Kurtosis Adjustments for Complex Noise-Induced Hearing Loss

1719 

Professional Poster 

W Gong¹

¹NIOSH, N/A

Wei Gong  
NIOSH

Complex noise poses a greater risk to hearing than continuous noise alone at equivalent energy levels. This study aimed to examine the associations between a potential metric of "kurtosis" and changes in hearing thresholds and explore how kurtosis might be integrated into occupational hearing loss prevention programs to better protect workers from noise‐induced hearing loss (NIHL).

Complex noise consists of impulse or impact sounds superimposed on a background of continuous hazardous noise. It has been repeatedly shown to pose a greater risk to hearing than continuous noise alone at equivalent energy levels. The typical A-weighted, time-weighted average noise measurement protocols used in industrial hygiene do not account for the additional hazard associated with complex noise. Kurtosis is a statistical metric which evaluates the "peakedness" of a signal which shows promise for evaluating the risk of complex noise‐induced hearing loss (NIHL). However, the methods for applying the kurtosis metric to noise measurements in order to accurately reflect the risk of hearing loss from complex noise remain unclear.

"This cross-sectional study analyzed data collected from 1520 non-noise-exposed workers and 2400 noise-exposed workers. The noise-exposed workers were from 13 industries and had 64 different job titles. A-weighted sound pressure level normalized to an 8-h working day (LAeq,8h), both arithmetic and geometric average kurtosis values (β_ARI, β_GEO) were calculated from personal noise exposure waveform recordings. Their noise exposure levels, measured as an LAeq,8h, ranged from 70 to 102 dBA.

Expected thresholds were calculated from the thresholds of the non-noise-exposed sample. The 50% point of the cumulative distribution for age-and sex-matched individuals in the non-exposed sample was subtracted from each threshold of the noise-exposed sample to estimate thresholds reflecting noise‐induced permanent threshold shift (NIPTS). Two sets of frequencies were analyzed: (a) NIPTS1234 (1, 2, 3, and 4 kHz) and (b) NIPTS346 (3, 4, and 6 kHz). Using a multiple regression model, the coefficients of kurtosis adjustment (λ) were calculated. Then the kurtosis‐adjusted LAeq,8h (L′Aeq,8h) was calculated.

LAeq,8h from the noise-exposed workers ranged from 70 to 102 dBA. Noise‐exposed workers were grouped into four categories based on A-weighted noise levels prior to kurtosis adjustment: 70–79, 80–87, 88–92, and 93–102 dBA. The average noise exposure levels for each category were 76, 85, 90, and 97 dBA. To identify variables that would be included in later linear regression model analysis, a general‐purpose procedure for regression (REG procedure) was employed. The average actual NIPTS1234 or NIPTS346 served as the dependent variable in the analysis. Independent variables included age, exposure duration in years, LAeq,8h, and log10(βARI/3) or log10(βGEO/3) to evaluate the relationship with hearing threshold levels. Multiple linear regression was used to examine the effect of kurtosis on NIPTS1234 or NIPTS346.

Statistical analyses were performed with SAS (release 9.4, SAS Institute Inc., Cary, NC, USA). A statistical significance level of p < 0.05 was adopted for all analyses."

"Result:
41.8% of workers were exposed to noise levels of 80-87 dBA and 32.3% to 88-92 dBA. Male participants generally had poorer thresholds than female participants, which is consistent with the results from non-noise exposed participants. Broader notches that extended towards 4 kHz and formed a U-shape were observed for the male workers, while sharp V-shaped audiometric notches at 6 kHz were more commonly observed for female workers.

Among workers in the same exposure categories, the mean NIPTS1234 and NIPTS346 increased as the βARI and βGEO values increased. The multiple linear regression analysis confirmed that both LAeq,8h and average kurtosis values strongly predict NIHL.

The adjustment coefficients of kurtosis ranged from 6.1 and 7.9, when workers are exposed to average noise levels between 80 dBA and 92 dBA. However, there is less certainty about how kurtosis affects NIHL at lower (70–79 dBA) and higher exposure levels (93–102 dBA). After applying kurtosis-adjusted LAeq,8h values (L'Aeq,8h), the estimated adjusted NIPTS are closer to the actual NIPTS. However, none of the four adjusted NIPTS are perfectly matched to the actual NIPTS, especially at frequencies of 2 kHz or below. The arithmetic average kurtosis coefficients (λ) of 6.1 resulted in the most reasonable estimate of actual NIPTS.

When arithmetic average kurtosis coefficient (λ) of 6.1 is applied to adjust measured LAeq,8h values, the number of participants exposed to 85 dBA and above increased from 1,311 to 1,757. Similarly, the number of participants exposed to 90 dBA and above increased from 393 to 1,379. This means that an additional 446 participants (a 25% increase) would need to be included in the hearing loss protection program (HLPP) when following the NIOSH recommendations. Alternatively, an additional 1,386 (78% increase) would need to be included in a hearing loss conservation program if adhering to OSHA PEL of 90 dBA TWA.

Conclusions
These findings suggest that kurtosis plays a significant role in hearing loss when workers are exposed to average noise levels of 85 dBA and 90 dBA (i.e., NIOSH REL and OSHA PEL, respectively). This highlights the importance of considering the enrollment process for complex noise-exposed workers into hearing conservation programs. Risk assessment analyses could help further examine the excess risk of hearing loss using kurtosis as a predictor when workers are exposed to complex noise. In addition, more data for noise exposure levels ranging from 70-79 dBA would be beneficial to enhance understanding within this range."

Noise and Hearing Loss Prevention