyclo-oxygenase 1 and 2 or the P2Y₁₂-receptors, respectively. Despite rapid clearance of aspirin from the circulation, the antiplatelet effects of aspirin last for several days (Clin Appl Thromb Hemost 2009,15:523–528), and the effect is terminated by production of new active thrombocytes. Thrombocyte survival time ranges from 8 to 12 days (Paulus JM. Platelet kinetics. Amsterdam: North Holland;1971:317–323), which indicates the time frame in which a therapeutic change of platelet function will be present.

Recent epidemiologic evidence has raised the concern that aspirin withdrawal for treatment noncompliance, surgery, or side effects can carry an increased thrombotic risk in the first days after withdrawal. The delay to the thrombotic event was between 7 and 30 days in most reports and most frequently 7–10 days (Neurology 2004;62:1187–1189; J Intern Med 2005;257:399–414; Clin Appl Thromb Hemost 2009;15:523–528). ASS nonadherence or withdrawal is associated with a 3-fold higher risk for major adverse cardiac events (Eur Heart J 2006;27:2667–2674). The mechanism underlying this effect is poorly understood. However, it is speculated that ultra low doses of ASS might induce a prothrombotic effect and this could be mediated via inhibition of cyclo-oxygenase 2 (Pathophysiol Haemost Thromb 2007;36:40–44; Thromb Haemost 2010;3:171–180).

The guidelines recommend the discontinuation of 5–7 days and to restart therapy as soon as possible (Ann Intern Med 2010;152:101–113). Sung et al also discussed that aspirin can be discontinued for 3–5 days after index bleeding and resumed after stabilization. This strategy is not proven, but tries to minimize the risk of bleeding and of vascular ischemia. However, based on the low mortality in the aspirin group, the strategy to discontinue aspirin needs confirmation and the optimal time to restart antiplatelet therapy has to be determined.