Pre-treatment with anti-platelet agents, the co-administration of aspirin and a second agent such as P2Y12
or glycoprotein IIb/IIIa inhibitors prior to coronary angiography in patients with acute coronary syndromes without ST elevation, remains a complex and controversial topic in the current era [1
]. The rationale for pre-treatment includes decreasing ischemic events prior to coronary angiography and a possible reduction in the risk of distal embolization and stent thrombosis during PCI by allowing adequate time for greater platelet inhibition to occur prior to balloon-injury [3
]. However, pre-treatment may increase bleeding risk prior to angiography, at the time of coronary angiography during arterial puncture and sheath removal, and during cardiac surgery [3
Withholding a second anti-platelet agent until the coronary anatomy is known may attenuate the risk of bleeding among those ultimately receiving cardiac surgery; however, it does not afford the ischemic protection of these agents prior to coronary angiography and revascularization. Furthermore, withholding pre-treatment prevents patients with an alternative diagnosis being over-treated and exposing them to increased bleeding risks without any potential for ischemic protection.
The net benefit of pre-treatment is therefore a balance between a reduction in ischemic events versus an increase in bleeding complications. This benefit may be modulated by specific patient factors, such as individual ischemic and bleeding risk, as well as institutional factors. The factors include the time delay to coronary angiography, which may increase the risk of ischemic events prior to revascularization, and local propensity for coronary artery bypass grafting with associated bleeding risks.
Many of the randomised clinical trials examining the efficacy of pre-treatment with P2Y12
inhibition are heterogeneous in their study populations, including patients with stable angina [7
], not scheduled for angiography [5
], randomised after angiography [8
], post fibrinolysis [9
], and only with ST-elevation acute coronary syndrome (STE-ACS) [10
]. The ‘Comparison of Prasugrel at the Time of Percutaneous Coronary Intervention (PCI) or as Pre-treatment at the Time of Diagnosis in Patients with Non-ST Elevation Myocardial Infarction’ (ACCOAST) study is the only contemporary randomised trial of the timing of administration of a second antiplatelet agent in patients with non-ST elevation acute coronary syndromes [11
]. The ACCOAST trial observed no ischemic benefit of pre-treatment with prasugrel; however, an increase in major bleeding complications was demonstrated [11
Due to the lack of randomised clinical trial evidence, current European Society of Cardiology (ESC) guidelines are unable to recommend for or against pre-treatment with a P2Y12
inhibitor (in addition to aspirin) for patients with non-ST-elevation acute coronary syndrome (NSTE-ACS) [12
]. Given the evidence gap in this area, we applied a decision-analytic model based on the available evidence from randomised trials to explore the possible clinical and institutional factors that may be associated with greater benefit with a pre-treatment strategy.
2. Experimental Section
2.1. Model Assumptions and Data Sources
We developed a decision-analytic model to explore the impact of pre-treatment with P2Y12
inhibitors administered at the time of initial NSTE-ACS diagnosis compared with administration at the time of PCI in patients treated with a planned invasive management strategy. The model examined the effects of death or myocardial infarction (MI), and major bleeding, as well as net clinical benefit (i.e., combined death or MI and major bleeding events) within 30 days (Figure 1
). Outcomes of the model included the ischemic outcomes of death or recurrent myocardial infarction as a summation of the estimated events prior to angiography, and then for each of the three treatment strategies (i.e., PCI, CABG and medical management) up to 30 days, with bleeding events before and after angiography similarly summated. To estimate the overall benefit of the strategy, we also report the net benefit of pre-treatment with each of the P2Y12
inhibitors, calculated as the absolute reduction in a 30 day ischemic event rate minus the absolute increase in bleeding event rate, (i.e., net clinical benefit (NCB)) between the pre-treatment and non-pre-treatment groups. For the purposes of this analysis, a clinically significant benefit was defined as a ≥1% increase in NCB at 30 days. The clinical relevance of this can be conceptualized as a Number Needed to Treat [NNT] of less than or equal to 100.
2.2. Baseline Assumptions of Patient Risk
Within this model, the baseline of death or MI within 30 days of diagnosis was set at 7% [13
]. Half of this risk was experienced within the first seven days, consistent with early trials of NSTE-ACS where very early angiography (within 48 h) was not commonplace [5
]. The risk of a major bleeding event within 30 days was estimated to be 2.5%, again with half of the risks experienced within the first seven days [13
]. These risks were converted to hourly risks in order to allow separate estimations for event rates before and after angiography.
2.3. Procedural Rates, Timing and Impact on Risk
The median time to angiography was set at 24 h, with the proportion of patients selected for surgical revascularization (CABG) initially set at 10%, and the medical management without PCI set at 30%, with the remainder of the patients undergoing PCI, reflecting rates observed in local registries [13
]. Revascularization was assumed to occur without subsequent delay and any associated ischemic and bleeding events occurring between angiography and revascularization with each strategy was included in the risk associated with that strategy. Among patients undergoing CABG, a relative risk reduction of death or MI by 30 days of 40% but a four-fold increase in major bleeding events compared to medical management was assumed. Similarly, for patients undergoing PCI, a 30% relative risk reduction in death or MI at 30 days, with a 1.25-fold increase in major bleeding, was assumed.
2.4. Modification of Risks Associated with Initiation of P2Y12 Inhibition
The relative effect of clopidogrel pre-treatment versus placebo, ticagrelor versus clopidogrel, and prasugrel versus clopidogrel were drawn from the results of the key large-scale randomized trials or subsequent published subgroup analyses where available [3
]. For base-case analysis, estimated relative effects of death or MI, and major bleeding used the overall randomized trial findings, since these provide the best estimates for treatment-related effects of each anti-platelet agent since they are not subject to the usual limitations of subgroup analyses that apply to post-randomization sub-groups (Table 1
). When examining the effect of prasugrel, we use the results of ACCOAST in the base-case since this was the only trial directly examining this question despite the fact that the TRITON TIMI 38 study was a much larger trial [3
2.5. Sensitivity Analysis
To provide estimates of potential risks and benefits, 10,000 runs of the model were undertaken with bleeding events, ischemic events, and the NCB for each agent was calculated for each run. The proportion of scenarios where the net clinical benefit or net mortality benefit was positive relative to the comparator for each of the literature-based therapy comparisons (clopidogrel versus placebo, prasugrel versus clopidogrel, and ticagrelor versus clopidogrel) was summed and reported as a percentage. These estimates were then used to calculate the relative effect of ticagrelor and prasugrel over placebo and reported similarly. In order to quantify the magnitude of potential benefit, the absolute differences in the overall 30 day rate of ischemic events and bleeding events associated with pre-treatment compared with no pre-treatment were used to calculate the number needed to treat (NNT) for each run of the model.
The treatment effects associated with each of the P2Y12 inhibitors was allowed to vary across the range of the reported 95% confidence bounds for both death or MI and major bleeding. However, due to the differences in reported bleeding events associated with both ticagrelor and prasugrel, the bleeding risk associated with the proportion of patients undergoing CABG were modeled separately using relative hazards that differed from the overall population. For all of these analyses, P2Y12 inhibition was considered not to have any effect on peri-operative ischemic events.
To explore the relationship between varying degrees of baseline ischemic and bleeding risk, time to angiography, institutional rates of CABG and medical management, and treatment effects associated with the specific therapies, multi-way sensitivity analysis was employed. Within each iteration of the model, baseline ischemic risk was allowed to randomly vary between 0% and 15%, while the risk of major bleeding was varied between 0% and 6% with 70% of this variation linked to the ischemic risk given the strong correlation between ischemic and bleeding risks observed among individual patients clinically. Again, with half the 30 day risk experienced within the first seven days, both the bleeding and ischemic risks were converted to an hourly risk.
The time to angiography was allowed to vary randomly between 0 and 96 h, while the proportion of patients deemed appropriate for CABG and medical management were varied between 7% and 15%, and 35%–45%, respectively. However, recognizing that patients with an increased ischemic risk were more likely to undergo CABG rather than PCI, the likelihood of CABG was adjusted by a factor of 1.25 in those with an increased ischemic risk and 0.5 for those models with an ischemic risk of <5%. Therefore, a CABG rate of 10% was increased to 12.5% if ischemic risk was >7%.
The relationship between the likely NCB rates and the estimates of the NNT with varying degrees of: ischemic risk; bleeding risk; time to angiography; and the local surgical revascularization rate were explored. The number needed to harm (NNH) was calculated as 100/absolute event increase %. All models were constructed in STATA 14.0 (Manufacturer, College Station, TX, USA).
Major society guidelines either support the use of pre-treatment in patients with acute coronary syndromes or refrain from advising on pre-treatment due to the lack of support from clinical trial data [12
]. ACCOAST, a contemporary randomized clinical trial assessing prasugrel pre-treatment in NSTE-ACS, was unable to demonstrate a reduction of ischemic outcomes [11
]. We built a decision analysis model for the clinical utility of pre-treatment based on the relative risk reductions from major clinical trials on the currently recommended dual anti-platelet agents [5
]. The likelihood of providing a substantial benefit with pre-treatment (NNT < 100) was not observed unless the baseline ischemic risk exceeded 11% and 14% for ticagrelor and clopidogrel, respectively. It should be noted that these estimates are sensitive to the baseline risk of bleeding, and since clopidogrel was modelled with less impact on bleeding, this allows for a greater tolerance of baseline bleeding risk using this agent. Institutional characteristics of time to angiography were not associated with increased net clinical benefit and institutional CABG rates had only minimal effects on net clinical benefit with pre-treatment using each of the agents. This analysis suggests that a pre-treatment strategy with P2Y12
inhibitors is confined only to patients at relatively high predicted ischemic risk. Consequently, decisions to initiate pre-treatment should be dependent on careful and potentially objective evaluation of ischemic and bleeding risk. This can be done easily at the patient bedside with well-validated risk scores such as the Global Registry of Acute Coronary Events (GRACE) score.
Large-scale randomized studies of intravenous glycoprotein IIb/IIIa inhibition given “upstream” before coronary angiography in NSTE-ACS have failed to demonstrate a significant reduction in ischemic events when compared with these agents used in the peri-procedural setting [17
]. Pre-treatment with P2Y12
inhibition in NSTE-ACS remains controversial because much of the evidence supporting its practice is derived from trials which included patients with stable angina, STE-ACS, and those having received fibrinolysis for STE-ACS [7
]. Data from small observational studies suggest reductions in peri-procedural ischemic events associated with pre-treatment with clopidogrel, with similar incremental benefits observed with large loading doses compared with small loading doses of this agent, but these analyses are confined to those actually receiving PCI [7
]. In contrast, the first randomized comparison of this question, Clopidogrel for the Reduction of Events during Observations (CREDO) study [8
] included patients who were planned for PCI (and therefore included elective patients), and the PerCutaneous Intervention-CLopidogrel as Adjunctive ReperfusIon TherapY (PCI-CLARITY) trial [9
] included only those with ST elevation who had received thrombolysis. A sub-study of the large Clopidogrel in Unstable angina to prevent Recurrent Events (CURE) trial [15
] reported a median time from randomisation to angiography was 10 days, making direct comparisons to the modern era where time to angiography is generally less than 48 h, less reliable. However, the patient population in CURE only included those with acute coronary syndromes without ST elevation, and therefore represented the specific group of interest for this discussion [5
Several meta-analyses attempting to address this question have included these studies, and have reported a benefit of pre-treatment with clopidogrel in PCI patients [20
]. This was demonstrated in the 2012 meta-analysis of 37,814 patients by the Action group, which showed a statistically significant reduction in major cardiac events (9.83% versus 12.35% p
< 0.001) in those pre-treated with clopidogrel [20
]. This study also found no significant difference in mortality with pre-treatment and no significant association between pre-treatment and major bleeding [20
]. However, these studies have focused on the subgroup of patients who actually received PCI, and excluded those who underwent coronary artery bypass grafting or continued medical management. Hence, these meta-analyses focus on the group of patients who are likely to have the most favourable balance between ischemic benefit and bleeding risk. In contrast, a meta-analysis of patients undergoing coronary artery bypass grafting from both observational and randomized studies demonstrates a clear increase in bleeding events with an increase in mortality among these patients [6
]. In a meta-analysis of three randomised trials and 17 observational studies, Biancari et al. demonstrated that pre-operative exposure to clopidogrel was associated with an increased risk of death (RR: 1.3; CI: 1.02–1.67), re-operation for bleeding (RR: 1.88; CI: 1.37–2.58) and need for packed red blood cells (RR: 1.23; CI: 1.10–1.37) [6
ACCOAST is the only other randomized control trial of pre-treatment in ACS patients without ST elevation. This study compared a half load of prasugrel as pre-treatment with a further half dose after the coronary anatomy had been identified by angiography versus 60 mg of prasugrel post-angiography among those undergoing PCI, with mean time from pre-treatment to coronary angiogram times of 4 h [11
]. PCI was performed in 69% of patients, CABG in 6.2%, and medical management in 25%. Pre-treatment did not reduce ischemic outcomes of NSTE-ACS; however, it resulted in increased bleeding events [11
]. As a result of this trial, the ESC guidelines caution against pre-treatment with prasugrel; however, they are unable to advise for or against pre-treatment with other P2Y12
inhibition in the management of NSTE-ACS [12
]. It is worth mentioning that ticagrelor pre-treatment was studied in the randomized control trial, “Administration of Ticagrelor in the Cath Lab or in the Ambulance for New ST Elevation Myocardial Infarction to Open the Coronary Artery” (ATLANTIC) among patients with STE-ACS, and resulted in no reduction in the composite end point of death, myocardial infarction, stroke, urgent revascularization or stent thrombosis, though a very short time difference in dosing times between the two arms (i.e., 31 min) occurred in this study [10
Based on our analysis, pre-treatment with ticagrelor or clopidogrel is anticipated to be of benefit in specific populations of patients presenting with NSTE-ACS, particularly those with a high risk of ischemic events. Pre-treatment with ticagrelor and clopidogrel could result in a greater than 50% probability of achieving a positive 30 day NCB at lower levels of individual bleeding risk combined with higher baseline ischemic risk. This emphasises the need for the clinician to be able to accurately identify a patient’s ischemic risk using well-validated tools such as the GRACE risk scoring system. For example, a GRACE score of approximately 150 translates to an ischemic risk of >11%, representing the threshold for ticagrelor pretreatment [25
]. Hence, while the validity of internationally derived risk scores requires local calibration, this risk-score threshold based decision is analogous to using the CHA2DS2-VASc score for identifying patients at sufficient stroke risk warranting anticoagulation in non-valvular AF. Nevertheless, prospective confirmation of pretreatment thresholds is desirable, and future studies of pretreatment should clearly stratify patients by baseline ischemic and bleeding risk.
Given the greater levels of platelet inhibition, ticagrelor was associated with an increase in NCB at lower thresholds of ischemic risk than observed with clopidogrel. However, among patients with the highest bleeding rates (5%), ticagrelor pre-treatment was only beneficial when the ischemic risk exceeded 18%. Conversely, in the same high risk bleeding population, clopidogrel pre-treatment required an ischemic risk of only 16% in order to obtain a greater than 50% probability of achieving a positive 30 day NCB.
Given the paucity of clinical trial data specifically addressing the question of pre-treatment in patients with NSTE-ACS, this decision-making analysis relies upon assumptions from numerous clinical trials. Firstly, the initial assumptions for clopidogrel are based on the CURE trial, which reflected clinical practice of over a decade ago, before the introduction of radial arterial access, longer delays to angiography, less refined coronary stent technology and less adjunctive anti-ischemic therapies available [5
]. The net effect of the aforementioned differences may in fact be an overestimation of the ischemic benefit of pre-treatment, as suggested in a recent meta-analysis in which a modest effect of major cardiovascular events was noted in older randomized control trials such as CURE, but was not found in more recent studies [5
]. Hence, our analysis may well over estimate the benefits of pre-treatment with clopidogrel and then, consequently, the newer P2Y12
agents. For the purposes of the decision-making analysis, the model assumes therapeutic efficacy at the time of administration of the P2Y12
inhibitor. In the clinical setting, there would be a delay prior to peak therapeutic efficacy with the time dependent on the P2Y12
Secondary, the pre-treatment effects of ticagrelor were modelled on the PLATO trial, which did not formally examine pre-treatment of this agent [4
]. Nevertheless, it offers the best available evidence for the pre-angiography use of ticagrelor compared with clopidogrel in the NSTE-ACS population.
This analysis does not take into account patients with presumed NSTE-ACS who go on to have normal coronaries with an alternative diagnosis such as myocarditis. Given these patients do not proceed to PCI or CABG, the absolute effect on their ischemic benefit or bleeding complication rate is likely to be negligible. Again, as a consequence, actual benefits of pre-treatment in real world practice may be less striking than what is presented in this decision-analytic model.