1471 Journal of Food Protection, Vol. 77, No. 9, 2014, Pages 1471 1480 doi:10.4315/0362-028x.jfp-14-021 Copyright G, International Association for Food Protection Survey of Salmonella Contamination in Chicken Layer Farms in Three Caribbean Countries ABIODUN ADESIYUN, 1 * LLOYD WEBB, 2 LISA MUSAI, 3 BOWEN LOUISON, 4 GEORGE JOSEPH, 5 ALVA STEWART-JOHNSON, 1 SANNANDAN SAMLAL, 1 AND SHELLY RODRIGO 6 1 School of Veterinary Medicine, Faculty of Medical Sciences, University of the West Indies, St. Augustine, Trinidad and Tobago; 2 Institute of Public Health Studies, College of Veterinary Medicine, Nursing and Allied Health, Tuskegee University, Tuskegee, Alabama 36088, USA; and 3 Poultry Surveillance Unit, Ministry of Food Production, Port of Spain, Trinidad and Tobago; 4 Ministry of Agriculture, Forestry and Fisheries, Ministerial Complex, Tanteen, St. George s Grenada; 5 Veterinary and Livestock Services, Ministry of Agriculture, Food Production, Rural Development and Fisheries, Castries, St. Lucia; and 6 Department of Public Health and Preventive Medicine, School of Medicine, St. George s University, University Centre, Grenada, West Indies MS 14-021: Received 14 January 2014/Accepted 7 March 2014 ABSTRACT This study was conducted to investigate the demography, management, and production practices on layer chicken farms in Trinidad and Tobago, Grenada, and St. Lucia and the frequency of risk factors for Salmonella infection. The frequency of isolation of Salmonella from the layer farm environment, eggs, feeds, hatchery, and imported day-old chicks was determined using standard methods. Of the eight risk factors (farm size, age group of layers, source of day-old chicks, vaccination, sanitation practices, biosecurity measures, presence of pests, and previous disease outbreaks) for Salmonella infection investigated, farm size was the only risk factor significantly associated (P ~ 0.031) with the prevalence of Salmonella; 77.8% of large farms were this pathogen compared with 33.3 and 26.1% of medium and small farms, respectively. The overall isolation rate of Salmonella from 35 layer farms was 40.0%. Salmonella was isolated at a significantly higher rate (P, 0.05) from farm environments than from the cloacae. Only in Trinidad and Tobago did feeds (6.5% of samples) and pooled egg contents (12.5% of samples) yield Salmonella; however, all egg samples from hotels, hatcheries, and airports in this country were negative. Salmonella Anatum, Salmonella group C, and Salmonella Kentucky were the predominant serotypes in Trinidad and Tobago, Grenada, and St. Lucia, respectively. Although Salmonella infections were found in layer birds sampled, table eggs appear to pose minimal risk to consumers. However, the detection of Salmonella-contaminated farm environments and feeds cannot be ignored. Only 2.9% of the isolates belonged to Salmonella Enteritidis, a finding that may reflect the impact of changes in farm management and poultry production in the region. Worldwide, table eggs are used in the preparation of numerous commercial and homemade products (7, 18). Salmonellosis resulting from the consumption of eggs or egg products has been reported in outbreaks globally, with both public health and economic implications (11, 19, 26, 30, 32). Rearing of layer birds in Salmonella-free farm environments and obtaining hatching eggs and day-old chicks free of Salmonella have been reported as imperative for preventing Salmonella from entering the human food chain (27, 31, 35). Salmonella can contaminate eggs by both vertical and horizontal transmission (14, 15, 27, 31). The frequency of isolation of Salmonella reported elsewhere has ranged from 4 to 12% in layer birds (3, 16), from 0.95 to 33.3% for layer farm environments or litters (6, 34), from 12.6 to 19.2% for imported day-old chicks (6, 31), from 0.35 to 1.24% for hatching eggs (6), and from 13 to 29% for hatcheries (13, 35). Rodents also have been documented as a potential source of Salmonella in layer farm environments (22, 23, 38). * Author for correspondence. Tel: 868-645-4481; Fax: 868-645-7428; E-mail: abiodun.adesiyun@sta.uwi.edu. Although all serotypes of Salmonella can cause eggborne salmonellosis, Salmonella Enteritidis has emerged as the serotype most frequently associated globally with most human epidemics (25, 33, 39). Since 1989, Salmonella Enteritidis had emerged in the Caribbean as a pathogen of public health concern, causing sporadic cases and outbreaks of diarrhea in both local and tourist populations. In 1996, this serotype was the most frequent cause of salmonellosis in the Caribbean (21). Salmonellosis has been increasing in frequency since 1986 in Trinidad and Tobago followed by Jamaica and Barbados (8). Indar et al. (20) reported that the prevalence of Salmonella on layer farms in Trinidad and Tobago was significantly higher on egg shells (4.7%) than in egg contents (1.2%). Adesiyun et al. (2) found that Salmonella Enteritidis was the serotype of 58.3% of the Salmonella isolates from composite eggs (shells, content, or both) collected from layer farms and supermarkets in Trinidad. To date, no comprehensive surveys of Salmonella in table eggs have been conducted in the Caribbean subregion, and the impact that changes in farm management and poultry production systems in the Caribbean region would have on the distribution of Salmonella is unknown.
1472 ADESIYUN ET AL. J. Food Prot., Vol. 77, No. 9 Several Caribbean countries import table and hatching eggs from within and outside the region, and the risk to consumers of Salmonella in eggs and egg products is of concern. The current study was conducted to survey Salmonella in layer birds and their farm environments, feeds, table eggs, hatcheries, imported eggs, and day-old chicks in three countries. The risk factors for Salmonella contamination of farm environments and eggs and infection of layers were also investigated, and the serotypes of the Salmonella isolates recovered were determined. MATERIALS AND METHODS Study strategy. The project was designed to be performed in the poultry industry in the Caribbean subregion. The first phase of the study involved dissemination of an electronic questionnaire to each regional chief veterinary officer, and the second phase involved sampling for Salmonella in countries that consented to participate in the study. Questionnaire. Questionnaires were sent to the chief veterinary officers or their designates in 27 Caribbean countries to elicit information for 1999 to 2010 regarding foodborne and eggborne outbreaks and agents implicated; importation of eggs (table or hatching), egg products, and day-old chicks (country of origin); distribution of imported eggs (hatcheries, supermarkets, local markets, or hotels) or chicks (layer or broiler farms); microbial testing of table eggs (local or imported) and birds (layers or broilers); and pathogens isolated. Only 11 (40.7%) of the 27 countries returned completed questionnaires. Selection of countries for the study. After analysis of the questionnaire data, the sample collection phase was conducted in three countries that consented to participate: Trinidad and Tobago (7 September to 3 October 2011), Grenada (17 to 21 October 2011), and St. Lucia (14 to 18 November 2011). Logistical problems (permission to conduct study and availability of technical support to identify and visit poultry farms and other venues) associated with coordinating the project in the three countries made it impossible for simultaneous (time and duration) sampling in these three countries. Types, numbers, and sources of samples. Samples were collected from layer farms, hatcheries, ports of entry, and the kitchens of hotels. The sample sizes were determined for each country using the following formula (37): n~1:96 2 P exp 1{P exp =d 2 where n is the estimated number of samples, P exp is the expected prevalence of Salmonella, and d is the precision level. Based on the questionnaire data, the farms were classified as small (less than 5,000 birds), medium (5,001 to 10,000 birds), and large (more than 10,000 birds). Overall, the questionnaire was administered to a total of 81 layer farms in Grenada (21 farms), Trinidad and Tobago (23 farms), and St. Lucia (37 farms), but samples were collected from only 35 (43.2%) of the farms. In each of the three countries studied, the following types of samples were collected, wherever possible: (i) On layer farms, cloacal swabs of laying birds (19 per small farm, 56 per medium farm, and 57 per large farm), drag samples of each pen in use (deep litter system), feeds in use at the time of farm visit, and freshly laid eggs (6 per small farm and 18 per medium and large farm) were collected. A questionnaire was administered at each farm to obtain demographic data and information about general operations. (ii) At the port of entry (airports), imported eggs (table and hatching), samples of fecal matter, and samples of bedding papers in boxes used to import day-old chicks were collected. (iii) At the hatchery, fertile hatching eggs (preincubation), meconium samples from day-old chicks, broken egg shells, and environmental swabs from hatchers and incubators were collected. Of the six layer hatcheries in the three countries (two in Trinidad and Tobago, one in Grenada, and three in St. Lucia), only one hatchery in St. Lucia was in operation (i.e., hatching eggs) during the scheduled sampling periods. A total of 38 samples (6 pooled hatching egg contents, 6 pooled egg shells, 20 meconium, and 6 environmental) were collected. (iv) In hotel kitchens, eggs stored in refrigerators and environmental swabs of all refrigerators and freezers were collected during the visits. Collection of samples: layer farms. To collect cloacal samples, sterile swabs were inserted into the cloacae of a randomly selected number of birds. The swabs were rotated inside the cloaca, placed into transport medium, and held in a cooler for transportation to the laboratory. To collect environmental samples (drag swabs), a sterile surgical gauze swab (10 by 10 cm) attached to 110 cm of string and stapled to a wooden tongue depressor was moistened in sterile buffered peptone water (BPW) and then dragged along the sides and through the middle of the pen. The swab was then placed in 20 ml of BPW as transport medium and held in a cooler with ice packs for transportation to the laboratory. Table eggs selected randomly from stacks of crates were removed with sterile gloves and placed in sterile containers. The containers were then wrapped in sterile aluminum foil and placed in a cooler for transportation to the laboratory. Approximately 50 g each of different types of feed used on the farm was aseptically collected and placed into sterile plastic containers. Collection of samples: port of entry. Twenty boxes (or as many boxes available if there were fewer than 20) of day-old chicks were selected randomly at the airport. From each box, two swabs of fecal material were collected and placed in commercially available transport medium (Copan, Brescia, Italy). Both stained and unstained shredded bedding paper from each box was removed and placed in a sterile fecal cup and appropriately labeled. Twelve eggs selected randomly were removed with sterile gloves, placed in sterile containers, and held in a cooler for transportation to the laboratory. Collection of samples: hotel kitchens. Twelve table eggs selected randomly from stack of crates were removed with sterile gloves and placed into sterile containers. The containers were then wrapped in sterile aluminum foil and placed in a cooler for transportation to the laboratory. To collect environmental samples, sterile cotton-tipped applicators were moistened in sterile saline and rotated along the entire area of the handles of all refrigerators, using a new swab for each refrigerator. Each swab was placed in 20 ml of sterile saline and transported to the laboratory on ice. A similar procedure was used for freezers. Collection of samples: hatcheries. To collect environmental samples, sterile cotton-tipped applicators were moistened in sterile saline and used to swab the entire interior of each hatcher. The
J. Food Prot., Vol. 77, No. 9 SALMONELLA IN LAYER FARMS IN THE CARIBBEAN 1473 swabs taken for each hatcher were then placed in 20 ml of sterile saline and transported to the laboratory on ice. Six eggs selected randomly were removed with sterile gloves and placed into sterile containers. The containers were then wrapped in sterile aluminum foil and placed in a cooler for transportation to the laboratory. To collect meconium samples from hatched birds, a sterile swab was rotated on the meconium samples, placed into transport medium, and transported to the laboratory on ice. Five samples of broken egg shells from the hatcher trays were aseptically collected and compressed into sterile fecal cups for transportation to the laboratory. Processing of samples. Samples preenriched in BPW (eggs, environmental drag samples, feeds, broken egg shells, and environmental swabs from hotels and hatchers) were selectively enriched in tetrathionate (TT) broth (Oxoid, Basingstoke, UK) and Rappaport-Vassiliadis (RV) broth (Oxoid) and incubated for 18 to 24 h at 37 and 42uC, respectively. Processing of samples: shell eggs. Six eggs were pooled for each sample. Moistened sterile swabs were applied to the entire external surface of each egg and subsequently dipped in 20 ml of sterile saline as previously described (2). The eggs were then disinfected by submerging them in 70% alcohol for 10 min and then air dried. Each egg was then flamed at the pointed end and cracked with a sterile scalpel blade, and the contents were poured into a stomacher bag. The pooled egg contents were homogenized for 60 s at low speed in a Seward 400 Lab Stomacher (Seward, London, UK). Twenty-five milliliters of pooled egg contents was then mixed with 225 ml of BPW and incubated at 35uC for 18 h. Ten milliliters of the pooled egg shell swab rinsate was placed in 90 ml of BPW and incubated for 35uC for 18 h. Processing of samples: environmental drag samples. The drag swab sample from each pen was placed in 250 ml of BPW and incubated at 35uC for 18 h for enrichment. Processing of samples: cloacal, meconium, and fecal swabs. Each swab was mixed with 10 ml of sterile saline, and 0.5 ml of this suspension was then added to 9.5 ml of TT broth and 9.5 ml of RV broth and incubated for 24 h at 35 and 42uC, respectively. Processing of samples: chick bedding paper. Approximately 5 g of the shredded chick bedding paper with stained and unstained areas was added to 15 ml of sterile saline and mixed with a vortex mixer, and 0.5 ml of this suspension was added to 9.5 ml of TT broth and 9.5 ml of RV broth and incubated as described. Processing of samples: feed and broken egg shells. Each 25-g sample of feed and egg shells was aseptically weighed, placed in a stomacher bag containing 225 ml of BPW, and homogenized for 2 min. The homogenized mixture was then placed into a sterile flask and incubated at 35uC for 18 h. Processing of samples: environmental swabs (hotel kitchens and hatcheries). Ten milliliters of suspension from each swab was added to 90 ml of BPW and incubated at 35uC for 18 h. Isolation, identification, and serotyping of Salmonella. All selective enrichment broths were subcultured onto xylose-lysinedesoxycholate agar (XLD; Oxoid) and brilliant green agar (BGA; Oxoid) and incubated aerobically at 37uC for 24 h. Suspect colonies of Salmonella on XLD and BGA were subjected to standard biochemical tests (4, 24). Isolates biochemically identified as Salmonella were subjected to a slide agglutination test using a commercial kit with Salmonella polyvalent antiserum (A-I & Vi; Difco, BD, Detroit, MI). All slide-positive isolates were confirmed as Salmonella and serotyped at the Caribbean Epidemiology Center (Port of Spain, Trinidad and Tobago). Statistical analyses. Data were analyzed with the Statistical Package for Social Sciences (SPSS), version 22.0 (IBM, Armonk, NY). Logistic regression analysis was performed to determine which independent variables, i.e., country of origin, type of farms (large, medium and small), type of samples (cloacal swabs, environmental swabs, feed, eggs, hatchery, imported day-old chicks), and risk factors for Salmonella were predictive for the presence of Salmonella, the dependent variable. A forward stepwise method was used to test variables for entry into the model (at P, 0.05), and independent variables were removed (P. 0.10) based on the significance of the change in the likelihood ratio. Because explanatory variables are often correlated, the stepwise method was used because it combines both the forward selection and backward elimination. RESULTS Demographic data and management and production practices in layer farms in three Caribbean countries. The demographic data and management and production practices for the three countries are shown in Table 1. Layer farms in Trinidad and Tobago were significantly larger (P ~ 0.000); 6 (26.1%) of 23 farms in this country had over 20,000 birds, whereas none of the 21 and 37 layer farms in Grenada and St. Lucia, respectively, were that large. The frequencies of four practices (housing system, age pullets start to lay, age when layers are culled, and preparation of eggs on farms) were significant within each of the three countries (P, 0.05) but not across countries (P. 0.05), but the frequencies of two practices (preparation of eggs on farms and storage of feeds) were significantly different (P, 0.05) both within and across countries. Practices used to prevent disease on farms. The frequency of disease preventive measures is displayed in Table 2. Of the 10 practices investigated, the frequencies of 4 (vaccine administered, sanitation procedure, types of pest control, and biosecurity measures) were significantly different (P, 0.05) within and across the three countries. Risk factors for Salmonella infection in layers, feed, and farm environments. Table 3 shows the frequency of isolation of Salmonella based on pooled data from layers, feeds, and farm environments in the three countries. Of the eight factors studied, farm size was the only one that significantly affected (P ~ 0.031) the frequency of isolation of Salmonella. Isolation of Salmonella from layer farms. In Trinidad and Tobago, the frequency of detection of Salmonella on poultry farms and their environments was significantly higher (P ~ 0.024) on large farms (100.0%) than on medium farms (0.0%) and small farms (33.3%); however, these differences were not significant in Grenada (P ~ 0.400) and St. Lucia (P ~ 0.275) (Table 4). The overall frequency of isolation of Salmonella was 40.0% (14 of 35
1474 ADESIYUN ET AL. J. Food Prot., Vol. 77, No. 9 TABLE 1. Demographic data and management and production practices on layer farms in three countries Trinidad and Tobago Grenada St. Lucia Parameter No (%) P value No (%) P value No (%) P value All countries (P value) farms 23 21 37 layers on farm 0.637 0.012 0.001 0.000,1,000 2 (8.7) 12 (57.1) 13 (35.1) 1,001 5,000 6 (26.1) 8 (38.1) 17 (45.9) 5,001 10,000 4 (17.4) 1 (4.8) 4 (10.8) 10,000 20,000 4 (17.4) 0 0.20,000 6 (26.1) 0 0 No response 1 (4.3) 0 1 (2.7) Breed of layer hens used 0.043 0.228 0.001 0.000 Hyline Brown 4 (17.4) 7 (33.3) 15 (40.5) Bovan Brown 3 (13.0) 4 (19.0) 0 ISA Brown 5 (21.7) 0 16 (43.2) Mixed 10 (43.5) 8 (38.1) 3 (8.1) Other or no response 1 (4.3) 2 (9.5) 3 (8.1) Other animals on farm a 0.144 0.050 0.869 0.187 Yes 15 (65.2) 15 (71.4) 18 (48.6) No 8 (34.8) 6 (28.6) 19 (51.4) Housing system 0.000 0.000 0.000 0.345 Cages 1 (4.3) 0 0 Pens, deep litter 20 (87.0) 21 (100.0) 34 (91.9) Mixture 2 (8.7) 0 3 (8.1) Age groups present on farm 0.002 0.000 0.732 0.004 One age group 4 (17.4) 2 (9.5) 16 (43.2) Several age groups 19 (82.6) 19 (90.5) 18 (48.6) No response 0 0 3 (8.1) Supplier of chicks 0.144 0.009 0.441 0.000 Local 8 (34.8) 3 (14.3) 15 (40.5) Imported from the United States 15 (65.2) 1 (4.8) 4 (10.8) Imported from the Caribbean region 0 9 (42.9) 3 (8.1) Local and imported from the Caribbean region 0 6 (28.6) 0 Imported from both the Caribbean region and the United States 0 1 (4.8) 6 (16.2) Others or no response 0 1 (4.8) 9 (24.3) Age pullets start to lay (wk) 0.022 0.000 0.000 0.057 16 17 6 (26.1) 1 (4.8) 3 (8.1) 18 24 17 (73.9) 20 (95.2) 30 (81.1) 25 30 0 0 3 (8.1) No response 0 0 1 (2.7) Age at culling of layers (mo) 0.000 0.000 0.000 0.412 12 13 0 1 (4.8) 0 14 17 1 (4.3) 1 (4.8) 3 (8.1) 18 24 22 (95.7) 17 (81.0) 30 (81.1).25 or no response 0 2 (9.5) 4 (10.8) Preparation of eggs on farm 0.000 0.001 0.000 0.191 Yes 21 (91.3) 18 (85.7) 33 (89.2) No 2 (8.7) 3 (14.3) 1 (2.7) No response 0 0 3 (8.1) Methods used to prepare eggs for sale b 0.000 0.000 0.000 0.000 Dry clean by hand 10 (47.6) 0 21 (58.3) Wiping with damp cloth 3 (14.3) 20 (95.2) 5 (13.9) Machine wash with warm water 1 (4.8) 0 0 Washed with water 2 (9.5) 0 0 Wipe with bleach 1 (4.8) 1 (4.8) 0 Wash with water and hand sanitizer 1 (4.8) 0 1 (2.8) Other or no response c 3 (14.3) 0 10 (27.8) Destination of eggs from farm d 0.022 0.275 0.040 0.000 Supermarkets 21 (91.3) 8 (38.1) 11 (29.7) Community 3 (13.0) 12 (57.1) 18 (48.6)
J. Food Prot., Vol. 77, No. 9 SALMONELLA IN LAYER FARMS IN THE CARIBBEAN 1475 TABLE 1. Continued Trinidad and Tobago Grenada St. Lucia Parameter No (%) P value No (%) P value No (%) P value All countries (P value) Hotel 0 2 (9.5) 11 (29.7) Other or no response 4 (17.4) 6 (28.6) 8 (21.6) Supplier of feeds NA e NA 0.000 0.032 Local 23 (100.0) 21 (100.0) 29 (78.4) Imported 0 0 7 (18.9) Mixture of local and imported 0 0 1 (2.7) Storage of feeds 0.000 0.000 0.000 0.010 Store room or shed 16 (69.7) 17 (81.0) 34 (91.9) Other 6 (26.1) 1 (4.8) 1 (2.7) Pen 0 3 (14.3) 2 (5.4) No response 1 (4.3) 0 0 a Trinidad and Tobago: pigs (13.3%), dogs (86.7%), and cats (40.0%). Grenada: pigs (14.3%), dogs (47.6%), cats (19.0%), cows (14.3%), sheep (9.5%), and goats (28.6%). St. Lucia: pigs (22.2%), dogs (66.7%), cats (27.8%), cows (5.6%), sheep (22.2%), and goats (27.8%). b For Trinidad and Tobago, only 21 farms were asked for responses to this question. c Trinidad and Tobago: wash, oil, and crate (4.8%); dry clean, oil, and grade (9.6%). St. Lucia: wash, oil, and crate (2.8%); clean, pack, and distribute (13.9%); preclean with steel wool and wash in egg washer (2.8%); no response (8.3%). d Eggs from some farms had multiple destinations. e NA, not applicable because the variable was constant and the chi-square test could not be performed. farms) for the three countries, with a range of 26.1% (small farms) to 77.8% (large farms), and this difference was significant (P ~ 0.025). Frequency of isolation of Salmonella by source on layer farms. Within each of the three countries, the frequency of isolation of Salmonella was higher in environmental samples than in cloacal swabs and feeds, but the differences were significant only in Trinidad and Tobago (P ~ 0.000) and St. Lucia (P ~ 0.000) (Table 5). Although the frequency of isolation of Salmonella was highest for cloacal swabs in St. Lucia (1.6%), for environmental swabs in Trinidad and Tobago (26.7%), and for feeds in Trinidad and Tobago (6.5%), the differences for each type of sample across countries were not significant (P. 0.05). Pooled samples of freshly laid eggs (6 eggs per pooled sample) were Salmonella only in Trinidad and Tobago, with a frequency of 12.5% (3 of 24 pooled samples [18 of 144 eggs]), and all Salmonella isolates were recovered from egg shell surfaces (Table 6). All 12 pooled samples (72 eggs) from Grenada and 23 pooled samples (138 eggs) from St. Lucia were negative for Salmonella. Frequency of isolation of Salmonella from samples collected at ports of entry (airports). All 200 samples collected from the airports (fecal swabs of chicks, bedding paper in shipment boxes, and eggs) were negative for Salmonella. Frequency of isolation of Salmonella from a hatchery in St. Lucia. All 38 samples, i.e., 6 pooled egg samples, 6 egg shell samples, 20 meconium samples, and 6 environmental samples, were negative for Salmonella. Frequency of isolation of Salmonella from table eggs and kitchen environments of hotels. All samples collected from hotel kitchens in all three countries were negative for Salmonella. Serotypes of Salmonella isolates recovered from various sources. For cloacal samples from layers, the Salmonella serotypes recovered were Anatum 3,10:e,h:1,6 in Trinidad and Tobago, group C in Grenada, and Kentucky (8),20:i:z 6 in St. Lucia (Table 6). Three Salmonella serotypes (Mbandaka 6,7:z 10 :e,n,z 15, Polyvalent A negative, and Montevideo 6,7:g,m,s:) were isolated from freshly laid eggs on layer farms only in Trinidad and Tobago. Eight Salmonella serotypes were isolated from environmental swabs in Trinidad and Tobago, and Anatum 3,10:e,h:1,6 was dominant. In St. Lucia, the dominant Salmonella serotype was Kentucky (8),20:i:z 6. From all sources in all the three countries, 35 Salmonella isolates were recovered, only one of which (2.9%) was identified as Salmonella Enteritidis 1,9,12:g,m:, which represented only 6.6% of the total 15 serotypes isolated. DISCUSSION The demographic data revealed that the layer poultry industry in these three countries were similar but with some differences that could affect the exposure of farms to Salmonella. For example, farms in Trinidad and Tobago were considerably larger than those in the other two countries, and the isolation rate for Salmonella was significantly higher on large farms (77.7%) than on medium and small farms (26.9%). This finding is in agreement with published reports in which poultry farm size was reported as a risk factor for Salmonella infection (12, 36). In Trinidad and Tobago, a majority of the farms imported day-old chicks from the United States, whereas farmers in Grenada and St. Lucia imported chicks primarily
1476 ADESIYUN ET AL. J. Food Prot., Vol. 77, No. 9 TABLE 2. Farm practices used to prevent disease in three countries Trinidad and Tobago Grenada St. Lucia Practice No. (%) P value No. (%) P value No. (%) P value All countries (P value) farms 23 21 37 Vaccine administered a 0.000 0.000 0.000 0.000 Yes 20 (87.0) 1 (4.8) 1 (2.7) No 2 (8.7) 20 (95.2) 35 (94.6) Do not know or no response 1 (4.3) 0 1 (2.7) Location of pullets 0.007 0.000 0.000 0.132 Transferred to a laying unit 5 (21.7) 7 (33.3) 3 (8.1) Stay in same pen, cage, area 18 (78.3) 13 (61.9) 33 (89.2) Other or no response 0 1 (4.80 1 (2.7) Sanitization procedures 0.000 0.000 0.000 0.000 Change litter 1 (4.3) 1 (4.8) 2 (5.4) Change litter and wash 0 0 25 (67.6) Change litter and clean with disinfectant or bleach 16 (69.6) 19 (90.5) 6 (16.2) Other or no response 6 (26.1) 1 (4.8) 4 (10.8) Water supply NA b 0.000 0.000 0.356 Government pipe 23 (100.0) 20 (95.2) 36 (97.3) Rain 0 2 (9.5) 3 (8.1) Water system 0.003 NA 0.000 0.000 Automatic 15 (65.2) 21 (100.0) 2 (5.4) Manual 5 (21.7) 0 31 (83.8) Automatic and manual 0 0 1 (2.7) Other or no response 3 (13.0) 0 3 (8.1) Presence of pests 0.000 0.000 0.732 0.000 Yes 20 (87.0) 20 (95.2) 18 (48.6) No 3 (13.0) 1 (4.8) 16 (43.2) No response 0 0 3 (8.1) Type of pests present c 0.002 0.000 0.059 0.130 Rats 18 (90.0) 18 (90.0) 14 (77.8) Birds 8 (40.0) 12 (60.0) 6 (33.3) Mice 0 3 (15.0) 7 (38.9) Other 2 (10.0) 3 (15.0) 3 (16.7) Type of pest control 0.000 0.005 0.003 0.000 Poison 1 (5.0) 19 (95.0) 5 (27.8) Bait 17 (85.0) 2 (10.0) 3 (16.7) Other 3 (15.0) 3 (15.0) 9 (50.0) Biosecurity measures 0.003 0.000 0.001 0.000 Foot dips 8 (34.8) 0 4 (10.8) Restricted access 18 (78.3) 20 (95.2) 5 (13.5) Protective clothing 1 (4.3) 1 (4.8) 21 (56.8) Other or no response 1 (4.3) 0 8 (21.6) Disease outbreak in last 10 yr 0.000 0.000 0.001 0.580 Yes 2 (8.7) 1 (4.8) 1 (2.7) No 21 (91.3) 20 (95.2) 36 (97.3) a Vaccination against Marek s disease, Newcastle disease, pox, bursal disease, bronchitis, and egg drop syndrome. b NA, not applicable because the variable was constant and the chi-square test could not be performed. c For Trinidad and Tobago, Grenada, and St. Lucia, only 20, 20, and 18 farms, respectively, responded to this question. from regional suppliers. These differences in sources of dayold chicks can affect the introduction of Salmonella into layer farms in these three countries. However, all day-old chicks and table eggs imported from the United States and regional sources were negative for Salmonella. Imported day-old chicks have been responsible for the introduction of Salmonella into other countries, with carriage rates of 12.6 to 19.2% (6, 31). The overall frequency of isolation of Salmonella from the cloacae of laying birds was relatively low (0.9%) in these three countries, which was considerably lower than the range (4 to 12%) reported for other countries (3, 16). Swabs of litter in pens on layer farms in Trinidad and Tobago had significantly higher frequency (26.7%) of isolation of Salmonella than did those in Grenada (6.5%) and St. Lucia (16.1%), which is an indication that freshly
J. Food Prot., Vol. 77, No. 9 SALMONELLA IN LAYER FARMS IN THE CARIBBEAN 1477 TABLE 3. Risk factors for Salmonella on all farms in Trinidad and Tobago, Grenada, and St. Lucia Risk factor farms tested No. (%) of farms Salmonella P value Farm size a 0.031 Large 9 7 (77.8) Medium 3 1 (33.3) Small 23 6 (26.1) Age groups of layers on 0.627 farms One age group only 5 1 (20.0) Several age groups 30 13 (43.3) Source of day old chicks 0.156 United States 15 7 (46.7) Caribbean region 14 3 (21.4) Other 6 4 (66.7) Vaccination b 0.206 Yes 11 6 (54.5) No 24 8 (33.3) Sanitization procedures 0.119 Change litter 2 0 Change litter and wash 4 2 (50.0) Change litter and clean with disinfectant, bleach 14 3 (18.8) Other or no response c 15 9 (60.0) Presence of pests 0.321 Yes 30 13 (43.3) No 5 1 (20.0) Biosecurity measures 0.069 Restricted access 17 7 (41.2) Protective clothing (aprons, boots, gloves) 2 0 Multiple (various combinations of the above) 10 2 (20.0) None 4 3 (75.0) Other 2 2 (100.0) Disease outbreaks in last 10 yr Yes 3 2 (66.7) No 32 12 (37.5) 0.348 a Large,.10,000 birds; medium, 5,001 to 10,000 birds; small,,5,000 birds. b Vaccination against Marek s disease, Newcastle disease, pox, bursal disease, bronchitis, and egg drop syndrome. c Includes a combination of practices. laid eggs in Trinidad and Tobago have a higher risk of being contaminated by Salmonella. This risk was confirmed by the finding that in only Trinidad and Tobago were the shells of freshly laid eggs contaminated with Salmonella. Wood shavings and sand used as litter on layer farms may be contaminated by pests such as rodents (22, 23, 38) and could have occurred in Trinidad and Tobago, where 90% of the farms reported a rat problem. The need for good biosecurity practices and disinfection cannot be overemphasized, particularly between introducing new batches of birds to the pens to avoid residual contamination from the preceding flocks (17). The overall frequency of isolation of TABLE 4. Frequency of isolation of Salmonella from layer farms in three countries Farm size Total no. of farms No. (%) of farms sampled No. (%) of farms Salmonella a P value Trinidad and Tobago Large 5 5 (100.0) Medium 2 0 Small 3 1 (33.3) Total 23 10 (43.5) 6 (60.0) 0.024 Grenada Large 0 NA Medium 1 1 (100.0) Small 9 3 (33.3) Total 21 10 (47.6) 4 (40.0) 0.400 St. Lucia Large 4 2 (50.0) Medium 0 NA Small 11 2 (18.2) Total 37 15 (40.5) 4 (26.7) 0.275 All countries Large 9 7 (77.8) Medium 3 1 (33.3) Small 23 6 (26.1) Total 81 35 (43.2) 14 (40.0) 0.025 a Based on results from cloacal swabs, eggs, feeds, and environmental samples from the layer farms. NA, not applicable. Salmonella from farm environments in the current study (17.8%) is within the range of 0.95 to 33.3% reported for other countries (6, 34). The finding that only the pooled egg shells (12.5%) and not the pooled egg contents yielded Salmonella in Trinidad and Tobago suggests horizontal rather than transovarian transfer. This finding differs from those of two previous studies conducted in Trinidad and Tobago. Indar et al. (20) isolated Salmonella from 4.7% of egg shell samples and 1.2% of egg content samples, and Adesiyun et al. (2) reported Salmonella isolation rates of 6.5% for both pooled egg shells and egg contents. Because the same isolation techniques were used in all three studies, the differences found should not have been due to the isolation protocols, although differences in techniques can affect the recovery rates and serotypes of Salmonella (5). Therefore, the increase in horizontal transmission of Salmonella in in Trinidad and Tobago appears to be real. Reports elsewhere have documented contamination rates of eggshells at 0 to 34% (3, 6, 28). All 31 pooled samples of eggs (shells and egg content) from 19 hotel kitchens in the three countries were negative for Salmonella, which is important for food safety. Although 12.5% of the freshly laid eggs sampled on farms in Trinidad and Tobago were Salmonella, all table eggs sampled from hotel kitchens in that country were negative for the pathogen. This difference could be in part due to the practice of preparing eggs (cleaning and sometimes disinfecting) for sale to hotels, supermarkets, and other destinations. Another possibility is that the farms
1478 ADESIYUN ET AL. J. Food Prot., Vol. 77, No. 9 TABLE 5. Frequency of isolation of Salmonella by source on layer farms in three countries Trinidad and Tobago Grenada St. Lucia All countries (P value) No. (%) samples Salmonella P value samples tested No. (%) samples Salmonella P value samples tested No. (%) samples Salmonella P value samples tested Sample type Cloacal swabs of layers 454 1 (0.2) 227 2 (0.9) 437 7 (1.6) 0.091 Environmental swabs 45 12 (26.7) 31 2 (6.5) 31 5 (16.1) 0.074 Feed 31 2 (6.5) 13 0 20 0 0.333 Total 530 15 (2.8) 0.000 271 4 (1.5) 0.111 488 12 (2.5) 0.000 0.494 with Salmonella-positive eggs did not supply eggs to the hotels that provided eggs for this study. The low prevalence (12.5%) may reflect a reduction of contamination of eggs on poultry farms in Trinidad and Tobago, this hypothesis must be confirmed with further work because the present prevalence study provided information on the contamination of eggs at only a single time point. Salmonellosis associated with contaminated eggs and egg products has been reported in the Caribbean region (8, 21) and elsewhere (11, 19, 26, 30, 32). However, sample size, isolation techniques used, and time frame should be considered when comparing prevalence data from various studies. Feeds, particularly as part of a deep litter management system, can be a source of contamination of eggs laid in the litter; 6.5% of feed samples from Trinidad and Tobago were Salmonella. Other researchers have recovered Salmonella from feeds provided for layers on farms at frequencies of 0.1 to 5.6% (6, 17, 34), indicating that feed can serve as a vehicle for Salmonella transmission on poultry farms. The recovery of distinctly different Salmonella serotypes from various poultry sources in these three countries has epidemiological relevance and may indicate the origins and spreading mechanisms of Salmonella serotypes to the layer farms. Eggborne human salmonellosis in these countries is most likely caused by these country-prevalent serotypes. The relatedness of the Salmonella isolates recovered from cloacae and environmental sources cannot be conclusively established because genetic fingerprinting was not performed. Adesiyun et al. (1) used pulse-field gel electrophoresis to determine the relationships among Salmonella Enteritidis isolates recovered from each of the Caribbean countries studied. Salmonella serotypes differ among layer farms and countries (9, 28, 29, 34), and all serotypes recovered in the present study have been documented previously in the Caribbean region (1, 10, 20). Only 2.9% (1 of 35) of all isolates were Salmonella Enteritidis, representing 6.7% of the 15 serotypes recovered in the current study. This finding is not in accordance with published reports from Trinidad and Tobago, in which this the serotype accounted for 58.3% of all the Salmonella isolates recovered from table eggs in this country (2), nor with reports from outside the region (25, 33, 39). This difference may reflect a change in production and management practices, resulting in a change in the epidemiology of this egg- and layer-associated serotype, which needs further investigation. However, the difference also might be an artifact of the limited time period over which the study was conducted and the small number of samples that were Salmonella. This survey revealed that the poultry farm environments in the three countries studied pose a risk for contaminating freshly laid eggs, although table eggs (local and imported) and imported day-old chicks appear to pose minimal risk for Salmonella transmission through consumption of eggs or the introduction of the pathogen into the respective countries. The very low frequency of detection of Salmonella Enteritidis from all sources indicates that this serotype may not be commonly associated with eggborne salmonellosis in the Caribbean region. These results revealed the presence of country-specific Salmonella serotypes associated with
J. Food Prot., Vol. 77, No. 9 SALMONELLA IN LAYER FARMS IN THE CARIBBEAN 1479 TABLE 6. Serotypes of Salmonella isolated from various sources in three countries a Trinidad and Tobago Grenada St. Lucia Sample type No. (%) of samples Salmonella Salmonella serotypes isolates No. (%) of samples Salmonella Salmonella serotypes isolates No. (%) of samples Salmonella Salmonella serotypes isolates Cloacal swabs of layers 1 Anatum 3,10:e,h:1,6 1 2 Group C 2 7 Kentucky (8),20:i:z 6 7 Eggs on the farm b 3 Mbandaka 6,7:z10:e,n,z15 1 0 NA 0 NA Polyvalent A negative 1 Montevideo 6,7:g,m,s: 1 Environmental swabs 13 Groups I O (1) 1 2 Group C 1 5 Kentucky (8),20:i:z6 3 Enteritidis 1,9,12:g,m: 1 Group D 1 Group B 2 Anatum 3,10:e,h:1,6 4 Group E 1 Caracas 6,14.(25):g,m,s 2 Braenderup 1 Uganda 3,10:l,z 13 :1,5 2 Muenster 3,10:e,h:1,5 1 Feed 2 Anatum 3,10:e,h:1,6 1 0 NA 0 NA Uganda 3,10:l,z 13 :1,5 1 Total 19 4 12 a Salmonella isolates were recovered from different farms. b Pooled samples of freshly laid eggs (6 eggs per pooled sample) were tested for Salmonella. In Trinidad and Tobago, 3 of 24 pooled samples (18 of 144 eggs) were positive (frequency of 12.5%), and all Salmonella isolates were recovered from egg shell surfaces. None of the 12 pooled samples (72 eggs) from Grenada and none of the 23 pooled samples (138 eggs) from St. Lucia were Salmonella.
1480 ADESIYUN ET AL. J. Food Prot., Vol. 77, No. 9 poultry farms in these three countries, and these serotypes may play a role in salmonellosis in humans and animals. ACKNOWLEDGMENTS The authors are grateful to the Caribbean Food Research Institute (CFRI), Jamaica, and the Pan American Health Organization (PAHO) (grant 65503 52163) for funding the project. The assistance provided by Ms. Margaret Kalloo (Caribbean Community Secretariat) by circulating the questionnaire to the regional chief veterinary officers is appreciated. We thank the CVOs of Trinidad and Tobago, Grenada, and St. Lucia, who assisted in sample collection. REFERENCES 1. Adesiyun, A. A., A. Carson, K. McAdoo, and C. Bailey. 2001. Molecular analysis of Salmonella Enteritidis isolates from the Caribbean region by pulse field gel electrophoresis. Pan Am. J. Public Health 8:342 347. 2. Adesiyun, A. A., V. N. Offiah, N. Seepersadsingh, S. Rodrigo, V. Lashley, L. Musai, and K. Georges. 2005. Microbial health risk posed by table eggs in Trinidad. Epidemiol. Infect. 133:1049 1056. 3. Ammar, A., N. Alloui, O. Bennoune, and A. Kassah-Laouar. 2010. Survey of Salmonella serovars in broilers and laying breeding reproducers in East of Algeria. J. Infect. Dev. Ctries. 4:103 106. 4. Andrews, W. 1992. Manuals of food quality control. 4. Microbiological analysis. FAO food and nutrition paper 14/4, rev. 1, p. 27 48. Food and Agriculture Organization, Rome. 5. Arnaut-Rollier, I., L. De Zutter, and J. Van Hoof. 1993. Comparison of four isolation methods for the recovery of Salmonella from freshly slaughtered poultry, p. 505 509. In Proceedings of the 11th International Symposium of the World Association of Veterinary Food Hygienists, Bangkok, Thailand, 24 to 29 October. 6. Barbour, E. K., and N. H. Nabbut. 1982. Isolation of Salmonella and other potential pathogens from two chicken breeding farms in Saudi Arabia. Avian Dis. 26:234 244. 7. Blumenthal, D. 1990. Salmonella Enteritidis: from the chicken to the egg. FDA Consumer 24(3):6 10. 8. Caribbean Epidemiology Centre. 1997. Annual bacteriology reports 1986 1997, p. 42 52. In CAREC annual reports. Caribbean Epidemiology Centre, Port of Spain, Trinidad and Tobago. 9. Caribbean Epidemiology Centre. 2002. Annual report, 2002. Caribbean Epidemiology Centre, Port of Spain, Trinidad and Tobago. 10. Caribbean Epidemiology Centre. 2003. Annual report, 2003. Caribbean Epidemiology Centre, Port of Spain, Trinidad and Tobago. 11. Centers for Disease Control. 1990. Update: Salmonella Enteritidis infections and shell eggs in the United States. Morb. Mortal. Wkly. Rep. 39:909 912. 12. Chemaly, M., A. Huneau-Salaün, A. Labbe, C. Houdayer, and P. Fravalo. 2009. Isolation of Salmonella enterica in laying-hen flocks and assessment of eggshell contamination in France. J. Food Prot. 72:2071 2077. 13. Chen, S. J., T. E. Lee, E. M. Wang, T. J. Cho, and C. H. Wang. 2002. Monitoring the hygiene of chicken hatcheries in Taiwan during 1999 2001. J. Microbiol. Immunol. Infect. 35:236 242. 14. Davies, R., and M. Breslin. 2001. Environmental contamination and detection of Salmonella enterica serovar Enteritidis in laying flocks. Vet. Rec. 149:699 704. 15. Davies, R., E. Liebana, and M. Breslin. 2003. Investigation of the distribution and control of Salmonella enterica serovar Enteritidis PT6 in layer breeding and egg production. Avian Pathol. 32:225 237. 16. Garcia, C., J. M. Soriano, V. Benitez, and P. Catalá-Gregori. 2011. Assessment of Salmonella spp. in faeces, cloacal swabs, and eggs (egg shell and content separately) from a layer hen farm. Poult. Sci. 90:1581 1585. 17. Hacking, W. C., W. R. Mitchell, and H. C. Carson. 1978. Sources of salmonellae in broiler chickens in Ontario. Can. J. Comp. Med. 42: 392 399. 18. Hang ombe, B. M., R. N. Sharma, E. Skjerve, and L. M. Tuchili. 1999. Occurrence of Salmonella Enteritidis in pooled table eggs and market ready chicken carcasses in Zambia. Avian Dis. 43:597 599. 19. Hrivniaková, L.,D.Schmid,A.Luckner-Hornischer,H.Lassnig,C. Kornschober, J. Angermayer, and F. Allerberger. 2011. Salmonellosis outbreak due to Salmonella Enteritidis phage type 14b resistant to nalidixic acid, Austria, September 2010. Euro. Surveill. 16:pii~19952. 20. Indar, L., G. Baccus-Taylor, E. A. Comissiong, P. Prabhakar, and H. Reid. 1998. Salmonellosis in Trinidad: evidence of transovarian transmission of Salmonella in farm eggs. West Indian Med. J. 47: 50 53. 21. Indar-Harrinauth, L., N. Daniels, P. Prabhakar, C. Brown, G. Baccus- Taylor, E. Comissiong, and J. Hospedales. 2001. Emergence of Salmonella Enteritidis phage type 4 in the Caribbean: case-controlled study in Trinidad and Tobago, West Indies. Clin. Infect. Dis. 32:890 896. 22. Lapuz, R., H. Tani, K. Sasai, K. Shirota, H. Katoh, and E. Baba. 2008. The role of roof rats (Rattus rattus) in the spread of Salmonella Enteritidis and S. Infantis contamination in layer farms in East Japan. Epidemiol. Infect. 136:1235 1243. 23. Lapuz, R. R., D. V. Umali, T. Suzuki, K. Shiroto, and H. Katoh. 2012. Comparison of the prevalence of Salmonella infection in layer hens from commercial layer farms with high and low rodent densities. Avian Dis. 56:29 34. 24. MacFaddin, J. F. 2002. Biochemical tests for identification of medical bacteria. Williams and Wilkins, New York. 25. Mishu, B., J. Roehler, L. A. Lee, D. Rodrigue, F. Hickman Brenner, P. Blake, and R. V. Tauxe. 1994. Outbreak of Salmonella Enteritidis infections in the United States, 1985 1991. J. Infect. Dis. 169:547 552. 26. Much, P., C. Berghold, G. Krassnig, H. Schweighardt, H. Wenzl, and F. Allerberger. 2005. An Austrian outbreak of Salmonella Enteritidis phage type 36 in 2004. Wien. Klin. Wochenschr. 117:599 603. 27. Mueller-Doblies, D., C. Clouting, and R. H. Davies. 2013. Investigations of the distribution and persistence of Salmonella and ciprofloxacin-resistant Escherichia coli in turkey hatcheries in the UK. Zoonoses Public Health 60:296 303. 28. Murchie, L., P. Whyte, B. Xia, S. Horrigan, L. Kelly, and R. H. Madden. 2007. Prevalence of Salmonella in grade A whole shell eggs in the island of Ireland. J. Food Prot. 70:1238 1240. 29. Musgrove, M. T., J. D. Shaw, and M. A. Harrison. 2012. Salmonella collected from nest run cart shelves in commercial egg processing facilities. Poult. Sci. 91:2386 2389. 30. Oboegulem, S. I., P. W. Collier, J. S. Sharp, and W. J. Reilly. 1993. Epidemiological aspects of outbreaks of food-borne salmonellosis in Scotland between 1980 and 1989. Rev. Sci. Tech. 12:957 967. 31. Osman, K. M., A. M. Yousef, M. M. Aly, and M. I. Radwan. 2010. Salmonella spp. infection in imported 1-day chicks, ducklings, and turkey poults: a public health risk. Foodborne Pathog. Dis. 7:383 390. 32. Peresi, J. T., I. A. Almeida, S. I. Lima, D. F. Marques, E. C. Rodrigues, S. A. Fernandes, D. S. Gelli, and K. Irino. 1998. Food borne disease outbreaks caused by Salmonella Enteritidis. Rev. Saude Publica 32:477 483. (In Portuguese.) 33. Poppe, C. 1994. Salmonella Enteritidis in Canada. Int. J. Food Microbiol. 21:1 5. 34. Shirota, K., D. V. Umali, T. Suzuki, and H. Katoh. 2012. Epizootiologic role of feeds in the epidemiology of Salmonella Senftenberg contamination in commercial layer farms in eastern Japan. Avian Dis. 56:516 520. 35. Sivaramalingam, T., S. A. McEwen, D. L. Pearl, D. Ojkic, and M. T. Guerin. 2013. A temporal study of Salmonella serovars from environmental samples from poultry breeder flocks in Ontario between 1998 and 2008. Can. J. Vet. Res. 77:1 11. 36. Snow, L. C., R. H. Davies, K. H. Christiansen, J. J. Carrique-Mas, A. J. C. Cook, and S. J. Evans. 2012. Investigation of risk factors of Salmonella on commercial egg-laying farms in Great Britain, 2004 2006. Vet. Rec. 166:579 586. 37. Thrusfield, M. V. 2005. Veterinary epidemiology. Blackwell Science, Oxford. 38. Umali, D. V., R. R. Lapuz, T. Suzuki, K. Shirota, and H. Katoh. 2012. Transmission and shedding patterns of Salmonella in naturally infected captive wild roof rats (Rattus rattus) from a Salmonellacontaminated layer farm. Avian Dis. 56:288 294. 39. Van Loock, F., G. Ducoffre, J. M. Dumont, M. L. Libotte-Chasseur, H. Imberechts, M. Gouffaux, J. Houins-Roulet, G. Lamsens, K. De Schrijver, N. Bin, A. Moreau, L. De Zutter, and G. Daube. 2000. Analysis of foodborne disease in Belgium. Acta Clin. Belg. 55:300 306.