1
Client
Report
1982
Prepared for MAF Policy
Act
March 2006
Key indicators of
poultry welfare in
New Zealand
C. S. Bagshaw
Information
L. R. Matthews
A. Rogers
Official
the
under
Released
1
2
EXECUTIVE SUMMARY
• There is currently a high level of interest by policy/regulatory
authorities, the public, welfare organisations, growers and lobby
groups in the welfare of poultry kept for meat production.
• Thus, there is an urgent need to identify key welfare indicators
relevant to New Zealand conditions, use these measures to
assess the welfare status of broilers under common farming
practices, and, where necessary, revise practices in industry, and
1982
standards and recommendations in the relevant welfare codes. Act
• At the beginning of the study, a Broiler Welfare Research
Committee (BWRC) was formed (with representation from
Poultry Industry Association of New Zealand (PIANZ), poultry
veterinarians and representatives from the major meat chicken
companies in New Zealand, a scientist from AgResearch and an
Information
independent scientist to oversee the research programme.
• The aim of the research was to identify appropriate welfare
Official
indicators for broilers, and apply them to quantify the welfare
status of broilers in New Zealand.
the
• This study used a suite of measures to assess welfare including:
the incidence (and causes) of mortalities and culls; skin disorders
under
of the feet, hocks and breasts; gait score; and the incidence of
back scratches, birds dead on arrival (DOA), rejects.
• The potential influence of management factors on welfare was
Released
examined through information collected in a grower survey.
• The potential influence of stocking density was examined by
correlating welfare measures with variations in stocking density
that arose naturally during the study.
2
3
• The research was carried out on 36 New Zealand housed
commercial broiler farms in each of the four seasons and within
the context of the typical management of the birds. There were
few differences in husbandry practices between farms.
• Overall mortality was 3.8%, which is at the lower end of levels
typically seen in the industry overseas. Of all the birds in the
study, 0.3% died as leg culls.
•
1982
The contribution of leg culls to total mortality was relatively low
(8%) and about half that reported for birds in two recent UK
Act
studies (15%).
• The proportion of birds with severe leg weakness (gait score 4
(0.78%) and 5 (0.1%)) was about one quarter that reported for
broilers in the UK, Denmark and Sweden.
Information
• Considering both the leg culling and gait scoring information
together, our results demonstrate that leg health in New Zealand
broilers is better than for birds in Europe.
Official
•
the
Prevalence of all forms of contact dermatitis (footpad, hock burn
and breast blisters) was within the range reported internationally.
Incidences, such as occurred, were mostly of a relatively minor
under
nature. These data indicate that the litter in the houses is
maintained in good condition, although there were seasonal
variations.
• Pre-slaughter handling procedures, as assessed by Dead On
Released
Arrivals and reject birds, were generally better than reported for
other countries. The proportion of birds with back scratches
averaged 39%. Meaningful comparisons between our data and
the incidence of scratches reported elsewhere were not possible
due to the lack details of the measurement criteria provided in
other published reports
3
4
• Stocking density varied between 32.9 and 39.3 kg m-2. Stocking
density was not correlated with any of the welfare measures, or
variables in the epidemiological survey, suggesting that broiler
welfare is either unaffected by stocking density, or that the range
in densities seen in the current study was too small to enable
correlations to be detectable. Stocking density (at the levels
currently used in this study) would appear appropriate as the
welfare of the birds is good.
• The welfare of broilers in New Zealand is equal to or superior to
1982
that of broilers kept elsewhere.
Act
Information
Official
the
under
Released
4
TABLE OF CONTENTS
Executive Summary .............................................................................................................. 2
Table of contents............................................................................................................... 1
INTRODUCTION .................................................................................................................. 3
METHODS ............................................................................................................................ 9
Farms ................................................................................................................................ 9
Animals ........................................................................................................................... 10
1982
Timing of measurements ............................................................................................. 10
Measures......................................................................................................................... 11
Act
Epidemiological questionnaire ..................................................................................... 11
Mortality........................................................................................................................... 11
Leg heath..................................................................................................................... 12
Contact dermatitis ........................................................................................................... 13
Information
Foot pad dermatitis ...................................................................................................... 13
Hock Burn .................................................................................................................... 13
Breast blisters (blisters) ............................................................................................... 13
Official
Pre-slaughter handling .................................................................................................... 14
the
Back Scratches............................................................................................................ 14
Dead on arrival (DOA) and rejects............................................................................... 14
under
Rejects......................................................................................................................... 14
STATISTICAL ANALYSES ................................................................................................. 15
Response data ................................................................................................................ 15
Epidemiological survey data............................................................................................ 15
Released
The model ....................................................................................................................... 15
Training data ................................................................................................................... 16
Reliabilty.......................................................................................................................... 16
RESULTS ........................................................................................................................... 17
Common husbandry procedures and welfare measures ................................................. 17
2
Welfare measures and correlation with husbandry procedures....................................... 18
Incidence of mortality ................................................................................................... 18
Incidence of leg weakness........................................................................................... 25
Incidence of contact dermatitis .................................................................................... 28
Incidence of back scratches, DOAs (birds dead on arrival) and
rejected birds. .............................................................................................................. 34
DISCUSSION...................................................................................................................... 39
References ......................................................................................................................... 52
1982
References ......................................................................................................................... 52
Act
Information
Official
the
under
Released
2
3
INTRODUCTION
There is currently a high level of interest by policy/regulatory
authorities, the public, welfare organisations, growers and lobby
groups in the welfare of poultry kept for meat production. Much of
this interest has been brought into focus by the New Zealand
Animal Welfare Act (1999) and associated Code development. In
order for all interested parties to be confident that the standards of
welfare specified in Codes and practised on farms are at
acceptable levels, scientifically-defensible information on the
welfare of broilers in New Zealand is required. Published scientific
1982
information has been reviewed for MAF (Bagshaw and Matthews,
2001). There is a considerable international literature on broiler
Act
welfare but, unfortunately, there are no relevant published details
on the welfare status of poultry in New Zealand. Bagshaw and
Matthews (2001) indicated that while there are many similarities in
the way that poultry are farmed in NZ and elsewhere, there are
likely to be important differences (e.g. in disease status) which may
Information
influence overall welfare. Thus, there is an urgent need to identify
key welfare indicators relevant to New Zealand conditions, use
these measures to assess the welfare status of broilers under
common farming practices, and, where necessary, revise practices
Official
in industry, and standards and recommendations in the relevant
welfare codes.
the
In the United Kingdom a survey of the welfare status of broilers
has been undertaken at Bristol University. The data generated by
under
Bristol would benchmark broiler welfare in the UK. It was proposed
that a New Zealand study on the welfare of broilers be of a similar
format to that in the Bristol study. It is important that the work
carried out in NZ is internationally credible. In addition, comparing
welfare standards between countries (e.g. UK and NZ) would be
Released
useful and will be carried out when both reports are available.
The main causes of poor welfare in other countries are leg
disorders, diseases, skin disorders (primarily associated with poor
environmental conditions) and damage during catching and
transport to slaughter. The low levels of activity and a limited
3
4
behavioural repertoire are also considered by some authors to
indicate poor welfare (Berg et al., 2000).
Leg weakness can arise from infectious and non-infectious causes
and its prevalence can be influenced by a wide range of
management factors (e.g. genotype, live weight, feeding level,
lighting regime, activity levels) (Butterworth, 2004).
Typically, leg weakness has been assessed using the gait scoring
procedure developed by Kestin et al.(1992) which uses a 6 point
1982
gait score (GS) scale (0 being normal, through to 5 being
immobile). Using this system with the Ross genotype and birds
Act
aged 42 to 49 days and weighing 2.6 to 2.8 kg, Kestin and
colleagues reported that the percentage of birds with gait scores 0
to 5 were 8.5, 22, 43, 21, 4 and 2.5, respectively. Sanotra et al.
(2003) reported percentages of 26, 25, 25, 19, 4 and <1% in these
GS categories for Ross and Cobb genotypes slaughtered at 35
days of age and with a mean weight of 1.8 kg . It should be noted
Information
that the percentage with gait score 0 (26%) is extremely high
compared with other published reports.
Official
Leg weakness might adversely affect welfare either directly (e.g.
through pain) or indirectly (e.g. through preventing the fulfilment of
the
essential functions like feeding).
There is some evidence, though not unequivocal, to suggest that
higher gait scores (3 and over) result in broilers experiencing pain.
under
The evidence has been derived from two types of studies: in one,
the effects of analgesics on behaviour have been measured; and
in the second, the effects of gait on the propensity to exhibit
specific important activities have been assessed. Using an
analgesic/anti-inflammatory drug, carprofen, (McGeown et al.,
Released
1999) compared the time taken by broilers with gait scores of 0 or
3 to traverse an obstacle course, with and without prior drug
administration. Gait score 3 birds birds with and without carprofen
negotiated the course in 18 and 38 seconds, respectively, while
birds with sound legs took 11 seconds. The times taken for the
sound birds and carprofen treated gait score 3 birds were not
significantly different. Danbury et al. (2000) reported two
4
5
experiments in which gait score 0 and 3 birds were offered a
choice of carprofen and untreated food. In one study, there was no
difference in the consumption of drug-treated and untreated food
between gait score treatments and, in the second, the amount of
drug-treated food eaten increased with gait score. These data
were interpreted as indicating an increasing level of drug self-
selection (and therefore pain relief) with increasing gait score.
However, this study requires replication before firm conclusions
can be drawn due to: the inconsistencies between the two
experiments; the lack of a significant difference in drug
1982
concentration in the plasma between birds with sound and
unsound legs; the correlation between consumption of drug-treated Act
food and gait score apparently being driven entirely by a lower
consumption rate by the gait score 0 birds; and the measures of
gait score used in the analyses being taken at the end of the study,
meaning that drug consumption and gait score were totally
confounded. Further, there were no differences in the plasma
concentrations of carprofen between birds with different gait
Information
scores.
Indirect measures of the welfare impacts of variations in gait score
Official
include the research by Bokkers and Koene (2004) and Weeks
and Kestin (1997) where the behaviour of birds with different gait
the
scores has been compared. While some aspects of broiler
behaviour are modified in a way that suggests that lameness (GS
2 and 3) imposes a cost on the birds (e.g. increased feeding bout
under
duration, reduced motivation to feed), there is no effect of
lameness on total time spent feeding or drinking (Ross genotype;
Weeks et al., 2000). Nonetheless, as Mench (2004) has aptly
stated, many gait disorders must be painful as a consequence of
the clinical manifestations involving inflammation, spinal cord
Released
damage, tension on the joints or rupture of tendons.
Broilers live on a litter substrate. If the litter is too damp, then the
birds may develop contact dermatitis on the foot (footpad
dermatitis), hock (hock burn) or breast (breast blister). Dryness of
litter is influenced by a wide range of factors including the
effectives of the ventilation system, climate, food composition, litter
5
6
type, stocking density, drinker system and bird health (Berg, 2004).
The mean percentage of birds seen with the various forms of
contact dermatitis is 7 to 19% (range up to 90%) for hock burn
(Dawkins et al., 2004; Menzies et al., 1998), 19% for foot pad
dermatitis (Dawkins et al., 2004), and 0.0 to 0.4% for breast
blisters (Berg, 2004; Hall, 2001).
Average total mortality in flocks is reported to vary from a low of
about 2.5% (Butterworth, 2004), but is more likely to be in the
range from 4 to 6.5% (Hall, 2001, Dawkins et al., 2004). About
1982
15% of the total mortality arises due to culling for leg weakness,
35% due to culling for other reasons, and 50% from natural deaths Act
(Hall, 2001, Dawkins et al., 2004). As a proportion of the flock, leg
culling rates have been observed in the range 0 to 4% (McNamee
and Smyth 2000; Dawkins et al., 2004).
The most common causes of leg weakness in broilers are reported
to be femoral head necrosis (17 to 38%), limb deformities (11 to
Information
14%), tibial dyschondroplasia (7 to 8%), Infections (13%), and
spondylolisthesis (kinky back) (3 to 8%) (Butterworth, 1999;
McNamee and Smyth, 2000). Official
Pre-slaughter handling is comprised mainly of the catching,
the
containment and transport of birds to slaughter. Measures to
assess the impact of these procedures on bird welfare include
mortalities post-transport (dead on arrival, DOA) (Mitchell and
under
Kettlewell, 2004), rejects or downgrades of the carcasses due to,
for example, trauma or bruising (Scott, 1993), and other measures
of injury (e.g. back scratches (Hall, 2001). Of course, some
traumatic events and other variables (e.g. disease conditions,
(Ekstrand, 1998) occurring prior to the pre-slaughter handling
Released
period influence the incidence of, and therefore confound, these
measures. Ekstrand (1998) reported the range for DOA is typically
0.1 to 0.6%, and for rejects/downgrades is very wide at 0.5 to 20%.
Hall (2001) observed between 0.3 and 0.5 % of birds at slaughter
with back scratches (Ross genotype). Frankenhuis et al. (1991)
demonstrated that scabby hip can be caused by scratches to the
6
7
back; the incidence of scabby hip varies greatly between flocks
and can be up to 60%.
Stocking density is frequently raised in discussions of broiler
welfare, although the evidence is equivocal on the relationship
between stocking density and welfare. Bagshaw and Matthews
(2001) suggested that high litter temperatures rather than stocking
density per se may contribute to poor welfare.
This study used a suite of measures to assess the welfare of
1982
broilers in NZ. Health status was assessed from the incidence of
mortalities (and causes) and culls; skin disorders of the feet (Berg, Act
1998), hocks (Butterworth, 2002) and breasts; leg health was
assessed with a gait scoring procedure (Kestin et al 1992);
environmental conditions were assessed by measuring the
incidence of skin disorders; and pre slaughter handling was
measured from the incidence of birds dead on arrival (DOA),
rejects and back scratches.
Information
The potential influence of management factors on welfare was
examined through association of management parameters
Official
(obtained in an epidemiological questionnaire) with the health and
welfare measures. It was beyond the scope of the present study to
the
measure directly the behavioural repertoire of the broilers or the
effects of stocking density on welfare. The potential influence of
stocking density was examined by correlating welfare measures
under
with variations in stocking density that arose naturally during the
study.
Aims: To identify appropriate welfare indicators for broilers, and
apply them to quantify the welfare status of broilers in New
Released
Zealand.
At the beginning of the study, a Broiler Welfare Research
Committee (BWRC) was formed (with representation from Poultry
Industry Association of New Zealand (PIANZ), poultry veterinarians
and representatives from the major meat chicken companies in
7
8
New Zealand, a scientist from AgResearch and an independent
scientist to oversee the research programme.
1982
Act
Information
Official
the
under
Released
8
9
METHODS
In conjunction with BWRC meetings, liaison with staff from the
University of Bristol, visits to NZ slaughter plants, and attendance
at a gait scoring course, protocols were prepared for the collection
of data for the following measures: mortality, leg health, contact
dermatitis and pre-slaughter handling and catching damage. In
addition, information on management practices was collected
using a questionnaire.
FARMS
1982
The research was carried out on commercial broiler farms and
within the context of the typical management of the birds.
Act
To determine the number of sheds to study, historical mortality
data was analysed to assess the most appropriate sample size
(number of sheds) required to measure mortality reliably and to
obtain some indication of seasonal variability in health and welfare
of the birds. A shed refers to one growing cycle of a flock of
Information
broilers from placement to slaughter. These analyses indicated
that, ideally, 36 sheds (one shed per farm) over the four seasons
ought to be examined each year. However, for logistical reasons
Official
(i.e. availability of veterinarians, financial constraints), a total of 36
sheds only could be used in the study (12 sheds were examined in
the
the first year, and 24 shed in the second year). Sheds were
allocated to companies in proportion to the number of birds they
process over the industry. Nine sheds were examined in each
under
season, three for each season in the first year and six in the
second year.
In summary, only one shed was examined on each farm in the
study. A total of 36 farms/sheds were sampled over the 18 months
Released
of the study. Each shed was visited by researchers between 1 to 3
times, depending on slaughter dates for the flock. During the study
a total of 85 visits to the sheds were carried out (36 sheds had one
visit, and of those 32 had a second visit, and of those 17 sheds
had a third visit). The experimental protocol was tested on one
additional farm prior to the start of the project (these data were
included in the analyses).
9
10
There were three common sets of management practice
(designated A, B and C). They varied in terms of feeding,
genotype, lighting regime and other factors. For commercial
sensitivity reasons, these cannot be described in detail.
Kestin et al. (1992) demonstrated that a sample size of at least 250
birds per shed was required to measure reliably walking ability (leg
health). In our study, samples of at least 250 birds per shed were
used.
1982
In the slaughter plant, approximately 100 birds were used to
Act
measure skin disorders (Ekstrand et al., 1998) and back scratches.
ANIMALS
There were two genotypes (breeds) used in this study, Ross and
Cobb. Table 1 shows the average number of chicks, the average
Information
shed size, stocking rate and stocking density for birds of each
breed for the sheds used in the study.
Official
Table 1. Averages (across sheds) for the numbers of chicks at
placement, shed sizes, stocking rates and maximum densities
the
attained for each genotype.
Breed Ave.
number Ave. shed
Range of Max stocking
Average stocking rate
Average stocking
Of chicks
size m2
density (kgm-2)
Chicks m-2
density (kgm-2)
under
Cobb 27235 1360
32.9-37.5
20
36.1
Ross 23368 1094
34.3-39.3
21
36.3
Average 25354
1230
32.9-39.3
21
36.2
Timing of measurements
Released
Normal farming practice was carried out on the farms in this study.
During the time a flock was in a shed, typically a proportion of the
flock was removed for slaughter (thinning) once or twice (and very
occasionally three times) before the remainder of the birds in the
flock were slaughtered (terminal slaughter). Gait scores were
measured on a sample of birds in the shed two days before a thin
or the terminal slaughter. When more than two thins were
10
11
scheduled, measurements were taken for two thins only (at the first
thin and a mid thin). Additional measures were taken on a sample
of these birds at the slaughter plant. These measurements were
contact dermatitis (foot, hock and breast), pre-slaughter handling
(back scratches), dead on arrival (DOA), and reject birds.
MEASURES
Epidemiological questionnaire
A questionnaire was used to obtain information about
1982
management and husbandry practices and policies for each
shed used in the study. This provided details of factors such as Act
number and weight of chicks placed; sex; time of year; age at
assessment and slaughter; stocking density, thinning practices;
brooding conditions; nutritional profile; vitamin and mineral
levels; litter substrates; feeder and drinker design/type; water
source; lighting programmes; age of sheds; medication policy,
Information
coccidiostat usage, vaccination programmes, bird to stock
person ratio; training /qualifications; biosecurity measures.
The questionnaire was completed by the grower (farm manager)
Official
on a standard form either on their own (if the questionnaire was
sent to them) or in conversation with a research assistant, at the
the
first visit to the farm, when the chicks were placed in the shed. On
the second visit to the farm, growers were asked a set of follow up
questions specifically related to the management of the shed used
under
in this study.
MORTALITY
Birds that died on days 1, 3, 5, 7, in the first week, and thereafter
Released
once per week (e.g. day 10 in week 2, day 17 in week 3) were kept
for post mortem (PM) examination. Birds that died in the shed
(“natural death”) were distinguished from those that were culled.
Birds that were culled were separated into two groups; leg culls
and other culls. The grower ascribed reasons for culling the birds.
Leg culls were birds that were perceived to have a walking ability
11
12
problem by the grower, whereas other culls were birds with
problems other than leg weakness.
It was agreed to adapt the criteria for cause of death from the
procedure used by one of the poultry veterinarians in his
commercial practice. The mortality descriptions were agreed by all
the veterinarians involved in the study, and a booklet containing a
list of common pathologies, detailed descriptions and colour
photographs was completed. This booklet was made available to
all personnel undertaking post-mortem analyses (and is available
1982
on request).
Act
For logistical reasons, it was decided that all dead birds would be
frozen prior to examination to allow the autopsies to be carried out
as time permitted by the veterinarians involved. However, one
veterinarian subsequently found it was not possible to follow this
policy and conducted 55% of his autopsies on fresh birds.
Information
Leg heath
Gait scoring
Official
Leg weakness was assessed using a recognised method
developed by Kestin et al. (1992), which comprises a 6 point gait
the
scoring system (scores of 0 to 5, with 5 being the poorest gait).
Techniques developed in the Bristol study were used to train
observers. Dr Andy Butterworth ran a course on gait scoring at
under
Ruakura for scorers in the present study and representatives from
the poultry industry prior to the start of the study. In addition,
similar to the Bristol study, the reliability of gait scoring ability by
the assessors was monitored by regular independent assessment
of observers scoring standard video taped sequences.
Released
Consistency of individual scoring and degree of agreement
between scorers (reliability) was undertaken prior to and twice
during the study.
On the farm visit a gait scorer and recorder would gait score 250
birds, selected at random within one shed, by reference to a pre
12
13
randomised location identifier. Birds were selected from 10
locations, in groups of 25 to 30, by corralling at each location using
a hinged catching pen. Each bird was individually encouraged to
walk out of the pen and was scored as it did so.
CONTACT DERMATITIS
Standard photographs were used to describe the scoring system
for each type of dermatitis (foot pad, hock burn, breast blisters).
Back scratches were assessed according to number present (see
1982
below). Consistency of individual scoring and degree of agreement
were obtained for each measure before and during the study for all Act
observers.
The sample of birds scored for hock burn, breast blisters and back
scratches were assessed after plucking and prior to evisceration.
Once the feet had been removed (after plucking and evisceration),
50 right feet and 50 left feet were randomly collected for scoring
Information
foot pad dermatitis.
Foot pad dermatitis
Official
The foot pad dermatitis scores were on a five point scale of 0 (no
the
foot pad dermatitis) to 4 (severe foot pad dermatitis).
Hock Burn
under
The hocks were assessed on a four point scale (0 to 3). Ranging
from 0 (no hock burn) to 3 (severe hock burn).
Breast blisters (blisters)
Released
Breast blisters were given a score of 0 (no breast blisters), 1
(presence of breast blisters). There was no account taken of the
size of the breast blister.
13
14
PRE-SLAUGHTER HANDLING
Back Scratches
Scratches longer than 2.5 cm were scored according to the
following scale: 0 ( no scratches); 1 (1 to 2 scratches); 2 (3 or more
scratches). No account was taken of age or depth of scratch. A
marker 2.5 cm in length was held up near the bird being scored to
assist observers to make their assessment of the length of the
scratch. Consistency of individual scoring and degree of
agreement were obtained for each measure before and during the
1982
study for all observers.
Act
Dead on arrival (DOA) and rejects
The number of DOA and rejects for each shed were taken from
the slaughter plant records.
Information
Rejects
Birds that were rejected as part of normal company policy were
collected and ascribed a cause of death using the definitions
Official
described previously on farm mortalities.
the
under
Released
14
15
STATISTICAL ANALYSES
The data analysed were welfare measures (response data) and
epidemiological survey information, which were put into a model to
identify significant correlations between the two. Correlations
between welfare measures were also assessed.
RESPONSE DATA
Because there were large age and live weight differences between
thins, on different farms, and these clearly affected the measures
1982
of gait score and contact dermatitis, the gait and contact dermatitis
scores were estimated on a farm basis adjusting for age
Act
differences using a linear mixed model. In this model, the thins
were split into 3 age groups with farms modelled as a random
effect. The adjusted farm effects were the best linear unbiased
predictors (BLUPs) from the model. Scorer was included in the
model. The contact dermatitis data, except for breast blisters, were
not normally distributed, so for these variables, the data were log-
Information
transformed.
Official
EPIDEMIOLOGICAL SURVEY DATA
Some of the responses to particular questions were the same for
the
most farms, or for all farms within a common type of management
practice. This precluded their use as potential explanatory
variables. The remaining variables were used in multiple
under
regression models to determine which combination showed the
best association with each welfare measure taken on the farm or
slaughter plant.
Released
THE MODEL
The model included season, stocking rate, stocking density,
genotype (breed) and management practice (which includes
factors such as nutrition, lighting schedules, and genotype). Only
those factors which were statistically significant will be presented.
15
16
BLUP analyses were used to identify the linkage of welfare
measures to management practices. With a study such as this a
large number of statistical tests have been carried out, which
increased the likelihood that one or more of these would produce a
significant result merely by chance. Results of borderline
significance should be interpreted with caution. For this reason,
only those associations (BLUPs) where p < 0.01 are presented.
The R2 value measures the proportion of the variance in the
response variable (welfare measure taken on the farm or slaughter
1982
plant) explained by the explanatory terms (responses to the
questionnaire). Because of the nature of the data and the
Act
analytical techniques are correlational we cannot infer any causal
relationship between the response and explanatory variables.
TRAINING DATA
Differences between scorers were tested using an analysis of
Information
variance for each of the response variables used in the main study.
RELIABILTY
Official
There were no significant differences between scorers for gait
the
score and they were consistent over time. The gait scorers were
on average scoring 0.15 of a score less than the reference scores
from Bristol University.
under
For the measures of dermatitis and back scratches, the individual
scorers were consistent over time and were not significantly
different from each other.
Released
Two of the veterinarians who completed 45% of the autopsies
agreed 98% of the time. The agreement between all three
veterinarians was 70%, which was deemed acceptable for the
purposes of this study.
16
17
RESULTS
We firstly describe husbandry procedures that were common
across farms, then practices that differed, then data on the welfare
measures, then present significant correlations between welfare
measures and the husbandry procedures.
COMMON HUSBANDRY PROCEDURES AND WELFARE MEASURES
A number of factors were common to the majority of the farms
studied. Most farms had smooth concrete floor (94%), the
1982
remaining 6% had bitumen or wooden floors. The ventilation
systems on all farms had air extracted not blown into the shed. Act
All sheds used pans to feed the chickens, and the lines
between the pans were emptied out and the pans were
disinfected before each new flock entered the shed. Most farms
had nipple drinkers (93%), which were fed by bore water (83%),
with chlorine added to the water (86%). A manual weighing
Information
system was used to weigh birds (94% of farms). All farms had a
formal rodent control policy. The sheds were washed and
disinfected before the placement of a new flock on all farms,
and the majority of growers applied the disinfectant with a spray
Official
(86%). Growers either replaced or laundered their work clothes
the
(86%) and replaced or cleaned and disinfected their boots
(97%) prior to the placement of a new flock. Trafficked and non-
trafficked areas outside the shed were very clean (97% sheds).
under
The cleanliness outside the shed deteriorated on some farms
by the second visit (first thin), with 72% of farms having very
clean trafficked areas, and 66% farms having very clean non-
trafficked areas. Most growers changed their boots before
entering the shed (86%) on the first visit (when the birds were
Released
placed in the shed), however by the second visit (first thin)
fewer growers changed their boots before entering the shed
(64%).
For all farms, feed was supplied to the growers from company
feed mills, and all of the feed contained antibiotics (100%) and
17
18
enzymes. For all flocks feed was provided
ad libitum to chicks
less than 7 days old. The majority of flocks were not vaccinated
at the hatchery (83%) or the farm (94%). Eight percent of the 36
study farms examined were diagnosed with a disease (such as
E.coli.) during this study. All sheds had automated lighting
systems.
Wood shavings were used in all of the sheds, there was no new
litter added before the first thin, and in a few sheds (16%) litter
was added before the second thin.
1982
At the first thin, 14% of sheds examined had system failures Act
and another 8% of sheds had system failures by the second
thin (e.g. 3 sheds had a feeder line failure, and 2 sheds had
water leaks). There were no behavioural enrichment devices
provided for the chickens in any of the sheds.
Information
As there was so little variation in these procedures across farms,
no correlation between them and welfare measures was expected,
and none was found. There were many other husbandry
procedures which varied between farms. Correlational analyses
Official
between these procedures and the welfare measures were under
the
taken and the results are described when there was a significant
effect detected.
under
WELFARE MEASURES AND CORRELATION WITH HUSBANDRY
PROCEDURES
Incidence of mortality
A total sample of 35,892 deaths was recorded by the growers for
Released
the flocks studied. Averaged over all the farms, the percentage
mortality was 3.8%, the percent of birds that died of ‘natural
causes’ was 2.7%, the percent of birds that were culled due to leg
weakness (leg culls) was 0.3% and the percent of birds that were
culled for reasons other than leg weakness (other culls) was 0.8%.
18
19
Farm mortality and Post mortem examination birds
Cause of death data are presented from 31 of the farms (there
were no data from five of the farms (4 out of 6 farms for
management type A) due to unforeseen technical reasons. The
post-mortem information comprised data for 6503 birds (18% of
the recorded farm deaths). Post-mortems were carried out on
14.8% of ‘natural’ farm deaths, 23.6% of leg culls and 27.5% of
other culls. Table 2 shows percentage of weekly farm deaths by
type of death. Overall, 71% of autopsied birds died of natural
causes, 8% were leg culls and 21% were other culls (Table 2).
1982
These figures did not vary much from week to week, but leg culls
were higher and other culls lower over the last three weeks..
Act
Table 2. Percentage of weekly farm deaths by type of death
Week
Week Week
Week
Week Week Week
Average over
1
2
3
4
5
6
7
grow out
“natural deaths”
69%
73% 74% 72% 72% 70% 74%
71%
Leg
culls
6%
7% 8%
8% 12% 13% 14%
8%
Information
Other culls
25%
20%
18% 20% 16% 16% 11%
21%
Total
100% 100% 100% 100% 100% 100% 100%
Of the birds that died over the 7 weeks of the flocks’ life (grow-out
Official
period), most occurred in the 1st week of life (35%), with only 3%
the
of all deaths occurring in the 7th week of life (Table 3). Of all the
birds that died of ‘natural’ causes 34% died in the first week.
Similarly, 24% and 42% of all leg culls and other culls,
under
respectively, died in the first week. The proportions of deaths in
each category in the seventh week were 3, 5, and 2% for natural
deaths, leg culls and other culls, respectively.
Table 3. Percentage of each mortality type (‘natural’ deaths and
Released
culls) occurring in each week of the grow-out period.
Week
Week Week
Week
Week Week Week
1
2
3
4
5
6
7 Total
“natural deaths”
34%
18% 15% 12% 11% 7% 3% 100%
Leg culls
24%
16%
13%
13% 17% 12% 5% 100%
Other culls
42%
18%
12% 12% 9% 5% 2% 100%
Average mortality
35%
18%
14% 12% 11% 7% 3% 100%
19
20
Note: Culls are birds removed by grower; average mortality is average of
‘natural deaths and culls.
1982
Act
Information
Official
the
under
Released
20
21
Table 4 shows the 10 most frequent causes of death in the
autopsies of ‘natural deaths. The highest percentage of deaths
was attributed to acute death syndrome (ADS) (17.6%). The
percentage of birds with “no diagnosis” was 11.43%. The second
highest cause of death was attributed to navel infection (9%)
(Table 4).
Table 4 Top 10 causes of death for the autopsied birds of natural
deaths
Post mortem condition
% of birds autopsied
1982
Acute Death Syndrome
17.6
No Diagnosis
11.4
Act
Navel Infection
9.0
Cull Runt
8.8
Non Starters
7.8
Colisepticaemia 7.2
Other 6.0
Infected Yolk Sac
5.8
Information
Visceral Gout
4.8
Too Decomposed
3.2
Table 5 shows the percentage of all birds (live and dead)
Official
autopsied in the study affected by each condition. The percent of
the
birds dying from the two highest ranking conditions were acute
death syndrome at 0.69% navel infection at 0.35%.
Table 5. The percentage of all birds in the study affected by each
under
condition
Autopsy condition
Number of
Overall %
autopsies
Acute death syndrome
1141
0.69
No diagnosis
743
0.45
Released
Naval infection
583
0.35
Cull runt
574
0.35
Non starters
510
0.31
Colisepticaemia 467
0.28
Other 390
0.23
Infected yolk sac
375
0.23
21
22
Gout 315
0.19
Ascites 199
0.12
Twisted leg
175
0.11
Femoral head necrosis
140
0.08
Cull no obvious
138 0.08
abnormality
Trauma 103
0.06
Joint infection
83
0.05
Others 61
0.04
Kinky back
47
0.03
1982
Dehydration 45 0.03
Pasty vent
27
0.02
Act
Pecked 40
0.02
Impacted bowel
35
0.02
Legs other
33
0.02
Infection (other)
24
0.01
Abnormalities 15 0.01
Cellulitis 9
0.01
Information
Pendulous crop
8
0.00
Hepatitis 7
0.00
Rickets 3
0.00
Septicaemia 3 0.00
Official
T.D.P. 2
0.00
Osteomyelitis 2
0.00
the
Liver absess
1
0.00
Flks 0
0.00
Enteritis
0
0.00
under
Cholangiohepatitis 0
0.00
Correlations between cause of death and husbandry
procedures
Released
Acute Death Syndrome
The correlations between cause of death and husbandry
procedures are presented for the two most common causes only
(as the rates of occurrence of the other causes were too small to
be useful in the model).
22
23
Husbandry procedures which were correlated with acute death
syndrome were: hock burn (Figure 1); the number of days the birds
were fed diet 1 (the food the birds were fed for the first few weeks
of the grow out period) (Figure 2); and use of foot baths by growers
in the annex to the sheds (Figure 3). These three variables
accounted for 41% (R2) of the variability (
p = 0.0059).
There was an increasing incidence of ADS with increasing average
hock burn score.
f
1982
50
e
ing o
om 40
Act
ndr 30
d dy
y
ie
20
th s
ops
a 10
ut
de
a
e
0
of
ut
c
a
0.0
0.1
0.2
0.3
0.4
0.5
%
Average hock burn scores
Information
Figure 1 Percentage of autopsied birds that died of acute death
syndrome versus average hock burn scores
Official
The trend was for fewer deaths ascribed to ADS with increasing
time on diet 1.
the
g
in
50
y
th
e 40
es d
ea
m
o 30
si
under
p
te d
u
ndr 20
to
y
s 10
f au
f ac
o
o
0
%
0
5
10
15
20
Number of days on diet 1
Released
Figure 2 Percentage of autopsied birds that died of acute death
syndrome versus number of days fed on diet 1.
In those farms with a foot bath in the annex, there appeared to be
less ADS compared with farms were there was no foot bath in the
annex.
23
24
50
ing
40
dy
th
s
a
e
ie
30
de
om
e
ops
ut
ndr 20
ut
c
y
a
a
s 10
of
of
%
0
No
Yes
Foot baths in annex
Figure 3 Percentage of autopsied birds that died of acute death
syndrome versus the use of foot baths in the annex of the sheds
1982
Navel infection
Act
Husbandry procedures that were correlated with navel infections
were the type of ventilation system used in the shed, and the
genotype of the bird. These two variables accounted for 64% (R2)
of the variation (
p < 0.001).
Sheds with cross flow ventilation had lower percentages of navel
infection compared with sheds with other types of ventilation
Information
systems (Figure 4).
40
ing
n
Official
dy
tio 30
s
c
ie
fe
20
the
ops
l in
e
ut
v
a
a 10
of
f n
o
%
0
Cross flow
Other
under
Ventilation system
Figure 4 Percentage of autopsied birds that died of navel infection
versus ventilation system
Released
While genotype was correlated with death rate due to navel
infection, this factor (genotype) is inextricably confounded with
other aspects of the management practices associated with
particular genotypes. Thus, navel infection death rates are
presented as a function of management practice (Figure 5).
Management practice C had a lower percentage of navel infection
compared with management practices A and B.
24
25
40
ing
n
dy
tio 30
s
c
ie
fe
20
ops
l in
e
ut
v
a
a 10
of
f n
o
%
0
A
B
C
Management practice
Figure 5 Percentage of autopsied birds that died of navel infection
1982
versus management practice
Act
Incidence of leg weakness
Of the birds culled on the farm (0.3% of all birds) ostensibly due to
a leg problem (refer to Table 2 for percentage of birds each week),
over a quarter (28%) of the birds (Table 4) could not be ascribed a
Information
“cause of death” (other than culling) using the post-mortem
techniques and criteria used in this study. Just less than a quarter
of birds (24%) culled for leg problems were ascribed as having
twisted legs, 10% joint infections, 9.5% femoral head necrosis and
Official
6.2% as having spondylolisthesis (kinky back) (Table 6).
the
Table 6 Top 10 conditions associated with leg culls on the farm
Condition %
of
leg
culls
under No Diagnosis
28.1
Varus valgus deformity (sp) and twisted
24.1
tibia
Joint Infection
10.1
Released Femoral Head Necrosis
9.5
Spondylolisthesis (Kinky Back)
6.2
Legs Other
4.3
Colisepticaemia 3.3
Trauma 2.7
25
26
Acute Death Syndrome
2.2
The factors from the epidemiological survey which were correlated
with percentage of birds culled for leg problems were genotype
and target temperature at week 3 (R2 = 48%) (
p < 0.001). As
already mentioned, genotype is confounded with other
management practices. Variation in leg culls is, thus, presented as
a function of management practice (Figure 6).
The percentage of autopsied birds that were culled due to leg
1982
weakness was highest with management practice A.
Act
e
u
d
50
lled
ss
e 40
n
s cu
ak 30
e
e
si
w 20
p
to
eg 10
l
0
f au
to
A
B
C
Information
% o
Management practice
Figure 6 Percentage of autopsied birds that were culled due to leg
weakness versus management practice
Official
the
The graph of percentage of autopsied birds that were culled due to
leg weakness (leg culls) against target temperature at week 3
(Figure 7) shows two outlier points (sheds). The birds in these two
sheds had a higher percentage of leg culls with lower than average
under
target temperature at week 3, compared with the other sheds in
the study. Further, these two sheds were the only data in this data
set for management practice A.
d
le
s
s 25
Released
ul ne ck 20
s
a
ie
e 15
g w 10
tops
u
le
o
5
a
t
of
0
%
due
0
10
20
30
Target temperature at week 2
26
27
Figure 7 Percentage of autopsied birds that were culled due to leg
weakness versus target temperature at week 3.
Gait score
Gait scorers in this study scored within half a score of each other,
and scored similarly to Bristol University standards (our scorers
were on average 0.15 of a score less than the Bristol reference
score).
The average flock gait score across all farms and all live weight
1982
groups was 2.14.
Gait score is shown by live weight category (live weight group 1, Act
1.64 to 1.99kg; live weight group 2, 2.06 kg to 2.89 kg; and live
weight group 3, 2.94 kg to 3.53 kg) in Figure 8. The average
percentage birds with gait score 0 was 0.1% for live weight group
1, and 0.0% for groups 2 and 3. Over the three live weight groups,
the majority of the birds had gait scores 1 or 2 (75.6%). The
percentage of birds with gait score 2 were 70.1%, 66.4% and
Information
62.7% for groups 1, 2 and 3 respectively (
p < 0.01).
There were also significant differences between the live weight
Official
groups (
p < 0.001) in percentage of birds with gait scores in
categories 1, 3 and 4. The lightest birds (group 1) showed a higher
the
percentage of gait score 1 (16.5%) compared with heavier birds
(group 2, 3%; group 3, 0.6%). In contrast, heavier birds, had a
higher percentage in gait score 3 (group 2, 29%; group 3, 34%)
under
compared with lighter birds (group 1, 12%) (
p < 0.001). There were
very few birds with a gait score of 4 (1.3%, averaged over the three
live weight groups). The highest percentage with gait score 4
occurred in group 3 (2.3%) (
p < 0.001). The corresponding
percentages for groups 1 and 2 were 0.4% and 1.7%, respectively.
Released
The percentages of birds with gait score 5 in each weight category
was 0.1%. There was no difference between the live weight groups
in the percentage of birds with gait score 5 (average 0.1%).
27
28
Live weight Grp 1
100
Live weight Grp 2
Live weight Grp 3
80
average
s
ird
60
f b
o
e
g
***
ta
40
rcen
e
P
20
***
1982
***
0
0
1
2
3
4
5
Act
Gait score
Figure 8 Percentage of birds by gait scores in 3 live weight groups
Note: * = p < 0.05; ** = p < 0.01; *** = p < 0.001; sed are shown for each data
point. Live weight group 1, 1.64 to 1.99kg; live weight group 2, 2.06 kg to 2.89
kg; and live weight group 3, 2.94 kg to 3.53 kg
There were no significant correlations between husbandry factors
Information
and gait score.
Official
Incidence of contact dermatitis
the
Foot pad dermatitis scores
The average foot pad dermatitis score was 0.37. On average, 71%
of birds had no foot pad dermatitis, 89.5% had scores of 0 or 1,
under
and 97.9% had scores of 0, 1, or 2. A higher proportion of the
heaviest birds (live weight group 3, 83.7%) had a score of 0
compared with the lighter birds (live weight groups 1 and 2, mean
68%) (
p < 0.05). In contrast, lower proportions of the heaviest birds
had scores of 1 (12.9%) or 2 (2.6%) compared with lighter birds
Released
(score 1: live weight group 1, 21.9%; live weight group 2, 17.9%) (
p
< 0.01); score 2: live weight group 1, 9.9 %; live weight group 2,
9.5%) (
p < 0.01). Figure 9 shows the percentage of birds across
foot pad dermatitis scores in each live weight group.
28
29
100
Live weight Grp 1
*
Live weight Grp 2
80
Live weight Grp 3
rds
Average
bi
60
of
ge
ta
40
n
e
rc
**
e
P
20
**
1982
0
0
1
2
3
4
Foot pad Score
Act
Figure 9 Percentage of birds by foot pad scores in 3 live weight
groups
Note: * = p < 0.05; ** = p<0.01; *** = p<0.001, sem’s are shown for each data
point Live weight group 1, 1.64 to 1.99kg; live weight group 2, 2.06 kg to 2.89
kg; and live weight group 3, 2.94 kg to 3.53 kg
Information
The factors from the epidemiological survey which were
correlated with foot pad dermatitis scores were; hock burn,
season, and system failures. These variables accounted for
55% (R2) (
p < 0.001) of the variation. Birds with little foot pad
Official
dermatitis also had low hock burn scores (Figure 10).
the
under
1.5
d
pa
s 1.0
re
Foot
o
ge
sc 0.5
ra
Released
evA 0.0
0.0
0.2
0.4
0.6
0.8
1.0
Average Hock burn scores
Figure 10 Foot pad dermatitis versus hock burn scores
Foot pad dermatitis was correlated with system failures in the
shed. Over the study, there were five system failures, four of which
29
30
showed high foot pad dermatitis scores. Of these four system
failures, three were feed line faults and one was a water line leak.
Birds in one shed with a system failure giving a large water leak
had a low incidence of foot pad dermatitis (Figure 11)
1.6
ore 1.4
c
1.2
d s 1.0
pa 0.8
1982
0.6
Foot 0.4
ge
ra 0.2
Act
e
v 0.0
A
No
Yes
System failure
Figure 11 Foot pad dermatitis scores as a function of incidence of
system failure in the shed
Information
Season was associated with foot pad dermatitis score, with
scores in winter/spring being higher than summer/autumn
(Figure 12).
Official
e 2.0
the
or
c
1.5
d s
pa 1.0 under
Foot
0.5
ge
ra
e
v
A 0.0
Summer/Autumn
Winter/Spring
season
Released
Figure 12 Foot pad dermatitis scores versus season
Hock Burn
Average hock burn scores were less than 0.5 on a scale from 1
(mild) to 3 (severe). On average, 72.4% of birds had no hock burn,
30
31
97.3% had a score of 0 or 1, and 99.9% had a score of 0, 1 or 2.
There was no difference in hock burn between the three live weight
groups for none (0), mild (1) or severe (3) hock burn scores. There
was a higher proportion of birds with hock burn score 2 in the
medium live weight group (group 2, 2.06 kg to 2.89 kg) compared
with lightest birds (group 1, 1.64 to 1.99kg) (
p < 0.001). Figure 13
shows the hock burn scores for birds within each of the three live
weight groups.
1982
100
Live weight Grp 1
Live weight Grp 2
80
Act
s
Live weight Grp 3
ird
Average
f b
60
o
e
g
ta
40
rcen
e
P
20
**
Information
0
0
1
2
3
Hock Burn Score
Official
Figure 13 Percentage of birds by hock burn scores in 3 live weight
groups
the
Note: * = p < 0.05; ** = p<0.01; *** = p<0.001, sem’s are shown for each data
point Live weight group 1, 1.64 to 1.99kg; live weight group 2, 2.06 kg to 2.89
kg; and live weight group 3, 2.94 kg to 3.53 kg
under
The factors from the epidemiological survey which were
correlated with incidence of hock burn were: foot pad
dermatitis (see Figure 10); brooding method in the shed; the
subjective description of air quality and chicks m-2. These
four variables accounted for 57% (R2) of the variation in the
Released
model (
p < 0.001).
Average Hock burn scores were higher in sheds were there was
half shed brooding, compared with two thirds, three quarters or full
shed brooding (Figure 14).
31
32
1.0
rn
u
0.8
b
ck
0.6
o
res
H
e
0.4
sco
ag
0.2
er
v
A
0.0
1/2 shed
Whole
Brooding
shed
Figure 14 Hock burn scores versus brooding method used in the
1982
shed
Act
Farms with “very good” or “average” air quality was associated with
low average hock burn scores, whereas “good air quality” was
associated with relatively higher average hock burn scores (Figure
15)
Information
n
1.0
0.8
bur
k
s 0.6
oc
re
H
o 0.4
ge
sc
Official
0.2
ra
e
v
0.0
A
the
0
1
2
3
4
5
Air quality
Figure 15 Average hock burn scores versus air quality
under
Note: 1 = smelly; 2 = dusty; 3 = average; 4 = good; 5 = very good
Hock burn scores appear to increase with increasing number of
chick per metre square (Figure 16)
Released
32
33
1.0
rn
u
0.8
b
ck
0.6
o
res
H
0.4
e
sco
ag
0.2
er
v
0.0
A
0
5
10
15
20
25
chick m-2
Figure 16 Hock burn scores versus chicks per m2
1982
Breast blisters
Act
On average 99% of birds had no breast blisters (score = 0) (Figure
17). There were no differences between the live weight groups.
Live weight Grp 1
100
Live weight Grp 2
Information
Live weight Grp 3
80
s
Average
ird
f b
60
o
e
g
ta
Official
40
rcen
e
P
20
the
0
no blister
blister
under
Breast blister Score
Figure 17 Percentage of birds by breast blister scores in 3 live
weight groups
Note: * = p < 0.05; ** = p<0.01; *** = p<0.001, sem’s are shown for each data
point Live weight group 1, 1.64 to 1.99kg; live weight group 2, 2.06 kg to 2.89
Released
kg; and live weight group 3, 2.94 kg to 3.53 kg
Factors in the epidemiological survey were not associated with the
percent of birds with breast blisters.
33
34
Incidence of back scratches, DOAs (birds dead on arrival)
and rejected birds.
Back scratches
The average back scratch score was 0.64. On average, 59.4% of
birds had no back scratches, 91.2% had a score of 0 or 1. The
majority of birds with no back scratches were in the heaviest live
weight group (75%), while less than half of the lightest birds had no
back scratches (44.2%) (
p < 0.001). A higher proportion of the
lightest birds had scores of 1 (41%) and 2 (14 %) compared with
1982
live weight group 2 (score 1, 34% and score 2, 6.9%) and live
weight group 3 (score 1 22%; score 2 2.5%) (
p < 0.001). Figure18 Act
shows the back scratch scores for birds within each of the three
live weight groups.
Live weight Grp 1
100
Information
Live weight Grp 2
Live weight Grp 3
***
80
s
Average
rd
60
Official
of bi
ge
***
ta
40
n
e
the
rc
e
P
20
***
0
under 0
1
2
Back Scratch Score
Figure 18 Percentage of birds by back scratch scores in 3 live
weight groups
Note: * = p < 0.05; ** = p<0.01; *** = p<0.001, sem’s are shown for each data
Released
point. Live weight group 1, 1.64 to 1.99kg; live weight group 2, 2.06 kg to 2.89
kg; and live weight group 3, 2.94 kg to 3.53 kg
Factors in the epidemiological survey which were correlated with
back scratches were: season; management practices; and the
material used to construct the shed. These three variables
accounted for 38% (R2) of the variation in the model (
p = 0.0075).
34
35
Management practice A had higher back scratch scores than B
and C (Figure 19).
1.0
tch
ra
sc
s
ck
re
a
o 0.5
e b
sc
g
ra
1982
ve
0.0
A
A
B
C
Management practice
Act
Figure 19 Average back scratch score across management
practices
Average back scratch scores were higher in winter/spring
compared with summer/Autumn season (range from 0.43 to 0.79)
Information
compared with the winter/spring season (range from 0.54 to 0.78)
(Figure 20. All female sheds had a lower average back scratch
score compared with as hatched (mixed gender) or all male
sheds).
Official
1
the
tch
ra
0.8
sc
0.6
ck
res
a
b
0.4
under
e
sco
ag
0.2
er
v
0
A
Summer/
Winter/
Season
Autumn
Spring
Figure 20 Average back scratch score across season
Released
35
36
Back scratch scores were higher in sheds with wooden framing
(Figure 21).
s 1.0
e
or
c 0.8
h s
tc
ra 0.6
c
s
k
c 0.4
ba
0.2
ge
ra
e
v 0.0
A
1982
steel
wood
Material used to construct shed
Act
Figure 21 Average back scratch score by construction material of
shed
Birds dead on arrival (DOA) and rejected birds (rejects)
Table 7 shows the percentage of birds dead on arrival (DOA) or
rejected at the slaughter plant for each live weight group. The
Information
average proportion of birds dead on arrival (DOA) was 0.15% with
the heaviest birds having the highest proportion (0.20%) (
p <
0.001). The average number of rejects was 0.22%. There was no
Official
difference in the proportion of reject birds between weight
categories. The average number of DOAs and rejects combined
the
was 0.37%, with no difference in proportions between the weight
categories.
under
Table 7. The percentage of birds dead on arrival (DOA) or rejected
at the slaughter plant by live weight group
DOA &
Live weight
DOA%
SEM
Reject%
SEM
Reject% SEM
Group 1
0.12
0.03
0.22
0.04
0.34
0.06
Released
Group 2
0.16
0.02
0.23
0.06
0.39
0.06
Group 3
0.20
0.04
0.18
0.03
0.38
0.04
Average 0.15
0.02
0.22
0.03 0.37 0.04
Note: SEMs are shown for each data point. Live weight group 1, 1.64 to
1.99kg; live weight group 2, 2.06 kg to 2.89 kg; and live weight group 3, 2.94
kg to 3.53 kg
36
37
Factors in the epidemiological survey were not associated with the
percent of birds dead on arrival at the slaughter plant.
The factors in the epidemiological survey which were correlated
with proportion of birds rejected were; hock burn, expected age of
slaughter and litter depth at placement (R2 = 42.6,
p = 0.0037).
Rejection rates increased at higher average hock burn scores
(Figure 22).
1982
d
0.10
te
c
je
Act
s
f re
rd 0.05
t o
bi
n
e
rc
e
0.00
P
0.0
0.1
0.2
0.3
0.4
Average hock burn scores
Information
Figure 22 Percentage of rejected birds versus average hock burn
scores
Official
Percentages of rejected birds increased with increases in expected
age of slaughter (Figure 23).
the
0.3
ed
ct
e
ejunder
s 0.2
f r
rd
t o
bi 0.1
en
c
er
0.0
P
25
35
45
55
Expected age at slaughter
Released
Figure 23 Percentage of rejected birds versus expected age of
slaughter
Percentages of rejected birds increased with litter depth at
placement (Figure 24).
37
38
0.3
jected
e
0.2
f r
rds
t o
bi 0.1
rcen
e
0.0
P
0
2
4
6
8
10
Litter depth at placement
Figure 24 Percentage of rejected birds versus litter depth at
1982
placement
Act
Information
Official
the
under
Released
38
39
DISCUSSION
This study has used the best measures available to objectively
assess the management and welfare of broiler chickens in New
Zealand, and establishes a baseline for welfare standards in the
industry.
Some of the husbandry procedures are common across the
industry, include: type of flooring (concrete); and litter type (wood
1982
shavings); ventilation system (extraction); use of pans for feeding
and nipple drinkers; rodent control; hygiene policies; types of and Act
frequency of use antibiotics; and enzymes added to feed; type of
feed available to chicks less than 7 days old; and absence of
specific behavioural enrichment strategies.
Variations in the welfare measures recorded in the present study
were correlated with each other and with variations in the
Information
procedures identified in the epidemiological survey. It is important
to remember that some of the significant correlations may have
occurred by chance, and if a reasonable (logical) explanation
Official
cannot be found for the correlations, it is not sensible to attribute
them any significance. That said, these results remain useful in
the
identifying potential causes of variation in the welfare status of
broilers.
under
Overall mortality was 3.8%, which is at the lower end of levels
typically seen in the industry overseas, and a little higher than the
best that has been reported (Butterworth, 2004). Of all the birds in
the study, 0.3% died as leg culls. The contribution of leg culls to
total mortality was relatively low (8%) and about half that reported
Released
for birds in two recent UK studies (15% in UK) (Hall, 2001,
Dawkins et al. 2004). Further, the proportion of birds with severe
leg weakness (gait score 4 and 5) in our NZ research was about
one quarter that reported for broilers in the UK, Denmark and
Sweden (1.4% versus 5 to 6.5%; Kestin et al. 1992; Sanotra et al.
2003). Moreover, these studies are the only published ones that
have used the same gait scoring procedure as used in our
39
40
research. Considering both the leg culling and gait scoring
information together, our results demonstrate that leg health in
New Zealand broilers is better than for birds in Europe.
In the UK studies mentioned above, there appears to be a higher
proportion of “other culls” (35% UK versus 21% NZ) and a lower
proportion of “natural deaths” (50% UK versus 71% NZ) relative to
total mortality compared with the New Zealand industry. This
suggests that there are differences between NZ and the UK not
only in leg health, but also in the conditions leading to culling for
1982
non-leg related disorders and/or natural deaths. The post-mortem
data can potentially provide insights into likely differences in non-Act
leg related conditions between NZ and elsewhere but,
unfortunately, there are no other studies with appropriate autopsy
data to make a detailed comparison. Further, we should exercise
caution when interpreting the New Zealand industry post mortem
data for the natural deaths, as the proportion of birds autopsied (for
natural deaths, 14%) was much lower than the proportions
Information
autopsied for leg and other culls, about 25%). Further, there were
a relatively high proportion of post mortems accorded “no
diagnosis” for birds in all categories (mortality, leg cull and other
Official
cull). The reason for this may have been due to the process of
freezing the birds (which had to occur for logistical reasons)
the
altering some of the internal structures, making it more difficult for
an accurate diagnosis of death. In addition, histological and other
examinations of tissue were not undertaken due to financial
under
constraints of the project.
Generally speaking, the range of conditions seen in the autopsies
of the natural deaths and other culls in our study is similar to that
seen elsewhere (Butterworth, 2004; Julian, 2004), and the
Released
prevalence of ascites (0.12%) was similar to that quoted elsewhere
(0.15%, Butterworth, 2004), whilst kinky back (0.03% NZ versus
0.3 % elsewhere), acute death syndrome (0.69% NZ versus >2%
elsewhere), pendulous crop (0.00% NZ versus 0.1 % elsewhere)
were lower in our study. Other diseases such as enteritis and
cellulitis were rare or absent in our post-mortem samples.
40
41
Of the birds autopsied, acute death syndrome (ADS) (also known
as sudden death syndrome) was the highest ranking cause of
death (17.6%). Bagshaw and Matthews (2001) reported that
incidence of ADS (or sudden death syndrome) may be effected by
fast growth rate, nutrition and other environmental factors. In the
current study, there was a lower incidence of ADS on farms with a
foot bath in the annex. On those farms without a foot bath in the
annex, the growers changed boots; therefore it is unlikely that
hygiene was a factor in incidence of ADS. Indeed, the presence of
a foot bath cannot be assumed to reduce ADS from this study. The
1982
significance of the correlation between ADS and hock burn score is
not clear, particularly since factors usually associated with hock
Act
burn (e.g. density and season (Imaeda, 2000) were not correlated
with ADS. The lower rate of ADS with increased time on diet 1
could be an area for further fruitful research.
Of the birds autopsied, 9% died from navel infection. The
percentage of birds with navel infection was correlated with
Information
genotype and ventilation system in the shed. There were higher
percentages of birds which were autopsied with navel infections
under management practice types A and B compared with type C.
Official
The difference between management practices is likely due to
factors which are highly confounded such as nutrition, lighting
the
schedules, stockmanship, and genotype. Further exploration of
the differences in management practices between the three
systems might yield useful information on ways to reduce the
under
incidence of navel infection. Navel infection was generally lower in
cross-flow ventilation sheds compared with other types of
ventilation in sheds (e.g. roof, side and tunnel). However, there
were insufficient numbers of any other types of ventilation system
to make a direct comparison between different systems.
Released
The high rate of non diagnosis of pathologies for the leg culls
(28%) implies that we should exercise caution when interpreting
these data. Of the leg culls autopsied, the most common pathology
identified was deformed limbs (Varus valgus and twisted tibia)
(24%), with the next three most common conditions being
infections (joint (10%) and femoral head necrosis (9.5%)) and
41
42
kinky back (6%). These pathologies are typical of those seen
elsewhere, but the relative frequency of the infectious causes
appears lower in the New Zealand birds than elsewhere (e.g.
femoral head necrosis rates in the UK of 17 to 38%; (Butterworth,
1999; McNamee and Smyth, 2000). Conversely, deformed limbs
occurred relatively more often in the New Zealand industry than in
the UK (11 to 14 % Butterworth, 1999; McNamee and Smyth,
2000). In absolute terms, this latter observation does not imply that
deformed legs are a greater welfare problem for New Zealand
birds as the culling rate for leg disorders in New Zealand is only
1982
about half or less than half that seen elsewhere (e.g. Dawkins,
2004; Hall, 2001; McNamee and Smyth, 2000).
Act
The percentage of autopsied birds that were culled for leg
weakness was correlated with type of management practice
(higher in type A) and target temperature at week 3 (occurring in
type A sheds). In this data set, management practice A was
represented by two sheds only, therefore clear conclusions cannot
Information
be drawn from these results. However, further investigations may
help to understand if there is any causal link between leg
weakness and target temperature at week 3, or other specific
Official
management practices.
the
The gait scoring procedure used was highly reliable and consistent
over time. On average, there were relatively few birds with a
normal or near normal gait (8.1%, gait scores 0 and 1 combined),
under
with two-thirds having a gait score of 2, 23% with a gait score of 3,
and 1.4% with gait scores of 4 or5). There was a clear effect of live
weight on gait score, with proportionately fewer of the heaviest
(also older) birds in categories 1 and 2, and proportionately more
with gait scores 3 and 4. Kestin et al. (2001) has also reported
Released
poorer gait scores with heavier birds, and demonstrated that the
effect is due to weight rather than age.
A number of UK researchers (Danbury et al., 2000; Hocking et al.,
1999; McGeown et al., 1999;) have attempted to assess the
welfare impacts of different gait scores by measuring time budgets
and determining if the degree of pain experienced varies with gait
42
43
score (see Introduction). While it would be surprising if lame birds
didn’t experience some pain (at least for some pathologies and
particularly at high gait scores), the interpretation of these data is
not unequivocal
In addition, a single gait score can be associated with a range of
different pathologies (Mench, 2004) and there may be varying
levels of pain associated with different pathologies. This has also
not been examined scientifically. Others have attempted to assess
the welfare implications of high gait score by looking at walking
1982
speed and time budgets of animals with different gait scores (e.g.
Weeks et al, 2000; Bokkers and Koene, 2004). High gait score
Act
birds walk more slowly through obstacle courses, rest for longer
periods and have a reduced visitation rate to feeders and drinkers.
These results suggest impaired walking ability but do not resolve
the issue of the presence or degree of pain experienced. We do
not know if pathologies causing a particular gait score in New
Zealand broilers are the same as those causing altered gait in
Information
birds in other countries. Thus, it is difficult to know if, or what,
proportion of say gait score 3 birds in New Zealand are
experiencing pain (as seems likely for a proportion of gait score 3
Official
birds in the UK). It would be instructive to conduct studies in New
Zealand of gait score before and after pain relief to determine the
the
relationship between gait score and pain in broilers under New
Zealand conditions. Further, more detailed studies of the
pathology leading to gait abnormalities in New Zealand broilers
under
would provide additional insights into the likely welfare impacts
caused by those pathologies.
Because both leg culling rates and the occurrence of gait scores 4
and 5 (about 1%) were relatively low in our study, it seems that
Released
severe leg weakness is less of a problem in the New Zealand
industry than elsewhere. As birds with gait scores 4 and 5 are
likely to experience poor to very poor welfare (including pain) it is
desirable to cull birds in these categories as soon as possible,
thereby reducing their incidence in the flock to as close to zero as
is practically possible. Research and practical management
procedures to achieve this goal are a priority.
43
44
While research conducted overseas has implicated a large variety
of management factors in the aetiology of leg weakness (e.g.
stocking density, lighting cycles and intensity, ambient temperature
and humidity, ammonia levels, litter quality and floor type; (Reiter
and Bessei, 1998) no significant correlations between such factors
and gait score were seen in our study. Possible reasons for the
absence of correlations in the present study include: these factors
do not influence leg weakness under New Zealand conditions;
and/or there was too little variation in the relevant
1982
management/husbandry procedures.
Act
Contact dermatitis is thought to be caused by a combination of
moisture and chemical irritants in the litter of broiler houses
(Ekstrand et al., 1998). Thus, the incidence of contact dermatitis is
an indicator of the condition of litter substrate, and this in turn
provides a measure of the suitability of the environmental
Information
conditions in the houses. Berg (2004) reported that a number of
risk factors for contact dermatitis in temperate climates have been
identified. These are: litter material; litter type; floor type; stocking
Official
density; drinker system and feed composition.
the
The majority (71%) of birds in this study had no foot pad dermatitis,
which is similar to the levels reported in the literature (e.g.
Dawkins, 2004), although it is difficult to compare directly the
under
results from our study with others as the measurement protocols
differed. Previously used protocols have been lacking in precision
and not often tested for reliability and consistency in their
application. In our study, an average of 89.5% of birds had scores
of 0 or 1 (none or mild). These results imply that litter conditions
Released
are well maintained in NZ, although the incidence was slightly
higher in winter/spring months when humidity is higher. During the
winter/spring season, four of the sheds had a system failure and
these birds had relatively high foot pad dermatitis. In contrast,
during the summer/autumn season, one shed had a system failure
these birds had relatively low foot pad dermatitis. Thus, although
44
45
the seasonal data were confounded by system failures, the higher
incidence during wetter times of the year it is consistent with other
studies showing that the relative humidity inside and outside the
house has an influence on litter quality, with the highest prevalence
of foot pad dermatitis occurring in the autumn/winter months in
Europe (Berg, 2004).
The incidences of foot pad dermatitis and hock burn were
correlated, as has been observed by others (Greene et al., 1985).
A higher proportion of the heavier (older) birds had no foot pad
1982
dermatitis, suggesting that this condition improves over the grow-Act
out period. David Marks (pers comm.) suggests that this could
occur as a result of an earlier bacterial infection healing over the
life of the bird when litter quality is good.
The majority of birds (72%) in this study had no hock burn, which is
similar to the levels reported in the literature (Menzies et al., 1998),
Information
although it is difficult to compare directly the results from our study
with others as the measurement protocols differed. Previously
used protocols have been lacking in precision and not often tested
Official
for reliability and consistency in their application. An average of
97.3% of birds had no or mild hock burn. These data indicate that
the
litter conditions are well maintained in NZ. There was no difference
in the percentage of birds with hock burn across the live weight
groups for none, mild and severe scores. However, birds of
under
medium weight had higher hock burn scores compared with the
lightest or heaviest birds in the study. The reason for this is
unclear.
Average hock burn scores were correlated with foot pad dermatitis,
Released
the type of brooding in the shed, air quality, and chick m-2 at
placement. Average hock burn scores tended to increase with
average foot pad scores and this relationship was discussed
earlier. Average hock burn scores are higher in sheds where there
is half shed brooding, compared with two thirds, three quarters or
full shed brooding. Average hock burn scores varied little across
45
46
the subjective scores of air quality. However there appeared to be
slightly lower average hock burn scores when the air quality was
deemed very good (score 5). There was a trend toward increased
hock burn with increased density of chicks at placement but not
density at the time of assessment. There are no obvious
explanations for these correlations.
The incidence of breast blisters was extremely low (0.5%), is
consistent with that reported internationally (Berg, 2004), and
implies that litter conditions were well maintained.
1982
Act
The incidence of contact dermatitis (foot pad dermatitis, hock burn
and breast blisters) is similar to that reported internationally and
implies that New Zealand broiler sheds have good environmental
conditions. In this study, there were no correlations found between
litter depth or type, floor type, or stocking density, drinker system
and welfare measures. The litter material used in most sheds in
Information
our study was wood shavings, and this has been found to result in
lower incidence of foot pad dermatitis compared with straw
(Ekstrand et al., 1997). Nipple drinkers with cups were used in the
Official
sheds in our study, and this type of drinker system has been
shown to reduce the risk of wet litter and the risk of contact
the
dermatitis (Berg, 2004).
Back scratches, birds dead on arrival (DOA) and reject birds were
under
used in this study as indicators of catching and other damage
during the pre-slaughter period, although the activities and factors
prior to this period may influence and, therefore, confound these
measures.
Released
The frequency of occurrence of back scratches is only infrequently
reported in the literature, and the incidence of scratches and
associated conditions varies widely from 0.3 to 60% (Hall
2001;Feddes et al., 2002; Frankenhuis et al., 1991). The average
incidence of back scratches in the present study was 41%, being
much higher than the lowest overseas estimates. It is not clear why
46
47
there would be such a difference, although it is possible that the
low rates mentioned elsewhere are underestimates. This could
occur if the measurement systems used were not reliable. For all
the data reported by others,(e.g. Hall, 2001), the definitions of
scratches have been imprecise and there have been, no
assessments of the reliability or consistency of the scoring systems
used. Thus, it is difficult to make meaningful comparisons between
our data and the incidence of scratches reported elsewhere. Back
scratching during the grow out period is an important welfare issue
as it can lead to scabby hip syndrome and other lesions of the skin
1982
(Proudfoot and Hulan, 1985). The welfare impacts of scratches
that occur during catching and pre-slaughter handling are not
Act
known, but would likely cause some pain or discomfort.
In our study, a higher percentage of lighter birds had back
scratches compared with the heaviest birds. There are many
factors which could contribute to this difference including stocking
Information
density (heavier birds were at a lower density due to prior thinning
of a portion of the flock), feather cover (older birds have better
feather cover), behaviour (older are less flighty), and gender of the
bird (males are more likely to scratch and have later development
Official
of feathers). The significant correlations between back scratch
incidence and management practices and season could be
the
attributable to behavioural differences in these particular
environments, but additional research would be required to
ascertain any causal relationships. Average back scratch scores
under
were correlated with the type of construction material of the shed; it
is likely that this correlation occurred by chance.
The average percent of birds dead on arrival was 0.15% and the
average percent of bird rejected at the slaughter plant was 0.22%,
Released
both values being at the lower end of the internationally reported
levels (Ekstrand, 1998), suggesting that the pre-slaughter handling
of birds in New Zealand is well conducted.
Rejected birds at the slaughter plant were correlated with average
hock burn scores, expected age of slaughter and litter depth when
47
48
the chicks were placed in the shed. As average hock burn score
increased there was an increase in percentage of rejected birds.
However, it is important to note that the variation in hock burn
scores was very small, and there were very small increases in the
percentage of rejected birds (less than 0.1%). There was a trend
towards increasing percentage of rejected birds with increasing
expected age of slaughter, but the meaning of this is not clear as
there was no relationship between live weight and reject rate. Litter
depth varied from 3 to 12.5cm, but explanations for an effect of
litter depth at placement on reject rates is not apparent.
1982
Act
Apart from a relatively high incidence of back scratches, the
measures taken indicate good standards of pre-slaughter handling
of broilers are achieved in New Zealand.
Stocking density and animal welfare has often been a contentious
issue in the literature. Bagshaw and Matthews ( 2001) reported in
Information
a literature review that stocking density per se does not appear to
be a major direct cause of most welfare problems, traditionally
associated with high stocking density rates. From the published
Official
information it appears that other environmental conditions are
more directly implicated. In this study, stocking density varied
the
between 32.9 and 39.3 kg m-2. Stocking density was not correlated
with any of the welfare measures, or variables in the
epidemiological survey, suggesting that broiler welfare is either
under
unaffected by stocking density or that the range in densities seen
in the current study was too small to enable correlations to be
detectable. Stocking density (at the levels currently used in this
study) would appear appropriate as the welfare of the birds is
good.
Released
48
49
CONCLUSIONS AND FURTHER RESEARCH
1. General flock health is good as demonstrated by the relatively
low total mortality rate and low incidence of infectious
pathologies in birds culled for leg disorders.
2. There is a very low rate of culling for leg disorders in New
Zealand broilers and substantially fewer birds with severe leg
weakness (gait score 4 and 5). Most birds had gait scores of 1
or 2. Considering both the leg culling and gait scoring
information together, our results demonstrate that leg health in
1982
New Zealand broilers is better than for birds in Europe. Given
the high level of international concern with leg weakness in
Act
broilers, particularly in relation to the degree of pain
experienced, it would be pertinent to undertake further
research on the welfare impacts at different gait scores and of
the common pathologies. The frequencies of various
pathologies associated with leg weakness appears different for
our broilers, implying that, in order to make further research on
Information
questions such as pain relevant to our conditions, it should be
undertaken in New Zealand.
3. Prevalence of all forms of contact dermatitis (footpad, hock
Official
burn and breast blisters) appears within the range reported
internationally. Incidences, such as occurred, were mostly of a
the
relatively minor nature. These data indicate that the litter in the
houses is maintained in good condition, with small seasonal
variation
under
4. Pre-slaughter handling procedures as assessed by Dead On
Arrivals and reject birds were generally better than reported for
other countries. The proportion of birds with back scratches
averaged 39%. Meaningful comparisons between our data and
the incidence of scratches reported elsewhere were not
Released
possible due to the lack details of the measurement criteria
provided in other published reports. The welfare impacts of
back scratches under New Zealand conditions, and procedures
to reduce their incidence, would merit further study.
5. In this study, stocking density varied between 32.9 and 39.3 kg
m-2. Stocking density was not correlated with any of the welfare
49
50
measures, or variables in the epidemiological survey,
suggesting that broiler welfare is either unaffected by stocking
density, or that the range in densities seen in the current study
was too small to enable correlations to be detectable.
Controlled studies using a wider range of stocking densities
would need to be undertaken to identify any effects of stocking
density on welfare. Stocking density (at the levels currently
used in this study) would appear appropriate as the welfare of
the birds is good.
6. Correlations undertaken between the welfare measures and
1982
rather simple features of the husbandry system or
environmental conditions revealed some significant and
Act
interesting results. Some of these appeared logical and
reasonable (e.g. correlation between foot pad dermatitis and
season) and provide valuable insights into possible ways to
improve welfare through alteration of husbandry procedures.
Variations in more complex factors such as management type
Information
(A, B, C), which includes genotype, were sometimes
associated significantly with variation in welfare measures. For
example, management practice type C had a higher incidence
of ADS and a lower incidence of navel infection compared with
Official
the other types of management practice. This, again, suggests
that there would be considerable opportunity to improve broiler
the
welfare through identifying cause and effect relationships, such
as they may exist, that underlie such correlations. It is
recognised that this is not a trivial task due to the large number
under
of highly confounded (and commercially sensitive) variables
that make up, for example, a particular type of management
practice.
7. One of the often-expressed concerns with intensive indoor
broiler growing conditions is the lack of behavioural activity
Released
(both in the range of activities and time spent engaged in them)
of broilers. The current study did not attempt to assess this
aspect of broiler welfare, but we did record in the survey that
there are no procedures in place to encourage a wider
behavioural repertoire of birds in sheds.
50
51
8. Apart from the expected correlation between two of the
measures of contact dermatitis (footpad dermatitis and hock
burn), there were no significant correlations between the
various welfare measures. This suggests that each measure
was assessing independent welfare traits and, thus, all should
be used in studies where a comprehensive description of
broiler welfare status is required (e.g. in audits).
9. This study has demonstrated that the welfare of broilers in New
Zealand is equal to or superior to that of broilers kept
elsewhere.
1982
Act
Information
Official
the
under
Released
51
52
REFERENCES
Bagshaw, C. S., and L. R. Matthews. 2001. Broiler welfare - a
review of latest research and projects in progress internationally,
AgResearch, Hamilton, New Zealand.
Berg, C. 2004. Pododermatitits and hock burn in broiler chickens.
In: C. A. Weeks and A. Butterworth (eds.) Measuring and auditing
broiler welfare. p 37-50. CABI publishing, UK.
Berg, C. et al. 2000. The welfare of chickens kept for meat
productions (broilers). SANCO.B.3/AH/R15/2000, Scientific
Committee on Animal Health and Animal Welfare.
1982
Berg, C. C. 1998. Foot-pad dermatitis in broilers and turkeys -
Act
prevalence, risk factors and prevention. Acta Universitatis
Agriculturae Sueciae Veterinaria No. 36: 43 pp.
Bokkers, E. A. M., and P. Koene. 2004. Motivation and ability to
walk for a food reward in fast and slow growing Broilers to 12
weeks of age. Behavioural Processes 67: 121-130.
Butterworth, A. 1999. Infectious components of broiler lameness: A
Information
review. Worlds Poultry Science Journal 55: 327-352.
Butterworth, A. 2002. Photographs for hock burn. In: C. S. M.
Bagshaw, L R Matthews (ed.), Hamilton, New Zealand.
Butterworth, A. 2004. Infectious disease: Morbidity and mortality.
Official
In: C. A. Weeks and A. Butterworth (eds.) Measuring and auditing
broiler welfare. p 61-70. CABI Publishing, Wallingford.
the
Danbury, T. C., C. A. Weeks, J. P. Chambers, A. E. Waterman-
Pearson, and S. C. Kestin. 2000. Self-selection fo the analgesic
drug carprofen by lame broiler chickens. The Veterinary Record
under
146: 307-311.
Dawkins, M. S., C. A. Donnelly, and T. A. Jones. 2004. Chicken
welfare is influenced more by housing conditions than by stocking
density. Nature 427: 342 - 344.
Ekstrand, C. 1998. An observational cohort study of the effects of
Released
catching method on carcase rejection rates in broilers. Animal
Welfare 7: 87-96.
Ekstrand, C., B. Algers, and J. Svedberg. 1997. Rearing conditions
and foot-pad dermatitis in swedish broiler chickens. Preventive
Veterinary Medicine 31: 167-174.
52
53
Ekstrand, C., T. E. Carpenter, I. Andersson, and B. Algers. 1998.
Prevalence and control of foot-pad dermatitis in broilers in Sweden.
British Poultry Science 39: 318-324.
Feddes, J. J., Emmanuel, E.R. and Zuidhoft, M.J. 2002. Broiler
performance, body weight variance, feed and water intake, and
carcass quality at different stocking densities. Poultry Science
81:774-779.
Frankenhuis, M. T., M. H. Vertommen, and H. Hemminga. 1991.
Influence of claw clipping, stocking density and feeding space on
the incidence of scabby hips in broilers. British poultry science 32:
1982
227-230.
Greene, J. A., R. M. McCracken, and R. T. Evans. 1985. A contact Act
dermatitis of broilers - clinical and pathological findings. Avian
Pathology 14: 23-38.
Hall, A. L. 2001. The effect of stocking density on the welfare and
behaviour of broiler chickens reared commercially. Animal Welfare
10: 23-40.
Hocking, P. M., R. Bernard, and M. H. Maxwell. 1999. Assessment
Information
of pain during locomotion and the welfare of adult male turkeys
with destructive cartilage loss of the hip joint. British Poultry
Science 40: 30-34.
Official
Imaeda, N. 2000. Influence of the stocking density and rearing
season on incidence of sudden death syndrome in broiler chickens.
the
Poultry Science 79: 201-204.
Kestin, S. C., T. G. Knowles, A. E. Tinch, and N. G. Gregory. 1992.
Prevalence of leg weakness in broiler chickens and its relationship
under
with genotype. Veterinary Record 131: 190-194.
McGeown, D., T. C. Danbury, A. E. Waterman Pearson, and S. C.
Kestin. 1999. Effect of carprofen on lameness in broiler chickens.
Veterinary Record 144: 668-671.
McNamee, P. T., and J. A. Smyth. 2000. Bacterial chondronecrosis
Released
with osteomyelitis ('femoral head necrosis') of broiler chickens: A
review. Avian Pathology 29: 253-270.
Mench, J. 2004. Lameness. In: C. A. Weeks and A. Butterworth
(eds.) Measuring and auditing broiler welfare. p 3-18. Cromwell
Press., Trowbridge.
53
54
Menzies, F. D., E. A. Goodall, D. A. McConaghy, and M. J. Alcorn.
1998. An update on the epidemiology of contact dermatitis in
commercial broilers. Avian Pathology 27: 174-180.
Mitchell, M. A., and P. J. Kettlewell. 2004. Transport and handling.
In: C. A. Weeks and A. Butterworth (eds.) Measuring and auditing
broiler welfare. p 145-160. CABI publishing, UK.
Proudfoot, F.G. and Hulan, H.W. 1985. Effects of stocking density
on the incidence of scabby hip syndrome among broiler chickens.
Poultry Science 64:2001-2003.
Reiter, K., and W. Bessei. 1998. Possibilities of reducing leg
1982
disorders in broilers and turkeys (review). Moglichkeiten zur
Verringerung von Beinschaden bei Broilern und Puten (Ubersicht) Act
62: 145-149.
Sanotra, G. S., C. Berg, and J. Damkjer Lund. 2003. A comparison
between leg problems in Danish and Swedish broiler production.
Animal Welfare 12: 677-683.
Weeks, C. A., T. D. Danbury, H. C. Davies, P. Hunt, and S. C.
Kestin. 2000. The behaviour of broiler chickens and its modification
Information
by lameness. Applied Animal Behaviour Science 67: 111-125.
Weeks, C. A., and S. C. Kestin. 1997. The effect of leg weakness
on the behaviour of broilers. In: Proceedings of the 5th European
Official
Symposium on Poultry Welfare, Wageningen, Netherlands. p 117-
118.
the
Acknowledgements
We would like to thank the Broiler Welfare Research Committee for
under
their lively discussions and support for this study. Thanks to the
scorers in the field, in particular Diane Sutton, Frankie Hubbard,
Katie Carnie. A special thanks to Dr Andy Butterworth, for his
discussions and running the gait scoring workshop.
Released
54