Differential Diagnosis of Vertigo in the Emergency Department: A Prospective Validation Study of the STANDING Algorithm.

OBJECTIVE: We investigated the reliability and accuracy of a bedside diagnostic algorithm for patients presenting with vertigo/unsteadiness to the emergency department.

METHODS: We enrolled consecutive adult patients presenting with vertigo/unsteadiness at a tertiary hospital. STANDING, the acronym for the four-step algorithm we have previously described, based on nystagmus observation and well-known diagnostic maneuvers includes (1) the discrimination between S pon TA neous and positional nystagmus, (2) the evaluation of the N ystagmus D irection, (3) the head I mpulse test, and (4) the evaluation of equilibrium (sta N din G ). Reliability of each step was analyzed by Fleiss' K calculation. The reference standard (central vertigo) was a composite of brain disease including stroke, demyelinating disease, neoplasm, or other brain disease diagnosed by initial imaging or during 3-month follow-up.

RESULTS: Three hundred and fifty-two patients were included. The incidence of central vertigo was 11.4% [95% confidence interval (CI) 8.2-15.2%]. The leading cause was ischemic stroke (70%). The STANDING showed a good reliability (overall Fleiss K 0.83), the second step showing the highest (0.95), and the third step the lowest (0.74) agreement. The overall accuracy of the algorithm was 88% (95% CI 85-88%), showing high sensitivity (95%, 95% CI 83-99%) and specificity (87%, 95% CI 85-87%), very high-negative predictive value (99%, 95% CI 97-100%), and a positive predictive value of 48% (95% CI 41-50%) for central vertigo.

CONCLUSION: Using the STANDING algorithm, non-sub-specialists achieved good reliability and high accuracy in excluding stroke and other threatening causes of vertigo/unsteadiness.