The effect of crust freezing on the quality of raw chicken breasts, with or without skin, during aerobic, refrigerated storage for up to 18 days was assessed by means of International Commission on Illumination (CIE) color parameters L*, a*, and b*; tenderness; and total aerobic (APC) and yeasts and molds counts (YMC). Skin-on breasts had significantly higher L* values compared to skinless units (average 75 vs. 55), whereas a* and b* remained relatively constant regardless of the presence or absence of skin, freezing, or time. For a* and b*, values oscillated between -2.10 to 0.78 and 1.38 to 3.77, respectively. Shear energy varied erratically for skinless samples but tended to remain constant throughout time for skin-on breasts. APC increased over time and exceeded 8.0 log10CFU/ml of rinse, which occurred between 6 and 12 days of storage. Under the experimental conditions used, crust freezing did not affect color or tenderness of raw chicken breasts, with or without skin but also did not extend microbiological shelf life.
Key words: crust freezing, shelf life, chicken breasts, skin
Raw poultry products are susceptible to spoilage and deterioration. Several freezing applications at the commercial level claim toimprove quality attributes or extend shelf life of these products. Crust freezing, where the product is semi-frozen only at superficial layers of tissue, is one such application. The current study found that crust freezing did not affect quality attributes of raw chicken breasts, with or without skin, but neither did it extend their shelf life under aerobic, refrigerated storage, based on aerobic plate count values. Although further investigation is needed, this research serves as the foundation for other studies to benefit industry. Since crust freezing is widely accepted as a shelf life extending treatment, the results of this study imply that the freezing method, length of the crust freeze time, and the product type and form may influence the effectiveness of crust freezing for shelf life extension.
The increase in consumption and production of poultry meat has progressively moved from whole carcasses to cut-up parts and boneless meat. Boneless, skinless chicken breast fillets are of primary economical importance and more than 60% of consumers purchase boneless chicken breast meat when purchasing chicken products (Rotabakk et al. 2006; Zhuang et al. 2007; del Rio et al. 2007). However, poultry products are highly perishable foods. Depending on the degree of processing following slaughter, their spoilage varies between 4 and 10 days under refrigeration. Mesophilic aerobic counts, psychrotrophs, Enterobacteriaceae, coliforms, and yeasts and molds are general indicators of processing hygiene, storage quality and potential shelf life in aerobic atmospheres (Charles et al. 2006; Patsias et al., 2008; Guevara-Franco et al. 2010). Quality, including taste, color, freshness and tenderness of chicken meat are major components of consumer satisfaction and the major marketing emphasis by chicken processors (Kumar et al. 2007). Freezing is a widely accepted preservation method used to store meat and poultry and is the safest and most efficient way to maintain product quality for long-term storage. However, poultry meat quality may decline during long periods of frozen storage (Lee et al. 2008b). Crust freezing is defined as an operation in which only the upper 10 to 20 mm deep in the tissue is frozen (James and James, 2002). The core deep in the unit remains unfrozen and it is used to produce optimum temperature in a chilled product so that it is suitable for cutting, portioning or cubing. In this case, the product is semi-frozen so that it is stiff enough to be portioned. Crust freezing can be used to temporarily stiffen products which are not to be totally frozen but are subsequently kept in refrigeration. The crust allows handling and packing with less damage and according to commercial claims, it could potentially improve quality attributes and extend the shelf life of raw poultry products. The goal of this study was to evaluate the effect of crust freezing on the color, tenderness and total microbial loads of raw chicken breasts with and without skin.
Materials and Methods
Materials and sample preparation
Fresh, clean chicken carcasses were kindly provided by a local commercial processing plant in Greenville, SC and transported in ice directly to the laboratory about one hour after slaughter. All of the chicken carcasses used for a given experimental replicate were procured from the same process lot in order to reduce variability among samples. Samples were not frozen prior to performing the experiments so that the normal microflora and microbial load would remain as intact as possible. Upon arrival to the laboratory, the chicken carcasses were kept under refrigeration at 4Â±2 Â°C for no more than four hours until preparation. Using aseptic technique, whole chicken breasts in each batch were manually separated from the carcasses and randomly assigned to a "skin-on" or a "skinless" group then randomly assigned to one of five possible storage times (control, or 18 h; 3, 6, 12, or 18 days). Five breasts with either skin-on or skin-off were then split longitudinally into two halves and each half was randomly assigned to one of two possible freezing treatments (control or crust-freezing).
Crust freezing, thawing, and sampling periods
All chicken breasts assigned to the crust freezing group, with or without skin, were placed in clear, plastic freezing bags and individually frozen at -85 Â°C for 20 min, time needed for a visible crust to be formed homogenously around the product (approximately 20 mm deep, as determined by preliminary tests). After the crust freezing treatment, bags containing the samples were placed under refrigeration at 4Â±2 Â°C, along with their non-frozen counterparts, prior to sampling and quality analyses. Control or non-frozen samples were handled in the same way except that they were not crust frozen but placed directly under refrigeration for 18 h to ensure complete thawing. Microbial and quality analyses were performed on treatment and control samples at 18 h (control), 3, 6, 12, and 18 days of refrigerated storage.
Refrigerated chicken breasts were cooked prior to measuring instrumental texture in order to determine treatment effects on meat tenderness. Samples were wrapped in aluminum foil, placed on a tray and steamed in an autoclave for 15 min so that the internal temperature in the thickest part of the breast reached at least 74 °C. In order to determine the internal temperature after steaming, a fine thermocouple (Omega Engineering, Stamford, CT) was inserted in the thickest part of chicken breasts of similar size during preliminary studies (data not shown) and the temperature measured using a digital multimeter (Omega, Engineering, Stamford, CT). After cooling down to room temperature (22-23 °C) in open air for at least 60 min, instrumental texture was measured on each breast as described below. Since the samples were rinsed prior to cooking (see below), excess water was removed by draining before wrapping in aluminum foil and proceeding with the cooking step.
Microbiological and quality analyses
Instrumental color and texture, aerobic plate counts (APC), and total yeasts and molds counts were measured at every sampling period. All of the measurements were taken on a breast in the following order: color, APC along with yeasts and molds, and finally texture, which was the only destructive assay.
Color. Instrumental color analysis was based on at least five measurements of light reflected from the surface of each raw chicken breast, measured at the center and edges of every sample. International Commission of Illumination (CIE) lightness (L*), redness (a*), and yellowness (b*) values were obtained using a Chroma Meter with an 8 mm viewing port and illuminant D65 (CR-300, Minolta Corp, Ramsey, NJ). The instrument was calibrated against a standard white ceramic tile immediately before the measurements were taken.
Microbial analyses. Fifty milliliters of sterile Bacto Peptone water (0.1% wt/vol; Becton Dickinson, Sparks, MD) were added to each freezing bag containing a raw chicken breast. The pieces were then massaged gently by hand for 30 s to disperse bacteria into the peptone water within the bag. The rinse solution was then serially diluted in sterile Bacto Peptone water (0.1% wt/vol; Becton Dickinson, Sparks, MD) and handled as follows:
ï‚§Â For Aerobic Plate Counts, serial dilutions were surface-plated in duplicate on Difco Plate Count Agar (PCA; Becton Dickinson, Sparks, MD). Colonies were counted on plates with 25 to 250 colonies after incubation at 37 °C for 48 h on a Quebec colony counter. Bacterial populations were converted to log10 CFU/ml of sample rinse.
ï‚§Â For yeasts and molds, serial dilutions were surface-plated in duplicate on Difco Dichloran-Rose Bengal-Chloramphenicol (DRBC; Becton Dickinson, Sparks, MD). Plates were incubated at room temperature (22-23 °C) protected from light. Colonies were counted on plates with 10 to 150 colonies after 5 days of incubation on a Quebec colony counter. Microbial populations were converted to log10 CFU/ml of sample rinse.
Instrumental Texture Analysis. Chicken meat tenderness was evaluated by shear energy using a TX.XTTexture Analyzer (Stable Micro Systems Inc, Surrey, UK) connected to a PC for data logging via Texture Exponent (TEE) 32 version 18.104.22.168, following the procedure described by Cavitt et al. (2004) with modifications. Briefly, a razor blade probe with a height of 24 mm and a width of 8.9 mm set to a penetration depth of 20 mm was used to compress the muscle tissue perpendicularly to the muscle fibers after equilibration to room temperature. Instrumental blades were replaced after every 50 samples and recalibrated to eliminate error due to dulling of blades. Shear energy and shear force on intact cooked breasts were recorded in at least five different locations on each breast to obtain mean and standard error data. The instrument settings were: maximum cell load: 2 kg; probe pre-test speed: 2 mm/s; test speed: 10 mm/s; post-test speed: 10 mm/s; trigger force: 10-g contact force. Shear force (g) was defined as the maximum force recorded and shear energy (g*mm) was considered to be the area under the force-deformation curve from the beginning to the end of the test.
The experiment was designed as a split-plot study with two whole plots (presence of skin and storage time) and one sub-plot (freezing type) and replicated two times from lots collected on different days of processing using a total of ten chicken carcasses per replicate (see Material and Sample Preparation above for an explanation of how chicken breasts were allocated to treatments.) The lot from which the carcasses were obtained was considered a blocking factor. Analysis of variance of the APC, yeasts and molds counts, CIE color parameters L*, a* and b*, and shear energy were performed using the PROC MIXED command on the Statistical Analysis System (SAS) version 9.2 (SAS Institute, Cary, NC). Bonferroni's inequality was used to estimate multiple comparison error rates at a 5% level of significance.
Results and Discussion
Crust freezing had little effect on the color of chicken breasts as determined by CIE instrumental chromatic attributes. Unfrozen, skinless samples did not have significantly different L* values (p>0.05) relative to their crust-frozen, skinless counterparts (Table 1). However, average lightness differences were smaller than two units. Unfrozen and crust-frozen samples that retained skin did not show significant differences in L* values (average 74.64 and 75.14, respectively). Differences in L* values were observed between skinless and skin-on chicken breasts (p<0.0001). On average, skin-on samples were 20 units lighter than skin-off units (75 vs. 55), regardless of the application of a freezing treatment. These results are in agreement with those of Sirri et al. (2010) who found an average L*value of 75.40 for 2,300 yellow-skinned broiler chicken breasts in Italy. Likewise, the mean L* values of approximately 55 for skinless and 75 for skin-on samples reported here are comparable to those obtained by Mielnik et al. (1999) of between 66 and 77 lightness units, depending on whether the measurement was made on the front or back side of the breast and on the type and rate of chilling used. This suggested a possible interaction between skin and freezing method in determining L* values. This interaction was found to be significant in the present study (p=0.0046), mainly due to major differences among samples with skin off compared to those with skin on, as described above. In this study, all of the values were taken from the front or upper side of the chicken breasts in order to be consistent with the location of the measurements and to be able to provide objective evidence of treatment effects.
On the other hand, only small variations were observed in L* values throughout the storage period. Like in the present study, Rotabakk et al. (2006) found that average L* values of skinless chicken breasts fillets stored aerobically in refrigeration for up to 24 days was 56 and did not change significantly over storage time. Similarly, in a study on the effect of freeze chilling and modified atmosphere packaging on quality parameters of raw chicken fillets, Patsias et al. (2008) did not find differences in L* values during aerobic, refrigerated storage with values ranging from 50 to 60, similar to those found in the present investigation.
Redness values (a*) remained fairly stable throughout storage with a notable difference at 18 days of storage, when samples with skin on had significantly higher a* values than their counterparts sampled at day 12 for both unfrozen and crust-frozen units (Table 2). This indicates a significant interaction (p<0.0001) between presence of skin and freezing in a* values of chicken breasts, particularly at the latter stages of aerobic, refrigerated storage. However, there was no discernible pattern to indicate why the interaction was significant. The behavior of yellowness values (b*) on chicken breasts were determined by the three-way interaction between the factors, which turned out significant (p<0.01). No differences in yellowness values (b*) were seen between unfrozen and crust-frozen units on days 3, 6, and 12, regardless of the presence of skin (Table 3). The only difference occurred at day 18, when skin-on, crust-frozen units had higher b* values than their corresponding skin-on, unfrozen pairs. Average b* values for unfrozen, skin-on samples were consistently lower than their crust-frozen counterparts and b* values of chicken breasts tended to increase throughout time becoming significantly higher by day 12 of storage; however, values tended to decrease by day 18.
The a* values obtained for samples in this study are similar to those reported by Sirri et al. (2010) who obtained a mean of 1.16 for chicken breasts, whereas thighs and shanks had lower values. However, b* values here were much lower than the average 22.77 reported by the same authors. The overall range for redness was -2.10 to 0.78. This range was considered small compared to reports in the literature of up to 12 units (Petracci et al. 2004). One possible reason is the natural difference in the concentration of xanthophylls and other pigments present in the feed and later deposited in the epidermis. Furthermore, low myoglobin content of the chicken breasts might have influenced their low a* values. Still, visually, samples were considered to be slightly red or pink. The range for yellowness was 1.38 to 3.79, consistent with an average b* of 3.62 obtained by Wattanachant et al. (2004) and 2.08 reported by Petracci et al. (2004) for broilers chicken breasts in Thailand and Italy, respectively. The latter authors also report a range for b* values from -3 to 12, indicating that high natural variability among samples play an important role in color determination. Studies have shown higher a* and b* values for chicken breasts subjected to blast chilling, of approximately 3 and 12, respectively (Patsias et al. 2008), whereas freezing and frozen storage for up to eight months yielded a* and b* mean values of 3.5 and 2.5, respectively (Lee et al. 2008b). Reduced freezer burn compared to prolonged frozen storage could potentially be the reason why crust-frozen units in this study presented lower b* values than in other reports.
Color variation is a major problem in the retail environment because consumers are more sensitive to color variation than to absolute color. Instrumental color data was supported by the fact that no visual discoloration was observed throughout the storage period. The processing method, the packaging conditions, the degree of exposure to light and other interactive effects can influence changes in visual color of the products (Petracci et al. 2004; Modi et al. 2006). In the present investigation, the reflectance was stable during storage. Thus, differences observed in other studies could be attributed to natural variability among the chicken meat or skin, to random location of color readings on each unit, or to process variability during scalding, and not necessarily to the treatments (Ellis et al. 2006). The results indicate that chicken breast lightness, redness and yellowness undergo only small changes during aerobic, refrigerated storage and therefore, they are not conclusive parameters in determining shelf life of this kind of food product.
Tenderness is another critical quality and palatability attribute for chicken breast meat (Barbanti and Pasquini 2005; Zhuang et al. 2007). Chicken meat texture has been measured instrumentally using several different probes, the most common being Warner-Bratzler shear-type blade, Allo-Kramer shear, Razor Blade shear and needle puncture. The use of a razor blade shear, like the one used in this study, has the advantage of being less time-consuming than the other tests as it requires no weighing or further sample preparation other than cooking (Young and Lyon 1997; Cavitt et al. 2004; Cavitt et al. 2005a; Thielke et al. 2005; Zhuang et al. 2007; Del Olmo et al. 2010). Additionally, the razor blade probe has proven to perform similar to Warner-Bratzler shear-type blade, the typical reference method. Studies by Cavitt et al. (2005b) and Xiong et al. (2006) conclude that all of the instrumental shear stress probes previously mentioned perform similarly for predicting the tenderness of cooked broiler breast meat and correlate well with descriptive and sensory analysis, therefore making the razor blade probe suitable option for measuring tenderness.
As shown in Table 4, differences in tenderness of unfrozen and crust-frozen samples were observed on days 6 and 12 of aerobic, refrigerated storage, particularly for skin-on units (p<0.05). When these differences were present, crust-frozen samples tended to have significantly lower shear energy than unfrozen units, in other words, the crust-frozen samples were more tender (McKee 2007). Shear energy values for skin-on samples remained stable throughout time likely because the test accounts for the firmness levels of the skin, not the meat directly. Therefore, even when the meat underneath the skin could potentially lack firmness, the skin on top may remain firm, as reflected by the data showed. The values measured at day 18 for were higher than their crust-frozen counterparts at day 12, regardless of the presence of skin but the differences were significant only for skinless units. On the other hand, skinless samples showed erratic behavior and tended to have lower, though not significantly, shear energy values than their skin-on pairs, which is consistent with the fact that skin would oppose higher resistance to biting and chewing than the meat tissue itself. Skinless, crust-frozen samples had unusually lower shear energy by day 6 and regained firmness by day 18 of storage. Differences noted intenderness measured as shear energy may be attributed to greater or lower activation rates of calpains acting on proteolysis of the meat muscle. This change in calpains activity may be due to pH variation in the muscle as the tissue ages and to change in the concentration of Ca2+ ion during storage (Lee et al. 2008b).
According to Lee et al. (2008b), refrigerated storage does not cause significant muscle shrinkage and therefore, softening of the tissue is expected, as opposed to toughening, which occurs during prolonged frozen storage. Loss of firmness as determined by decreasing shear energy values may be a consequence of microbial enzymatic activities, particularly proteolysis caused by Pseudomonas and yeasts and molds. Charles et al. (2006) indicated that when total microbial counts reach 108 logCFU/g, decomposition of the muscle tissue is evident by surface slime formation. In the present investigation, the loss of firmness of the tissue as storage time increases may be due to increasing levels of bacteria, yeasts, and molds throughout time. This relationship, however, was not evaluated statistically.
Table 5 shows the results for aerobic plate counts (APC) of chicken breasts subjected to crust freezing and stored in refrigeration for up to 18 days. Freezing along with prolonged frozen storage have proven to be effective in reducing the number of bacteria in processed chicken products. This was reported by Modi et al. (2006), who froze chicken curry for up to 6 months. In the present study, rapid surface freezing of skinless chicken breasts followed by refrigerated storage did not cause differences in microbial growth between crust-frozen units and their unfrozen pairs. For skin-on samples, the only difference was observed between 3 and 6 days of storage. Time of storage, along with freezing, significantly influenced APC values of chicken breasts. At the two first stages of refrigerated storage, higher bacterial counts (p<0.05) were detected on skin-on crust-frozen chicken breasts compared to their unfrozen pairs. However, the same trend was not evident at 12 and 18 days of storage, when the counts stabilized across time. This means that freezing also interacted significantly with skin for total aerobic bacterial counts of the chicken breasts (p=0.0062). Skin-on, crust frozen samples had significantly higher bacterial counts than their skin-on, unfrozen counterparts at days 3 and 6 but not for days 12 and 18, despite being higher. It has been hypothesized that the quick decrease in temperature to the freezing state opens up the product structure, resulting in a greater recovery of bacteria trapped in crevices and this may be the reason why crust-frozen samples showed higher bacterial counts that their unfrozen counterparts (Thomas and McMeekin 1981; Lillard 1988; Fagan et al. 2003; Patsias et al. 2008).
Initial low levels of APC (mean of 3.5 log10CFU/ml 0.1% peptone water) in chicken breasts either with skin off or skin on are indicative of good hygiene during processing. Since samples were brought directly to the laboratory shortly after slaughter and kept on ice during transport, handling was minimal, thus reducing the potential for contamination and bacterial growth (Modi et al. 2006). An average increase of 3 log10CFU/ml between 6 and 12 days was observed for all groups. No significant differences were noted in APC values between 12 and 18 days of storage, as described before. The average increase between these two sampling times was 1.4 log10CFU/ml. Maximum APC values were in the order of 1010 with an average increase of 2.0 log10CFU/ml between consecutive sampling times. Similarly, Rotabakk et al. (2006)found that time of storage was a significant factor in determining APC values, no matter what other factors were tested. Their results, however, tended to be lower for equivalent storage times compared to the results in the present investigation. These differences might be due to methodology employed for bacterial recovery or initial levels of contamination and handling practices prior to enumeration. In this study, bacterial recovery was done by rinsing. Therefore, the results are expressed as log10CFU/ml 0.1% rinse water, whereas in the Rotabakk et al. study, the recovery was done by excision of muscle tissue.
For a given sampling time, no significant differences were seen among skin-off or skin-on samples no matter the application of a freezing treatment. Initially, this was not considered to be the expected outcome, since the skin of poultry is known to retain a great proportion of the total number of bacteria, and certainly the bacterial levels are expected to be higher for skin than for muscle (Daud et al. 1979; Thomas and McMeekin 1981). However, the results of this study are supported by the observations of Berrang et al. (2000) who noted that bacterial populations on skin plus meat, skin alone or meat alone, recovered from split breasts, thighs and drumsticks of broilers purchased at a retail outlet and aseptically skinned in the laboratory were not significantly different from one another. The authors conclude that processing, particularly immersion in the chill tank, allows the counts on the bone-in meat, which is mainly covered by skin, to equalize to that on the skin itself (Thomas and McMeekin, 1981). Additionally, when cutting-up the carcass the muscle surfaces are compromised by exposing them to skin and allowing transfer of water and dissolved substances from skin to meat tissue (Berrang et al. 2001). This is certainly a potential hypothesis for the results of this investigation. Other temperature reduction methods, such as evaporative air chilling and freeze-chilling have also proven ineffective in reducing spoilage microorganisms in chicken carcasses (Mielnik, et al. 1999; Patsias et al. 2008).
Finally, total Yeasts and Molds Counts (YMC) results paralleled those of Aerobic Plate Counts (Table 6). In general, no differences were observed between unfrozen and crust-frozen chicken breasts, regardless of the presence of skin. As an individual factor, the presence of skin did not influence yeasts and molds counts but interacted significantly with freezing and time (p=0.0023), which were individually significant. The average increase in counts from 6 to 12 days was 1.6 log10 CFU/ml. From 12 to 18 days of storage, the APC counts increased 1 log10 CFU/ml in average. Maximum YMC were in the order of 105 CFU/ml with an average increase of 1.0 log10CFU/ml observed between consecutive sampling times. Initial mean population of yeasts and molds in raw, unprocessed chicken breasts was 3.0 log10CFU/ml, which are similar to those reported by Ismail et al. (2000). The Ismail et al. study found that after two weeks of storage, the final mean yeasts and molds counts of raw chicken breasts and carcasses were 3.7 and 5.0 log10CFU/g, respectively, equivalent to the results of the present study. The same authors also identified Yarrowia lipolytica, Candida zelanoides and basidiomycetous yeasts as major isolates, the first one being partly responsible for proteolytic and lipolytic spoilage of chicken meat surfaces. Total counts of yeasts and molds remain a small part of the spoilage microflora throughout time, as noted above. According to Thomas and McMeekin (1981), thesemicroorganisms are able to grow on the skin and muscle of chicken products stored under refrigeration but fail to compete with the pseudomonads and remain an insignificant proportion of the spoilage microflora. The equilibration in competitive flora may be the reason why yeasts and molds counts tended to decrease from day 3 to 6 and then increased again by day 12, although these changes did not turn significant.
In general, microbial counts increased during storage time as expected, with total bacteria loads being higher than those of yeasts and molds. Before 12 days of aerobic, refrigerated storage, APC values reached unacceptable levels of 8 log10CFU/ml 0.1% peptone water, which is generally considered a cut-off point for shelf life of raw chicken meat (Charles et al. 2006; Patsias et al. 2008; Rotabakk et al. 2006), with or without skin, unfrozen or subjected to crust freezing.
Crust freezing is a commercially available application aimed to improve quality and extend shelf life of highly perishable foods, such as raw poultry products. In the present study, color attributes L*, a*, and b* did not change greatly over 18 days of refrigerated, aerobic storage and were mostly affected by the presence of skin. Bacterial counts reached unacceptable levels between 6 and 12 days of storage. Yeasts and molds counts remained low throughout storage time. Finally, tenderness, measured by razor blade shear energy, tended to decrease progressively due to deterioration of the meat tissue and was the quality measure most affected by storage time. Under the conditions used, crust freezing did not extend shelf life nor affect quality of raw chicken breasts, with or without skin. In the current study, a static crust freezing method was employed that was closely monitored to mimic the process used commercially and to adhere to the "fresh, not frozen" criteria used commercially. Other factors used in commercial crust freezing of chicken may impact shelf life and may warrant further investigation.