Discontinuation of P2Y12 antagonists before coronary bypass surgery: is 5 days really required?

This editorial refers to ‘Coronary artery bypass grafting-related bleeding complications in patients treated with ticagrelor or clopidogrel: a nationwide study’†, by E.C. Hannson et al., on page 189. Bleeding in patients requiring urgent coronary artery bypass graft (CABG) surgery and dual antiplatelet therapy (DAPT) is a matter of concern and represents a major predictor of outcome. Prasugrel and ticagrelor are potent P2Y12 antagonists and are superior to clopidogrel in preventing major cardiovascular events in patients with acute coronary syndromes (ACS).1,2 Both agents are also associated with an increased rate of major bleeding.3 In the Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition with Prasugrel–Thrombolysis in Myocardial Infarction 38 (TRITON-TIMI-38), the CABG-related bleeding rate was 4.7-fold higher with prasugrel in clopidogrel-naive patients,1 whereas in the Platelet Inhibition and Patient Outcomes (PLATO) trial, no such increase was found with ticagrelor.2 In the PLATO trial, patients assigned to ticagrelor undergoing CABG within 7 days of the last dose of study medication had a 50% relative reduction in mortality compared with clopidogrel,4 and these deaths were related to cardiovascular, bleeding and infection events.5 The optimal discontinuation time of P2Y12 antagonists before CABG is still poorly defined. Current guidelines recommend a minimum of 5 days for ticagrelor and clopidogrel and a minimum of 7 days for prasugrel.6 This may be difficult to follow in the real world, and few data exist to show whether these delays are safe in terms of ischaemic complications and necessary in terms of a true reduction in relevant bleeding. Patients with non-ST-elevation myocardial ACS (NSTE-ACS) having complex multivessel coronary disease may require acute or urgent CABG within <5 days. It is for these situations in particular that data are lacking to show the true bleeding incidence with shorter discontinuation periods of P2Y12 antagonists. In this issue of the Journal, Hansson et al.7 report data from the SWEDEHEART registry on CABG-related bleeding complications in ACS patients on DAPT with either ticagrelor or clopidogrel. The main question asked was whether a discontinuation time <120 h is associated with increased bleeding and whether differences can be identified between ticagrelor and clopidogrel in real life after adjustment for factors that influence bleeding risk. They included ACS patients (presumably all NSTE-ACS, although not stated) who underwent CABG during 2012–3 in Sweden: 1266 patients on DAPT with ticagrelor and 978 patients with clopidogrel. Major bleeding occurred in 38% of patients with ticagrelor and in 31% of patients with clopidogrel if the drugs were discontinued <24 h before surgery. Within the ticagrelor group, there was no difference in major bleeding when comparing discontinuation times of 72–120 h vs. >120 h [unadjusted odds ratio (OR) 0.93 (95% CI 0.53, 1.64)], but increased bleeding rates in the clopidogrel-treated patients [unadjusted OR 1.71 (95% CI 1.04, 2.79)]. The authors conclude that discontinuation 3 days before surgery, as opposed to 5 days, did not increase the incidence of major bleeding with ticagrelor but increased the risk with clopidogrel. The overall risk of major CABG-related bleeding was lower with ticagrelor than with clopidogrel [adjusted OR 0.72 (95% CI 0.56, 0.92)]. This well-performed study provides important information on real-world experience with ticagrelor in patients with ACS. It is the largest contemporary observational study investigating bleeding complications in ACS patients requiring urgent CABG, using data from one of the most efficient currently active registries. Bleeding is thoroughly documented using three bleeding definitions as well as mediastinal drainage volume. However, some limitations have to be considered. First, as in any observational study, there may be considerable bias due to unmeasured confounders, i.e. pre-treatment characteristics that may vary in the two study groups, most frequently co-morbidities. Indeed, a history of bleeding, renal insufficiency, liver disease and heart failure are known predictors of bleeding,3 not all of which were measured in the current cohort. Adjustment was performed using a regression model rather than propensity score analysis, which may be a disadvantage,8 although there is no consensus as to when propensity score analysis should be used instead of traditional regression models.9 Second, no correction was made for multiple comparisons, which may lead to an overestimation of the number of significant findings. Third, the difference in bleeding between clopidogrel and ticagrelor may simply be due to the faster onset and offset of action of ticagrelor vs. clopidogrel.10 Thus, with proper timing of discontinuation, if at all related to clinical outcome,11 this difference may no longer be apparent. Finally, the practices of cardiac surgeons in controlling haemostasis likely changed over time and may have impacted clinical outcomes. Cardiac surgeons are increasingly confronted with patients who have received ticagrelor within the previous 24–72 h. Meticulous haemostasis and the use of antifibrinolytic agents such as tranexamic acid may limit bleeding in patients on ticagrelor.13 The protective effect of DAPT against ischaemic complications may be increasingly recognized among cardiac surgeons. This led to an 2012 update of the Society of Thoracic Surgeons guidelines suggesting a shortened interruption of DAPT.12 Thus current surgical practice, together with multidisciplinary clinical judgement, may obviate some of the current guideline recommendations. In conclusion, the present study shows that bleeding is not increased if ticagrelor is interrupted 3 days (compared with 5 days) before CABG in patients with ACS in a large nationwide Swedish registry. Because of the limitations discussed, these findings are at best hypothesis generating rather than conclusive and require further confirmation before current recommendations regarding DAPT interruption can be changed. However, pending such data either from randomized controlled trials or more observational studies, these findings are reassuring for those physicians who have to make the decision for acute or urgent CABG in patients on ticagrelor. Open questions in this context include the usefulness of bridging with short-acting P2Y12 antagonists such as cangrelor or elinogrel14 and the role of platelet function testing to guide the appropriate withdrawal period,15 which were not addressed in the present study. I thank Sigrid Sandner, MD, for her advice regarding surgical measures and Michael Edlinger, PhD, for valuable comments on epidemiological issues. Conflict of interest: F.W. has received personal fees from AstraZeneca and Daiichi Sankyo/Eli Lilly.