Acute cardiac care & resuscitationCohort study
Phenotypic clustering of myocardial infarction complicated by out-of-hospital cardiac arrest using unsupervised machine learning
- prospective cohort
- unsupervised machine learning analysis using k-means clustering
Authors' conclusion
Unsupervised machine learning identified three phenotypes among patients with MI-related OHCA associated with distinct outcomes. This phenotypic classification may facilitate personalized management and refined prognostic assessment in this high-risk population.
Study context
- Population
- Patients admitted after OHCA related to myocardial infarction (MI)Cluster 1 260 · Cluster 2 118 · Cluster 3 100
- Participants
- 478 patients
- Exposure
- Three clusters identified by k-means clustering, compared with cluster 1
- Cluster 1n = 260reference
- Cluster 2n = 118
- Cluster 3n = 100
Main finding
As reported in the abstract
Adverse in-hospital events:
- All-cause mortality
- Bleeding
- Stent thrombosis
- Most frequent in cluster 3
Observational data: associations, not effects.
Further findings
| Outcome | Cluster 1 | Cluster 2 | Cluster 3 | hazard ratio (95% CI) | |
|---|---|---|---|---|---|
| All-cause mortality · 90 days — Cluster 2 vs Cluster 1 | 22.5% | 53.0% | 77.2% | 2.97 (2.07–4.26) | |
| All-cause mortality · 90 days — Cluster 3 vs Cluster 1 | 6.75 (4.74–9.60) | ||||
| 0.10.31310Hazard ratio (log scale) |
- All-cause mortality: p < 0.001 across groups
Original publication
Phenotypic clustering of myocardial infarction complicated by out-of-hospital cardiac arrest using unsupervised machine learning. Resuscitation. Published 5 March 2026. doi:10.1016/j.resuscitation.2026.111037
Summarised from the published abstract; every number is checked against it