Clustering of vascular risk factors and in-hospital death after acute ischemic stroke

Objective to evaluate the influence of clustering of vascular risk factors (VRF) on in-hospital mortality in patients with acute ischemic stroke (AIS) developing a risk score for mortality prediction. Methods clinical data from 1527 patients admitted to hospital with a first-ever AIS were prospectively evaluated from 1997 to 2005 assessing the presence of six VRF: diabetes, hypertension, hyperlipidemia, ischemic heart disease, atrial fibrillation and peripheral arterial disease. A composite vascular risk score (VRS) was created using five risk factors.Hyperlipidemia was excluded from the score due to its protective impact on mortality. Two modified VRS models were created and assessed for their accuracy as predictors for in-hospital mortality based on the odds ratio for each VRF obtained in the univariate analysis. Results 197 patients (12.9 %) died during the acute hospitalization period. Stroke severity increased with each additional VRF (p = 0.002). Cox proportional hazards model adjusted for confounders showed an association between the composite VRS and in-hospital mortality (p = 0.045). According to the clustering of VRS, the risk for in-hospital death increased from 1.951 (95 % CI 1.041-3.665) for patients having one VRS to 2.343 (95% CI 1.081-5.076) for those having a VRS >= 4. ROC curves showed that the modified VRS model based on a given value of one for each accumulated VRF, including the absence of hyperlipidemia, had the highest predictive capability for in-hospital mortality (p < 0.0001). Conclusions the presence of multiple VRF in patients with acute ischemic stroke increases the stroke severity and the risk of in-hospital death.