[Analysis with the logistic regression mode to the audiologic and vestibular functions in Meniere's disease]

To investigate the clinical features, characteristic of cochlear and vestibular dysfunction in Meniere's disease (MD) by using the multiple factor Logistic regression analysis.

The clinical data of 36 patients with the diagnosis of Meniere's disease according to 2006 Guiyang standard and 30 patients with peripheral vestibular disorders were investigated. All subjects received audiologic and vestibular assessments, including pure tone audiometry, Metz's recruitment test, electrocochleography (ECochG), auditory brainstem response (ABR), glycerol test, vestibular caloric test, head shaking nystagmus (HSN) and fukuda stepping test. The clinical features and audiovestibular tests in Meniere's disease were investigated by chi square test, and then analyzed by using the multiple factor Logistic regression mode.

(1) The fluctuating sensorineural hearing loss, the number of active symptoms, Tullio phenomenon,--SP/AP ratio, the Metz's recruitment test and the glycerol test demonstrated as the main characteristics in the discrimination between the MD group and non-MD group, the difference was statistically significant (P < 0.05). (2) The Logistic regression predictive equation for Meniere's disease was: Logit(p) = 9.443 + 3.110 X1 + 5.015 X2 + 2.506 X3 + 3.963 X4, in which the concomitant variables were curve of electrocochleography (EcochG) (X1), glycerol test (X2), the number of active symptoms (X3), the fluctuating sensorineural hearing loss (X4). The area under the ROC curve was 0.993. The percentage of correct prediction was 95.5%.

The clinical features of Meniere's disease are distinguished, combined with audiovestibular tests, which can be differentiated from other peripheral vestibular disorders. The multiple factor Logistic regression predictive equation for Meniere's disease is an auxiliary diagnosis method.