Welfare characteristics of laying hens in aviary and cage systems

The poultry production industry is progressively phasing out cage systems (CS) due to increasing concerns regarding animal welfare (Kollenda et al., 2020). The absence of nests in CS is particularly detrimental to welfare (Duncan, 1992), as hens naturally prefer to lay eggs in nest boxes (Reed, 1994). Moreover, the CS limits natural behaviors such as roosting and scratching, affecting hens both physically and psychologically (Nicol, 1987; Baxter, 1994). Consequently, the industry has shifted its focus toward housing systems that better meet welfare standards (Broom, 2010; Butterworth, 2013; Sandøe et al., 2020).

Among the alternative systems, the aviary system (AS) has gained popularity among farmers due to its efficient use of space and ease of management compared to floor or free-range systems. The AS is a multi-tiered structure that offers an enriched environment, featuring a littered ground floor and multiple levels. This setup allows hens to engage in behaviors like running, wing flapping, and dustbathing, which are often restricted in cage settings (Leyendecker et al., 2005; Papageorgiou et al., 2023). Ahammed et al. (2014) suggest that AS can be a welfare-enhancing option for commercial farms, especially when considering the higher costs associated with animal welfare-certified eggs.

However, the AS also presents its own set of welfare challenges. For instance, non-beak-trimmed birds housed in AS may experience higher rates of feather damage and mortality (Vasdal et al., 2023). Furthermore, some studies suggest that AS may be less efficient than CS in terms of egg production (Philippe et al., 2020; Kato et al., 2022). Philippe et al. (2020) found that CS outperformed AS in egg-laying productivity as well as certain aspects of animal welfare. Despite these findings, the overall differences between the AS and the CS in both productivity and welfare remain unclear. Previous studies have demonstrated significant variations in egg quality (Cepero and Hernándiz, 2015), stress levels (Lay Jr et al., 2011), and feather-pecking behaviors (Heerkens et al., 2015) among housing systems. However, these variations could be influenced by factors such as experimental period, chicken breed, and beak trimming practices, as noted by Tahamtani et al. (2014), Blatchford et al. (2016), and Schwarzer et al. (2021). Therefore, to appropriately compare and evaluate the welfare and productivity of laying hens in AS and CS, it is necessary to control for these factors and conduct a comprehensive assessment that includes physical, behavioral, and physiological indicators.

This study compares the egg production performance and welfare indicators, encompassing physical condition scores, behavior, and stress hormone levels, of hens housed in cage and aviary systems. To minimize external variables, such as flock management and environmental conditions, the study was conducted on a single commercial farm that operates both CS and AS concurrently, with both systems adhering to Korean legal space requirements. We hypothesized that, when controlling for factors such as breed, nutrition, handling practices, and climatic conditions, laying hens will exhibit distinct production and welfare outcomes depending on their housing system. Furthermore, we expect this study to demonstrate that an approach combining physical condition evaluation and behavioral observation provides an effective means of assessing hen welfare and comparing it between housing systems.

The study was conducted between February and August 2023 on a commercial farm located in eastern South Korea. The experimental procedure was reviewed and approved by the animal experimental ethics committee of Chonnam National University (approval number: CNU IACUC-YB-2023-05).

This study provides valuable insights into the welfare and productivity of laying hens housed in different systems, specifically comparing cage and aviary systems. Notably, hens in the AS consistently exhibited better feather conditions on their back and wings compared to those in the CS at all ages. The AS hens also exhibited greater movement, as indicated by higher ADT values, which likely reflects their ability to move freely rather than merely heightened fear. Their more frequent interactions with novel objects in the NOT suggest increased curiosity and potentially lower fear responses. These behaviors, along with lower yolk CORT levels, suggest that AS provide a more enriched and welfare-friendly environment. In contrast, CS hens, while producing heavier eggs, exhibited more negative behaviors and had higher yolk CORT levels. These findings emphasize the need to balance productivity with welfare considerations in housing systems and improve housing environments to promote both animal welfare and egg production. Further research should focus on additional factors influencing welfare outcomes within various housing systems, with the goal of further optimizing welfare strategies and enhancing the overall well-being of laying hens.

Plumage condition is a key indicator of hen health, reflecting the effects of their interactions with both the housing environment and other birds (Pichová and Bilčík, 2017). Hence, assessing plumage conditions offers valuable insights into the welfare status of laying hens across various housing systems. Previous research has indicated that feather pecking, including behaviors such as cannibalistic attacks and pecking that damage skin and tissue, is associated with stress, discomfort, and fear (Bilč ík and Keeling, 1999; Gunnarsson, 1999; Lambton et al., 2013). This could explain our current findings, where the higher scores observed in all feather pecking parameters in the CS suggest increased stress levels in the hens. This is in line with previous studies by de Haas et al. (2013) and El-Lethey et al. (2000). It is plausible to speculate that these elevated scores may be attributed to the stressful conditions present within the housing environments. Several authors have observed that housing systems affect plumage conditions (Appleby et al., 2002; Weitzenbürger et al., 2006), while others found no evidence of such impacts (Moinard et al., 1998). Differences in feather pecking scores were evident as early as 28 weeks of age in the present study, suggesting that the disparity in stress levels may have begun early in the rearing period, possibly starting from the pullet stage, rather than between 28 and 48 weeks. The lack of change in feather pecking scores between 38 and 48 weeks suggests that the hens may have already experienced significant feather damage by 38 weeks and that feather pecking behavior stabilized thereafter. In several studies, the condition of the hens’ plumage deteriorated with increasing age (Hinrichsen et al., 2016; Sokołowicz et al., 2023). Such deterioration could be due to plumage wear, the accumulation of damage, or an increase in severe feather pecking behavior, but most damage has been confirmed to be caused by feather pecking (Bilč ík and Keeling, 1999). These findings highlight the importance of early intervention during the rearing process to manage stress and prevent severe feather damage over time.

Evaluating comb wounds in laying hens is commonly used to assess aggression, often interpreted as a response to stress (Roenchen et al., 2007). However, we found a higher incidence of comb pecking wounds in the AS group, suggesting they may result from the dynamics of increased space and social interactions. In AS, hens have more freedom to move, which could lead to more frequent comb pecking due to heightened opportunities for social interaction (Engel et al., 2019; Tünaydın and Dikmen, 2019). In contrast, hens housed in CS likely could only rarely engage in comb pecking behavior due to the physical limitations on their movement within the confined space. However, Elson and Croxall (2006) observed a higher incidence of comb injuries in CS, indicating that aggression can also occur in more restricted environments. Our study found that despite the increased frequency of comb pecking in AS, there were no severe injuries over the 48-week period. Direct behavioral observations were not conducted to confirm whether comb pecking was primarily associated with dominance interactions, so our interpretation remains hypothetical. However, the absence of severe injuries suggests that, in AS, pecking may function as a low-intensity social interaction rather than a form of harmful aggression. This suggests that while the AS environment allows for more natural behaviors, including pecking as a form of social interaction, the behavior might be less intense or more regulated, possibly due to the larger space and a more complex social structure. Therefore, the presence of comb wounds should not be interpreted solely as a sign of aggression or poor welfare but rather as a natural behavior related to social hierarchy. Further research is needed to directly assess the role of comb pecking in social dynamics and its implications for welfare, as well as to develop effective strategies to manage this behavior.

The QBA evaluates both negative and positive aspects of animal experiences, focusing on their affective states and behavioral expressions (Fleming et al., 2016). Hence, it provides insight into animals’ emotional and behavioral states, which makes it an effective tool for assessing animal welfare (Wemelsfelder and Lawrence, 2001). In the present study, CS hens exhibited behaviors indicative of stress, fear, and boredom more frequently, whereas AS hens demonstrated behaviors associated with happiness and activity more often. These findings are consistent with previous studies (Shimmura et al., 2010; Nenadović et al., 2022) comparing CS and AS. Lay Jr et al. (2011) suggested that the behavioral restrictions in CS can lead to increases in negative behaviors and decreases in positive behaviors. The greater expression of positive behaviors observed in AS hens in this study may be attributed to their increased opportunities for natural behaviors and a wider available range of movement.

Similarly, Niekirk et al. (2012) found that QBA scores were lowest in conventional cages and highest in AS, likely due to limited space for behavioral expression in cages and the potential influence of the observer effect. In our study, we analyze the QBA data using PCA, which is commonly used to reduce dimensionality and highlight key behavioral patterns that contribute to overall welfare scores. This enabled us to group related behavioral expressions and identify primary factors driving the differences in QBA scores between housing systems. The PCA-derived scores in our study suggest that, while the QBA effectively captures the behavioral expressions and emotional valence of the hens, it does not directly correlate with their physical condition. This aligns with the findings of Vasdal et al. (2022), who also found no direct relationship between QBA scores and physical condition. The lack of this correlation suggests that, while QBA are effective for assessing behavioral and emotional states, they may not fully reflect the physical health of animals. However, it is important to acknowledge that QBA relies on subjective assessment, and the observer’s awareness of housing conditions could introduce potential bias. Further research incorporating blinded assessments or automated behavioral analysis is necessary to better understand the relationship between behavioral assessments and physical welfare indicators in laying hens.

The NOT assesses the competing motivations of animals to either approach or avoid new objects (Miller, 1944), and it is widely employed to gauge levels of fearfulness in poultry (Jones, 1996). We found that, on average, 2.8 fewer laying hens approached a novel object in the CS (6.6 hens) than in the AS (9.4 hens). The placement of the novel object in the CS, which was in front of the feed trough, might have reduced its visibility to the hens inside the cages, limiting the number of hens that approached it. Additionally, the present results showed a tendency for fewer CS hens to gather around the novel object as they aged, which may indicate a reduction in environmental stimulation or a decrease in their ability to exhibit natural behaviors as they grow older. In contrast, AS hens consistently showed a willingness to gather around the novel object, with significantly more responding to the object in the AS than in the CS at 38 and 48 weeks of age. This supports a previous finding that the AS better fosters hens’ natural behaviors, with reduced fearfulness and increased exploration linked to superior physical health (Nannoni et al., 2022). This is further evidenced by the correlation observed in this study between the extent to which the laying hens approached novel objects and their breast plumage condition. A recent study has demonstrated that flocks experiencing feather damage and cannibalism exhibit greater fear of novel objects (Hüttner et al., 2023). Indeed, an increased fear response in hens has been associated with feather pecking (Rodenburg and Koene, 2004; Uitdehaag et al., 2008), which in turn can cause feather damage and cannibalism (Rodenburg et al., 2013). However, it is also possible that hens experiencing feather pecking become more fearful as a result of this behavior, suggesting a bidirectional relationship. As numerous studies have suggested, the results of this study further confirm that the NOT is a suitable method for assessing fear responses in laying hens (Uitdehaag et al., 2008; Rozempolska-Rucinska et al., 2017; Tahamtani et al., 2023; Hüttner et al., 2023).

Avoidance behavior in hens serves as a critical indicator of their fearfulness towards humans (Graml et al., 2008). The observed differences in avoidance behavior between hens in CS and AS suggest varying levels of habituation to humans and comfort around them, likely influenced by the specific environmental conditions of each housing system. These findings are consistent with those of Graml et al. (2008), which compared cage and free-range systems and revealed a similar pattern, indicating a consistent trend across various housing environments. The greater avoidance distances observed in AS hens may indicate increased stress and social anxiety, making these hens more likely to evade close human contact. However, it is important to note that the AS offers a more intricate environment than the CS, with greater opportunities for interactions with other hens and enrichment materials, which could affect their reaction to human approach. Conversely, the limited space and restricted capacity for retreat in CS due to the cage design might have led to shorter avoidance distances. Further research is necessary to explore the underlying factors contributing to these behavioral differences in hens and to develop strategies for accurately assessing their stress levels and comfort in the presence of humans across various housing environments.

The H:L ratio is commonly used to assess stress in laying hens, following the method by Gross et al. (1983). Earlier studies focusing on chickens suggest that a higher H:L ratio is related to higher levels of environmental stress (McFarlane and Curtis, 1989). Cotter (2015) found that H:L ratios observed in enriched cage systems (ECS) were significantly higher than those recorded in AS or CS, suggesting that hens in ECS may experience greater stress levels than those in either of the other systems (Cotter, 2015). However, another study recorded consistent H:L ratios in both conventional and ECS, suggesting that stress levels were low and comparable, likely due to the birds’ prolonged familiarity with their cage mates and environment throughout the study’s 11-month production period (Tactacan et al., 2009). Similarly, studies on laying hens have shown that stocking density does not affect hematological values, such as leukocyte differential counts and the H:L ratio (Thaxton et al., 2006; Tuerkyilmaz, 2008). The impact of feed restriction on immune parameters, including blood values, is also debated, with some studies finding effects (de Jong et al., 2002; Khajavi et al., 2003), while others do not (Hocking et al., 1994; Liew et al., 2003; Fassbinder-Orth and Karasov, 2006). These variations may be influenced by stress-related factors, such as the duration and intensity of stress exposure, and measurement-related factors, such as the specific indicators used and the timing of measurements (Mumma et al., 2006). In our study, the H:L ratio observed in the AS was greater than that in the CS, which may be due to the increased physical activity and movement inherent in AS. However, further research is needed to fully understand the specific factors within different housing systems that contribute to specific stress responses.

While the WBC count alone does not directly assess stress in laying hens, the granulocyte and agranulocyte components provide insight into the immune status of birds and their ability to respond to infections and other health challenges (Alabi et al., 2014). Thus, the WBC count is more effective at indicating a bird’s capacity to combat diseases rather than assessing stress related to physical and physiological factors. The higher WBC count observed in AS hens may suggest a stronger immune response or greater exposure to pathogens, likely resulting from the more dynamic environment of the AS. Cotter (2015) found that the normalized total WBC frequency distribution for ECS significantly differed from those of AS and CS, with each group possessing a distinct distribution. This underscores the potential impact of housing conditions on immune responses. Moreover, the findings align with those of Thaxton et al. (2006), who also reported that housing environments significantly influence hematological parameters in laying hens. As immune function is crucial to enhancing laying hen welfare and health, these findings demonstrate the importance of further research into the impact of different housing environments on their immune status.

Evaluating blood characteristics and serum CORT levels can provide valuable insights into the physiological responses of laying hens to different housing environments (Maxwell, 1993). Mashaly et al. (1984) demonstrated that higher stocking densities increased plasma CORT levels in hens, whereas Davis et al. (2000) reported no change in CORT levels with increased stocking density. Additionally, Tactacan et al. (2009) reported that plasma CORT levels were equivalent in birds housed in CS and ECS, implying that stress levels were similar between the two caging systems. However, it is important to note that the levels of CORT in the blood can be influenced by various factors, including the time of day, bird handling, and individual variability, which may obscure subtle differences between the housing systems (Beuving and Vonder, 1978; Tactacan et al., 2009). Additionally, factors such as the method and timing of blood sampling, the hens’ previous exposure to stressors, and even the ambient temperature at the time of collection can significantly impact CORT levels (Beuving and Vonder, 1978). Therefore, while our study did not find significant differences in CORT levels between CS and AS hens, the possibility that subtle variations were masked by external factors cannot be entirely excluded.

It is well-established that age and housing system are key factors influencing egg quality (Ahmadi and Rahimi, 2011; Holt et al., 2011). Housing systems affect egg weight by influencing feed efficiency and controlling the environment, while age impacts egg weight through improved nutrient utilization and physiological maturity, with older hens generally producing heavier eggs (Zita et al., 2009). This age-related increase in egg weight is supported by various studies (e.g., Bozkurt and Tekerli, 2009; Krawczyk, 2009). Regarding housing systems, Lewko and Gornowicz (2011) reported that eggs from CS are heavier than eggs from litter systems, and Philippe et al. (2020) reported that laying rate and egg weight were significantly lower in AS than CS, likely due to higher levels of animal activity and competition for resources in AS. These findings are consistent with our study, where eggs produced in CS were heavier than those produced in AS. This may be the result of better feed efficiency, reduced physical activity, and a more controlled environment, benefits of CS noted by Widowski et al. (2022). However, some studies have presented contrasting findings. For instance, Zita et al. (2018) found that eggs from litter systems were heavier than those from cages, while Vits et al. (2005) reported heavier egg weights from floor pens than from CS. In these cases, lower egg weights in CS may be related to higher egg production rates, as suggested by Michel and Huonnic (2003). In contrast, studies by Ahammed et al. (2014) and Ghanina et al. (2020) found no significant differences in average egg weights between AS and CS. The inconsistencies between these studies may be due to variations in experimental conditions, including hen breed, age, diet, and environmental management, emphasizing the need for further investigations into the impact of housing systems on egg quality in which these factors are controlled for.

CORT is the primary glucocorticoid found in hen plasma, and its concentrations in avian eggs increase when this hormone is experimentally elevated in the mother’s bloodstream (Hayward and Wingfield, 2004; Love et al., 2005) and under stimulated stress conditions during egg production (Saino et al., 2005). In birds, yolk CORT preferentially accumulates from plasma during early follicular development under low-stress conditions, while stress elevates its accumulation in later stages of egg formation (Okuliarová et al., 2010). For instance, seven days after experimental treatment, increased plasma CORT in Japanese quail led to elevated CORT levels in their egg yolks (Hayward and Wingfield, 2004). Factors such as stress (Lee et al., 2022), age (Widowski et al., 2022), and health status (Dai et al., 2019) can all influence hormone deposition in yolks. Our results, which show higher CORT levels in the eggs of hens in CS compared to those in AS at 48 weeks of age, align with the results of previous studies (Engel et al., 2019; Son et al., 2022). These elevated CORT levels are likely due to the cumulative stress effects associated with the housing conditions over time (Goymann, 2005). We specifically analyzed yolk CORT at 48 weeks because this period likely allowed enough time for any cumulative effects of the housing environment to manifest in the yolk CORT levels. The absence of significant differences in yolk CORT at 28 and 38 weeks suggests that other factors, such as age and environmental conditions, may also influence hormone levels during the laying period (Widowski et al., 2022). Although this study found differences between housing systems in yolk CORT, the specific mechanisms behind these variations were not investigated. We propose that the observed differences in CORT levels may be attributed to increased stress, which was also reflected in physical indicators, such as feather quality, and body condition, and behavioral responses. However, further research is needed to confirm this hypothesis and to fully understand the underlying mechanisms.

This study compared and evaluated the welfare and productivity of laying hens housed in cage and aviary systems by examining various parameters, including physical condition, behavior, blood properties, egg quality, and CORT levels in serum and egg yolks. While the CS offered certain benefits to productivity, it appears to restrict the hen’s ability to engage in natural behaviors, as evidenced by increased feather damage and the expression of negative behaviors. The findings emphasize the necessity of conducting such in-depth comparisons while controlling for environmental variables to accurately assess the impact of different housing systems on hen welfare. Considering these controlled conditions, our current findings suggest that the AS provides a more favorable environment for laying hen welfare than the CS. These insights offer valuable guidance for future housing designs and management practices aimed at enhancing both animal welfare and egg production. However, as this study was conducted on hens from a single farm, larger studies across multiple farms are needed to confirm these findings and ensure broader applicability.