Miljø- og Fødevareudvalget 2019-20
MOF Alm.del
Offentligt
2103134_0001.png
Environment International 132 (2019) 105009
Contents lists available at
ScienceDirect
Environment International
journal homepage:
www.elsevier.com/locate/envint
Endotoxin and particulate matter emitted by livestock farms and respiratory
health effects in neighboring residents
Myrna M.T. de Rooij
a,
, Lidwien A.M. Smit
a
, Hans J. Erbrink
b
, Thomas J. Hagenaars
c
,
Gerard Hoek
a
, Nico W.M. Ogink
d
, Albert Winkel
d
, Dick J.J. Heederik
a
, Inge M. Wouters
a
⁎
a
T
Institute for Risk Assessment Sciences, Utrecht University, the Netherlands
Erbrink Advies, Arnhem, the Netherlands
c
Wageningen Bioveterinary Research, Wageningen University and Research, the Netherlands
d
Wageningen Livestock Research, Wageningen University and Research, the Netherlands
b
A R TICL E INFO
Handling Editor: Hanna Boogaard
Keywords:
Livestock farming
Emissions
Air pollution
Public health
Spatial modelling
Endotoxin
A BSTR A CT
Background:
Living in livestock-dense areas has been associated with health effects, suggesting airborne exposures to
livestock farm emissions to be relevant for public health. Livestock farm emissions involve complex mixtures of
various gases and particles. Endotoxin, a pro-inflammatory agent of microbial origin, is a constituent of livestock farm
emitted particulate matter (PM) that is potentially related to the observed health effects. Quantification of livestock
associated endotoxin exposure at residential addresses in relation to health outcomes has not been performed earlier.
Objectives:
We aimed to assess exposure-response relations for a range of respiratory endpoints and atopic
sensitization in relation to livestock farm associated PM
10
and endotoxin levels.
Methods:
Self-reported respiratory symptoms of 12,117 persons participating in a population-based cross-sec-
tional study were analyzed. For 2494 persons, data on lung function (spirometry) and serologically assessed
atopic sensitization was additionally available. Annual-average PM
10
and endotoxin concentrations at home
addresses were predicted by dispersion modelling and land-use regression (LUR) modelling. Exposure-response
relations were analyzed with generalized additive models.
Results:
Health outcomes were generally more strongly associated with exposure to livestock farm emitted en-
dotoxin compared to PM
10
. An inverse association was observed for dispersion modelled exposure with atopic
sensitization (endotoxin:
p
= .004, PM
10
:
p
= .07) and asthma (endotoxin:
p
= .029, PM
10
:
p
= .022).
Prevalence of respiratory symptoms decreased with increasing endotoxin concentration at the lower range, while
at the higher range prevalence increased with increasing concentration (p < .05). Associations between lung
function parameters with exposure to PM
10
and endotoxin were not statistically significant (p > .05).
Conclusions:
Exposure to livestock farm emitted particulate matter is associated with respiratory health effects
and atopic sensitization in non-farming residents. Results indicate endotoxin to be a potentially plausible etio-
logic agent, suggesting non-infectious aspects of microbial emissions from livestock farms to be important with
respect to public health.
1. Introduction
Epidemiological studies performed worldwide have shown asso-
ciations between living in livestock dense areas and health effects,
suggesting airborne exposures to livestock farms at residential level to
be relevant for public health (Borlée
et al., 2015, 2017a, 2017b, 2018;
Douglas et al., 2018; Elliott et al., 2004; Mirabelli et al., 2006; Pavilonis
et al., 2013; Radon et al., 2007; Rasmussen et al., 2017; Schinasi et al.,
2011; Schulze et al., 2011; Sigurdarson and Kline, 2006; Smit et al.,
2014).
Adverse effects on health identified included increased
respiratory symptoms (wheezing, cough) and decreased lung function
(Borlée
et al., 2017a; Radon et al., 2007; Schulze et al., 2011);
pro-
tective effects included lower prevalence of asthma and atopy (Borlée
et al., 2015; Elliott et al., 2004; Smit et al., 2014).
Most studies have
used exposure proxies such as distance to nearest farm and farm den-
sities in the surroundings to represent exposure to livestock-related air
pollution at residences. Underlying mechanisms and etiologic agents
are difficult to establish as livestock farm emissions consist of complex
mixtures of various gases and particles (Cambra-López
et al., 2010;
Hamon et al., 2012).
⁎
Corresponding author at: Yalelaan 2, 3584CM Utrecht, the Netherlands.
E-mail address:
[email protected]
(M.M.T. de Rooij).
https://doi.org/10.1016/j.envint.2019.105009
Received 4 April 2019; Received in revised form 9 July 2019; Accepted 10 July 2019
Available online 03 August 2019
0160-4120/ © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
Organic particulate matter (PM), also called biological particulate
matter or bio-aerosols, are aerosolized solid or liquid particles of bio-
logical origin like bacteria, fungi, plant materials etc. Organic PM is an
important air pollutant emitted by livestock farms. A constituent of
organic PM is endotoxin, a potent pro-inflammatory component of the
cell wall of gram-negative bacteria (Liu,
2002).
Since PM concentra-
tions in livestock farms are high and buildings are intensively ventilated
(Winkel
et al., 2015),
endotoxins are emitted into the atmosphere in
large quantities (Pillai
and Ricke, 2002; Seedorf et al., 1998; Thorne
et al., 2009).
Endotoxins are dispersed in the near surroundings and
beyond, resulting in exposure to livestock farm emitted endotoxin at
residential sites (De
Rooij et al., 2017; Hiranuma et al., 2011).
Endotoxin exposure has been linked to both adverse and protective
effects on respiratory health (Farokhi
et al., 2018; May et al., 2012).
Adverse effects, such as upper respiratory symptoms and chronic
bronchitis, are caused by endotoxin inducing production of cytokines
and proteins that cause airway inflammation (Liu,
2002; Poole and
Romberger, 2012).
Understanding of protective effects, such as de-
creased atopy, allergic rhinitis, and atopic asthma, is incomplete, but
endotoxin induced alterations in immune responses leading to sup-
pression of allergy-promoting responses seems to play a role (Liu,
2002;
Poole and Romberger, 2012).
Health effects of endotoxin have been
firmly established for exposure to high concentrations (predominantly
short term) in the various experimental/occupational studies performed
(Castellan
et al., 1987; May et al., 2012; Michel et al., 2002; Smid et al.,
1994).
However, less is known on health effects of environmental ex-
posure to considerably lower airborne endotoxin concentrations as
measured at residential sites (Farokhi
et al., 2018).
Quantification of exposure to livestock-emitted PM and endotoxin at
residential addresses is needed to increase insight in the mechanisms
behind livestock-related health effects observed in non-farming popu-
lations. State of the art modelling and monitoring methods aimed at
endotoxin are needed to specifically predict ambient concentrations
enabling exposure-response analyses. Land-use regression (LUR) mod-
elling and dispersion modelling are extensively used modelling tech-
niques in urban air pollution studies to predict exposures (De
Hoogh
et al., 2014).
Implementation of these techniques to model exposure to
livestock farm emitted endotoxin at residential addresses is novel. LUR
modelling uses geospatial predictor variables to explain spatial con-
trasts in measured airborne concentrations. (De
Rooij et al., 2018)
Dispersion modelling uses mathematical equations to compute airborne
distribution of pollutants resulting from emission from a source to the
surroundings conditional on aerial and meteorological circumstances
(De
Hoogh et al., 2014).
Methodological principles are distinctly dif-
ferent between both approaches, each faced with their own challenges
and validation issues. Both approaches are expected to be of use to
model exposure to livestock farm emitted endotoxin at residential ad-
dresses, but it cannot be foreseen if and how predictive ability would
differ. We hypothesize that livestock farmemitted PM, and especially its
endotoxin content, plays a role in observed health effects among re-
sidents living in livestock dense areas. We performed exposure-response
analyses for a range of respiratory endpoints and atopic sensitization in
relation to livestock farm emitted PM
10
(particulate matter with a
nominal aerodynamic diameter below 10 μm) and endotoxin among
12,117 participants of the VGO project (Dutch acronym for Livestock
Farming and Neighboring Residents' Health Study). Both dispersion
modelling and LUR modelling were applied to predict exposure at re-
sidential addresses.
2. Methods
2.1. Study design and population
Exposure-response analyses were performed in a population based
cross-sectional study which was part of the “Livestock Farming and
Neighboring Residents' Health” (VGO) project. The VGO study
2
population is a general, non-farming population sample enrolled in
2012 through a questionnaire survey among randomly selected patients
from 21 general practices (aged 18–70 years) living in a rural area in
the southeast of the Netherlands (the VGO study area; 3000km
2
in size)
(Borlée
et al., 2015).
Analyses were conducted on 12,117 responders
who lived at their home address for at least 1 year. Responders who
were willing to participate in a follow-up study (asked in ques-
tionnaire), and who were living within 10 km of one of twelve tem-
porary research centres were invited to the nearest centre for a medical
examination (n = 7180). From March 2014 to February 2015, 2494
subjects participated in the medical examination that included spiro-
metry and peripheral blood collection, see Borlée et al. for a flow chart
of the data collection(Borlée
et al., 2017b).
Detailed non-response
analyses indicated no signs that selection bias influenced the observed
associations (Borlée
et al., 2015, 2017a, 2017b, 2018).
The study pro-
tocol (13/533) was approved by the Medical Ethical Committee of the
University Medical Centre Utrecht. All 2494 subjects signed informed
consent. Initial analyses within the VGO project involved associating
exposure proxies (distance to farms, farm density) of livestock farm
emissions to health outcomes (Borlée
et al., 2015, 2017a, 2017b, 2018).
In the current study, analyses on health outcomes in relation to pre-
dicted residential exposures to livestock farm emitted PM
10
and en-
dotoxin are included.
2.2. Description of study area
The study area (3000 km
2
in size) contained regions of the
Netherlands with high livestock densities. The study area was situated
in the provinces of Noord-Brabant and Limburg, these provinces have a
combined surface area of 7290km
2
on which in total 17,250 farms are
present (provincial databases on farm licenses of 2015,
http://bvb.
brabant.nl
and
http://limburg.vaa.com/webbvb).
Farms are not geo-
graphically evenly spread; instead farms are mainly concentrated in the
region where the two provinces border. In the Netherlands, livestock is
commonly kept in enclosed animal houses, apart from some dairy cows,
sheep, and horses that also are kept on pastures during parts of the year
(Central
Bureau of Statistics, 2015).
Most farms are highly specialized
intensive operations meaning that only one animal species is kept
aimed at a specific production type (e.g., broiler farms, laying hen
farms). The number of animals kept on a farm vary highly; on average,
the number of animals kept in the study area on pig farms, chicken
farms, cattle farms were 2659; 76,718; 205, respectively (Borlée
et al.,
2017a).
2.3. Exposure assessment
2.3.1. Dispersion modelling
Dispersion modelling was applied to estimate livestock farm emitted
annual average PM
10
mass concentration and annual average endotoxin
concentrations in PM
10
fraction at residential addresses. The dispersion
model used is based on the Gaussian plume model and implements the
Netherlands New National Model (see Supplementary Methods tables
M1–3 for details). The model was applied to estimate dispersion of li-
vestock farm emitted PM and its endotoxin content. PM dispersion was
modelled using farm-type specific PM
10
emission factors, farm-type
specific distribution of PM size-fractions, barn characteristics and
emission rate (both mass and heat), meteorological conditions, and
terrain roughness. To obtain endotoxin dispersion, additional model
input included farm-type specific endotoxin content per PM size-frac-
tion (see Supplementary methods Tables M1–3).
To estimate endotoxin concentrations at residential addresses of
VGO study participants, we applied the dispersion model to individual
barns within 10 km of each address using provincial livestock farm data
including the exact location, farm-type and licensed animal numbers.
Per residential address' geographic coordinate (based on centroid of the
home), annual averages of PM
10
and endotoxin concentrations were
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
obtained by summation of contributions of individual barns.
Meteorological data used was matched to the period of the study, being
2012 for the questionnaire study and 2014–2015 for the medical ex-
amination study. Data on licensed animal numbers were available for
the years 2012 and 2015, as provided by the provinces of Noord-
Brabant and Limburg.
2.3.2. LUR modelling
Annual average endotoxin exposure at residential addresses were
estimated through application of a previously described endotoxin LUR
model developed for the VGO study area (De
Rooij et al., 2018).
In
brief, the LUR model was developed based on repeatedly measured
endotoxin concentrations in PM
10
fraction at 61 residential sites geo-
graphically distributed in the VGO study area. The measurement cam-
paign started in May 2014 and ended in December 2015, 2-week
average air samples were collected. The 61 sites were sampled three to
five times spread over the four seasons and in addition a background
location was included which was sampled consecutively (see
De Rooij
et al., 2018
for details). GIS predictors used were based on provincial
livestock data of 2015. A forward supervised stepwise selection pro-
cedure was performed to select GIS predictors. The resulting LUR model
contained the following variables: ‘number of sows weighted to dis-
tance in a 1000 m buffer’, ‘number of laying hens weighted to distance
in a 3000 m buffer’, ‘number of poultry animals in a 500 m buffer’, and
‘number of horse farms in a 3000 m buffer’. Together these variables
explained 64% of spatial variation in endotoxin concentrations. In the
current study, geocoded residential addresses of VGO study participants
were combined with geocoded livestock data to compute livestock
characteristics of the residential address' surroundings enabling appli-
cation of the endotoxin LUR model function. As described in
De Rooij
et al., 2018,
LUR modelling on measured PM10 concentrations using
livestock-related GIS predictors was evaluated. As the resulting LUR
model explained the spatial variation in PM10 concentrations only
modestly (19%), model performance was regarded too limited to be
applied for predicting exposure.
2.3.3. Validation and comparative analyses
Dispersion modelling was also applied to estimate endotoxin con-
centrations at the 61 VGO sites where endotoxin concentrations had
been measured. This enabled validation of dispersion modelled en-
dotoxin concentrations against measured endotoxin concentrations.
Furthermore, dispersion modelled endotoxin concentrations were
compared against LUR modelled endotoxin concentrations for both the
61 VGO measurement sites as well as residential addresses of the health
study participants.
2.4. Health outcomes
2.4.1. Questionnaire study
In 2012, a questionnaire survey was performed including questions
on respiratory health adapted from the European Community
Respiratory Health Survey (ECRHS)-III postal questionnaire (Jarvis
et al., 2018).
The following outcomes were included in the current
analysis: current asthma, doctor diagnosed COPD, wheeze (with or
without shortness of breath/having a cold), and daily cough (See
Supplementary Methods Table M.4. for the definition of the self-re-
ported respiratory conditions and symptoms).
2.4.2. Medical examination study
The medical examination included spirometry and peripheral blood
collection (Borlée
et al., 2017a, 2017b, 2018).
Pre- and post-broncho-
dilator spirometry was conducted as described before (Borlée
et al.,
2017a)
according to European Respiratory Society (ERS) guidelines.
The current study analyzes the pre-bronchodilator lung function para-
meters forced expiratory volume in 1 s (FEV
1
), Forced Vital Capacity
(FVC), FEV
1
/FVC and Maximum Mid-Expiratory Flow (MMEF). Lung
3
function parameters were expressed as percentage predicted based on
the GLI-2012 reference equations; thus adjusting for age, sex and height
(Quanjer
et al., 2012).
Atopy was defined as elevated levels of specific
serum IgE antibodies (> 0.35 U/ml) to one or more common allergens
(cat, dog, grass and house dust mite) and/or a total IgE higher than
100 IU/ml, assessed by ELISA as previously described (Borlée
et al.,
2018; Doekes et al., 1996).
2.5. Exposure-response analyses
Analyses were performed using R studio (version 3.0.2) (R.
Core-
Team, 2017).
Predicted concentrations at residential addresses were
truncated to 99.5 percentile for both LUR modelled endotoxin and
dispersion modelled endotoxin and PM
10
to avoid outlying un-
realistically high values (see
Table 1
for values of 99.5 percentile).
Associations between exposures and health outcomes were analyzed by
means of regression splines as previously performed analyses with ex-
posure proxies showed associations to be non-linear (Borlée
et al.,
2015, 2017a, 2018).
Penalized regression splines using the (default)
“thin plate” basis as implemented in the mgcv R-package were used. To
enable comparisons between predicted concentrations at residential
addresses and earlier reported exposure proxies, analyses on health
outcomes in relation to ‘number of farms within 1000 m buffer’ and
‘distance to nearest farm’ were included. Shape of the associations,
smooth term
p-values,
Akaike Information Criterion (AIC) values were
compared between the livestock exposure parameters.
Previously identified confounders were included. Associations with
respiratory symptoms were adjusted for age, sex and smoking habits
(never smoker, ex-smoker, current smoker) (Borlée
et al., 2015).
As-
sociations with lung function parameters were adjusted for living on a
farm during childhood, born in study area and smoking habits (Borlée
et al., 2017a).
The week average ambient ammonia concentration prior
to lung function testing was taken into account in addition (visualized
by means of dashed lines) as this was found to be inversely associated
with lung function (Borlée
et al., 2017a).
Associations with atopic
sensitization were adjusted for sex, age, smoking habits (ever smoking
and pack years), education (high versus middle/low education), born in
study area, and history of living on a farm during childhood (Borlée
et al., 2018).
3. Results
3.1. Study population characteristics
Of the total study population, participants had on average 8 farms in
a buffer of 1000 m around their house and on average the closest farm
was at 490 m distance (see
Table 1).
Modelled annual average con-
centrations of livestock farm emitted PM
10
at residential addresses
ranged from the 10th percentile to the 90th percentile from 0.07 to
0.40 μg/m
3
. The mean value of modelled annual average concentra-
tions of endotoxin at residential addresses was for LUR modelling,
dispersion modelling, 0.25, 0.18 endotoxin units (EU) per m
3
, respec-
tively (see Supplementary Table S.1. for other study population char-
acteristics). At relatively close distances, distinctly different endotoxin
concentrations were predicted, resulting in a substantial contrast in
predicted residential exposures across the study area (see Supplemen-
tary Fig. S.1. for a geographical overview).
3.2. Comparative analyses: LUR versus dispersion
Comparisons between dispersion modelled endotoxin concentra-
tions and measured endotoxin concentrations at the VGO measurement
sites (n = 61) showed reasonable agreement overall and absolute con-
centrations to match despite clear differences at certain sites (Pearson
correlation 0.51, see Supplementary Fig. S.2.). The level of agreement
between LUR modelled endotoxin concentration and dispersion
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
2103134_0004.png
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
90%ile
modelled endotoxin concentration at residential addresses was sub-
stantial (Pearson correlation for the full study population was 0.64, for
the subset with medical examination 0.68; see
Fig. 1).
Dispersion
modelled endotoxin concentrations showed a broader range, a lower
minimum and higher maximum, than LUR modelled concentrations.
3.3. Livestock exposure parameters: proxies, PM
10
and endotoxin
Pearson correlations between modelled livestock farm emitted PM
10
concentrations and modelled endotoxin concentrations were 0.84 for
dispersion modelled endotoxin, and 0.74 for LUR modelled endotoxin
in the total study population (see Supplementary Fig. S.3). Correlation
between modelled concentrations (PM
10
and endotoxin) with the ex-
posure proxy ‘number of farms within 1000 m buffer’ was moderate
(Pearson correlation: 0.48–0.58). The correlation between modelled
concentrations and the exposure proxy ‘distance to nearest farm’ was
weaker (Pearson correlation: 0.37–0.41).
3.4. Exposure-response analyses
For the majority of health outcomes, strength of associations of
exposure-response analyses followed a similar pattern: strongest asso-
ciations (lowest
p-values,
lowest AIC values) were found with endotoxin
concentrations, weakest associations with livestock exposure proxies,
and intermediate with livestock farm emitted PM
10
concentrations (see
Figs. 2–5).
3.4.1. Atopic sensitization
A significant inverse association between dispersion modelled en-
dotoxin concentration and atopic sensitization (based on serological
data, available for subset of total study population) was observed (see
Fig. 2).
Predicted prevalence decreased from 0.32 to 0.24 when com-
paring the 5th percentile of dispersion modelled endotoxin concentra-
tion to the 95th percentile, thus a prevalence ratio of 0.75. The pro-
tective association was also observed with the proxy ‘distance to nearest
farm’ however the association was weaker. Associations with LUR
modelled endotoxin concentrations and dispersion modelled PM
10
concentrations showed a similar trend (non-significant) of decreased
prevalence with increased exposure.
3.4.2. Respiratory symptoms
Modelled endotoxin concentrations (both LUR modelled and disper-
sion modelled) were significantly associated with prevalence of wheeze
with shortness of breath and daily cough (based on questionnaire data
collected from total study population), see
Fig. 3.
At the lower range of
endotoxin concentrations, predicted prevalence of wheeze with shortness
of breath decreased with increasing endotoxin concentration from 0.11
for the lowest exposed (5th percentile) to 0.08 for dispersion modelled
endotoxin exposure of 0.32 EU/m
3
(prevalence ratio of 0.80), for daily
cough the predicted prevalence decreased from 0.14 for the lowest ex-
posed (5th percentile) to 0.12 for dispersion modelled endotoxin ex-
posure of 0.32 EU/m
3
(prevalence ratio of 0.82). At the higher range of
endotoxin concentrations, predicted prevalence of wheeze with shortness
of breath/daily cough increased with increasing endotoxin concentration
but confidence intervals were wide due to a lower number of observa-
tions. The pattern with livestock farm emitted PM
10
concentration was
comparable, but non-significant. The shape of the association between
‘number of farms within 1000 m buffer’ and prevalence of wheeze with
shortness of breath was distinctly different, no change in prevalence was
observed up to around 24 farms followed by a sharp increase in pre-
valence with increasing number of farms.
3.4.3. Asthma and COPD
Fig. 4
shows the associations between increased livestock exposure
with decreased prevalence of both asthma and COPD (based on ques-
tionnaire data collected from total study population). The shape of the
4
99.5%ile
50%ile
10%ile
Mean
SD
Table 1
Residential characteristics of study population.
Subset of study population (n = 2494, medical examination participants)
Total study population (n = 12,117)
Distance to nearest farm in meters
Number of farms within 1000 m buffer
Predicted annual average endotoxin concentration (EU/m
3
) by LUR model
Predicted annual average endotoxin concentration (EU/m
3
) by dispersion model
Predicted annual average PM
10
concentration (μg/m
3
) by dispersion model
Distance to nearest farm in meters
Number of farms within 1000 m buffer
Predicted annual average endotoxin concentration (EU/m
3
) by LUR model
Predicted annual average endotoxin concentration (EU/m
3
) by dispersion model
Predicted annual average PM
10
concentration (μg/m
3
) by dispersion model
488
8
0.25
0.18
0.20
439
9
0.27
0.25
0.31
275
6
0.09
0.13
0.15
266
6
0.10
0.16
0.18
170
1
0.16
0.07
0.07
135
2
0.17
0.09
0.12
440
7
0.24
0.15
0.18
401
9
0.25
0.22
0.29
870
15
0.35
0.33
0.40
794
17
0.38
0.42
0.53
1374
26
0.70
0.98
0.94
1360
27
0.72
1.26
1.22
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
2103134_0005.png
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
Fig. 1.
Comparison between dispersion modelled versus LUR modelled annual average endotoxin concentrations (in EU/m
3
); left panel for years 2014–2015, in line
with medical examination period; right panel for year 2012, in line with period when questionnaires were collected.
Note. Modelled concentrations for years 2014-2015: Pearson correlation is 0.68, Spearman correlation is 0.65.
Modelled concentrations for year 2012: Pearson correlation is 0.64, Spearman correlation is 0.66.
Solid grey line is linear regression fit.
Dashed grey line is line of unity.
Fig. 2.
Splines of associations between different livestock exposure parameters with atopic sensitization.
Model
LUR modelled [endotoxin]
(EU/m
3
)
2813.11
0.278
Dispersion modelled [endotoxin]
(EU/m
3
)
2806.69
0.00391*
Dispersion modelled [PM
10
]
(μg/m
3
)
2812.25
0.0695
N farms within 1000 m
buffer
2814.17
0.414
Distance to nearest
farm (m)
2806.89
0.0245*
AIC
P-value smooth term
Note. Associations were adjusted for sex, age, smoking habits (ever smoking and pack years), education (high versus middle/low education), born in study area, and
history of living on a farm during childhood. Rug plot shown on lower x-axis, percentiles shown on upper x-axis. Predicted concentrations at residential addresses
were truncated to 99.5 percentile for both LUR modelled endotoxin and dispersion modelled endotoxin and PM
10
. Panel ‘number of farms within 1000 m buffer’ and
‘distance to nearest farm’ reprinted in adapted format with permission of the BMJ Publishing group Ltd. Copyright © 2019.Borlée F., Yzermans C.J., Krop E.J.M.,
Maassen C.B.M., Schellevis F.G., Heederik D.J.J., Smit L.A.M.; Residential proximity to livestock farms is associated with a lower prevalence of atopy,
Occup Env. Med
75,
2018, 453–460,
https://doi.org/10.1136/oemed-2017-104769.
association with COPD prevalence was comparable between all live-
stock exposure parameters (‘distance to nearest farm’ lowest
P-value:
0.015). For asthma prevalence, the protective effect with modelled
concentrations (PM
10
and endotoxin) was observed in the lower con-
centration range; predicted prevalence decreased from 0.13 for the
lowest exposed (5th percentile) to 0.10 (prevalence ratio of 0.80).
3.4.4. Lung function
Analyses with FEV
1
and FEV
1
/FVC (based on spirometry data,
available for subset of total study population) showed no significant
associations with livestock exposure parameters (see Supplementary
Fig. S.4). Dispersion modelled endotoxin concentration was borderline
5
significantly (p = .06) associated with FVC (see
Fig. 5).
This downward
trend was not observed with the other parameters of livestock exposure.
The significant association between MMEF and ‘number of farms within
1000 m buffer’, was not found for the other livestock exposure para-
meters (see
Fig. 5).
4. Discussion
This study is the first to use both dispersion modelling and LUR
modelling to study health effects in relation to livestock farm emissions.
Results indicate livestock farm emitted PM
10
, and especially its en-
dotoxin content, play a role in the earlier identified associations
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
2103134_0006.png
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
Fig. 3.
Splines of associations between different livestock exposure parameters with respiratory symptoms; upper row – wheeze and shortness of breath, lower row -
daily cough.
Health outcome
Model
LUR modelled [endotoxin] Dispersion modelled [endo-
toxin] (EU/m
3
)
(EU/m
3
)
7483.02
7482.71
0.0361*
Dispersion modelled
[PM
10
] (μg/m
3
)
7485.25
0.0855
N farms within
1000m buffer
7478.73
0.007*
Distance to nearest
farm (m)
7488.56
0.142
Wheeze + SOB
AIC
P-value smooth
0.0379*
term
Daily cough
AIC
9391.19
9395.54
0.0089*
9400.58
0.0603
9403.11
0.062
9404.50
0.142
P-value smooth
0.0026*
term
Note. Associations were adjusted for age, sex and smoking habits (never smoker, ex-smoker, current smoker). Smoothed plot for probability of daily cough in relation
to LUR modelled endotoxin concentrations based on the “thin plate” basis showed an overfitted pattern (see dashed line), the non-overfitted pattern depicted (solid
line) was based on number of knots set at 5. Rug plot shown on lower x-axis, percentiles shown on upper x-axis. Predicted concentrations at residential addresses
were truncated to 99.5 percentile for both LUR modelled endotoxin and dispersion modelled endotoxin and PM
10
.
between livestock farming exposure proxies and respiratory health of
persons living in livestock dense areas. Associations with modelled
endotoxin concentrations were stronger than with modelled PM
10
concentrations for atopic sensitization, asthma and respiratory symp-
toms; supporting our hypothesis of the relevance of endotoxin with
respect to observed health effects. A clear decrease in prevalence of
atopic sensitization was found in relation to increased endotoxin con-
centration at the residential address, indicative of a protective effect.
Also for prevalence of asthma, indications for a protective effect of
endotoxin exposure were observed. Analyses of respiratory symptoms
suggested a protective effect at the lower range of endotoxin con-
centrations and an possibly adverse effect at the higher range of con-
centrations given the power limitations at the higher range of exposure
(minority of participants highly exposed). Spirometry results showed
some suggestions for a small decrease in lung function in relation to
livestock exposure, but associations were weak with all livestock ex-
posure parameters.
4.1. Concentrations: PM
10
and endotoxin
Both dispersion modelling and LUR modelling were successful in
quantifying endotoxin concentrations in ambient air at residential ad-
dresses. Model performance (expressed as measured spatial variation
explained) was comparable between both approaches (dispersion R
2
:
0.26, LUR hold-out validation R
2
: 0.32). A good level of agreement was
6
observed between dispersion modelled concentrations and measured
concentrations. The overall good level of agreement between dispersion
modelled concentrations and LUR modelled concentrations at the ma-
jority of residential addresses indicates robustness of exposure predic-
tions between two distinctly different modelling approaches. At certain
individual addresses, however, modelled concentrations were distinctly
different. In a LUR model, only a restricted number of variables are
selected that describe the general pattern in spatial variation of re-
sidential concentrations well. The endotoxin LUR model contains four
variables (De
Rooij et al., 2018),
farm types not included are re-
presented indirectly by means of overall spatial correlation in the study
area. Dispersion modelling uses emission from a single farm as the
starting point of modelling of exposure and thus includes all farm types
in a direct manner, however emission characterization is generalized
per farm type.
Dispersion modelled PM
10
concentrations and dispersion modelled
endotoxin concentrations showed some divergence. This can be ex-
plained by variability in endotoxin content per size fraction depending
on the farm type. For example, at addresses with many poultry farms in
the vicinity, endotoxin concentrations are relatively low compared to
PM
10
concentrations, whereas the opposite holds at addresses with
many pig farms in the vicinity.
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
2103134_0007.png
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
Fig. 4.
Splines of associations between different livestock exposure parameters with asthma (upper row) and COPD (lower row).
Health outcome
Model
LUR modelled [endotoxin] Dispersion modelled [endo-
(EU/m
3
)
toxin] (EU/m
3
)
8442.66
0.0348*
8442.67
0.0289*
Dispersion modelled
[PM
10
] (μg/m
3
)
8441.51
0.0222*
N farms within
1000m buffer
8448.52
0.14
Distance to nearest
farm (m)
8446.65
0.0428*
Current asthma
AIC
P-value smooth
term
COPD
AIC
P-value smooth
term
4274.19
0.307
4271.22
0.0523
4271.52
0.0607
4274.56
0.427
4269.41
0.0147*
Note. Associations were adjusted for age, sex and smoking habits (never smoker, ex-smoker, current smoker). Rug plot shown on lower x-axis, percentiles shown on upper
x-axis. Predicted concentrations at residential addresses were truncated to 99.5 percentile for both LUR modelled endotoxin and dispersion modelled endotoxin and PM
10
.
4.2. Exposure-response relations
For atopic sensitization, asthma and respiratory symptoms, weakest
exposure-response associations were found with livestock exposure
proxies, strongest with endotoxin and intermediate with PM
10
.
Considering attenuation of associations in epidemiological studies due
to misclassification, this pattern supports the hypothesis of the etiolo-
gical relevance of livestock farm emitted PM
10
and in particular its
endotoxin content. Earlier work showed livestock exposure proxies to
only moderately explain spatial variation in endotoxin concentrations
measured at residential sites (r-squared: 15.5%, 12.5% for ‘number of
farms within 1000 m buffer’, ‘distance to nearest farm’; respectively)
(De
Rooij et al., 2018).
Modelled concentrations are naturally an ap-
proximation of true levels, hence some misclassification will always be
present. Exposure-response relations for endotoxin concentration
modelled independently by both approaches corresponded well, this
substantiates identified associations.
The earlier identified significant negative association between the
lung function parameter MMEF and ‘number of farms within 1000 m
buffer’ (Borlée
et al., 2017a)
was not observed with modelled con-
centrations. On the contrary, FVC appeared (borderline) significantly
negatively associated with dispersion modelled endotoxin concentra-
tion while livestock exposure proxies did not suggest a trend for FVC.
4.3. Endotoxin and health
Significant associations between endotoxin exposure and health
outcomes were identified, even though predicted annual average levels
7
of endotoxin exposure (on average 0.25 EU/m
3
) were markedly below
proposed health based exposure limits (Dutch Health Council: occu-
pational limit of 90 EU/m
3
, tentative limit of 30 EU/m
3
for the general
population) (Health
Council of the Netherlands, 2010, 2012).
However,
these cannot be directly compared as exposure limits specifically aim at
short-term exposures (6-hour exposure) in inhalable dust. Occurrence
of short-term peak concentrations at residential sites are likely, espe-
cially at sites with elevated long-term average concentrations, since
farm emissions are not constant over time (Winkel
et al., 2015)
and
dispersion depends on ever-changing atmospheric conditions.
Our results showed an inverse association between endotoxin ex-
posure and atopic sensitization and asthma prevalence. For children, an
inverse association between livestock exposure and allergic diseases has
been firmly established, this protective association is believed to be
mediated by exposure to microbes (Campbell
et al., 2015).
Recent
studies also identified this inverse association in adult populations
(adjusted for childhood farm exposure) suggesting continuous exposure
to be the most effective to prevent development of allergic diseases
(Elholm
et al., 2013, 2018; Smit et al., 2010; Spierenburg et al., 2017).
For both atopic sensitization as asthma, the prevalence ratio observed
in our study was substantial (0.75, 0.80; respectively) suggesting the
size of the effect of livestock-related environmental exposure in a non-
farming population to be considerable. The majority of studies looking
into respiratory symptoms in relation to low levels of airborne en-
dotoxin exposure (defined as < 100 EU/m
3
) found an increase in
symptoms with increased exposure but most did not reach statistical
significance (Farokhi
et al., 2018).
Most of these studies focused on
short-term exposure, research regarding long-term exposure and
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
2103134_0008.png
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
Fig. 5.
Splines of associations between different livestock exposure parameters with lung function parameters; upper row – MMEF, lower row - FVC.
Health outcome
Model
LUR modelled [endo-
toxin] (EU/m
3
)
22009.60
0.506
Dispersion modelled [endo-
toxin] (EU/m
3
)
22010.77
0.955
Dispersion modelled
[PM
10
] (μg/m
3
)
22008.92
0.337
N farms within
1000m buffer
22003.61
0.0456*
Distance to nearest
farm (m)
22008.46
0.129
MMEF % predicted value
AIC
P-value
smooth term
FVC % predicted value
AIC
P-value
smooth term
17914.32
0.216
17913.96
0.0613
17917.18
0.588
17915.49
0.317
17917.27
0.655
Note. Adjustment for sex, age and height was made by calculating percentage predicted spirometry variables based on GLI-reference values. Associations are also
adjusted for farm childhood, smoking habits and born in study area. The dotted lines show the models for further adjustment for week-average ambient NH3 levels
(ug/m3) prior to the lung function test. Rug plot shown on lower x-axis, percentiles shown on upper x-axis. Predicted concentrations at residential addresses were
truncated to 99.5 percentile for both LUR modelled endotoxin and dispersion modelled endotoxin and PM
10
. Panel ‘number of farms within 1000m buffer’ reprinted
in adapted format with permission of the American Thoracic Society. Copyright © 2019 American Thoracic Society. Cite: Borlée F., Yzermans C.J., Aalders B.,
Rooijackers J., Krop E.J.M., Maassen C.B.M., Schellevis F.G., Brunekreef B., Heederik D.J.J., and Smit L.A.M.; Air pollution from livestock farms is associated with
airway obstruction in neighboring residents,
Am. J. Respir. Crit. Care Med.
196,
2017a, 1152–1161,
https://doi.org/10.1164/rccm.201701-0021OC.
The
American
Journal of Respiratory and Critical Care Medicine
is an official journal of the American Thoracic Society.
potential health effects is lacking. Splines displaying the relation be-
tween endotoxin exposure and respiratory symptoms in our study po-
pulation had a remarkable shape, displaying inverse associations at the
lower range and adverse associations at the higher range. It might be
that this pattern represents protective and adverse effects of endotoxin
and that exposure characteristics determine the effect. This would be in
line with suggestions made that (besides individual susceptibility) dose,
duration and frequency of exposure largely determines the response
(Gautam
et al., 2018; Liu, 2002; Wunschel and Poole, 2016).
Besides a
plausible causative agent, endotoxin is also a marker for livestock-re-
lated microbial exposures in general. Other microbial components like
glucans and peptidoglycans could also play a role in observed health
effects among residents (Poole
and Romberger, 2012).
An inverse association was observed in our study for self-reported
COPD prevalence with livestock exposure parameters. This inverse as-
sociation has been observed before in this rural area in the Netherlands
by a study assessing associations with livestock exposure proxies in
relation to health using electronic medical records of over 90,000 pa-
tients (Smit
et al., 2014).
Our study findings showed the association
with COPD prevalence to be strongest with ‘distance to nearest farm’,
dispersion modelled concentrations were borderline significant. It
might be that associations are influenced by self-reported COPD
8
overlapping with asthma (Borlée
et al., 2015).
Biologically, it does not
seem plausible that endotoxin exposure protects against COPD. Taking
in mind misclassification induced attenuation, this also argues against
livestock emitted PM underlying the protective association between
‘distance to nearest farm’ and COPD.
A clear decline in lung function after short term exposure to organic
dust, specifically endotoxin, has been firmly established in experi-
mental and occupational studies (Castellan
et al., 1987; Michel et al.,
2002; Möller et al., 2012; Smid et al., 1994).
Evidence with respect to
long-term exposure and accelerated lung function decline is limited
(Bolund
et al., 2017).
In our study population, we did not observe a
clear decrease in lung function in relation to long-term endotoxin ex-
posure. Findings reported by
Borlée et al. (2017a)
indicated short-term
elevated levels of exposure to livestock farm emissions, assessed by
variations in ammonia concentrations over time, being potentially re-
levant with respect to lung function in this study population.
4.4. Strengths and limitations
Exposures were estimated based on two distinct state-of-the-art
modelling approaches, both being validated against measured con-
centrations. Comparability of exposure-response relations for endotoxin
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
concentration modelled independently by both approaches, sub-
stantiates identified associations. Health outcomes used were obtained
from a large population based study including validated questionnaires,
high quality lung function testing and objectively assessed atopic sen-
sitization.
The main limitation of this study is its cross-sectional design ham-
pering interpretation of results with respect to causality and timing of
effects. No insight was gained in short-term variation of exposures to
livestock farm emissions and health effects, which remains a major
knowledge gap. Longitudinal health studies in combination with time-
resolved concentrations at residential addresses are warranted. Other
limitations of this study include the lack of information on home-re-
lated exposures and the use of self-reported data to determine current
asthma/COPD status which might not reflect the actual status in all
cases.
4.5. Implications
This study supports the hypothesis that emissions from livestock
farms are of public health relevance. Results indicated microbial
emissions from farms to have sizable effects on the respiratory system of
people living in livestock dense areas. Besides relevance of microbial
emissions with respect to transmission of infectious disease, this also
draws attention to non-infectious health effects (Leibler
et al., 2017).
Microbial air pollution is, compared to other types of air pollution,
understudied (Nazaroff,
2019).
Additional research is required to sub-
stantiate findings of this study, most importantly shape of the associa-
tions and effect sizes given the confidence intervals observed at the
higher range of exposure. As agricultural practices and rural popula-
tions vary, performing similar research in various countries/regions
enabling comparisons would be valuable. Besides studying effects in the
general population, there is a need to focus on specific groups poten-
tially more vulnerable to livestock-related air pollution (e.g. respiratory
patients, elderly and children).
4.6. Conclusion
Exposure to livestock farm emitted particulate matter seems to af-
fect respiratory health and atopic sensitization of non-farming re-
sidents. Results indicate endotoxin as the particulate matter constituent
potentially underlying observed health effects. Exposure-response
analyses suggested both protective and adverse health effects to be
related to exposure to livestock farm emitted endotoxin, indications of a
protective effect were particularly strong for atopic sensitization. In
view of residential health effects of microbial exposures related to li-
vestock farming, focus thus far lies on zoonotic infections. Our findings
draw attention to non-infectious aspects of microbial emissions with
respect to public health. It is essential to gain further insights in this,
especially since many people worldwide live in rural areas of which
some are potentially more vulnerable to livestock farm emissions.
Funding information
The study was funded by the Ministry of Infrastructure and Water
Management; the Province of Noord-Brabant; the Ministry of Health,
Welfare and Sports and the Ministry of Economic Affairs of the
Netherlands; and supported by a grant from the Lung Foundation
Netherlands (3.2.11.022).
Declaration of Competing Interest
No conflicts of interest to declare.
Acknowledgements
Colleagues of the Institute for Risk Assessment Sciences (IRAS) are
9
acknowledged for their contribution to the VGO study, especially Floor
Borlée and Azadèh Farokhi.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://
doi.org/10.1016/j.envint.2019.105009.
References
Bolund, A.C., Miller, M.R., Sigsgaard, T., Schlünssen, V., 2017. The effect of organic dust
exposure on long-term change in lung function: a systematic review and meta-ana-
lysis. Occup Env. Med 74, 531–542.
https://doi.org/10.1136/oemed-2016-103963.
Borlée, F., Yzermans, C.J., Van Dijk, C.E., Heederik, D., Smit, L.A.M., 2015. Increased
respiratory symptoms in COPD patients living in the vicinity of livestock farms. Eur.
Respir. J. 46, 1605–1614.
https://doi.org/10.1183/13993003.00265-2015.
Borlée, F., Joris Yzermans, C., Aalders, B., Rooijackers, J., Krop, E., Maassen, C.B.M.,
Schellevis, F., Brunekreef, B., Heederik, D., Smit, L.A.M., 2017a. Air pollution from
livestock farms is associated with airway obstruction in neighboring residents. Am. J.
Respir. Crit. Care Med. 196, 1152–1161.
https://doi.org/10.1164/rccm.201701-
0021OC.
Borlée, F., Yzermans, C.J., Krop, E., Aalders, B., Rooijackers, J., Zock, J.P., Van Dijk, C.E.,
Maassen, C.B.M., Schellevis, F., Heederik, D., Smit, L.A.M., 2017b. Spirometry,
questionnaire and electronic medical record based COPD in a population survey:
comparing prevalence, level of agreement and associations with potential risk factors.
PLoS One 12, e0171494.
https://doi.org/10.1371/journal.pone.0171494.
Borlée, F., Yzermans, J., Krop, E.J., Maassen, C.B., Schellevis, F., J Heederik, D.J., Smit,
L.A.M., 2018. Residential proximity to livestock farms is associated with a lower
prevalence of atopy. Occup Env. Med 75, 453–460.
https://doi.org/10.1136/oemed-
2017-104769.
Cambra-López, M., Aarnink, A.J.A., Zhao, Y., Calvet, S., Torres, A.G., 2010. Airborne
particulate matter from livestock production systems: a review of an air pollution
problem. Environ. Pollut. 158, 1–17.
https://doi.org/10.1016/j.envpol.2009.07.011.
Campbell, B.E., Lodge, C.J., Lowe, A.J., Burgess, J.A., Matheson, M.C., Dharmage, S.C.,
2015. Exposure to “farming” and objective markers of atopy: a systematic review and
meta-analysis. Clin. Exp. Allergy 45, 744–757.
https://doi.org/10.1111/cea.12429.
Castellan, R.M., Olenchock, S.A., Kinsley, K.B., Hankinson, J.L., 1987. Inhaled endotoxin
and decreased spirometric values. N. Engl. J. Med. 317, 605–610.
https://doi.org/10.
1056/NEJM198709033171005.
Central Bureau of Statistics, 2015. Agriculture; crops, livestock, land use per region.
Published In: Statline, CBS, .
http://statline.cbs.nl/StatWeb,
Accessed date: 8 April
2019.
Core-Team, R., 2017. R: A Language and Environment for Statistical Computing.
De Hoogh, K., Korek, M., Vienneau, D., Keuken, M., Kukkonen, J., Nieuwenhuijsen, M.J.,
Badaloni, C., Beelen, R., Bolignano, A., Cesaroni, G., Cirach Pradas, M., Cyrys, J.,
Douros, J., Eeftens, M., Forastiere, F., Forsberg, B., Fuks, K., Gehring, U., Gryparis, A.,
Gulliver, J., Hansell, A.L., Hoffmann, B., Johansson, C., Jonkers, S., Kangas, L.,
Katsouyanni, K., Künzli, N., Lanki, T., Memmesheimer, M., Moussiopoulos, N.,
Modig, L., Pershagen, G., Probst-Hensch, N., Schindler, C., Schikowski, T., Sugiri, D.,
Teixidó, O., Tsai, M.-Y., Yli-Tuomi, T., Brunekreef, B., Hoek, G., Bellander, T., 2014.
Comparing land use regression and dispersion modelling to assess residential ex-
posure to ambient air pollution for epidemiological studies. Environ. Int. 73,
382–392.
https://doi.org/10.1016/j.envint.2014.08.011.
De Rooij, M.M.T., Heederik, D.J.J., Borlée, F., Hoek, G., Wouters, I.M., 2017. Spatial and
temporal variation in endotoxin and PM10 concentrations in ambient air in a live-
stock dense area. Environ. Res. 153, 161–170.
https://doi.org/10.1016/j.envres.
2016.12.004.
De Rooij, M.M.T., Heederik, D.J.J., Van Nunen, E.J.H.M., Van Schothorst, I.J., Maassen,
C.B.M., Hoek, G., Wouters, I.M., 2018. Spatial variation of endotoxin concentrations
measured in ambient PM
10
in a livestock-dense area: implementation of a land-use
regression approach. Environ. Health Perspect. 126, 017003.
Doekes, G., Douwes, J., Wouters, I., de Wind, S., Houba, R., Hollander, A., 1996. Enzyme
immunoassays for total and allergen specific IgE in population studies. Occup.
Environ. Med. 53, 63–70.
Douglas, P., Robertson, S., Gay, R., Hansell, A.L., Gant, T.W., 2018. A systematic review of
the public health risks of bioaerosols from intensive farming. Int. J. Hyg. Environ.
Health 221, 134–173.
https://doi.org/10.1016/J.IJHEH.2017.10.019.
Elholm, G., Schlünssen, V., Doekes, G., Basinas, I., Martin Bibby, B., Hjort, C., Milvang
Grønager, P., Omland, Ø., Sigsgaard, T., 2013. Become a Farmer and Avoid New
Allergic Sensitization: Adult Farming Exposures Protect Against New-onset Atopic
Sensitization.
https://doi.org/10.1016/j.jaci.2013.06.044.
Elholm, G., Schlünssen, V., Doekes, G., Basinas, I., Bolund, A.C.S., Hjort, C., Grønager,
P.M., Omland, Ø., Sigsgaard, T., 2018. High exposure to endotoxin in farming is
associated with less new-onset pollen sensitisation. Occup. Environ. Med.
https://doi.
org/10.1136/oemed-2017-104384.
Elliott, L., Yeatts, K., Loomis, D., 2004. Ecological associations between asthma pre-
valence and potential exposure to farming. Eur. Respir. J. 24, 938–941.
https://doi.
org/10.1183/09031936.04.00006404.
Farokhi, A., Heederik, D., Smit, L.A.M., 2018. Respiratory health effects of exposure to
low levels of airborne endotoxin – a systematic review. Environ. Health 17, 14.
https://doi.org/10.1186/s12940-018-0360-7.
Gautam, R., Heo, Y., Lim, G., Song, E., Roque, K., Lee, J., Kim, Y., Cho, A., Shin, S., Kim,
C., Bang, G., Bahng, J., Kim, H., 2018. Altered immune responses in broiler chicken
 MOF, Alm.del - 2019-20 - Endeligt svar på spørgsmål 30: Spm. om ministeren vil bede de hollandske myndigheder om at få indsigt i den forskning, som viser, at der kan være helbredsmæssige konsekvenser forbundet med at bo tæt på en stor svineproduktion, som det fremgår af artiklen Nabo til svinefarm har gået til behandling i flere år: - De kan ikke finde ud af, hvad jeg fejler, til miljøministeren
M.M.T. de Rooij, et al.
Environment International 132 (2019) 105009
2012 equations. Eur. Respir. J. 40, 1324–1343.
https://doi.org/10.1183/09031936.
00080312.
Radon, K., Schulze, A., Ehrenstein, V., Van Strien, R.T., Praml, G., Nowak, D., 2007.
Environmental exposure to confined animal feeding operations and respiratory
health of neighboring residents. Epidemiology 18, 300–308.
https://doi.org/10.
1097/01.ede.0000259966.62137.84.
Rasmussen, S., Casey, J., Bandeen-Roche, K., Schwartz, B., Rasmussen, S.G., Casey, J.A.,
Bandeen-Roche, K., Schwartz, B.S., 2017. Proximity to industrial food animal pro-
duction and asthma exacerbations in Pennsylvania, 2005–2012. Int. J. Environ. Res.
Public Health 14, 362.
https://doi.org/10.3390/ijerph14040362.
Schinasi, L., Horton, R.A., Guidry, V.T., Wing, S., Marshall, S.W., Morland, K.B., 2011. Air
pollution, lung function, and physical symptoms in communities near concentrated
swine feeding operations. Epidemiology 22, 208–215.
https://doi.org/10.1097/EDE.
0b013e3182093c8b.
Schulze, A., Rommelt, H., Ehrenstein, V., van Strien, R., Praml, G., Kuchenhoff, H.,
Nowak, D., Radon, K., 2011. Effects on pulmonary health of neighboring residents of
concentrated animal feeding operations: exposure assessed using optimized estima-
tion technique. Arch. Environ. Occup. Health 66, 146–154. doi.
https://doi.org/10.
1080/19338244.2010.539635.
Seedorf, J., Hartung, J., Schröder, M., Linkert, K.H., Phillips, V.R., Holden, M.R., Sneath,
R.W., Short, J.L., White, R.P., Pedersen, S., Takai, H., Johnsen, J.O., Metz, J.H.M.,
Groot Koerkamp, P.W.G., Uenk, G.H., Wathes, C.M., 1998. Concentrations and
emissions of airborne endotoxins and microorganisms in livestock buildings in
Northern Europe. J. Agric. Eng. Res. 70, 97–109.
https://doi.org/10.1006/jaer.1997.
0281.
Sigurdarson, S.T., Kline, J.N., 2006. School proximity to concentrated animal feeding
operations and prevalence of asthma in students. Chest 129, 1486–1491.
https://doi.
org/10.1378/chest.129.6.1486.
Smid, T., Heederik, D., Houba, R., Quanjer, P.H., 1994. Dust- and endotoxin-related acute
lung function changes and work-related symptoms in workers in the animal feed
industry. Am. J. Ind. Med. 25, 877–888.
https://doi.org/10.1002/ajim.4700250611.
Smit, L.A.M., Heederik, D., Doekes, G., Lammers, J.W.J., Wouters, I.M., 2010.
Occupational endotoxin exposure reduces the risk of atopic sensitization but in-
creases the risk of bronchial hyperresponsiveness. Int. Arch. Allergy Immunol. 152,
151–158.
https://doi.org/10.1159/000265536.
Smit, L.A.M., Hooiveld, M., van der Sman-de Beer, F., Opstal-van Winden, A.W.J.,
Beekhuizen, J., Wouters, I.M., Yzermans, C.J., Heederik, D., 2014. Air pollution from
livestock farms, and asthma, allergic rhinitis and COPD among neighbouring re-
sidents. Occup. Environ. Med. 71, 134–140.
https://doi.org/10.1136/oemed-2013-
101485.
Spierenburg, E.A.J., Smit, L.A.M., Krop, E.J.M., Heederik, D., Hylkema, M.N., Wouters,
I.M., 2017. Occupational endotoxin exposure in association with atopic sensitization
and respiratory health in adults: results of a 5-year follow-up. PLoS One 12,
e0189097.
https://doi.org/10.1371/journal.pone.0189097.
Thorne, P.S., Ansley, A.C., Perry, S.S., 2009. Concentrations of bioaerosols, odors, and
hydrogen sulfide inside and downwind from two types of swine livestock operations.
J. Occup. Environ. Hyg. 6, 211–220.
https://doi.org/10.1080/15459620902729184.
Winkel, A., Mosquera, J., Groot Koerkamp, P.W., Ogink, N.W., Aarnink, A.J., 2015.
Emissions of particulate matter from animal houses in the Netherlands. Atmos.
Environ. 111, 202–212.
https://doi.org/10.1016/j.atmosenv.2015.03.047.
Wunschel, J., Poole, J.A., 2016. Occupational agriculture organic dust exposure and its
relationship to asthma and airway inflammation in adults. J. Asthma 53, 471–477.
https://doi.org/10.3109/02770903.2015.1116089.
husbandry workers and their association with endotoxin exposure. Ind. Health 56,
10–19.
https://doi.org/10.2486/indhealth.2017-0049.
Hamon, L., Andrès, Y., Dumont, E., 2012. Aerial pollutants in swine buildings: a review of
their characterization and methods to reduce them. Environ. Sci. Technol. 46,
12287–12301.
https://doi.org/10.1021/es3025758.
Health Council of the Netherlands, 2010. Endotoxins: Health-based Recommended
Occupational Exposure Limit. [WWW Document]. URL.
https://www.
gezondheidsraad.nl/nl/node/1958.
Health Council of the Netherlands, 2012. Health risks related to proximity of livestock
farming. [WWW Document]. URL.
https://www.gezondheidsraad.nl/sites/default/
files/201227Gezondheidsrisicoveehouderijen.pdf.
Hiranuma, N., Brooks, S.D., Gramann, J., Auvermann, B.W., 2011. High concentrations of
coarse particles emitted from a cattle feeding operation. Atmos. Chem. Phys. 11,
8809–8823.
https://doi.org/10.5194/acp-11-8809-2011.
Jarvis, D., Newson, R., Janson, C., Corsico, A., Heinrich, J., Anto, J.M., Abramson, M.J.,
Kirsten, A.-M., Zock, J.P., Bono, R., Demoly, P., Leynaert, B., Raherison, C., Pin, I.,
Gislason, T., Jogi, R., Schlunssen, V., Svanes, C., Watkins, J., Weyler, J., Pereira-Vega,
A., Urrutia, I., Gullón, J.A., Forsberg, B., Probst-Hensch, N., Boezen, H.M., Martinez-
Moratalla Rovira, J., Accordini, S., de Marco, R., Burney, P., 2018. Prevalence of
asthma-like symptoms with ageing. Thorax 73, 37–48.
https://doi.org/10.1136/
thoraxjnl-2016-209596.
Leibler, J.H., Dalton, K., Pekosz, A., Gray, G.C., Silbergeld, E.K., 2017. Epizootics in in-
dustrial livestock production: preventable gaps in biosecurity and biocontainment.
Zoonoses Public Health 64, 137–145.
https://doi.org/10.1111/zph.12292.
Liu, A.H., 2002. Endotoxin exposure in allergy and asthma: reconciling a paradox. J.
Allergy Clin. Immunol. 109, 379–392.
https://doi.org/10.1067/MAI.2002.122157.
May, S., Romberger, D.J., Poole, J.A., 2012. Respiratory health effects of large animal
farming environments. J Toxicol Env. Heal. B Crit Rev 15, 744288.
https://doi.org/
10.1080/10937404.2012.744288.
Michel, O., Rylander, R., Dentener, M., Corazza, F., Buurman, W., 2002. Healthy subjects
express differences in clinical responses to inhaled lipopolysaccharide that are related
with inflammation and with atopy. J. Allergy Clin. Immunol. 107, 797–804.
https://
doi.org/10.1067/mai.2001.114249.
Mirabelli, M.C., Wing, S., Marshall, S.W., Wilcosky, T.C., 2006. Asthma symptoms among
adolescents who attend public schools that are located near confined swine feeding
operations. Pediatrics 118, e66–e75.
https://doi.org/10.1542/peds.2005-2812.
Möller, W., Heimbeck, I., Hofer, T.P.J., Khadem Saba, G., Neiswirth, M., Frankenberger,
M., Ziegler-Heitbrock, L., 2012. Differential inflammatory response to inhaled lipo-
polysaccharide targeted either to the airways or the alveoli in man. PLoS One 7,
e33505.
https://doi.org/10.1371/journal.pone.0033505.
Nazaroff, W.W., 2019. Embracing microbes in exposure science. J. Expo. Sci. Environ.
Epidemiol. 29, 1–10.
https://doi.org/10.1038/s41370-018-0075-4.
Pavilonis, B.T., Sanderson, W.T., Merchant, J.A., 2013. Relative exposure to swine animal
feeding operations and childhood asthma prevalence in an agricultural cohort.
Environ. Res. 122, 74–80. doi.
https://doi.org/10.1016/j.envres.2012.12.008.
Pillai, S.D., Ricke, S.C., 2002. Bioaerosols from municipal and animal wastes: background
and contemporary issues. Can. J. Microbiol. 48, 681–696.
https://doi.org/10.1139/
w02-070.
Poole, J.A., Romberger, D.J., 2012. Immunological and inflammatory responses to or-
ganic dust in agriculture. Curr. Opin. Allergy Clin. Immunol. 12, 126–132.
https://
doi.org/10.1097/ACI.0b013e3283511d0e.
Quanjer, P.H., Stanojevic, S., Cole, T.J., Baur, X., Hall, G.L., Culver, B.H., Enright, P.L.,
Hankinson, J.L., Ip, M.S.M., Zheng, J., Stocks, J., Schindler, C., 2012. Multi-ethnic
reference values for spirometry for the 3–95-yr age range: the global lung function
10