Weekly Journal Scan: What should we CELEBRATE and what remains uncertain about early platelet inhibition in ST-elevation myocardial infarction

This comment refers to ‘Zalunfiban at First Medical Contact for ST-Elevation Myocardial Infarction’ which was published in the New England Journal of Medicine, https://doi.org/10.1056/EVIDoa2500268. CELEBRATE (CeleCor Blinded Randomized Trial in ST-Elevation Myocardial Infarction) was a phase 3, multicentre, industry-funded, randomized, double-blind, placebo-controlled trial assessing whether a single subcutaneous administration of zalunfiban at first medical contact improves early coronary reperfusion and clinical outcomes in patients with suspected ST-elevation myocardial infarction (STEMI). Zalunfiban is a novel, fast-acting, reversible, small-molecule glycoprotein IIb/IIIa inhibitor (GPI) designed for pre-hospital use.1 The primary efficacy endpoint was a 30-day hierarchical composite clinical outcome analysed using a proportional odds model. The hierarchy, from worst to best outcome, included (vii) all-cause death; (vi) stroke; (v) recurrent myocardial infarction (MI, types 1–4); (iv) acute stent thrombosis within 24 h after PCI; (iii) new-onset or worsening heart failure requiring rehospitalization; (ii) large MI, defined by high-sensitivity cardiac troponin T levels ≥30 times the upper limit of normal within 24 h of drug administration; and (i) absence of any of the above events. Key secondary efficacy endpoints included TIMI flow grade in the infarct-related artery at index angiography, ST-segment resolution 1 h after PCI, and the need for blinded bailout therapy with intravenous GPIs or P2Y12 blockers. The primary safety endpoint was GUSTO-defined severe or life-threatening bleeding at 30 days, with secondary safety outcomes including moderate or mild bleeding, thrombocytopenia, and injection-site reactions. A total of 2467 participants from eight countries were randomized in a 1:1:1 ratio to receive zalunfiban 0.11 mg/kg (n = 853), 0.13 mg/kg (n = 818), or placebo (n = 796). Study drug administration was performed predominantly in the pre-hospital setting (enrolment occurred in the ambulance in 87% of cases) by trained emergency medical personnel. Baseline characteristics were well balanced between treatment groups: mean age was 63 years, 21% of participants were women, and 95% were Whites. Most patients were allowed standard of care antithrombotic therapy, including aspirin, heparin, and/or a P2Y12 blocker (mostly ticagrelor) before or concomitant with study drug administration; overall, 95% of trial participants were treated with heparin and a P2Y12 blocker in addition to the study drug. The median time from symptom onset to angiography was 130 min, from study drug administration to angiography 36 min, and from study drug administration to PCI 50 min. PCI was performed in up to 92% of the patients, the radial artery was the preferred access site (95%), and approximately 17% of patients had three-vessel disease. Compared with placebo, patients in the pooled zalunfiban group had significantly lower odds of the primary hierarchical composite endpoint at 30 days [adjusted odds ratio (OR), 0.79; 95% confidence interval (CI), 0.65–0.98; P = .028], mostly driven by a significant reduction in large MI (OR 0.71; 95% CI 0.54–0.94). Acute stent thrombosis occurred less frequently with zalunfiban (0.2% vs 1.0%; OR, 0.18; 95% CI, 0.03–0.74), whereas reinfarctions were numerically more frequent with GPI but not statistically significantly different vs placebo (1.9% vs 1.2%; OR, 1.65; 95% CI, 0.76–3.97). The absence of any major adverse clinical endpoint was observed more frequently in the zalunfiban group than in the placebo group (13.3% vs 9.8%; OR, 1.41; 95% CI, 1.06–1.90). Treatment effects were consistent across prespecified subgroups. Zalunfiban was also associated with improved early coronary reperfusion, with a higher likelihood of achieving TIMI 2–3 flow at index angiography (52.0% vs 45.1%) and greater resolution of ST-segment deviation prior to angiography. The incidence of GUSTO severe or life-threatening bleeding at 30 days did not differ significantly between zalunfiban and placebo (1.2% vs 0.8%; P = .40), whereas GUSTO mild or moderate bleeding occurred more than twice as frequently in the pooled zalunfiban group than in the placebo group (6.4% vs 2.5%; P < .001). Bleeding Academic Research Consortium (BARC) 2–5 bleedings also occurred more frequently in the zalunfiban group (6.1% vs 2.8%, P < .001) Early and timely restoration of coronary artery patency remains a central goal in STEMI management; yet a substantial proportion of patients still arrive at the catheterization laboratory with an occluded infarct-related artery, a high thrombotic burden and ongoing myocardial ischaemia. Contemporary guidelines recommend early dual antiplatelet therapy, usually aspirin plus a P2Y12 inhibitor, but evidence supporting their initiation before coronary angiography is limited.2 Oral agents such as ticagrelor or prasugrel reach maximal platelet inhibition only several hours after administration. In addition, their onset of action may be further delayed in the acute STEMI setting owing to impaired absorption in haemodynamically unstable patients and the frequent use of opioids for symptom relief.3 Prior attempts to accelerate reperfusion using intravenous GPIs demonstrated mixed results: Abciximab before Direct Angioplasty and Stenting in Myocardial Infarction Regarding Acute and Long-Term Follow-up (ADMIRAL) and Ongoing Tirofiban in Myocardial Evaluation (ON-TIME 2) trials showed improvements in early ischaemic surrogates, whereas Facilitated Intervention with Enhanced Reperfusion Speed to Stop Events (FINESSE) trial failed to demonstrate a clear reduction in hard clinical outcomes and revealed instead an increased bleeding risk.4–6 The practical limitations of intravenous GPIs, including the requirement for continuous infusion and the resulting logistical challenges in the pre-hospital setting, have driven the need for rapid, effective, and reversible platelet inhibition available at first medical contact.7,8 The CELEBRATE trial tries to address this unmet need by evaluating zalunfiban, a novel subcutaneous small-molecule GPI, designed for very early administration by emergency medical personnel. Like traditional intravenous GPIs, subcutaneous delivery also enables rapid platelet inhibition (within 15 min) and a reversible effect (half-life of ∼1 h and a reduction in antiplatelet effect by roughly 50% within 2 h). These pharmacological characteristics are intended to bridge the gap between first medical contact and maximal platelet inhibition by oral agents, restoring early coronary flow and reducing procedural thrombotic complications.9 CELEBRATE used a hierarchical composite endpoint to capture a spectrum of ischaemic events. Of note, the observed benefit was primarily driven by improvements in surrogate markers of reperfusion, such as TIMI flow at angiography and ST-segment resolution, and by a reduction in acute stent thrombosis. Numerically fewer early stent thromboses were counterbalanced by a higher rate (although not statistically significant) of 30-day recurrent MI. Hard clinical endpoints, including 30-day mortality and reinfarction, were not significantly affected. Notably, the primary endpoint was dominated by the least clinically relevant component in the hierarchical design (a large MI), which substantially weakens the strength of the evidence and raises uncertainty regarding the presence of a true clinical benefit. This concern is reinforced when considering the upper limit of the 95% CI for the overall OR of the primary composite endpoint (0.98), indicating that statistical significance rests on a very narrow margin and could be lost with the misclassification of only a small number of events. Moreover, the interpretation of efficacy is further complicated by the trial design, which tested two different doses of the investigational drug, without reporting results separately for each, raising the possibility that neither dose alone clearly achieved statistical significance vs placebo, with the pooled analysis used to increase the statistical power of the study. In this regard, if the primary object was to maximize statistical power, a 2:1 randomization to a single selected dose would have been a more coherent design choice. Safety findings reflect the well-known trade-off between antithrombotic efficacy and bleeding liability. CELEBRATE showed no excess of severe or life-threatening bleeding, but clinically relevant bleeding (BARC types 2, 3, 5) and GUSTO mild-to-moderate bleeding were significantly higher with zalunfiban than with placebo, despite the majority of study participants undergoing radial access for PCI. The lack of statistical significance for the primary safety endpoint likely reflects limited statistical power, given the very low (0.8%) incidence of severe bleeding in the control group. In addition, thrombocytopenia occurred numerically more often in the zalunfiban group, with an increase (although not statistically significant) of approximately one-third, a pattern consistent with the known class effects of GPIs. These findings underscore the importance of careful patient selection, as individuals with a lower atherothrombotic risk may be exposed to an excess of bleeding events, which, irrespective of severity, have been consistently associated with increased 1-year mortality.10 Generalizability represents another limitation. CELEBRATE was conducted mainly in well-organized emergency medical systems, with short delays from first medical contact to angiography and high adherence to guideline-recommended therapies. Moreover, women and non-White populations were underrepresented, which may limit extrapolation to less-organized emergency settings or different demographic groups. Finally, background antithrombotic pre-treatment was not standardized, resulting in substantial treatment heterogeneity, and it remains unclear which specific antithrombotic drugs were used as background therapy and how they may have influenced the results. Importantly, aspirin was not administered as pre-treatment, and approximately 95% of participants received heparin and a P2Y12 inhibitor before or at the time of study drug administration, further complicating the interpretation of the incremental effects of zalunfiban. In conclusion, CELEBRATE may reopen a chapter in STEMI pharmacotherapy previously thought to be closed. The trial provides reasons to celebrate a step forward in pre-hospital STEMI management, while simultaneously warranting caution, as the suggested benefit is largely driven by surrogate endpoints. Accordingly, CELEBRATE should be regarded as a proof-of-concept rather than practice-changing evidence. The study does not allow identification of patients most likely to derive net clinical benefit, nor does it define the role of this strategy within contemporary antithrombotic therapy. Further adequately powered trials are required to clarify patient selection, interactions with modern antithrombotic regimens, and whether very early pharmacologic reperfusion improves meaningful clinical outcomes. G.L. received grant support (to the Institution) for investigator-initiated research from American Heart Association, Italian National Health Service and Italian Minister of Education, University and Research. She is currently involved in the research programmes of the Italian Cardiovascular Network. She received personal fees from AstraZeneca, Boehringer Ingelheim, Novo Nordisk, Daiichi Sankyo, Sanofi, and Novartis. D.P. received speaker’s fees from Daiichi Sankyo.