Can bedside assessment reliably exclude aspiration following acute stroke?

OBJECTIVE To investigate the ability of a bedside swallowing assessment to reliably exclude aspiration following acute stroke. SUBJECTS Consecutive patients admitted within 24 h of stroke onset to two hospitals. METHODS A prospective study. Where possible, all patients had their ability to swallow assessed on the day of admission by both a doctor and a speech and language therapist using a standardized proforma. A videofluoroscopy examination was conducted within 3 days of admission. RESULTS 94 patients underwent videofluoroscopy; 20 (21%) were seen to be aspirating, although this was not detected at the bedside in 10. In 18 (22%) of the patients the speech and language therapist considered the swallow to be unsafe. In the medical assessment, 39 patients (41%) had an unsafe swallow. Bedside assessment by a speech and language therapist gave a sensitivity of 47%, a specificity of 86%, positive predictive value (PPV) of 50% and a negative predictive value (NPV) of 85% for the presence of aspiration. Multiple logistic regression was used to identify the optimum elements of the bedside assessments for predicting the presence of aspiration. A weak voluntary cough and any alteration in conscious level gave a sensitivity of 75%, specificity of 72%, PPV of 41% and NPV of 91% for aspiration. CONCLUSION Bedside assessment of swallowing lacks the necessary sensitivity to be used as a screening instrument in acute stroke, but there are concerns about the use of videofluoroscopy as a gold standard. The relative importance of aspiration and bedside assessment in predicting complications and outcome needs to be studied.