INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Epidemiologic studies have shown that swine confinement workers are at increased risk of developing respiratory symptoms, reductions in expired flow rates, and airways hyperresponsiveness (1). Acute changes in lung function have been reported among swine confinement workers after a 4- to 8-h work shift (8). Swine confinement workers have greater annual loss in lung function in comparison with nonfarming control subjects (11) and baseline shift changes in FEV₁ are significant predictors of annual loss in swine confinement workers (12). In a longitudinal study of a selected number of swine confinement workers, there was an apparent trend toward developing asthma (14). Elevated ambient levels of dusts from animal feed and gases from animal waste are found in the indoor environment of swine confinement buildings (15). Several cross-sectional studies have shown that dust and gases in the indoor environment are related to the respiratory health of the swine confinement workers (9, 18). Recent cross-sectional and longitudinal studies have shown that endotoxin is related to respiratory symptoms, lower lung function levels, and lung function decline in swine confinement workers (9, 13, 21). Dose-response relationships were observed between environmental measurements and cross-shift changes in pulmonary function in cross-sectional and longitudinal studies of swine confinement workers (22).

Dust particles vary in size and shape in swine confinement buildings. Particles smaller than 0.5 µm in mean aerodynamic diameter (diminutive dust) are primarily from incoming air, not produced in the barn (24). Although diminutive dust is respirable, it is likely exhaled and not deposited in the lungs. Therefore, there has been an interest in controlling "modified" respirable dust (0.5 to 5 µm in mean aerodynamic diameter) and "modified" inhalable dust (> 0.5 µm). Modified inhalable dust includes modified respirable dust. The use of canola oil sprinkling in the facilities to control dust is attractive to the swine producer because it is relatively inexpensive as well as practical and easy to apply. Spraying a mixture of 5% rapeseed oil and 95% water in swine buildings reduced dust mass concentration 60 to 90% (25). Sprinkling a mineral oil at 15 ml/m²/d at 20 kPa pressure onto the floor area of a swine building resulted in the reduction of respirable and inhalable dust concentrations in the animal room by 75% (24). Alternative oils were also tested and all were found to be suitable for sprinkling in animal buildings at various sprinkling pressures (70 to 500 kPa) and temperatures (10 to 40° C) (24). The critical pressures and temperatures were defined to prevent oil mist during sprinkling, and optimal sprinkling frequency and quantity were proposed for swine buildings (26, 27).

No previous study to our knowledge has investigated the benefit to human health from controlling the dust in swine confinement facilities by sprinkling oil. We conducted a crossover trial to investigate the differences in human health from sprinkling canola oil in a swine confinement building. The study utilized two identical swine grower/finisher rooms in the Prairie Swine Centre Inc. (Saskatoon, SK, Canada). One room was sprinkled with canola oil and the other room was used as a control area. Baseline measurements were taken on 20 naive subjects at the Centre for Agricultural Medicine. After the baseline visit, subjects spent 5 h in each barn room on two different days that were at least 7 d apart. We report here the changes in across-shift pulmonary function measurements, methacholine challenge tests, white blood cell counts, nasal lavage cell concentration, serum and nasal cytokines, and indoor environmental measurements, including dust, gas, and endotoxin concentrations.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study was conducted at the research facilities of the Centre for Agricultural Medicine and the Prairie Swine Centre Inc. over 11 d in the winter months of November and December, 1995. The study protocol and the consent form were approved by the University of Saskatchewan Advisory Committee on Ethics in Human Experimentation.

Recruitment

Twenty lifetime nonsmoking male subjects 18 to 35 yr of age were recruited by placing advertisements in local newspapers and the University of Saskatchewan Campus newspaper and by placing postings in the Royal University Hospital and University of Saskatchewan. Subjects gave informed consent before the prescreening. If female subjects had been included in the study, a larger sample size would have been required to control for sex in the statistical analysis and to have adequate statistical power to detect differences in the outcomes between the two exposure days. Therefore, female subjects were not recruited for the study.

Screening

Screening included a questionnaire that retrieved information on previous swine barn exposure, medical and allergy history, skin prick tests of reaction to pollens from ragweed, mixed grasses, tress (box elder, birch, poplar, willow, mixed), and mixed weeds; foods (milk, eggs, peanut, shell fish); dust mite; dusts from house, grain, and wheat; animals (feathers, cat, dog, cattle, horse, hog hair); molds (Aspergillus, Helmiathosporium, Cladosporium, Alternaria); control and histamine.

Subjects with previous smoking history, history of asthma or allergies, positive reaction (indicated by a 3-mm wheal) to any of the allergens in skin prick testing except histamine, history of previous swine barn exposure, or any adverse medical history were excluded from the study.

Training Day

The 20 subjects who were chosen for the study attended a training session at which they completed a consent form and a questionnaire that elucidated information on previous occupational exposures, respiratory symptoms, past illnesses including asthma, allergy, and other respiratory conditions including toxic organic dust syndrome. In an attempt to minimize the "learning effect" of performing pulmonary function measurements (28) and nasal lavage the subjects practiced these procedures during this training session. The subjects were also instructed on appropriate protocol barn entry procedures at the Prairie Swine Centre Inc.

Baseline

Baseline measurements were taken on the laboratory day at the Centre for Agricultural Medicine. The minimum duration between training day and baseline day and the subsequent two exposure days was at least 7 d. The subjects arrived at the Centre at 7:00 A.M. for pulmonary function tests. Subjects returned to the Centre at 11:00 A.M. for pulmonary function tests, blood sample, and nasal lavage. They returned again to the Centre at 4:00 P.M. for methacholine challenge studies.

Crossover Design

In order to create minimal disruption in the barns, only two subjects were assigned to each of the two animal rooms in a single day. Subjects were randomly assigned to one of the rooms, and after at least a 7-d interval, they were placed in the opposite room. A randomization list was prepared for assigning the subjects to the treatment room or the control room. Subjects were unaware of the designation of the room to which they were assigned.

Exposure Day 1

Subjects arrived at the swine barn at 7:00 A.M. for pulmonary function measurements. Then personal air samplers were attached to each subject and subjects were randomly assigned to the treatment or the control room starting at 7:30 A.M. The subjects entered the barn rooms randomly at 15-min intervals; they spent 5 h in the room, leaving every 2 h for 10 min for pulmonary function tests, which were conducted in a common room adjacent to the barn rooms. Each subject spent only 20 min out of the exposure room during the entire 5-h exposure. In order to simulate the usual work load in a typical swine barn, the subjects were asked to ride a stationary bike for 3 km at 18 km/h for each hour during their stay in the barn. In addition to taking an oral temperature reading each hour until 10:00 P.M., subjects recorded the severity of cough, phlegm, shortness of breath, chest tightness, nasal irritation, eye irritation, chills, and headache using a Likert scale ranging from 0 to 5. At the end of the exposure at 12:30 P.M., pulmonary function was measured, blood was drawn, and nasal lavage was conducted. Subjects returned to the Centre at 4:00 P.M. for methacholine challenge studies.

Exposure Day 2

After a minimum period of 7 d, subjects were assigned to the opposite room from that on Exposure Day 1. The procedures on Exposure Day 2 were identical to those on Exposure Day 1.

Animal Facilities and Management

The exposure studies were conducted in two identical swine grower/ finisher rooms at the Prairie Swine Centre Inc. Each room used in this study measured 14.3 m by 11.0 m by 3.0 m. The pen floor was partially slotted (30% of the pen area). A 0.6 m deep manure collection channel was located beneath the slotted portion of the floor. Interior walls had plywood sheathing on both sides of a stud wall frame. A bank of propeller fans (total air delivery capacity, 6,500 L /s at 20 Pa) exhausted air from one end wall of the room. Unit block air inlets were located in the ceiling. Fresh air entered the attic through screened soffit openings. Heat in each room was supplied by an unvented natural gas unit heater. An electronic controller regulated the sequencing and speed of exhaust fans, the opening area of supply inlet modules, and the operation of the heaters using a proportional control algorithm.

A total of 288 pigs were housed in the two rooms, 144 pigs in each. The average body mass of the animals was 25 ± 5 kg per pig when they were admitted to the rooms. Pellet feed was filled daily to a single-space dry feeder in each of 12 pens. Management for the two rooms was virtually identical, and production methods conformed to those commonly used in the pork industry in Saskatchewan. For example, if one pen in a treatment room was cleaned, one pen in the control room was also cleaned so that the dust in all three rooms was disturbed approximately to the same extent. In the treatment room canola oil was sprinkled daily, and in the control room no oil was sprinkled.

Oil Sprinkling

A variable sprinkling schedule was followed from the second week of admission of pigs into the rooms: 40 ml/m²/d for the first 2 d, 20 ml/m²/d for the second 2 d, and 5 ml/m²/d for the following days (27). On every fifteenth day, each room was treated with an oil "surge," on which each room was sprinkled at a rate of 20 ml/m²/d. The oil sprinkling rate remained at 5 ml/m²/d during our study period. A backpack sprayer (SP Systems SP1; SP Systems Inc., Los Angeles, CA) designed for chemical spraying was used in this study. Sprinkling was conducted at each day at 7:00 A.M., 30 min before the human subjects began entering the room. When sprinkling, the nozzle was approximately 0.8 m above the floor (pen partition level), and the entire floor area was sprinkled, including the pen area, pig bodies, and operator walkway. Canola oil is sprinkled below 0.8 m to reduce the dust originating from the feed. Dust from other areas cannot be controlled by sprinkling canola oil as sprinkling above that level could allow oil particles easier access to human and animal breathing zones.

Lung Function

A SensorMedics volume displacement spirometer (SensorMedics, Anaheim, CA) was used for lung function test measurements that were made according to American Thoracic Society specifications (29). Each subject performed the lung function tests in the sitting position with a noseclip in place. The pulmonary test variables, FVC, FEV₁, FEV₁/ FVC ratio, and maximal midexpiratory flow rate (FEF₂₅₇₅) were measured. The percentage changes in pulmonary function from the first measurement to the last measurement on baseline and on each barn exposure day were calculated.

Methacholine Challenge

Methacholine challenge studies were performed with inhalation of a diluent followed by inhalation of increasing doses of methacholine starting at 1 mg/ml, with each increment representing doubling of the dose to a maximum final concentration of 256 mg/ml (30). The bronchial challenges were performed with a Bennett Twin Jet nebulizer (Puritan Bennett Corp. of California, Carlsbad, CA) which at a driving pressure of 50 pounds per square inch (psi) produces an output of 0.13 mg/min. The subjects, in the sitting position, inhaled the nebulized solution through a mask held close, but not tightly applied, to the face and breathed the mist quietly at tidal volume for 2 min. The FEV₁ was measured at 30 and 90 s after the 2-min inhalation of methacholine, and the doses of methacholine were given at 5-min intervals. A Clinical Pulmonary Function Spirometer (MCG; Medigraphics Corp., St. Paul, MN) was used to measure FEV₁. The PC₂₀ (methacholine concentration that reduced the FEV₁ by 20%) was interpolated from the log concentration-response curve or extrapolated from the last two responses to 256 mg/ml. PC₂₀ was used as an indicator of airway responsiveness.

Nasal Lavage Procedure and Analysis

The nasal lavage procedure was adapted from Naclerio and coworkers (31). The subjects extended their necks approximately 30 degrees from the horizontal while in a sitting position. Five milliliters of normal saline (0.9%) were instilled into each nostril while the subjects did not breathe or swallow while closing the oropharynx. After a minimum of 10 s and a maximum of 2 min, the subjects flexed their necks, passively expelling the mixture of mucus and saline into a sterile specimen container, which was stored on ice until analysis. For analysis the total sample volume of the nasal lavage fluid was recorded. The sample was then centrifuged (RT600B Refrigerated Centrifuge; Sorvall, Newtown, CT) at 400 rpm for 10 min at 4° C. The supernatant (2-ml aliquots) was removed and placed into plastic tubes, which were frozen at 70° C for later analysis. The sediment was gently resuspended and washed once with phosphate-buffered saline solution, the supernatant was removed by vacuum suction, and the sediment was gently resuspended to a volume of 0.5 or 1 ml, depending on visual estimation of cellularity. Cell count was then performed on a Neubauer hemocytometer (Bright-Line, Americas Optical, Buffalo, NY). The cell count was calculated on the hemocytometer by the number of cells divided by the original fluid volume to give the number of cells per milliliter. A direct smear of the sediment was made for staining with Wright-Giemsa stain (Diff Quik; Jade Diagnostics, Aguada, PR), and another direct smear of the sediment was made for esterase stain. A cytospin preparation (Cytospin 2; Shandon Southern Instruments, Sewickley, PA) of the sediment was performed at 500 rpm for 10 min at room temperature. The cytospin serves as a gentle centrifuge to concentrate cell-poor fluids for microscopic examination (31).

White Blood Cell Methodology

Methodology of counting white blood cells utilized the impotence principle with the use of the Coulter Counter STKS model (Coulter Electronics, Hialeah, FL). Analysis and classification of white blood cells were based on the Coulter method of leukocyte differential counting using three measurements: individual cell volume (V), high-frequency conductivity (C), and laser light scatter (S). The well-clotted blood sample was centrifuged at room temperature at a speed of 1,500 rpm for 10 min, which separated the serum from the cells. The serum was then divided into 1-ml aliquots and placed into a polypropylene container. The sample was then frozen at 70 to 80° C for later cytokine analysis.

Nasal and Serum Cytokines

Proinflammatory cytokines IL-1, IL-6, and IL-8 were measured in cell-free nasal wash supernatant and in sera by commercially available immunoassay kits (IL-1 and IL-8 from Perseptive Diagnostics, Cambridge, MA and IL-6 "high sensitivity" immunoassay from R&D Systems, Minneapolis, MN). These cytokines were chosen to examine acute-phase response and to confirm the previous finding that the reactions to swine dust might be mediated by the cytokine IL-6 (32).

Measurements of Air Quality and Environment

Dust mass concentration (mg/m³) was measured continuously by using personal aerosol samplers (Dupont Air Sampler Model P4L; Canada Safety Supply, Saskatoon, SK), with a 37-mm glass fiber filter of 0.8 µm porosity (Type A-E, no. 61652; Gelman, Montréal, PQ) at a flow rate of 3.94 L/min, which were carried by the study subjects. The air inlet of the sampler was attached at the shoulder near the subject's nose level. Dust filters were dried and weighed before and after the dust collection using an electronic balance. In addition to personal aerosol sampling, dust particle counts (particles /ml) were also measured using a laser particle counter (Model 227B; MetOne Inc., Grants Pass, OR), which was hung from the ceiling in the center of the room at 1.6 m above the floor, at the human breathing zone. The laser particle counter provided four particle-size (in optical diameter) ranges: 0.3 to 0.49 µm (diminutive), 0.5 to 0.99 and 1.0 to 4.99 µm (modified respirable), and > 0.5 µm (modified inhalable dust). Modified inhalable dust includes modified respirable dust. Counts were taken twice a day at 8:00 A.M. and at 12:30 P.M. Five measurements lasting 20 s each were taken at each of the measurement times. Means of dust particle counts taken at 8:00 A.M. and at 12:30 P.M. were used as indicators of respirable and inhalable dust particles in the control and treatment rooms during the exposure period.

Aerial dust mass concentration was measured using an air sampler (Gilian AirCon2; Gilian Instrument Corp., Caldwell, NJ). An air sample was drawn into the sampler through a 37-mm glass fiber filter (Gelman, Type A-E, no. 61652), which has a porosity of 0.8 µm. Gas levels were measured in treated and control rooms on each exposure day.

Hydrogen sulfide concentrations were measured every 10 s using a Toxilog atmospheric monitor (Model 1801 DL; Biosystems Inc., Rockfall, CT). An air sample was collected continuously into an air bag (Model H-01409-14; Cole-Parmer, Chicago, IL), and then colorimetric tubes were used to measure the daily mean ammonia (Model 105Sd; Matheson Gas Products, Edmonton, AB) and carbon dioxide (Model 126SC; Matheson Gas Products) concentrations. Three tubes were used for each air sample, and a mean value was obtained from the three readings.

Room temperatures and pressures across the exhaust fans and air inlets were measured and recorded continuously. Relative humidity was measured using a psychrometer located at the center of the operator alley. Air velocity was measured using an anemometer at animal levels in three pens located at the center and at both ends of the room. Ventilation rates were calculated from room temperature, fan schedule, pressure, and fan curves. As an additional indicator of ventilation rate, carbon dioxide concentrations were also measured in the two experimental rooms.

Endotoxin analysis was performed in the Immunology Section, National Institute for Occupational Safety and Health (Morgantown, WV) (33, 34). (We are grateful to Dr. S. Olenchok for facilitating endotoxin analysis.) Filters with collected dusts were extracted individually in 10 ml of sterile, nonpyrogenic water (LAL Reagent Water; BioWhittaker, Walkersville, MD) in the original 50-ml centrifuge tube by rocking at room temperature for 60 min (Labquake shaker; Labindustries, Berkeley, CA). The extracts were then centrifuged at 1,000 g for 10 min, and dilutions of the supernatant fluids were analyzed in duplicate for gram-negative bacterial endotoxin by the kinetic chromogenic modification of the Limulus amebocyte lysate assay (Kinetic-QCL; BioWhittaker). The concentrations of endotoxin in dust were converted to concentrations of endotoxin in air and reported in endotoxin units per cubic meter (EU/m³).

Statistical Methods

Two independent sample t tests were used to compare dust and gas concentrations between treatment and control rooms. Statistical significance of differences in the measurements between baseline and the two exposure days were tested by using an F test based on Pillai's statistic for analysis of variance of repeated measures (35). Paired t tests were used for comparison of measurements between any 2 d. Because the statistical comparisons were preplanned, p values were not adjusted for multiple comparisons. The statistical testings were carried out on pulmonary function measurements: FEV₁, FVC, FEV₁/ FVC ratio and FEF₂₅₇₅, PC₂₀, blood counts, nasal lavage cell counts, and serum and nasal lavage cytokines. Shapiro-Wilk tests were used to test whether the distributions of environmental variables were normally distributed (36). If the distribution of an environmental variable was not normally distributed, then the variable was log-transformed for statistical testing.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Demographic Characteristics

Twenty nonsmoking male subjects participated in the study. Mean values for age, weight, and height were 23.9 yr (SD, 4.6 yr), 83.5 kg (SD, 18.4 kg), and 177.2 cm (SD, 7.0 cm), respectively. Thirteen of the 20 subjects were university students.

Dust and Gas Concentrations

Mean dust and gas concentrations in treatment and control rooms are shown in Table 1. Mean modified respirable and inhalable dust particle counts and dust and endotoxin levels in the samples collected by area and personal samplers were significantly lower in the treatment room than in the control room. Interestingly, mean diminutive particle counts and endotoxin concentrations (EU/mg) measured by personal samplers were significantly greater in the treatment room than in the control room. When considering gases, mean ammonia, carbon dioxide, and hydrogen sulphide levels were lower in the treatment room than in the control room, but only the difference in ammonia levels between the treatment and control rooms was statistically significant (p < 0.001).

Symptom Scores

Mean differences between symptom score before entering the barn (7:00 A.M.) and the mean symptom score during the exposure (8:00 A.M. to 1:00 P.M.) in the treatment and control rooms are shown in Figure 1. Subjects recorded the severity of symptoms on a Likert scale before entering the barn and at each hour during exposure. The Likert scale ranged from zero for no symptoms experienced to 5 for severe symptoms. As seen in Figure 1, differences in scores were significant for the respiratory symptoms: cough, chest tightness, phlegm, shortness of breath, and nasal congestion. No significant differences were observed in scores for headache.

View larger version (22K): [in this window] [in a new window]   Figure 1.   Mean differences between the symptom score before entering the barn (7:00 A.M.) and the mean score during the exposure (8:00 A.M. to 1:00 P.M.) in treatment and control barns. Definition of abbreviations: Chest = chest tightness; SOB = shortness of breath; NASAL = nasal congestion.

Pulmonary Function Test Values

As shown in Figure 2 mean percent shift change in FEV₁ (standard error) for the 20 subjects was: 9.9% (1.12%) in the control room in comparison with the smaller mean percent shift change of 1.9% (0.63%) in the treatment room, and a positive mean shift change of 1.1% (0.63%) on the baseline day. The differences in mean shift change in FEV₁ were statistically significant: control versus treatment (p < 0.001), baseline versus control (p = 0.001), and baseline versus treatment (p < 0.001). Similar differences were observed for FVC, FEV₁/ FVC ratio, and FEF₂₅₇₅ (Table 2).

View larger version (12K): [in this window] [in a new window]   Figure 2.   Shift changes in FEV1 on baseline day and during exposure to treatment and control rooms. Pillai's test, p < 0.001. *p < 0.0001, treatment day compared with baseline day. **p < 0.0001, control day compared with baseline day and p = 0.001, control day compared with treatment day.

Bronchial Responsiveness

Subjects were more responsive to methacholine after exposure to the control room than after exposure to the treatment room. Airway responsiveness is expressed as that concentration of methacholine that will provoke a fall from baseline FEV₁ of 20% (PC₂₀). As shown in Figure 3, the mean value for PC₂₀ on the baseline day was 340.0 mg/ml (SE, 34.3 mg/ml), which was significantly greater than that observed on the treatment day (243.3 mg/ml; SE, 40.7 mg/ml; p = 0.04) and on the control day (158.5 mg/ml; SE, 36.0 mg/ml; p < 0.0001). Nevertheless, the differences in PC₂₀ after exposure to treatment and control rooms was not statistically significant (p = 0.06).

View larger version (12K): [in this window] [in a new window]   Figure 3.   Methacholine concentrations that reduced FEV1 by 20% (PC20) on baseline day and after exposure to treatment and control rooms. Pillai's test, p = 0.001. *p = 0.04, treatment day compared with baseline day. **p = 0.006, control day compared with baseline day.

Nasal Lavage Cell Counts

After exposure to the control room, subjects showed a higher mean concentration of total cells in nasal lavage (42,895 cells/ ml; SE, 8,802 cells/ml) compared with that in the treatment room (10,186 cells/ml; SE, 3,413 cells/ml; p = 0.001) and at baseline (15,628 cells/ml; SE, 6,449 cells/ml; p = 0.004). Concentration of total cells in nasal lavage on the baseline day was not significantly different from that observed after exposure to the treatment room (p = 0.12).

White Blood Cell Counts

White blood cell count was significantly greater in the subjects after exposure to the control room than on other occasions (Table 3). No significant difference was observed in white blood cell counts between the baseline day and after exposure to the treatment room. Mean neutrophil count after exposure to the control room was almost twice that observed on the baseline day and after exposure to the treatment room. Interestingly, neutrophil counts on the baseline day and after exposure to the treatment room were very similar. Lymphocyte counts were significantly reduced after exposure to the control room (Table 3) compared with those at baseline and after exposure to the treatment room. No significant differences were observed in eosinophil or monocyte counts.

Nasal and Serum Cytokines

Increases in IL-1, IL-6, and IL-8 were observed in nasal wash supernatant after exposure to the control room (Table 4). However, those levels were significantly lower on the baseline day (p < 0.0001) and when subjects were exposed to the treatment room (p < 0.001). IL-8 and IL-1 levels in blood sera after exposure to the control room were not significantly different from those observed on the baseline day and after exposure to the treatment room, whereas IL-6 blood content increased significantly, from 1.51 pg/ml on the baseline day to 10.0 pg/ml after exposure to the control room.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Several issues arise from the results that we have presented. These include: (1) effects of acute exposures on various indicators of human health, (2) the reproducibility of these indicators of human health in an intervention crossover design, (3) remarkable reduction in airborne particulate from the relatively inexpensive and simple intervention of sprinkling canola oil, and (4) accompanying reduction in observed health effects across a number of human health indicators.

In this study, we have confirmed and extended the observations of others on the human health effects of acute exposures to airborne particulate in swine-confinement facilities. As demonstrated in Table 1, the environment in the control room at the Prairie Swine Centre Inc. showed relatively high levels of dusts, endotoxin, and ammonia, and these concentrations were presumably causal for the significant increases in almost all the health indicators over baseline values.

We have recently demonstrated that over a 4-yr period, relatively young workers in swine-confinement facilities lose approximately 26 ml in FEV₁ and 34 ml in FVC per year over nonfarming rural-dwelling men who are not exposed, after adjusting for potential confounders (11). In a longitudinal study conducted in Iowa, swine workers had a persistently lower FEV₁ in comparison with neighborhood farming control subjects during the 2-yr period (13). However, no accelerated loss of lung function was observed among swine farmers. Shift change in FEV₁ at baseline study has also shown to be a predictor of longitudinal decline in these workers (12, 13). Thus, it seems reasonable to assume that an acute shift change in FEV₁ might also predict future development of chronic airway limitation in exposed workers. The cross-shift changes in FEV₁ and FVC that we observed in the control room, the room that had not been treated with canola oil, are similar to those that have been described previously by ourselves (12) and by others (8, 10, 13).

In the study that we herein report, all the health indicators showed a remarkably directional similarity of shift-change and intervention response, including spirometry, methacholine challenge studies, nasal lavage, and white blood cell count measurements. We and others have demonstrated increases in tests of airway responsiveness in swine workers versus nonexposed control subjects (6), and acute increases in tests of airway responsiveness on exposure to swine-barn contaminants (7), in all likelihood, as a result of the inflammatory response that has been demonstrated on acute exposures (9).

Our recent study has shown that exposure of volunteers to a swine-confinement building induces an intense airway inflammation characterized by an increased cellular infiltration as observed by nasal lavage, bronchoalveolar lavage, and blood tests and that these observations are quite reproducible between exposures (37). In the present study, a fourfold increase was observed in the concentration of total cells in nasal lavage after exposure to the control room in comparison with the baseline day and after exposure to the treatment room. A similar trend in another biologic exposure situation, that of grain dust, has been shown by a twofold increase in total cell concentration in nasal lavage, when these grain workers were compared with postal workers (38).

In the present study, a twofold increase was observed in blood neutrophil counts after exposure to the control room. Increases have also been reported in neutrophil cell counts in nasal lavage in grain workers (38) and in bronchial lavage among healthy subjects exposed to swine-confinement building (37) or to swine dust (7). However, no significant increases were found in blood eosinophils after exposure to the control room in our study.

With regard to nasal mucosal function on exposure, a clear increase in IL-8, a specific chemotactic factor for neutrophils was detected in nasal lavage fluid. Increases in IL-1 and IL-6 proinflammatory cytokines were also observed. It was demonstrated that after migration of leukocytes to the local target site, these cells release their own chemoattractant mediators, thus expanding the inflammatory response (39). The increase in inflammatory cytokines was most evident when the subjects were exposed to the control room compared with the nonexposed baseline evaluation or the treatment room. In the serum, no significant increase in IL-8 or IL-1 was detected, but a clear increase of IL-6 was measured, suggesting that the systemic reactions to swine dust might be mediated by IL-6. Malmberg and coworkers (32) reported an increase in IL-6 production in serum after exposure of naive subjects to a swine-confinement building. IL-6 has been found in elevated amounts in blood in autoimmune diseases, malignancy, and sepsis as well (40). The inability to detect IL-1 and IL-8 in serum might be due to the presence of antagonists such as IL-1 receptor antagonist (IL-IRA) that might interfere with the acute reaction and serve as regulator of inflammation (41, 42). A likely mechanism for the accumulation of inflammatory cells is an increased production of proinflammatory mediators from structural cells such as fibroblasts and epithelial cells after exposure to airborne contaminants in the barns.

In our study, ammonia, carbon dioxide, and hydrogen sulfide concentrations were lower in the room treated with canola oil in comparison with those observed in the control room. The reasons for the reductions in gas concentrations are not known, but some gas molecules might have become attached to the dust particles sprinkled with oil and emission of gases from the floor and fecal materials might have been reduced because of the "masking" effect of the oil film. Although mean respirable and inhalable particle counts, dust concentrations, and airborne endotoxin levels were lower in the treated room, diminutive particle counts and endotoxin concentration (EU/mg) measured by personal samplers were greater in the treated room than those observed in the control room. The greater concentration of diminutive particles in the treatment room may have been due to the agglomeration effect. When concentration of large particles was reduced in the treatment room, diminutive particles, which are primarily from the incoming air, have less chance to agglomerate with the large particles. The reasons for increased endotoxin concentration (EU/mg) in the treated room need further investigation, but an explanation for the increase may be that dust in the treated room has a higher fecal material content, which in turn has a higher endotoxin content.

The endotoxin levels that we describe are higher than those reported by others in Iowa (13, 22, 23). Part of the explanation could be the differences in outdoor temperature and consequent reduction in ventilation rates that occurred in order to conserve heat in the "finisher" barns that we studied. Our study took place in December 1995 with outdoor temperature varying between 31.5° C and 5.6° C (mean temperature, 17.6° C). The reduction of ventilation rates was suggested by the indoor carbon dioxide levels of approximately 3,700 ppm in the facilities that we studied, as compared with values reported by Kiekhaefer and coworkers (43) of 1,930 ppm for a "finisher" barn at 2.2° C in Iowa.

Zhang and coworkers (26) investigated sprinkled oil particles in the air by collecting samples on glass plates at different heights during the oil sprinkling (zero sample) and 10 s after the oil sprinkling (residue sample). They found oil mist particles in zero and residue samples collected at 0.1 and 0.8 m above the floor, which is within the animal breathing zone. However, they did not find oil mist particles at 1.6 m above the floor level, which is within the human breathing zone. As the oil particles were greater than 44 µm in diameter, they would not reach a height of 1.6 m and remain suspended in the air after the sprinkling. Although the oil particles may be inhaled by the pigs, pathologic examinations of pigs have shown that sprinkling rapeseed oil has no effects on the lungs and that there was no evidence of oil resorption to lung tissues, lymph nodes, or upper respiratory systems in the pigs (27).

In our study, we have shown that sprinkling canola oil reduces acute health effects of healthy naive subjects exposed to airborne contaminants in swine barns. Long-term health effects of sprinkling canola oil to swine-confinement workers and animals are not known. Further studies are required to investigate the long-term effects prior to recommending sprinkling canola oil as a method for controlling the environment in swine-confinement buildings.