Abstract WP479: Bedside Modified Volume-Viscosity Swallow Test Might be Reliably to Predict Risk of Aspiration Pneumonia After Stroke

Abstract

Background: Dysphagia is a common complication after stroke, which could lead to aspiration. Early and reliable bedside detection and intervention of dysphagia is necessary to prevent dysphagia-related aspiration pneumonia. Compared to water swallow test (WST) , modified volume-viscosity swallow test(V-VST) provides more detail information of management. To predict the risk of aspiration pneumonia from dysphagia tool in acute stroke patients, we developed modified V-VST and validated its predictability of risk of aspiration pneumonia. Methods: This was a prospective observational study enrolled acute stroke within 14 days. Only patients with Glasgow Coma Scale score higher than 10 were included into analyze. All patients had bedside WST and if failed would have a V-VST after admission was documented. Modified V-VST assessment, includes tongue movement to V-VST, including tongue rolling, extending to front, left and right. The primary outcome was the rate of pneumonia, which was diagnosed by either chest CT or chest X-ray. Multiple logistic regression models were performed in patients who failed the tests. The area under curves (AUCs) were compared among all tests. Results: From 7/2017-6/2018, 562 hospital-based acute stroke patients were enrolled. The overall incidence of pneumonia after acute stroke was 15.9 % (95%CI 13.0-19.2%). Of 157 patients who failed WST, 101 had a positive V-VST, and 173 had a positive modified VVST. The modified VVST was the most powerful to predict the risk of pneumonia (OR 6.156 in WST, 3.058 in VVST, 18.076 in modified VVST, P<.001), after adjusted age, sex, NIHSS score, length of stay, interval of onset to admission and smoke status. The AUC of modified V-VST was 0.851(95% 0.806-0.897) Conclusions: Bedside modified V-VST is an easy and sensitive tool to predict post-stroke aspiration pneumonia. A large prospective interventional study is needed to confirm our findings.