Cervical VEMP threshold response curve in the identification of Ménière's disease.

ABSTRACT

PURPOSE: To investigate the sensitivity/specificity of a shift upward in the most sensitive frequency of the cervical vestibular evoked myogenic potential (cVEMP) threshold-response curve in the identification of Ménière's disease (MD). A secondary purpose was to investigate the clinical characteristics that had an impact on the sensitivity/specificity and to adjust the criteria for a positive shift upward in the cVEMP curve to maximize performance of the test. RESEARCH DESIGN: A retrospective review of patients diagnosed with MD and those without MD. STUDY SAMPLE Two hundred ninety-four patients met the inclusion criteria of symptom complaints of spontaneous events of vertigo and a full vestibular and balance evaluation with cVEMP threshold-response curve testing. Two hundred six of these patients were diagnosed with MD, and 88 patients were determined to be non-MD. DATA COLLECTION AND ANALYSIS: Review of the patients' medical records was used to extract data on the results of the cVEMP curve, age, gender, duration from time of onset of spontaneous events, pure tone average from hearing test, and water caloric asymmetry. Student's t-test, χ² test, receiver operating characteristic (ROC) curve with area under the curve (AUC), Pearson correlation coefficient, and sensitivity/specificity from 2 × 2 tables were all used in the analysis. RESULTS: Basic sensitivity/specificity for a shift upward in the most sensitive frequency to 1000 Hz in the cVEMP threshold-response curve was 0.47/0.64 respectively. Clinical characteristics that were found to have a significant impact on the sensitivity/specificity were age equal to or above 60 yr and a caloric asymmetry ≥25%. Various combinations of age and caloric with the requirement of a shift upward in the cVEMP curve most sensitive frequency to 1000 Hz resulted in significant but modest improvements in sensitivity/specificity. However, the overall performance was not shown acceptable for routine clinical use with maximum sensitivity at 0.73. Therefore, placing an emphasis on specificity over sensitivity results showed specificity of 0.95 for those under 60 yr and 0.90 for those 60 yr of age or older with sensitivity at 0.20, but only in the context of a ≥25% caloric asymmetry. CONCLUSIONS: We recommend the use of the shift upward to 1000 Hz with a caloric asymmetry as the clinical protocol to maximize the use of the cVEMP threshold-response curve for assistance in the identification of MD, in the context of a ≥25% caloric asymmetry. This implies that if the test is negative no interpretation of identification of MD can be made. If the test is positive the results can be used to increase the argument for MD since the probability of the result being a false positive is only 5-10%.