• Uwe Walter, Rupert Knoblich, Volker Steinhagen, Martina Donat, Reiner Benecke, and Antje Kloth.
• Dept. of Neurology, University of Rostock, Gehlsheimer Str. 20, 18147, Rostock, Germany. uwe.walter@med.uni-rostock.de
• J. Neurol. 2007 Oct 1; 254 (10): 1323-9.

ObjectiveTo determine independent clinical predictors of stroke-associated pneumonia (SAP) that are available in all patients on day of hospital admission.MethodsWe studied 236 patients with acute ischemic stroke admitted to the neurological intensive care unit at our university hospital. Risk factors of SAP and of non-responsivity of early-onset pneumonia (EOP; onset within 72 hours after admission) to initial antibacterial treatment were analyzed.ResultsIncidence of SAP was 22%. The following independent risk factors were found to predict SAP with 76% (EOP: 90%) sensitivity and 88% specificity: dysphagia (RR, 9.92; 95% CI, 5.28-18.7), National Institute of Health Stroke Scale > or = 10 (RR, 6.57; CI, 3.36-12.9), non-lacunar basal-ganglia infarction (RR, 3.10; CI, 1.17-5.62), and any other infection present on admission (RR, 3.78; CI, 2.45-5.83). Excluding the patients with other infections on admission, the same independent risk factors (except infection) were found. Further, but not independent risk factors were: combined brainstem and cerebellar infarction, infarction affecting more than 66% of middle cerebral artery territory, hemispheric infarction exceeding middle cerebral artery territory, impaired vigilance, mechanical ventilation, age > or = 73 years, current malignoma, and cardioembolic stroke, whereas patients with lacunar infarctions had significantly lower risk. In contrast to previous reports, no impact of male gender or diabetes was found. Initial vomiting, especially if associated with impaired vigilance, predicted antibacterial treatment non-responsivity of EOP. In nonresponders exclusively fungal pathogens were identified.ConclusionIncreased risk of pneumonia in acute stroke patients can be sufficiently predicted by a small set of clinical risk factors.

20 keywords...

hide…