A study of the effects of cochlear loss on the auditory brainstem response (ABR) specificity and false positive rate in retrocochlear assessment.

The auditory brainstem response (ABR) test is widely used as a screen for retrocochlear dysfunction in individuals suffering idiopathic auditory and vestibular disturbance. Although its sensitivity for lesions of the lower brainstem is well established, hearing loss is known to have significant effects upon the test's specificity and false positive rate. This study analysed the effects of aspects of cochlear hearing loss including level, slope and general audiogram shape on these properties of the ABR test in a large clinical group. The study comprised 306 patients (153 male, 153 female) referred for ABR screening to exclude retrocochlear dysfunction. Of 566 ears meeting the inclusion criteria, 85 (15 per cent) failed the test. On follow-up eight ears (3 per cent) were confirmed to have some form of retrocochlear abnormality, seven of which were pathological. All had abnormal ABR results. No cases of retrocochlear abnormality were found in the ABR pass group. In those patients diagnosed as free of retrocochlear problems, the importance of the I-V interval as a primary index of function was confirmed. It displayed a high specificity (>90 per cent) which was largely independent of the level or slope of high-frequency hearing loss. In contrast wave V specificity decreased with increasing loss and increasing slope. Its value as a functional index was limited with losses greater than 70 dB HL in whom specificity fell below 70 per cent. The study concluded that, despite reducing interval availability above 60 dB HL, wave V latency cannot be used as an alternative index because it displays weakening specificity over this range of loss. With the problems inherent in applying latency corrections, the author advocates the use of improved procedures for wave I identification in these cases.