ABSTRACT

Concentrations of airborne microorganisms were measured at four poultry slaughtering plants during four one-day visits using Anderson single- and six-stage viable particle (bioaerosol) samplers. Nonselective bioaerosols including total aerobic mesophilic bacteria, psychrotrophic

bacteria, and yeasts and molds were measured at approximately 40 sites in and around each plant. Bioaerosol concentrations were highest in shackling and defeathering areas and decreased with product flow through the plant until they approached outside levels in packaging areas. Bioaerosols exhibited a lognormal particle size distribution, with a count median diameter of 4.26 [+ or -] 0.19 m (n = 59) for total aerobic mesophilic bacteria, and 4.17 [+ or -] 0.33 m (n = 33) for yeasts and molds. Selected bacteria were speciated from some samples taken during each visit. Listeria monocytogenes, Staphylococcus aureas spp., Escherichia coli, Salmonella spp., and Lactobacilli were detected on 47% (n = 188), 54% (n = 92), 18% (n = 208), 31% (n = 206), and 89% (n = 100) of the samples, respectively. Some of the detections occurred in areas of the plant where meat was exposed following the chilling process.

INTRODUCTION AND OBJECTIVES

In fiscal year 2004, 8.9 billion chickens, 255 million turkeys, and 26 million ducks were slaughtered in the United States (USDA 2005). Foodborne illness is responsible for about 9,000 deaths per year in the United States alone. It has been estimated to cost $5 to $25 billion per year, and cause 24 to 81 million cases per year (Luchansky 2004). Therefore, in addition to packaging a clean product to enhance shelf life, poultry slaughtering plants must provide a safe product for consumers.

Poultry slaughtering plants typically consist of seven major processing areas: live bird shackling (LBS), defeathering, evisceration, chilling, processing, and packaging. The shackling and defeathering rooms have inherently high particulate matter and microorganism concentrations. The feathers of birds brought to the plants for slaughter are typically infested with pathogens such as Salmonella, Staphylococcus aureus spp., and Listeria monocytogenes. These pathogens become airborne when dislodged from the birds in the LBS area. Carcass contamination is supposed to be the lowest as the birds leave the chiller. However, recontamination of the carcass surface can potentially occur by deposition of large pathogen-carrying solid or liquid particles (Burge 1995). It is therefore important to keep exposure of chilled meat in the processing and packaging areas as low as possible to reduce the risk of foodborne illness and food spoilage. The objectives of this study of poultry slaughtering plants were to:

1. Characterize concentrations, size distribution, and sources of total aerobic mesophilic bacteria, psychrotrophic bacteria, and yeasts and molds.

2. Identify predominant yeasts and molds to the genus level.

3. Characterize concentration and sources of Pseudomonas spp., Listeria monocytogenes, Staphylococcus aureus spp., Escherichia coli, Salmonella spp., and Lactobacilli.

METHODS AND PROCEDURE

This study consisted of bioaerosol sampling of four different poultry processing plants, hereafter referred to as plants T, U, C, and K. Plant T processed about 37,000 turkeys daily and employed approximately 775 workers. The plant operated 16 hours per day in two 8-hour shifts. Plant U processed about 18,800 turkeys daily during the four visits and employed approximately 525 workers. The plant operated 16 hours per day in two 8-hour shifts. Plant C processed about 21,200 ducks daily during the four visits and employed approximately 250 workers. The plant operated 8 hours per day in one 8-hour shift. Plant K processed about 19,700 ducks daily during the four visits and employed approximately 270 workers. The plant operated 8 hours per day in one 8-hour shift. Plant floor plans and sampling locations are presented in Figures 1 through 4. Ventilation strategies and performance were described by Heber et al. (2002).

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

Each plant was evaluated four times, once during each season. The first visit to each plant occurred in the fall and was designated with the number "1" following the plant letter (e.g., T1, U1, C1, and K1). The subsequent visits in the winter, spring, and summer were denoted by 2, 3, and 4, respectively (e.g., T2, T3, and T4). Measured variables included air temperature, relative humidity, and bioaerosol concentrations. Nonselective bioaerosols sampled in this study included total aerobic mesophilic bacteria (TPC), psychrotrophic bacteria (PPC), and yeasts and molds (YAM). Sampled selective bioaerosols included Pseudomonas spp., Listeria monocytogenes, Staphylococcus aureus spp., Escherichia coli, Salmonella, spp., and Lactobacilli. The concentrations of selective bioaerosols were expected to be observed at much lower concentrations than nonselective bioaerosols.

The Andersen viable particle samplers were chosen due to the broad range of concentrations expected in poultry processing environments and the large numbers of samples to be handled and analyzed. The six stages of the standard Andersen impactor were designed to simulate the human respiratory system by the cutoff diameters of each stage. The single-stage impactor simply collects all samples on the sixth stage, for a size range of 0.65-7.0 [micro]m and above.

Size selective sampling was conducted for total aerobic mesophilic bacteria in six areas of each plant for each visit. Fungi were sampled in three areas of the plant during each visit. In each area, two samples were taken at different sampling times to obtain countable plates. The counts per stage were plotted by cumulative frequency distribution against the cutoff diameter of each stage to obtain the count median diameter and geometric standard deviation (Hinds 1982), parameters that describe the lognormal distribution, which these data seemed to fit well.

The flow rate of each single- and six-stage sampler was calibrated before and after each plant visit with a primary flow calibrator. Flow checks were conducted with precision rota-meters during plant visits. Duplicate samples were taken for each target microorganism each time sampling was conducted at a sampling location. Duplicates consisted of samples taken at two different sampling times (morning and afternoon). A standard procedure was followed at each site, beginning with cleaning the sampler with sterile pads. Petri plates were then labeled and loaded into the sampler, and digital timers were used to measure sampling time. Plant activity, temperature, and relative humidity were observed during sampling. The height of the bioaerosol sampling equipment was chosen to represent both the height of the product line as well as the employee's breathing zone. The sampled plates were transferred into mobile incubators using aseptic techniques until they could be placed in laboratory incubators upon return to campus. Incubation conditions for the various samples are presented in Table 1.

Total aerobic mesophilic bacteria and psychrotrophic bacteria were isolated and cultivated using tryptic soy agar (TSA), a non-selective general purpose medium. Bacto[TM] rose bengal agar (RBA) base supplemented with Bacto[TM] rose bengal antimicrobic supplement C was used for the selective isolation and enumeration of yeasts and molds. For selective bioaerosol detection, several different agars were used, each with a different target organism. Modified Oxford, Pseudomonas, MRS, Mannitol, XLD, and MacConkey agars were used to detect Listeria monocytogenes, Pseudomonas spp., Lactobacilli, Staphylococcus aureus spp., Salmonella spp., and Escherichia coli, respectively.

Predominant molds were selected according to their predominance on the plates. Classification was based on observation of the color and morphology of the viable culturable cell or colony forming unit (CFU). The predominant molds were transferred by point inoculation to plates filled with 20 mL of MEA to allow for better mold growth and easier identification. After incubation, the molds were identified to the genus level using mold identification keys (Pitt and Hocking 1985) with the assistance of a Foodborne Fungi Lab Manual (Cousin 1995; Samson et al. 1984).

The Andersen sampling method allowed for enumeration of the CFU. Two different counting methods generally employed in the field of bioaerosol sampling are the standard plate count method and the positive hole correction method.

With the standard plate count method, all distinct CFU on the surface of the medium are counted either with the naked eye or with the aid of magnification to some degree. Whenever possible, plates with less than 250 CFU were used. Plates with greater than 250 CFU were recorded as "too numerous to count."

The second method is the positive hole correction factor method (Macher 1989). As the name implies, it involves determining the number of potential impaction sites on the medium surface that were filled and applying a statistical correction factor to account for colony masking. The number of impaction sites is based on the number of jets per stage of the impactor. After the appropriate incubation time, the CFU on the plates were counted visually. All plates were enumerated using an AO Quebec Darkfield Colony Counter with 1.5X magnification. The quantity of CFU used to calculate nonselective bioaerosol concentrations (CFU/[m.sup.3]) were selected as the average of plates with less than 250 CFU; if neither plate yielded counts below 250 CFU, the positive hole correction technique was applied (Anderson 1958).

Nonselective bioaerosol concentrations (CFU/[m.sup.3]) were calculated as the product of plate count, flow rate ([m.sup.3]/s), and sampling time (s). Tukey's studentized range (Tukey's Honestly Significant Difference [HSD]) (Steel and Torrie 1980) tests were performed for the main plant effect in order to rank the four plants from highest to lowest for each of the three contaminant concentration means.

General observations of the concentrations of the selective bioaerosols were made and the averages were calculated among duplicate samples since a significant number of samples had no CFUs. For example, if samples A and B had no detected colonies, the concentration was reported as <X, where X is the concentration calculated if one colony was detected. If the concentration of sample A was <10 and sample B was <100, then the "average" concentration of the two samples would be reported as <10. For samples that were countable, a simple arithmetic average was calculated.

RESULTS--NONSELECTIVE BIOAEROSOLS

Nonselective bioaerosol concentrations in the four poultry slaughtering plants ranged from log 2 CFU/[m.sup.3] (0.45 log CFU/[ft.sup.3]) outside the plant for all three contaminants to log 6 CFU/[m.sup.3] (4.45 log CFU/[ft.sup.3]) for mesophilic bacteria and psychrotrophic bacteria and log 4 CFU/[m.sup.3] (2.45 log CFU/[ft.sup.3]) for YAM in the shackling and picking areas. The concentrations of mesophilic bacteria ranged from log 3 CFU/[m.sup.3] (1.45 log CFU/[ft.sup.3]) in evisceration areas to log 2.5 CFU/[m.sup.3] (0.95 log CFU/[ft.sup.3]) in the cut-up and packaging areas (Lutgring et al. 1997). Concentrations of psychrotrophic bacteria were usually within 1 log lower than mesophilic bacteria. The concentration of yeasts and molds represented approximately 1% of the total bioaerosol population and was within 2 logs lower than mesophilic bacteria (Figure 5).

Turkey Plant T

Indoor bioaerosol concentrations were highest in the live bird shackling (LBS), defeathering, and offal areas at the north end of the plant. Total plate counts of mesophilic bacteria were available only during visit T3 because concentrations usually exceeded the upper limits of the sampling equipment during other visits. Psychrotrophic bacteria levels were one to two orders of magnitude higher in the LBS, defeathering, and offal areas than the rest of the plant. Fungal (YAM) concentrations were typically one order of magnitude greater in LBS, defeathering, and offal areas than the rest of the plant except during visit T2 when they were nearly the same order of magnitude throughout the entire plant. Visit T3 was the only time when mesophilic bacteria concentrations were available in the LBS area, and the levels were two to three orders of magnitude greater than the rest of the plant.

Outdoor mesophilic bacteria concentrations were not significantly different from those observed in the whole bird packaging (WBP) area. In the defeathering areas, TPC were two to three orders of magnitude higher than outside and further processing areas. Mesophilic bacteria levels were significantly lower in the evisceration area and dropped to their lowest in the chilling, cut-up, and portions packaging (PP) areas (Figure 5).

Psychrotrophic bacteria concentrations in the defeathering areas were significantly higher than outside and other areas of the plant (P = 0.0001). Levels of PPC in evisceration, chilling, cut-up, and both packaging areas were not significantly different from levels outside the plant.

For most areas of the plant, YAM concentrations were lowest in winter and highest in summer. The shackling and defeathering areas had higher levels than other plant areas (P = 0.0001) but were not significantly different than outside. YAM concentrations in the further processing areas (evisceration, chilling, cut-up, and packaging) decreased to levels below outdoors (Figure 5).

Turkey Plant U

High bioaerosol concentrations were expected only in the LBS area and were expected to decrease as the birds moved through the LBS area and picking room into evisceration because of the lack of bioaerosol generation from feathers. However, bioaerosol concentrations remained high in the evisceration area because of air movement from picking to evisceration.

Air was drawn from the shackling area into the picking room. A portion of the air that entered the picking room was exhausted by the roof-mounted fans, and the remaining air entered evisceration through wall openings. This caused the transfer of bioaerosols into the evisceration area. Unfortunately, the dirtiest area in the plant actually supplied a portion of the ventilation air to the rest of the plant.

Live bird shackling had the highest TPC, but the levels decreased only slightly in the picking and evisceration areas and were not significantly different from the chilling and WBP areas. The physically separated cut-up and PP areas had mesophilic bacteria concentrations significantly lower than the other five processing areas but, again, were significantly higher than outdoor levels (Figure 5).

Outdoor psychrotrophic bacteria concentrations were significantly lower than any indoor processing area, with levels in the shackling area significantly the highest (p < 0.0001). Processing areas located in the same room (picking, evisceration, chilling, and WBP), or that were supplied by the same air, had psychrotrophic bacteria concentrations that were not significantly different from each other. Cut-up and PP operations located in a separate room had significantly lower psychrotrophic bacteria concentrations but were still higher than outdoor levels.

[FIGURE 5A OMITTED]

[FIGURE 5B OMITTED]

Concentrations of YAM were lowest outdoors and increased one to two orders of magnitude in the shackling area. Concentrations remained high throughout the plant, never decreasing significantly from the levels observed in the shackling area (Figure 5).

Duck Plant C

Concentrations were highest in LBS, scalding, and picking. Total aerobic and psychrotrophic concentrations were not available in the LBS and scalding areas due to levels exceeding the upper measurement limits of the sampling equipment. These concentrations were only available in the picking area during visit C3 and were two to three orders of magnitude higher than outdoors. Fungal concentrations were available in these areas during visits C1, C2, and C3 and were one order of magnitude higher than outdoor levels. In the cut-up and evisceration areas, concentrations were of the same order or one order of magnitude higher than outdoor levels during all visits.

The outdoor mesophilic bacteria levels were significantly lower than all processing areas of plant C. The shackling and picking areas had the highest mesophilic bacteria concentrations, serving as the exhaust port for the remainder of the plant. TPC in the evisceration, chilling tanks, and WBP, which were located in the same room, were not significantly different than those in the cut-up and PP areas located in the adjacent room (Figure 5).

The lowest psychrotrophic bacteria concentrations were observed in the west side of evisceration by the chill tanks and WBP and were not significantly different than outside concentrations. These fluctuations were not similar to TPC and perhaps reflected the pattern of airflow from the storage and freezer areas where psychrotrophic bacteria may exist, through cut-up and evisceration toward picking, bypassing the chill tank and WBP.

Yeast and mold concentrations were significantly lower outdoors than in the defeathering areas by approximately one order of magnitude (p < 0.0001). Levels in the picking room were highest but were not significantly different from levels in evisceration, chilling, cut-up, and both packaging areas. Overall, YAM levels in plant C were lower than the other duck plant and lower than turkey plant U (Figure 5).

Duck Plant K

Bioaerosol concentrations were lowest outside and in the packaging area. Bioaerosol levels were highest in the LBS, kill, and scalding/picking areas, where mesophilic and psychrotrophic bacteria concentrations were at least three to four orders of magnitude higher than outdoors. During visits K1, K2, and K3, mesophilic and psychrotrophic bacteria concentrations exceeded the upper measurement limits of the sampling equipment. Inside and outside YAM concentrations were similar during all visits, but TPC and PPC decreased with product flow through the plant. The cut-up and PP areas were located at the south end of evisceration, and mesophilic and psychrotrophic bacteria concentrations were one to two orders of magnitude higher in these areas than the WBP area in the packaging room (Figure 5).

Outdoor mesophilic bacteria concentrations were significantly lower than all processing areas of plant K except the WBP area, which was located in a separate room at the west end of the plant. The shackling and picking areas had the highest mesophilic bacteria concentrations, serving as the exhaust port for the remainder of the plant, but were not significantly higher than the further processing areas. Mesophilic bacteria concentrations in the evisceration, chilling, cut-up and PP areas, which were located in the same room, were not significantly different than those in the shackling and picking areas, although a slight decreasing trend existed (Figure 5). Only the WBP room had significantly lower mesophilic bacteria concentrations.

The lowest concentrations of psychrotrophic bacteria were observed outside and in the WBP area, similar to mesophilic bacteria. Concentrations in the shackling, picking, and all subsequent operations in evisceration were not significantly different from one another. These data, along with the mesophilic bacteria concentrations, suggest that although air was exhausted through evisceration to the defeathering room, some mixing or introduction of additional sources occurred.

The YAM concentrations were significantly lower outdoors than in the shackling and picking areas by one to two orders of magnitude. YAM concentrations in the shackling room were highest but were not significantly different from the picking room nor from levels in the cut-up and PP area in the south end of evisceration, which were one to two orders of magnitude above those in the WBP area. Other areas in evisceration and near the chilling operation showed YAM concentrations significantly lower than shackling. Again, only the WBP room had YAM concentrations similar to those observed outside (Figure 5). The data suggest some type of air exchange between the shackling, defeathering, and evisceration rooms (Heber et al. 2002).

SELECTIVE BIOAEROSOLS

No colonies of Pseudomonas were observed in 89 samples collected among four plants. Based on sampling times, concentrations were less than 4 CFU/[m.sup.3] (0.1 CFU/[ft.sup.3]) throughout the plants, even in the "dirty" areas such as the shackling, scalding, picking, and defeathering areas. Listeria monocytogenes was detected on only 88 out of 188 samples, and concentrations ranged from 1 to 1,932 CFU/[m.sup.3] (0.03 to 55 CFU/[ft.sup.3]). Staphylococcus aureus spp. was detected on 50 out of 92 samples (Figure 6). Escherichia coli was detected in 37 out of 208 samples, of which 11 samples were between 100 and 624 CFU/[m.sup.3] (2.83 to 17.7 CFU/[ft.sup.3]). The highest measurements occurred in the dirtiest areas of the plant, e.g., the picking, scalding, offal, and shackling areas. Salmonella was detected in 63 of 206 samples, ranging from 2 to 598 CFU/[m.sup.3] (0.06 to 16.9 CFU/[ft.sup.3]). The highest levels were observed in the "dirty" areas, especially the picking room. Salmonella was not detected in most of the clean areas of the plant.

Lactobacilli (lactic acid bacteria) were detected on 89 of 100 samples (Figure 6). The highest measured concentration was 6516 CFU/[m.sup.3] (185 CFU/[ft.sup.3]). The colonies on 29 samples, taken mostly in shackling, picking, evisceration, and defeathering rooms, were too numerous to count. The highest measured concentration of Lactobacilli was greater than 6516 CFU/[m.sup.3] (185 CFU/[ft.sup.3]).

[FIGURE 6A OMITTED]

[FIGURE 6B OMITTED]

Turkey Plant T

Cladosporium and Fusarium were predominant outdoors and in the shackling and defeathering areas. They were joined in the evisceration area by Penicillium, which continued to be observed in the cooler areas of the plant toward the end of the process.

Listeria was detected in the shackling room at concentrations of 20 to 894 CFU/[m.sup.3] (0.57 to 25 CFU/[ft.sup.3]), but was less than 4 CFU/[m.sup.3] (0.11 CFU/[ft.sup.3]) in all other areas (Figure 6).

Turkey Plant U

Predominant outdoor molds for visit U4 were Scopulariopsis, Fusarium, and Cladosporium. The shackling area had a similar predominance, although Scopulariopsis was outnumbered by Penicillium. The defeathering area showed a predominance of Geotrichum, commonly known as a "machinery mold," which may indicate equipment sanitation problems. The further processing areas showed no predominance of the molds observed outside plant U but instead introduced a variety of Penicillium species, which were observed throughout the cooler areas of the plant in high numbers.

Listeria was 1 to 10 CFU/[m.sup.3] (0.03 to 0.28 CFU/[ft.sup.3]) during visit U3 in the cut-up and evisceration areas but was less than 3 CFU/[m.sup.3] (0.09 CFU/[ft.sup.3]) during visit U1 and less than 1 CFU/[m.sup.3] (0.03 CFU/[ft.sup.3]) during visit U4 (Figure 6).

Duck Plant C

Predominant molds for visit C4 primarily reflected those commonly occurring outdoors, evidenced by the predominance of Cladosporium throughout the plant. Fusarium and Talaromyces, both common soil molds, were observed outside the plant and in the shackling and defeathering areas of the plant, where movement of ducks stirred up soil and litter. Penicillium outnumbered soil molds indoors and, along with Cladosporium, was observed throughout the process, even in the packaging areas.

During visit C1, Listeria concentrations ranged from 3 to 1,932 CFU/[m.sup.3] (0.09 to 55 CFU/[ft.sup.3]), with the highest concentrations observed in the picking, shackling, and offal rooms. It was less than 4 CFU/[m.sup.3] (0.11 CFU/[ft.sup.3]) in other areas during visit C1. Listeria concentrations ranged from 0 to 1,932 CFU/[m.sup.3] (0 to 55 CFU/[ft.sup.3]) in the picking, shackling, and offal rooms, with the highest levels measured during visit C2 (Figure 6). Listeria was detected at low levels in the "clean" areas such as the refrigeration and deboning (DB) rooms and the breast trim area.

Duck Plant K

Predominant molds identified for the summer showed the predominance of common outdoor molds Cladosporium and Penicillium outside and throughout the processing areas of the plant. However, the predominance shifted to Geotrichum, beginning in the picking area of the defeathering room and continuing into the chilling and packaging areas. Plant K was the only plant with a predominance of Geotrichum that continued beyond the highly mechanized picking area.

The geometric mean of Listeria concentrations was 257 CFU/[m.sup.3] (7.3 CFU/[ft.sup.3]) in the live bird holding (LBH) area and ranged from 128 to 1074 CFU/[m.sup.3] (3.6 to 30.4 CFU/[ft.sup.3]) (Figure 6). In evisceration, Listeria concentrations ranged from less than 3 CFU/[m.sup.3] (0.09 CFU/[ft.sup.3]) during visit K2 to 143 CFU/[m.sup.3] (0.06 to 4.1 CFU/[ft.sup.3]) during visit K4, but most of the measurements were less than 10 CFU/[m.sup.3] (0.28 CFU/[ft.sup.3]). Listeria was detected in 7 of 12 defeathering room samples. The seven samples ranged from 7 to 49 CFU/[m.sup.3] (0.20 to 1.39 CFU/[ft.sup.3]). Out of eight measurements in the packing room, Listeria was detected in only two samples and the concentrations were only 3 and 7 CFU/[m.sup.3] (0.09 to 0.20 CFU/[ft.sup.3]), respectively.

Staphylococcus aureus spp. concentrations ranged from less than 2 to 598 CFU/[m.sup.3] (0.06 to 16.9 CFU/[ft.sup.3]) in the LBH area. Most of the highest Staphylococcus aureus spp. concentrations occurred in the "dirty" areas of shackling, defeathering, picking, and evisceration areas. The highest levels in the "clean" areas were observed during visit K2.

DISCUSSION

The mesophilic and psychrotrophic bacteria concentrations varied in each plant depending on the process. Active and frantic movement of birds' wings and feet in the shackling area dispersed bioaerosols from litter and the birds themselves. Rapidly rotating mechanical picking apparatus in the picking area coupled with large quantities of warm water spray or mist may have caused additional aerosolization of contaminants from bird carcasses, water, and feather material. Mesophilic bacteria concentrations were the highest in these areas. In most cases, concentrations then decreased toward and into the evisceration area where the nature of the bioaerosols changed with their mode of generation and dispersion. Aerosolization occurred through the use of water sprays and vacuum apparatus in the process. From this point into the plant, mesophilic and psychrotrophic bacteria concentrations decreased unless they were spatially oriented nearby the evisceration area or were supplied with air from a more contaminated process area.

Beyond the shackling and picking areas, YAM concentrations in plants T, C, and K reflected concentrations and composition of the outside air with little exception. Predominant molds were also similar between outdoors and the processing areas in three of the four plants. Molds observed at high levels inside the plants that were not predominant outside the plants may indicate an internal source or conditions that favor mold growth.

Although each plant had difficulty controlling humidity and condensation to some degree, plant U had the most problems. Both YAM concentrations and mold composition for plant U indicated potential growth of Penicillium indoors that was not predominant outdoors. The predominance of Geotrichum indicated potential equipment sanitation problems at plants U and K.

Major sources of bioaerosols for each poultry processing plant were the LBH and LBS areas. These areas had concentrations of bioaerosols that were several orders of magnitude greater than upwind outdoor locations. General airflow within the plants consisted of air movement from all processing areas, including LBH and LBS, toward the large exhaust outlets in the defeathering areas.

Plant layout and the separation or lack of separation of processing areas was a very good indicator of which areas would have similar mesophilic and psychrotrophic bacteria concentrations. Areas that were physically separated, such as PP at plant T or WBP at plant K, had lower mesophilic and psychrotrophic bacteria concentrations by 1-2 log differences.

The four plants, although unique in their layouts, showed basically the same trends in nonselective bioaerosol concentrations. Concentrations of mesophilic and psychrotrophic bacteria varied with processing activities, whereas yeasts and molds varied with environmental factors. The time of day and season had a very significant effect on the variation of the concentrations but did not mask these trends.

Pseudomonas was not detected, whereas the concentrations of the other selective bioaerosols followed similar patterns throughout the plant as the nonselective bioaerosols. Maximum concentrations of Listeria monocytogenes, Staphylococcus aureus spp., Escherichia coli, Salmonella, and Lactobacilli were 1932, 598, 624, 598, and 6516 CFU/[m.sup.3] (55, 17, 18, 17, and 185 CFU/[ft.sup.3]), respectively, all in the "dirty" areas of the plants. Many samples of each of these microorganisms resulted in no detections.

The particle size distribution of bioaerosols in the plants was consistent between plants, with the largest particles aerosolized in the LBS areas (Figure 7). The particle size was higher in and near the shackling areas because of the nature of the source, e.g., feathers, litter, etc. The count median diameter gradually decreased from greater than 5 [mu] to less than 4 [micro] through the processing areas, most likely due to a loss of larger particles due to deposition and sedimentation (Table 2). Packaging was slightly larger than cut-up in both cases, most likely due to the location of packaging areas in some of the plants.

Workers in the LBH areas of the duck plants and in the LBS areas of duck and turkey plants had the highest occupational exposure to bioaerosols. Several mold types identified in these areas, such as Cladosporium and Fusarium, are known to cause allergenic and sensitization reactions in at least 10% of people.

The weather outside and conditions inside the plant were highly variable and nonrepeatable, as bioaerosol samples were taken and ventilation measurements were recorded throughout the day-long visits to each duck and turkey plant. Large differences in the amount of product or the number of people in processing areas occurred depending on the time of day. Processing intensity and people activity varied between samples at the same site. Periods when birds were not on the shackle line occurred due to staggered lunch and coffee breaks for the workers. Discrepancies in the data were expected because of the inherent variability.

[FIGURE 7 OMITTED]

The season of the year did not appear to affect bioaerosol levels, and data collected during only one sampling day per season was insufficient to detect differences. Higher bioaerosol concentrations in the shackling and defeathering areas during visits T4, U4, and K4 were recorded because positive hole correction was used during visit 4 for all plants. This provided a higher upper detection limit than during previous visits, during which many of the concentrations in the shackling and defeathering areas exceeded the upper detection limit. These were reported as greater than a certain level, e.g., >3.3 x [10.sup.5] CFU/[m.sup.3] (9.4 x [10.sup.3] CFU/[ft.sup.3]), and during visit 4, concentrations up to 4.0 x [10.sup.6] CFU/[m.sup.3] (9.4 x [10.sup.3] CFU/[ft.sup.3]) were measured.

Measured bioaerosol concentrations were affected by day of year, time of day, worker activity level, and frequency and duration of process operations. Concentrations measured with the single-stage particle sampler in the shackling and picking areas frequently reached the upper detection limit of 400 positive holes. The average concentrations in the eight major sampling areas were statistically different (P < 0.05). Trends in bioaerosol concentration correlated well with the overall room-to-room airflow and provided useful and important information for developing guidelines for duck and turkey slaughtering plant design.

The types of mold were identified to the genus level for the final (summer) visit to each plant. The molds are listed in Table 3 for the eight major sampling areas of each plant. All of the molds except Geotrichum and Penicillium were also predominant outdoors. Geotrichum was observed only in plants U and K and was consistently predominant in plant K. Plant T predominant molds reflected the outdoor molds. Plants C and U showed a predominance of Penicillium in the processing areas.

Cladosporium, observed in the outdoor environment of each of the four plants, is a common plant pathogen, or "field fungi." It is dominant in air because its spores are small, dry, and resistant to sunlight. It grows well at near freezing temperatures but not at body temperature (37[degrees]C). Because of this, it can spoil refrigerated meat. Fusarium was another "field fungi" that was isolated in this study. It also does not grow at body temperature but grows well around 5[degrees]C (41[degrees]F). Geotrichum needs a high water activity to grow, such as that which occurs on machinery where water pools and food residue accumulates. It does not grow at 37[degrees]C (98.6[degrees]F). Penicillium, a commonly isolated, ubiquitous mold, grows over a wide temperature range. Since many Penicilli are psychrotrophic, it causes spoilage of refrigerated foods (Cousin 1995).

For plant T, the predominance of Cladosporium and Fusarium was consistent throughout the process. For plants T and K, Cladosporium was predominant outdoors and remained predominant throughout the plant. For plant C, the same was true with the exception of evisceration, where Penicillium was predominant. For plant U, Cladosporium and Fusarium were not predominant past the picking area but were outnumbered by Penicillium in the rest of the plant. Penicillium was predominant in all four of the plants' chilling, cut-up, and packaging areas. This could have been related to the ability of Penicillium to survive in the cooler temperature and higher humidity in these processing areas, although both Penicillium and Cladosporium have been classified as psychrotrophic. This may have also indicated a lack of makeup air from the outside to dilute indoor bioaerosols and reduce humidity. The types of molds most capable of causing allergic response include all of the predominant molds identified in the four poultry processing plants: Alternaria, Cladosporium, Aspergillus, Fusarium, Penicillium, Geotrichum, and Scopulariopsis, among others. The highest potential for allergy and pulmonary disease existed for the shackling workers. This was due not only to high TPC and YAM but also to their repeated and prolonged exposure to endotoxins and proteins found in bird droppings, which have long been implicated in "bird handler's" or "breeder's lung" disease (Al-Doory and Domson 1984).

CONCLUSIONS

Specific conclusions from the results of this field investigation of poultry slaughtering and processing plants were as follows:

1. Mesophilic bacteria concentrations were highest in the shackling area of all plants and were typically 5 to 6 log CFU/[m.sup.3] (3.45 to 4.45 log CFU/[ft.sup.3]), decreasing to 3 log CFU/[m.sup.3] (1.45 log CFU/[ft.sup.3]) at later stages of the process.

2. Psychrotrophic bacteria concentrations were within one log less than mesophilic bacteria and followed the same trends throughout the plants; thus, for similar bioaerosol populations, mesophilic bacteria may be used as an indicator of psychrotrophic bacteria.

3. Concentrations of yeasts and molds varied from log 2 CFU/[m.sup.3] (0.45 log CFU/[ft.sup.3]) outdoors to log 4 CFU/[m.sup.3] (2.45 log CFU/[ft.sup.3]) in the shackling area. Yeast and mold concentrations were similar to outdoor levels once the product reached the packaging areas.

4. Bioaerosol particle size was largest in the live-bird holding area and decreased as the product moved through processing.

5. Employees working in the live bird shackling areas had the highest potential exposures to bioaerosols.

6. Predominant molds inside the plants typically reflected those observed outside each plant, with the exception of Penicillium.

7. Based on 89 samples, Pseudomonas was not airborne in the poultry processing plants.

8. Concentrations of Pseudomonas spp., Listeria monocytogenes, Staphylococcus aureus spp., Escherichia coli, Salmonella spp., and Lactobacilli followed similar patterns throughout the plant.

9. The maximum concentrations of Listeria monocytogenes, Staphylococcus aureus spp., Escherichia coli, Salmonella, and Lactobacilli were 1,932, 598, 624, 598, and 6516 CFU/[m.sup.3] (55, 17, 18, 17, and 185 CFU/[ft.sup.3]), respectively, all occurring in the "dirty" areas of the plants.

RECOMMENDATIONS

1. Bioaerosol dispersion in poultry processing plants should be controlled by physically separating process areas and controlling airflow parameters by design to ensure a flow direction opposite of product flow or by independent exhausts. The design may include positive pressure in clean areas to create these proper airflow patterns.

2. Based on the presence of allergenic molds at concentrations significantly higher than background outdoor levels, poultry plant employees in the live bird holding and shackling areas should be included in a medical surveillance program and provided with particulate respirators and training in accordance with OSHA regulations.

ACKNOWLEDGMENTS

This research project (RP-834) was supported by the American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc., and TC 10.9, Refrigeration Application for Foods and Beverages. Support from the Purdue University Agricultural Research Programs is also acknowledged.

NOMENCLATURE

C = duck plant 1

CFU = colony-forming units

CLAD = Cladosporium

CMD = count median diameter

CU = cut-up area or room

DF = defeathering (room)

EVIS = evisceration (room)

FUS = Fusarium

GEO = Geotrichum

GSD = geometric standard deviation

LBS = live bird shackling

LBH = live bird holding

K = duck plant 2

KILL = bird killing area or room

PEN = Penicillium

PIN = pin feather removal area or room (duck plants only)

PP = portions packaging (area or room)

PPC = psychrotrophic plate counts

SCALD = scalding room

SCO = Scopulariopsis

T = turkey plant 1

TPC = total plate counts of mesophilic bacteria

U = turkey plant 2

WAX = waxing area or room (duck plants only)

WBP = whole bird packaging (12)

YAM = yeasts and molds (20)

REFERENCES

Al-Doory, Y., and J.F. Domson. 1984. Mould Allergy. Lea & Febiger.

Andersen, A.A. 1958. New sampler for the collection, sizing, and enumeration of viable airborne particles. Journal of Bacteriology 76:471-84.

Burge, H.A. 1995. Bioaerosols. Boca Raton, FL: CRC Press.

Cousin, M. 1995. Foodborne Mold Identification Workshop. Laboratory manual. Food Science Department, Purdue University, West Lafayette, IN, March 6-10.

Heber, A.J., M.J. Peugh, N.J. Zimmerman, and R.L. Linton. 2002. Poultry slaughter plants: Ventilation system performance. ASHRAE Trans. 108(2):129-44.

Hinds, W.C. 1982. Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles. New York: John Wiley & Sons, Inc.

Luchansky, J.B. 2004. Recovery, characterization and control of food borne pathogens from slaughter through further processing. The National Food Centre. Athens, Greece. Meeting abstract.

Lutgring, K.R., R.H. Linton, N.J. Zimmerman, M. Peugh, and A.J. Heber. Distribution and quantification of bioaerosols in poultry-slaughtering plants. Journal of Food Protection 60(7):804-10.

Macher, J.M. 1989. Positive hole correction of multiple jet impactors for collecting viable microorganisms. American Industrial Hygiene Association Journal 50:561-68.

Pitt, J.I., and A.D. Hocking. 1985. TPCi and Food Spoilage. Orlando, FL: Academic Press.

Samson, R.A., E.S. Hoekstra, and C.A.N. van Oonschot. 1984. Introduction to foodborne TPCi. Centraalbureau voor Schimmelcultures, BAARN.

Steel, R.G.D., and J.H. Torrie. 1980. Principles and Procedures of Statistics--A Biometrical Approach. New York: MaGraw-Hill.

USDA. 2005. Statistical summary: Federal meat and poultry inspection for fiscal year 2004. US Department of Agriculture, Washington, DC.

Albert J. Heber, PhD, PE

Member ASHRAE

Michael W. Peugh

Member ASHRAE

Karen R. Lutgring

Neil J. Zimmerman, PhD, PE, CIH

Richard H. Linton, PhD

Albert J. Heber, Neil J. Zimmerman, and Richard H. Linton are professors at Purdue University, West Lafayette, IN. Michael W. Peugh is an HVAC design engineer at TMP Consulting Engineers, Inc., Boston, MA. Karen R. Lutgring is a safety director at Grain Processing Corporation, Washington, IN.

Table 1. Incubation Conditions for Target Microorganisms

                                    Incubation Temperature  Incubation
Agar             Microorganism      [degrees]C  [degrees]F  Time (Days)

Trypticase soy   TPC                37          98.6         2
Trypticase soy   PPC                 7          44.6         7
Rose Bengal      Yeast and molds    21          70.0        10
Malt extract     Molds              21          70.0         5
Pseudomonas      Pseudomonas spp.   37          98.6         2
Modified Oxford  Listeria           37          98.6         2
Mannitol Salts   Staphylococcus     37          98.6         2
                   aureus spp.
XLD, MacConkey   Escherichia coli,  37          98.6         1
                   Salmonella
MRS              Lactobacilli       37          98.6         1

Table 2. Particle Size Diameter of Total Aerobic Bacteria and YAM
(Mean [+ or -] Std Error)

Sampling   Total Aerobic Bacteria
Area       CMD, [micro]m           GSD, [micro]m

Shackling  4.97 [+ or -] 0.01 in.  1.98 [+ or -] 0.001 in.
Kill room  4.98 [+ or -] 0.03 in.  1.86 [+ or -] 0.005 in.
Picking    4.53 [+ or -] 0.20 in.  2.03 [+ or -] 0.019 in.
EVIS       4.21 [+ or -] 0.01 in.  1.68 [+ or -] 0.001 in.
Cut-up     3.64 [+ or -] 0.01 in.  1.69 [+ or -] 0.001 in.
Packaging  4.06 [+ or -] 0.01 in.  1.82 [+ or -] 0.002 in.

Sampling   Molds and Yeasts
Area       CMD, [micro]m            GSD, [micro]m

Shackling  5.12 [+ or -] 0.018 in.  1.67 [+ or -] 0.003 in.
Kill room  -                        -
Picking    -                        -
EVIS       3.69 [+ or -] 0.012 in.  1.54 [+ or -] 0.003 in.
Cut-up     3.79 [+ or -] 0.097 in.  1.55 [+ or -] 0.025 in.
Packaging  3.82 [+ or -] 0.030 in.  1.56 [+ or -] 0.001 in.

Table 3. Predominant Molds Observed in Sampling Areas of Each Plant

Area       Plant T           Plant U     Plant C            Plant K

Outside    CLAD, FUS         SCO, FUS,   CLAD, FUS          CLAD
                               CLAD
Shackling  CLAD, Alternaria  FUS, CLAD,  CLAD, Talaromyces  PEN
                               PEN
Picking    CLAD, FUS         GEO         CLAD               GEO, FUS
EVIS       CLAD, FUS, PEN    PEN         PEN                CLAD
Chilling   CLAD, ASP, PEN    PEN, ASP    PEN, CLAD          GEO, CLAD,
                                                              PEN
Cut-up     CLAD, PEN         PEN         PEN, CLAD          CLAD, PEN
Packaging  CLAD, PEN                     PEN, CLAD          CLAD, PEN,
                                                              GEO