Preslaughter handling factors affecting dead on arrival, condemnations, and bruising in broiler chickens raised without an antibiotic program

Abstract

In Thailand, knowledge about the factors affecting broiler losses during the preslaughter process is very limited, especially for broilers raised without an antibiotic program. The objective of this study was to determine the preslaughter factors that influence the incidence of dead on arrival (DOA), condemnations, and bruising in broilers raised without antibiotics. Data from 13,581 truckloads of broilers raised without an antibiotic program in 95 contract farms of one of Thailand's largest broiler producers in 2021 were analyzed using a generalized linear mixed model that accounted for farm as a random effect. Results showed that the following risk factors were associated with the occurrence of DOA, condemnations, and bruising: season, time of transport, sex, age at slaughter, mortality and culling rate, and weight per crate. While mean body weight affected the incidence of condemnations and bruising, transport time and lairage time affected DOA and bruising. Feed withdrawal time affected DOA and condemnations. Rearing stocking density only affected condemnation rate. Reducing or eliminating the effects of these risk factors could reduce production losses due to DOA, condemnations, and bruising, thereby improving animal welfare and producer profitability. Reducing weight per crate could reduce DOA, condemnations, and bruising. Reducing lairage time could reduce DOA and bruising, while reducing feed withdrawal time could reduce DOA and condemnations. Raising broilers at a younger age with a lower slaughter weight could prevent the occurrence of DOA, condemnations, and bruising.

INTRODUCTION

In poultry production, antibiotics are used not only for therapeutic purposes but also to promote growth at subtherapeutic doses (Diaz-Sanchez et al., 2015). Antibiotics provide effective treatment and control of infectious diseases. However, the excessive and incorrect use of antibiotics in animal feed has led to a rapid increase in antibiotic resistance (Ma et al., 2021). Antimicrobial resistance is defined by the World Health Organization as “an increase in the minimum inhibitory concentration of a compound for a previously sensitive strain” (World Health Organization, 2013). Antibiotic resistance impacts the effective prevention and treatment of poultry illnesses and increasingly poses a threat to global public health (Yang et al., 2019). In animal production, antibiotics are used in subtherapeutic doses for long periods of time, which creates optimal conditions for bacteria to incorporate resistance genes. These genes can be transferred to human-adapted pathogens or the gut microbiota, especially through contaminated food, humans, or the environment. They also provide optimal conditions for the propagation of genes that may have evolved in humans or in environment. The fact that antibiotics used in human and veterinary medicine are largely composed of similar compounds could lead to the spread of resistance between animals and humans (Robinson et al., 2016).

Because antibiotics are widely used in food-producing animals, they can serve as reservoirs for antibiotic-resistant bacteria that can be transmitted to humans (Marshall and Levy, 2011). In recent years, the emergence of antibiotic resistance around the world has led to concerns about the use of antibiotics in animals (Diaz-Sanchez et al., 2015; Singer et al., 2019). Concerns were not only for resistant zoonotic bacteria and veterinary pathogens, but also for residues that could be present in the animal-derived foods (Diaz-Sanchez et al., 2015). In response, the World Organization for the Humane Treatment of Animals has created a plan to reduce the development of drug resistance by researching tools and approaches for animal welfare management, and poultry producers in many countries have incorporated the principles into practice design (Iannetti et al., 2021). There are 3 rules for antibiotic-free animal husbandry 1) antibiotic residue-free: antibiotics were administered to these chickens but removed from the feed a few days before slaughter to ensure that no residues remained in the meat during processing, 2) raising without antibiotics (RWA): no antibiotics were used in the rearing of these chickens, and 3) organic rearing: this poultry must be raised without antibiotics, fed 100% organic feed, and have access to the outdoors (USDA, 2013; Diaz-Sanchez et al., 2015; Karavolias et al., 2018; Singer et al., 2019).

As Thailand is the fourth largest exporter in the world (Department of Trade Negotiations, 2023), Thai broiler producers must comply with exporters’ rules and regulations to meet importers’ requirements. Raising chickens without an antibiotic program while adhering to animal welfare practices is required, especially for export to Europe (Bracke et al., 2019). Thus, the Thailand's broiler producers that would like to export especially to Europe need to produce broilers under the RWA program which broilers are not allowed to receive antibiotics through feed, water, or injections. However, chemical coccidiostats can be used to prevent coccidia infections and also some alternatives to antibiotics such as essential oils and probiotics can be used in broiler production (Fancher et al., 2020). Animals with diseases requiring antibiotic treatment will be removed from the RWA program and need to find alternative distribution. Thai broiler producers produce broilers without antibiotics and in compliance with animal welfare, consideration should be given to the various risk factors during rearing, preprocessing, transportation to the slaughterhouse, and the slaughter process that affect the economic losses of broiler products. Several studies mentioned the 3 major criteria related to animal welfare and economic loss during preslaughter handling in broiler production were dead on arrival (DOA), condemnations, and bruising (Nijdam et al., 2004; Jacobs et al., 2017a; Buzdugan et al., 2020; Junghans et al., 2022; Torma et al., 2022).

Dead on arrival refers to broilers that died between the time of catching and the time of slaughter (Nijdam et al, 2004). Several studies have examined DOA % as a key indicator of broiler welfare in the preslaughter period. On-farm risk factors for DOA % include flock size and vaccination schedule. Preslaughter risk factors include catching method, season, weather conditions, temperature, duration of transport, and lairage time. The mean value of DOA % varied considerably across studies, with values ranging from 0.14% to 0.46% (Nijdam et al., 2004; Jacobs et al., 2017a; Buzdugan et al., 2020; Junghans et al., 2022). This wide range may be due to differences in the population studied, rearing system, transport conditions, and preslaughter handling practices.

All broiler carcasses found unfit for human consumption due to macroscopic lesions during meat inspection in a poultry slaughterhouse are referred to as “broiler condemnation” (Lupo et al., 2009). Several studies have reported various broiler carcass conditions requiring whole carcass condemnation: ascites, abnormal carcass color, perihepatitis, cellulitis, hard breast, tumors, and “dead on arrival” (Buzdugan et al., 2020; Junghans et al., 2022; Torma et al., 2022). Although the conditions that lead to condemnation are multifactorial, identifying the factors that influence condemnation could be useful in developing strategies to reduce overall condemnation. Condemnation rates in broilers depend on several factors, including health status, diseases, weather conditions, weight and age of broilers, rearing stocking density, and preslaughter handling, for example, time of transport, feed withdrawal time, and lairage time (Lupo et al., 2009; Burdugan et al., 2020; Junghans et al., 2022).

The term “bruising in broilers” refers to damage to the skin and underlying tissues of broilers caused by physical trauma (Northcutt et al., 2000). Bruising can affect the quality and appearance of meat and lead to economic losses in the poultry industry (Cockram et al., 2020). Postmortem examination of injuries is a method to evaluate the welfare of poultry during rearing and preslaughter handling (Gouveia et al., 2009). Identifying the factors that influence the occurrence of certain injuries appears to be essential for improving animal welfare. Risk factors that influence the percentage of bruising include mean body weight at rearing, age at slaughter, sex, catching team and catching method, season, stocking density in crates, and time of transport (Mayes 1980; Nijdam et al., 2004; Yalcin et al., 2004; Nijdam et al., 2005; Jacobs et al., 2017b). Bruising not only indicates poor animal welfare—it is a direct result of trauma—but also devalues the carcass and should be avoided.

Regarding the limited information on risk factors for production losses due to DOA, condemnations, and bruising that affect economic losses in broiler production, especially for broilers raised without an antibiotic program, analysis of preslaughter handling data to identify risk factors for these losses in broiler production needs to be investigated, and the results could help producers develop strategies and husbandry improvements to reduce economic losses. Therefore, the objective of this study was to identify the preslaughter risk factors affecting DOA, condemnations, and bruising of broilers raised without an antibiotic program and slaughtered in 2021 by one of the largest Thai broiler producers. The results of this study can serve as a guide to reduce economic losses for broiler producers, especially those who raise broilers without an antibiotic program.

MATERIALS AND METHODS

Study Population, Data Collection, and Variable Description

This study used only data collected after the animals were slaughtered and no procedures or interventions were performed on live animals, ethical review and approval were not required for the animal study.

Data were collected from one of Thailand's largest broiler producers in the form of 13,581 truckloads from 95 contract farms. The chickens were raised in evaporative cooling houses without antibiotics (RWA), received chemical coccidiostats and alternatives to antibiotics such as essential oils and probiotics, and were slaughtered in 2021. According to the animal welfare practice and legal standards of the Poultry Meat and Poultry Products Inspection Regulations B.E. 2548 (2005) of the Department of Livestock Development of Thailand and the exporter's guidelines (Department of Livestock Development, 2005), not only DOA but also condemnation and bruising of broilers during the postmortem process were recorded. In this study, there were 3 dependent variables: DOA, condemnation, and bruising, which were calculated as percentage (%). The total number of DOA, condemnation, and bruising of each truckload was counted at the processing plant as the numerator, and the total number of broilers transported per load was the denominator, which was then multiplied by 100. The causes of carcass condemnation were purulent abscesses, cellulitis, viscera unfit for human consumption, cachexia and emaciated carcasses, abnormal color carcasses, and carcasses without offal before meat inspection. Bruising was discoloration of the skin or under the skin due to blood admixtures larger than 1/3 breast area and leg and larger than 2 cm² for wings. Bruises on breast, wings, and/or legs were counted as 1 per carcass.

In the present study, there were 11 independent variables or factors: 4 categorical variables and 7 continuous variables. The categorical variables included sex (male, female, and mixed sex), age at slaughter (40–43 d, 44–47 d, 48–51 d, 52–55 d, and 56–71 d), seasons (winter [November–February], summer [March–May], and rainy [June–October]), time of transport or time of day the vehicle left the farm after loading (night = 18:00–04:00, morning = 04:00–08:00, day = 8:00–18:00). Continuous variables were mean body weight (g), rearing stocking density (kg/m²), percentage of mortality and culling (%), weight per crate (kg), feed withdrawal time (min), transport duration (min), and lairage time (min). Mortality and culling rate (%) was obtained from farm data at processing and calculated as ([total number of dead and cull broilers * 100]/total number of chicks placed).

Handling and Transport Procedures

The following procedures were used in handling and transporting broilers. First, the feed and water in the rearing houses were removed before loading to slaughterhouse. The broilers were then caught inside the house using a manual catching system by conveying the transport crates into the houses, and a catcher picked up 2 broilers with whole body catching. The trunks of 2 broilers were caught simultaneously with the bodies standing upright and supported laterally under the breast with the thumbs on the back so that the catcher could hold the wings, and loaded into the transport crate, which was approximately 0.72 m long × 0.54 m wide × 0.31 m high. The stocking density of the crates was based on the weight of the birds and the environmental conditions. The number of birds per crate ranged from 4 to 10 birds, depending on the weight of the broilers, which weighed not more than 18 kg per crate. Then the transport crates were taken outside and stacked on trucks. All 495 crates were stacked on top of each other in a truck. While loading the crates, each stack was sprayed with water until every chicken feather was wet. After loading, the trucks were driven to the slaughterhouse. During the day, especially in the summer season, the trucks were sprayed with water every 150 km on the way from the farm to the slaughterhouse. Once at the slaughterhouse, the truckloads were taken to a holding area away from the main slaughterhouse. Here, the ante-mortem inspections were performed. The broilers were kept in crates on the trucks in the same position as during transport while they were ventilated with fans and evaporative cooling. After remaining in the holding area for 30 to 120 min, the trucks were moved from the holding area to the unloading area, the crates were unloaded from the truck, and the broilers were manually removed from the crates and placed on a shackle line. During the ante-mortem inspection, the number of dead broilers was recorded as the percentage of dead on arrival per truckload by the quality control inspectors. Broilers were then transported to slaughter sites. The stunning method was reversible electrical stunning at 800 HZ, 50 to 100 volts in a water bath for 10 s without rhythmic breathing. Then, the broilers were slaughtered by cutting the carotid artery and jugular vein with a sharp knife according to the halal method. After scalding, defeathering, and eviscerating, the total number of condemnations and bruises was checked and recorded by meat hygiene inspectors under the guidance of veterinarians.

Slaughterhouse Record and Data Handling

The slaughterhouse submitted data in the form of digital spreadsheets and forms that included handwritten records of flock reports, driver reports, and holding area reports for each truckload of broilers transported from each farm to the slaughterhouse from January to December 2021. Duration was calculated for each stage of preslaughter handling during transportation of broilers to slaughter in this study as follows: Transport duration was the time from the end of loading to arrival at the slaughterhouse; lairage time was the time from entering the holding area to leaving the holding area; and feed withdrawal time was the time from feed removal on the farm to leaving the holding area. Each load of broilers transported from each farm to the slaughterhouse was uniquely identified by date, farm number, house number, truck number, and number of broilers loaded. These data were compiled to provide detailed information on each stage of the transport of the broilers to the slaughterhouse. A series of checks were made to ensure that the data had been entered correctly. Some measurements, such as the duration of the different stages and stocking density, were calculated from the raw data. Seasons were defined using winter, summer, and rainy season data for the year 2021. When there was doubt about the correct interpretation of the original source material for descriptive or analytical statistics, these data were not included. The actual time of slaughter was not recorded in the submitted data, so for some time calculations the time when the broilers left the holding area was used as the end time.

Statistical Analysis

Descriptive statistics of original data were obtained. The general linear mixed model was used to analyze the data (Agresti, 2015; Finch et al., 2019). The statistical model was as follows.

$${\mathrm{y}}_{\text{ijt}}={\beta }_0+\sum _{\mathrm{k}=1}^{\mathrm{n}}{\beta }_{\mathrm{k}}{\mathrm{x}}_{\text{ij}}+{\mathrm{u}}_{\mathrm{j}}+{\epsilon }_{\text{ijt}}$$ (1)

where ${\mathrm{y}}_{\text{ij}}$ was the $\mathrm{t}$ observation (e.g., % DOA) of truckload $\mathrm{i}$ from farm $\mathrm{j}$, and ${\beta }_0$ was the intercept. The term ${\beta }_{\mathrm{k}}$ represented the fixed effects and n denoted the total number of fixed effects. The term ${\mathrm{x}}_{\text{ij}}$ represented the variable included in the model, while ${\mathrm{u}}_{\mathrm{j}}$ was the random effect on the intercept for farm $\mathrm{j}$. It was assumed that

${\mathrm{u}}_{\mathrm{j}}\sim \text{NID}\left(0,{\sigma }_{\text{farm}}^2\right)$ and ${\epsilon }_{\text{ijt}}\sim \text{NID}\left(0,{\sigma }^2\right)$.

In this study, the variables defined as fixed effects consisted of 11 variables. Model building was performed in 2 steps that included univariable and multivariable analysis. Variables with P ≤ 0.2 from the univariable analysis were included in the multivariable analysis. Model selection was performed using the backward method. The multicollinearity problem between variables was assessed during model selection by testing the variance inflation factor (Goldstein, 1993). Multicollinearity was determined when the variance inflation factor for a variable was greater than 4 (Mansfield and Helms, 1982). In the case of multicollinearity, the variable with the higher biological plausibility was retained in the model (Singhla et al., 2017; Sansamur et al., 2020).

Originally, the models were developed using original data including % DOA, % condemnation, and % bruising. However, the residuals ${\epsilon }_{\text{ijt}}$ did not meet the normality assumption. Therefore, these original data were transformed with a logarithm similar to previous studies (Agresti, 2015). In addition, models were fitted with transformed data, including ln(% DOA), ln(% condemnation), and ln(% bruising). Thus, in Equation (1), the term ${\mathrm{y}}_{\text{ijt}}$ was replaced by $\text{ln}\left({\mathrm{y}}_{\text{ijt}}\right)$ and the statistical model was updated to the following equation.

$$\text{ln}\left({\mathrm{y}}_{\text{ijt}}\right)={\beta }_0+\sum _{\mathrm{k}=1}^{\mathrm{n}}{\beta }_{\mathrm{k}}{\mathrm{x}}_{\text{ij}}+{\mathrm{u}}_{\mathrm{j}}+{\epsilon }_{\text{ijt}}$$ (2)

For each final model, the error term, or residual ${\epsilon }_{\text{ijt}}$ was tested for model assumptions such as homogeneity of variance and normality. Homogeneity of variance was assessed using the plot of the residuals versus the fitted values, while normality was assessed using the normal Q–Q plot of the residuals.

Because the model included farms as random effects and observations (e.g., % DOA) were clustered by farm, the intraclass correlation coefficient (ICC) (Finch et al., 2019) was determined for each model. The calculated ICC indicated the proportion of the total variance of the observations that was attributable to clustering. A high ICC indicated that the observations were highly dependent on the farms to which they belonged. A low ICC, on the other hand, indicated that the observations were not overly dependent on the farms to which they belonged.

In addition, models were built based on the dependent variable ($\mathrm{y}$), transformed into a logarithmic scale. In this approach, the interpretation of the slope β should be taken into account. The estimated coefficient for the dependent variable was β, such that a one-unit increase in any predictor was associated with a 100*(e^β-1) percent change in % DOA, % condemnation, or % bruising.

The statistical software R and the package “lme4” were used for the general linear mixed model. Backward model selection was performed using the functions of the package “lmerTest,” whereas the functions of the package “emmeans” were used to determine the differences in the means of observation between groups for categorical variables.

RESULTS

Descriptive Statistic Results of Studied Variables

The dataset used in this study was from 13,581 truckloads of broiler chickens raised without an antibiotic program on 95 contract farms of one of Thailand's largest broiler producers and slaughtered in 2021. Summary statistics for categorical and continuous variables, original data on the percentage of DOA, condemnation, and bruising are presented in Tables 1 and 2. Due to the COVID-19 outbreak, there was a shortage of labor to slaughter the broilers, which resulted in the need to extend the rearing period and therefore some of the older 52- to 71-day-old broilers were slaughtered. The overall mean of DOA for all truckloads was 0.20%, with minimum and maximum values of 0.03% and 1.02%, respectively. The overall mean of condemnation for all truckloads was 0.89%, with minimum and maximum values of 0.06% and 6.09%, respectively. The overall mean of bruising for all truckloads was 0.74%, with minimum and maximum values of 0.26% and 1.88%, respectively. There were 6 individual causes of condemnations in the current study, as shown in Table 3. The 3 most common causes of condemnations were viscera unfit for human consumption (65.17%), purulent abscesses (12.36%), and cachexia and emaciated carcasses (11.23%). The distribution of DOA, condemnation, and bruising percentages per truckload are shown in Figure 1. The relationship between original and ln-transformed %DOA, ln-transformed % condemnation, and ln-transformed % bruising with the predictor variables were presented in the supplement.

Table 1.

Mean percentage of dead on arrival (DOA), condemnation, and bruising within each category variable of 13,581 truckloads (n, %) of broilers raised by one of Thailand's largest broiler producers without an antibiotic program and slaughtered in 2021.

Variable	Truckload		% DOA	% Condemnation	% Bruising
	n	%
Season
Winter	5,057	37.24	0.210	0.977	0.705
Summer	3,241	23.86	0.190	0.681	0.714
Rainy	5,283	38.9	0.202	0.925	0.779
Time of transport1
Night	6,398	47.11	0.192	0.886	0.72
Morning	1,043	7.68	0.168	0.842	0.716
Day	6,140	45.21	0.218	0.893	0.743
Sex
Male	2,048	15.08	0.238	0.873	0.720
Female	1,998	14.71	0.169	0.709	0.716
Mixed sex	9,535	70.21	0.201	0.926	0.743
Age
40–43 d	5,962	43.91	0.214	0.844	0.724
44–47 d	5,505	40.53	0.195	0.776	0.715
48–51 d	200	1.47	0.240	0.920	0.791
52–55 d	49	0.36	0.258	0.692	0.823
56–71 d	1,865	13.73	0.251	1.346	0.828

¹Time of transport (night = 18:00–04:00, morning = 04:00–08:00, day = 8:00–18:00).

Table 2.

Descriptive statistics of continuous variables and percentage of dead on arrival (DOA), condemnation, and bruising of 13,581 truckloads of broilers raised by one of Thailand's largest broiler producers without an antibiotic program and slaughtered in 2021.

Variable	Mean	SD	Min	Max	%CV
Continuous variables
Mean body weight (g)	2,936.80	305.07	2,230	4,562	10.39
Rearing stocking density (kg/m2)	28.95	2.96	21.43	41.23	10.21
Mortality and culling (%)	4.96	2.42	1.40	11.68	48.42
Weight per crate (kg)	17.94	2.11	12.31	30.45	11.75
Transport duration (min)	125.04	78.12	37	379	62.48
Lairage time (min)	119.30	54.54	20	289	45.72
Feed withdrawal time (min)	528.11	82.71	302	792	15.66
DOA (%)	0.20	0.16	0.03	1.02	80.20
Condemnation (%)	0.89	0.93	0.06	6.09	104.69
Bruising (%)	0.74	0.28	0.26	1.88	38.23

Table 3.

Percentages of condemnation causes.

Causes of condemnation	% of total slaughter	% of causes of condemnation
Purulent abscesses	0.11	12.36
Cellulitis	0.02	2.25
Viscera not fit for human consumption	0.58	65.17
Cachexia and emaciated carcasses	0.10	11.23
Abnormal color carcasses	0.04	4.49
Carcass without offal	0.04	4.49
Total condemnation	0.89	100

Figure 1.

The distribution for DOA, condemnation, and bruising percentages per truckload.

Factors Affecting DOA Percentage

In the final model for the percentage of DOA, 9 independent variables or factors were associated with the ln-transformed percentage of DOA (Table 4). Compared with the rainy season, the percentage of DOA increased in the winter and decreased in the summer (β = 0.149 and −0.154, respectively). When broilers were transported in the morning and at night, the percentage of DOA was lower than when they were transported during the day (β = −0.235 and −0.178, respectively). The percentage of male broilers DOA significantly increased compared to mixed sexes (β = 0.199), while the percentage of female broilers DOA significantly decreased (β = −0.119). The percentages of DOA were significantly lower at 40 to 43 d and 44 to 47 d compared to 56 to 71 d (β = −0.181 and −0.152, respectively). For continuous variables, the percentage of DOA increased significantly with increasing mortality and culling rate (β = 0.0512), weight per crate (β = 0.0515), transport duration (β = 0.0011), lairage time (β = 0.0009), and feed withdrawal time (β = 0.0008). In addition, the statistical model was built based on the dependent variable (y) transformed into a logarithmic scale. In this approach, the interpretation of the slope β should be taken into account. Accordingly, the coefficient β for the dependent variable was estimated, such that a one unit increase in any predictor was associated with a 100*(e^β-1) percent change in % DOA. For example, each additional increase in mortality and culling rate was associated with an increase of 100*(e^β-1) = 100*(e^0.0512-1) = 5.25% in % DOA.

Table 4.

Factor associated with the percentage of dead on arrival (DOA) from 13,581 truckloads of broilers raised by one of Thailand's largest broiler producers without an antibiotic program and slaughtered in 2021.

Variable	Estimates1	SE	Statistic	P-value	95% confidence interval
Season
Winter	0.149	0.019	7.942	<0.001	0.1120	0.1860
Summer	−0.154	0.019	−8.037	<0.001	−0.1160	−0.1910
Rainy	Ref
Time of transport2
Night	−0.178	0.013	−14.102	<0.001	−0.2030	−0.1530
Morning	−0.235	0.024	−9.940	<0.001	−0.2810	−0.1880
Day	Ref
Sex
Male	0.199	0.020	9.814	<0.001	0.1590	0.2380
Female	−0.119	0.020	−5.976	<0.001	−0.1580	−0.0799
Mixed sex	Ref
Age
40–43 d	−0.181	0.030	−6.115	<0.001	−0.2390	−0.1230
44–47 d	−0.152	0.029	−5.280	<0.001	−0.2090	−0.0959
48–51 d	0.085	0.056	1.514	0.130	−0.0252	0.1960
52–55 d	−0.066	0.106	−0.627	0.531	−0.2740	0.1410
56–71 d	Ref
Mortality and culling (%)	0.0512	0.0033	15.566	<0.001	0.0448	0.0577
Weight per crate (kg)	0.0515	0.0033	14.446	<0.001	0.0445	0.0585
Transport duration (min)	0.0011	0.0002	5.304	<0.001	0.0007	0.0011
Lairage time (min)	0.0009	0.0001	7.693	<0.001	0.0006	0.0009
Feed withdrawal time (min)	0.0008	0.0001	10.011	<0.001	0.0006	0.0014

Intraclass correlation coefficient (ICC) = 0.08.

¹Estimates (β) indicate how likely an outcome is to occur in one context relative to another or one unit increase or decrease in the predictor to a β value increase or decrease in the outcome.

²Time of transport (night = 18:00–04:00, morning = 04:00–08:00, day = 8:00–8:00).

Factor Affecting Condemnation Percentage

There were 9 risk factors associated with the ln-transformed condemnation percentage (Table 5). Compared with the rainy season, the condemnation rate was statistically significantly higher in the winter (β = 0.386) and lower in the summer (β = −0.054). When broilers were transported in the morning, the condemnation rate was lower than during the day (β = −0.078). The condemnation rate was higher for male broilers than for mixed sex broilers, but statistically significantly lower for female broilers (β = 0.200 and −0.139 for male and female broilers, respectively). The percentage of condemnation decreased statistically significantly when broilers were younger (β = −0.446, −0.443, −0.783, and −0.815 for 52–55 d, 48–51 d, 44–47 d, and 40–43 d, respectively). For continuous variables, the percentage of condemnation decreased significantly with the increase in mean body weight (β = −0.0005). While the percentage of condemnation significantly increased with increasing rearing stock density (β = 0.0186), mortality and culling rate (β = 0.0617), weight per crate (β = 0.0277), and feed withdrawal time (β = 0.0014).

Table 5.

Factor associated with the percentage of condemnations of 13,581 truckloads of broilers raised by one of Thailand's largest broiler producers without an antibiotic program and slaughtered in 2021.

Variable	Estimates1	SE	Statistic	P-value	95% confidence interval
Season
Winter	0.386	0.022	17.994	<0.001	0.3440	0.4280
Summer	−0.054	0.022	−2.478	0.013	−0.0114	−0.0973
Rainy	Ref
Time of transport2
Night	−0.020	0.014	−1.413	0.158	−0.0479	−0.0077
Morning	−0.078	0.027	2.934	0.003	−0.1300	−0.0250
Day	Ref
Sex
Male	0.200	0.023	8.673	<0.001	0.1540	0.2460
Female	−0.139	0.023	−6.076	<0.001	−0.1840	−0.0944
Mixed sex	Ref
Age
40–43 d	−0.815	0.039	−20.667	<0.001	−0.8920	−0.7380
44–47 d	−0.783	0.037	−21.226	<0.001	−0.8550	−0.7110
48–51 d	−0.443	0.064	−6.894	<0.001	−0.5680	−0.3170
52–55 d	−0.446	0.119	−3.737	<0.001	−0.6800	−0.2120
56–71 d	Ref
Mean body weight (g)	−0.0005	0.0001	−11.071	<0.001	−0.0006	−0.0004
Rearing stocking density (kg /m2)	0.0186	0.0042	4.400	<0.001	0.0103	0.0269
Mortality and culling (%)	0.0617	0.0039	15.831	<0.001	0.0540	0.0693
Weight per crate (kg)	0.0277	0.0051	5.418	<0.001	0.0177	0.03777
Feed withdrawal time (min)	0.0014	0.0000	16.136	<0.001	0.0012	0.0015

Intraclass correlation coefficient (ICC) = 0.09.

¹Estimates (β) indicate how likely an outcome is to occur in one context relative to another or one unit increase or decrease in the predictor to a β value increase or decrease in the outcome.

²Time of transport (night = 18:00–04:00, morning = 04:00–08:00, day = 8:00–8:00).

Factors Affecting Bruising Percentage

There were 9 risk factors associated with the ln-transformed percentage of bruising (Table 6). The percentage of bruising was statistically lower in winter (β = −0.067) and summer (β = −0.055) than in the rainy season. The percentage of bruising was lower in morning (β = −0.032) and night (β = −0.031) transports than in daytime transports. The percentage of bruising in male (β = −0.043) and female (β = −0.026) birds decreased significantly compared to mixed sexes. When comparing the ages of 40 to 43 d and 44 to 47 d with 56 to 71 d, the percentage of bruising significantly decreased (β = −0.07 and −0.074, respectively). For continuous variables, the percentage of bruising significantly increase with the increase in mean body weight (β = 0.0002), transport duration (β = 0.0003), and lairage time (β = 0.0002). While the percentage of bruising significantly decreased with the increased in mortality and culling rate (β = −0.0053) and weight per crate (β = −0.0194).

Table 6.

Factor associated with the percentage of bruising of 13,581 truckloads of broilers by one of Thailand's largest broiler producers raised without an antibiotic program and slaughtered in 2021.

Variable	Estimates1	SE	Statistic	P value	95% confidence interval
Season
Winter	−0.067	0.010	−6.546	<0.001	−0.0871	−0.0470
Summer	−0.055	0.011	−5.202	<0.001	−0.0754	−0.0341
Rainy	Ref
Time of transport2
Night	−0.031	0.007	−4.498	<0.001	−0.0449	−0.0176
Morning	−0.032	0.013	−2.429	0.015	−0.0572	−0.0061
Day	Ref
Sex
Male	−0.043	0.011	−3.811	<0.001	−0.0643	−0.0206
Female	−0.026	0.011	−2.399	0.016	−0.0480	−0.0048
Mixed sex	Ref
Age
40–43 d	−0.070	0.018	−3.832	<0.001	−0.1060	−0.0342
44–47 d	−0.074	0.017	−4.278	<0.001	−0.1080	−0.0402
48–51 d	−0.027	0.031	−0.865	0.387	−0.0882	0.0342
52–55 d	−0.019	0.058	−0.321	0.748	−0.1320	0.0948
56–71 d	Ref
Mean body weight (g)	0.0002	0.0000	7.964	<0.001	0.0001	0.0002
Mortality and culling (%)	−0.0053	0.0018	−2.956	0.003	−0.0088	−0.0017
Weight per crate (kg)	−0.0194	0.0025	−7.919	<0.001	−0.0242	−0.0146
Transport duration (min)	0.0003	0.0001	3.024	0.002	0.00009	0.00042
Lairage time (min)	0.0002	0.0001	3.665	<0.001	0.0001	0.0003

Intraclass correlation coefficient (ICC) = 0.03.

¹Estimates (β) indicate how likely an outcome is to occur in one context relative to another or one unit increase or decrease in the predictor to a β value increase or decrease in the outcome.

²Time of transport (night = 18:00–04:00, morning = 04:00–08:00, day = 8:00–8:00).

DISCUSSION

In the present study, the overall mean DOA, condemnation, and bruising in broilers raised without antibiotics by a large broiler producer in Thailand in 2021 were 0.20%, 0.89%, and 0.74%, respectively. The overall mean of DOA in this study was within the range of other studies that reported DOA percentages between 0.14% and 0.46 % (Nijdam et al., 2004; Jacobs et al., 2017a; Buzdugan et al., 2020; Junghans et al., 2022). The condemnation rate in the current study was lower than the result reported by Junghans et al., 2022 (1.48%). The difference could be due to the different causes of condemnation as in this study DOA did not include in condemnation, while Junghans et al., 2022 did. Nijdam et al. (2004) reported 2.2% bruising which was higher than in this study. The difference could be due to the number of bruises counted as they counted each bruise on a single leg and wing, which may contribute to the total number of bruises in a flock.

The Thai broiler producer who provided data for this study, has not used antibiotics in broiler production since 2018. The percentages of DOA and condemnation were reported as 0.26% and 2.01%, respectively, in the year prior to the cessation of antibiotic use in broiler production (Sunfood, 2017). The DOA percentage was comparable to the DOA rate of broilers raised without antibiotics in this study. While the condemnation rate in the RWA production system was lower than the previous system raised with antibiotics. A possible explanation for this could be that producers change farm management when raising broilers without antibiotics, for example, lower rearing stock density, use of alternative antibiotics such as essential oils and probiotics, and improved preslaughter handling (e.g., reduced weight per crate), and lower slaughter age associated with lower slaughter weight. In agreement with Buzdugan et al. (2020) who reported that weight at slaughter and number of broilers per crate were associated with more condemnation outcomes. The main cause of condemnation in this study was that the offal was unsuitable for human consumption. One possible explanation could be the husbandry practices and hot and humid climate in Thailand, which increased the susceptibility of broiler chickens to respiratory diseases and then secondary infections caused by Escherichia coli (Fancher et al., 2020). Purulent abscesses were the second main cause of condemnation in this study. According to Lorenzoni, (2022), purulent abscesses were mainly caused by staphylococci. Cachexia and emaciated carcasses, the third major cause of condemnation in the current study was probably the result of poor flock management, which was also related to poor early management, poor chick quality, and ambient temperature. More attention should be paid to health management, water and feed intake, stocking density, and stunting syndrome during the first week of life (Nery et al., 2017).

Factor Influencing DOA, Condemnation, and Bruising

Season

In this study, the percentage of DOA, condemnation, and bruising in broiler rearing differed in different seasons. DOA and condemnation were higher in the winter but lower in the summer than in the rainy season. From a health point of view, the main cause of condemnation in the processing plant was bacterial infection due to E. coli (Lutful Kabir, 2010; Wijesurendra et al., 2017; Muchon et al., 2019). The broilers in the present study were raised in broiler houses with evaporative cooling. During the winter season, farmers turned off some fans at night to regulate the temperature in the barns, which resulted in a higher accumulation of ammonia gas that caused respiratory irritation and infection. After an initial viral infection, the broilers became sick, but their mortality rate was low unless they were later infected with another pathogen, especially E. coli, which caused them to become unhealthy and die during catching and transport, leading to an increase in the percentage DOA (Aliabad et al., 1998; Cockram and Dulal, 2018). The main cause of condemnation in this study was viscera unfit for human consumption. Husbandry practices and climate could increase the susceptibility of broilers to respiratory diseases, especially secondary infections caused by E. coli (Fancher et al., 2020). In the present study, a lower condemnation rate was observed in the summer, which could be due to the fact that E. coli infections decreased in the summer because temperature and humidity were stable during this season and had less influence on health, resulting in a decrease in the condemnation rate.

The broilers in this study had a lower percentage of bruising in the winter and summer seasons than in the rainy season. This was because ambient temperature had a significant effect on the broilers’ susceptibility to bruising, hemoglobin breakdown, and the rate of healing of injured areas (Hamdy et al., 1961). In the rainy season, the temperature dropped suddenly after the rain, and due to the temperature fluctuations, the sudden change to lower temperatures resulted in a slight decrease in susceptibility to bruising, which was probably due to vasoconstriction, which requires several hours to a day for the skin to adapt to the new environment, but the healing rate was also lower in the rainy season, which may be the reason why bruising is more common in the rainy season (Hamdy et al., 1961).

Time of Transport

In the present study, it was found that the percentage of DOA was lower when broilers were transported at night and in the morning than when they were transported during the day. The hypothesis for this relationship could be the lower ambient temperature at night and in the morning compared to daytime, which is in agreement with the study of Nijdam et al. (2004). Lower heat stress and tonic movements in broilers indicated less anxious behavior during transport of broilers in the dark and at lower temperatures (Nijdam et al., 2004). Broilers transported during the day were exposed to elevated temperatures and heat stress caused by the heat and sunrise during the day. The birds could not survive the trip because they were unable to adapt, and previous infections in broilers may have been a risk factor. The major bacterial pathogen, E. coli, can cause respiratory infections and ultimately death, leading to high condemnation rates in broilers after slaughter (Junghans et al., 2022). The observation that the percentage of DOA was higher during the day may be explained by the fact that broilers were exposed to higher temperatures at this time, resulting in heat stress, heart problems, congestion, etc. In addition, most broilers caught during the dark hours and lower temperatures were less active, resulting in fewer physical injuries during catching and transport and a lower percentage of bruising (Hamdy et al., 1961; Nijdam et al., 2004).

Sex

The sex of broilers influenced the percentage of DOA, condemnation, and bruising. The DOA rate was higher when only male broilers were raised in the flock than when both sexes were. In contrast, when only female broilers were raised, the DOA and condemnation rates were lower than when both sexes were raised in the flock. When only male or female broilers were raised in the flock, the rate of bruising was lower than when mixed-sex broilers were raised. In this study, these results could be due to the lower body weight of female broilers compared to mixed sex and male broilers, so they should have better viability and fewer health problems, that is, a lower risk factor for bacterial infections resulting in a lower percentage of condemnations, while higher body weight male broilers are more sensitive to thermal stress and have more health problems, which may result in a higher percentage of DOA and condemnations (Bayliss and Hinton, 1990; Ritz et al., 2005). Sex-segregated rearing results in uniform live weight (Aviagen, 2018). The uniform weight of broilers reduced the injury problem during catching (Ritz et al., 2005). The hypothesis that the percentage of bruising was lower in male and female broilers than in mixed sex broilers in this study may be due to the fact that sex-segregated broilers have a more uniform body weight, which makes them easier to catch or manipulate and therefore have fewer bruises.

Age

It was found that with the increasing age of the broilers, the average body weight increased in chicken rearing. Thermal stress negatively affected the ability of chickens to regulate their body temperature. As broilers age and their body weight increase, their ability to thermoregulate decreases. Therefore, broilers with higher body weights were more susceptible to heat stress resulting in a higher percentage of DOA (Nijdam et al., 2004). In this study, younger broilers had a lower percentage of condemnations, which may indicate better health and higher immunity than older birds, as explained by Bayliss and Hinton (1990) and Ritz et al. (2005).

The reason why younger broilers had a lower percentage of bruising could be due to the higher metabolic activities in younger broilers and the higher cell proliferation rate compared to older broilers, so that older broilers were more susceptible to bruising and had more damaged tissue after 2 d of injury than younger broilers. Therefore, the percentage of bruising was lower in younger broilers than in older broilers (Hamdy et al., 1961). In addition, younger broilers were easier for the catching team to handle during rearing, manual catching, and transport due to their lighter weight, which resulted in fewer injuries and therefore less bruising (Langkabel et al., 2015).

Mortality and Culling Rate

Many factors, including high rearing density, high litter moisture, high ammonia, low ventilation, high temperatures, and problems with the disinfection process, showed clinical signs of decreased feed intake and performance in colibacillosis, a disease caused by E. coli that led to a high mortality rate during rearing (Whiting et al., 2007; Chauvin et al., 2011). Some of the broilers survived, but pathology developed in their gastrointestinal tract that could spread to other organs and caused respiratory infections. The pathology could also cause septicemia by colonizing internal organs (Muchon et al., 2019). Due to their lower physiological tolerance, they were more likely to die during transport than healthy broilers, and the percentage of condemnation increased with the characteristic of viscera unfit for human consumption (Lupo et al., 2009; Lupo et al., 2010; Cockram and Dulal, 2018). The higher mortality and culling rate during the rearing period could be due to the unhealthy broilers. The remaining broilers were weak and of low weight, so intensive catching and handling were important. Intensive care may result in fewer injuries, which translates into fewer bruises.

Weight per Create

The increase in broiler weight per crate was caused by either a higher number of broilers per crate or a higher weight per crate (Nijdam et al., 2004; Whiting et al., 2007; Chikwa et al., 2019) which resulted in a high percentage of DOA. The percentage of DOA increased during transport when the weight per crate increased and broilers were exposed to microclimate stress (Nijdam et al., 2004; Chikwa et al., 2019). In addition, the percentage of condemnation increased because flocks transported in poor conditions were more likely to have congested carcasses, which in turn led to higher condemnation rates (Lupo et al., 2009). High stocking density in the crate resulted in less space per broiler, but injuries and bruising were lower due to less movement of broilers in the crate during the vibration of the transport vehicle (Randall et al., 1993; Yalcin et al., 2004; Edgar et al., 2013).

Rearing Stocking Density

This study found that the percentage of condemnation increased with stocking density. This was consistent with the study by Kanabata et al. (2023), which reported that high stocking density was associated with an increase in ascites condemnation, unattractive appearance condemnation, and overall total condemnation. The acceptable stocking density for rearing should be 33 to 39 kg/m² in Europe (The Council of The European Union, 2007). In this study, the stocking density ranged from 21 to 41 kg/m², which was quite high. Factors such as limited access to feed and water, excessive heat, high ammonia levels, and air pollution due to insufficient air exchange rates were the main stressors at high stocking density that contributed to poor health and performance (Cengiz et al., 2015). The poor health status could be related to the decreased immune response observed at high stocking density, which could be due to the higher corticosterone affecting cytokine production, lymphocyte proliferation, and production of anti-inflammatory substances (Houshmand et al., 2012). Lupo et al. (2010) found that the condemnation rate was higher in flocks with health disorders.

Mean Body Weight

In this study, mean body weight did not affect the percentage of DOA, whereas higher body weight decreased the percentage of condemnation but increased the percentage of bruising. Both age and body weight increased sensitivity to heat stress during production. In addition, environmental changes in barns can affect broiler stress levels and susceptibility to disease. Humidity and high temperatures increased disease virulence (Fancher et al., 2021). Resistance to colibacillosis and growth rate were inversely related (Johnson et al., 2008). Emphasis on nutrient redistribution and maximum development competing with immune system maturation and function contributed to high infection rates (Rama Rao et al., 2003). Cleaning and management practices in broiler houses were the first steps in controlling the environment and preventing E. coli infections that led to condemnations and resulted in lower growth rate and mean bodyweight (Fancher et al., 2021).

Manual catching in the present study could increase the percentage of bruising, which could be because manual catching of broilers with a higher mean body weight caused the catching team to be fatigued for a longer period and injured the broilers while filling and moving the crates, resulting in a higher percentage of bruising (Delezie et al., 2006). In addition, Caffrey et al. (2017) reported that broilers were manually loaded into crates on the truck, that there was a significant effect of the catching team on mortality risk, and that differences between catching teams in how they caught, carried, and loaded the broilers likely had a significant impact on injuries and mortality.

Transport Duration

Compared to a farm near the slaughterhouse, the longer transport distance would require more time for transport. When transporting broilers to the slaughterhouse, they must be placed in a crate with limited space and reduce their body temperature by panting through the skin and excretions, so an increase in weight per crate affected the increasing ambient humidity. It was found that a higher average body weight made it more difficult to reduce body temperature under these conditions. These high temperatures and high humidity contributed to heat stress, hyperthermia, and a higher percentage of DOA (Nijdam et al., 2004; Vieira et al., 2010; Luptakova et al., 2012). According to the study by Kranen et al. (2000), injuries were more common during transportation due to impact from moving vehicles, muscle strain from maintaining balance while driving, or overheating.

Lairage Time

The fact that the DOA percentage within a load was affected by prolonged lairage time in the present study could be due to the fact that the broilers have limited space in the crate and heat accumulates in the crate, which has a significant impact on the thermal environment experienced. In a lairage area, inadequate ventilation could be a challenge for all broilers in the load. The temperature inside the crates could rise above the outside temperature, causing a high percentage of DOA with an increase in body temperature of more than 1°C (Bayliss and Hinton, 1990; Mitchell and Kettlewell, 1998; Nijdam et al., 2004; Jacobs et al., 2017a). Some deaths during lairage were due to injuries during catching, while others were due to problems during transport (Chauvin et al., 2011).

This study also noted the effect of lairage time on the incidence of bruising. More than 40% of bruises were caused by catching and crating (Griffiths, 1985). Therefore, these bruises could be detected during post-mortem inspection. If the color of the bruise was red within the last 12 h, it most likely originated from the slaughterhouse, and after that, it became dark red and purple, which happened during catching and was easily detected during post-mortem inspection (Aviagen, 2018). The longer the lairage time, the darker the color of the bruise.

Feed Withdrawal Time

Feed withdrawal prior loading serves to give the digestive tract time to empty before processing, leaving less ingesta, and fecal matter for possible contamination of the carcass. Within 4 to 6 h of fasting, there was a significant reduction in the frequency of defecation and weight of intestinal contents (May and Deaton, 1989; Warriss et al., 2004).

Higher DOA and condemnation rates were the result of prolonged feed deprivation. According to Mitchell and Kettlewell (1998) and MacCaluim et al. (2003), high temperatures and high humidity affected the mortality risk of broilers during fasting. The lower air velocity due to thermal conditions in the holding barn, where birds were kept in crates on the truck and in lairage circumstances, caused the respiration rate to increase with increasing temperature and relative humidity, which eventually led to panting (Gleeson, 1985). The condition ended fatally once body temperature reached 46°C (Kettlewell, 1989).

The reason for the longer feed withdrawal period leading to more condemnations was fecal contamination of the carcass and muscle injury. The process of fecal contamination was caused by the longer feed withdrawal time, which caused the intestine to become fragile because much of the inner layers of the mucosa and submucosa became ruptured. During evisceration, the abdominal cavity and carcass became contaminated with fecal matter (Lopez, 2010). In addition, depletion of hepatic glycogen stored during feed deprivation resulted in hypoglycemia (low blood glucose), and animals were under exhaustive stress, making them more susceptible to muscle injury (Kranen, 2000).

Limitations

The study presented here may have limitations, even though a representative sample of the integrator broiler population was selected and the structure of the data was considered at several levels. The methodology of the study did not allow extrapolation of the results to a larger population of Thai broilers. Similarly, the use of data from a single integrator may reveal low variance in practices and traits, making it difficult to identify risk factors for poor outcomes. However, the results of the study may be generalizable to other broiler populations in similar environments because of the relative uniformity of standard broiler production practices.

Another potential limitation of the present study is that the data used in this study were from the year in which the Covid-19 pandemic occurred, so older broilers were included in the records due to staff shortages at farms and slaughterhouses and lower demand from buyers. Results showed that older broilers with higher body weights had a higher percentage of DOA, condemnation, and bruising, suggesting that these variables may be more prevalent than usual.

The data in the present study were from one year, so longitudinal data may be needed to examine variations in DOA, condemnation, and bruising. The causes of condemnations in this study differed from those in Buzdugan et al. (2020) and Junghans et al., 2022 because Thailand is a tropical region and condemnations such as ascites are not common, so they were not included in this study, which is different from other studies in temperate regions. Thus, some of the individual condemnations recorded in this study were based on commercial practices and buyer requirements. Therefore, the results of this study may not be extrapolable.

Finally, our study identified risk factors for DOA, condemnations, and bruising that may be of limited value for control strategies because they are difficult to change, such as chicken sex, because all DOA, condemnations, and bruising occurred less frequently in female chickens, but it is not possible to raise only female chickens in commercial practice.

In summary, the 3 major production losses in broilers raised without antibiotic use, death on arrival, condemnation, and bruising, are influenced by several factors. Some risk factors could be reduced by changing management prior to slaughter. The main risk factors that may be easily controlled are slaughtering broilers at a younger age (40–47 d) with low average body weight, reducing the weight per crate, shortening the lairage time, and shortening the feed withdrawal time, which may help improve animal welfare and broiler producer productivity. Controlling the environment when the season changes, transporting broilers at night and in the morning, or providing a place to spray water during daytime transport could help minimize losses.