This prospective study was undertaken to determine the accuracy of bedside clinical methods compared with fiberoptic endoscopic examination of swallowing (FEES) for detecting aspiration in acute stroke patients. Fifty patients underwent an examination of their ability to swallow 50 ml of water in 10-ml aliquots. Later their oxygen saturation levels before and after swallowing 10 ml of water were measured using a pulse oximeter. Oxygen desaturation of more than 2%, was considered to be clinically significant. All patients then underwent a FEES assessment by a speech therapist and were followed up during their inpatient stay for evidence of aspiration pneumonia. The oxygen desaturation test had a sensitivity of 76.9% and specificity of 83.3% (chi2 = 18.154, p = 0.00002), while the 50-ml water swallow test had a sensitivity of 84.6% and specificity of 75.0% (chi2 = 18.001, p = 0.00002). However, when these two tests were combined into one test called "bedside aspiration," the sensitivity rose to 100% with a specificity of 70.8% (chi2 = 27.9, p = 0.000001). Five (10%) patients developed pneumonia during their inpatient stay. The relative risk (RR) of developing pneumonia, if there was evidence of aspiration on FEES, was 1.24 (1.03 < RR < 1.49). We conclude that the oxygen desaturation test combined with the 50-ml water swallow test is suitable as a screening test to identify all acute stroke patients at risk of aspiration for further evaluation and management.