Exposures to Quartz, Diesel, Dust, and Welding Fumes During Heavy and Highway Construction

ABSTRACT

Personal samples for exposure to dust, diesel exhaust, quartz, and welding fume were collected on heavy and highway construction workers. The respirable, thoracic, and inhalable fractions of dust and quartz exposures were estimated from 260 personal impactor samples. Respirable quartz exposures exceeded the National Institute for Occupational Safety and Health (NIOSH) recommended exposure limit (REL) in 7-31% of cases for the trades sampled. More than 50% of the samples in the installation of drop ceilings and wall tiles and concrete finish operations exceeded the NIOSH REL for quartz. 

Thoracic exposures to quartz and dust exceeded respirable exposures by a factor of 4.5 and 2.8, respectively. Inhalable exposures to quartz and dust exceeded respirable exposures by a factor of 25.6 and 9.3, respectively. These findings are important due to the identification of quartz as a carcinogen by the National Toxicology Program and the International Agency for Research on Cancer. Fourteen percent of the personal samples for EC (n=261), collected as a marker for diesel exhaust, exceeded the American Conference of Governmental Industrial Hygienists (ACGIH) threshold limit value (TLV®) for diesel exhaust. Seventeen of the 22 (77%) samples taken during a partially enclosed welding operation reached or exceeded the ACGIH TLV of 5 mg/m3 for welding fume.

METHODS

With the exception of a few area samples taken in the cabs of operating engineers, personal samplers covering approxi­mated 6 hours of a typical 8-hour workday were used in this study. The Construction Occupational Health Program’s sampling strat­egy requires researchers to collect data as randomly as possible, by choosing operations within a site and randomly sampling workers in those operations at the work site. Therefore, the collected sam­ples cover a wide range of construction phases, operations, and exposure levels.

Samples were collected with a number of aerosol collection devices. The first was a BGI-4 respirable cyclone (BGI, Waltham, Mass.) run at 2.2 L/min. The cyclone sampler with a heat treated quartz filter (Gelman Sciences Pallflex, Ann Arbor, Mich.) was used to collect samples for elemental and organic carbon analysis (diesel particulate). In high silica exposure operations that would overload the impactor, the cyclone was used with a 5-p.m PVC filter (SKC, Eighty-Four, Pa.). The second aerosol sampler was an IOM inhalable sampler (SKC) with a 5-pm PVC filter, run at 2 L/min. This sampler was used to collect welding fume particulate when placed near the breathing zone outside the welding hood. A third sampler was used to collect fluoride samples in the breath­ ing zone outside the welding helmet. 

This was a 37-mm closed- face cassette with a 0.8 pm cellulose ester membrane filter and a sodium carbonate treated cellulose backup pad (National Institutefor Occupational Safety and Health [NIOSH] method 7902).131 The fourth sampler was a Marple 290 series personal four-stage cascade impactor with sampling cowl and visor, which was run at 2 L/min. Personal impactor samples were collected to estimate exposures to construction dust and crystalline silica (quartz). The impactor stages used a mylar substrate (Grasby Anderson, Atlanta, Ga.) coated with less than 1 mg of Apezion L grease (Supelco, Bellefonte, Pa.) applied in a toluene solution using an artist’s air brush.

Use of the personal impactor permitted estimation of the in­halable, thoracic, and respirable size-selective particulate fractions. The method uses Simpson’s rule in a tabular-graphical approach to estimating the contribution of each impactor stage to the over­ all particle size fraction of interest. Due to the relatively wide range of particle sizes represented on each stage, this approach may contain some error; however, it represents a reasonable esti­mate of these biologically relevant size fractions. Appendix A contains the correction factor, particle size range, and midpoint and fraction of each stage used to estimate the respirable, thoracic, and inhalable particulate mass concentrations.

Sample Analysis

All particulate samples, except the diesel samples, were analyzed gravimetrically using a Cahn C-30 microbalance after equilibrating the filters for at least 3 hours in a temperature (21±1°C) and humidity (50 ±3% relative humidity) controlled environmental chamber. The limit of detection for the gravimetric analysis was 14 pg for Mylar® filters and 12 pg for PVC filters. Quartz analysis was done using a micropellet technique (5 mm) and an FTIR spectrophotometer. Spectral manipulations such as Fourier self­ deconvolution and derivatizations were performed on the aver­ aged spectrum to improve quantification.

Peak height for quartz was measured at 798 cm and 696 cm-1. Separate calibration curves were generated for each stage of the four-stage personal impactor, by using the impactor to sample quartz dust (Min-U- Sil 5 and Min-U-Sil 30) in an aerosol-gencrating loop. The limit of detection for the quartz analysis is 1 pg for respirable particles and 3 or 4 pg for thoracic and inhalable particles, respectively.

Diesel particulate samples were analyzed by evolved gas analysis using a thermal-optical analyzer (NIOSH method 5040).311 Anal­ ysis was done bv either the DataChem Laboratories (Salt Lake City, Utah) or the Sunset Laboratory (Forest Grove, Ore.) The resultant elemental and organic carbon levels are reported.

Analysis for metals and elements in the welding samples was done using ICP-AES and NIOSH method 7300 (Liberty Mu­ tual Labs, Hopkington, Mass.). Fluoride samples were analyzed using ion chromatography and NIOSH method 7906 (ESA labs, Chelmsford, Mass.).


Data Analysis

Personal exposures for construction workers were grouped by trade union of the subject and by the construction operation being performed. The trades included operating engineers who operate heavy equipment; laborers who do manual tasks throughout all phases of construction; and miscellaneous trades including iron workers, carpenters, plumbers, electricians, tile setters, piledrivers, and boilermakers who were put together because of an insufficient number in any one group.

The data were examined for the underlying distribution using the Shapiro-Wilks statistic and by graphing probability plots and histograms using the SAS System for the PC (Chapel Hill, N.C.). The data fit the lognormal distribution better than the normal distribution. Hence, the natural logarithms of the measured ex­posures were used for all analyses. Although it is common to use the geometric mean (GM) and geometric standard deviation (GSD) to describe lognormal exposure distributions, for epide­miologic purposes it is often desirable to estimate the mean ex­posure of an exposure group

However, the simple arithmetic mean is a poor estimator of the population mean when the dis­tribution is lognormal. Hence, the minimum variance unbiased estimator (MVUE) of the population mean was calculated for each operation.'361 When analytical results were less than the limit of detection (LOD) for a method, the LOD/square root of 2 was used to permit these samples to be included in the data analysis.1’71 To calculate the fraction of respirable dust that was quartz, the samples that were less than the LOD for respirable dust were ex­cluded (n= 14). There were 170 paired personal respirable dust and elemental carbon (EC) samples used to calculate the fraction of respirable dust that was diesel.

RESULTS

Samples for this study were collected during 113 site visits be­ tween June 1994 and April 1999. Examination of 260 respi­rable dust and respirable quartz concentrations by trade found them to be more lognormal than normal. Nevertheless, the dis­tributions were highly skewed, with large GSDs in many cases (Table I). There were 206 workers sampled for respirable dust and silica. The laborers had the highest mean exposures to respirable dust and to respirable quartz. The laborers and operating engi­neers had the highest percentages of quartz as part of their respi­rable dust exposures. The operating engineers had the highest per­centage of EC in their particulate and the lowest percentage of samples over either the federal OSHA PEL or the NIOSH REL for quartz. Conversely, the laborers had the highest percentage of samples over these occupational exposure limits (Table I).

Examination of the distributions of respirable dust and respi­rable quartz concentrations by operation also found them to be more lognormal than normal. Many operations, like trades, had highly skewed exposure distributions (Table II and III). The op­erations with the highest mean respirable dust exposures were found in concrete finish work, excavation support work, and in­stallation of drop ceilings and wall tiles. The highest percentages of quartz in the respirable dust were found in concrete finish work operations, demolition, and installation of drop ceilings and wall tiles. The highest percentages of EC particulate in the respirable dust were found in the excavation and concrete pouring opera­tions (Table II).

The operations with the highest mean respirable quartz expo­sures were found in concrete finish work, demolition, and instal­lation of drop ceilings and wall tiles. More than 30% of the samples taken in the operations of concrete finish work, pipejacking, and installation of drop ceilings and wall tiles exceeded the federal OSHA PEL for quartz exposure. More than 50% of the samples in the installation of drop ceilings and wall tiles and concrete finish operations exceeded the NIOSH REL for quartz (Table III).

Thoracic exposures were estimated for both dust and quartz in a variety of construction operations (Table IV). In the thoracic dust samples, the highest percentage of quartz was found in the samples from concrete pouring and demolition work, followed by pipejacking and site support operations. The highest mean tho­racic dust exposures were found in installation of drop ceilings and wall tiles, excavation support work, pipejacking, and laying conduit/pipe in trenches. The highest mean thoracic quartz expo­sures were found in pipejacking, installation of drop ceilings and wall tiles, laving conduit/pipe in trenches, and concrete finish work (Table IV).

Inhalable exposures were estimated for both dust and quartz in a variety of construction operations (Table V). In the inhalable dust samples, as with the thoracic samples, the highest percentage of quartz was found in the samples from concrete pouring and demolition work, followed by site support and pipejacking oper­ations. The highest mean inhalable dust exposures were found in installation of drop ceilings and wall tiles, laying conduit/pipe in trenches, pipejacking, as well as concrete finish work and excava­tion support work. The highest mean inhalable quartz exposures were found in pipejacking, laying conduit/pipe in trenches, con­crete pouring, and concrete finish work (Table V).

For both dust and quartz, thoracic exposures exceeded respi­rable exposures. In the case of dust it was by a factor of almost three (Table VI). In the case of quartz, thoracic exposures ex­ceeded respirable exposures by a factor of more than 4.5. It is also important to note that the regression coefficient (R2) between the thoracic and respirable concentrations was less than 0.90 (0.83 for quartz and 0.76 for dust) (Table VI ). An R2 of more than 0.90 has been suggested as the value indicating potential collinearity among predictors in multiple regression analysis. Another way to state this is that the thoracic concentration is not a simple linear function of the respirable concentration or the regression coeffi­cient (squared correlation coefficient) would be greater than 0.90.

Inhalable exposures greatly exceeded respirable exposures for both dust and quartz. For dust, the mean ratio of inhalable to respirable dust was more than 9 (Table VI). For quartz, on av­erage, the inhalable concentration exceeded the respirable quartz concentration by almost 26 fold (Table VI). As with the thoracic data, the regression coefficient (R2) between the inhalable and respirable concentrations was less than 0.90 (Table VI). The rel­atively low regression coefficients (0.42 for quartz and 0.48 for dust) suggest that these are not a simple linear function of each other (Table VI).

EC levels were measured as an indication of diesel particulate concentrations. There were 261 diesel particulate samples collect­ ed, representing 204 workers. The miscellaneous trades group had the highest mean exposure to EC, followed by the operating en­gineers (Table VII). The operations with the highest EC concen­rations were the installation of drop ceilings and wall tiles, con­crete pouring, concrete finish work, laying conduit/pipe in trenches, and excavation work (Table VII). The percentage of samples exceeding the American Conference of Governmental In­dustrial Hygienists (ACGIH) proposed threshold limit value (TLV®) for diesel exhaust of 20 p.g/m3 EC was calculated.

The miscellaneous trade group had the highest percentage of samples over the proposed TLV. The operations with the highest percent­ age of samples over the TLV included installing drop ceilings and wall tiles and concrete pouring. The miscellaneous trades group and the laborers had the highest mean exposures to organic car­ bon. Operations with the highest mean organic carbon exposures were installation of drop ceilings and wall tiles, pipejacking, and laving conduit/pile in trenches (Table VII).

Welding exposures primarily occurred in the excavation sup­port operation when steel crossbeams are welded in place (Table VIII). Seventeen of the 22 (77%) samples taken during welding operations reached or exceeded the 5 mg/m3 ACGIH TLV for welding fume. Eleven percent (2) of the fluoride samples exceeded the 2.5 mg/m’ fluoride TLV7. Sixteen percent (3) of the manganese samples reached the level of the ACGIH TLV of 0.2 mg/m3. For the remaining elements, none of the samples reached or exceeded their respective TLVs.
