1

Introduction

Patients who undergo percutaneous coronary intervention (PCI) are less likely to experience cardiovascular death, myocardial infarction, or ischemic events necessitating revascularization when they receive a platelet adenosine diphosphate (ADP) P2Y12 receptor inhibitor in addition to aspirin before the procedure . Dual anti-platelet therapy remains a cornerstone in pharmacotherapy for PCI and is critical in prevention of ischemic events. Ticlopidine was the first P2Y12 inhibitor to be used in conjunction with aspirin to reduce stent thrombosis after PCI , but its use has decreased dramatically due to safety concerns and the emergence of clopidogrel as a superior alternative . Newer oral P2Y12 inhibitors prasugrel and ticagrelor have demonstrated greater reductions in ischemic events after PCI compared to clopidogrel. Cangrelor is an intravenously administered P2Y12 inhibitor that rapidly provides potent inhibition of platelet ADP-mediated activation and aggregation with a very short half-life, allowing for rapid onset and offset of platelet inhibition . This reversibility is especially desirable when performing angiography on emergent acute coronary syndrome (ACS) patients with unknown coronary anatomy who may require surgical revascularization. Additionally, the intravenous delivery of cangrelor offers other advantages compared to oral P2Y12 inhibitors, especially in patients with emesis or inability to take oral medications. However, cangrelor is expensive and requires that clopidogrel or prasugrel loading dose be administered after discontinuing the cangrelor infusion. This raises the possibility of developing ischemic events in the period of time after discontinuing the cangrelor infusion and prior to achieving adequate platelet inhibition with oral P2Y12 inhibitors. Importantly, ticagrelor can be administered concomitantly with cangrelor and helps minimize this risk. Importantly, the safety and efficacy of cangrelor have not been directly compared to either prasugrel or ticagrelor in randomized controlled trials (RCTs). Therefore, we performed a network meta-analysis to compare the clinical outcomes of patients receiving clopidogrel, prasugrel, ticagrelor, and cangrelor prior to or during PCI.

2

Methods

The primary objective of this network meta-analysis was to compare the effects of P2Y12 inhibitors clopidogrel, prasugrel, ticagrelor, and cangrelor on clinical outcomes including death, cardiovascular death, myocardial infarction (MI), major adverse cardiac events (MACE), stroke, probable/definite stent thrombosis, major bleeding, and minor bleeding after PCI. Two authors (P.C. W and M.J.L.) independently and systematically searched (Feb 2016) Medline/PubMed and ClinicalTrials.gov , applying the search terms ‘clopidogrel’ or ‘prasugrel’ or ‘ticagrelor’ or ‘cangrelor’. We included only: (i) patients from RCTs comparing at least two of the aforementioned P2Y12 inhibitors (ii) in whom PCI was planned or performed emergently and (iii) studies that provided clinical outcome data that could be abstracted. We excluded studies that utilized a crossover design, were duplication from other studies, or were available only in abstract form. Data were abstracted and verified by P.C.W. and M.J.L. in accordance with PRISMA guidelines for network meta-analysis . Study definitions of MACE were preferentially accepted, and composites of ischemic cardiac events were preferred over composite endpoints that included cerebrovascular events. We preferentially abstracted 30 day outcomes when available. Additionally, we preferentially abstracted data from patients who underwent PCI, when possible. For example, instead of abstracting data from the overall PLATO cohort , we chose to abstract data from only the cohort which underwent PCI . We accepted the study definitions of all outcomes, and all studies had event adjudication by physicians or independent clinical events committees. The Cochrane Collaboration tool for assessing risk of bias was used to assess for different forms of bias within the included studies.

Dichotomous variables were reported as percentages whereas continuous variables were reported as mean ± standard deviation (SD) or median (interquartile range). Network meta-analysis with a Bayesian framework was used to compare outcome frequency between different treatments. Mixed treatment comparison model generation was performed to directly and indirectly compare the frequency of dichotomous outcomes experienced with the different P2Y12 inhibitors using GeMTC 0.14.3 software (GeMTC, http://drugis.org/software/r-packages/gemtc ). Bayesian hierarchical random-effects models with directed acyclic graph models for general-purpose Markov chain Monte Carlo analysis were created with 50,000 tuning iterations and 100,000 simulation iterations. Results are presented as odds ratios (OR) [credible intervals (CI)]. Convergence was appraised graphically according to Gelman and Rubin. Data from a consistency model are presented and direction of findings was confirmed with an inconsistency model to serve as a sensitivity analysis. Further sensitivity analyses were performed by removing one study at a time to confirm directionality of findings. We also performed sensitivity analyses limiting studies to those patients with ST elevation myocardial infarction (STEMI) and non-ST elevation acute coronary syndrome (NSTE-ACS).

2

Methods

The primary objective of this network meta-analysis was to compare the effects of P2Y12 inhibitors clopidogrel, prasugrel, ticagrelor, and cangrelor on clinical outcomes including death, cardiovascular death, myocardial infarction (MI), major adverse cardiac events (MACE), stroke, probable/definite stent thrombosis, major bleeding, and minor bleeding after PCI. Two authors (P.C. W and M.J.L.) independently and systematically searched (Feb 2016) Medline/PubMed and ClinicalTrials.gov , applying the search terms ‘clopidogrel’ or ‘prasugrel’ or ‘ticagrelor’ or ‘cangrelor’. We included only: (i) patients from RCTs comparing at least two of the aforementioned P2Y12 inhibitors (ii) in whom PCI was planned or performed emergently and (iii) studies that provided clinical outcome data that could be abstracted. We excluded studies that utilized a crossover design, were duplication from other studies, or were available only in abstract form. Data were abstracted and verified by P.C.W. and M.J.L. in accordance with PRISMA guidelines for network meta-analysis . Study definitions of MACE were preferentially accepted, and composites of ischemic cardiac events were preferred over composite endpoints that included cerebrovascular events. We preferentially abstracted 30 day outcomes when available. Additionally, we preferentially abstracted data from patients who underwent PCI, when possible. For example, instead of abstracting data from the overall PLATO cohort , we chose to abstract data from only the cohort which underwent PCI . We accepted the study definitions of all outcomes, and all studies had event adjudication by physicians or independent clinical events committees. The Cochrane Collaboration tool for assessing risk of bias was used to assess for different forms of bias within the included studies.

Dichotomous variables were reported as percentages whereas continuous variables were reported as mean ± standard deviation (SD) or median (interquartile range). Network meta-analysis with a Bayesian framework was used to compare outcome frequency between different treatments. Mixed treatment comparison model generation was performed to directly and indirectly compare the frequency of dichotomous outcomes experienced with the different P2Y12 inhibitors using GeMTC 0.14.3 software (GeMTC, http://drugis.org/software/r-packages/gemtc ). Bayesian hierarchical random-effects models with directed acyclic graph models for general-purpose Markov chain Monte Carlo analysis were created with 50,000 tuning iterations and 100,000 simulation iterations. Results are presented as odds ratios (OR) [credible intervals (CI)]. Convergence was appraised graphically according to Gelman and Rubin. Data from a consistency model are presented and direction of findings was confirmed with an inconsistency model to serve as a sensitivity analysis. Further sensitivity analyses were performed by removing one study at a time to confirm directionality of findings. We also performed sensitivity analyses limiting studies to those patients with ST elevation myocardial infarction (STEMI) and non-ST elevation acute coronary syndrome (NSTE-ACS).

3

Results

We searched Medline/PubMed and ClinicalTrials.gov using the previously defined criteria and identified 3077 citations. Utilizing our inclusion/exclusion criteria, we identified 108 abstracts, of which we evaluated 60 as full-text publications. We excluded studies due to non-randomization , crossover design , duplication of data , genotyping , no PCI , no outcomes , allowing switching of therapy , randomization after PCI , subtherapeutic dosing , and publication of rationale/design . However, in order to perform analyses on patients with STEMI and NSTE-ACS, we re-included two sub-studies separately from the main analysis . NSTE-ACS data for TRITON-TIMI 38 were calculated by subtracting the numbers of events at 15 months in the STEMI cohort from the numbers of events at 15 months in overall TRITON-TIMI 38 cohort . Similarly, NSTE-ACS data for PLATO were calculated by subtracting the numbers of events at one year in the STEMI cohort from the numbers of events at one year in the overall PLATO PCI cohort . Our study flow diagram is displayed in Fig. 1 . For our meta-analysis, we included 15 RCTs with 54,025 patients. Of these, 7192 patients were randomized to ticagrelor , 26,903 to clopidogrel , 7455 to prasugrel , and 12,475 to cangrelor . Baseline study characteristics are found in Table 1 . Among the included patients in our analysis, mean age was 63 ± 10, 74% were male, 73% were hypertensive, 20% had a prior MI, 30% had diabetes mellitus, and 45% were smokers. Baseline patient characteristics can be found in Table 2 . As seen in Tables 3 , 29.4% of included patients presented with STEMI, 87.2% presented with ACS, and 92.4% underwent PCI.

Study flow diagram.

Table 3

Patient index hospitalization and procedural characteristics.

Study	Therapies	ACS (%)	STEMI (%)	SA (%)	PCI (%)	GPI use (%)
Angiolillo et al	T v Cl	100	0	nr	100	nr
Bonello et al	P v Cl	100	0	nr	100	17.5
CHAMPION PCI	Ca v Cl	85.0	11.8	15.0	97.4	26.5
CHAMPION PHOENIX	Ca v Cl	43.9	18.2	56.1	98.2	nr
CHAMPION PLATFORM	Ca v Cl	94.7	0	5.3	98.9	9.2
Dasbiswas et al	P v Cl	100	nr	nr	100	nr
DISPERSE-2	T v Cl	100	0	0	42	31
Dridi et al	P v Cl	nr	nr	nr	100	nr
ETAMI	P v Cl	100	92.0	0	87.1	9.7
ExcelsiorLOAD	P v Cl	0	0	nr	100	nr
JUMBO-TIMI 26	P v Cl	40	0	nr	100	69.0
Laine et al	T v P	100	nr	0	100	0
PLATO	T v Cl	99.8	49.0	nr	76.8	35.3
RAPID	T v P	100	100	0	100	0
TRITON-TIMI 38	P v Cl	100	26	0	99	54.5

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Predicting readmission risk following percutaneous coronary intervention at the time of admission
2. Association of remote ischemic peri-conditioning with reduced incidence of clinical heart failure after primary percutaneous coronary intervention
3. Long-term outcome of full plastic jacket treatment for bare metal in-stent restenosis
4. Role of optical coherence tomography in managing patients with ST elevation myocardial infarction

Stay updated, free articles. Join our Telegram channel

Join