[Prediction of cerebrovascular event risk following myocardial infarction]

Introduction: Patients with coronary artery disease (CAD) are at increased risk of stroke. The aim of this study was to analyze the prognostic accuracy of selected clinical and laboratory variables in stroke risk prediction following discharge after myocardial infarction (MI).

Methods: We analyzed 404 consecutive patients (aged 68.1±13.7 years; 63.4% male; 37.4% with diabetes) without previous stroke who were discharged in sinus rhythm after being admitted for MI. The following data were collected: cardiovascular risk factors, admission blood glucose (BG), HbA1c, creatinine, peak troponin levels; glomerular filtration rate (GFR) by the MDRD formula; maximum Killip class; GRACE score for in-hospital and 6-month mortality; and extent of CAD. Patients were followed for two years and each variable was tested as a possible predictor of cerebrovascular events (stroke or transient ischemic attack [TIA]).

Results: During follow-up, 27 patients were admitted for stroke or TIA. The presence of diabetes, hypertension, dyslipidemia and previously known CAD, type of MI (STEMI vs NSTEMI) and extent of CAD did not predict cerebrovascular risk. The following variables were associated with higher stroke risk: GFR <60ml/min/m(2) (p=0.029, OR 2.65, 95% CI 1.07-6.55); maximum Killip class >1 (p=0.025, OR 2.71, 95% CI 1.10-6.69); GRACE in-hospital mortality >180 (p=0.001, OR 4.09, 95% CI 1.64-10.22); admission BG >140 mg/dl (p=0.001, OR 5.74, 95% CI 1.87-17.58); GRACE 6-month mortality >150 (p=0.001, OR 4.50, 95% CI 1.80-6.27); and peak troponin >42ng/ml (p=0.032, OR 2.64, 95% CI 1.06-6.59). Logistic regression analysis produced a model with the predictors GRACE 6-month mortality >150 (OR 3.26; p=0.014) and admission BG >7.7mmol/l (OR 4.09; p=0.017) that fitted the data well (Hosmer-Lemeshow: p=0.916).

Discussion/conclusions: In patients with MI, variables known to be predictors of in-hospital mortality, including admission BG, renal function, acute heart failure and GRACE score, were found to be useful predictors of stroke during 2-year follow-up. While both GRACE score for 6-month mortality >150 and admission BG >7.7 mmol/l were independent predictors of stroke, CV risk factors, previously known CAD, and extent of CAD assessed by coronary angiography did not improve stroke risk prediction. This study highlights the need for even more aggressive secondary prevention in patients most at risk.