The Effectiveness of Applying Different Permissible Exposure Limits
Transcripción
The Effectiveness of Applying Different Permissible Exposure Limits
Balachandar S. SAYAPATHI, et al.: Review of Different PELs for Preserving HTL J Occup Health 2014; 56: 1–11 1 Journal of Occupational Health Review The Effectiveness of Applying Different Permissible Exposure Limits in Preserving the Hearing Threshold Level: A Systematic Review Balachandar S. Sayapathi1, Anselm Ting Su1,2 and David Koh3,4 Centre for Occupational and Environmental Health, University of Malaya, Malaysia, 2Department of Hygiene, School of Medicine, Wakayama Medical University, Japan, 3Occupational Health and Medicine, PAPRSB Institute of Health Sciences, University of Brunei Darussalam and 4SSH School of Public Health, National University of Singapore, Singapore 1 Abstract: The Effectiveness of Applying Different Permissible Exposure Limits in Preserving the Hearing Threshold Level: A Systematic Review; Balachandar S. SAYAPATHI, et al. Centre for Occupational and Environmental Health, University of Malaya, Malaysia—Objectives: A systematic review was conducted to identify the effectiveness of different permissible exposure limits in preserving the hearing threshold level. This review compared the limits of the US National Institute of Occupational Safety and Health with those of the US Occupational Safety and Health Administration. The prevalence of occupational noiseinduced hearing loss is on an increasing trend globally. This review was performed to reduce the prevalence of noise-induced hearing loss. Methods: We searched 3 major databases, i.e., PubMed, Embase and Lippincott Williams & Wilkins Journals@Ovid, for studies published up until 1May 2013 without language restrictions. All study designs were included in this review. The studies were identified and retrieved by two independent authors. Results: Of 118 titles scanned, 14 duplicates were removed, and a total of 13 abstracts from all three databases were identified for full-text retrieval. From the full text, eight articles met the inclusion criteria for this systematic review. These articles showed acceptable quality based on our scoring system. Most of the studies indicated that temporary threshold shifts were much lower when subjects were exposed to a noise level of 85 dBA or lower. Conclusions: There were more threshold shifts in subjects adopting 90 dBA compared with 85 dBA. These temporary threshold shifts may progress to permanent shifts over time. Action curtailing noise exposure among employees would be taken Received Jun 18, 2013; Accepted Sept 27, 2013 Published online in J-STAGE Nov 22, 2013 Correspondence to: B.S. Sayapathi, Centre for Occupational and Environmental Health, University of Malaya, 50603 Kuala Lumpur, Malaysia (e-mail: [email protected]) earlier on adoption of 85 dBA as the permissible exposure limit, and hence prevalence of noise-induced hearing loss may be reduced. (J Occup Health 2014; 56: 1–11) Key words: Effects 85 or 90dBA, Noise, Threshold shift The progression to noise-induced hearing loss is reliant on a few factors such as frequency, intensity and duration of noise exposure1, 2). Cumulative and repetitive exposure to high-intensity noise levels and long duration of contact with noise levels beyond the permissible exposure limit may aggravate the development of this occupational disease 3). Noise-induced hearing loss is a slow and progressive disease. The duration needed for an employee to develop this irremediable disease depends on several factors. Besides being influenced by high noise levels, the degree of hearing loss of an employee is dependent on the institution of a hearing conservation program in an industry. This kind of program utilizes various approaches to curb noise exposure. The gold standard for measuring hearing loss is use of an audiometer. The pure-tone audiometer is capable of detecting hearing loss at various frequencies ranging from 250 to 8,000 Hertz (Hz)4). Noiseinduced hearing loss shows few characteristic features in an audiogram. Permanent shifts will be noted relative to baseline findings of the employees5). These shifts usually occur in both ears and involve frequencies ranging from 3,000 to 6,000 Hz1, 3). There will be a dip at these frequencies in early phases of the disease with recovery at 8,000 Hz. In long-standing disease, the lower frequencies would be involved too. These features distinguish it from a condition known as presbycusis, in which no recovery at 8,000 Hz is 2 noted. According to the US Occupational Safety and Health Administration (OSHA), which is a regulatory body, the permissible exposure limit for noise is 90 dBA6). An employee should not be exposed beyond this level for more than eight hours. As for the National Institute of Occupational Safety and Health (NIOSH), USA, which is a research body, the exposure limit recommended is 85 dBA for the same duration in hours6). There has been an increase in the prevalence of occupational noise-induced hearing loss globally. The occurrence of this occupational disease has doubled from 120 million in 1995 to 250 million in 20047). Millions of workers in the US are exposed to noise levels above the permissible exposure limit, 90 dBA. Besides the US, Malaysia is also adopting 90 dBA8) as the permissible exposure limit. There were a total of 663 cases of occupational diseases investigated in Malaysia for the year 2010. From this total, around 70% of the cases were diagnosed as having noise-induced hearing loss, making it the most common occupational disease9). There is clear evidence associating noise intensity and hearing loss. A hearing conservation program is instituted once the noise level is above the action level8), which is a noise dose of half the permissible exposure limit. We systematically reviewed the evidence in order to compare the effects of 85 and 90 dBA on temporary threshold shifts to reduce the prevalence of noise-induced hearing loss. Methods We considered observational studies (trials and any other comparative design). We used medical subject headings with the keywords “Effects 85 or 90 dBA AND Noise AND Threshold shift” for all the databases searched. The outcome measures were temporary threshold shifts on an audiogram. Search strategy We searched 3 major databases, i.e., PubMed, E m b a s e a n d L i p p i n c o t t Wi l l i a m s & Wi l k i n s Journals@Ovid, for studies published up until 1 May 2013. We applied no language restrictions. There were a few inclusion criteria that needed to be achieved for the papers to be included in this systematic review. The studies were required to include exposure information regarding the noise source, such as industrial or experimental noise, noise levels and outcome of threshold levels. Initially the titles of the papers were reviewed for applicability in this review. Later, the abstracts were studied to find out further whether the studies fulfilled the requirements for the review. Finally, the full texts of the papers were appraised to confirm that the contents met the require- J Occup Health, Vol. 56, 2014 ments of this review. On evaluation of the full text, once a paper was accepted for this review, information was gathered from the paper such as the names of authors, year of publication, country where the study was conducted, the setting (experimental or industrial), number of participants or subjects, exposure information including the sources of noise, duration of exposure and usage of hearing protection devices, the noise levels in dB and finally the outcome of the hearing threshold level in the subjects. However, unpublished works such as conference proceedings were not included in this assessment. Quality assessment The scoring system used to assess the quality of a paper was as depicted in Table 1. This scoring system was a modified form of that of Su et al.10). The standards included in this scoring system were divided into three major subheadings, i.e., population, exposure and outcome. Under the population subheading, probability for bias among participants, nonrespondent bias and confounding variables were taken into account, whereas under the exposure subheading, control of confounding variables and use of effective measurement tools to quantify exposure were recorded. Finally, the outcome subheading, use of effective measurement tools to quantify outcome and appositeness of blinding were justified for the score. A paper in this review was categorized as either acceptable or poor quality based on the score gained. Acceptable quality of a paper would mean that a score of 2 was attained for each measurement of confounding variables, effective measurement to quantify exposure and outcome. If any score was below 2 for these standards, a score of at least 2 was necessary for probability for bias on selection of participants or subjects and for avoidance of nonrespondent bias standards. Data extraction We (Sayapathi BS and Su AT) approved and retrieved studies independently based on the study’s inclusion criteria. We then extracted data from the included studies. These data later were gathered depending on the country in which the study was done, setting, number of subjects, exposure information, noise level and outcomes. Results After a thorough search by titles, we identified a total of 47 pertinent articles through the PubMed database and 47 articles from Lippincott Williams & Wilkins Journals@Ovid. The search of the Embase database identified 24 articles. Fourteen duplicates found in Lippincott Williams & Wilkins Journals@Ovid were then eliminated. There were a Balachandar S. SAYAPATHI, et al.: Review of Different PELs for Preserving HTL 3 Table 1. Scoring system for ranking the quality of papers Scope Description Population I) Probability for bias in selected subjects 3 2 1 Survivor bias avoided Volunteer bias avoided Selection method not reported II) Nonrespondent bias 2 1 Response rate ≥75% Response rate <75% or not reported III) Potential confounding variables between groups 2 1 NA Potential confounders measured or both groups exposed to similar confounders Potential confounders not measured No comparison group Exposure IV) Confounding exposure controlled 2 1 Exposure confounders measured or both groups exposed to similar confounders Exposure confounders not measured nor reported V) Effective measurement tool used to quantify exposure 2 1 Exposure measured and reported Exposure not measured nor reported Outcome VI) Effective measurement tool used to quantify outcome 2 1 Outcome measured and reported Outcome not measured nor reported VII) Blinding 2 1 Relevant blinding No blinding done or reported Judgment criteria There were threats to the validity of a study if criteria III), V) and VI) were scored as 1. The quality of a study was acceptable if these criteria were scores as at least 2. The scores for criteria I) and II) should be at least 2 if criterion III) is scored as NA (not available) in order for the quality of a study to be considered as acceptable. total of 36 relevant titles retrieved from the three databases. Of them, 20 relevant abstracts were obtained. Finally, a total of 13 relevant full texts were retrieved from these databases. After thorough assessment, it was found that two similar papers from Lippincott Williams & Wilkins@Ovid and three similar papers from Embase had also been retrieved from PubMed. Hence, after removing the five similar papers from these databases, there were a total of eight papers that fulfilled the requirements for this systematic review. The process of identifying these eight articles is shown in Fig. 1. All eight articles in the current review had effective measurement tools used to quantify noise exposure and an instrument to measure outcome due to noise. Only three studies had not controlled for confounding bias, namely, those of McBride et al.11), Kvaerner et al.12) and Chen and Tsai13). However, the three studies had avoided selection bias, and the response rate of respondents in these studies was at least 75%. Hence, all eight articles showed acceptable quality based on the scoring system as outlined in Table 2. Most of the studies were carried out in the US. There was also a study conducted in the east, in Taiwan by Chen 4 J Occup Health, Vol. 56, 2014 Fig. 1. Flow diagram of search strategy a Resources included Journals@Ovid Full Text, UM Library Full Text Journals & LWW, EBM Reviews-ACP Journal Club (Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews 2005, Cochrane Methodology Register, Database of Abstracts of Reviews of Effects, Health Technology Assessment, NHS Economic Evaluation Database and ACP Journal Club), Biological Abstracts, CAB Abstracts, EconLit, GEOBASE, ICONDA and Ovid MEDLINE (In-Process & Other Non-Indexed Citations, without Revisions, Daily Update, Ovid OLDMEDLINE). b Total number of papers searched for systematic review; 2 similar papers from Lippincott Williams & Wilkins@Ovid were retrieved from PubMed; 3 similar papers from Embase were retrieved from PubMed. and Tsai13). In Europe, there were two studies, which were conducted in Denmark and Norway, while one study was conducted in the southwest Pacific, i.e., New Zealand. Therefore, in this review most of the regions of the world were covered with respect to the effects of different noise levels on hearing threshold level, thereby eliminating effects of ethnic variability. A summary of studies is shown in Table 3. Stephenson et al. 14) conducted an experimental study to identify the minimum noise intensity level that would result in an asymptotic temporary thresh- old shift. Below this level, one may be exposed to noise for an unlimited period of time, since continuous exposure would not result in permanent damage to hearing. This study was conducted among 12 volunteers who happened to be all males. These volunteers had normal hearing threshold levels, 15 dB or below at each frequency tested, before the study was conducted. They were then exposed to noise levels of 65, 70, 75, 80 and 85 dBA in a room using a random noise generator. They were exposed to different noise levels for 24 hours with a one-week 5 Balachandar S. SAYAPATHI, et al.: Review of Different PELs for Preserving HTL Table 2. Quality assessment of included papers in this review Paper Stephenson et al., 1980 Mc Bride et al., 2003 Mills et al., 1983 Yates et al., 1976 Kvaerner et al., 1995 Chen and Tsai, 2003 Rubak et al., 2006 Melnick, 1977 Avoid Avoid nonselection respondent bias bias Comparable study and control groups Control for confounding bias Valid exposure measures Valid outcome measures Blinding Overall quality and validity 1 2 2 2 2 2 2 A 2 2 1 1 2 2 1 A 2 2 3 2 2 2 2 2 1 2 2 1 2 2 2 2 2 2 1 1 1 A A A 3 2 1 1 2 2 1 A 2 1 2 2 2 2 1 A 1 2 2 2 2 2 1 A A, acceptable interval. Hearing threshold levels were measured over the left ear at various intervals upon exposure and after exposure to noise. It was noted that the mean asymptotic of threshold shifts more than two minutes after exposure to noise were higher among subjects exposed to 85 dBA, 10−13 dB at 4 kHz, compared with those exposed to other noise intensities. These significant changes were also noted for other test frequencies among subjects exposed to 85 dBA, but they were not evident when subjects were exposed to less than 80 dBA. It was also found that upon exposure to 85 dBA, the subjects tended to have more threshold shifts, with the average for six frequencies being 4.4 dB and the recovery period of these shifts being prolonged. This was contrary to the results for subjects exposed to 80 dBA, in which the mean threshold shift was only 1.9 dB and the recovery period was much faster. The investigators came to the general conclusion that one may be exposed to noise levels below 80 dBA for an indefinite period of time without causing permanent changes in hearing. This study also showed that the action level should be 80 dBA for institution of a hearing conservative program and that the permissible exposure limit should preferably be 85 dBA. In New Zealand, McBride et al. 11) performed a cross-sectional study in an engineering industry. Ninety-two employees of a railway workshop took part in this study. These employees were exposed to levels either below 85 dBA, 85 to 90 dBA or more than 90 dBA in the workplace. The noise exposures of these subjects were recorded using a noise dosimeter. Baseline audiometry was performed for these employees initially, followed by post-shift audiometry using manual pure-tone audiometry. The standard for the threshold shifts was based on the NIOSH’s criteria for a recommended standard on noise exposure. Usage of hearing protection devices during the study was not reported by the authors. A total of 50% of the subjects had temporary threshold shifts upon exposure to noise levels below 85 dBA compared with 29% when exposed to noise levels between 85 and 90 dBA and 39% when exposed to noise levels beyond 90 dBA. However, there were no statistically significant differences in threshold shifts between these three groups. In this study, no changes were observed in occurrence of temporary threshold shifts regardless of adoption of different permissible exposure limits. However, the lengths of time these workers had worked in this industry were not mentioned by the authors. Moreover, the post-shift audiometry was performed a day after the baseline audiometry; noise in any industry usually fluctuates and hence, a longer duration of exposure is usually required to observe any significant changes in threshold shifts upon adopting different permissible exposure limits. Mills et al.15) conducted an experiment to identify temporary threshold shifts among students after exposure to low-frequency noises. There were 52 students; all participants in the study were male. Prior to the study, their hearing threshold levels were within the accepted range. There were primarily divided into two groups, whereby the first group of students was exposed to three octave bands centered at 63, 125 and 250 Hz at 84 dBA for a period of 24 hours, while the second was exposed at 90 dBA for a period of 8 hours. Another group of students were exposed to 90 dBA at only 63 Hz for 8 hours. Noise was gener- Experimental Industry Experimental Experimental Industry Industry Industry Experimental New Zealand, McBride et al., 2003 USA, Mills et al., 1983 USA, Yates et al., 1976 Norway, Kvarner et al., 1995 Taiwan, Chen and Tsai, 2003 Denmark, Rubak et al., 2006 USA, Melnick, 1977 Setting USA, Stephenson et al., 1980 Country, authors, date n=9 n=649 n=384 n=13 (excluded one ear) i. n1=6 at 85 dBA ii. n2=6 at 90 dBA i. n1=7−8 each at 84 dBA ii. n2=7−8 each at 90 dBA i. n1=30 ii. n2=34 iii. n3=28 n=12 Study subjects Exposure information Noise generated through loudspeakers (24 hours) All noisy trades (20 years and more) Oil refinery (8 hours) Global iron works (7 hours each day for 3 days) Noise generating equipment (7 hours) Noise generating equipment with loudspeakers (24 hours for 84 dBA; 8 hours for 90 dBA) Railway workshop (7 hours 13 minutes) General Radio Model 1,382 Random Noise Generator (24 hours) Noise source (duration of exposure) Table 3. Summary of studies on different noise intensities ≤85 dBA ≤85 dBA and>85 dBA ≤85 dBA >85 dBA ≤85 dBA and>85 dBA ≤85 dBA and>85 dBA ≤85 dBA and>85 dBA ≤85 dBA Noise level Exposure to 80 dBA: i. 9.3 dB at 4 kHz 7.2 dB at 6 kHz Exposure to 85 dBA: i. 17.8 dB at 4 kHz ii. 14.6 dB at 6 kHz i. Exposure between 80 and 84 dBA, no statistically significant change, OR 1.92 (95 % CI, 0.77 to 4.80) ii. Exposure between 85 and 89 dBA, OR 3.05 (95% CI, 1.33 to 6.99) Hearing loss (average hearing threshold level more than 25 dBA), mean noise level was 81 dBA: i. 9.6% - low frequencies (0.5 to 2 kHz) ii. 38.3% - high frequencies (3 to 6 kHz) a) Temporary threshold shifts, exposure between 85−90 dBA: i. Median=3.8 dB (range=−10.0 to 16.7 dB), p=0.001 at 4 kHz ii. Median=5.5 dB (range=−5.0 to 11.7 dB), p<0.001 at 6 kHz b) Transient evoked otoacoustic emissions showing median=0.65 (range=−0.8 to 3.3 dB), p=0.001 a) Mean temporary threshold shifts (combined frequencies): i. For 85 dBA - mean (SD)=6.26 (9.15) ii. For 90 dBA - mean (SD)=10.13 (7.45) b) No statistically significant difference in temporary threshold shifts for individual frequencies. Temporary threshold shifts: i. 84 dBA−13, 9 and 9 dB at 63, 125 and 250-Hz ii. 90 dBA−17, 14 and 18 dB at 63, 125 and 250-Hz Recovery of threshold shifts: i. 48 hours for 84 dBA ii. 12 hours for 90 dBA Temporary threshold shifts were not statistically significant among the three groups (70.4−84.9; 85.1−89.8; 90.2−104.2), x2=2.39, p=0.24 Mean asymptotic threshold shifts of various noise intensities from six frequencies (0.5−6 KHz): 85 dBA−4.4 dB 80 dBA−1.9 dB 75 dBA−0.73 dB 70 dBA−0.7 dB 65 dBA−0.06 dB Outcomes 6 J Occup Health, Vol. 56, 2014 Balachandar S. SAYAPATHI, et al.: Review of Different PELs for Preserving HTL ated in a room using noise generating equipment with loudspeakers. The threshold shifts were recorded an average of four minutes after exposure to noise among the subjects by using a Bekesy type of audiometer. The temporary threshold shifts were found to be increased among students exposed to 90 dBA by more than 10 dB for all three conditions. Among students exposed to 84 dBA, there were threshold shifts too, amounting to more than 10 dB for all three conditions; the threshold shifts increased under the 250 Hz condition but decreased in the 63 and 125 Hz settings after 12 hours of exposure. Crude estimation by comparing the threshold shifts among subjects exposed to two different noise levels, showed a more damaging effect among those exposed to noise levels of 90 dBA; that is, the subjects showed threshold shifts of 17, 14 and 18 dB for the 63, 125 and 250 Hz conditions compared with 13, 9 and 9 dB for 84 dBA. Recovery from threshold shifts took up to 48 hours for subjects exposed to 84 dBA, since the duration of exposure was longer, while those exposed to noise levels of 90 dBA took around 12 hours. This study had showed that there were temporary threshold shifts at 84 dBA but that they were comparatively worse when subjects were exposed to noise levels of 90 dBA. The author also reported that a median shift of 5 dB would be produced if noise levels were between 78−80 dBA at 250 Hz. So it is advisable to have the action level at 80 dBA with the permissible exposure limit at 85 dBA rather than 85 and 90 dBA for the action level and permissible exposure limit respectively. Yates et al.16) evaluated the damage risk of noise exposure at 85 and 90 dBA. The damage risk was evaluated by comparing temporary threshold shifts when exposed to these two noise intensities. There were six subjects in two groups, one group exposed to 85 and the other exposed to 90 dBA. There were equal numbers among genders in both groups. This experiment was conducted in a room where noise was produced by using either one of two noise generators with four loudspeakers. All the subjects had normal hearing threshold levels, which was prerequisite to be selected for the study. Each half of the subjects in a group was exposed to noise either for half a day or a full day. The Bekesy type of audiometer was used to measure threshold shifts ranging from 1 to 8 kHz for both ears, with the measurements being performed more than two minutes after exposure among the subjects. The outcome of this study showed that the mean temporary threshold shifts for the frequencies combined more than two minutes after exposure were higher on exposure to 90 dBA (mean [SD]=10.13 [7.45]) compared with 85 dBA (mean [SD]=6.26 [9.15]) among subjects exposed for a full day. This result was found to be significantly different upon conduct- 7 ing Duncan’s multiple range tests. This test however showed no statistically significant difference when comparing the individual frequencies under the fullday exposure condition. It also revealed that there were no significant differences between exposure to 85 dBA for a full day and 90 dBA for a half day. This indicated that the damages occurred on exposure to 85 dBA full-day was similar to that of 90 dBA halfday. Yates et al.16) also noted that the mean threshold shift measured at various time intervals by individual frequencies or frequencies together were similar to those at two minutes after noise exposure. This study emphasized that the effect on hearing upon exposure to 90 dBA was far more detrimental than that of 85 dBA. Hence, based on this study, it was advisable to adopt NIOSH’s recommended exposure limit. In a paper published by Kvaerner et al. 12), the authors evaluated temporary threshold shifts upon exposure to noise in an industry. Thirteen employees took part in this intervention study. These employees were tested and found to have normal hearing threshold levels and tympanograms prior to selection for the study. The subjects were exposed to noise levels between 85 and 90 dBA. Their hearing levels were examined for three days, and hearing protection devices were used for the first two days. The hearing levels were measured at 1 to 6 kHz using air conduction pure-tone audiometry, tympanometry and transient evoked otoacoustic emissions. Both ears were examined, since the correlations between them were low. The outcome showed that there were significant threshold shifts at 4 kHz (p=0.001) with a median value of 3.8 dB (range=−10.0 to 16.7 dB) and at 6 kHz (p < 0.001) with a median value of 5.5 dB (range= −5.0 to 11.7 dB) when comparing before and after work exposure. Consequently, there was a reduction in amplitude in the otoacoustic emissions (median=0.65 dB, range=−0.8 to 3.3 dB, p=0.001). The results showed that there were significant temporary threshold shifts when workers were exposed to noise levels beyond 85 dBA. This study again supports the notion that the permissible exposure limit preferably should be at 85 dBA. In Taiwan, Chen and Tsai 13) performed a crosssectional study on hearing loss among workers. The authors included all the workers in an oil refinery, but only 384 of them were selected for the study, as they were exposed to noise levels above 80 dBA and had no past history of chronic otitis media. All of them happened to be males. The average noise exposure among the subjects was 81 dBA. The left ears were analyzed using audiometry in the range of 0.5 to 8 kHz. The workers were categorized as having hearing loss if their average hearing threshold levels were more than 25 dBA between frequencies 0.5 to 2 kHz 8 and 3 to 6 kHz. The use of hearing protection devices during the study was not reported by the authors. The results showed that 9.6% of the workers had hearing loss at low frequencies compared with 38.3% who had hearing loss at high frequencies. Most of the threshold shifts were noted at 6 kHz among the subjects. This study highlighted that the action level should be 80 dBA, since slight hearing losses were observed among subjects at a noise exposure below 85 dBA, and that the permissible exposure limit should preferably be 85 dBA. In Denmark, Rubak et al.17) performed a prevalence study on risk of noise-induced hearing loss. The authors included 91 companies in the study, which comprised of 649 workers from noisy environments and 104 residents as a reference population. Dosimeters were used to measure noise areas. Questionnaires were distributed to extract information such as duration of work, ear diseases and exposure to noise during relaxation hours. The hearing threshold levels of the subjects were measured using pure-tone audiometry by a qualified technician. If the average hearing threshold levels were more than 20 dB from frequencies 2 to 4 kHz, then the subject was considered to have hearing impairment. The workers were wearing hearing protection devices. The results showed that risk of hearing impairment was statistically significant among subjects exposed to noise levels between 85 and 89 dBA [OR 3.05 (95% CI 1.33−6.99)]. These subjects worked for more than 20 years. The risk was mainly seen among subjects working in construction and basic metals trades. The subjects exposed to noise levels between 80 and 84 dBA and subjects who had worked less than 20 years showed no significant changes in hearing impairment. This finding indicated that if employees were exposed to noise levels beyond 85 dBA and continued to be exposed for many years, then the risk of hearing loss among them was much higher. Hence, the NIOSH’s recommended limit is far more appropriate based on the outcome from this study. Melnick18) performed an experiment on development of temporary threshold shifts on exposure to noise for 24 hours. This study was conducted among nine subjects having normal hearing threshold levels that were not more than 15 dB. Hearing threshold levels were measured among subjects during three time periods: before, during and after exposure to noise. The hearing threshold levels were as the average of the levels at 8, 12, 16, 20 and 24 hours of noise exposure. All nine subjects were exposed to 80 and then 85 dB through loudspeakers, with the octave band of the noise being centered at 4 kHz. The workers were given hearing protection devices if they were exposed high noise levels in the workplace. When the subjects J Occup Health, Vol. 56, 2014 were exposed to 80 dBA, they developed asymptotic threshold shifts of 9.3 dB at 4 kHz and 7.2 dB at 6 kHz, while at 85 dBA, the threshold shifts were 17.8 dB at 4 kHz and 14.6 dB at 6 kHz. Recovery of the threshold when subjects were exposed to 85 dBA took more than a day compared with those exposed to noise levels of 80 dBA. This indicated that more damaging effects were seen on exposure to noise levels of 85 dBA. Hence, the action level should be lower preferably at 80 dBA, as threshold shifts were also noted at this noise intensity. Discussion We conducted a systematic review of sources of evidence to compare adoption of 85 or 90 dBA for preserving the hearing threshold. A total of eight papers were found in our systematic search of the three main databases. These studies were considered to have acceptable quality. Half of the studies in this review were experimental in design, increasing the implications with regard to causality and outcome. According to DeVries and Berlet19), the evidence from these experimental studies was of high quality. The other studies in this review though were cross-sectional in design and were appropriate for reporting on the occurrence of temporary threshold shifts, as they fulfill the aim of this systematic review, which was to compare the effects of 85 dBA and 90 dBA on temporary threshold shifts. There was only one study 14) that used a blinding process. However, all the studies in this review had valid exposure and outcome measures. Probability of bias on selection may be noted in two studies14,18); however, the exposure and control groups were comparable, and confounding bias was controlled for in these studies. The search was not restricted to a specific study design, since only limited numbers of published studies were available. Any discrepancies were resolved by consensus through discussion before the studies were included. All four experimental studies, those of Stephenson et al.14), Mills et al.15), Yates et al.16) and Melnick18), were conducted on human subjects. The subjects were exposed to noise generated by equipment with loudspeakers. The noises introduced were of a continuous type. The other studies, those of McBride et al.11), Kvaerner et al.12), Chen and Tsai13) and Rubak et al.17), were conducted in an industrial setting in which employees were exposed to a fluctuating type of noise. The precise machines involved in generating noise in these industries were not reported. Hearing protection devices were used by the workers during the studies of Kvaerner et al.12) and Rubak et al.17) but use of hearing protection devices was not reported by McBride et al.11) and Chen and Tsai13). There were Balachandar S. SAYAPATHI, et al.: Review of Different PELs for Preserving HTL studies comparing 80 dBA and 85 dBA, such as those of Stephenson et al.14) and Melnick18), and there was also a study on subjects exposed to 81 dBA, that of Chen and Tsai13). We included these studies to show the detrimental effects of exposure to 85 dBA and that the negative effects would be greater if employees were exposed to noise levels reaching 90 dBA. Our study recommends adoption of 85 dBA for conservation of the hearing threshold. Most of the studies indicated that the temporary threshold shifts were much lower when subjects were exposed to noise levels of 85 dBA or lower. However, the study of McBride et al.11) showed that there were no statistically significant changes in temporary threshold shifts among subjects exposed to less than 85, between 85 and 90 dBA and more than 90 dBA. These findings of McBride et al.11) may not be applicable, since post-shift audiometry was done within a day and the employment durations of the workers exposed to these noise levels were not reported. The temporary threshold shifts among the employees may result in permanent shifts after continuous exposure to high noise levels20). These permanent shifts may result due to damage to hair cells in the organ of Corti2) which is irreversible. The cochlea contains the organ of Corti, which is embedded with sensory and non-sensory cells. The sensory cells appear to be hair-like cells referred to as outer or inner cells. The outer sensory cells are not only more abundant but are also arranged in three rows compared with the inner cells, which are arranged in only one row5). The outer hair cells are more prone to damage from noise and ototoxic drugs. Prolonged exposure to sounds ranging 3,000 to 6,000 Hz may damage these cells1,3). In the current review, there are four comparative studies among subjects exposed to noise levels of 85 and 90 dBA. Of them, three studies showed that there were more temporary threshold shifts with 90 dBA than with 85 dBA. One of the studies showed that when subjects were exposed to noise levels between 85 and 89 dBA, the OR was 3.05 (95% CI, 1.33 to 6.99), whereas no change in threshold shifts was observed on exposure to noise levels between 80 and 84 dBA17). Another study, that of McBride et al.11), showed that there was no statistically significant change in threshold shifts, as data collection was performed within a day after baseline audiometry in an industry setting. Our findings are consistent with a survey done earlier in the US. According to the US NIOSH21), the survey was conducted to estimate material hearing impairment in industries over a 40-year working lifetime. Employees from 13 industries were surveyed from 1968 to 1971. Material hearing impairment is said to occur if the hearing threshold level is beyond 25 dBA at 1,000, 2,000, 3,000 and 4,000 Hz. Only 9 1,172 audiograms were taken for data analysis from more than 4,000 audiograms collected, since the others did not fulfill the inclusion criteria. The abnormal hearing levels of the employees and levels of noise exposure were obtained through questionnaires and audiometry assessment. Excess risk estimates were then calculated; the difference between employees exposed to noise resulting in levels beyond the maximum acceptable hearing loss and the unexposed population exceeded these levels in percentage. The excess risk estimates increased as duration of exposure to noise increased among the population exposed to noise levels of 85 and 90 dBA. Employees exposed to noise levels of 85 dBA had excess risk estimates of 1.4 (95% CI, 0.3 to 3.2), 2.6 (95% CI, 0.6 to 6.0), 4.0 (95% CI, 0.9 to 9.3) and 4.9 (95% CI, 1.0 to 11.5) at 30, 40, 50 and 60 years old, respectively, for a duration of exposure to noise of between 5 to 10 years, compared with those exposed to 90 dBA, who had excess risks of 5.4 (95% CI, 2.1 to 9.5), 9.7 (95% CI, 3.7 to 16.5), 14.3 (95% CI, 5.5 to 24.4), and 15.9 (95% CI, 6.2 to 26.2). For employees who were exposed to noise for more than 10 years, the excess risk estimates for 85 dBA were 2.3 (95% CI, 0.7 to 5.3), 4.3 (95% CI, 1.3 to 9.4), 6.7 (95% CI, 2.0 to 13.9) and 7.9 (95% CI, 2.3 to 16.6) at 30, 40, 50 and 60 years old, respectively, compared with those exposed to 90 dBA, who had excess risks of 10.3 (95% CI, 5.8 to 16.2), 17.5 (95% CI, 10.7 to 25.3), 24.1 (95% CI, 14.6 to 33.5), and 24.7 (95% CI, 14.9 to 34.3). In summary, at 60 years old, the excess risk estimates were 8% for employees exposed to 85 dBA and 25% for employees exposed to 90 dBA. It is claimed that subjects or employees exposed to noise with 90 dBA considered the permissible exposure limit may experience a more damaging effect on hearing compared with those exposed 85 dBA. Temporary threshold shifts may result in permanent shift of hearing thresholds over time if continuous exposure to noise ensues20). According to Kryter22), temporary threshold shifts that developed in his study reflected one day’s exposure to noise. These temporary threshold shifts occurred two minutes after exposure to noise (TTS2). The author also stressed that temporary threshold shifts at two minutes may result in permanent threshold shifts if one is exposed to noise continuously for a longer duration. It is hypothesized that permanent threshold shifts would occur at about 10 years due to this continuous noise insult. Kryter22) also revealed that if the temporary threshold shifts at two minutes was beyond 40 dB, the recovery period of threshold shifts were prolonged and development of permanent threshold shifts may ensue. According to Lawton20), noise levels beyond 80 dBA produced temporary threshold shifts that recovered 10 quickly upon cessation of the noise insult. Noise levels at or below 80 dBA may not produce hearing loss at 4,000 Hz, the frequency that most susceptible to noise. Lawton also mentioned that employees exposed to noise levels beyond 85 dBA over a period of eight hours would acquire some degree of hearing loss. Hence, it would be more appropriate to institute a hearing conservation program1) at 80 dBA in which action can be taken to reduce noise exposure among employees, also known as the action level, and to adopt 85 dBA as the permissible exposure limit. The findings from this systematic review also suggest adoption of the recommendations of the US NIOSH rather than the US OSHA’s occupational exposure limit. The Japan Society for Occupational Health23) has also recommended the permissible exposure limit of the US NIOSH, that is, 85 dBA for a period of eight hours. Another experimental study24) has been conducted on noise-induced temporary threshold shifts in Japan. In the exposure group, ten subjects were exposed to noise levels in between 65 and 86 dB for 24 hours. The control group was not exposed to noise. The hearing threshold levels were measured over the right ear for both groups. There were no statistically significant changes in threshold levels among the control group. However, among the exposure group, temporary threshold shifts increased as noise levels increased. The threshold shifts at 86 dBA were below 10, 15, 25, 25 and 20 dB as compared with subjects exposed at 83 dBA, the threshold shifts of which were 5, 12.5, 15, 20 and 15 dB at 2, 3, 4, 6 and 8 kHz, respectively, after 24 hours of noise exposure. The study of by Yates et al.16) showed that the damaging effects on threshold shifts of subjects exposed to a full day of noise at 85 dBA were equivalent to those of subjects exposed to a half day of noise at 90 dBA. Moreover, most of the detrimental effects were seen at 4 kHz16,18). Conclusions There were more threshold shifts in subjects exposed to 90 dBA compared with 85 dBA. These temporary threshold shifts over time may lead to noise-induced hearing loss, which is irreversible. By adopting 85 dBA as the permissible exposure limit, action to minimize noise levels from sources and exposure among employees would be taken earlier. Early action taken may reduce occurrence of temporary threshold shifts and hence reduce the prevalence of noise-induced hearing loss. Acknowledgments: This study was supported by a Postgraduate Research Grant (Grant number PV106/ 2011A) from the University of Malaya. J Occup Health, Vol. 56, 2014 References 1) Kirchner DB, Evenson E, Dobie RA, et al. ACOEM guidance statement: occupational noise-induced hearing loss. JOEM 2012; 54: 106−8. 2) Morioka I, Miyashita K, Takeda S. Noise-induced hearing loss in working environment and its background. J Occup Health 1997; 39: 5−17. 3) Haboosheh R, Brown S. Workplace hearing loss. BCMJ 2012; 54: 175. 4) British Society of Audiology. Recommended procedure: pure tone air and bone conduction threshold audiometry with and without masking and determination of uncomfortable loudness levels; 2004. 5) Rutka J. Discussion paper on hearing loss: Prepared for the veterans review and Appeal Board: Veterans Review and Appeal Board Canada. 2011. 6) Franks JR, Stephenson MR, Merry CJ. Preventing occupational hearing loss—A practical guide U.S. Department of Health and Human Services. 1996. 7) Nelson DI, Nelson RY, Concha-Barrientos M, Fingerhut M. The global burden of occupational noise-induced hearing loss. Am J Ind Med 2005: 1−15. 8) Laws of Malaysia. Factories and Machinery Act 1967 (Act 139), Regulations & Rules: International Law Book Services. 2010. 9) Occupational diseases and poisoning investigation. Ministry of Human Resources, Malaysia, 2013. [Online]. 2013; URL: http://www.dosh.gov. my/index.php?option=com_content&view=articl e&id=392:occupational-diseases-and-poisoninginvestigation&catid=349&Itemid=760&lang=en 10) Su AT, Darus A, Bulgiba A, Maeda S, Miyashita K. The clinical features of hand-arm vibration syndrome in a warm environment- a review of the literature. J Occup Health 2012; 54: 349−60. 11) McBride D, Gilmore T, Waite D. The immediate retest in temporary threshold shift audiometry. JOEM 2003; 45: 1211−2. 12) Kvaerner KJ, Engdahl B, Arnesen AR, Mair IW. Temporary threshold shift and otoacoustic emissions after industrial noise exposure. Scand Audiol 1995; 24: 137−41. 13) Chen JD, Tsai JY. Hearing loss among workers at an oil refinery in Taiwan. Arch Environ Health 2003; 58: 55−8. 14) S t e p h e n s o n M R , N i x o n C W, J o h n s o n D L . Identification of the minimum noise level capable of producing an asymptotic temporary threshold shift. Aviat Space Environ Med 1980; 51: 391−6. 15) Mills JH, Osguthorpe JD, Burdick CK. Temporary t h r e s h o l d s h i f t s p r o d u c e d b y ex p o s u r e t o low-frequency noises. J Acoust Soc Am 1983; 73: 918−23. 16) Yates JT, Ramsey JD, Holland JW. Damage risk: an evaluation of the effects of exposure to 85 versus 90 dBA of noise. J Speech Hear Res 1976; 19: 216−24. 17) Rubak T, Kock SA, Koefoed-Nielsen B, Bonde JP, Balachandar S. SAYAPATHI, et al.: Review of Different PELs for Preserving HTL 18) 19) 20) 21) Kolstad HA. The risk of noise-induced hearing loss in the Danish workforce. Noise Health 2006; 8: 80−7. Melnick W. Temporary threshold shift following 24-hour noise exposure. Ann Otol Rhinol Laryngol 1977; 86: 821−6. DeVries JG, Berlet GC. In the profession: understanding levels of evidence for scientific communication. Foot Ankle Spec 2010; 3: 205−9. Lawton BW. A noise exposure threshold value for hearing conservation. Brussels: CONCAWE; 2001. National Institute for Occupational Safety and Health. Criteria for a Recommended Standard: 11 Occupational Noise Exposure Revised Criteria 1998: U.S. Department of Health and Human Services; 1998. 22) Kryter KD. Hazardous exposure to intermittent and steady-state noise NAS-NRC Committee on Hearing, Bioacoustics, and Biomechanics. 1965. 23) The Japan Society for Occupational Health. Recommendation of occupational exposure limits. J Occup Health 2012; 54: 387−404. 24) Takagi K, Hiramatsu K, Yamamoto T, Ito A. Prediction of noise-induced temporary threshold shift. J Sound Vib 1988; 127: 513−9.