To identify the predictive capabilities of caloric and rotational testing to distinguish peripheral from nonperipheral vestibular causes of vertigo.
Retrospective case review.
Academic tertiary care vestibular function test center.
Two hundred dizzy patients (132 with and 68 without suspected peripheral vestibular dizziness) evaluated with bithermal binaural caloric and sinusoidal and step-velocity rotary chair (RC) tests.
Caloric and RC tests.
1) Receiver operating characteristic (ROC) analysis of individual caloric and rotary parameters for area under the curve (AUC) as indication of predictive value, and 2) logistic regression analysis of parameter combinations for identification of optimal test battery for predicting peripheral vestibular dysfunction
Analysis of the individual receiver operating characteristic (ROC) curves for each caloric and RC test parameter for prediction of peripheral vestibulopathy revealed that the best overall predictive parameter was percentage of caloric weakness (CW; AUC-ROC, 0.73; 95% confidence interval, 0.66-0.80). Using 29.5% as the optimal cutoff point for CW, a specificity of 84% and sensitivity of 55% were obtained. Because of the reduced AUC for the RC parameters, no optimal cutoff points were chosen. Using a logistic regression model in predicting patients with peripheral vestibulopathy demonstrated that either comprehensive caloric and RC testing or the combination of caloric and 0.025 Hz, 0.5 Hz, and time-constant rotational testing yielded the highest predictive value (AUC-ROC, 0.79)
Calculation of CW was the most promising single marker for identifying peripheral from nonperipheral vestibulopathic patients. Individual RC parameters showed limited predictive value to differentiate peripheral from nonperipheral vestibular dysfunction. However, the combination of subtests yielded an AUC-ROC only slightly less than caloric testing. The combination of caloric and RC testing revealed the strongest predictive capabilities for identifying peripheral vestibular injury. Furthermore, the combination of an abbreviated form of RC testing with caloric testing yielded nearly identical results as full caloric and rotational testing.
Department of Otolaryngology, Washington University School of Medicine, St. Louis, Missouri, U.S.A.
Address correspondence and reprint requests to Joel A. Goebel, M.D., Department of Otolaryngology, Head and Neck Surgery, Washington University School of Medicine, 660 South Euclid Ave., Campus Box 8115, St. Louis, MO; E-mail: email@example.com