Specificity of auditory brainstem response audiometry criteria in acoustic neuroma screening as a function of deviations of reference values in patients with cochlear hearing loss.

On the basis of 79 patients with cochlear hearing loss, the statistical distribution of two criteria commonly used in auditory brainstem response audiometry (ABR) was analyzed: the interaural V latency difference (ILD V) and the interaural difference of I-V interpeak interval (ID I-V). The distribution of both criteria was Gaussian. By evaluating their standard deviations the percentages of statistical false-positives were estimated. The estimated results were 24% false-positive findings using the decision criterion ILD V greater than 0.2 ms and 5.4% false-positive findings using ID I-V greater than 0.3 ms. This corresponds closely to the actual false-positive ABR rates obtained in this sample: 21.5% and 6.3%, respectively. In a separate series of 301 unselected cases with asymmetric sensorineural hearing loss, 29 ABRs were suspect for retrocochlear pathology. In 20 patients, ABRs were absent due to severe hearing loss. Retrocochlear pathology could be confirmed in only 2 cases (both from the group with ABR present). Thus, 47 ABRs (15.7% of 299) were false-positive.