Validation of the SMART-REACH model after stroke and the effect of colchicine by atherosclerotic cardiovascular disease risk category: a secondary analysis of the CONVINCE randomised clinical trial.
The Colchicine for prevention of vascular inflammation in Non-CardioEmbolic stroke (CONVINCE) trial showed that recurrent events were significantly reduced among colchicine-adherent non-cardioembolic stroke patients in the on-treatment analysis. This study aimed to validate the SMART-REACH risk score in stroke patients, and to determine whether colchicine's efficacy varies by baseline atherosclerotic cardiovascular disease (ASCVD) risk.
Patients with non-severe non-cardioembolic ischaemic stroke/transient ischaemic attack (TIA) were randomised to colchicine 0.5 mg plus usual care or usual care alone. Participants were stratified into moderate (10%-19%), high (20%-30%) and very high (≥30%) 10-year ASCVD risk categories using the SMART-REACH model. Model performance was assessed using the C-statistic and calibration plots. The primary endpoint (major adverse cardiovascular events [MACE]) was a composite of fatal or non-fatal recurrent ischaemic stroke, myocardial infarction, cardiac arrest or hospitalisation for unstable angina.
Among 3144 patients, MACE incidence significantly increased with ASCVD risk levels: 7.2% (moderate), 8.8% (high) and 13.8% (very high) (P < .01). The C-statistic for 3-year risk of MACE was 0.59 (95% CI, 0.56-0.63). While no statistically significant treatment interaction was found (P = .88), absolute risk reductions (ARRs) were more pronounced in higher-risk groups: moderate risk 7.2% (colchicine) vs 7.2% (usual care) (hazard ratio [HR] 1.01; 95% CI, 0.55-1.83); high risk 7.7% vs 9.8% (ARR 2.1%; HR 0.79; 95% CI, 0.53-1.18); very high risk 12.5% vs 15.2% (ARR 2.7%; HR 0.85; 95% CI, 0.64-1.12).
We identified an association between very high baseline ASCVD risk (≥30%) assigned by the SMART-REACH score and increased recurrent MACE. Although no significant treatment interaction was observed, patients in higher risk categories may represent a more promising target population for secondary prevention with colchicine.
ClinicalTrials.gov Identifier: NCT02898610.
Patients with non-severe non-cardioembolic ischaemic stroke/transient ischaemic attack (TIA) were randomised to colchicine 0.5 mg plus usual care or usual care alone. Participants were stratified into moderate (10%-19%), high (20%-30%) and very high (≥30%) 10-year ASCVD risk categories using the SMART-REACH model. Model performance was assessed using the C-statistic and calibration plots. The primary endpoint (major adverse cardiovascular events [MACE]) was a composite of fatal or non-fatal recurrent ischaemic stroke, myocardial infarction, cardiac arrest or hospitalisation for unstable angina.
Among 3144 patients, MACE incidence significantly increased with ASCVD risk levels: 7.2% (moderate), 8.8% (high) and 13.8% (very high) (P < .01). The C-statistic for 3-year risk of MACE was 0.59 (95% CI, 0.56-0.63). While no statistically significant treatment interaction was found (P = .88), absolute risk reductions (ARRs) were more pronounced in higher-risk groups: moderate risk 7.2% (colchicine) vs 7.2% (usual care) (hazard ratio [HR] 1.01; 95% CI, 0.55-1.83); high risk 7.7% vs 9.8% (ARR 2.1%; HR 0.79; 95% CI, 0.53-1.18); very high risk 12.5% vs 15.2% (ARR 2.7%; HR 0.85; 95% CI, 0.64-1.12).
We identified an association between very high baseline ASCVD risk (≥30%) assigned by the SMART-REACH score and increased recurrent MACE. Although no significant treatment interaction was observed, patients in higher risk categories may represent a more promising target population for secondary prevention with colchicine.
ClinicalTrials.gov Identifier: NCT02898610.
Authors
Maes Maes, Verschuere Verschuere, Walsh Walsh, Weimar Weimar, Purroy Purroy, Price Price, Fonseca Fonseca, Hill Hill, Jatuzis Jatuzis, Kõrv Kõrv, Kruuse Kruuse, Mikulik Mikulik, Nederkoorn Nederkoorn, Czlonkowska Czlonkowska, Fischer Fischer, McCormick McCormick, Castellanos Castellanos, Baptista Baptista, Marto Marto, Thavanesan Thavanesan, Williams Williams, Kelly Kelly, Lemmens Lemmens,
View on Pubmed