Identifying Patients at High Risk of Coronary Events After Stroke

A Prerequisite for Better Prevention

Peter J. Kelly, MD; Chris Price, MD


Stroke. 2019;50(12):3335-3336. 

Despite high rates of treatment with modern preventive medications, about 25% to 30% of ischemic stroke survivors will experience a recurrent stroke or coronary event by 5 years after their initial stroke.[1–3] Coronary events (myocardial infarction [MI] and sudden cardiac death) account for about one-third of these and are more likely to be fatal compared with recurrent stroke.[3] Although the risk of poststroke coronary recurrence is the highest in patients with known coronary disease, these account for a minority (12%–25%) of patients in population-based studies.[4,5] Consequently, the absolute number of new coronary events is likely to be at least as high in patients with no previous coronary history.[6] However, despite the high risk and serious consequences, no randomized trials exist of coronary screening strategies in asymptomatic patients after stroke. Guidelines for coronary screening during routine care vary, with some not giving explicit guidance on the issue of routine screening and others providing mainly opinion-based recommendations, which have not been recently renewed.[7–9] In part, this reflects the absence of evidence that percutaneous coronary intervention benefits asymptomatic patients after stroke, although subgroup analysis from early clinical trials suggests that coronary bypass surgery might be beneficial in asymptomatic patients with left main or 3-vessel disease.[10] Other currently available treatment options supported by randomized trial data include intensification of existing prevention strategies such as smoking cessation and lipid-lowering (via dietary measures and statins) and selection for newer drugs such as proprotein convertase subtilisin-kexin 9 inhibitors.[11–13] Observational data also suggest that knowledge of coronary disease is associated with improved compliance with preventive interventions in asymptomatic individuals.

A major barrier to improving personalized coronary disease risk management among stroke patients without cardiac symptoms has been the difficulty of identifying patients who are most likely to benefit because detailed investigation of all patients is unlikely to be feasible. In this issue of Stroke, Boulanger et al[14] report a study investigating coronary risk stratification in 2555 ischemic stroke/transient ischemic attack survivors followed for up to 10 years in the population-based OXVASC (Oxford Vascular Study). Rather than attempt to derive a new risk score, they reasoned that the existing Essen Score (which includes several risk factors for atherosclerosis) might identify patients who subsequently developed MI but had no known coronary disease at the time of stroke/transient ischemic attack onset (almost 80% of their cohort). The Essen Score improved identification of patients with subsequent MI better than chance (c-statistic 0.64, 95% CI, 0.57–0.71). Asymptomatic patients with a high score (4 or 5) had an estimated 10-year MI risk of 17%, similar to that in those with known coronary disease.

The study demonstrates that a simple risk score based on variables which are widely available in clinical practice may improve poststroke coronary risk prediction. Strengths include the rigorous methodology and population-based design (minimizing selection bias), large sample size, long and near-complete follow-up, high rates of modern preventive medication use, analysis of stroke subtypes, and choice of a highly relevant outcome measure (MI).

Although the study is an important step forward, some limitations must be mentioned. As acknowledged by the authors, the study sample was mainly white and reflected the population of a single service, so the findings may not be directly generalizable to other clinical settings or ethnic groups. Because the Essen score was originally developed for prediction of late recurrent stroke, several variables which may have greater prognostic information for recurrent coronary events caused by atherosclerosis are not included (eg, sex and low-density lipoprotein cholesterol). Although the score identified patients with future MI better than chance, the c-statistic of 0.64 indicates only modest discrimination. In the Boulanger study, about 70% of patients (48 of 70) without known coronary disease who developed future MI had low or medium scores (3 or less). The c-statistic was increased (to 0.74) by the addition of large artery disease to the Essen Score, consistent with earlier studies which added information about carotid stenosis to the Framingham risk score.[15,16] As a further step, addition of information from imaging studies such as coronary artery calcium score or lumen stenosis on CTCA might further improve identification of high-risk patients. Future studies should investigate whether a strategy of further refining the Essen score or derivation of a new score more specific for coronary events is most useful for coronary risk prediction.

Improved cardiac risk stratification may also offer the opportunity for stroke patients to be included in randomized trials of new secondary prevention interventions from which they are frequently excluded. Targeting new antiatherosclerotic therapies toward selected high-risk patients may increase the likelihood of demonstrating efficacy and reduce the sample size required in randomized trials, thereby shortening the time to bring new evidence into practice. It is time to move toward a more positive view about the potential benefits from identification and treatment of asymptomatic coronary disease after stroke.