Exposure to Respirable Crystalline Silica in Eastern North Carolina Farm Workers

ABSTRACT

Occupational exposure to crystalline silica has been linked to silicosis, some forms of cancer, and certain autoimmune diseases. Little information exists on exposure levels of respirable silica in the agricultural industry. This study assessed respirable silica exposure of farm workers in eastern North Carolina. Sandy soils in this region have been shown to contain high levels of respirable silica. Personal breathing zone samples (n = 37) were collected from 27 workers at seven farms during various agricultural activities. The highest respirable silica concentrations were measured during sweet potato transplanting (3.91 +- 2.07 mg/m3 ). Respirable silica exposure was observed to be associated with agricultural activity, soil moisture, relative humidity, and wind speed. Most of the variation in exposure (79%) was explained by agricultural activity and soil moisture. The observed percentage of silica levels (mean 34.7%) were almost twice as high as was reported in studies of California agriculture. This may be due to the loamy sand and sandy loam soil types in the regions in this study. In agriculture, respirable silica exposure is highly variable, but the potential for exposures above the threshold limit value of 0.05 mg/m3 exists during particular agricultural activities.

MATERIALS AND METHODS

Study Population

Farm workers from seven different farms were recruited by the North Carolina agricultural extension agents in three eastern North Carolina counties (Pitt, Lenoir, and Wayne). These three counties were among those with the highest potential for exposure to respirable quartz due to the soil type, respirable quartz content of the soil, and amount of harvested cropland (Stopford, personal communication). All measurements were performed between May and November 1999. Written informed consent was obtained from all participants in the study. The study was approved by the institutional review boards of the University of North Carolina School of Public Health and the National Institute of Environmental Health Sciences.

Sample Collection

Personal breathing-zone measurements of approximately 4 hours in length (mean 230 min) were collected for each worker. Sampling for respirable crystalline silica was conducted as specified in National Institute for Occupational Safety and Health analytical method 7500 (Silica, Crystalline, by X-Ray Diffraction). Samples were collected using preweighed 37-mm polyvinyl chloride filters (5.0 um; SKC Inc., Eighty-Four, Pa.) housed in open-faced polystyrene filter cassettes. Aluminum cyclones (SKC Inc.) were affixed to the open-faced cassettes and provided a 50% cut point of 4.0 um. Filter assemblies were connected with Tygont tubing to personal sampling pumps (SKC model 224-PCXR4; SKC Inc.), which were field calibrated at 2.5 L/min. After sampling, cyclones were removed and filter cassettes were closed and capped for shipment. Filters were equilibrated in an environmentally controlled weighing area at the University of North Carolina-Chapel Hill for at least 2 hours before pre- and postweighing on a microbalance (SN 1114410420, Mettler-Toledo, Columbus, Ohio). This procedure provided the respirable dust concentration for each sample.

After postweighing, filters were sealed and shipped to Laboratory Corporation of America (Richmond, Va.; AIHA/PAT certification #175, ID 100531) for respirable silica analysis by X-ray powder diffraction. The analysis was conducted for the presence of three forms of silica (quartz, cristobalite, and tridymite) using quartz as the primary analyte. Results were reported for respirable silica mass and concentration, and percentage silica was determined by dividing respirable silica mass by respirable dust mass.

A portable Weather Monitor II (Davis Instruments, Hayward, Calif.) was used to record temperature, humidity, wind speed, and wind direction during each sampling period. Soil samples used for soil moisture determination were collected from each farm location in metal soil containers from the soil surface to a depth of 3-4 inches. These measurements were used as potential determinants of respirable silica exposure in the statistical modeling.

Sampling was performed at each of the seven farms at least once. The number of sampling trips to each farm varied due to the types of crops grown and weather conditions. Data on the specific activities (e.g., planting, harvesting), job functions (e.g., tractor driving), and conditions (e.g., cabbed or uncabbed tractors) were recorded for each sample using a standardized form.

Statistical Analysis

General descriptive statistics were calculated, using Microsoft Excel, for the three outcome variables (respirable dust, respirable silica, and percentage silica content). Statistical analyses were performed using SAS System Software (SAS Institute, Cary, N.C.) and evaluated at a two-tailed significance level of 0.05. Respirable silica concentrations were natural log-transformed prior to statistical analysis. Three of the 37 measurements of respirable silica concentration and percent silica were found to be below the analytical limit of detection (LOD) of 0.005 mg quartz; these were assigned a value of LOD/Ï2 prior to statistical analysis.

Potential determinants of respirable silica exposure were examined separately in univariate models using PROC GLM (SAS Institute, Cary, N.C.), with log-transformed respirable silica concentration as the dependent variable. The main determinants investigated were (1) farm, (2) county, (3) job function, and (4) environmental effects (percentage soil moisture, average air temperature, relative humidity, and average wind speed). In addition, four specific agricultural activities were examined in the univariate model: (1) cotton planting, (2) sweet potato planting, (3) mechanical harvesting (tobacco and cotton), and (4) hand harvesting (tobacco). Job function and agricultural activity were treated as categorical variables in the analysis of covariance. To reduce the number of empty cells in the data matrix, samples collected during field preparation, cultivation, and soybean planting activities (n 5 5) were excluded from the statistical analysis, bringing the sample size to 32. A forward selection method was used for model building in which each covariate was considered separately in a model, and only those covariates with p-values of less than 0.05 were used to obtain final model.


RESULTS

A total of 37 personal breathing-zone samples from 27 farm workers were collected. Measured respirable silica and respirable dust concentrations as well as percentage silica content by farm, county, agricultural activity, and crop type are given in Table I. The overall mean respirable dust, respirable silica, and percentage silica values were 1.3+-2.9 mg/m3 (n = 37), 0.7+-1.6 mg/m3 (n=34), and 34.4+-15.8% (n = 34), respectively.

Respirable dust exposures were relatively low at each of the farms, except at Farm 3, where high respirable dust and respirable silica concentrations were measured during sweet potato planting (transplanting). Four of the five workers had exposures significantly above the TLV of 3 mg/m3 for respirable particulates (containing less than 1% silica), as well as the Occupational Safety and Health Administration (OSHA) permissible exposure limit (PEL) of 5 mg/m  for nuisance respirable dust.

Figure 1 shows the percentage of samples collected at each farm, indicating exposures exceeding the TLV time-weighted average (TWA) of 0.05 mg/m3 for respirable silica. Calculations were performed for personal breathing-zone samples using two methods. First, 8-hour TWAs were calculated assuming zero exposure for the remaining time up to 8 hours. Alternatively, worst-case 8-hour TWAs were calculated assuming uniform exposure over the remaining time period up to 8 hours. Using the conservative assumption of no exposure for the remaining time up to 8 hours, 32% of the samples exceeded the TLV-TWA. Assuming uniform exposure for 8 hours, 51% of the samples (at least one sample at each farm, except at Farm 6) exceeded the TLV-TWA. When the results were compared with individually calculated OSHA PELs [PEL 5 10/(%SiO2 + 2)], 16% of the samples exceeded the PEL assuming no exposure, and 22% exceeded the PEL assuming uniform exposure.

Cotton planting, sweet potato planting, mechanical harvesting, and hand harvesting were the most common agricultural activities performed during the sampling periods. Respirable silica exposure varied considerably between these activities ranging from the limit of detection for cotton planting to 3.9 mg/m3 for sweet potato planting (Table II). Significant differences were observed between sweet potato planting and cotton planting (p<0.0001), mechanical harvesting (p<0.0001), and hand harvesting (p = 0.0001). In addition, significant differences were observed between cotton planting and hand harvesting (p = 0.0214) and mechanical harvesting (p = 0.0530). Similar respirable silica concentrations were observed in the farm workers’ breathing zones during mechanical harvesting (tobacco and cotton) and hand harvesting (tobacco) (p = 0.535).

The three major job functions performed by the farm workers were: (1) driving a cabbed tractor, (2) driving or riding on an uncabbed tractor, or (3) performing tasks on foot (Table III). Mean respirable dust and respirable silica concentrations for workers on uncabbed tractors were at least an order of magnitude higher than for the other two job functions. Significant differences were observed between uncabbed drivers and riders and cabbed tractor drivers (p = 0.017) and workers on foot (p = 0.044).

Farm, agricultural activity, and relative humidity were the most significant determinants of respirable silica exposure in the regression analyses (Table IV). All other determinants except air temperature were also significant predictors of respirable silica exposure. Because agricultural activity was observed to be the most significant predictor of respirable silica exposure in the univariate model (F-value 26.25, R2 = 0.738), it was included in the base model. Determinants for farm and county were not included in the base model and subsequent univariate analyses, because the types of agricultural activities were not consistent across the counties. Furthermore, the underlying influence of agricultural activity (i.e., the different activities performed at the different farms), not the farm or county per se, was the most likely reason for the observed variability in exposure.

The significance of the other predictor variables (job function, soil moisture, air temperature, relative humidity, and wind speed) was investigated by adding each variable into the base model separately (Table IV). Only soil moisture improved the fit of the model (F = 25.60, R2 = 0.791) and retained its significant predictive effect on respirable silica exposure (F = 6.95, p = 0.0137). Furthermore, the residual error (MSE) was reduced from 1.127 to 0.930 by the addition of soil moisture to the model (Table IV). The addition of the other predictor variables (job function, relative humidity, or wind speed) did not improve the fit of the model, as the R2 values remained essentially unchanged. Therefore, the best-fit model, which explained 79% of the variation in respirable silica exposure, included only agricultural activity and soil moisture as predictor variables.
