Personal Exposures to Inorganic and Organic Dust in Manual Harvest of California Citrus and Table Grapes

ABSTRACT

The aim of this study was to determine characteristics of personal exposure to inorganic and organic dust during manual harvest operations of California citrus and table grapes. Personal exposures to inhalable dust and respirable dust were measured five times over a 4-month period of harvesting season. We analyzed components of the dust samples for mineralogy, respirable quartz, endotoxin, and total and culturable microorganisms. Workers manually harvesting were exposed to a complex mixture of inorganic and organic dust. Exposures for citrus harvest had geometric means of 39.7 mg/m3 for inhalable dust and 1.14 mg/m3 for respirable dust. These exposures were significantly higher than those for table grape operations and exceeded the threshold limit value for inhalable dust and respirable quartz. Exposures for table grape operations were lower than the threshold limit value, except inhalable dust exposure during leaf pulling. Considered independently, exposures to inhalable dust and respirable quartz in citrus harvest may be high enough to cause respiratory health effects. The degree of vigorous contact with foliage appeared to be a significant determining factor of exposures in manual harvesting.

METHODS

Experimental Design

Personal exposure samples were collected and analyzed from two groups of agricultural field workers in the San Joaquin Valley of California. The first group was involved in citrus harvest in Tulare County. The second group was involved in table grape operations in Kern County. Airborne dust exposures were measured for a total of 5 working days for each crop over the peak labor months of June through September 1992. Sampling dates were approximately 1 month apart, except in September, when 2 consecutive days were monitored for each crop. Tasks were selected to be representative of those performed by labor crews throughout the summer months. For citrus, harvesting was the task measured for all five sampling visits. For table grapes, the task on the first sampling visit was leaf pulling; harvesting table grapes was conducted for the remaining four sampling visits. For both crops, 14 workers were monitored during each visit. Ten workers were fitted with respirable dust samplers to measure respirable dust and respirable quartz exposures. Four workers were fitted with inhalable dust samplers to measure inhalable dust, endotoxin, and total bacteria and fungi exposures.

Selection of Farm Locations and Worker Participants

A fruit packing company in Tulare County provided access to all the citrus harvest sites. A producer of table grapes in Kern County provided access to all the sites visited for table grape production. Tulare and Kern counties were chosen because they are very significant producers of citrus and table grapes, and because they were the main field locations in a previous epidemiologic study that had found reduced forced vital capacity in grape harvest workers. From each farm location, and on each sampling date, a convenience sample of individual farm workers was used, based on the willingness of individuals to participate. No attempt was made to follow the same individuals throughout the study.

Collection of Samples

We collected inhalable dust, endotoxin, and total bacteria and fungi samples with an open face polystyrene cassette of 37 mm diameter with a 0.4 μm pore size polycarbonate filter, and a pump flow rate of 2.0 L/min. The samples were collected in the breathing zone. Sampling times for endotoxin, total bacteria and fungi, and inhalable dust were regulated with the goal of collecting approximately 1 mg of total mass. On the first sampling day, eight of nine inhalable samples exceeded a total mass of 2 mg. The concentrations for that day are probably conservative because some of the sampled material may have not adhered to the filters. Except on the first sampling day, samples were less than 2 mg of total mass. About half of the samples had less than 1 mg of total mass. For citrus harvest a typical sample time was 15 min. For table grape operations a typical sample time was 1 hour. For respirable dust and respirable quartz samples, a 10-mm nylon cyclone was used with a 5 μm pore size PVC filter, and a pump flow rate of 1.7 L/min. Respirable fraction samples were typically collected for 6 to 8 hours.

A two-stage Andersen sampler (Graseby Andersen, Atlanta, Ga.) was used to collect culturable airborne bioaerosol samples within 2 feet of employees. Fungal samples were collected on Sabouraud media. Bacterial samples were collected on trypticase soy agar with 5% sheep's blood (TSA II). Sampling times averaged 30 sec. One sample and one duplicate (consisting of two plates each) were collected daily for bacteria and for fungi. Field blanks were generated daily, and one laboratory blank was generated per medium batch.

We collected dust from leaf surfaces by agitating the foliage in front of a canopy connected to a high-volume pump (Gast Model 1532, SKC, Fullerton, Calif.). This approach was used to provide sufficient dust samples for chemical characterization of the bulk dust. The pump was operated at 30 L/min and was fitted with a 20 cm × 25 cm polycarbonate filter with 0.4 μm pore size. All dust shaken off the leaves was collected on the filter. Sampling time was approximately 90 min per sample, with two samples collected from each farm.

Four surface soil samples were collected daily. Two samples each were collected from under plants and from rows between plants. Sample locations were selected to represent the limits of the area worked by employees on that day. For example, samples would be collected from the northeast corner and the southwest corner of the section of orchard or vineyard being harvested. Samples were collected with a stainless steel trowel from approximately 0–3 inches below the soil surface and placed in plastic bags.

Analysis of Samples

Inhalable Dust, Respirable Dust, and Respirable Quartz

The filters were pre- and postweighed, controlling for moisture effects, to determine sample weight and dust concentration. Quantitative analyses of respirable quartz were performed using powder X-ray diffraction. Filters for respirable dust were ashed in a muffle furnace, and the residue from the filter was transferred to a 25 mm diameter silver membrane filter. The silver membrane was step-scanned from about 26 to 27 degrees 2 theta (quartz diffraction maximum) using a Diano XRD-8000 (Waburn, Mass.). The diffraction peak area for each sample was compared with the quartz peak area from a set of prepared standards (20 to 200 μg quartz per filter) to determine the mass of quartz and the corresponding percentage in the original dust (Method 7500). (11)

Dust from Foliage and Soil

Two samples each of the citrus dust and the grape dust were analyzed for total sulfur content by inductively coupled plasma spectrometry (Perkin Elmer Optima 3000 DV ICP, Perkin Elmer, Shelton, Conn.). The mineral composition of the foliar total dusts was determined by X-ray diffraction (Siemens D5000 X-Ray Diffractometer, Bruker AXS, Madison, Wis.) after the dusts were ashed in a muffle furnace and transferred to silver membranes. Total foliar dust samples were dry-sieved to obtain a <10 μm diameter size fraction for particle component analysis.

For particle component analysis, an aliquot of the sample was ashed in a Denton PE-120 Plasma Asher (Cherry Hill, N.J.) for 24 hours to remove all organic matter, and the residue was made up to 100 mL with double distilled water with a drop of Aerosol OT. After sonication, aliquots were filtered through 0.1 μm pore size polycarbonate filters and mounted on carbon planchettes. Using a scanning electron microscope in back-scattered electronic-imaging mode an average of 1000 or more particles in each sample were randomly analyzed for elemental analysis using an energy dispersive X-ray analyzer (Kevex 7000 EDXA System, Kevex International, Foster City, Calif.) at 1000× according to the methods described by Stettler et al. 

Soil particle size-distribution (% sand, silt, and clay) of Na-dispersed samples was determined by the standard pipette and sieving method. 

Airborne Endotoxin

Filters with collected dusts were analyzed for endotoxin content by means of the kinetic chromogenic modification of the Limulus amebocyte lysate assay (Kinetic-QCL; BioWhittaker, Walkersville, Md.) as previously described. Sterile nonpyrogenic plasticware was used throughout the analyses. Each filter was extracted separately in 10 mL of sterile nonpyrogenic water (LAL Reagent Water; BioWhittaker) by rocking at room temperature for 60 min. The extracts were decanted into separate plastic tubes, centrifuged for 10 min at 1000 × g, and the resulting supernatant fluids were assayed in duplicate for the presence of endotoxin. The results are reported in terms of endotoxin units (EU) per milligram of dust, and calculations were made to express the airborne levels in terms of EU per cubic meter of air.

Airborne Total Bacteria and Fungi

Total culturable and nonculturable bacteria and fungi were quantified using the fluorescence microscopy NFE method. For each filter collected for total fungi and bacteria, the collected microorganisms were extracted by washing three times with a filter sterilized aqueous solution of 0.01% Tween 80 and 1% formaldehyde. For each washing, 1.5 mL of wash solution was injected into the support pad through the outlet connection of the cassette, after which the connection was plugged. Five mL of the wash solution was pipetted into the inlet hold, after which the cassette was replugged and vigorously shaken on a shaking table for 5 min. The cassette was then opened and the suspension removed with a syringe. The three wash suspensions were pooled, and serial dilutions of this solution were filtered through black polycarbonate filters. The filter and the adhered microorganisms were stained for 2 min using a filtered acridine orange solution (0.1 mg/mL, pH = 7.2 in phosphate buffer).

The filters were removed, dried in a laminar flow hood, and mounted on a microscope slide with Cargile A immersion oil and a cover-slip. The number of microorganisms on the filter surface was counted at a magnification of 1000× with an epifluorescence microscope. Spores and bacteria were counted until all organisms in 40 high-power fields had been evaluated.

Airborne Culturable Bacteria and Fungi

The collected culture plates were incubated at 30°C and 25°C for bacterial plates and fungal plates, respectively. Colonies were counted using a colony counter. Results were expressed as colony forming units and were adjusted for coincident collection.
RESULTS

Inorganic and organic dust exposures were obtained for both citrus harvest and table grape operations (see Table I). Inhalable dust exposures in citrus harvest had a geometric mean of 39.7 mg/m3, whereas inhalable dust exposure in table grape operations had a geometric mean of 3.5 mg/m3. Table II shows inhalable dust exposures for citrus harvest and table grape harvest. The inhalable dust exposures in citrus harvest appeared to increase through the season. Exposures for table grape operations on June 17 were higher than the other days, despite potential loss due to overloading. Exposure on that day was measured during leaf pulling, which causes significant contact between the workers and foliage.

Geometric means of respirable dust exposures were 1.14 mg/m3 for citrus harvest and 0.23 mg/m3 for table grape operation (Table III). Geometric means of respirable quartz exposures were 0.08 mg/m3 for citrus harvest and 0.02 mg/m3 for table grape operations. The proportion of respirable quartz to respirable dust was similar in both crops. The percentage of respirable quartz in the respirable dust averaged 7% for citrus harvest samples and 9% for table grape operation samples. Respirable dust exposure for table grape operations on June 17 (leaf pulling) was higher than on the other days. However, respirable quartz exposures were not higher on June 17.

Temporal variability of personal exposures was determined by analysis of variance (ANOVA). All exposures to inhalable dust, respirable dust, and respirable quartz were significantly different among the 5 sampling days in both crops (p < 0.01). Inhalable dust exposures in citrus increased through the season, but exposures to respirable dust and quartz did not. When we excluded the exposure data from the first day of table grape leaf pulling from the ANOVA, the inhalable, respirable dust, and respirable quartz exposures were not significantly different among the 4 sampling days (p > 0.1).

Inhalable dust samples were analyzed for endotoxin and total microorganisms (Table I). Endotoxin exposure in citrus was significantly higher than in table grapes throughout sampling periods (ANOVA, p < 0.01). Geometric means of endotoxin exposures for the citrus and table grape samples were 201 and 11 EU/m3, respectively. The ratios of total and culturable microorganisms between two crops were much less substantial than the ratio in endotoxin. The geometric mean of total bacteria and fungi number in citrus was 1.32 × 10 8 organisms/m3 while the geometric mean in table grapes was 0.42 × 10 8 organisms/m3. Culturable bacteria exposures in citrus and table grapes were 8700 CFU/m3 and 6300 CFU/m3, respectively. Culturable fungi exposures in citrus and table grapes were 11,000 CFU/m3 and 6200 CFU/m3, respectively. Total bacteria and fungi were represented morphologically by 51.3% cocci, 44.3% bacilli, and 4.4% spores. The morphologic distributions were similar between citrus and table grape operations.

Correlation between dust exposures was determined by Pearson's correlation coefficients with log transformed data, using STATA software. The strongest correlations were observed between inhalable dust and endotoxin in citrus (r = 0.877) and inhalable dust and total microorganisms in table grapes (r = 0.894). The moderate correlations were between respirable dust and respirable quartz in citrus (r = 0.6944) and in table grapes (r = 0.796), between inhalable dust and total microorganisms in citrus (r = 0.677) and between fungi and bacteria in table grapes (r = 0.743). The weak correlations were between fungi and bacteria in citrus (r = 0.460) and between inhalable dust and endotoxin in table grapes (r = 0.548).

The two citrus dusts contained 2.55 and 4.36 g kg−1 total S (average 3.46 g kg−1). The two table grape dust samples contained 4.07 and 3.37 g kg−1 total S (average 3.74 g kg−1). Qualitatively, the mineralogy of the total foliar dusts from citrus and table grapes is similar (Figure 1). Both foliar dusts are dominated by quartz, feldspar, and layer silicates. Previous work showed that the citrus foliar dusts contained 32 to 37% quartz, whereas the table grape total foliar dust contained 39 to 41% quartz. 

The distribution of minerals in the <10 μm diameter foliar dust samples was similar for citrus orchards and table grape vineyards based on the electron microscope particle analysis (Table IV). In both locations, aluminum silicates predominated, followed by quartz, with the ratio of aluminum silicates to quartz of roughly 10:1. Aluminum silicates and quartz comprised almost 86% of the foliar dust. The aluminum silicates are most likely dominated by feldspars and layer silicates such as biotite and smectite (based on X-ray results, above). The percentages of quartz in this fraction were 10.1% for citrus dust and 7.9% for table grape dust. These quartz values are similar to the quartz contents of the respirable dust fractions reported above but are much lower than the quartz contents of the total foliar dusts. Quartz is physically resistant to being broken into finer particles due to its crystal structure and tends to persist in the coarser soil and dust fractions and be less abundant in finer fractions. 

We assessed correlations (least-squares linear regression) between soil particle size distribution (soil texture, proportions of sand, silt, and clay) in the citrus orchard or vineyard and exposures to respirable dust and quartz. For citrus dust we found the strongest correlation between respirable quartz and soil clay content (mg/m3 respirable quartz = 0.0015 [% clay in the soil] + 0.044, R2 = 0.72, p = 0.002). The correlation between respirable dust and clay content was weaker but still significant (R2 = 0.55, p = 0.014). For table grape dust we found no significant correlation between respirable quartz or respirable dust and soil texture when all table grape operations were included in the analysis. Leaf pulling activities early in the season produced the highest dust and quartz exposures and obscured any relationship with soil texture. When we excluded leaf pulling we found a stronger correlation between respirable dust levels and soil clay content (mg/m3 respirable dust = 0.014 [% clay in the soil] −0.045, R2 = 0.51, p = 0.048). The correlation between soil clay content and respirable quartz was much weaker (R2 = 0.24) and not significant (p = 0.21).
