Unsupervised phenomapping of perioperative risk in stable coronary artery disease: revealing limitations of the Revised Cardiac Risk Index
Authors' conclusion
Perioperative risk in stable CAD is phenotype-dependent. In this single-center cohort, RCRI performance varied by phenotype and may misclassify risk in clinically important high-risk phenotypes, particularly those characterized by occult ischemia or metabolic dysregulation. CHALSA showed more stable performance across phenotypes
Study context
- Population
- Patients with stable coronary artery disease (CAD) undergoing elective non-cardiac surgery, 2013–2023
- Participants
- 9,171 patients
- Exposure
- CHALSA vs Revised Cardiac Risk Index (RCRI) for perioperative major adverse cardiovascular events (MACE) prediction
- Outcome
- Perioperative MACE:
- Discrimination of RCRI and CHALSA
- Overall
- By phenotype
Methodological detailssingle-centre, unsupervised machine learning phenomapping
- Single-centre
- Unsupervised machine learning phenomapping
Main finding
As reported in the abstract
Discrimination for perioperative MACE, overall
| Group | AUROC |
|---|---|
| CHALSA | 0.816 |
| RCRI | 0.711 |
AUROC
CHALSA0.816
RCRI0.711
- CHALSA vs RCRI: P < 0.001 — 0.816 vs 0.711
- Overall perioperative MACE
- 514 (5.6%)
- Six phenotypes identified; MACE incidence range
- 1.0% to 16.4%
Observational data: associations, not effects.
Further findings
Incremental discrimination of CHALSA over RCRI, by phenotype
- Non-revascularized Ischemia ΔAUROC
- 0.221
- Severe Metabolic Burden ΔAUROC
- 0.264
- Stable Revascularized ΔAUROC
- 0.205
- Low-Risk ΔAUROC
- 0.172
- Ischemic HF ΔAUROC
- 0.086
- Elderly Frailty ΔAUROC
- 0.142
Original publication
Unsupervised phenomapping of perioperative risk in stable coronary artery disease: revealing limitations of the Revised Cardiac Risk Index. Eur J Prev Cardiol. Published 24 March 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →