J. A. Cason

Broiler Carcass Contamination with Campylobacter from Feces during Defeathering

Three sets of experiments were conducted to explore the increase in recovery of Campylobacter from broiler carcasses after defeathering. In the first set of experiments, live broilers obtained from a commercial processor were transported to a pilot plant, and breast skin was sampled by a sponge wipe method before and after defeathering. One of 120 broiler breast skin samples was positive for Campylobacter before defeathering, and 95 of 120 were positive after defeathering. In the second set of experiments, Campylobacter-free flocks were identified, subjected to feed withdrawal, and transported to the pilot plant. Carcasses were intracloacally inoculated with Campylobacter (10(7) CFU) just prior to entering the scald tank. Breast skin sponge samples were negative for Campylobacter before carcasses entered the picker (0 of 120 samples). After defeathering, 69 of 120 samples were positive for Campylobacter, with an average of log10 2.7 CFU per sample (approximately 30 cm2). The third set of experiments was conducted using Campylobacter-positive broilers obtained at a commercial processing plant and transported live to the pilot plant. Just prior to scalding, the cloacae were plugged with tampons and sutured shut on half of the carcasses. Plugged carcasses were scalded, and breast skin samples taken before and after defeathering were compared with those collected from control broilers from the same flock. Prior to defeathering, 1 of 120 breast skin sponge samples were positive for the control carcasses, and 0 of 120 were positive for the plugged carcasses. After passing through the picker, 120 of 120 control carcasses had positive breast skin sponge samples, with an average of log10 4.2 CFU per sample (approximately 30 cm2). Only 13 of 120 plugged carcasses had detectable numbers of Campylobacter on the breast skin sponge, with an average of log10 2.5 CFU per sample. These data indicate that an increase in the recovery of Campylobacter after defeathering can be related to the escape of contaminated feces from the cloaca during defeathering.

Effect of Fecal Contamination and Cross-Contamination on Numbers of Coliform, Escherichia coli, Campylobacter, and Salmonella on Immersion-Chilled Broiler Carcasses

The effect of prechill fecal contamination on numbers of bacteria on immersion-chilled carcasses was tested in each of three replicate trials. For each trial, 16 eviscerated broiler carcasses were split into 32 halves and assigned to one of two groups. Cecal contents (0.1 g inoculated with Campylobacter and nalidixic acid-resistant Salmonella) were applied to each of eight halves in one group (direct contamination) that were placed into one paddle chiller (contaminated), whereas the other paired halves were placed into another chiller (control). From the second group of eight split birds, one of each paired half was placed in the contaminated chiller (to determine cross-contamination) and the other half was placed in the control chiller. Postchill carcass halves were sampled by a 1-min rinse in sterile water, which was collected and cultured. Bacterial counts were reported as log CFU per milliliter of rinsate. There were no significant statistical differences (paired t test, P < 0.05) from direct contamination for coliforms (mean 3.0 log CFU) and Escherichia coli (mean 2.7 log CFU), although Campylobacter numbers significantly increased from control values because of direct contamination (1.5 versus 2.1 log CFU), and the incidence increased from 79 to 100%. There was no significant effect of cross-contamination on coliform (mean 2.9 log CFU) or E. coli (mean 2.6 log CFU) numbers. Nevertheless, Campylobacter levels were significantly higher after exposure to cross-contamination (1.6 versus 2.0 log CFU), and the incidence of this bacterium increased from 75 to 100%. Salmonella-positive halves increased from 0 to 42% postchill because of direct contamination and from 0 to 25% as a result of cross-contamination after chilling. Water samples and surface swabs taken postchill from the contaminated chiller were higher for Campylobacter than those taken from the control chiller. Immersion chilling equilibrated bacterial numbers between contaminated and control halves subjected to either direct contamination or cross-contamination for coliforms and E. coli. Campylobacter numbers, Campylobacter incidence, and Salmonella incidence increased because of both direct contamination and cross-contamination in the chiller. Postchill E. coli numbers did not indicate which carcass halves were contaminated with feces before chilling.

Recovery of Salmonella from retail broilers by a whole-carcass enrichment procedure.

Fresh whole broiler carcasses were purchased from grocery stores over a 20-week period. Carcasses were selected on the basis of their having intact packages and unique U.S. Department of Agriculture (USDA) plant numbers and sell-by dates, such that each bird represented a single processing plant-processing day combination. Carcasses were tested for Salmonella with a rinse aliquot obtained after whole-bird incubation in the rinse media for 24 h. On the basis of the number of unique processing plants (USDA plant numbers) and expiration dates involved, the number of birds available each week ranged from 6 to 17. Over the 20-week period, 251 independent carcasses from 14 processing plants were tested. The percentages of carcasses testing positive for Salmonella ranged from 0 (for 1 week) to >60% (for 3 weeks). For only 4 of the 20 weeks was an incidence of Salmonella-positive carcasses of <20% found. For the entire 20-week study, 85 (33.9%) of the 251 carcasses tested were found to be Salmonella positive. For those processing plants from which >10 carcasses were obtained, the percentages of carcasses testing positive for Salmonella ranged from <20 (two plants) to >40% (four plants). These results indicate that a whole-carcass enrichment may be more sensitive for the detection of Salmonella-positive carcasses than the traditional whole-carcass rinse followed by immediate testing of a subsample aliquot when small numbers of Salmonella are expected.

Coliform, Escherichia coli, and Salmonellae Concentrations in a Multiple-Tank, Counterflow Poultry Scalder

Scald water samples from a commercial broiler processing plant were tested for coliforms, Escherichia coli, and salmonellae to evaluate the numbers of suspended bacteria in a multiple-tank, counterflow scalder. Water samples were taken from each of three tanks on 8 different days after 6-week-old broilers had been processed for 8 h. Coliforms and E. coli were counted using Petrifilm, and the most probable number (MPN) of salmonellae was determined both in water samples and in rinses of defeathered carcasses that were removed from the processing line immediately after taking the water samples. Mean coliform concentrations in tanks 1, 2, and 3 (the last tank that carcasses pass through before being defeathered) were 3.4, 2.0, and 1.2 log10(CFU/ml), respectively. E. coli concentrations followed the same pattern with means of 3.2, 1.5, and 0.8 in tanks 1, 2, and 3, respectively, with significant differences (P < 0.02) in the concentrations of both coliforms and E. coli between the tanks. Sixteen of 24 scald-water samples were positive for salmonellae with a geometric mean of 10.9 MPN/100 ml in the positive samples. Salmonellae were isolated from seven of eight water samples from both tanks 1 and 2, but in only two of eight water samples from tank 3, the last tank that carcasses pass through. It appears that most bacteria removed from carcasses during scalding are washed off during the early part of scalding.

Transmission of Salmonella to broilers by contaminated larval and adult lesser mealworms, Alphitobius diaperinus (Coleoptera: Tenebrionidae)

The ability of the lesser mealworm, Alphitobius diaperinus (Panzer), commonly known as the darkling beetle, to transmit marker Salmonella Typhimurium to day-of-hatch broiler chicks was evaluated, as well as the spread to nonchallenged pen mates. In trial 1, day-of-hatch chicks were orally gavaged with 4 larval or 4 adult beetles that had been exposed to marker Salmonella-inoculated feed for 72 h. In addition, chicks were gavaged with the marker Salmonella in saline solution. These chicks were then placed into pens to serve as challenged broilers. In trial 2, all pens received 2 challenged chicks that were gavaged with larvae or beetles that had been exposed to marker Salmonella-inoculated feed for 24 h and then removed from the inoculated feed for a period of 7 d. At 3 wk of age, cecal samples from the marker Salmonella-challenged broilers and from 5 pen mates in trial 1, or 10 pen mates in trial 2, were evaluated for the presence of the marker Salmonella in their ceca, and at 6 wk of age, all remaining pen mates were sampled. To monitor the presence of the marker Salmonella within pens, stepped-on drag swab litter samples were taken weekly. For the Salmonella-saline pens, 29 to 33% of the broilers that had been challenged and 10 to 55% of the pen mates were positive at 3 wk of age, and only 2 to 6% had positive ceca at 6 wk. For the pens challenged with adult beetles, 0 to 57% of the challenged broilers and 20 to 40% of the pen mates had positive ceca at 3 wk, and 4 to 7% were positive at 6 wk. The pens challenged with larvae had the greatest percentage of marker Salmonella-positive broilers; 25 to 33% of the challenged broilers and 45 to 58% of pen mates were positive at 3 wk, and 11 to 27% were positive at 6 wk. These results demonstrated that ingestion of larval or adult beetles contaminated with a marker Salmonella could be a significant vector for transmission to broilers.

Effect of Prechill Fecal Contamination on Numbers of Bacteria Recovered from Broiler Chicken Carcasses Before and After Immersion Chilling

Paired carcass halves were used to test whether fecal contamination of skin during processing of broiler chickens can be detected by increased bacterial counts in samples taken before and after immersion chilling. In each of three trials, six freshly defeathered and eviscerated carcasses were cut in half, and a rectangle (3 by 5 cm) was marked with dots of ink on the breast skin of each half. One half of each pair was chosen randomly, and 0.1 g of freshly collected feces was spread over the rectangle with a spatula. After 10 min, both halves were sprayed with tap water for 10 to 15 s until feces could no longer be seen in the marked area. Both halves were sampled with a 1-min carcass rinse and were then put in a paddle chiller with other eviscerated carcasses for 45 min to simulate industrial immersion chilling. Immediately after chilling, each carcass half was subjected to another 1-min rinse, after which the skin within the rectangle was aseptically removed from the carcass halves and stomached. Rinses of fecally contaminated halves had significantly higher Enterobacteriaceae immediately before chilling, but there were no differences in coliform and Escherichia coli counts. After chilling, there were no differences in Enterobacteriaceae, coliform, and E. coli counts in rinse or skin samples from the paired carcass halves. Correlations were generally poor between counts in rinse and skin samples but were significant between prechill and postchill rinses for both control and fecally contaminated halves. Correlations were also significant between counts in rinses of control and contaminated halves of the same carcass after chilling. Bacterial counts in postchill carcass rinses did not indicate that fecal contamination occurred before chilling.

Comparison of Sampling Methods for the Detection of Salmonella on Whole Broiler Carcasses Purchased from Retail Outlets

An experiment was conducted to compare the effectiveness levels of two methods in recovering Salmonella from the same carcass. One hundred fresh whole broiler chickens were purchased from retail outlets over a 5-week period (20 carcasses per week). After carcasses had been aseptically removed from the packages and giblets had been removed, the carcasses were placed in sterile bags containing 400 ml of 1% buffered peptone water, the bags were shaken for 60 s, and a 30-ml aliquot was removed and incubated for 24 h at 37 degrees C (aliquot sample). Then, an additional 130 ml of 1% buffered peptone water was immediately added to the bag with the carcass (bringing the volume to 500 ml), the bag was reshaken, and the carcass and rinse were incubated for 24 h at 37 degrees C (whole-carcass enrichment sample). Following incubation, 0.5-ml samples for the two methods were placed into 10 ml of Rappaport-Vassiliadis broth and into 10 ml of tetrathionate (Hajna) broth and incubated at 42 degrees C for 24 h. Each broth was then streaked onto BG Sulfa agar and modified lysine iron agar and incubated for 24 h at 35 degrees C. Suspected Salmonella colonies were inoculated onto triple sugar iron and lysine iron agar slants and incubated at 35 degrees C for 24 h. Presumptive positive results were confirmed by Poly O and Poly H agglutination tests. Over the 5-week period, 13% of the aliquot samples tested positive for Salmonella, compared with 38% of the whole-carcass enrichment samples from the same carcasses. Recovery rates ranged from 0 of 20 samples to 4 of 20 samples for aliquot method and from 4 of 20 samples to 10 of 20 samples for the whole-carcass enrichment method over the 5-week period. These results indicate that when small numbers of Salmonella are expected, the sampling method has a major influence on the identification of Salmonella-positive carcasses.

Comparison of shell bacteria from unwashed and washed table eggs harvested from caged laying hens and cage-free floor-housed laying hens

These studies evaluated the bacterial level of unwashed and washed shell eggs from caged and cage-free laying hens. Hy-Line W-36 White and Hy-Line Brown laying hens were housed on all wire slats or all shavings floor systems. On the sampling days for experiments 1, 2, and 3, 20 eggs were collected from each pen for bacterial analyses. Ten of the eggs collected from each pen were washed for 1 min with a commercial egg-washing solution, whereas the remaining 10 eggs were unwashed before sampling the eggshell and shell membranes for aerobic bacteria and coliforms (experiment 1 only). In experiment 1, the aerobic plate counts (APC) of unwashed eggs produced in the shavings, slats, and caged-housing systems were 4.0, 3.6, and 3.1 log(10) cfu/mL of rinsate, respectively. Washing eggs significantly (P < 0.05) reduced APC by 1.6 log(10) cfu/mL and reduced the prevalence of coliforms by 12%. In experiment 2, unwashed eggs produced by hens in triple-deck cages from 57 to 62 wk (previously housed on shavings, slats, and cages) did not differ, with APC ranging from 0.6 to 0.8 log(10) cfu/mL. Washing eggs continued to significantly reduce APC to below 0.2 log(10) cfu/mL. In experiment 3, the APC for unwashed eggs were within 0.4 log below the APC attained for unwashed eggs in experiment 1, although hen density was 28% of that used in experiment 1. Washing eggs further lowered the APC to 0.4 to 0.7 log(10) cfu/mL, a 2.7-log reduction. These results indicate that shell bacterial levels are similar after washing for eggs from hens housed in these caged and cage-free environments. However, housing hens in cages with manure removal belts resulted in lower APC for both unwashed and washed eggs (compared with eggs from hens housed in a room with shavings, slats, and cages).

Partitioning of external and internal bacteria carried by broiler chickens before processing

Broiler chickens from the loading dock of a commercial processing plant were sampled to determine the incidence and counts of coliforms, Escherichia coli, and pathogenic bacteria. Feathers were removed by hand from ten 6-week-old chickens from each of seven different flocks and rinsed in 400 ml of 0.1% peptone water. Heads and feet were removed and rinsed, and the picked carcass was also rinsed, each in 200 ml. The ceca, colon, and crop were aseptically removed and stomached separately in 100 ml of peptone water. Campylobacter was present in six of the seven flocks. Salmonella was isolated from 50 of the 70 carcasses, with at least 2 positive carcasses in each flock, and five-tube most-probable-number (MPN) assays were performed on positive samples. Significantly (P < 0.05) more coliforms and E. coli were found in the ceca than in the feathers, which in turn carried more than the other samples, but total external and internal counts were roughly equivalent. Counts of Campylobacter were higher in the ceca and colon than in the other samples. Salmonella was isolated in external samples from 46 of the 50 positive carcasses compared with 26 positive internal samples or 17 positives in the ceca alone. The total MPN of Salmonella was approximately equivalent in all samples, indicating that contamination was distributed through all external and internal sampling locations. Salmonella-positive samples did not carry higher counts of coliforms or E. coli, and there were no significant correlations between the indicators and pathogens in any sample. Campylobacter numbers in the ceca were correlated with Campylobacter numbers in the feathers and colon, but Salmonella numbers in those samples were not correlated. The pattern of bacterial contamination before processing is complex and highly variable.

Recovery of Bacteria from Broiler Carcasses After Immersion Chilling in Different Volumes of Water, Part 2

Experiments were conducted to determine the relationship between poultry chilling water volume and carcass microbiology. In the first study, the volume of water used during immersion chilling was found to have a significant effect on the counts of bacteria recovered from broiler carcass halves; however, these volumes (2.1 and 16.8 L/kg) were extreme and did not reflect commercial levels. A second study using commercial chilling volumes was conducted with 3.3 L/kg (low) or 6.7 L/kg (high) distilled water in the chiller. Prechill broiler carcasses were removed from a commercial processing line, cut into left and right halves, and one-half of each pair was individually chilled in a bag containing low or high volume of water. Bags containing halves were submersed in a secondary chill tank containing approximately 150 L of an ice-water mix (0.6 degrees C). After 45 min, halves were removed, allowed to drip for 5 min, and rinsed with 100 mL of sterile water for 1 min. Rinses were analyzed for total aerobic bacteria, Escherichia coli, Enterobacteriaceae, and Campylobacter. When the numbers of bacteria in the half-carcass rinses (HCR) were compared, counts recovered from halves chilled in a low volume of water were the same as those recovered from the halves chilled with a high volume of water (P > 0.05). Levels found in the HCR ranged from 4.0 to 4.2 log(10) cfu/mL for aerobic bacteria, 3.3 to 3.5 log(10) cfu/mL for E. coli, 3.6 to 3.8 log(10) cfu/mL for Enterobacteriaceae, and 2.4 to 2.6 log(10) cfu/mL for Campylobacter. Data were also analyzed using a paired comparison t-test, and this analysis showed that there was no difference (P > 0.05) in the numbers of aerobic bacteria, E. coli, Enterobacteriaceae, or Campylobacter recovered from paired-halves chilled in different volumes of water. The present study shows that under the conditions outlined in this experiment, doubling the amount of water during immersion chilling (3.3 vs. 6.7 L/kg) did not improve the removal of bacteria from the surfaces of chilled carcasses.