Temporal trends of flour dust exposure in the United Kingdom, 1985-2003.

Temporal trends of flour dust exposure
Martie van Tongeren,*ab Karen S. Galea,a John Ticker,c D
Received 26th March 2009, Accepted 15th June 2009
First published as an Advance Article on the web 6th July 2009
DOI: 10.1039/b906055c
The aim of the study was to examine the long-term changes in inhala
se (
nsf
han
ust
ob
on
d i
dev
occupational exposure limit of 3 mg m�3, Germany have set the
�3
occupation were grouped according to Table 1.
the data, a number of data
sing the following criteria:
n 5 years for an industry or
vailable across all years for an
ts available for an individual
xposure determinants (data
ation, season).
nsformed prior to statistical
analyses. For twomeasurements the flour dust level was recorded
PAPER www.rsc.org/jem | Journal of Environmental MonitoringThe Netherlandsas 0 mg m�3. As information on the specific detection limits of
the methods used was unavailable, the value of 0 mg m�3 was
replaced by an arbitrary value of 0.005 mg m�3, which was half
the lowest observed value (0.01 mg m�3). Statistical analyses were
carried out with linear mixed effect models using the Proc Mixed
procedure in SAS (vs 8.02). Workers were grouped by occupa-
tion. Static measurements were also grouped using the same
occupation classification, linking the most relevant occupation to
aInstitute of Occupational Medicine, Research Avenue North, Riccarton,
Edinburgh, UK EH14 4AP. E-mail: martie.van.tongeren@iom-world.org;
Fax: +44 (0)870 850 5132; Tel: +44 (0)131 449 8097
bCentre of Occupational and Environmental Health, University of
Manchester, UK
cIndependent consultant, 7 Daneway, Ainsdale, Southport, UK
dCentre for Suicide Prevention, University of Manchester, UK
eInstitute for Risk Assessment Sciences, University of Utrecht, Utrecht,limit at 4 mg m , while Finland, Iceland and Norway have
set exposure limits for total organic dust at 5 mg m�3. In the
US, the ACGIH have suggested a TLV of 0.5 mg m�3 for flour
dust.7
As part of a larger project aimed at identifying the long-term
changes in inhalation exposure to various hazardous substances
in a number of industries within the UK8 we investigated the
temporal changes in flour dust exposure in the UK from 1985 to
2003.
Prior to statistical analyses of
exclusion steps were carried out u
(i) Data available for less tha
occupation;
(ii) Less than 50 measurements a
industry or occupation category;
(iii) Less than 50 measuremen
category of other potentially e
source, sample type, sampling dur
All exposure data were log-traUK. Flour dust data held in the UK National Exposure DataBa
company from 1985 to 2003 were reviewed. Analysis of the log-tra
linear mixed effect models and expressed as the relative annual c
flour dust measurements were analysed. The overall mean flour d
the bakeries to 17.9 mg m�3 in the flour mills. Analysis of the data
revealed no statistically significant temporal trends in exposure. C
persisted over the last 20 years, there is a need for government an
measures aimed at reducing flour dust exposures and the risk of
Introduction
Flour is produced by milling and otherwise processing cereals
and is used for human and animal consumption. Occupational
exposure occurs mainly in bakeries, but also in confectionary,
pasta and pizza bakeries, animal feed plants, malt factories and
agriculture. Flour dust contains a number of agents and many
studies have investigated the association between exposure to
flour dust and related allergens, such as alpha-amylase, and
prevalence or incidence of sensitization, respiratory symptoms
and occupational asthma.1–4 A review of the available literature
reported by the European Scientific Committee on Occupa-
tional Exposure Limits (SCOEL) for flour dust did not provide
evidence of a threshold of flour-dust induced effects,5 whilst
Heederik and Houba3 suggested that the lowest observed effect
level for sensitization to flour dust was in the order of 0.5 and
1 mg m�3. In the UK there is currently a Workplace Exposure
Limit (WEL) of 10 mg m�3.6 Some other countries have set
a lower limit, for example, Denmark and Sweden have set an1492 | J. Environ. Monit., 2009, 11, 1492–1497in the United Kingdom, 1985–2003
avid While,bd Hans Kromhoute and John W. Cherriea
tion exposure to flour dust in the
NEDB) and from one large
ormed data was carried out using
ge in exposure. 1451 inhalable
levels ranged from 7.8 mg m�3 in
tained from NEDB and industry
sidering these high levels have
ndustry to implement further
eloping occupational asthma.
Methods
Data on flour dust exposure were obtained from the UK
National Exposure DataBase (NEDB) which contains occupa-
tional exposure data from a wide range of chemical and bio-
logical agents.9 The data were predominantly collected for HSE
inspection purposes, although data from industry-wide surveys
and data from industry sources were also included in NEDB. All
flour dust exposure data from NEDB were extracted. In addi-
tion, data were obtained from one large industrial sized bakery
in the UK. Results from personal and static inhalable dust
measurements were included as were data from short- and long-
term measurements.
The quality of the data was assessed using guidelines based
upon occupational hygiene information, variability and preci-
sion issues and internal and external validity of the data.10,11
Several essential core data criteria were identified, such as
sampling date; sampling device; sample duration; analytical
method; measured concentration and units. Industry sector andThis journal is ª The Royal Society of Chemistry 2009

and plant (i). It was assumed that c(i), dj(i) and 3k(ij) are normally
distributed with zero means, and that these random effects
are statistically independent. It was assumed that any two
measurements on the same worker have equal correlation irre-
spective of the time interval between the measurements, whilst
measurements on different workers are uncorrelated (compound
symmetry covariance structure).12
The fit of the models was assessed by visually examining
residuals plots (predicted versus residuals, residuals versus year,
histogram of residuals) and a plot of actual versus predicted
concentrations. The temporal trends were expressed as the
annual change in geometric mean exposure using the following
expression:
% change per year ¼ 100 � (exp[b1] �1)
Results
Descriptive resultseach location of measurement. Three categories of variables were
included in the mixed models:
1) A continuous variable indicating the temporal trend (year).
2) Variables likely to be associated with exposure and which
could potentially confound or modify the real temporal trend in
exposure. These include the following categorical variables:
occupation; industry; data source (NEDB-inspection data;
NEDB-non inspection data; industry data); sample type
(personal or static); sampling duration (long-term ($60 minutes)
or short-term measurements (<60 minutes)); and season).
3) Variables that represent random effects (plant and worker).
Table 1 Number of inhalable flour dust measurements and number of
years with data by occupation and industry for time period 1985 to 2003a
# Occupation/industry n Yearsb
Occupation
1 Bakery workers 283 14
2 Cleaners 68 7
3 Dough shapers/dusters 143 13
4 Drivers/forklift truck 6 3
5 Kitchen staff 7 3
6 Laboratory/quality assurance assistants 20 3
7 Maintenance workers 2 2
8 Millers 131 5
9 Mixers/weighers 679 14
10 Oven operators 75 11
11 Packers/slicers 100 10
12 Production non-baking 41 4
99 Unknown 1 1
Industry
1 Bakeries 1156 14
2 Food products – jam, sauces, etc. 39 3
3 Food products – pizza, pastries, chicken nuggets 95 8
4 Mills and ingredients 266 5
a n ¼ number of measurements. b Total number of years for which data
were available.Unique identifiers for plant were available in the NEDB and
industry data sets, but unique identifier for worker was only
available for the industry data set. It was assumed that NEDB
contained no repeated measurements on the same worker.
Year was entered first as a continuous variable and the final
model was developed using a systematic forward stepwise
modeling procedure based on the improvement in fit between
2 successive models. Likelihood ratio tests were used to deter-
mine if inclusion of variables significantly improved the fit of the
model (p < 0.05). Modeling of the main effects ended when there
was no further improvement in fit from systematically testing
main effects. The following model was used to describe the
various data sets:
ln(Ek(ij)) ¼ b0 + b1$Year + bC1$C1 +.+ bCn$Cn + ci + dj(i) + 3k(ij) (1)
for i ¼ 1,2,.,ni plants; j ¼ 1,2,.,nj(i) workers (nested within
plant); and k ¼ 1,2,.,nk(ij) measurements (nested within
worker). The fixed effects C1,.,Cn represent the variables that
can potentially confound or modify the temporal trend, with
bC1,.,bCn as the regression coefficients for these fixed effects. cI
represents the random effect of plant (i); and dj(i) represents the
random effect of worker (j) nested within plant (i), whilst 3k(ij)
represents the random effect of day (k) nested within worker (j)
This journal is ª The Royal Society of Chemistry 2009In total, 1,556 inhalable dust measurements were available from
1985 to 2003 (Table 1). Applying the exclusion criteria resulted
in exclusion of 105 measurements due to insufficient data in some
occupations (driver, kitchen staff, laboratory, maintenance,
production-non baking and unknown) and industries (‘Food
products – jam, sauces, etc.’). Therefore, 1,451 inhalable flour
dust measurements were available for the statistical analyses. The
majority of these measurements were collected by industry after
1995 (n ¼ 911, 63%). Prior to 1995, all exposure data were from
the NEDB (Table 2). The dataset include results from personal
(n ¼ 1,084) as well as static measurements (n ¼ 367) and results
from long and short-term (<1 hours) measurements.
The arithmetic mean (AM) of the whole data set was
9.7 mg m�3, with a geometric mean (GM) of 3.4 mg m�3 and
a geometric standard deviation (GSD) of 4.1. Some very high
dust levels were included in the database, with a maximum value
of 1,148.7 mg m�3. Personal measurements generally resulted
in higher levels than those measured by static sampling (GM ¼
Table 2 Inhalable flour dust exposure (mg m�3) by potentially con-
founding variablesa
n AM GM GSD Min Max
Data source
NEDB Inspection 471 11.4 4.7 3.6 0.1 571.9
NEDB Non-inspection 69 5.6 3.6 2.7 0.5 34.3
Industry data 911 9.1 2.8 4.4 0.0 1148.7
Sample type
Personal 1084 11.3 4.3 3.6 0.1 1148.7
Static 367 5.0 1.6 4.8 0.0 115.0
Sampling duration
Short term 163 21.5 8.9 3.8 0.3 571.9
Long term 1288 8.2 3.0 4.0 0.0 1148.7
Season
Winter 426 9.6 3.7 4.2 0.04 213
Spring 393 8.1 3.2 3.6 0.01 571.9
Summer 360 6.4 2.7 4.2 0.01 111.4
Autumn 272 16.5 3.9 4.6 0.04 1148.7
a n: number of measurements, AM: Arithmetic mean, GM: Geometric
mean, GSD: Geometric standard deviation, Min: Minimum exposure,
Max: Maximum exposure.J. Environ. Monit., 2009, 11, 1492–1497 | 1493

tion
Dt
mea
ic s4.3 mgm�3 vs 1.6 mg m�3, respectively), whilst results from short-
term measurements were generally higher than those from
long-term measurements (GM ¼ 8.9 mg m�3 vs 3.0 mg m�3,
respectively). Exposure levels were generally higher for the
NEDB data (GM ¼ 3.6 mg m�3 for non-inspection data and
4.7mgm�3 for inspection data, respectively), than for the industry
data (GM ¼ 2.8 mg m�3). No data from inspection visits were
available from 1993 to 2000. Data from non-inspection visits are
only sporadically available, whilst data from industry are only
available from 1995 onwards. Inhalable flour dust exposure
appears to be higher during the autumn (GM¼ 3.9 mg m�3) with
lowest levels observed during the summer (GM ¼ 2.7 mg m�3).
Table 3 details the descriptive statistics for industry and occu-
pation. High exposure levels were observed in all sectors, but were
particularly high in ‘Mills and ingredients’ (GM¼ 4.0mgm�3) and
in ‘Food products – pizza, etc’ (GM ¼ 5.5 mg m�3). The highest
exposure levels for occupations were observed for ‘Cleaning staff’
with aGMof 6.9mgm�3. The lowest but still substantial exposure
levels were observed for the ‘oven operators’ (GM ¼ 2.2 mg m�3)
and the ‘Packers/slicers’ (GM ¼ 1.6 mg m�3). After excluding
short-term and static measurements very similar results were
Table 3 Inhalable flour dust exposure (mg m�3) by industry and occupa
n AM GM GS
Industry
Bakeries 1148 7.8 3.2 3.9
Food products – pizza, etc. 67 13.0 5.5 4.2
Mills and Ingredients 236 17.9 4.0 5.1
Occupation
Bakery workers 283 7.9 3.5 3.6
Cleaners 67 42.4 6.9 5.8
Dough shapers/dusters 143 8.5 4.5 3.4
Millers 131 8.2 3.3 4.3
Mixers/weighers 662 9.2 3.4 4.3
Oven operators 75 4.0 2.2 3.3
Packers/slicers 90 3.4 1.6 3.6
a n: number of measurements, AM: Arithmetic mean, GM: Geometric
geometric standard deviation, GSDbw: Between worker/location geometr
deviation, Min: Minimum exposure, Max: Maximum exposure.obtained, with GM for ‘Mills and ingredients’ being 5.8 mg m�3
and the GM for ‘Food products – pizza, etc’ being 5.7 mg m�3.
When summarizing the result by occupation using only long-term
personal measurements, the results were again very similar to
the whole dataset with the GM exposure level ranging from
1.9 mgm�3 for ‘Packers/slicers’ to 7.0 mgm�3 for ‘Cleaning staff ’.
Table 3 also shows the between-plant, between-worker/loca-
tion and within-worker/location or day-to-day variability. No
repeated measurements on individual workers were available for
the ‘Food products – pizza, etc.’ sector and for the occupations
‘Dough shapers/dusters’ and ‘Packers/slicers’, hence no within-
worker GSD could be estimated. The between-plant GSD was
relatively high for the ‘Cleaners’, indicating that exposure levels
for this group varied widely between different sites. The between-
worker GSDs were similar across the jobs, ranging from 2.2 for
‘Cleaners’ to 2.8 for ‘Packers/slicers’. Relatively high within-
worker GSDs were observed for ‘Millers’, ‘Mixers/weighers’ and
‘Cleaners’, indicating that exposure levels for the same worker
could vary significantly from one day to the next.
1494 | J. Environ. Monit., 2009, 11, 1492–1497Mixed effect models
Table 4 provides information on the model fit for the various
stages in the model development, together with parameter esti-
mates for annual trend and the variance components in the
exposure data. In the order that they were entered, the final
model includes the main effects of year, sample type (personal vs
static samples), sampling duration (long-term vs short-term
measurements), occupation, season and industry. Inclusion of
year into the model did not significantly improve the fit of the
model, nor was the effect of year statistically significantly asso-
ciated with flour dust exposure at any point in the model devel-
opment. The between-plant variance was decreased by just 7% in
the final model compared to model that only included random
effects, whilst the between-worker variance decreased by 39%.
The within-worker variance was not reduced in the final model
compared to the random-effects model. The mixed effects anal-
yses were repeated after excluding all the static and short-term
measurements, which did not alter the annual trend in flour dust
exposure; a non-significant annual trend of �0.1% was found.
Table 5 provides the parameter estimates from the final mixed
a
ot GSDbp GSDbw GSDww Min Max
2.0 2.3 2.3 0.0 571.9
2.2 3.6 — 0.0 175.0
2.2 2.7 3.0 0.0 1148.7
2.2 2.5 1.7 0.0 213.0
3.7 2.2 2.4 0.1 1148.7
2.4 2.5 — 0.2 70.1
1.8 2.3 2.9 0.0 112.6
2.3 2.2 2.6 0.0 571.9
1.9 2.3 1.8 0.0 26.8
2.3 2.8 — 0.1 29.1
n, GSDtot: Total geometric standard deviation, GSDbp: Between plant
tandard deviation, GSDww: Within worker/location geometric standardmodel. The results from the mixed effects model confirms that
the geometric mean for the static measurements was lower than
the geometric mean for personal samples and that geometric
mean from long-term measurements was lower than short-term
measurements. Exposure levels were higher during the winter
season than during other seasons. For industry, it appears that
the ‘Mills and Ingredients’ have the highest exposure levels whilst
‘Bakeries’ have the lowest. The ‘Packers/slicers’ have lower
exposure levels compared to the other occupations.
Stratified analyses for ‘Bakeries’ and ‘Mills and ingredients’
were also carried out. After correcting for sample type, sampling
duration, occupation and season, a non-significant annual trend
of �1.0% was observed for inhalable flour dust exposure for
‘Bakeries’. In contrast, for ‘Mills and ingredients’ a non-signifi-
cant annual increase of nearly 15% was observed for inhalable
flour dust exposure in this industry, suggesting that exposure
levels may have increased in recent years. Finally, stratified
analyses were carried out for data source (NEDB inspection
survey, NEDB non-inspection survey and industry); no annual
This journal is ª The Royal Society of Chemistry 2009

A2.
2.8
0.8
0.5
0.5
0.1
n cr
viou
em
twe
el;
Ye
¼Table 4 Information on stages of mixed effects model developmenta
Modelb -2RLL AIC Diff df p-value byear SE
0 4912.5 4918.5
1 4915.4 4921.4 �2.8 1 NEc 0.024 0.012
2 4803.9 4809.9 111.5 1 <0.001 0.027 0.011
3 4734.3 4740.3 69.6 1 <0.001 0.008 0.011
4 4709.9 4715.9 24.4 6 <0.001 0.005 0.012
5 4695.1 4701.1 14.8 3 0.002 0.005 0.012
6 4688.3 4694.3 6.8 2 0.034 0.001 0.012
a �2RLL: �2 Residual Log Likelihood estimate; AIC: akaike informatio
inclusion of the additional variable (ie difference with the model in the pre
variable; p-value: p-value of log-likelihood ratio test, to determine improv
Between-plant: between plant variance in exposure; Between-worker: be
worker or day-to-day variance. b 0: no fixed effects included in the mod
duration; 4: Year, Sample type, Sampling duration and Occupation; 5:
Sample type, Sampling duration, Occupation, Season and Industry. c NEchange in flour dust exposure was observed in any of these three
data sources.
Discussion
Results from the analyses presented in this paper suggest that the
flour dust exposure levels in the UK are relatively high, with
some individual measurements results in excess of 100 mg m�3.
Furthermore, no downward temporal trend in flour dust expo-
sure in the UK was observed over the period 1985–2003. This is
in contrast to many other occupational exposures in other
industry sectors where annual trends in the range of approxi-
mately �5 and �15% have been observed.13 When stratified
analyses were carried out by industrial sector, a non-statistically
significant increasing trend (15% per year) was observed between
Table 5 Parameter estimates, standard errors and p-values for the final
mixed effects regression model for log-transformed inhalable flour dust
exposureb
Description of effect b SE
Intercept 2.234 0.339
Yeara 0.001 0.012
Sample type Static �0.873 0.079
Personal Base
Sampling duration Long-term �0.929 0.112
Short-term Base
Occupation Bakery workers 0.675 0.165
Cleaners 0.621 0.223
Dough shapers/dusters 0.749 0.178
Millers 0.305 0.245
Mixers/weighers 0.880 0.149
Oven operators 0.492 0.205
Packers/slicers Base
Season Autumn �0.243 0.135
Spring �0.325 0.116
Summer �0.496 0.113
Winter Base
Industry Bakeries �0.610 0.220
Food products - pizza, etc �0.138 0.319
Mills and ingredients Base
a The values for the variable year were changed from 1985 to 2003 to 0 to
18. b b: Parameter estimate for fixed effects, SE: Standard error of
parameter estimate.
This journal is ª The Royal Society of Chemistry 20091999 and 2003 in ‘Mills and Ingredients’. Unfortunately, no data
were available before 1999; therefore, it was not possible to
determine the temporal trend before that year, although it seems
unlikely that this trend of increasing flour dust exposure over
time had existed for many years prior to 1999. In the Netherlands
comparison of exposure levels from studies performed between
1993 and 2005 also indicate no clear trend in exposure levels for
flour dust, wheat allergens and a-amylase.14,15 However, de Pater
et al.14 reported that for flour mills, data on flour dust exposure
available from 1988 and 2001 seemed to indicate an overall
downward trend in exposure.
We assume that inhalation exposure levels in industry tend to
decline exponentially, rather than, for example, in a stepwise
manner. Although it is possible that exposure levels within one
plant may change abruptly at a number of time points, when
analysing data across one or several industrial sectors, it is
unlikely that these changes all occur at the same time. Hence, we
feel that the use of a continuous variable to indicate the temporal
change in exposure is justified.
Flour dust exposure levels vary between occupation and
industry and therefore the analyses of trends were corrected for
these variables as the frequency of measurements in different
occupations and industry sectors may have changed over time. In
nnual Trend (%) Between-Plant Between-worker Within-worker
0.5395 0.7300 0.8106
5 0.5501 0.7293 0.8060
0.4931 0.5556 0.8613
0.4739 0.5526 0.7965
0.5148 0.5085 0.8021
0.5356 0.4437 0.8427
0.5039 0.4443 0.8410
iterion (smaller is better); Diff: difference in �2RLL associated with the
s row); df: degrees of freedom associated with inclusion of the additional
ent in fit; byear: parameter estimate for year; SE: standard error for byear;
en worker (within plant) variance in exposure; Within-worker: within-
1: Year; 2: Year and Sample type; 3: Year, Sample type and Sampling
ar, Sample type, Sampling duration, Occupation and Season; 6: Year,
p-value not estimable.addition, any systematic changes over time in the sample type,
sample duration or season may have resulted in an apparent
change in exposure, even if in reality the exposure levels remained
the same. Therefore, all analyses of temporal trends were
adjusted for these variables. However, this approach could
potentially disguise any temporal trends that exist in subsets of
the data. To review this, several stratified analyses were also
carried out, by occupation, industry, data source, sample type
and sample duration. None of these stratified analyses revealed
any statistically significant temporal trends in subsets of the
exposure data, although for industry (‘Mills and ingredients’)
a non-significant annual increase of nearly 15% was observed.
Guidance on dust control in bakeries has been issued in the
UK. For example, the Health and Safety in Bakeries Liaison
Committee (HSBLC), which consists of representatives from all
the relevant trade associations and unions, together with the
HSE produced the publication ‘Guidance on dust control and
health surveillance in bakeries’.16 The industry, in cooperation
J. Environ. Monit., 2009, 11, 1492–1497 | 1495

with the HSE also produced an interactive training package in
1998 entitled ‘‘Breathe Easy’’ which includes a copy of the
guidance, training notes and a video illustrating the control
measures specified in the guidance.17 However, there is currently
no strong evidence that these initiatives had a major impact on
exposure to flour dust. This was confirmed by a recent UK wide
flour dust survey, involving 55 randomly selected bakeries which
suggested that there was limited knowledge of good working
practices.18 Note that these data were not included in our data-
base. Activities such as dry sweeping and flour dusting by hand
were still repeatedly undertaken by the majority of bakeries, local
ventilation was present in only 28% of companies and less than
half of the companies were judged to have adequate control
measures.
Exposure to flour dust has been associated with increased risk
for developing sensitization and occupational asthma, with no
evidence of a trustworthy threshold of flour-dust induced
effects,5 possibly because the allergenic components of flour dust
linked to asthma do not exist in constant proportion to the flour
dust itself.19 In the UK, it is estimated that in the last 10 years at
least 600 people have developed asthma as a result of flour (or
grain) exposure.20 The HSE also reports that between the years
2000 and 2002 the number of new asthma cases caused by flour
and grain remained at about the same level, while new reported
cases of occupational asthma from most causes fell during the
same period.20 A limited analysis of the actual and estimated
cases of occupational asthma attributed to flour dust reported to
the Surveillance of Work-Related & Occupational Respiratory also to Hilary Cowie, IOM, for her helpful comments during
A. Lorusso, A. Mariano, P. L. Paggiaro, D. Talini, G. Pisati,Disease (SWORD) by chest physicians between 1997 and 2006
also suggest that the numbers of actual and estimated cases are
not reducing in number (Table 6).
Table 6 Actual and estimated cases of occupational asthma attributed to
flour dust reported to SWORDby chest physicians by years 1997–2006 a,b
Year Actual cases Estimated cases
1997 17 83
1998 14 14
1999 13 24
2000 7 29
2001 11 44
2002 23 67
2003 12 12
2004 23 34
2005 19 52
2006 23 34
Total 162 393
a Source personal communication from The Health and Occupation
Reporting network (THOR) (http://www.medicine.manchester.ac.uk/
coeh/thor). b Note: The cases were reported by a selection of the UKs
chest physicians with an interest in occupationally related diseases.
Some physicians (core group) reported cases to the reporting scheme
throughout the year, while others (sample group) only reported during
one (randomly selected) month of the year. The actual number of cases
is the number of cases reported to the scheme. The estimated number
of cases is the total number of cases that would have been reported if
all physicians participating in the scheme were reporting throughout
the year: Estimated cases ¼ (cases reported by core group) + (cases
reported by sample group during a single randomly allocated month
per year � 12).1496 | J. Environ. Monit., 2009, 11, 1492–1497C. Romano and F. Sulotto, Int. Arch. Occup. Environ. Health.,
1999, 72, 335–37.
3 D. Heederik and R. Houba, Ann. Occup. Hyg., 2001, 45, 175–85.
4 R. Houba, D. Heederik and G. Doekes, Am. J. Resp. Crit. Care.
Med., 1998, 158, 1499–1503.
5 SCOEL, Recommendation from the Scientific Committee on
Occupational Exposure Limits for flour dust. SCOEL, 2007.
Available at URL: http://ec.europa.eu/employment_social/health_
safety/docs/sum_123.pdf (accessed September 2007).
6 HSE, EH40/2005 Workplace Exposure Limits: Containing the list of
workplace exposure limits for use with the Control of Substances
Hazardous to Health Regulations 2002. HSE Books, 2008.
7 E. Mounier-Geyssant, J.-F. Barth�elemy, L. Mouchot, C. Paris and
D. Zmirou-Navier, BMC Public Health, 2007, 7, 311.
8 K. S. Creely, M. van Tongeren, D. While, A. J. Soutar, J. Tickner,
M. Agostini, F. de Vocht, H. Kromhout, M. Graham, A. Bolton,
H. Cowie and J. W. Cherrie, Trends in inhalation exposure: Mid
1980s till present. HSE Research Report 460, HSE Books, 2006.
Available at URL http://www.hse.gov.uk/research/rrpdf/rr460.pdf
(accessed September 2007).
9 D. K. Burns and P. L. Beaumont, Ann. Occup. Hyg., 1989, 33, 1–14.
10 P. J. Ritchie and J. W. Cherrie, Appl. Occup. Environ. Hyg., 2001, 216,
295–99.
11 E. Tielemans, H. Marquart, J. De Cock, M. Groenewold and J. van
Hemmen, Ann. Occup. Hyg., 2002, 46, 287–97.
12 C. Peretz, A. Goren, T. Smid and H. Kromhout, Ann. Occup. Hyg.,
2002, 46, 69–77.the preparation of this manuscript. The work was funded by the
Health and Safety Executive in the UK under Contract No
D4807.
References
1 P. Cullinan, A. Cook,M. J. Nieuwenhuijsen, C. Sandiford, R. D. Tee,
K. M. Venables, J. C. McDonald and A. J. Newman Taylor, Ann.
Occup. Hyg., 2001, 45, 97–103.
2 R. De Zotti, M. Bovenzi, C. Negro, A. Sirla, A. Innocenti,It is clear that there is a need for government and industry to
implement further measures to improve the communication and
promotion of good practice in controlling flour dust exposure so
that possible future downward trends, both in flour dust exposure
and the incidence of occupational asthma attributable to this, are
observed. Given that no data were available for analysis after
2003, the effect of the introduction in the UK of the 10 mg m�3
flour dust exposure limit in 2001 is difficult to determine.
However, we believe that this limit must be substantially lower if
progress is to be made in reducing the incidence of occupational
asthma from flour dust exposure.
In the UK the collection and storage of new measurement data
appears to be decreasing. For example, as Tickner21 discussed in
his review of NEDB, the addition of new data within this data-
base is diminishing, in part due to changes in operational
procedures and also due to the costs associated with data
collection. It should be recognized that there is a continuing need
to collect and systematically store new exposure measurement
data to help assess the effectiveness or otherwise of the Work-
place Exposure Limits (WELs) as well as any future initiatives to
ensure that inhalation exposure to flour dust decreases.
Acknowledgements
We would like to thank Andy Phillips, Celia Elliott-Minty and
Peter Griffin at the UK Health and Safety Executive (HSE) for
their help and encouragement throughout the project. ThanksThis journal is ª The Royal Society of Chemistry 2009