Distortion product otoacoustic emissions (DPOAE), permanent threshold shifts (PTS) and outer hair cell (OHC) losses were analyzed in a population of 187 noise-exposed chinchillas to determine the predictive accuracy (sensitivity and specificity) of the DPOAE for PTS and OHC loss. Auditory evoked potentials (AEP) recorded from the inferior colliculus of the brainstem were used to estimate hearing thresholds and surface preparation histology Was used to determine sensory cell loss. The overlapping cumulative distributions and high variability in emission responses for both PTS and OHC loss made it difficult to predict AEP threshold and OHC loss from DPOAE level measurements alone. Using a strict criterion (i.e. emissions better than the 5th percentile of the preexposure DPOAE level, and PTSless than or equal to5 dB or OHC loss less than or equal to5%), it was found that the postexposure DPOAE level could be used with reasonable confidence to determine if the status of peripheral auditory system was either normal (i.e. PTS:5 5 dB) or abnormal (PTS>30 dB or OHC loss >40%). However, the high variability of individual DPOAE responses resulted in a broad region of 'uncertainty' (i.e. 5<PTS <= 30 dB and 5%<OHC lossless than or equal to40%) making it difficult in the chinchilla model to use the postexposure DPOAE level with confidence to predict in individual subjects the amount of PTS or OHC loss. Our results also indicate that significant reductions in the amplitude of the DPOAE are related primarily to a systematic loss of OHCs, and that a postexposure DPOAE level less than or equal to10 dB SPL, obtained with a low frequency primary level of 65 dB SPL, represents a criterion value which can serve as an indication of significant OHC loss (greater than or equal to50%) or PTS (greater than or equal to35 dB) in noise-exposed chinchillas. Based on an exponential regression analysis of individual subjects, correlations were higher for PTS/DPOAE than for OHC loss/DPOAE. (C) 2003 Elsevier B.V. All rights reserved.