| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of General Practice and Social Medicine (Drs. Vogelzang, van der Gulden, and van Schayck), and the Department of Pulmonology Dekkerswald (Dr. Folgering), University of Nijmegen; the Department of Environmental and Occupational Health (Dr. Heederik), University of Wageningen; and the Department of Herd Health and Reproduction (Dr. Tielen), University of Utrecht, The Netherlands.
Correspondence to: Peter F. J. Vogelzang, MD, PhD, Department of General Practice and Social Medicine, HSV 229, University of Nijmegen, PO Box 9101, 6500 HB Nijmegen, The Netherlands; e-mail P.Vogelzang@hsv.kun.nl
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: A cohort of 171 pig farmers was followed over a 3-year period. Bronchial responsiveness was assessed by a histamine provocation test. Long-term average exposure to inhalable dust and endotoxin was determined by personal monitoring in summer and winter, using data on farm characteristics and activities. Time-weighted average (TWA) personal exposure to ammonia was measured. Data on farm characteristics were gathered in the same period.
Results: Mean increase in responsiveness was 2.52 doubling concentrations of histamine for a 10% decrease in FEV1 and 1.36 doubling concentrations for a 20% decrease in FEV1. Long-term average exposure to dust was 2.63 mg/m3 and to endotoxin was 105 ng/m3. TWA exposure to ammonia was 1.60 mg/m3. After adjusting for age and smoking behavior, long-term average exposure to inhalable dust was associated with increases in bronchial responsiveness expressed as steps for provocative concentration causing 10% fall in FEV1. TWA exposure to ammonia, use of wood shavings as bedding, and automated dry feeding were associated with increases in responsiveness expressed as steps for provocative concentration causing a 20% fall in FEV1.
Conclusions: Exposure to dust and ammonia in pig farms contributes to chronic inflammation of the airways and should be reduced.
Key Words: ammonia • occupational epidemiology • occupational exposure • organic dust • pig farmer
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Bronchial responsiveness was shown to be associated cross-sectionally with certain specific characteristics of the pig farm (quaternary ammonium compounds, wood shavings, automated dry feeding), but not with measured personal exposure of the farmer.13 Bias from various sources makes it difficult to interpret cross-sectional studies reporting associations between exposure and bronchial responsiveness.11 12 13 To assess the effect of exposure, longitudinal studies that suffer less from such bias are required.
The present study evaluates longitudinal changes in bronchial responsiveness in a 3-year follow-up in a cohort of 171 pig farmers. Associations with exposure to inhalable dust, endotoxin, ammonia, and farm characteristics determining exposure are shown to exist.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There were no systematic differences in age, number of years worked as pig farmer, baseline FEV1, smoking behavior or type of farm between the 171 pig farmers completing the follow-up, and those either not included in the cohort (n = 201) or not completing the follow-up (n = 27).
Histamine Challenge
Lung function was measured according to European Respiratory
Society guidelines,15
and corrected to body temperature
and pressure, saturated with water vapor.
Bronchial responsiveness was assessed by a shortened procedure as described by Hargreave et al.16 Participants inhaled aerosolized histamine for 2 min, starting with saline solution control, and then with doubling concentrations from 0.03 to 16 mg/mL histamine. The nebulizer was calibrated for an output of 0.13 mL/min. The provocative concentration causing a 10% fall in FEV1 (PC10) and the prococative concentration causing a 20% fall in FEV1 (PC20) were assessed by linear intrapolation in a log-dose-response plot. In case a 10% decrease was reached after inhaling saline solution control, PC10 was arbitrarily set at 0.01. If a 10% decrease in FEV1 was not reached at 16 mg/mL, PC10 was arbitrarily set at 32 mg/mL. This is similar to the approach used by Rijcken and coauthors.17
The study was approved by the Committee for Ethical Research of the University of Nijmegen.
Validation
To validate the comparison of data from 1992 and 1995, a short
validation experiment was carried out. The experimental setups were
reconstructed, and five volunteers were tested between 2 days and 5
days apart using the actual nebulizers of 1992 and 1995 (which were
both still available) and the same study protocol. As the pig farmers
in the study did not have severely obstructed airways, healthy
volunteers were chosen for this validation as well.
Exposure and Farm Characteristics
Measurements of exposure were carried out on the farms of all
participants during full work shifts of on average 8.3 h on 2 days
in summer 1991 and winter 1992. Personal exposure to inhalable dust
(50% cutoff diameter of 30 µm, meeting requirements for total dust
sampling) was determined using a dust sampler with a 6-mm diameter
inlet opening and an airflow of 2 L/min. Teflon filters with a pore
size of 1 µm (Millipore; Bedford, MA) were used. Endotoxin in the
inhalable dust samples was analyzed with a modified kinetic Limulus
amoebocyte lysate test.18
Methods for measurement of dust
and endotoxin are described in detail elsewhere.19
20
Because the day-to-day variations of exposure to dust in individual
participants were considerable compared with the variations of exposure
in the entire group, the long-term average exposure was predicted by a
mathematic modeling technique: long-term average exposure to dust and
endotoxin of each individual farmer was estimated using data on farm
characteristics and time spent on activities in pig farming of all
cohort members combined. A separate article on this approach
illustrated that the exposure predicted with the modeling approach
correlates well with the average exposure measured over more than eight
occasions, and that error is reduced and remains random.21
Personal ammonia exposure was determined in duplicate by using a passive monitoring method.22 Samples were analyzed according to a modified indophenol detection method, and concentration was assessed by spectrophotometry. Results were expressed as time-weighted average (TWA) exposure. From the total number of 171 cohort members, complete data on exposure to dust, endotoxin, or ammonia were available for 146, 146, and 140 participants, respectively.
In the winter of 1992, an inventory of farm characteristics was made by walk-through surveys on the farms of participants. In the summer as well as in the winter, the farmers completed a diary on time spent on activities in pig farming, thus covering two 7-day periods. Based on the information from the surveys and the diaries, the average time a farmer was exposed to a particular characteristic was arrived at. The methods used to estimate exposure to these farm characteristics are fully described by Preller et al.18
Analysis
Changes in bronchial responsiveness between 1992 and 1995
were calculated by taking the log of base 2 of the ratio of the two
concentrations causing a fall of 10% or 20%, respectively, in
FEV1: 
PC10 = 2log
(PC10 1995/PC10 1992)
In this way, the changes were expressed in terms of doubling concentrations of histamine.
As exposure was log-normally distributed, log-transformed values with base 2 were used in the analyses of associations with health effects. Bronchial responsiveness can show fluctuations over short periods, especially following occasional heavy exposure.11 12 Pig farmers experiencing heavy exposure on one of the days with a histamine responsiveness test, therefore, may show very large increases or decreases during the follow-up. This may seriously distort estimates of longitudinal decline and consequently hamper exposure-response analysis. Therefore, trimmed means were used in exposure-response analyses, by excluding 12 pig farmers prior to analyses with concentration steps PC10, and 6 prior to analyses with concentration steps PC20.23 Mean values of the exposure variables were not affected by this procedure. Most pig farmers were not hyperreactive, at least at the start of the follow-up. Room for decreases in responsiveness was limited, as we used no concentration of histamine > 16 mg/mL. Consequently, more farmers with very large increases in responsiveness were excluded from exposure-response analysis than farmers with very large decreases in responsiveness.
Associations between exposure and increases in responsiveness were tested with linear regression analysis, adjusting for age and smoking behavior, the latter defined as pack-years of cigarettes. For current and past cigarette smokers, the numbers of pack-years were calculated by the number of cigarettes smoked per day multiplied by the number of years smoked, divided by 25. Initial small caliber is associated with greater responsiveness, as any further contraction of obstructed airways inevitably leads to a more rapid decline in diameter. Entering baseline FEV1 into multivariate analysis is done to adjust for a confounding effect from initial caliber. However, prevalence of small caliber was low in this group of basically healthy, working men, and consequently the geometrical effect on responsiveness could not play a major role. As baseline FEV1 was itself associated with exposure,22 associations between exposure and increases in responsiveness were not adjusted for baseline FEV1. As baseline atopy (IgE to common allergens measured in sera) had little effect on the associations tested (< 10% change for any statistically significant point estimate), it was left out of the final models as well.
In the regression analysis with farm characteristics, the farmers who used a certain characteristic were compared to those without. For every characteristic, the group not using it formed the reference category. Our previous reports had provided ample evidence that these characteristics are detrimental, both to respiratory health of pig farmers in general and also to responsiveness cross-sectionally.13 14 22 Considering the a priori evidence, a one-sided test was chosen. By using a multivariate regression, the results could be adjusted for confounders and the other farm characteristics.
Computations were completed with Statistix for Windows (Analytical Software; Tallahassee, FL) for personal computer.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
80%
predicted at the start of the follow-up (n = 16) were excluded, the
temporal trend in responsiveness is largely unaffected (2.44
concentration steps for PC10 and 1.16 for
PC20). The temporal trend in responsiveness was
seen in both the highest- and lowest-exposed pig farmers. Responsiveness in 1995 correlated with the number of years worked as pig farmer (Spearman R2 for PC10, - 0.23, p < 0.05; for PC20, - 0.26, p < 0.05), which we observed already in 1992.13
Validation
The individual results for PC20 were within
the two- to threefold reproducibility required by European Respiratory
Society standards.15
The mean results from the five
volunteers, however, showed a nonsignificant tendency toward greater
responsiveness when using the 1995 nebulizer (1.3 concentration steps
for PC10 and 1.1 concentration steps for
PC20), suggesting that the histamine yield of the
1992 nebulizer may have been smaller than that of the 1995 nebulizer.
We cannot exclude, therefore, that the increase in responsiveness was
overestimated to some extent, despite the use of the same protocol,
setup, and type of nebulizer at the two measurements. If the
differences found at the validation experiment were applied as a
correction factor to the measured increases in responsiveness, the
corrected mean increase in responsiveness was 1.48 doubling
concentrations for a 10% decrease in FEV1
(p < 0.001) and 0.72 doubling concentrations for a 20% decrease in
FEV1 (p < 0.001).
Exposure and Associations With Respiratory Effects
Estimated long-term average exposure to inhalable dust was 2.63
mg/m3 (geometric mean) and to endotoxin was 105
ng/m3. TWA exposure to ammonia was 1.60
mg/m3 (Table 1)
.
Increases in bronchial responsiveness in the 3-year period, measured as doubling concentration steps for PC10, were significantly associated with exposure to inhalable dust, but not with exposure to endotoxin and not significantly to ammonia (Table 2 ). The associations with exposure were stronger for the consistently symptomatic farmers than for those without symptoms. The association between dust exposure and increases in responsiveness, based on these data, is represented in Figure 3 .
|
|
|
If the associations with exposure were reanalyzed with differences corrected for the results of the validation experiment, the strength of associations was of the same order and remained statistically significant.
Farm Characteristics and Associations With Respiratory Effects
Both the use of wood shavings as bedding material and the use of
an automated dry feeding system were associated with a larger increase
in bronchial responsiveness, expressed as concentration steps for
PC20 (Table 3
). The use of quaternary ammonium compounds as active substance in
disinfectants was not associated with increases in responsiveness,
though it was cross-sectionally associated with hyperresponsiveness
both in 199213
and in 1995 (not shown). As was the case
with exposure to dust and ammonia, associations tended to be stronger
for the group of symptomatic pig farmers. Associations were weaker for
responsiveness when expressed as concentration steps
PC10, but pointed in the same direction.
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Log-transformed exposure with base 2 was used in analysis. This means that the regression coefficients in Table 2 represent the effect of an increase in exposure with a factor 2. For instance a doubling of long-term average exposure to inhalable dust was associated with an extra increase in responsiveness of 0.59 concentration steps for PC10 during follow-up. For the farm characteristics in Table 3 , the regression coefficients represent the effect of using the characteristic. For instance, the farmers using wood shavings as bedding (n = 5) had an extra increase in responsiveness of 1.54 concentration steps for PC20 during follow up, when compared to the farmers who used no bedding (n = 130).
The increase in responsiveness over the 3-year period was very large. At this rate of increase, all farmers should have been hyperresponsive by now considering the average of 16.7 years experience in pig farming. So, the actual increase in responsiveness in this group of pig farmers must have been smaller than the measured results. That the pig farmers did indeed increase in responsiveness, though at a smaller rate, is supported by the following arguments:
The associations with exposure were not affected by the potential bias in the results. This could have occurred if the magnitude of bias was determined by exposure. However, the trend in responsiveness was seen in both higher and lower exposed farmers and was unaffected when obstructed pig farmers were excluded. Such differential bias is, therefore, highly unlikely. This supports the notion that the exposure in pig farming leads to prolonged inflammation of the airways.
Some of the observed increase in bronchial responsiveness may have been
due to the reduction in FEV1. As shown in Table 1
, the symptomatic farmers had a lower FEV1 at
the start than the asymptomatic ones and had a larger increase in
responsiveness. This could lead to overestimation of the direct effect
of exposure on responsiveness. However, exclusion of pig farmers with a
baseline FEV1 80% predicted left the
temporal trend in responsiveness unaffected. Furthermore, inclusion of
baseline FEV1 in the analyses of associations
with exposure yielded lower results for ammonia exposure only,
demonstrating that this factor had a limited influence on the results.
Large changes in exposure during the observation period are a potential explanation for a rapid increase in responsiveness. However, at the second medical examination, questions on farming methods were asked and the results suggested no major shifts in exposure.
The effect of endotoxin on lung function in pig farmers is very
clear.1
2
A possible role of other dust components is
less established. In the present study, we found no association between
endotoxin exposure and increases in responsiveness, but we did for dust
exposure, though not consistently for PC10 and
PC20. Earlier, we have shown that besides
endotoxin, inhalable dust has an effect on lung function as
well.2
This supports the hypothesis that more substances
in organic dust than endotoxin alone, for instance ß(1
3)-D-glucan
from fungi25
and peptidoglycan from Gram-positive
bacteria,26
contribute to the development of respiratory
disease in exposed workers. This is further corroborated by the
observation that the use of automated dry feeding systems, known to be
associated with higher exposure to dust18
27
and with
health effects,13
14
28
contributed to increases in
responsiveness. In many countries, occupational threshold exposure
levels for exposure to dust are as high as 10
mg/m3. The threshold limit value for grain dust
in the United States is 4 mg/m3. Recently, a lower
threshold of 2.8 mg/m3 for organic dust was
proposed.3
The health effects demonstrated in this study,
at an average exposure level to dust of 2.6 mg/m3, suggest
that such a threshold level for organic dust exposure should certainly
not be higher.
Associations between TWA exposure to ammonia and increases in responsiveness were significant for concentration steps PC20. Earlier, we demonstrated an inverse association between ammonia exposure and baseline lung function in the cross-sectional part of our studies. This was strongest for the symptomatic farmers as well.22 Based on associations between ammonia exposure and FEV1 decrements over a work shift, a threshold of 5.4 mg/m3 for ammonia in confinement farming was recommended.3 29 The present study provides additional evidence that a threshold of 18 mg/m3 currently valid in many countries is too high for this multiexposure environment.
Analyses of associations with exposure were performed after exclusion of subjects with increases in responsiveness deemed unreliable. If these subjects were included in the analyses, associations were weaker, though in the same direction, for all exposure variables.
In conclusion, we find evidence for increases in responsiveness in pig farmers associated with exposure to dust and ammonia. This stresses the need for reduction of exposure, for which measures are proposed in an earlier publication.18 Few pig farmers leave their job due to health reasons alone. Earlier, we demonstrated it is likely that self-selection out of this industry due to asthma occurs in an early stage.30 Such a self-selection would attenuate observed exposure-response associations. This group with on average 16.7 years of pig farming, therefore, forms a relatively insensitive group. That clear effects of exposure are shown in such a group emphasizes the need for prevention.
|
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the Dutch prevention fund (grant No. 2821440) and the Netherlands Asthma Foundation (grant No. 94.42).
Received for publication September 29, 1999. Accepted for publication November 30, 1999.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. C. Mirabelli, S. Wing, S. W. Marshall, and T. C. Wilcosky Asthma Symptoms Among Adolescents Who Attend Public Schools That Are Located Near Confined Swine Feeding Operations Pediatrics, July 1, 2006; 118(1): e66 - e75. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S J Langley, S Goldthorpe, M Craven, A Woodcock, and A Custovic Relationship between exposure to domestic allergens and bronchial hyperresponsiveness in non-sensitised, atopic asthmatic subjects Thorax, January 1, 2005; 60(1): 17 - 21. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.A. Dosman, J.A. Lawson, S.P. Kirychuk, Y. Cormier, J. Biem, and N. Koehncke Occupational asthma in newly employed workers in intensive swine confinement facilities Eur. Respir. J., October 1, 2004; 24(4): 698 - 702. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W Karmaus and C Botezan Does a higher number of siblings protect against the development of allergy and asthma? A review J. Epidemiol. Community Health, March 1, 2002; 56(3): 209 - 217. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
