Background
Randomized trials did not consistently support superiority of ticagrelor, as monotherapy or in combination with aspirin, in terms of efficacy or safety, in patients with atherosclerotic artery disease.
Methods
Medline, EMBASE, the Cochrane Central Register of Controlled Trials, and scientific session abstracts were searched for trials of patients with coronary or peripheral artery disease (with >1,000 participants and a follow-up ≥3 months) randomly assigned to ticagrelor-based or conventional antiplatelet therapies. Trial-level hazard ratios (HRs) were pooled using a fixed- or random-effect model (in case of significant heterogeneity) with the inverse variance weighting. The primary outcome was all-cause mortality. Other outcomes were myocardial infarction (MI), stroke, and major bleeding.
Results
Overall 77,489 patients received either ticagrelor-based (n = 38,721) or conventional antiplatelet regimens (n = 38,768) in 6 trials. The primary outcome occurred in 4.5% of patients treated with experimental therapy and 4.9% of patients treated with control therapy (HR = 0.91, 95% CI 0.81-1.01; P = .07). Overall, patients treated with ticagrelor-based versus conventional antiplatelet regimens showed no significant difference in terms of all-cause death, MI, stroke, or major bleeding after 20 months. However, in trials of patients with coronary artery disease as primary diagnosis, the risk for all-cause death (HR = 0.84 [0.77-0.91], P < .001) and MI (HR = 0.87 [0.80-0.94], P = .007) was significantly reduced by experimental therapy.
Conclusions
In patients with atherosclerotic artery disease, the benefit of ticagrelor-based therapies was confined to patients treated for coronary artery disease. The drug significantly reduced the risk for all-cause death and MI without excess risk of bleeding in these patients. In consideration of limitations of subgroup analyses, these results need further validation.
Antiplatelet agents represent the cornerstone of pharmacological prevention of recurrent ischemic events in patients with vascular manifestations of atherosclerosis. Indeed, in patients treated with a percutaneous intervention for stable coronary artery disease (CAD) or acute coronary syndrome (ACS) and not requiring an oral anticoagulation, guideline-writing authorities recommend a strategy of dual antiplatelet therapy (DAPT) for a period of time ranging between 6 and 12 months followed by at least 1 antiplatelet agent lifelong. Similarly, patients with clinically managed or revascularized symptomatic peripheral artery disease (PAD) receive at least 1 antiplatelet agent lifelong or 1-month DAPT as standard of care. The direct-acting, oral cyclo-pentyltriazolo-pyrimidine ticagrelor, which reversibly blocks the adenosine diphosphate receptor P2Y12 on platelets, has become an essential component of antithrombotic therapy for patients with ACS or with stabilized heart attack because it improves clinical outcomes as compared to other thienopyridines (namely, clopidogrel) or aspirin alone. The lower ischemic risk associated with a more potent and/or prolonged platelet inhibition may be offset by an increased risk of bleeding. In this regard, the search for alternative antiplatelet regimens ensuring intensified platelet inhibition while having a wide therapeutic window is a matter of broad clinical interest. Recent randomized clinical trials did not consistently support the superior efficacy and safety of ticagrelor as monotherapy or in combination with aspirin compared with other antiplatelet regimens for secondary prevention of ischemic events in patients with CAD and/or PAD. Against this background, we performed this meta-analysis to investigate the clinical impact of ticagrelor-based versus conventional antiplatelet regimens in this setting.
Methods
Data sources and searches
We searched Medline, EMBASE, the Cochrane Central Register of Controlled Trials, scientific session abstracts, and relevant Web sites ( www.cardiosource.com , www.clinicaltrialresults.org , www.escardio.org , www.tctmd.com , www.theheart.org ) without restricting language or publication status. The references listed in all eligible studies were checked to identify further citations. The last search was performed on 6 January 2019. Search terms included the keywords and the corresponding Medical Subject Headings for “ticagrelor,” “clopidogrel,” “aspirin,” “antiplatelet,” “secondary prevention,” “trial,” and “randomized trial.” Inclusion criteria were (1) randomized design, (2) total number of participants >1,000, and (3) follow-up duration ≥3 months. Comparisons of ticagrelor- versus prasugrel-based antiplatelet regimens and trials of ticagrelor-based versus conventional antiplatelet regimens after routine administration of thrombolytic agents were ineligible.
Study selection and quality assessment
Two investigators (S. C. and G. N.) independently assessed publications for eligibility at title and/or abstract level, with divergences resolved by a third investigator (A. K.). Studies that met inclusion criteria were selected for further analysis. Freedom from bias was independently evaluated for each study by the same investigators in accordance with The Cochrane Collaboration method. Composite quality scores were not assigned.
Data extraction and outcome variables
The primary outcome was all-cause mortality. Main secondary outcome was myocardial infarction (MI). Other secondary outcomes were stroke and major bleeding. All end points were evaluated in the intention-to-treat population up to the maximum follow-up duration in accordance to definitions reported in the original protocols.
Data synthesis and analysis
We used hazard ratios (HRs) and 95% CIs to compare the outcomes of interest associated with ticagrelor-based therapy (experimental group) versus conventional antiplatelet regimens (control group). Trial-level logHRs and corresponding SEs were pooled using a fixed- or random-effect model (in case of significant heterogeneity) with the inverse variance weighting. Heterogeneity between the trials was quantified using the I 2 statistic: I 2 values around 25%, 50%, and 75% were considered to indicate low, moderate, or high heterogeneity, respectively. In addition, we estimated the between-study variance (τ ) according to DerSimonian-Laird and derived the 95% prediction interval of pooled estimates. An influence analysis, in which meta-analysis estimates are computed omitting one study at a time, was performed for the primary and main secondary outcomes. We performed a linear regression test for funnel plot asymmetry for main outcomes. We used a χ 2 test for treatment-by-subgroup interaction to investigate whether the treatment effect was dependent on the primary diagnosis of enrolled patients in the original trials (CAD or PAD). By means of random-effects meta-regression analysis, we assessed the impact of therapy adherence and duration in the experimental arm with respect to primary and main secondary outcomes. Finally, we determined the power of our random-effects meta-analysis to detect a prespecified 20% relative risk reduction (RRR) for primary and main secondary outcomes with the experimental therapy conditional on the observed precision of the pooled estimate. The 20% RRR threshold was chosen because it was the average risk reduction threshold (range 13.5%-22.5%) selected to power the individual study designs of randomized trials included in this meta-analysis. This study was reported in compliance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses statement (Supplemental Table I). All analyses were performed in R (version 3.3.2; R Foundation for Statistical Computing, Vienna, Austria). No extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.
Results
Eligible studies
The flow diagram for the trial selection process is shown in Supplemental Figure 1. After application of inclusion/exclusion criteria, 6 trials were included in meta-analysis: 5 trials published as full-length manuscripts and 1 trial as meeting presentation . In these trials, a total of 84,477 patients were randomly allocated and received at least 1 dose of either experimental (n = 45,709) or control therapy (n = 38,768). In 1 trial, patients received either ticagrelor 60 mg twice daily (n = 6,958) or ticagrelor 90 mg twice daily (n = 6,988): for the main analysis, we obtained from this study only the treatment arm prescribing ticagrelor 60 mg twice daily because the latter dose regimen is actually recommended from guideline-writing authorities for long-term secondary prevention after MI. In addition, we explored the treatment effect for all outcomes of interest by pooling outcomes data from treatment arms prescribing either ticagrelor 60 mg or 90 mg twice daily to derive summary risk estimates. Thus, a total of 77,489 participants were available for final analyses allocated to experimental (n = 38,721) or control (n = 38,768) therapy. All trials had a multicenter design. The main characteristics of the trials included are shown in Supplemental Table II. Briefly, patients with obstructive chronic/stable or unstable CAD and/or PAD were randomized to experimental or control therapy. Patients assigned to experimental therapy received ticagrelor 60 or 90 mg twice daily as monotherapy or in combination with aspirin 75-150 mg once daily for a period of time ranging between 3 and 33 months. Patients assigned to control therapy received clopidogrel 75 mg once daily as monotherapy or in combination with aspirin 75-150 mg once daily. The GLOBAL LEADERS trial had a peculiar 3-arm design: in this study, the experimental treatment consisted of ticagrelor 90 mg twice daily in combination with aspirin 75-150 mg once daily for 1 month followed by ticagrelor monotherapy for 23 months, whereas the control arms consisted of clopidogrel 75 mg once daily or ticagrelor 90 mg twice daily in combination with aspirin 75-150 mg once daily for 12 months followed by aspirin 75-150 mg once daily for further 12 months. For the current study, data belonging to these 2 latter groups were pooled in the control therapy in the main analysis. Additionally, 2 separate analyses explored the treatment effect for main outcomes associated with either ticagrelor (experimental therapy) or aspirin (control therapy) beyond 1 year and with either ticagrelor (experimental therapy) or clopidogrel (control therapy) up to 2 years. In 4 trials , , , the predominant diagnosis was CAD including patients with ACS or stabilized MI. The Ticagrelor in Coronary Artery Bypass Graft (TiCAB) trial allocated patients after surgical coronary revascularization to either ticagrelor or aspirin. One trial enrolled patients with nonsevere ischemic stroke or high-risk transient ischemic attack not due to a cardioembolic cause. In the experimental group, the adherence to assigned antiplatelet therapy at the longest available follow-up ranged between 69.9% and 86.2%. All patients received ancillary therapies for acute or chronic manifestations of CAD and/or PAD in accordance with standard of care. All trials had sufficient statistical power for composite clinical end points, which included overall mortality and MI in most of them. Three trials , , reported outcome data beyond 12 months. The definitions of outcomes are reported in Supplemental Table III, and the risk of bias among studies is reported in Supplemental Table IV.
Baseline characteristics are shown in the Table . Patients were more often male and had a median age of 65.5 years (interquartile range, 64.5-66.0), nearly a third of them had diabetes, and one fourth of them were active smokers at the time of inclusion in the primary trials. A previous MI was reported in 18.7% of patients and 5.8% of the patients have had cerebrovascular accidents before enrolment. The weighted median follow-up available for the assessment of outcomes of interest was 20 months (mean 19.0 ± 11.8).
Clinical outcomes
Primary outcome
Overall, all-cause deaths occurred in 3,607 patients (4.6%; Figure 1 , A ). The primary outcome occurred in 4.5% of patients treated with experimental therapy and 4.9% of patients treated with control therapy (HR = 0.91, 95% CI 0.81-1.01; P = .07). The random-effects meta-analysis had 98.4% power to detect a 20% RRR of all-cause death associated with the experimental therapy. Of note, the 95% prediction interval for this outcome contained the null (0.68-1.21), and there was evidence of significant heterogeneity. However, the risk for all-cause death was significantly reduced with the experimental therapy in trials that enrolled only patients with CAD as primary diagnosis (4.0% vs 4.8%; HR = 0.84 [0.77-0.91]; P < .001; Figure 1 , B ) compared with trials with PAD as primary diagnosis (5.1% vs 5.1%; 1.01 [0.91-1.12], P = .91; P for interaction [ P int ] = .008). The number needed to treat to avoid 1 case of all-cause death with the experimental therapy in trials that enrolled patients with CAD as primary diagnosis was 130 patients (90-231). Cardiac death occurred in 1,985 patients (3.2%; data available for 61,521 participants; Supplemental Figure 2). Patients treated with experimental versus control therapy showed a comparable risk of cardiac death (3.1% vs 3.4%; HR = 0.92 [0.77-1.09], P = .32). Of note, the risk for cardiac death was significantly reduced with experimental therapy in trials that enrolled patients having CAD as primary diagnosis (HR = 0.80 [0.72-0.90], P < .001) compared with trials with PAD as primary diagnosis (HR = 1.08 [0.94-1.24], P = .28; P int = .001).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
