Validation of Cooking Methods Using Shell Eggs Inoculated with ...

Validation of Cooking Methods Using Shell Eggs Inoculated with Salmonella Serotypes Enteritidis and Heidelberg A. L. Davis, P. A. Curtis,1 D. E. Conner, S. R. McKee, and L. K. Kerth Department of Poultry Science, Auburn University, Auburn, AL 36849 ABSTRACT Salmonella enterica serotype Enteritidis has long been associated with eggs, and more recently, Salmonella enterica serotype Heidelberg has also become associated with eggs. This study was undertaken to determine whether Salmonella Enteritidis and Salmonella Heidelberg are effectively eliminated from eggs by various cooking methods. Seven cooking methods were chosen—hard and soft cooked, scrambled, over easy, sunny-side up, poached, and free poached—and a pan insert and the

free-flowing method were used. Shell eggs, purchased from a grocery store, were inoculated with Salmonella and cooked. The cooked eggs were analyzed by USDAapproved methods for Salmonella recovery. Findings indicated that existing cooking methods for the hard-cooked, soft-cooked, and poaching methods were safe. However, the same was not true for the current sunny-side-up, overeasy, and scrambled egg cooking methods.

Key words: Salmonella, egg, cooking 2008 Poultry Science 87:1637–1642 doi:10.3382/ps.2007-00419

INTRODUCTION Salmonella enterica serotype Enteritidis (SE) is one of the most common Salmonella serotypes worldwide, particularly in developed countries (Patrick et al., 2004). Casecontrol studies of sporadic human Salmonella outbreaks and infections have shown that shell eggs are a major risk factor for disease (St Louis et al., 1988; Mishu et al., 1994; Passaro et al., 1996). However, when examining the natural occurrence of Salmonella in laying hens, the incidence of SE was found to be low in eggs. Poppe et al. (1992) found that less than 0.065% of eggs tested were positive for SE (roughly 2 positives from a sample size of 16,000 eggs), whereas Humphrey et al. (1989a, 1991) also isolated Salmonella serotypes other than SE from eggs. Salmonella enterica serotype Heidelberg (SH), like other nontyphoidal salmonellae, appears to be associated primarily with food. An average of 2,180 cases of SH infection were reported annually in the United States between 1993 and 1997, which accounted for approximately 6% of all Salmonella culture-confirmed infections (Centers for Disease Control and Prevention, 1998). A study done by Schoeni et al. (1995) showed the growth and penetration of SH in the egg, and Jones et al. (1995) also found SH on eggshells. In a study conducted by Gast et al. (2004), SH was shown to invade the reproductive tissues of lay-

ing hens, and SH could then be deposited inside the developing egg, similar to contamination with SE. It has been difficult to determine whether SH infections are a result of contamination of eggshells or from eating intact eggs contaminated before lay (transovarian transmission). Because of the potential presence of Salmonella in shell eggs, numerous reported outbreaks have been associated with eating raw or undercooked eggs (Centers for Disease Control and Prevention, 1990, 1996a,b). Cooking studies done by Humphrey et al. (1989b), Baker (1990), and Tharrington et al. (2003) showed that some, but not all, cooking methods will inactivate Salmonella should it be present. The purpose of this research was to evaluate current consumer cooking methods for safety. Thus, 7 cooking methods—hard and soft cooked, scrambled, over easy, sunny-side up, poached, and free poached—were chosen to determine whether they were adequate for the destruction of SE and SH. The cooking methods were chosen to simulate cooking techniques in the home. Microbiological tests were used to determine whether a cooking method could be considered safe (for any age or health group), reasonably safe (safe for a healthy person), or unsafe (not safe for consumption).

MATERIALS AND METHODS Preparation of Inoculum

©2008 Poultry Science Association Inc. Received October 11, 2007. Accepted March 28, 2008. 1 Corresponding author: [email protected]

The SE phage types 3, 4, 8, 9a, 22, 23, 24, and 28 and SH phage type 11 were used in the project. These phage types were used because they were isolated from poultry 1637

1638

DAVIS ET AL.

and egg products. For each SE and SH phage type tested, a 106 dilution of a 24-h culture was made by using 10 mL of tryptic soy broth (TSB; Hardy Diagnostics, Santa Maria, CA) as the diluent. A 3-mL aliquot from each of the 106 diluted-culture phage types was combined in a sterile glass bottle to make the Salmonella cocktail used for inoculation of tested eggs. Confirmation of the bacterial level in the SE and SH suspensions was obtained by the spiral plating technique (Microbiology International, Frederick, MD) with plate count agar (Hardy Diagnostics) and brilliant green sulfapyridine agar (BGS; Hardy Diagnostics) that were incubated for 24 h at 37°C before counting colonies. After incubation, the plates were counted by using the spiral function of a Q Count Plate Reader (Microbiology International, Bethesda, MD). The mean population of the SE and SH cocktail on BGS was 4.3 cfu/mL.

Poached and free-poached eggs were cooked differently. For the poached method, a 3-cup (711-mL) sterile nonstick poaching pan with insert was placed on a hot plate and the water was brought to a rolling boil (approximately 10 to 12 min). An egg was then removed from the refrigerator and cracked into the poaching insert (the insert was sprayed with a nonstick cooking spray to prevent the egg from sticking and to simulate cooking at home). Immediately after cracking the egg into the poaching insert, the lid was placed on the pan. The hot plate temperature was changed from high to low. The egg was cooked for 5 min, which was the time it took for the thick and thin albumen to set and the yolk to thicken. Freepoached eggs were cracked and dropped into boiling water. The hot plate temperature was reduced from high to low. The egg was cooked until the thick and thin albumen were set (approximately 3 min).

Collection and Handling of Egg Samples

Temperature Measurement and Inoculation Procedures

The eggs used in this study were large Grade A white eggs purchased from a local supermarket each Monday morning and stored in a refrigerator until they were needed later in the week. Eggs were candled with a Lyon Hi-Intensity Egg Candler (Lyon Electric Company Inc., Chula Vista, CA) to exclude any cracked and leaking eggs from further testing. Acceptable candled eggs were held at 4°C before cooking.

Cooking Methods All cooking methods used in this study were taken from Tharrington et al. (2003) in an attempt to simulate consumer cooking methods used in the home. Hard- and soft-cooked eggs were prepared by placing a single egg into an empty sterile sauce pan (1.4 L) and adding sterile water until the egg was covered. Eggs were then cooked on a single hot plate (Toastmaster Inc., Macon, MO). The water was brought to a boil, which took about 10 to 12 min, after which the pan was removed from the hot plate and allowed to sit for 15 min for hard-cooked eggs and 5 min for soft-cooked eggs. Scrambled eggs were scrambled in a frying pan by intermittently scraping the bottom surface toward the center of the skillet with a sterile spatula. This was continued until no visible liquid remained (this took approximately 2 min). The over easy and sunny-side up eggs were cracked into a preheated sterile 20.32-cm (8-in.) frying pan (heated to at least 121°C) and covered with a glass lid. The hot plate temperature was then switched from high to low. The eggs were cooked until the thick and thin albumen were completely set and the yolk began to thicken (approximately 1.5 min). After this took place, the lid was removed, and the egg was turned over. The lid was then replaced and the egg was cooked for an additional 30 s of cooking. Sunny-side up eggs were not turned after the original 1.5 min.

For hard- and soft-cooked eggs, a hypodermic type K thermocouple (Digi-Sense, Cole-Parmer, Vernon Hills, IL) attached to a Tempest Data Logger (Tangent Systems Inc., Charlotte, NC) was inserted through the eggshell into the large end of the egg through the air cell and into the center of the yolk of the egg with the aid of a Lyon HiIntensity Candler (Tharrington et al., 2003). Shells in all the eggs in this study were not surface-sterilized to replicate consumer cooking practices. Super Glue (Henkel Consumer Adhesives Inc., Avon, OH) was then used around the thermocouple to seal the hole in the eggshell. A sterile 21-gauge needle and syringe were used to deposit 0.1 mL of the Salmonella cocktail through the large end of the egg and air cell onto the vitelline membrane. Placement of the inoculua was confirmed with the aid of a Lyon Hi-Intensity Candler. After injection, the hole was sealed with Super Glue. Scrambled eggs were aseptically cracked on sterile foil and opened into a preheated sterile nonstick frying pan (heated to at least 121°C). Immediately after cracking the eggs into the frying pan, a pipette was used to deposit 0.1 mL of the Salmonella cocktail onto the top of the vitelline membrane of each test egg (2 eggs were used; Tharrington et al., 2003). Eggs were then mixed with a sterile spatula. The temperature of the eggs was measured during cooking with an infrared thermometer (infrared/type K thermometer, Fisher Scientific, Pittsburgh, PA) every 5 s in the centermost portion of the pan. After 2 min, the frying pan was removed from the hot plate (Toastmaster Inc., Macon, MO) and the eggs were scraped into a mound approximately 5.08 cm (2 in.) tall. A stainless steel pocket thermometer (Cole-Parmer) was immediately inserted into the thickest portion of the scrambled eggs to obtain the final temperature. After cracking the over-easy and sunny-side-up eggs into a preheated sterile nonstick frying pan (heated to at least 121°C), a pipette was used to deposit 0.1 mL of the Salmonella cocktail onto the top of the vitelline membrane

EFFECT OF COOKING ON SALMONELLA INACTIVATION IN SHELL EGGS

on the surface of the egg yolk. A wire type K thermocouple (Tangent Systems Inc.) attached to a Tempest data logger (Tangent Systems Inc.) was then placed into the vitelline membrane on top of the yolk, and a glass lid was placed on the frying pan. When the egg was ready to be turned, based on visual endpoint observations, the lid was removed and the thermocouple was gently removed so the egg could be turned. The thermocouple was placed back into the yolk through the cooked albumen, and the lid was replaced for an additional 30 s. The eggs cooked to sunny-side-up doneness were not turned. Immediately after cracking the egg into the 3-cup (711mL) sterile nonstick poaching pan insert, a pipette was used to deposit 0.1 mL of the Salmonella cocktail onto the top of the vitelline membrane surface of the egg yolk. A wire type K thermocouple attached to a Tempest data logger was then inserted into the vitelline membrane on top of the yolk, and the lid was placed on the pan. For the free-poached method, a temperature curve was made separately from each set of eggs that was inoculated. For the inoculated eggs, a sterile 21-gauge needle with the syringe prefilled with the Salmonella cocktail was inserted into the large end of the egg through the air cell. The eggs were inoculated in the shell because when the egg was cracked into boiling water, the inoculua could not be deposited on the egg surface. The eggs were aseptically cracked on sterile foil and opened into a ladle, and a type K wire thermocouple attached to a Tempest data logger was inserted into the yolk to record the temperature. The ladle was then submerged into the boiling water and the egg was allowed to cook for 3 min with the lid placed on the pan, as stated in the cooking section.

In-Shell Controls An in-shell control was used to validate the hard- and soft-cooked and free-poached cooking methods. A sterile syringe prefilled with sterile distilled water was inserted into the large end of the egg through the air cell, and 0.1 mL of sterile water was deposited onto the vitelline membrane of the yolk, using a Lyon Hi-Intensity Candler to confirm placement of the Salmonella suspension or distilled water. After injection, the hole was sealed with Super Glue. Eggs were cracked on sterile foil and opened into a sterile 1,000-mL beaker. The beaker was then covered with a piece of sterile foil, reweighed to obtain the egg weight, and enumerated for aerobic bacteria. A dilution of 1 to 10 was used, based on egg weight, to determine the amount of buffered peptone water (Biotrace, Muncie, IN) to add before mixing. A large mixer (Talboys/ Troemner, Thorofare, NJ) with sterile paddles was used to mix the egg and buffered peptone water dilution for 2 min. A sample (0.1 mL) of the diluted egg was removed and plated onto aerobic plate count (APC) 3M Petri Film (3M Microbiology Product, St. Paul, MN), which was incubated for 48 h at 37°C. The films were read by counting the red positive colonies.

1639

Broken-Out Controls This control was used to validate the scrambled, overeasy, sunny-side-up, and poached methods. Immediately after cracking the egg into a 1,000-mL beaker, a pipette was used to deposit 0.1 mL of sterile distilled water onto the vitelline membrane on the surface of the egg yolk. The beaker was then covered in a piece of sterile foil, reweighed to obtain the egg weight, and enumerated for aerobic bacteria.

Enumeration and Salmonella Detection For each cooking method, an APC was conducted to determine the number of aerobic bacteria, if any, that were present after cooking. The lowest detectable level of APC using 3M Petri Films (3M Microbiology Product) is 10 cfu/g. Control eggs were not taken past the enumeration step. Eggs in each cooking method also underwent microbiological testing to determine whether the samples were positive or negative for SE or SH. Another 0.1 mL was removed from the original diluted egg and placed into 10 mL of Rappaport-Vassiliadis R 10 broth (Hardy Diagnostics). A 0.5-mL sample was removed from the diluted egg and added to 10 mL of tetrathionate Hajna broth (Hardy Diagnostics, Santa Maria, CA). Both tubes were then placed in a 42°C incubator for 24 h. After 24 h, a 10-␮L loopful was removed from each tube and streaked for isolation onto BGS and modified lysine iron agar (MLIA; Hardy Diagnostics). The 4 total plates (each tetrathionate Hajna broth tube equaled one BGS plate and one MLIA plate; each Rappaport-Vassiliadis R 10 broth tube equaled one BGS plate and one MLIA plate) were then inverted and placed into an incubator and incubated for 24 h at 37°C. All positive plates (bright pink with colonies for BGS and black colonies for MLIA) were then stabbed and streaked with a sterile needle by stabbing the butt and streaking the slant onto triple-sugar iron agar and lysine iron agar slants and incubated for 24 h at 37°C. Positive tubes for lysine iron agar had a purple slant with a purple butt with or without hydrogen sulfide production. Positive triplesugar iron agar tubes had a red slant and a yellow butt with or without hydrogen sulfide production. Hydrogen sulfide production produces a black butt with a sulfur smell. A 10-␮L loopful was removed from the positives tubes and transferred into a cryovial that contained a mixture of 70% TSB and 30% glycerol that was added to the vial before and was placed into a −80°C ultracold freezer for confirmation at the end of the 3 wk. The positive samples were confirmed by using Salmonella O Antisera, H Antisera (BD Difco, Rockville, MD), Antiserum Vi group D and group B (BD Difco) to determine whether the positive samples belonged to group O, D, or B Salmonella. Groups B and D were used as controls to ensure that other naturally occurring Salmonella strains were not present and therefore did not confound the research. The positive cryovialed samples were regrown by using a 10-␮L loop from each cryovialed sample and

1640

DAVIS ET AL.

were placed into a 10-mL TSB tube to incubate for 24 h at 37°C. This was done 2 more times, after which a 10␮L loopful was removed and plated onto an APC plate. Agglutination was performed by using glass microscope slides. With a sterile plastic needle, a uniform colony was removed from an APC plate and mixed with the Salmonella antisera. The slide was then gently rocked for 1 min with a Fisher clinical rotator (Fisher Scientific). A test was considered positive if the background of the slide was slightly cloudy (coagulation), whereas in a negative test the sample showed no signs of coagulation.

Serotyping All positive samples were transferred to TSA slants and incubated for 24 h at 37°C. The samples were then shipped under refrigeration to the USDA National Veterinary Services Laboratory in Ames, Iowa, for phage typing.

for the inactivation of SE and SH. The results of this study showed that no Salmonella survived the hard- and softcooking methods. In fact, after 9 min of total cooking time, eggs from all 3 cooking methods reached 63°C, the temperature that was reported by Tharrington et al. (2003). For the scrambled cooking method, 1 egg out of the 30 that were cooked yielded viable SE phage types 4 and 30. Although phage type 4 was part of the suspension used in this study, phage type 30 was not; it must therefore be assumed that phage type 30 was in the egg before the inoculation. The APC following scrambled cooking was 1.1 log10 cfu/g (Table 1). Baker (1990) stated that scrambled eggs should be cooked for 1 min to be safe, whereas Tharrington et al. (2003) stated that 2 min was needed for the egg to reach the 63°C safe temperature. However, the results of this study showed that 2 min of cooking did not fully inactivate Salmonella. This could be because the variation in scrambling action did not allow all the egg mass to come in contact with the heat source long enough to be evenly heated.

Statistical Analysis A completely randomized design was used. An egg represented one experimental unit, except in the case of scrambled egg, where it was 2 eggs. For each of the 7 cooking treatments and controls, 30 eggs were used. The within-treatment effects on the percentage of Salmonellapositive samples were determined with a chi-square analysis by using the PROC FREQ procedure of SAS (SAS Institute, Cary, NC), whereas the total APC were analyzed by using the PROC GLM procedure of SAS. The PROC GLM was used to determine the P-values of each treatment to determine whether there was a significant difference within the weeks (3) of each treatment. Means were separated by using Fisher’s protected least significant differences.

RESULTS AND DISCUSSION The hard-cooked method had an APC reduction of 2.5 log10 cfu/g; the soft-cooked and poached methods had an APC reduction level of ≥2.5 log10 cfu/g; the scrambled and over-easy methods had an APC reduction of 2.4 log10 cfu/g; the free-poached method had an APC reduction level of 2.1 log10 cfu/g; and the sunny-side-up method had an APC reduction level of 0.7 log10 cfu/g. The sunnyside-up method had the lowest APC reduction level of all the cooking methods (Table 1).

Hard, Soft, and Scrambled Cooking Methods When the APC were done by using 3M Petri Films, the hard- and soft-cooked methods yielded ≤10 cfu/g (Table 1) and reached 63°C after 9 min of cooking. These results are similar to previous studies done by Licciardello et al. (1965), Humphrey et al. (1989b), Baker (1990), and Chantarapanont et al. (2001), which validated the use of the hard- and soft-cooking methods for eggs as reliable

Over Easy The over-easy method produced no eggs that were positive for SE; however, 1 egg yielded viable SH bacteria. The APC was 1.1 log10 cfu/g (Table 1) and resulted in a final temperature of 82.5°C in the center of the yolk or the coldest part of the egg. Baker (1990) determined that it would take a total of 5 min of cooking time to produce a SE-free egg. The 5-min total time describes the cooking method because the egg is cooked for 3 min and then turned over and cooked for an additional 2 min. The over-easy cooking method used in this study required that the egg be cooked for 1.5 min and then flipped and cooked for an additional 30 s. It could be speculated that an increase in cooking time after the egg was flipped could possibly eliminate the occurrence of Salmonella. When comparing the 2 studies, Baker (1990) produced a SE-free product but used a 5-min total cooking time for the over-easy egg, whereas in this study a 2-min total cooking time was used.

Sunny-Side Up Sunny-side up was much like the over-easy cooking method except that instead of flipping the egg at 1.5 min, the egg was removed from the heat completely. This method resulted in 19 positive samples out of the 30 samples that were cooked. Of those samples, SH was the most viable, and 16 of these were a combination of SE and SH. The 3 remaining samples yielded viable SE only. Of the SE phage types used, only 4 phage types, 4, 4a, 8, and 23, were recovered and no other phage types that were not in the suspension were found. The APC of the cooked product was 2.8 log10 cfu/g (Table 1) and the final temperature was 58.9°C. Baker (1990) observed that the time needed for complete kill with this cooking method was 7 min. Tharrington et al. (2003) determined that eggs cooked in this manner without the use of the basting

1641

EFFECT OF COOKING ON SALMONELLA INACTIVATION IN SHELL EGGS Table 1. Summary of cooking methods and recovery of viable bacteria

1

Initial bacterial population aerobic plate count2 (cfu/g, raw)

Initial bacterial population of Salmonella3 (cfu/g, raw)

Postcook population sample aerobic plate count (cfu/g)

Hard cooked Soft cooked Scrambled

2.0 2.0 2.0

3.5 3.5 3.5