The impact of scheduled cage cleaning on older hens (Gallus gallus)

Transcription

1 The impact of scheduled cage cleaning on older hens (Gallus gallus) Kenneth E. Anderson, BS, MS, PhD, Paul E. Mozdziak, BS, MS, PhD & James N. Petitte, AB, MS, PhD Researchers are increasingly using the domestic hen (Gallus gallus) as an animal model for ovarian cancer. The authors analyzed mortality rates of two large flocks of older hens that were being used for ovarian adenocarcinoma studies. All hens were fed the same maintenance diets, though some hens in each flock received experimental chemopreventive treatments. Per the request of a collaborating institution, partway through the study, the authors started to remove the hens in one of the flocks for cage changing once every 4 weeks. After the authors began cleaning some of the hens cages, the mortality rate in this flock increased significantly. Throughout the study, within each flock, hens in the treatment and control groups had similar mortality rates. These results suggest that regularly cleaning the cages of older hens may not promote better welfare or improve flock mortality. A major obstacle to ovarian cancer research has been the absence of a valid animal model 1 3. Human clinical trials that test prevention and treatment strategies for ovarian cancer are expensive, require many subjects and take many years to complete. To develop effective ovarian cancer chemopreventive strategies or therapies in a timely manner, researchers need to use animal models that closely mimic human ovarian cancer 1. Ideally, this research would allow scientists to identify the agents that have the greatest potential for ovarian cancer prevention and treatment, which could then be evaluated in human trials. Among the animal models of ovarian cancer that are currently being studied, the egg-laying domestic hen (Gallus gallus) could potentially be a useful model for chemoprevention research. The domestic hen has a high incidence of spontaneous ovarian cancer; in hens between the ages of 208 and 312 weeks, ovarian cancer incidence ranges between 11% and 35% (refs. 4,5). Researchers recently showed that at 200 weeks of age, 43% of hens had spontaneous ovarian cancer 2. Additionally, the egg-laying hen has a high ovulation rate, raising the possibility that chicken and human ovarian cancers have a common pathogenesis related to ovulation-induced genetic damage to ovarian epithelial cells 6,7. Domestic egg-laying hens that are older than 108 weeks of age have great potential to serve as an animal model for spontaneous formation of ovarian epithelial cancer. However, no one in the biomedical research community has thoroughly considered the recommendations provided by current guidelines 8 for the care and use of these hens in a biomedical setting. There are inconsistencies in the welfare recommendations put forth by the Guide for the Care and Use of Laboratory Animals (the ILAR Guide 8 ) and the Guide for the Care and Use of Agricultural Animals in Agricultural Research and Teaching (the FASS Guide 9 ). Whereas the FASS Guide 9 recommends that personnel clean the cages of hens before placing the hens in the cages and then only when necessary, the ILAR Guide 8 recommends that personnel clean the cages of domestic hens once every 2 weeks. The FASS Guide states that to minimize stress in hens and roosters, the birds may be kept in their cages for 18 months or longer, and that personnel should clean cages without handling the hens 9. On the other hand, the ILAR Guide 8 suggests that the cagecleaning frequency recommendations for other species, such as rodents, should be applied to birds held in cages as well. Additionally, the FASS Guide 9 recommends that each hen should have minimum floor space of 464 cm 2. Department of Poultry Science, North Carolina State University, Raleigh, NC. Correspondence should be addressed to K.E.A. (ken_anderson@ncsu.edu). 210 Volume 39, No. 7 JULY

2 In contrast, the ILAR Guide 8 recommends that each adult laying hen (weighing 1.5 kg) should have at least 929 cm 2 of floor space. In agricultural settings, personnel do not move laying hens until the end of their productive life. As researchers start to increasingly use older hens as biomedical animal models, these hens might be subject to more frequent handling for cage sanitation, per ILAR Guide recommendations 8. We are concerned that this increased handling could negatively impact research results and the welfare of the older hens, which have fragile bones. As a result, we decided to compare the mortality rates of two flocks of hens ( weeks old) being used for ovarian cancer adenocarcinoma studies. These studies were evaluating the effects of experimental chemopreventive treatments (vitamin D 3 and levonorgestrel) on ovarian cancer incidence. All hens in both flocks were fed the same maintenance diets, though some hens in each flock also received the experimental chemopreventive treatments. Initially, we followed the husbandry recommendations from the FASS Guide 9 and, once the studies began, did not handle the hens from either flock. At 48 weeks into the study of one of the flocks (when the hens were 156 weeks old), the IACUC of a collaborating institution insisted that we more closely follow the husbandry protocols suggested by the ILAR Guide 8. To better conform to recommendations from the ILAR Guide 8, we increased the amount of floor space for each hen in this flock (flock 2), and we began to clean the cages of the hens in flock 2 once every 4 weeks. We then compared the mortality rates of the hens in the two flocks, both before and after we implemented the husbandry changes for flock 2. Throughout the studies, we also recorded the ovulation rates and body weights of the hens. METHODS Study designs and animal husbandry We analyzed mortality data from two studies that were evaluating the impact of specific chemopreventive techniques on the rate of cancer formation in two different flocks of laying hens ( weeks old). The IACUC at North Carolina State University (NCSU) approved these studies, both of which were completely randomized. Before being used in these biomedical research studies, the flocks had been maintained under commercial dietary and management protocols 10,11. In our studies, we wanted to see if there was a difference in the incidence of ovarian cancer in hens that were fed with maintenance feed compared with hens that were fed maintenance feed plus additional vitamin D 3 (three times the amount in the maintenance feed) and levonorgestrel, a synthetic progesten. For both studies, we used an identical maintenance diet formulation, which had been previously developed for body maintenance of older hens 12. The maintenance diets consisted of 1,650 kcal per kg of feed, 9.8% crude protein, 1.33% calcium, 0.70% phosphorous, 0.42% lysine and 0.35% total sulfur amino acids. The feed was composed primarily of soybean and corn (Table 1). For the maintenance feed plus vitamin D 3 and levonorgestrel, personnel used a feed mixer to mix vitamin D 3 and levonorgestrel into the maintenance diet at the facility. We provided the feed and water to the hens ad libitum. For both flocks, personnel measured the body weights of the hens at 108 and 200 weeks of age and also halfway through the study, when the hens were 152 weeks old. We tried to keep the two flocks in similar, controlled environments (temperature of 78 ± 4 F; relative humidity of 70 ± 5%). The ventilation rate varied from 0.25 ft 3 per min per bird to 6.5 ft 3 per min per bird; we increased the ventilation rate when the temperature increased. We kept the animals on a 10-h:14-h light:dark schedule. During the light period of the day, we kept lighting at 20 lux. Each day, personnel counted and recorded the number of eggs produced by each flock. Personnel also checked for and recorded mortalities each day. We then added up the number of eggs produced during each 4-week period and calculated the number of ovulations per 100 hens during each time period (the number of eggs produced by a hen corresponds with the number of ovulations). We also added up the number of deaths for each 4-week period and calculated the mortality rate for each time period (as a percentage of the total number of hens in the flock). To try to determine the cause of death, we carried out a necropsy on each hen that died. Flock 1. Flock 1 consisted of 2,880 White Leghorn hens that had previously been at the North Carolina Layer Performance and Managements Test research laying houses at the Piedmont Research Station (Salisbury, NC). These hens were maintained using commercial standards for production research 9. The hens were negative for Mycoplasma gallisepticum and laryngotracheitis. When the hens were 104 weeks old, personnel moved the hens to another poultry house located at the Piedmont Research Station. This house was equipped with tri-deck battery style cages with dropping boards between the levels. To remove any accumulated dust, personnel vacuumed the house on a weekly basis. We randomly assigned the birds to different cages, allocating 2,880 hens to 192 replicates (4 cages per replicate). A replicate is a group of cages; we applied the same treatment to all hens within a replicate. We housed either three or four birds in each cage. The cages (Big Dutchman, Inc.; Holland, MI) were 61 cm 51 cm (width length), giving each bird 1,032 cm 2 (for three birds per cage) or 774 cm 2 (for four birds per cage) of floor space. In both the treatment group (n = 96 replicates) and the control group (n = 96 replicates), LAB ANIMAL Volume 39, No. 7 JULY

3 TABLE 1 Ingredients and composition of the maintenance diet Ingredient Composition (lb per ton of maintenance diet) Corn Soybean hulls 1, Wheat midds Calcium carbonate Mono-dicalcium phosphate Salt 9.15 Methionine 2.67 Choline chloride (60%) 1.00 Vitamin mixture a 1.00 Mineral mixture b 1.00 Fat 9.99 MYC-OUT 65 (mold inhibitor) 1.00 Selenium mixture c 1.00 Total 2,000 a The vitamin mixture supplied 5,500 IU vitamin A, 1,300 IU vitamin D, mg vitamin E, mg vitamin K (menadione), μg vitamin B, mg biotin, 750 mg choline, mg folic acid, 34.8 mg niacin, mg pantothenic acid, 1.2 mg pyridoxine, mg riboflavin and mg thiamin per pound of maintenance diet. b The mineral mixture supplied ppm manganese, ppm zinc, 0.034% iron, ppm copper, ppm iodine, ppm cobalt and ppm selenium per pound of maintenance diet. c The selenium mixture contained 0.06% selenium per pound of maintenance diet. half of the birds were housed three to a cage and the other half were housed four to a cage. For identification, we banded the wings of each hen with replicate number, cage number and hen number within each cage. We fed all of the hens the layer formulated diet (Table 1) and allowed them to adapt to their new environment for 2 weeks before starting the study. After 2 weeks, the hens were fasted for a short period (2.5 h) for gut clearance. We then began to feed the hens the body maintenance diet. Throughout the study, we monitored egg production, body weight and mortality. We housed the hens as per the agricultural standards, though we provided more floor space per hen than the FASS Guide recommended 8. In accordance with the FASS Guide recommendations 9, we only cleaned the hens cages before placing the hens in their cages. As a result, once we placed the hens in the cages, the hens remained there for the next 2 y. Flock 2. We acquired flock 2, a commercial flock of 5,184 White Leghorn hens (72 weeks old and at the end of their commercial productive life), from Ise America, Inc (Newberry, SC). We housed the hens at the Piedmont Research Station. The hens were negative for Mycoplasma gallisepticum and laryngotracheitis. Until they were 104 weeks old, we housed the hens in 216 replicates according to the North Carolina Layer Performance and Managements Test housing standards and husbandry practices, which comply with FASS Guide recommendations 10. We then carried out a baseline necropsy on 1,352 hens to look for the incidence of ovarian cancer in the flock. Due to the necropsies and to mortalities, the flock was reduced to 3,672 hens. To reduce the flock population, we removed hens from cages. We did not handle the hens that remained in the flock. When these hens were 108 weeks old, we started using them for the ovarian cancer research project. We assigned the hens in half of the 216 replicates to receive the maintenance (control) diet, while hens in other half of the replicates received the maintenance plus vitamin D 3 and levonorgestrel (treatment) diet. We housed hens in quad deck layer cages (Big Dutchman Inc), either four hens to a cage (61 cm 40.6 cm (width length); 619 cm 2 of floor space per hen) or five hens to a cage (81.3 cm 40.6 cm (width length); cm 2 of floor space per hen). In both the treatment group (n = 108 replicates) and the control group (n = 108 replicates), approximately half of the birds were housed four to a cage and the other half were housed five to a cage. Per the recommendations in the FASS Guide 9, once we placed hens in the cages, we did not clean the cages. Throughout the study, personnel vacuumed the house on a weekly basis to remove any accumulated dust. At 48 weeks into our study (when the hens were 156 weeks old), the IACUC of the collaborating institution insisted that we more closely follow the husbandry protocols suggested by the ILAR Guide 8. The collaborating institution s IACUC did not accept the regulatory oversight provided by NCSU IACUC, which had allowed us to roughly follow the FASS Guide 9 recommendations. To address the collaborating institution s request, we reduced the flock size to 2,592 hens (over a 4-week time period). We carried out necropsies on all hens removed from the flock (approximately 1,020 hens) to look for ovarian cancer incidence. By reducing the flock size, we were able to better conform to the floor space recommendations of the ILAR Guide 8. By the time the hens were 160 weeks old, they were housed either three to a cage (cage size of 61 cm 40.6 cm) or four to a cage (cage size of 81.3 cm 40.6 cm), yielding 826 cm 2 of floor space per hen. In both the treatment group (n = 108 replicates) and the control group (n = 108 replicates), half of the birds were now housed three to a cage and the other half were housed four to a cage. The ILAR Guide 8 recommends that cages be cleaned biweekly. The goal of cleaning cages is to maintain a reasonable degree of sanitation, which could thereby improve animal welfare. To mitigate any adverse effects of handling on older birds, we wanted to minimize the amount of bird handling. The collaborating institution s IACUC and the NCSU IACUC granted an exception 212 Volume 39, No. 7 JULY

4 a Mortality (%) b Mortality (%) Age (weeks) Age (weeks) FIGURE 1 Weekly mortality rates for leghorn hens age weeks in (a) flock 1 (cared for in accordance with the FASS Guide 6 ) and (b) flock 2 (cared for in accordance with the FASS Guide 6 at first but in accordance with the ILAR Guide 5 after the hens were about 160 weeks old). Mortality rates were significantly higher in flock 2 than in flock 1 from the time the hens were 172 weeks old (P < ). to the ILAR Guide 8 recommendation and allowed us to clean the cages once every 4 weeks instead of once every 2 weeks. To clean the cages, personnel removed hens from the cage and placed them in a holding coop. Personnel then cleaned any debris from the cages and sprayed the cages with a quaternary disinfectant. (BioSentry 904 Quaternary Disinfectant; BioSentry, Inc., Stone Mountain, GA). The disinfectant, which we mixed according to the manufacturer s instructions, had been approved for animal contact. Statistical analysis Mortality rate. We analyzed mortality data for each of the two flocks. Flock 1 had 192 replicates and flock 2 had 216 replicates. We used two methods to analyze the mortality data, both of which accounted for potential variability of replicates between flocks. For the first method, we created plots of percentage mortality versus time (Fig. 1). To analyze this data, we used the LIFETEST procedure (a χ 2 test statistic of 550.1; 1 degree of freedom; SAS Institute Inc., Cary, NC) to compute a common survivor function underlying both flocks. We then used the logrank test to compare the survival distributions of the two flocks. For the LIFETEST analysis, P < represents a statistically significant difference. For the second method, we used mortality data from the entire study period to calculate the mean percentage of surviving birds in each replicate, for any particular time range. We then tested the hypothesis that the populations of hens in flocks 1 and 2 were the same. We used the product-limit empirical estimator for the survival function of the birds in the two flocks. This method allows us to accommodate the right-truncation that occurred when we stopped recording the birds mortality rates (when the birds were 200 weeks old). For flock i, replicate j and time period t, the estimated mean lifetime ( ˆm ij ) is calculated by the following equation: t = 23 mij ˆ = Sˆ ( t ) t = 1 where Ŝi(t) denotes the product-limit estimator of the survivor function for flock i at time t: ˆ # birdsin replicate j surviving toor beyond timet Sij ( t ) = # birds that started out in replicate j At the end of the study, when the hens were 200 weeks old, each of the 408 replicates in both flocks had at least 7 birds surviving. Additionally, the lifetimes of all surviving birds are right-censored at the end of the study. This truncation figures in the estimation of the mean through the equation for the estimated mean lifetime given above. Once we calculated the estimated mean lifetime of the birds in all replicates in both flocks, we used the following formula to calculate μ i, the mean lifetime for the birds in a randomly sampled replicate from flock i: ˆm ij = m + E ij i ij E ij represents replicate-to-replicate variability within each flock. We assumed that these replicate lifetime errors were normally distributed but had different normal distributions, meaning they had inhomogeneous variance (E ij normal with mean 0, variance σ 2 i ). We carried out a simple t-test (PROC TTEST; SAS Institute Inc.) to test the hypothesis that the two flocks had equal mortality or equal mean lifetimes (H 0 : μ 1 = μ 2 ). Because of the pronounced inhomogeneity of the variance, we used the unequal variances t-test. Ovulation, body weight data and overall mortality data analysis. To analyze the ovulation, body weight and mortality data for both flocks, we used the General Linear Model (GLM) procedure (SAS Institute Inc.), a software program that uses the F-test to fit general linear models. The two components that entered into the error term were time period and dietary treatment. The analysis was based upon the replicate means. Within each flock analysis, means significantly different from one another (P < 0.05 by the F-test) were separated using the least squares means. RESULTS Responses to variable diets In general, both flocks had similar responses to the diets. Within each flock, the hens that were fed the LAB ANIMAL Volume 39, No. 7 JULY

5 TABLE 2 Ovulation frequency, body weight and average mortality in flock 1 and flock 2 throughout the entire study period. Flock 1 Ovulation frequency (per 100 hens) Body weight (kg) Mortality (%) Control 4.94 ± ± ± 0.09 Vitamin D ± ± ± 0.09 Flock 2 Control ± ± ± 0.03 Vitamin D ± ± ± 0.03 Values given are mean ± s.e. Ovulation frequency was significantly different between hens that were fed the maintenance diet only (control) and those that were fed the maintenance diet plus vitamin D 3 (P < 0.05) within each flock. maintenance diet had significantly higher numbers of ovulations per 100 hens than did the hens that were fed the maintenance diet with added vitamin D 3 (F-test in GLM procedure, P < 0.05; Table 2). Overall, the hens in flock 2 had significantly more ovulations per 100 hens than did the hens in flock 1 (F-test in GLM procedure, P < 0.05; Table 2). Within each flock, the hens in the control and vitamin D 3 groups had similar body weights and mortality rates (F-test in GLM procedure, P < 0.05). The differences between the flocks ovulation rates, body weights and mortality rates were consistent with the results seen before each flock was used for the chemoprevention studies, as shown in the layer test reports 10,11 from the facilities where the hens were housed prior to these studies. Lifetimes of hens in flocks 1 and 2 After accounting for right-censoring, we calculated the estimated mean lifetime of hens (± s.e.) in flock 1 to be ± 0.18 weeks and of hens in flock 2 to be 22.1 ± 0.06 weeks. The difference between the estimated mean lifetimes of hens in flock 1 and flock 2 is significant (LIFETEST, P < on d.f. = 230). The estimated population lifetime standard deviations, among cages, are 2.48 weeks for cages in flock 1 and 0.84 weeks for cages in flock 2. The LIFETEST results indicated that the mortality over time was greater in flock 1 than in flock 2 (P < ). Variability of flock mortality The mortality rate in flock 1 was significantly higher during the first 4 weeks after we moved the flock to the new house for the start of the first cancer study than it was throughout the rest of the study. The LIFETEST indicated that during the first 4 weeks, the mortality rates were significantly greater (P < ) in flock 1 than in flock 2. After the initial 4 weeks of the study, the mortality rate of hens in flock 1 decreased and remained relatively stable throughout the rest of the study (Fig. 1). The mortality rate of the hens in flock 1 peaked again when the hens were 152 weeks old, which was the week that we weighed the hens. The mortality rate of flock 1 when the hens were 152 weeks old was significantly higher than it was when the hens were weeks old and weeks old. At the start of the second cancer study, the hens in flock 2 had already been in their cages for 36 weeks. At the beginning of the study, the hens in flock 2 had a relatively low and stable average percentage mortality (Fig. 1). For the first 60 weeks of the study, the mortality profiles of flocks 1 and 2 did not differ significantly (LIFETEST, P > 0.05). There was a slight increase in mortality in flock 2 when the hens were 156 weeks old, which was 4 weeks after we had weighed these hens. Per the request by our collaborating institution, when the hens were 160 weeks old, we began to clean their cages once every 4 weeks. When the hens were 172 weeks old, the mortality rate in flock 2 was significantly higher than it had been at any other point in the study. The mortality rate of flock 2 continued to rise until 180 weeks of age, when it reached a plateau that was approximately three times greater than the mortality rate during the first year of the study (Fig. 1). The overall mortality rate of flock 1 (6.4%) was significantly higher than flock 2 (1.3%; F-test in GLM procedure, P < 0.05). Necropsies To try to determine cause of death, we carried out necropsies of the animals that had died in each of the flocks. The necropsies indicated that no animals had died from physical injuries that had resulted from being handled. After we started cleaning the cages in flock 2, the necropsy results found only one hen (in flock 2) that had a broken bone that could have arisen from handling. This broken bone could have caused the death of this hen. DISCUSSION We examined the mortality of two large flocks of birds in a biomedical study and analyzed the changes in mortality rates associated with handling the hens for cage cleaning. With the exception of the first 4 weeks of the study, the mortality rates of flocks 1 and 2 remained relatively constant throughout the first 60 weeks of the study. 44 weeks into the study our collaborating institution s IACUC required us to start following the cage size and sanitation recommendations from the ILAR Guide 8 for flock 2 instead of continuing to follow the FASS guidelines 9. To minimize the handling of these older hens, we implemented a modified sanitation program and cleaned the cages once every 4 weeks, instead of once every 2 weeks as the ILAR Guide 8 recommends. 214 Volume 39, No. 7 JULY

6 We also reduced the number of hens per cage by removing some hens from the cages between weeks 48 and 52 of the study (when the hens were between 156 and 160 weeks old). After the second cage cleaning, which occurred when the hens were 168 weeks old, the mortality rate of the hens in flock 2 increased significantly (P < 0.05). The mortality rate of flock 2 continued to rise until the hens were 180 weeks old, when mortality reached a plateau that was approximately three times greater than flock s mortality rate during the first year of the study (Fig. 1). Our analyses indicate that when we followed the husbandry protocols recommended by the FASS Guide 9, mortality in both flocks remained relatively stable. In both studies, mortality rates varied only slightly, except for periods in which we handled the birds. For example, flock 1 exhibited increased mortality after the birds were moved at the beginning of the research project. The LIFETEST analysis indicated that the mortality rates were significantly greater (P < ) in flock 1 than in flock 2 during the first 4 weeks of the study. When we began to handle the 168- to 200-week-old hens in flock 2 for cage cleaning, this flock s mortality rate increased significantly (172 weeks old; P < ). Additionally, the mortality rate was significantly higher (P < ) in flock 2 than in flock 1 at the same time point (when hens in flock 2 were 172 weeks old). These results suggest that, if possible, it is best to minimize the handling of older hens. In our studies, handling seemed to be associated with increased rates of mortality. For our flocks, the FASS Guide 9 husbandry recommendations promoted better animal welfare than did the ILAR Guide 8 husbandry guidelines, which recommend regular cage cleaning. When carrying out studies with older hens, researchers should consider the possible impacts of cage cleaning and other husbandry practices on animal welfare and study outcomes. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. Received 22 January; accepted 19 April 2010 Published online at 1. Johnson, K.A. The standard of perfection: thoughts about the laying hen model of ovarian cancer. Cancer Prev. Res. (Phila. Pa) 2, (2009). 2. Hakim, A.A. et al. Ovarian adenocarcinomas in the laying hen and women share similar alterations in p53, ras, and HER-2/neu. Cancer Prev. Res. (Phila. Pa) 2, (2009). 3. Vanderhyden, B.C., Shaw, T.J. & Ethier, J.F. Animal Models of ovarian cancer. Reprod. Biol. Endocrinol. 1, 67 (2003). 4. Wilson, J.E. Adeno-carcinomata in hens kept in a constant environment (abstract). Poult. Sci. 37, 1253 (1958). 5. Fredrickson, T.N. Ovarian tumors of the hen. Environ. Health Perspect. 73, (1987). 6. Casagrande, J.T. et al. Incessant ovulation and ovarian cancer. Lancet 314, (1979). 7. Rodriguez, G.C. et al. in Ovarian Cancer (eds. Jacobs, I.J. et al.) (Oxford University Press, 2002). 8. Institute for Laboratory Animal Research. Guide for the Care and Use of Laboratory Animals (National Academies Press, Washington, DC, 1996). 9. Federation of Animal Science Societies. Guide for the Care and Use of Agricultural Animals in Agricultural Research and Teaching 2nd edn. (Federation of Animal Science Societies, Savoy, IL, 1999). 10. Anderson, K.E. Final Report of the Thirty Second North Carolina Layer performance and Management Test: Production Report 32, July (North Carolina State University, Cooperative Extension Service, Raleigh, NC, 1998). Available: < layer_reports/32_final_report.pdf>. 11. Anderson, K.E. Final Report of the Thirty Fifth North Carolina Layer Performance and Management Test 35, May (North Carolina State University, Cooperative Extension Service, Raleigh, NC, 2005). Available: < 35_final_report.pdf>. 12. Anderson, K.E. & Havenstein, G.B. Effects of alternative molting programs and population on layer performance: results of the Thirty-fifth North Carolina Layer Performance and Management Test. J. Appl. Poult. Res. 16, (2007). LAB ANIMAL Volume 39, No. 7 JULY