Validity of Estimating the 3 kHz Audiometric Threshold by Averaging the Thresholds at 2 and 4 kHz in Cases of Noise-Induced Hearing Loss

• Published in: Archives of occupational health
• Main Authors: Philippe Henri, DeJonckere, Jean, Lebacq
• Format: Journal Article
• Language: English
• Published: 12.02.2022
• ISSN: 2588-3070; 2588-3690
• DOI: 10.18502/aoh.v6i1.8665

Background: Hearing thresholds at 3000 Hz are generally not measured in routine clinical audiometry. However, for purposes other than clinical diagnosis, the threshold at 3 kHz has many applications, in epidemiological studies in the field of occupational health and medicine, as well as in (medicolegal) quantification of physical impairment due to hearing loss, particularly noise-induced hearing loss (NIHL). The present study addressed the validity of estimating, in the case of NIHL, the 3 kHz-audiometric thresholds by averaging the thresholds at 2 and 4 kHz. Methods: All 200 investigated subjects (400 ears) had a well-documented noise exposure, moderate to severe NIHL, and underwent, as they were claiming for compensation, a detailed medicolegal audiological investigation, including beside pure tone audiometry, electrophysiological objective frequency-specific threshold definition using cortical evoked response audiometry (CERA) and auditory steady-state response (ASSR). Results: The study results showed a good correlation between the 2-4 kHz interpolation and the actual 3 kHz threshold; the error may be around 2 dB on average. However, in individual cases, the results demonstrated that the error due to interpolation exceeds 5 dB HL in about one-quarter of the cases. This error is predictable; the larger the 2- 4 KHz difference (which reflects the steepness of the left slope of the audiometric notch), the larger the error (on either side) made by interpolating. Conclusion: For epidemiological studies with large amounts of data, the interpolated threshold (average of 2 and 4 KHz) may be considered as a valid estimate of the true value of the 3 KHz threshold. More caution is required in individual cases: the error due to interpolation exceeds 5 dB HL in about one-quarter of the cases, but this error is predictable.