Sensitivity and Specificity of the Water-Swallowing test

Author/ Year/

Country/ N



Osawa et al.

Population: Stroke
Protocol: Patients with suspected dysphagia were assessed by videofluorography (VF). Results of this test were compared to a clinical assessment performed by a blinded SLP.

Scoring for clinical assessment: Choking or gargling voice or decrease in oxygen saturation (Sp02). Sp02 decrease of greater than 2% was classified as abnormal.

  1. The number of aspirations experienced by patients increased as the amount of liquid increased (17 occurrences at 5mL to 70 occurrences at 60mL).
  2. The sensitivity of the WST was greatest at 60mL (55.7%), while the specificity of the WST was greatest at 10mL (93.2%).
  3. Positive predictive value was greatest at 60mL (86.7%) and negative predictive value was greatest at 5mL (91.4%).
  4. A specific volume of water (3mL, 5mL, 10mL, 30mL, 60mL) did not stand out as most appropriate across all psychometric properties of the WST.

Kopey et al.

Population: Stroke.
Protocol: A retrospective review of patients admitted to an acute rehabilitation unit who were alert, and non-dysarthric. These patients received a 3-sip test on day 2 following admission. A portion of the patients underwent additional VMBS due to continued suspicion of dysphagia. The sensitivity, specificity, positive and negative predictive values (PPV, NPV) of the sip test compared with clinically relevant dysphagia, defined as VMBS findings that precipitated a diet change (i.e. minced or pureed solids) were calculated.

  1. 206 patients passed the 3-sip test.
  2. 67 (32.5%) patients had clinically significant dysphagia.
  3. The reported sensitivity and specificity were 20.8% and 98.7%, respectively.
  4. The PPV and NPV were 88.2% and 72.3%, respectively.
  5. A low (<60) Functional Independence Measure score was also predictive of clinically relevant dysphagia.


Nishiwaki et al. (2005)


Population: Stroke.
Protocol: Symptoms of oromotor function were evaluation. The water swallowing test (using 30 mL of water), saliva swallowing test and a VMBS examination were also conducted. Factor analysis was used to predict dysphagia in patients with stroke.

  1. Cough/voice change in the water swallowing test was the only variable that was significantly associated with aspiration on VMBS examination, with a sensitivity of 72% and a specificity of 67%.

Wu et al.

Population: Stroke.
Protocol: Outpatients with suspected dysphagia underwent a 100 mL water-swallowing test. Signs of choking or a wet sounding voice within 1 minute of completing the test were considered evidence of an abnormal swallow. Swallowing speed (< 10 mL/s or ≥ 10 mL/s) was also recorded. The results were compared to a VMBS study.

  1. 55 patients were identified as having some form of swallowing dysfunction on VMBS examination.
  2. An abnormal swallowing speed was detected in 49 of 55 patients; of these, 47 demonstrated abnormal swallowing.
  3. The sensitivity and specificity of the test was 85.5% and 50%, respectively.
  4. 33 patients either aspirated or demonstrated penetration on VMBS study. Of these 11 choked on the WST, while 3 patients with a normal VMBS result, choked on the WST.

The sensitivity and specificity of the test was 47.8% and 91.7%, respectively.

Chong et al.

Population: Stroke.
Protocol: Patients with suspected dysphagia, 65 years or older, received a WST (patients were asked to drink 50 mL of water in 10mL aliquots), an oxygen desaturation test (desaturation of ≥ 2% was considered clinically significant) and a fiberoptic endoscopic evaluation of swallowing (FEES), where episodes of aspiration or penetration of various food consistencies were noted. The consistency or results between the tests were compared.

  1. The WST had a sensitivity of 79.4% and specificity of 62.5% for the detection of aspiration, with a positive predictive value (PPV) of 81.8% and a negative predictive value (NPV) of 58.8%.
  2. The oxygen desaturation test had a sensitivity of 55.9% and a specificity of 100% with PPV of 100% and NPV of 51.6%.
  3. When both tests were combined, a sensitivity of 94.1% and a specificity of 62.5% were attained, with PPV of 84.2% and NPV of 83.3%.
  4. Using the clinical assessment test, three aspirators were detected who would otherwise have been missed if they were assessed with the water swallow test using thin fluids alone.

Lim et al.


Population: Stroke.
Protocol: Patients received a 50 mL WST (in 10 mL aliquots) and a fiberoptic endoscopic evaluation of swallowing examination. Patients also received an oxygen desaturation test.

  1. The 50-ml WST had a sensitivity of 84.6% and specificity of 75.0%.
  2. The oxygen desaturation test had a sensitivity of 76.9% and specificity of 83.3%.
  3. When the two tests were combined into one test called "bedside aspiration," the sensitivity rose to 100% with a specificity of 70.8%.

Garon et al.


Population: Stroke (50%).
Protocol: All patients were asked to drink 3 oz. of water from a cup without interruption. Any coughing or throat clearing was indicative of an abnormal WST. The results of the WST were compared to the VMBS study.

  1. Fifty-four patients aspirated. Of these, only nineteen (35%) coughed, leaving thirty-five, or 65%, who were not identified by the WST.
  2. There was a significant difference between patients identified by the WST compared with the VMBS (p<0.005).
  3. The sensitivity and specificity of the tool to identify confirmed aspirators were 54% and 79%, respectively.

DePippo et al., (1992)
N= 44

Population: Stroke.
Protocol: Patients were given 3 oz of water from a cup and asked to drink without interruption. Coughing for up to 1 minute after the test or a wet-horse voice was considered abnormal. Patients also received a VMBS study and the results from the 2 tests were compared.

  1. 27 patients had an abnormal WST. 
  2. 20 patients were proven aspirators on the basis of the VMBS study. The 3 oz. WST identified 16/20 aspirators. 11 patients without evidence of aspiration on VMBS had an abnormal WST.
  3. The sensitivity and specificity of the WST was 76% and 59%, respectively.

Note: WST=Water Swallowing Test; VMBS=Videofluoroscopic Modified Barium Swallow.