Fewer Gastrointestinal Bleeds With Ticagrelor and Prasugrel Compared With Clopidogrel in Patients With Acute Coronary Syndrome Following Percutaneous Coronary Intervention

Neena S. Abraham; Eric H. Yang; Peter A. Noseworthy; Jonathan Inselman; Xiaoxi Yao; Jeph Herrin; Lindsey R. Sangaralingham; Che Ngufor; Nilay D. Shah


Aliment Pharmacol Ther. 2020;52(4):646-654. 

In This Article


In this study of 37 019 real-world ACS patients who underwent PCI, we demonstrate fewer GIB events in patients treated with ticagrelor, when compared with clopidogrel in all subgroups (overall, NSTE-ACS and STEMI). Prasugrel was also associated with fewer GIB events overall and in the STEMI group, but not in the NSTE-ACS group. There was no observed difference in GIB events with ticagrelor vs prasugrel in all three subgroups (Figure 2). With all three antiplatelet agents, upper GIB was more frequent than lower GIB. The clinical severity of the bleeding event is demonstrated by inpatient admissions that exceeded 5 days regardless of exposure. The latter is not surprising given the significant co-morbidity burden of ACS patients. Interestingly, ticagrelor was associated with fewer transfusions when compared to clopidogrel, but no statistical significant difference was observed in transfusion rates between clopidogrel and prasugrel.

Data from prior randomised clinical trials of ticagrelor and prasugrel vs clopidogrel have shown variable GIB risks in the post-ACS population. Ticagrelor is associated with a nonsignificant trend towards increased GIB when compared with clopidogrel (relative risk [RR] 1.23; 95% CI: 0.93–1.64) in the Platelet Inhibition and Patient Outcome Trial (PLATO).[18] Prasugrel had a significantly higher risk of GIB (RR 1.31; 95% CI: 1.07–1.61) compared to clopidogrel in the Trial to Assess Improvement with Prasugrel-Thrombolysis in Myocardial Infarction 38 Trial (TRITON-TIMI 38).[19] A meta-analysis of all randomised controlled trials of ticagrelor and prasugrel (including trials for peripheral artery disease and medical management of ACS) demonstrated no significant increase in the risk of GIB with ticagrelor and an increased risk of GIB with prasugrel (RR = 1.28; 95% CI: 1.13–1.46).[20] However, the Intracoronary Stenting and Antithrombotic Regimen: Rapid Early Action for Coronary Treatment (ISAR-REACT) 5 trial[21] comparing prasugrel to ticagrelor showed a similar rate of major bleeding for patients treated with prasugrel compared to ticagrelor; but the use of a lower dose of prasugrel (5 mg vs 10 mg) in patients over the age of 75 or those weighing less than 60 kg likely contributed to the observed bleeding rates.[21]

Prior real-world studies have investigated the overall risk of bleeding with ticagrelor or prasugrel compared to clopidogrel but fail to clarify the risk of GIB. In a large study of 45 073 Swedish patients with ACS, an increased risk for readmission with bleeding is seen with ticagrelor when compared to clopidogrel (5.5% vs 5.2% HR 1.2 [1.04–1.40].[5] A Swiss registry of 7621 ACS patients showed that prasugrel was associated with an increased risk of bleeding compared to clopidogrel (4.1% vs. 2.8%, P = 0.024).[6] However, neither of these studies described the rates of GIB.

Our study provides real-world evidence to specifically investigate the risk of GIB in ACS patients following PCI and treatment with P2Y12 inhibitors. These data are complementary to the efficacy data obtained from RCTs and provides valuable outcomes in an unselected patient population (patients of all ages and with concomitant prescription of NSAIDs, ASA or anticoagulants) that would normally be excluded from an RCT. Ticagrelor was associated with a lower risk of GIB compared to clopidogrel and prasugrel was associated with lower rates of GIB in the STEMI subgroup.

Data from a large retrospective multicentre observational study of 19 913 ACS patients in the US between 2010 and 2013[22] may explain our observation of fewer adverse events associated with prasugrel (when compared with clopidogrel) in the STEMI population. In this large study, STEMI patients were more likely to receive prasugrel (vs clopidogrel) at discharge when compared to NSTE-ACS patients. The observed attenuated cardiac adverse events with prasugrel (when compared with clopidogrel) were attributed to preferential use of prasugrel in lower-risk patients.[22] A large French observational study further highlighted the narrow prescribing window recommended for a prasugrel prescription (younger, higher body mass index and less frequent stroke history).[23] In this latter study, the baseline characteristics differed between STEMI patients prescribed prasugrel vs clopidogrel, and this preferential prescription to 'lower risk patients' was likely the factor contributing to fewer cardiac adverse events.[23]

Before adjustment, STEMI patients prescribed prasugrel had fewer co-morbidities and were younger when compared with clopidogrel patients (data not shown). However, after the IPTW, baseline characteristics were equally distributed among exposure groups. It is possible that despite adjustment residual unknown confounding related to clinical heuristic may have contributed to this finding.

Taken together our results are compelling because, despite significant reductions in major adverse cardiac events in ACS patients undergoing PCI, many clinicians remain reticent to use ticagrelor due to concerns for increased risk of bleeding.[5,6] Our data suggest that limiting ticagrelor due to concerns of GIB may be unnecessary.

Strengths and Weaknesses

One must interpret our study within the framework of its observational study design. First, the associations between antiplatelet drug exposure and the outcome of interest may not be causal. Second, with the observational study design, there exists the potential for unmeasured confounders. However, our sensitivity analysis of falsification endpoints suggests no unmeasured confounders related to patient frailty or co-morbidity that could modify GIB risk. Third, the measurement of ASA exposure is limited to prescription drug claims and does not include potential over-the-counter (OTC) ASA use. Thus, measured ASA use in this population is likely under-representative of actual ASA use as part of dual antiplatelet therapy. However, we know from published registry data[24,25] that ASA use exceeds 95%-98% in patients in the first year following coronary reperfusion therapy, consistent with published cardiac guidelines.[1] Since we ascertained GIB outcomes within the first year of index prescription we assume that most (if not all) of our patients are exposed to dual antiplatelet therapy (with concomitant prescribed or OTC ASA) and have no reason to believe there exists a differential miss-classification bias of ASA exposure in our cohort that could influence the outcome of interest.

Despite these limitations, this study has important strengths worth considering. Our finding of a reduction in GIB adverse events associated with ticagrelor prescription is novel. Our ability to ascertain outcomes in a large, geographically diverse, national cohort of patients of different backgrounds provides clinically relevant data for physicians as they choose between the three P2Y12 inhibitors following coronary reperfusion therapy. Finally, these data help to clarify the risk reduction associated with the prescription of ticagrelor that has previously been difficult to ascertain in a real-world population.

Implications for Healthcare Professionals and Patients

In RCTs, ticagrelor and prasugrel use in ACS patients undergoing PCI reduce major adverse cardiac events compared to clopidogrel. However, concerns for increased bleeding have resulted in limited use of ticagrelor and prasugrel.[5,6] GIB is the most common cause of bleeding in post-PCI patients treated with DAPT. The results of the current study show that in a real-world population of ACS patients treated with PCI, ticagrelor appears to have a reduced rate of GIB compared to clopidogrel in both STEMI and NSTE-ACS patients. Prasugrel was found to have lower rates of GIB compared to clopidogrel in the select subgroup of STEMI patients.