Gastrointestinal Injury Caused by Aspirin or Clopidogrel Monotherapy Versus Dual Antiplatelet Therapy

John A. Bittl, MD; Loren Laine, MD


J Am Coll Cardiol. 2022;79(2):129-131. 

Patients who undergo percutaneous coronary intervention (PCI) with drug-eluting stents require dual antiplatelet therapy (DAPT) for at least 6 months for stable ischemic heart disease (SIHD) or 12 months for acute coronary syndromes (ACS) to prevent stent thrombosis and other cardiovascular events.[1] The rationale for using DAPT is that aspirin and a P2Y12 antagonist (such as clopidogrel, prasugrel, or ticagrelor) separately inhibit the cyclooxygenase-1 (COX-1) and the adenosine diphosphate–dependent pathways, and together confer greater protection against platelet activation and stent thrombosis than either antagonist alone. Unfortunately, DAPT approximately doubles the risk of gastrointestinal (GI) bleeding compared with single antiplatelet therapy (SAPT).[2] Mounting evidence suggests that a short course of DAPT followed by SAPT reduces bleeding without increasing ischemic events in patients with either SIHD or ACS,[3] but it is uncertain whether using either aspirin or a P2Y12 antagonist alone reduces GI injury.

To compare the separate and combined GI effects of aspirin and clopidogrel, Han et al[4] performed serial capsule endoscopy (CE) after treatment with 3 different antiplatelet regimens in patients who underwent PCI predominantly for ACS. In the trial reported in this issue of the Journal,[4] 505 patients who had no GI ulcerations or bleeding detected by CE after 6 months of DAPT were randomly assigned to either SAPT using low-dose aspirin alone (n = 168) or clopidogrel alone (n = 169) or to DAPT using low-dose aspirin plus clopidogrel (n = 168) for an additional 6 months. At the second CE, the incidence of the primary endpoint of GI mucosal injury (erosions, ulceration, or bleeding) was lower after using SAPT than after using DAPT (94.3% vs 99.2%; P = 0.02),[4] a finding that was driven by the high incidence of erosions. In the study,[4] almost every patient had erosions in the stomach or small intestine on every CE study (especially the last one), and 83% of patients had some evidence of GI injury at baseline despite the fact that 39% of patients were not on antiplatelet therapy before admission but were presumably on DAPT for the 30–120 hours from PCI to the time of CE. When the analysis focused on the subgroup of 68 patients without erosions, ulcers, or bleeding on the prerandomization CE, the investigators found that SAPT caused less GI mucosal injury (68.1% vs 95.2%; P = 0.006), including fewer new ulcers (8.5% vs 38.1%; P = 0.009), than did DAPT.[4]

The investigators[4] found that the secondary endpoint of any type of GI bleeding between 6 and 12 months was less with SAPT compared with DAPT (0.6% vs 5.4%; P = 0.001), without an increase in ischemic events (0% vs 0%). Moreover, overt GI bleeding was 90% lower with SAPT than with DAPT (incidence of 0.3% with SAPT [aspirin 0% vs clopidogrel 0.6%] vs 3.0% with DAPT; relative risk [RR]: 0.10; 95% confidence interval [CI]: 0.01–0.85). The other secondary outcome of clinically overt bleeding at any site, which was primarily Bleeding Academic Research Consortium 1, was also lower with SAPT than with DAPT (5.9% vs 11.9%; RR: 0.50; 95% CI: 0.28–0.90), which was similar to the treatment effect reported in most contemporary trials evaluating SAPT after a short course of DAPT.[3]

Han et al[4] should be commended for performing an excellent randomized trial in patients undergoing PCI. The study affirms that using aspirin or clopidogrel monotherapy after a 6-month course of DAPT in patients undergoing PCI predominantly for ACS has a better safety record than using a 12-month course of DAPT. This finding is in accord with a recent meta-analysis showing that using aspirin monotherapy (hazard ratio [HR]: 0.61; 95% CI: 0.30–1.36) or P2Y12 inhibitor monotherapy (HR: 0.63; 95% CI: 0.45–0.86) was associated with similar reductions in bleeding (Pinteraction = 0.95) and no increase in ischemic events compared with using a longer course of DAPT.[3] Although early trial evidence suggested that aspirin had greater GI toxicity than clopidogrel, such a finding has lately come into question. In the CAPRIE (Clopidogrel Versus Aspirin in Patients at Risk of Ischaemic Events) trial of 19,185 patients with vascular disease,[5] the risk of GI bleeding (0.72% vs 0.52%; P < 0.05) was statistically higher in patients randomized to aspirin. However, the number needed to harm (NNH) of 500 and the 325 mg dose of aspirin used in the trial, which is higher than the 81 mg dose used now, suggest that the bleeding differences are no longer relevant to current practice.[5] A recent meta-analysis comparing the safety of aspirin with a P2Y12 inhibitor found higher rates of GI bleeding with aspirin (1.56% vs 1.22%; odds ratio [OR]: 1.69; 95% CI: 1.12–2.56),[6] but this corresponds to an NNH of 152, again suggesting that differences between aspirin and P2Y12 inhibitors are tiny.

Comparing the GI mucosal effects of aspirin monotherapy with clopidogrel monotherapy, the investigators in the present study[4] found no difference in bleeding (3.6% vs 8.3%; RR: 0.43; 95% CI: 0.17–1.10) or ulceration in either the stomach (6.8% vs 6.1%) or small intestine (7.6% vs 10.6%), a finding that challenges the conventional wisdom that aspirin is the culprit in DAPT responsible for GI injury and raises the possibility that clopidogrel, which does not inhibit COX-1 but targets the P2Y12 receptor, may be ulcerogenic or impair the healing of mucosal injury due to other causes (eg, nonsteroidal anti-inflammatory drugs [NSAIDs] or Helicobacter pylori). Although high-dose aspirin used for pain behaves similarly to NSAIDs, low-dose aspirin incurs a lower incidence of ulcers, at least in part because although low-dose aspirin completely inhibits platelet COX-1, it is a weaker inhibitor of COX-2.[7] Furthermore, in a prospective endoscopic study of patients taking placebo, asymptomatic gastroduodenal ulcers spontaneously formed at a rate of up to 2% per month.[7] In such a setting, any antiplatelet agent can convert asymptomatic ulcers into bleeding ulcers.

The results of the study by Han et al[4] should compel cardiologists to start thinking like gastroenterologists. It is known that a history of GI bleeding in a patient undergoing PCI is the strongest risk factor for bleeding after hospital discharge, which in itself is a stronger predictor of mortality than is myocardial infarction.[8] Other predictors of bleeding include older age, lower baseline hemoglobin, and use of chronic oral anticoagulation therapy.[8] Patients at high bleeding risk undergoing PCI should thus be considered for SAPT after 1–3 months of DAPT after PCI, as long as they remain at low risk of ischemic events.[9] They should also receive proton pump inhibitors.[10]

A finding in the present study[4] that has investigational implications was the discovery of GI erosions in nearly all patients. It is possible that some erosions were due to other causes or constitute a finding that is too sensitive, too nonspecific, and prone to being over-read. Their nearly ubiquitous appearance suggests that erosions as a marker of GI injury are not really useful as a clinical endpoint or a discriminator of safety. This supports the clinical impression that patients and doctors should be more concerned with symptomatic ulceration or overt bleeding than with incidental endoscopic erosions.

The present study,[4] which explored the mechanisms of mucosal injury and GI bleeding with antiplatelet agents, demonstrates that investigations intersecting cardiology and gastroenterology can improve overall care. The results of the study[4] suggest that the antiplatelet effect of SAPT or DAPT is likely of greater importance as a cause of GI bleeding than the mucosal effects of low-dose aspirin. The study should continue to motivate the collaboration between cardiologists and gastroenterologist in the search for safer agents and regimens in patients who have GI bleeding and require antiplatelet agents after undergoing PCI.