Gastrointestinal bleeding in patients undergoing percutaneous coronary intervention

INTRODUCTION — Periprocedural and long-term antithrombotic (anticoagulant or antiplatelet) therapy is an integral part of percutaneous coronary intervention (with either coronary artery stenting or percutaneous transluminal coronary angioplasty). Antithrombotic therapy has been shown to prevent periprocedural and long-term cardiovascular events such as stent thrombosis and recurrent myocardial infarction. (See "Long-term antiplatelet therapy after coronary artery stenting in stable patients" and "Antithrombotic therapy for elective percutaneous coronary intervention: General use".)

Current antithrombotic regimens have the potential to result in either new gastrointestinal (GI) bleeding or to exacerbate chronic episodic GI bleeding. For those with GI bleeding, there is significant attributable morbidity and mortality. However, in many such patients, alternatives to percutaneous coronary intervention (PCI) may be limited. Given the fact that over one million PCI procedures are performed annually, and that antithrombotic therapy has become more aggressive, there is a need to understand the unique challenges in the prevention and management of GI bleeding in this patient population.

Furthermore, with the increasing adoption of a radial as compared to femoral artery approach, GI bleeding may be the most common type of bleeding associated with PCI.

This topic focuses on periprocedural and long-term GI bleeding in patients undergoing PCI. Issues of GI bleeding in patients on chronic anticoagulant or nonsteroidal antiinflammatory therapy are discussed elsewhere. (See "Coronary artery disease patients requiring combined anticoagulant and antiplatelet therapy" and "NSAIDs (including aspirin): Primary prevention of gastroduodenal toxicity".)

Periprocedural bleeding not related to the GI tract in patients undergoing PCI is discussed separately. (See "Periprocedural bleeding in patients undergoing percutaneous coronary intervention".)

DEFINITION — While strict definitions of peri-procedural bleeding in percutaneous coronary intervention (PCI) patients have been adopted, there is less consensus on what constitutes clinically significant gastrointestinal (GI) bleeding. (See "Periprocedural bleeding in patients undergoing percutaneous coronary intervention".)

Common types of GI bleeding in PCI patients are the occurrence of hematemesis, melena, or red blood per rectum, and endoscopic evidence of the source of GI bleeding. Significant reduction in hemoglobin (≥2 g/dL) or need for transfusion has also been required in some, but not all, studies [1-5].

INCIDENCE AND RISK FACTORS — The combined use of aspirin and a platelet P2Y₁₂ receptor blocker, which is common after percutaneous coronary intervention (PCI) with stenting, is associated with an increased risk of gastrointestinal (GI) bleeding (overall incidence of 0.7 to 2.4 percent at 30 days), which accounts for approximately 15 percent of all periprocedural bleeding events [1-3,5,6]. This increase in early GI bleeding has also been associated with the intensity of antithrombotic therapy, advanced age, primary PCI for ST-elevation myocardial infarction, shock or inotrope requirement, cardiac arrest, baseline anemia, smoking, diabetes, duration of antithrombotic study drug administration, female sex, prior stroke, chronic kidney disease, history of prior bleeding or complicated peptic ulcer disease, and heart failure [2,3,6]. Triple antithrombotic therapy (dual antiplatelet therapy plus an oral anticoagulant) significantly increased the risk of GI bleeding (15.6 percent in patients on triple therapy versus 6.7 in patients not on triple therapy) [7].

The risk factors for GI bleeding in patients on long-term therapy with aspirin are a history of complicated peptic ulcer disease, bleeding (15 percent recurrence of GI bleeding at one year on 100 mg aspirin per day for example), or at least two of the following: age >65 years (1 percent per year increase per decade), concurrent use of a second nonsteroidal antiinflammatory drug (two- to fourfold increase in risk), glucocorticoid therapy, or anticoagulant, high-dose aspirin or nonsteroidal antiinflammatory drug, a history of an uncomplicated ulcer, or Helicobacter pylori (H. pylori) infection (fivefold increase in risk) [8-12]. This issue is discussed in detail separately.

Aspirin — Aspirin (and other nonsteroidal antiinflammatory drugs) cause gastrointestinal bleeding due to their irritation of the gastroduodenal mucosa. However, for patients undergoing PCI with stenting, withholding aspirin is an option only in rare cases. These issues are discussed in detail elsewhere. (See "Antithrombotic therapy for elective percutaneous coronary intervention: General use" and "NSAIDs (including aspirin): Primary prevention of gastroduodenal toxicity".)

Platelet P2Y12 receptor blockers — Clopidogrel and other platelet P2Y₁₂ receptor blockers inhibit platelet activation by binding the platelet adenosine diphosphate (ADP) receptor, leading to inhibition of ADP-dependent activation of the glycoprotein IIb-IIIa complex. While the use of these drugs is associated with an increase in the risk of GI bleeding, the magnitude of the increase has not been well studied. In a large cohort study from Denmark that included over 75,000 patients who took clopidogrel, the long-term risk of GI events (bleeding, ulcer, or gastritis) increased from about 2 to about 6 percent, comparing never-users with users [13]. Among individuals taking no antiplatelet therapy, the crude risk of GI events was 1.6 percent; among non-clopidogrel aspirin users, the risk was 4.1 percent; among clopidogrel non-aspirin users, the risk was 6.1 percent; and among patients on both agents the risk was 6.6 percent.

Platelet P2Y₁₂ receptor blockers are not believed to be directly ulcerogenic. In a Scandinavian study of healthy volunteers with upper endoscopy, clopidogrel (75 mg per day), unlike aspirin (325 mg per day), was not associated with macroscopic changes in gastroduodenal mucosa after eight days of therapy [14]. Thus, some have advocated P2Y₁₂ receptor blocker monotherapy for patients with gastrointestinal intolerance to aspirin. However, clopidogrel and other adenosine diphosphate-receptor antagonists can impair ulcer healing, and have been associated with increased hemorrhagic risk in patients with a prior history of GI bleeding [15,16]. Platelet aggregation leads to the release of platelet-derived growth factors (eg, vascular endothelial growth factor or VEGF) that promote angiogenesis and endothelial cell proliferation required for ulcer healing [17].

The incidence of GI bleeding with the more potent drugs prasugrel and ticagrelor has not been well studied. There is some evidence suggesting that the risk is greater than with clopidogrel [18].

Cangrelor, an intravenous rapid onset/offset P2Y12 inhibitor that has been approved, is associated with increases in the rates of minor but not moderate or severe bleeding in comparison to clopidogrel in patients undergoing PCI [19].

Dual antiplatelet therapy — The incidence, predictors, and outcomes of GI bleeding were evaluated in an observational study of 1852 patients who underwent PCI with drug-eluting stent placement and were taking aspirin and clopidogrel [20]. The rate of hospitalized GI bleeding was 2.7 percent; a history of GI bleeding was the most important independent predictor of future GI bleeding (odds ratio 5.0). The 30-day mortality after GI bleeding was 3.7 compared to 0 percent in a group of comparable individuals who had not bled.

Little or no data exist on periprocedural incidence of GI bleeding, possibly due to the subacute time course required for pathogenesis in most patients.

OUTCOMES — Gastrointestinal (GI) bleeding is associated with increased early and late mortality, as well as ischemic outcomes and a higher incidence of stent thrombosis. The reasons for poorer outcome in patients with GI bleeding are not clear, but bleeding may be either directly causative or be a marker for morbidity and heightened patient risk. (See "Coronary artery stent thrombosis: Incidence and risk factors" and "Periprocedural bleeding in patients undergoing percutaneous coronary intervention", section on 'Outcomes after bleeding'.)

The following are examples of this association:

●A comprehensive evaluation of the prognostic impact of GI bleeding was performed in the ACUITY trial of 13,819 acute coronary syndrome (ACS) patients (7789 patients with percutaneous coronary intervention [PCI]) [3]. GI bleeding occurred in 1.3 percent and was a significant independent predictor of outcomes at 30 days with hazard ratios of 4.87 for mortality, 1.74 for nonfatal myocardial infarction, and 1.94 for composite ischemia. The increased risk of each of the end points persisted at one-year follow-up Moreover, the stent thrombosis rate was significantly higher than in those with, as compared to those without, GI bleeding (5.8 versus 2.4 percent) [3]. In a study examining patients enrolled in the Primary Angioplasty in Myocardial Infarction (PAMI) trials, GI bleeding was associated with a significantly longer hospital stay (12.6 ±13 versus 6.4 ±5.3 days) and a higher incidence of death both in-hospital (10 versus 2.8 percent) and at six-month follow-up (14 versus 4.6) [1].

●In a retrospective study of over 20,000 PCI patients, the 30-day mortality was significantly increased in those with GI bleeding compared to those without (20.5 versus 2.4 percent) [6].

●The CHARISMA trial enrolled patients with either established atherosclerosis or multiple risk factors, and randomized patients to either clopidogrel 75 mg/day or placebo, in addition to aspirin 75 mg to 162 mg/day. GI bleeding was more common in the clopidogrel-treated patients, occurred most frequently in the first year of the study, and was independently and significantly associated with all-cause mortality (1.82, 95% CI 1.24-2.69) [21].

●In a study of 6212 patients in the Bern PCI registry, GI bleeding was associated with increased all-cause mortality (adjusted hazard ratio 3.40, 95% CI 1.67-6.92) [7].

TREATMENT — Gastrointestinal (GI) bleeding in a patient who has recently undergone percutaneous coronary intervention (PCI) poses unique therapeutic challenges. The need to attain hemostasis often requires the premature discontinuation of antithrombotic therapies, which leads to an increased risk of recurrent myocardial ischemic events, including stent thrombosis [22,23]. (See "Long-term antiplatelet therapy after coronary artery stenting in stable patients" and "Long-term antiplatelet therapy after coronary artery stenting in stable patients", section on 'Duration and Type of Antiplatelet Treatment'.)

In addition, acute bleeding itself leads to platelet activation and initiation of the clotting cascade for hemostasis, potentially increasing the risk of thrombosis in the absence of antiplatelet therapy.

In the event of acute GI bleeding, the decision to discontinue aspirin, P2Y₁₂ receptor blockers, or other antithrombotic therapy is made on an individual patient basis, balancing the likelihood and consequences for either a thrombotic or a hemorrhagic event action [24]. The following should be kept in mind when deciding further management:

●Some (but not all) randomized trials have demonstrated increased risk of ischemic events in PCI patients receiving aspirin monotherapy in comparison with dual antiplatelet therapy (aspirin and P2Y₁₂ receptor blocker) [25-29].

●Efforts should be made to continue antiplatelet therapy unless the bleeding is life threatening, particularly if coronary stenting has been recently performed. For patients who have undergone PCI with drug-eluting stents who experience GI bleeding, it is reasonable to change from dual antiplatelet therapy to monotherapy at three months for stable ischemic heart disease and at six months for acute coronary syndromes.

●For patients with clinically suspected GI bleeding, a gastroenterology evaluation should be pursued in a timely fashion in accordance with the urgency of the clinical presentation, with appropriate endoscopy performed for both confirmation of the diagnosis and deliverance of therapy. It is important to note that bleeding from endoscopy itself can occur, with a risk that varies according to procedure type and whether mucosal biopsies are performed. (See "Management of antiplatelet agents in patients undergoing endoscopic procedures" and "Gastrointestinal endoscopy in patients with disorders of hemostasis" and "Gastrointestinal endoscopy in patients with disorders of hemostasis", section on 'Estimating procedure-related bleeding risk'.)

●In patients for whom it has been determined that antiplatelet therapy absolutely needs to be discontinued, there are no data on optimal timing for reintroduction of these agents or the appropriate doses that should be used. One small, prospective investigation (n = 156) examined reintroduction of low-dose aspirin in patients with cardiovascular disease and GI bleeding. All patients were treated with pantoprazole and then randomly assigned to either aspirin 80 mg daily or placebo after endoscopic therapy. Patients in the aspirin group had a nonsignificant increase in the incidence of recurrent GI bleeding (10.3 versus 5.4 percent; p = 0.25), but demonstrated a significantly lower all-cause mortality (1.3 versus 12.9 percent) [30]. In the placebo group, 5 of the 10 deaths were attributed to cardiovascular causes.

●Blood transfusion may be considered in patients with GI bleeding. However, the benefit of transfusion for periprocedural bleeding remains controversial, with some studies demonstrating the potential for harm. (See "Periprocedural bleeding in patients undergoing percutaneous coronary intervention", section on 'Blood transfusion'.)

●Proton pump inhibitors (PPI) promote ulcer healing and prevent recurrence. Thus, for patients with GI bleeding, we recommend treatment with a PPI. The role of PPI in the management of GI bleeding related to ulcer disease is discussed in detail elsewhere. (See "Peptic ulcer disease: Treatment and secondary prevention" and "Proton pump inhibitors: Overview of use and adverse effects in the treatment of acid related disorders" and "Approach to acute upper gastrointestinal bleeding in adults", section on 'Acid suppression'.)

●The potential interaction between PPI and clopidogrel is discussed elsewhere. (See "Clopidogrel resistance and clopidogrel treatment failure", section on 'Interaction with other drugs'.)

PREVENTION — The primary prevention of gastrointestinal toxicity from nonsteroidal antiinflammatory drugs such as aspirin is presented in detail elsewhere. (See "Stress ulcers in the intensive care unit: Diagnosis, management, and prevention" and "NSAIDs (including aspirin): Primary prevention of gastroduodenal toxicity".)

Proton pump inhibitors — With regard to bleeding after percutaneous coronary intervention (PCI), observational studies have suggested that the use of proton pump inhibitors (PPIs) is associated with a significantly lower incidence of upper gastrointestinal (GI) bleeding within the first 30 days [2]. The COGENT trial tested the hypothesis that PPI therapy reduces the risk of GI bleeding in patients on long-term dual antiplatelet therapy with aspirin and clopidogrel [31]. In this trial, 3761 patients with an acute coronary syndrome or undergoing percutaneous coronary intervention were randomly assigned to omeprazole 20 mg or placebo daily. The following findings were noted with omeprazole therapy:

●The primary GI end point (a composite of overt or occult GI bleeding, symptomatic gastroduodenal ulcers or erosion, obstruction, or perforation) was significantly reduced (1.1 versus 2.9 percent; hazard ratio 0.34, 95% CI 0.18-0.63) at 180 days.

●The rate of overt upper GI bleeding was significantly reduced (hazard ratio 0.13, 95% CI 0.03-0.56).

●There was no negative impact of omeprazole therapy on the primary composite cardiovascular end point in the COGENT trial (hazard ratio 0.99, 95% CI 0.68-1.44), despite some concern from observational studies that PPIs might have an adverse effect on cardiovascular outcomes in patients treated with clopidogrel [32]. (See 'Potential interaction with clopidogrel' below.)

However, because of the small number of cardiovascular events (109), the 95 percent confidence limits for this hazard ratio include a 32 percent decrease in cardiovascular events as well as a 44 percent increase in such events, leaving open the question of a clinically meaningful difference in cardiovascular events due to the use of a PPI. (See "Clopidogrel resistance and clopidogrel treatment failure", section on 'Interaction with other drugs'.)

●In a post-hoc analysis, the reduction in the rate of the primary GI end point with PPI therapy was seen in patients irrespective of whether they were receiving high-dose (150, 162, 300, or 325 mg) or low-dose (75 or 81 mg) aspirin [33].

In terms of secondary prevention, PPIs reduce the risk of recurrent upper gastrointestinal bleeding. (See "Aspirin for the secondary prevention of atherosclerotic cardiovascular disease".)

PPI should be given to the patient with a history of GI bleeding, need for triple antithrombotic therapy, or at high risk of bleeding. For patients in need of triple antithrombotic therapy, treatment should entail low-dose aspirin, clopidogrel (instead of prasugrel or ticagrelor), warfarin with target INR of 2.5, a PPI, and minimization of the duration of dual antiplatelet therapy.

Potential interaction with clopidogrel — Despite early concerns about an interaction between clopidogrel and PPIs, we think evidence does not support such an interaction. Meta-analyses have not clearly identified an increased risk of ischemic events with the use of PPIs in conjunction with clopidogrel. Some studies have cited increased rates of cardiovascular events attributable to PPIs, even without clopidogrel on board [32,34].

Reports in 2008 and 2009 raised concerns that some PPIs interfere with the ability of clopidogrel to inhibit platelet function and thereby increase the likelihood of coronary artery stent thrombosis (ST), myocardial infarction, or cardiovascular death. In addition, there may be genetic factors that influence the risk of drug interactions and the effectiveness of clopidogrel. More data are needed to clarify in which patients the risk of gastrointestinal bleeding outweighs the risk of a cardiovascular event associated with the possible drug interaction. For additional information on drug interactions, use the drug interaction program provided by UpToDate. (See "Clopidogrel resistance and clopidogrel treatment failure", section on 'Summary and recommendations' and "Clopidogrel resistance and clopidogrel treatment failure", section on 'Interaction with other drugs'.)

The COGENT trial, discussed in the section above, gives reassurance of the safety of PPI therapy in combination with clopidogrel.

Recommendations for use of proton pump inhibitors — We recommend gastrointestinal prophylaxis for all patients undergoing PCI who are at high risk of an adverse gastrointestinal event consequent to therapy with aspirin and clopidogrel (see 'Incidence and risk factors' above). In addition, we treat most patients taking prasugrel or ticagrelor with PPIs. (See 'Platelet P2Y12 receptor blockers' above.)

Use of a PPI in patients on aspirin and a P2Y₁₂ receptor blocker not at high risk is also reasonable. PPI should be started as soon as possible after the initiation of dual antiplatelet therapy and continued indefinitely. For these lower-risk patients, factors such as complexity of the medical regimen, cost, or concern for the potential interaction between clopidogrel and PPIs should be taken into account when deciding whether to prescribe PPIs indefinitely.

These recommendations are consistent with those made in the 2010 American College of Cardiology Foundation/American College of Gastroenterology/American Heart Association expert consensus document on the concomitant use of PPI and P2Y₁₂ receptor blockers [35]. The 2013 European Society of Cardiology expert position paper on the use of PPI in patients with cardiovascular disease and antithrombotic therapy makes similar recommendations and states that "PPIs with weaker inhibition of CYP219 are preferred in combination with clopidogrel compared with those with stronger inhibition" [36].

Other drugs — H2 receptor antagonists have not been shown to prevent ulcer bleeding in patients taking dual antiplatelet therapy, and are therefore of unproven benefit as an alternative to proton pump inhibitors. Because of the known ulcerogenic effects of aspirin, the lowest dose for therapeutic efficacy should be used. (See "Antithrombotic therapy for elective percutaneous coronary intervention: General use", section on 'Aspirin' and "Acute ST-elevation myocardial infarction: Antiplatelet therapy", section on 'Aspirin for all patients' and "Long-term antiplatelet therapy after coronary artery stenting in stable patients", section on 'Summary and recommendations'.)

SCREENING FOR LONG-TERM GASTROINTESTINAL BLEEDING — All patients treated with long-term oral antiplatelet therapy should have a complete blood count (CBC) either before or at the time of drug initiation. The finding of baseline anemia in a CBC should lead to testing the stool for occult blood. We do not perform period CBC to screen for GI bleeding unless there are symptoms or signs of bleeding.

SUMMARY AND RECOMMENDATIONS

●Incidence – Gastrointestinal (GI) bleeding occurs in 1.2 to 2.4 percent of percutaneous coronary intervention (PCI) patients. (See 'Incidence and risk factors' above.)

●Outcomes – GI bleeding is associated with increased early and late mortality, including a heightened risk of stent thrombosis. (See 'Outcomes' above.)

●Risk factors – The risk factors for GI bleeding can be identified pre-procedurally. (See 'Incidence and risk factors' above.)

●In patients being treated with aspirin and a P2Y₁₂ receptor blocker (see 'Prevention' above):

•For patients felt to be at high risk of GI bleeding, we recommend GI prophylaxis with a proton pump inhibitor (PPI) (Grade 1B).

•For patients felt not to be at high risk of GI bleeding, we suggest GI prophylaxis with a PPI (Grade 2B).

•Such therapy should be started as soon as possible after the initiation of dual antiplatelet therapy.

●Treatment – In the event of GI bleeding, appropriate management includes timely gastroenterology evaluation, endoscopy, and PPI administration. (See 'Treatment' above.)

Antithrombotic therapy should not be withheld unless the GI bleeding is life threatening. (See 'Treatment' above.)