An evaluation of short-term exposures of brake mechanics to asbestos during automotive and truck brake cleaning and machining activities

Historically, the greatest contributions to airborne asbestos concentrations during brake repair work were likely due to specific, short-duration, dust-generating activities. In this paper, the available short-term asbestos air sampling data for mechanics collected during the cleaning and machining of vehicle brakes are evaluated to determine their impact on both short-term and daily exposures. The high degree of variability and lack of transparency for most of the short-term samples limit their use in reconstructing past asbestos exposures for brake mechanics. However, the data are useful in evaluating how reducing short-term, dust-generating activities reduced long-term exposures, especially for auto brake mechanics. Using the short-term dose data for grinding brake linings from these same studies, in combination with existing time-weighted average (TWA) data collected in decades after grinding was commonplace in rebuilding brake shoes, an average 8-h TWA of approximately 0.10 f/cc was estimated for auto brake mechanics that performed arc grinding of linings during automobile brake repair (in the1960s or earlier). In the 1970s and early 1980s, a decline in machining activities led to a decrease in the 8-h TWA to approximately 0.063 f/cc. Improved cleaning methods in the late 1980s further reduced the 8-h TWA for most brake mechanics to about 0.0021 f/cc. It is noteworthy that when compared with the original OSHA excursion level, only 15 of the more than 300 short-term concentrations for brake mechanics measured during the 1970s and 1980s possibly exceeded the standard. Considering exposure duration, none of the short-term exposures were above the current OSHA excursion level.


Introduction
Epidemiology studies of insulation workers in the 1960s (e.g., Selikoff et al., 1964) were the first to show conclusively that use of an asbestos-containing product could pose a cancer risk to workers. As asbestos was also a major component of friction materials in automobiles and trucks (e.g., brake linings and clutch plates), researchers also began to investigate the degree of asbestos exposure experienced by vehicle mechanics during the servicing of brakes. The first survey of airborne asbestos levels associated with brake servicing was conducted in the United Kingdom in 1968 . The first exposure survey of brake mechanics in the United States was published by the National Institute of Occupational Safety and Health (NIOSH) in 1972 (Dement, 1972), followed by several additional NIOSH surveys in the 1970s and 1980s (Johnson et al., 1979;Roberts, 1980a, b;Roberts and Zumwalde, 1982;Sheehy et al., 1989). During this same time period, major studies of potential exposures of brake mechanics were also conducted in Germany (Ro¨delsperger et al., 1986), Sweden (Plato et al., 1995), and Finland (Kauppinen and Korhonen, 1987). Paustenbach et al. (2003) evaluated the published airborne asbestos concentrations for brake mechanics and developed 8-h time-weighted average (8-h TWA) asbestos concentrations from the personal air sampling data. This study found that the mean 8-h TWA asbestos concentration for automobile brake mechanics in the United States decreased from 0.063±0.032 fibers per cubic centimeter (f/cc) in the 1970s (n ¼ 42) to 0.0021±0.00036 f/cc in the late 1980s (n ¼ 42), due largely to the introduction of dust control technologies during cleaning activities.
As with many types of occupational dust exposures, the TWA asbestos concentrations associated with historical brake repair were likely due to short-term activities that generated the greatest quantities of dust. These short-duration activities involved the removal of brake-wear debris (e.g., brake dust) from brake assemblies (often using compressed air or a dry brush) and the machining of brake linings (often by grinding or beveling the lining surfaces to provide a better fit with the drum). During the 1970s and 1980s, hundreds of short-term airborne samples were collected in garage settings during brake cleaning and machining activities. Many of these samples were collected to assess the effectiveness of various dust control measures and/or to assess mechanic exposures and compliance with short-term exposure standards. However, to our knowledge, there has been no systematic attempt to assemble and evaluate these data.
We believe characterizing the short-term sampling data collected during brake servicing would directly answer several important questions. First, did introduction of dust control technologies over time result in decreasing airborne asbestos concentrations associated with the ''dust-generating'' shortterm activities? Second, can the short-term data be used to adjust TWA measurements to include specific activities that were not common when the long-term data were collected? And third, how do the measured concentrations compare with the short-term occupational exposure limits (OELs) for asbestos in place at the time? Regarding the second question, it has been noted that most of the available brake-job TWA concentration data were collected in the 1970s and 1980s, a time when machining of automobile brake linings was not as common as in prior years due to the introduction of bonded and pre-arced brake pads for drum brakes and the greater use of disc brakes on specific automobile models, which required no machining before installation. As a result, it is unclear whether TWAs in the 1960s and earlier decades, when machining was more common, might have been substantially different from those reported for brake mechanics in the post-1970s era. Therefore, it would be useful to determine whether incorporation of short-term machining data into the existing long-term brake-job measurements would yield a substantial change in the brake-job and 8-h TWA estimates for brake mechanics.
This analysis represents the first published summary and interpretation of all the available short-term asbestos air sampling data for vehicle mechanics during the cleaning and machining of vehicle brakes.

Brake-Servicing Activities
For vehicles with asbestos brake linings or pads, the potential for the release of airborne asbestos is thought to have been greatest during cleaning of the brake assemblies and machining of the brake linings. Table 1 lists the various short-term brakeservicing activities that are evaluated in this paper.

Cleaning Activities
Over the years, numerous cleaning methods have been used to remove brake wear debris, which is composed mostly of non-fibrous materials containing less than 1% asbestos by weight (Anderson et al., 1973;Jacko et al., 1973;Williams and Muhlbaier, 1982;Cha et al., 1983), from the brake assemblies. As it was quick and readily available at most facilities, mechanics often used compressed air to blow off the loose debris from the brake assemblies. If compressed air was not available, mechanics would often use a rag or brush to remove the accumulated debris. However, as concern over asbestos dust increased during the 1970s, brake maintenance facilities switched to cleaning methods that reduced the generation of airborne dust, including HEPA vacuuming, wet brushing, wet wiping, solvent cleaning, and various combinations (Sheehy et al., 1989).

Machining Activities
Over the years, changes in brake design resulted in significant decreases in the need for machining activities, especially for cars and light trucks (Sheehy et al., 1989). For example, from the late 1920s through the 1940s, molded linings were used to replace worn brake linings. The old linings were removed by punching out the rivets that held them to the brake shoes. The new linings either came in rolls or were precut. If not predrilled, holes for the rivets were drilled in the new linings. The edges sometimes required beveling to prevent the new linings from catching when the brakes were applied. If the brake drums had been turned to remove surface imperfections, it was often necessary to arc grind the linings so that the external radii of the new linings were the same as the new internal radii of the drums. All of these machining operations had the potential to generate airborne dust, which contained resin particles from the lining matrix, fibers with attached resin, and some free asbestos fibers . The introduction in 1948 of bonded linings for cars and light trucks, in which the linings came bonded to replacement shoes, dramatically reduced the need for the machining activities that produced dust. As beveling was sometimes a part of brake repair during the early years, by the mid-1950s, replacement brakes arrived pre-beveled from the factory (Sheehy et al., 1989). Also, by approximately 1950, replacement shoes could be obtained that were pre-arced to specific radii to match turned drums, thus negating the need for arc grinding. The introduction of disc brakes in the 1960s in the United States (1950s in Europe) further reduced the need for machining activities, because the old disc brake pads were simply removed and replaced with new ones. By the mid 1970s, most automobiles in the United States had a combination of front disc brakes and rear drum brakes. By the 1990s, most automobiles sold in the United States had disc brakes on all four wheels.
Due to the costs involved, some service facilities still rebuild the much larger bus and heavy-truck brakes by removing the old linings from the shoes and installing new ones. This process may still involve machining activities, including removing the old rivets and linings, and drilling and riveting the new linings.

Data Collection
To identify and collect relevant asbestos air sampling data for vehicle brake mechanics, literature searches of 35 databases from the Thomson DIALOG aggregator search system were conducted. This search system has international coverage of tens of thousands of journals, government studies, conference proceedings, technical reports, and monographs. Databases searched included Medline, EMBASE, and others related to medicine; toxicology, environment, materials, transportation, engineering, and other related technical fields. Keywords included various forms of asbestos terminology (e.g., asbestos, chrysotile, brakes) limited to the title and descriptor fields of the records.
Given the disparate nature of the studies in terms of methodology and reported level of detail, it was necessary to establish minimum criteria for identifying those data most appropriate for use in quantitative analyses. Specifically, the following criteria were used to identify the air sampling data that were most appropriate for use in quantifying short-term exposures during brake cleaning and machining activities: (1) the asbestos air sampling data were published in the peerreviewed literature or in government reports; (2) only personal sampling measurements collected in the breathing zone were evaluated; (3) samples had to be associated with a specific cleaning or machining activity; (4) samples had to have reported sampling durations; and (5) samples had to have a reported asbestos concentration or a mean and range of concentrations measured in standard units of fibers per cubic centimeter (fibers 45 mm in length) and analyzed using phase contrast microscopy methods. As the brake-servicing activities for cars and light trucks are known to have been quite different from those of heavy trucks and buses, these two categories of vehicles are considered separately.
A summary of the five government reports (Johnson et al., 1979;Roberts, 1980a, b;Roberts and Zumwalde, 1982;Sheehy et al., 1989) and ten published studies (Hatch, 1970;Knight and Hickish, 1970;Lorimer et al., 1976;Rohl et al., 1976;Jahn, 1983;Cheng and O'Kelly, 1986;Ro¨delsperger et al., 1986;Kauppinen and Korhonen, 1987;Plato et al., 1995; identified as containing short-term asbestos air sampling data for brake mechanics is presented in Table 2. These 15 studies, conducted in the United States, United Kingdom, Finland, Germany, andHong Kong between 1968 and2000, represent more than 300 short-term personal asbestos air samples for automobile and truck mechanics collected during brake machining and cleaning activities. Tables 3 and 4 present more detailed information on the individual air samples collected during specific brake repair activities involving cars and light trucks, and heavy trucks and buses, respectively. The manner in which each sample is evaluated in this paper is indicated in these tables in the ''Concentration analysis'' and ''Short-term dose analysis'' columns. Ro¨delsperger et al. (1986) was the only report that took into account the exposure duration, providing a summary of short-term asbestos doses for brake mechanics performing cleaning and machining activities for a duration of 3 to 60 minutes.
Seven other studies (Lloyd, 1975;Johnson, 1976;Rohl et al., 1977;Roberts and Zumwalde, 1980;Nicholson et al., 1983;Jahn et al., 1985;Ro¨delsperger, 1987) with short-term asbestos air sampling data were identified, but were excluded because these data duplicated those in the other evaluated studies. A few studies Dement, 1972;Lorimer et al., 1976) were not included in the quantitative evaluation because they lacked sufficient sampling information.

Data Analysis
The short-term asbestos air sampling data were interpreted using two different methods. First, all samples that were collected for less than 15 min were evaluated in the ''concentration analysis.'' As some of the concentration data were presented as summaries, it was only possible to determine the total number of samples and the averages, maximums, and minimums. Because most of the data for individual samples were unreported in these summaries, the summary data could not be aggregated with individual sample data for further statistical analyses. Second, all individual samples with both a concentration and a specific sampling duration were used to calculate short-term asbestos doses associated with machining and cleaning activities to account for the relatively wide range of sampling durations associated with specific activities. For the purposes of this paper, these calculations are termed ''short-term dose analyses.'' Specifically, short-term asbestos doses were calculated as time-integrated estimates in terms of f/cc*min: Short-term doseðf=cc Ã minÞ ¼concentrationðf=ccÞ Âsampling durationðminÞ

Reconstructing Past Long-Term Exposures
It seems reasonable that the short-term dose data for grinding brake linings could be used to adjust the available long-term TWA concentrations collected after grinding had decreased to estimate earlier TWA concentrations for mechanics performing brake jobs where the grinding of linings had occurred. Potential mechanic exposures that included grinding were estimated based on the available, later brake-job and 8-h TWA concentrations and short-term grinding dose and duration data. The TWA for a brake job including grinding was estimated as follows: TWA BJþG ¼ððTWA BJ ÂdurationÞ þ grinding doseÞ =ðbrake job duration þ grinding durationÞ    Assuming that the background concentration would have increased proportionately, the 8-h TWA for the brake job including grinding was calculated by extrapolating from the 8-h TWA assumed to be without grinding using the following equation:

Cleaning Methods
The data for car and light-truck mechanics consist of 69 air samples collected during the cleaning of drum brake assemblies (see Figure 1). The vast majority of these samples were collected during compressed air (n ¼ 39) and dry-brush (n ¼ 13) cleaning events, with relatively few samples collected for the other cleaning methods. The highest mean and maximum asbestos air concentrations were from samples collected during ''blow-out'' using compressed air (3.0 and 29 f/cc, respectively). Mean levels for samples collected during other brake cleaning methods for cars and light trucks ranged from 0.04 f/cc (vacuum) to 1.6 f/cc (wet brush).
The data for heavy-truck and bus mechanics consist of 42 samples collected during the use of seven different methods for cleaning of drum brake assemblies ( Figure 2). Interestingly, the mean airborne asbestos levels (which ranged from 0.25 f/cc for opening/water washing to 1.9 f/cc for damp cloth wiping) were comparable to the range of means calculated for the car and light-truck samples.
Of all the studies evaluated, only Ro¨delsperger (1987) reported measured durations of various cleaning activities for cars vs. trucks and buses. The reported mean and standard deviations per axle for eight activities are shown in Table 5. On the basis of these data, the cleaning of brake assemblies on cars using either compressed air or dry brushing took an average of about 5 min per vehicle (for all four wheels), while cleaning truck or bus brake assemblies took approximately 14 min. More recently, as part of a brake-change simulation for six cars with drum brakes, Blake et al. (2003) reported that the duration of compressed-air blowout for all four wheels ranged from 0.37 to 0.77 min, a much shorter duration than that reported by Ro¨delsperger (1987). As can be seen in Tables 3  and 4, while the majority of the reported sampling durations for cleaning activities are 10 min or less, the sampling durations for any given type of activity were highly variable across the different studies, ranging from a fraction of a minute up to 10 min (and in a few instances longer than 30 min).
When the task (or sampling) duration data were available, estimates of short-term doses were calculated (Table 6). For car and light-truck mechanics, compressed-air blowout, drybrushing, and wet-brushing samples yielded the highest mean short-term exposure values of all the cleaning methods. The relative magnitudes of some of these doses do not correspond to the concentration data. For example, the mean short-term  c Calculated by multiplying the asbestos concentration times the sampling duration.
For heavy-truck and bus mechanics, the mean short-term doses were similar among the three cleaning methods sampled (range of 3.8 f/cc*min for compressed air to 6.7 f/cc*min for damp cloth) and were similar to the range of means measured for cleaning cars and light trucks (Table 6). These values are similar to those Ro¨delsperger et al. (1986) who reported an average dose of 5.3 f/cc*min (n ¼ 8) for compressed-air blowout per axle.

Machining Methods
As can be seen in Tables 3 and 4, the majority of the car/lighttruck mechanic data were collected during cleaning activities, whereas the majority of heavy-truck/bus mechanic data were collected during machining activities. The relative paucity of the car/light-truck machining data (four samples) is likely due to the fact that when asbestos sampling for brake mechanics began in the early 1970s, new brake shoes for cars and light trucks were typically pre-assembled and did not require the use of machining methods during most brake-servicing operations. Measured airborne asbestos concentrations ranging from 0.15 to 0.50 f/cc were reported for turning brake discs and drums. Both of these machining activities remove wear debris from the metallic surfaces of the brake assemblies and do not actively machine the asbestos brake linings or pads.
More than 100 air samples were collected during a wide range of machining activities for heavy trucks and buses (Figure 3). For most machining activities, the means ranged from 0.20 f/cc (turning brake drums) to 3.8 f/cc (grinding used linings). Two operations in particular produced substantially greater airborne asbestos concentrationsFmachine beveling of new linings, with a mean fiber concentration  of 37 f/cc, and grinding new linings without exhaust controls, with a mean concentration of 56 f/cc. The data for grinding of linings are divided into five groups, because information is lacking as to whether the linings were new or used and whether exhaust controls were used. For example, Cheng and O'Kelly (1986) do not report whether the grinding was done with or without exhaust controls, whereas Ro¨delsperger et al. (1986) indicate that the grinding was done in an enclosed shroud, with the shoes mounted on the truck and exhaust controls in place. As indicated by the data, the use of exhaust controls during the grinding of truck brake linings did significantly reduce airborne asbestos concentrations; the mean level without exhaust (56 f/cc) was approximately 40 times higher than the mean level with exhaust (1.5 f/cc). By comparison, Ro¨delsperger et al. (1986) reported an average dose of 9.9 f/cc*min (n ¼ 4) for grinding of truck brakes per axle. The mean concentration for hand beveling of brake linings (0.40 f/cc) was almost 100 times less than that for machine beveling (37 f/cc). The hand method was the type used most often by car and light truck mechanics before the introduction of pre-beveled linings.
As indicated above, only the study by Ro¨delsperger (1987) reported measured durations of various machining activities for cars vs. trucks and buses (see Table 5). On the basis of these data, the average duration of grinding activities on truck brake linings (34 min) was substantially longer than that reported for car brake linings (11 min), which is simply due to the larger wheel assemblies on these vehicles and the complexity of the process. The more recent assessment by Blake et al. (2003) reported that arc grinding for two cars averaged about 19 min, which was about 8 min longer than what Ro¨delsperger (1987) reported. Both authors reported arc grinding durations that lasted much longer than the actual grinding of the brake lining which takes only a few minutes; the reported durations are, in fact, for the entire time required to mount the brake shoe, set the radius, arc the lining, and remove the brake shoe from the grinder.
When taking the sampling duration information into account, the calculation of short-term doses for brake mechanics performing machining activities is based on only four samples for cars and light trucks (Table 6), only one of which involves brake linings. The doses ranged from 1.5 f/cc*min for turning disc rotors and brake drums to 8.6 f/cc*min for grinding new linings. However, it should be noted that there are additional short-term dose data for grinding automobile brake linings, although these are not reported as individual samples. For one auto axle (two wheels) Ro¨delsperger et al. (1986) reported a mean short-term dose of 3.9 f/cc*min based on 10 samples for mechanical grinding and 3.9 f/cc*min for 5 samples for hand   grinding of brake-shoe linings. The authors stated that most of the bag filter systems for grinding were unsuitable for asbestos. These short-term machining doses are similar to those calculated for cleaning activities on cars and light trucks.
For heavy-truck and bus mechanics, seven samples were available to calculate doses during machining activities. As shown in Table 6, exposures were calculated for heavy-truck and bus mechanics grinding new linings, grinding linings with exhaust controls, and preparing brake shoes before riveting new linings. There was less than a 10-fold difference in doses for these activities, with mean values ranging from 1.4 f/cc*min (preparation of brake shoes before riveting new linings) to 10 f/cc*min (grinding linings on wheels with exhaust controls). As with cars and light trucks, the machining exposures during heavy truck brake repair are similar to the cleaning exposures. The mean estimate for grinding new linings on heavy trucks and buses (5.5 f/cc*min) is similar to that for the one sample for grinding linings on cars and light trucks (8.6 f/cc*min).

Reconstructing Earlier Long-Term Exposures
Earlier long-term exposures for automobile brake mechanics that used grinding in rebuilding brake shoes were estimated based on the available brake-job and 8-h TWA concentrations, short-term grinding dose data, and several assumptions. Due to the lack of statistical data for grinding automobile brake shoes, only the average TWA concentrations were determined.
From the Paustenbach et al. (2003) data, during the 1970s, the average brake-job TWA was 0.11 f/cc over a period of approximately 90 min, resulting in a dose of 9.9 f/cc*min. These exposures definitely included the use of compressed air and/or dry brushing, but may not have included the machining of brake shoes. The best estimate of the short-term dose for grinding is 3.9 f/cc*min per axle (or 7.8 f/cc*min for both axles and all four wheels) for the 10 samples from Ro¨delsperger et al. (1986). The adjusted brake-job TWA was then calculated by adding the brake-job dose of 9.9 f/cc*min and the grinding dose for two axles of 7.8 f/cc*min, and then dividing by the total adjusted time of 101 min. Adjusted for grinding, the brake-job TWA is approximately 0.18 f/cc, about a 60% increase over the brake-job TWA, for which it was assumed grinding was not conducted during brake repair. This estimate is comparable with the airborne concentrations reported by Blake et al. (2003) for two brake jobs (0.20 and 0.44 f/cc) where arc grinding was evaluated. Using the average 8-h TWA of Paustenbach et al. (2003) for the 1970s of 0.063 f/cc, when cleaning was not controlled and assuming that background airborne concentrations in the garage would have increased proportionately due to grinding, the estimated average 8-h TWA for auto mechanics who performed grinding could have been approximately 0.10 f/cc.

Data Variability
Our review indicates that for one of the most common dustgenerating activities, compressed air cleaning of automobile brakes, the short-term concentrations ranged over about three orders of magnitude, from 0.01 to 29 f/cc, whereas the doses ranged over approximately two orders of magnitude, from 0.1 to 7.5 f/cc*min. Thus, incorporating the sampling duration into the exposure estimates did little to reduce the variability in estimating exposures, which suggests that there may have been substantial differences in fiber emissions from compressed-air blowout by individual mechanics. A similar degree of variability was observed in the short-term doses of mechanics grinding brake linings on heavy trucks and buses, with values ranging from 0.4 to 8.3 f/cc*min. This variability is likely due in large part to differences in study objectives and variations in mechanic work practices, neither of which was usually described in detail. These apparent anomalies could be due to several different factors, such as: (1) sample durations were variable and short samples may have captured a high concentration but were of too short a duration to have consistently captured a majority of the emitted fibers, (2) limited data and/or high variability among samples did not permit meaningful comparisons, and/or (3) brake mechanics conduct specific cleaning and machining activities quite differently from one another.
These findings are similar to those reported in other studies. For example, Ro¨delsperger et al. (1986) provided a summary of asbestos doses (43 too60 min duration) for car and truck mechanics performing several cleaning and machining activities. In general, the estimates of dose from Ro¨delsperger et al. (1986) for comparable cleaning and machining activities were about twofold higher than the dose values calculated in this analysis; however, caution should be taken when comparing these results to this study, because it was not always clear how many wheels or axles were being worked on while the samples were collected.
Although some understanding of sample duration is available for all of the results that met the data selection criteria, most researchers, with the exception of Ro¨delsperger (1987), did not report the duration of the cleaning or machining activity relative to the sampling duration. Hence, it is highly likely that many samples were collected over a duration that was either shorter or longer than the actual activity itself. An airborne dust sample collected for a duration that is less than or equal to the ''peak'' of the dust-generating activity is likely to contain a higher dust concentration than a sample collected for several minutes following the completion of that same activity. Conversely, the short-term dose metric (f/cc*min) will increase as the sampling duration continues beyond the cessation of activity until dust dissipates from the breathing zone. Hence, sampling durations that do not last beyond the activity duration may yield higher concentrations but underestimate the total extent of exposure as Brake mechanics' short-term asbestos exposure Richter et al.
measured by the dose, because it often takes a few minutes after cleaning or grinding for airborne concentrations to return to background levels. Ideally, the sample duration should have extended beyond the activity duration to ensure that the residual suspended dust was collected and the total dose from a specific activity determined.

Usefulness of Short-Term Sample Data for Reconstructing Asbestos Exposures
The second purpose of this analysis was to determine whether the short-term sampling data could be used to develop estimates of pre-1970s exposures, either by using the shortterm data alone or by incorporating short-term machining data into the (post-1970s) brake-job and 8-h TWAs. It would be difficult to use these short-term data to estimate the degree of exposure due largely to the substantial variability in the reported measurements and the lack of descriptions of the specific activities that were sampled. Specifically, the variabilities in concentrations and doses for specific cleaning and machining activities are high, often spanning several orders of magnitude. This is most apparent for the measurements of compressed air cleaning of automobile brake assemblies and grinding of truck brake linings. Clearly, there are substantial differences in how the mechanics performed these activities, and in garage conditions, and a lack of documentation as to what contributes to the high and low values reported. Therefore, the combined short-term data from the studies evaluated provide a less than ideal basis for reconstructing short-term or long-term exposures historically experienced by a brake mechanic. However, it is clear that subsets of these data from groups intimately involved with sampling the brake servicing process, such as that proposed by Ro¨delsperger et al. (1986), have been used to estimate longterm TWAs based on a sum of short-term doses for specific activities, and that these data could be used by others evaluating similar activities and exposure conditions.
Although the variability of the data makes the results of total reconstruction quite uncertain, it has been shown that short-term grinding dose data could be used to adjust the long-term TWA values for mechanics performing brake jobs where the grinding of linings occurred, as that took place when brake shoes were rebuilt in the period before collection of the available TWA data (i.e., pre-1970s).

Impact of Short-term Activities on Long-Term Exposures Over Time
As evidenced by our analysis of the short-term exposure data, the decrease in dust-generating activities over time had a large impact on both the short-term and the long-term asbestos exposures for brake mechanics. The need for machining activities, especially for cars and light trucks, was minimized by the introduction of pre-arced, replacement brake shoes for drum brakes in the 1950s and 1960s, and the conversion to disc brakes in the 1970s and 1980s. The changes in cleaning methods during the 1980s, when employed correctly, further reduced asbestos dust generation to the point where long-term exposure concentrations were at or below the detection limits.
On the basis of our assumptions and the data from Paustenbach et al. (2003), the average brake-job and 8-h TWAs for automobile mechanics were 0.18 and 0.10 f/cc when grinding and compressed air or similar cleaning methods were used. Removing grinding activities from brake servicing reduced the average brake-job and 8-h TWA values to 0.11 and 0.063 f/cc, respectively. The introduction of dust control measures, such as HEPA vacuuming, during cleaning further reduced the average TWAs to 0.005 f/cc during the brake job and to 0.0021 f/cc for 8 h.

Comparison to OELs
The OSHA ceiling limit for asbestos in the workplace from 1972 to 1986 was 10 f/cc for a 15-min period (Martonik et al., 2001). There was no ceiling limit in place between June 1986 and August 1988, and from September 1988 to the present, an excursion limit of 1 f/cc for 30 min has been in place. The purposes of the short-term limit for asbestos, and most other cumulative toxicants, are to minimize the opportunity for the lifetime cumulative dose to rise above that permitted by the 8-h TWA, to reduce the nuisance associated with exposure to dusty conditions (the vast majority of dust associated with a brake change is either brake-wear debris or dust from the road), and to ensure that engineering controls are generally available to reduce exposure from these short-term activities. In fact, in reintroducing the short-term exposure limit in 1988, OSHA (1988) stated that, ''compliance would further reduce a significant health risky.'' However, while related to longterm health-based values, short-term exposures at or above the ceiling limit are not necessarily indicative of an increased acute or chronic health risk.
Most of the short-term samples for mechanics were collected during the time when the 10-f/cc OSHA standard was in place. Few of the approximately 300 short-term samples reported for mechanics were collected for a 15-min period, including those collected by NIOSH. This can complicate comparisons of the short-term exposure concentrations to the 15-min OSHA standard. Specifically, if a sample of less than 15 min is collected during a particular activity and that sample meets the standard, then one can conclude that the sample would also have met the standard if it had been collected for a full 15 min. However, if a sample of less than 15 min exceeds that standard, then it is possible that continued sample collection for a 15-min period could possibly have resulted in an exceedance of the standard. A review of the more than 130 personal samples for mechanics servicing cars and light trucks indicated that only five samples could possibly have exceeded the ceiling limit if they had persisted for 15 min. Of the more than 180 personal samples reported for brake servicing on Brake mechanics' short-term asbestos exposure Richter et al.
trucks and buses, only 10 samples had concentrations that could have exceeded the OSHA ceiling limit. Therefore, although direct comparison to the ceiling limit is hampered by a lack of samples of appropriate duration, the reported concentrations for the vast majority of short-term brake cleaning and machining samples were sufficiently low to support a conclusion that the ceiling limit would rarely have been exceeded in most brake maintenance facilities.
Comparisons to short-term limits can also be made on the basis of the dose (concentration Â time for the task). For example, assuming an exposure duration of 15 min and a maximum concentration of 10 f/cc yields an equivalent shortterm dose of 150 f/cc*min for the original excursion limit. Similarly, a duration of 30 min at a concentration of 1 f/cc yields a dose of 30 f/cc*min based on the current excursion level. As shown in Table 6, the highest calculated short-term doses were 12.8 f/cc*min for cars and light trucks mechanics and 19 f/cc*min for heavy-truck and bus mechanics. This analysis again shows that short-term doses historically experienced by mechanics cleaning brake assemblies and machining brake linings were below the acceptable shortterm dose (exposure limit) implied within the contemporaneous ceiling limit as well as the current excursion level.

Conclusions
Taking into consideration the strengths and limitations of the short-term sampling data for brake mechanics, the following conclusions were reached: Because of the lack of specificity in the reports describing these short-term samples, the substantial variability in concentration and doses, and the lack of data for some older machining activities, it would be difficult to reconstruct brake-job and 8-h mechanic exposures exclusively from these short-term data. However, by making several assumptions about the data it is possible to adjust the available 8-h TWA concentrations to include grinding of brake linings for automobiles. The average airborne concentration of asbestos for brake mechanics during the era when arc grinding was more routinely performed could have been about 60% higher than in the 1970s and later decades when arc grinding during typical brake servicing was less common. Over time, the decrease in machining activities, such as grinding, reduced the average 8-h TWAs for automobile brake mechanics from about 0.10 f/cc during the 1940s through 1960s, to 0.063 f/cc during the 1970s. The introduction of less dusty cleaning methods in the early 1980s caused a further reduction to an average of 0.0021 f/cc by the late 1980s. On the basis of the available literature, short-term exposures for brake mechanics cleaning or machining brakes were well below both the original and current OSHA standards.