Raona, Mehran et al.: 2024 ESC high-ischaemic-risk criteria in 15,336 chronic coronary syndrome patients

Summary

The study assessed 15,336 chronic coronary syndrome patients treated with drug-eluting stents at Mount Sinai between 2012 and 2022. The 2024 ESC high-ischaemic-risk (HIR) criteria used were left main, stent length > 60 mm, diabetes, CKD, bifurcation and chronic total occlusion (CTO). The primary endpoint was MACCE (death, MI or stroke) at 1 year.

10,952 patients (71.4%) met at least one criterion. MACCE rose with the number of criteria:

In adjusted analysis, two-stent bifurcation and CTO did not predict MACCE: HR 1.24 (0.89 to 1.74) and HR 1.24 (0.85 to 1.80).

Strengths

• Large real-world cohort on a frequent decision: extended DAPT or not.
• The authors openly report the loss of prognostic value of bifurcation and CTO. This fits second-generation DES, POT and kissing balloon as default, intracoronary imaging in harder cases, and CTO success near 90% in dedicated centres.
• The conclusion is measured: the authors flag the 71.4% prevalence as a problem and call for a more granular classifier.

Methodological issues

• A proxy tested, labelled as validation. Supplementary Table 1 excludes three criteria: proximal LAD or last patent vessel in the left main criterion, suboptimal stent deployment, and prior stent thrombosis on antithrombotic therapy. The discussion lists others not captured: recurrent MI, polyvascular disease and imaging-assessed underexpansion. What is tested is a six-criterion proxy.
• Unclear multivariable model. Table 4 adjusts the six criteria for each other. The text does not say whether age, LVEF, prior MI, prior CABG, anaemia, multivessel disease or calendar year entered the model. These differ across strata: LVEF 56.6% vs 52.5%, prior CABG 11% vs 25%, anaemia 21% vs 53%. Without adjustment, the HRs are stratified crude associations.
• MACCE composition. Death accounts for 246 of 476 events. Stroke contributes 35, with no trend across strata (p = 0.37). Diabetes and CKD predict death from other causes, without a thrombotic pathway. Part of the risk labelled ischaemic is general mortality risk.
• Causal reading of intracoronary imaging. The supplementary model with imaging gives HR 0.59 (0.39 to 0.89) for MACCE, and the authors use it to support broader adoption. At a single high-volume centre, imaging use clusters by operator, lesion subset and indication. The E-value for HR 0.59 is about 2.8, within reach of unmeasured confounders such as operator volume, procedure era and complexity beyond SYNTAX.
• Statistical scaffolding. No Schoenfeld test, no multiplicity correction across the many outcomes and predictors, and no calendar-period stratification over a 2012–2022 window that spans two DES generations and a shift in DAPT practice.
• Generalisability. High-volume tertiary US centre: rotational atherectomy in 18% overall and 38% in the ≥ 3 criteria stratum, radial access in 22%. Applicability to a community lab is open.