[Distortion product otoacoustic emissions for the assessment of auditory sensitivity].

Distortion product otoacoustic emissions (DPOAEs) were evaluated in 494 normal and 506 cochlear-impaired human ears, to determine whether DPOAEs depend on factors such as background noise, the shape of the pure tone audiogram, sex and aging, and whether a DPOAE test can perform well in distinguishing normal-hearing from hearing-impaired ears. The amplitudes of DPOAEs were measured at the frequency of 2f1-f2 (f1 < f2, f2/f1 = 1.22, f2 at 1, 2 and 4 kHz) using as stimuli two pure tones at level of 70 dB from an ILO92 Otoacoustic Emission Analyzer. The correlation coefficients between the DPOAE level and the auditory threshold decreased as the background noise levels at 1 kHz and 2 kHz increased. Therefore, it appeared that ears with large background noise levels would be inadequate for the study of DPOAEs predicting the hearing state. The sensitivity (normal-hearing ears identified as normal hearing) and the specificity (hearing-impaired ears identified as hearing impaired) at the equal-sensitivity-specificity condition were 80.7-86.7% at 1, 2 and 4 kHz, and the areas under the receiver operating characteristic (ROC) curves, which were used to estimate the test performance, were 0.88 for 1 kHz, 0.91 for 2 kHz and 0.92 for 4 kHz. Since these results suggest that a DPOAE can be used as a reliable technique for objective auditory tests, it is thought that actual values (DPOAE level: 4.3 dB at 1 kHz, 5.0 dB at 2 kHz and 2.9 dB at 4 kHz) of false-positive (hearing-impaired ears identified as normal hearing) rates corresponding to 5% can be used in clinical evaluation to separate normal hearing from hearing-impaired ears. There was, however, a significant age effect at 4 kHz on DPOAEs in the ears with the same pure tone hearing thresholds, and the areas of the ROC curves in subjects ranging from 10 to 29 years old were larger than in subjects over 50 years (1 kHz: 0.88 to 0.94 versus 0.83 to 0.84, 2 kHz: 0.95 versus 0.89, 4 kHz: 0.95 to 0.96 versus 0.88 to 0.89). Therefore, it is thought that age-adjusted norms may be necessary for the accurate interpretation of DPOAE results.