N. Mydin, Petronas, Bintulu, Sarawak, Malaysia.
It has long been accepted that exposure to welding fume can cause respiratory and pulmonary illnesses. More recently studies have demonstrated an association between welding and neurological and reproductive effects. The purpose of this study is to develop and validate a model for the estimation of welding fume exposure in the workplace. A total of eight companies in Aberdeen, Scotland participated in the study. Sixty-three personal and 13 background welding fume samples arising from tungsten inert gas (TIG), metal inert gas (MIG), manual metal arc (MMA), flux-cored arc (FCA), sub-arc and oxyacetylene gas welding processes were collected. Potential determinants of exposure were observed and recorded throughout the sampling duration. Descriptive statistics and multiple linear regression techniques were used for data analyses. The geometric mean of the overall personal and background welding fume concentration were 3.64mg.m-3 (GSD=3.75) and 0.69mg.m-3 (GSD=2.47) respectively. Four welding fume exposure models combining the fume concentrations with the workplace factors such as the types of welding processes, air movement within the work area, welder’s position relative to workpiece and use of ventilation were developed using 32 of the measurement dataset. Guidance values were proposed and the models were validated using the remaining data. The percentage of variation (16.4–129%) and R-squared values (0.47–0.83) between the predicted and measured concentrations for the models were determined. Model 3 provided the best estimate of the welding fume exposure. Model 3 uses the relationship between the weighted fume concentrations, welder’s position, use of ventilation, air movement and fuming factor to estimate welding fume exposures (R-squared=0.83; % variation=16.4). The manganese and nickel content of the fume sampled was analyzed and compared with the levels predicted using the current method EH541-procedure (UK regulations) and this resulted in poor correlation (p=0.01), suggesting that EH54 is not a reliable method.
P. Williams, ChemRisk, Boulder, CO; D. Paustenbach, ChemRisk, San Francisco, CA.
Despite attempts over the past 50 years to estimate airborne concentrations of asbestos during selected job tasks or activities, there has been no known attempt to reconstruct asbestos exposures for different craftsmen employed in various (nonmanufacturing) industries. We rely upon our extensive review of the literature, coupled with site-specific data and interviews with former workers, to characterize the likely distribution of asbestos exposures for 12 different trades at a petroleum refinery from the 1940s to the present. Specifically, we estimate 8-hour TWA airborne concentrations of asbestos for each trade during six distinct time periods at the Beaumont, Texas refinery using a probabilistic (Monte Carlo) model. The primary trades of interest include insulators, pipefitters, boilermakers, masons, welders, sheet metal workers, millwrights, electricians, carpenters, painters, laborers, and maintenance workers. We find that insulators had the highest estimated exposures, with 8-hour TWA asbestos concentrations at the 50th (and 95th) percentile ranging from 3.1 f/cc (8.4 f/cc) from 1945-1965 to 0.001 f/cc (0.005 f/cc) from 1994–2004 (a 1,000-fold difference). Most other crafts had estimated exposure levels approximately 2-10 times less that that of insulators. The most notable reduction in exposures for all craftsmen occurred during the 1985-1993 time period due to several factors, including the reduction in regulatory exposure limits, increased respirator use and training at the facility, and less time working with asbestos due to the use of substitute products. The two most sensitive input parameters in the Monte Carlo analysis are (1) the task-specific asbestos concentrations which were used to calculate the 8-hour TWA estimates, and (2) the fraction of time spent working with asbestos materials. Despite a number of data limitations, this analysis provides a practical approach for how to conduct an exposure reconstruction assessment and yields specific quantitative exposure estimates that can be used in future investigations involving asbestos.
S. Paskal, MAS, Severna Park, MD.
In this study, a mechanic performed arc-grinding on asbestos-containing brake shoes in a small chamber over a short period of time. Immediately thereafter, a very small mixing fan and a bolus of sulfur hexafluoride (SF6) were introduced into the chamber. Serial asbestos and SF6 measurements were then taken over time, and the respective decay curves of asbestos and SF6 were determined. They were substantially identical, indicating that the respirable asbestos behaved like a gas in the minimal air currents of the chamber environment. Given the known volume of the chamber and solving the asbestos decay equation for time zero allowed for the estimation of the total number of asbestos fibers released by the grinding task. Next, a field study was conducted to determine the effective ventilation of an actual, three-bay automotive garage using the SF6 decay method. Finally, the field and chamber data were combined. Given the garage’s known volume and measured exposure decay curve and the per-task asbestos release data from the chamber study, it was possible to determine the likely asbestos area exposure and subsequent decay curve had the grinding been performed in the garage. Simple integral calculus allowed for the calculation of asbestos dose for any given time period. In addition, exposure patterns and cumulative doses were determined for scenarios where the grinding was repeated at various intervals in the course of a workday. It was determined that significant asbestos doses would have been inhaled by all personnel present in the garage, not just the worker conducting the grinding. The concentration-decay method employed here is particularly suited to asbestos investigations since low asbestos exposures are not measurable in workplaces comprising appreciable background dusts and no instrument exists that can provide accurate, instantaneous asbestos concentrations.
P. Sheehan, E. Goswami, J. Greene, J. Hicks, Exponent, Oakland, CA.
Although the main focus of benzene exposure assessment remains on workers in manufacturing industries producing or using substantial quantities of benzene, recently there have been questions raised about workers exposed to benzene-containing products during nonmanufacturing activities. Mechanics often use mineral spirit solvents containing benzene during parts washing and degreasing. However, there are few data on benzene concentrations in mineral spirit solvents and in breathing zone air of workers using these solvents during parts washing. This study summarizes previously unpublished benzene data for parts washing activities and provides an evaluation of mechanics’ potential daily and cumulative exposures to benzene associated with the use of mineral spirit solvents in parts washers. The distribution of the potential total dose for mechanics from inhalation and dermal uptake of benzene from the use of parts washers was calculated using Monte Carlo probabilistic methods. The 50th and 95th percentile of the dose distribution for mechanics performing a single parts washing event during a workday are estimated to be 0.07 and 1.1 mg, respectively. These are equivalent to the dose received from 8-hour time-weighted average benzene concentrations of approximately 0.002 and 0.03 ppm, which are substantially less than the current workplace threshold limit value of 0.5 ppm. Results of the study indicate most of the benzene dose is from inhalation of airborne benzene rather than dermal contact with the benzene in parts washing solvent. Additionally, the 50th and 95th percentile of the distribution of occupational cumulative dose for mechanics are estimated to be 0.006 and 0.2 ppm-years, respectively.
P. Williams, J. Knutsen, ChemRisk, Boulder, CO; D. Paustenbach, ChemRisk, San Francisco, CA.
Benzene has historically been found as a contaminant in industrial and consumer products. The purpose of the current study was to characterize benzene exposures from the use of a common penetrant and de-rusting agent that contained benzene prior to the late 1970s. Specifically, we recreated several products that were similar in physical properties and chemical composition to that used in the past, and measured the airborne concentrations of benzene during the simulated use of these products under various conditions. A total of 219 air samples were collected during 11 product use scenarios, including 15-minute background samples (N=43), 15-minute personal samples (N=88), 1-hour personal samples (N=44), and 1-hour areas samples (N=44). We find that airborne concentrations of benzene varied by 100-fold or more depending on the scenario. Specifically, benzene concentrations were found to range from 0.04–4.9 ppm for all 15-minute personal samples, 0.04–4.1 ppm for all 1-hour personal samples, and 0.01–1.9 ppm for all 1-hour area samples. Airborne concentrations were most influenced by the benzene content of the product (1%, 3%, 14%, 30%) as well as the ventilation rate (low, average, high, outdoors), but were generally not affected by the type of product (historical versus current blend) or quantity of the product used (10 ml vs. 20 ml). We also developed a linear regression model based on the air sampling data which illustrates how several key parameters affect the measured air concentrations of benzene. This model provides a relatively good fit to the data, with the coefficient of variance (R2) ranging from 0.66–0.68 for the personal samples and 0.74 for the area samples. The current study represents the first known attempt to fully characterize the plausible range of benzene exposure levels associated with the past use of this type of product, and provides quantitative data for use in retrospective exposure assessments.
J. Lavoué, M. Gérin, Université de Montréal, Montreal, PQ, Canada.
Occupational exposure databanks have been cited as sources of exposure data for exposure surveillance and exposure assessment in epidemiology. In 2001, an extract was made from the U.S. databank IMIS, off all data on formaldehyde, a substance carcinogenic to humans. The data were analysed with linear mixed-effects models. 5280 personal concentrations were available for analysis, including short term and full shift measurements. The fixed effects of the final model explained 25% of the variance of log-transformed concentrations. In addition to differences between industries, measurements decreased with the year of sampling with a different time trend for short term (estimated reduction in exposures of 19 % per year from 1979 until 1987 and then 4% per year) and full shift measurements (7% per year until 1987 and then 4% per year). A difference between short- and long-term measurements was also observed, varying between industries and in time, with the ratio of short term to full shift levels decreasing in average from 6 in 1979 to 2 in 2001. Formaldehyde levels measured during programmed inspections were marginally lower than those measured during non programmed inspections. Measurements taken during summer were marginally higher than those taken during the other seasons. High correlation was found between measurements taken during the same inspection (0.7) and low correlation between measurements in the same state (0.05). Industry and type of measurement (short term/ full shift) appeared to have the greatest influence on exposure variablility. The highest geometric means, estimated for year 2001, were 0.39 mg/m3 in the funeral service and crematories sector for short term measurements, and 0.20 mg/m3 in the lumber and wood products sector for the full shift measurements. The analysis was limited by the small number of exposure-related variables in IMIS and by the codification scheme used for nondetects and jobs.
Posted May 30, 2006