Discontinuation of Beta-Blocker Therapy after Myocardial Infarction (SMART-DECISION)

Authors, Journal, Affiliations, Type, DOI

• Ki Hong Choi, Danbee Kang, Weon Kim, et al. for the SMART-DECISION Investigators
• N Engl J Med 2026;394:1302-12
• Heart Vascular Stroke Institute, Samsung Medical Center, Sungkyunkwan University (lead centre); 25 centres, South Korea
• Type: Open-label, multicenter, randomized, noninferiority trial
• DOI: https://doi.org/10.1056/NEJMoa2601005

Overview

The SMART-DECISION trial randomised 2,540 patients in stable condition after MI (LVEF ≥40%, no HF, ≥1 year of beta-blocker therapy) in a 1:1 ratio to discontinue or continue beta-blockers. At a median follow-up of 3.1 years, the primary composite endpoint (death from any cause, recurrent MI, or HF hospitalisation) occurred in 7.2% of the discontinuation group vs 9.0% of the continuation group (HR 0.80; 95% CI 0.57–1.13; P=0.001 for noninferiority), with the upper confidence bound of 1.13 below the prespecified margin of 1.4. The trial demonstrated that beta-blocker discontinuation is noninferior to continuation in this highly stabilised post-MI population — contrasting with the earlier ABYSS trial, which failed to show noninferiority due to differences in endpoint selection and trial design.

Keywords

Beta-blockers, myocardial infarction, discontinuation, noninferiority, left ventricular ejection fraction, heart failure, secondary prevention, STEMI, NSTEMI, SMART-DECISION

Key Takeaways

Background and Rationale

• Beta-blockers remain cornerstone therapy in HF or reduced LVEF (<40%) post-MI, but their benefit in patients with preserved EF (≥50%) has been questioned by multiple contemporary RCTs (REDUCE-AMI, REBOOT) and two meta-analyses showing no clinical benefit.
• Observational studies beyond the first year show inconsistent associations; the ABYSS trial did not show noninferiority of beta-blocker interruption (primary endpoint driven by hospitalisation for cardiovascular causes — susceptible to open-label bias).
• SMART-DECISION was designed specifically to address discontinuation of long-term beta-blocker therapy using hard, objective endpoints.

Trial Design and Population

• Open-label, multicenter RCT; 25 centres, South Korea; 1:1 randomisation (discontinuation vs continuation).
• Eligibility: stable post-MI, LVEF ≥40%, no ongoing HF treatment, ≥1 year of beta-blocker therapy, event-free from index MI to screening; no restriction on BB type/dose or time since MI.
• Exclusions: ongoing HF therapy, LVEF <40%, contraindications to BB, atrial fibrillation.
• Mean age 63.2 years; 12.8% women; 56.8% STEMI; 98.0% underwent coronary revascularisation at index MI.
• Median time from MI to randomisation: 4.7 years (IQR 2.4–8.4) — highly stabilised population.
• Most common BB at baseline: carvedilol 47.6%, bisoprolol 32.3%, nebivolol 19.9%.
• In discontinuation group, BB was stopped immediately after randomisation; continuation group maintained same agent and dose.

Primary Endpoint

• Composite of death from any cause, recurrent MI, or hospitalisation for heart failure.
• Prespecified noninferiority margin: upper limit of 95% CI for hazard ratio of 1.4.

Results

• Primary endpoint: 58 patients (7.2% 4-year KM estimate) in discontinuation vs 74 patients (9.0%) in continuation; HR 0.80 (95% CI 0.57–1.13); upper bound 1.13 < 1.4; P=0.001 for noninferiority.
• All-cause death: 2.4% vs 3.4%; HR 0.71 (95% CI 0.43–1.16).
• Recurrent MI: 2.3% vs 2.6%; HR 1.11 (95% CI 0.63–1.96).
• HF hospitalisation: 2.2% vs 2.1%; HR 0.82 (95% CI 0.42–1.57).
• Per-protocol analysis: HR 0.79 (95% CI 0.56–1.12) — consistent with ITT.
• Serious adverse events: 11.5% (discontinuation) vs 13.4% (continuation); serious cardiac events 5.7% vs 6.1%.
• Subgroup analyses: Consistent results across prespecified subgroups.
• LVEF and NT-proBNP: No clinically meaningful differences between groups.
• Quality of life (PROMIS-29): No apparent differences between groups.
• Blood pressure and heart rate: Both increased after BB discontinuation, but uncontrolled hypertension/tachycardia rates were not clinically alarming.

Comparison with ABYSS Trial

• ABYSS enrolled patients a median 2.9 years post-MI (vs 4.7 in SMART-DECISION) — less stabilised population.
• ABYSS primary endpoint included hospitalisation for broad cardiovascular causes (angina admissions, diagnostic angiography) — susceptible to open-label clinical decision-making; hard outcomes (death, MI, HF hospitalisation) were similar between groups in ABYSS.
• SMART-DECISION had more intensive background secondary prevention: more ezetimibe use, P2Y12 monotherapy (vs aspirin), lower LDL achieved.
• These differences in endpoint design, timing of enrolment, and secondary prevention intensity explain the divergent noninferiority conclusions.

Limitations of the Document

• Open-label design: mitigated by adjudicated, objective primary endpoints.
• Single-country (South Korea): limited generalisability to different healthcare systems, revascularisation practices, and background therapy patterns.
• Noninferiority margin of 1.4 is relatively wide; lower-than-expected event rate reduces precision and cannot exclude a clinically meaningful increase in risk.
• Small proportion of women (12.8%) and mildly reduced EF (40–49%) patients — subgroup analyses are exploratory and hypothesis-generating only.
• LVEF was based on most recent pre-randomisation echocardiogram; possible unmeasured LVEF change between imaging and randomisation.
• Trial enrolled at median 4.7 years post-MI — does not define a safe time point for earlier discontinuation.

Key Concepts Mentioned

• concepts/Beta-Blocker-Post-MI — central evidence base: noninferiority of discontinuation vs continuation
• concepts/Noninferiority-Trial-Design — methodology: HR upper CI < 1.4; one-sided 97.5% CI equivalence
• concepts/Secondary-Prevention-Post-MI — background therapy (P2Y12 monotherapy, ezetimibe, LDL targets)

Key Entities Mentioned

• entities/Acute-Coronary-Syndrome — index event; STEMI/NSTEMI enrollment
• entities/Chronic-Coronary-Disease — long-term management; BB de-escalation evidence update

Wiki Pages Updated

• wiki/sources/BB-SMARTDECISION-NEJM-2026.md (this file — created)
• wiki/sourceindex.md (updated)
• wiki/wikiindex.md (updated)
• wiki/concepts/Beta-Blocker-Post-MI.md (created)
• wiki/entities/Chronic-Coronary-Disease.md (updated — BB section)
• wiki/entities/Acute-Coronary-Syndrome.md (updated — long-term management)