Abstract
BACKGROUND: Benign paroxysmal positional vertigo (BPPV) is the most common vestibular disease. Both therapeutic and diagnostic benefits with mechanical rotation chairs (MRCs) for management of BPPV have been reported. No previous studies have compared diagnostics in MRCs to traditional diagnostics on an examination bed. OBJECTIVE: To investigate the agreement between BPPV diagnostics performed with an MRC and traditional diagnostics on an examination bed. Secondary objectives were to (1) examine if the two test modalities differ in diagnostic properties when diagnosing largely untreated patients referred from general practitioners (uncomplicated BPPV) compared to patients referred from private ENTs (complicated BPPV) and (2) examine whether impaired participant cooperation during Manual Diagnostics (MDs) alters agreement, sensitivity and specificity. METHOD: Prospective randomized clinical trial in which patients with a case history of BPPV were recruited by referrals from general practitioners, otorhinolaryngologists and other hospital departments in the Northern Region of Denmark. Participants underwent diagnostic examinations twice: once by traditional MDs on an examination bed and once with an MRC. Initial examiner and order of test modality were randomized. Examiners were blinded to each other's findings. RESULTS: When testing the ability to diagnose BPPV, agreement between the two test modalities, was 0.83, Cohen's kappa 0.66. When comparing MD diagnostics to MRC diagnostics (set as gold standard diagnostics following test result interpretation), values for MDs were: sensitivity 71%, specificity 98%, Negative Predictive Value 73%, and Positive Predictive Value 97%. Agreement regarding BPPV subtype classification was found to be 0.71, and Cohen's kappa 0.58. Agreement when isolating the diagnosis to posterior canalolithiasis (p-CAN) was 0.89, Cohen's kappa 0.78. CONCLUSION: Diagnostics, aided by an MRC, are more sensitive than traditional manual BPPV diagnostics. The overall agreement level between test modalities was found to be weak to moderate. When isolating diagnostics to p-CAN, the level of agreement increased to "moderate-strong." Results also showed higher agreement between test modalities and a significantly higher negative predictive value for MDs when examining patients referred directly from General Practitioners following no- or a single treatment attempt. The diagnostic properties of MDs improved in patients with a higher degree of cooperation.