In a longitudinal study of 76 noise-exposed industrial workers, 34 (68 ears) of them were available for final evaluation after 9 years of observation. Distortion-product otoacoustic emissions (DP) at 65/55 dB SPL and pure-tone audiometry were used to detect noise-induced inner ear changes. Repeated measures analyses of variance were made on the subjects and in a control group, whereas significant threshold shifts (STS) and significant emission shifts (SES) were calculated. Both hearing thresholds and DP showed a high-frequency deterioration, but there was absence of statistical correlation between elevation of hearing thresholds and decrease of DP. There was no clear pattern between individual changes in audiometric thresholds and DP, and all combinations were present: ears with only STS, ears with only SES, ears with both STS and SES and ears with absence of STS and SES. Audiometric changes were maximal at 4 and 6 kHz and DP changes at 2 and 3 kHz. Since significant individual emission changes do not necessarily follow the same pattern as the group-averaged results, the use of DP for monitoring of individual subjects is not advised. However, its use in conjunction with audiometry in hearing conservation programs is highly recommended.