A survey of broiler breast meat quality in the retail market of Quebec

Abstract In this study, 206 breast fillets were purchased from grocery stores in the province of Quebec and evaluated for the presence of different quality defects. Of these fillets, 48.5% showed breast muscle myopathies (BMM), 19.4% showed pale, soft, and exudative (PSE), and 6.8% showed dark, firm, and dry (DFD) attributes. BMM were equally present (P > 0.05) in fillets of economical, commercial, and high-quality brands, while PSE-like fillets were more present in economical brands (P < 0.0001). The combined effect of BMM and DFD induced significantly higher counts of Salmonella (P = 0.03) and Enterobacteriaceae (P = 0.03) in myopathic than in unaffected fillets. These quality defects also altered the nutritional quality of breast meat: BMM-affected fillets had greater fat content (P < 0.0001) and DFD fillets had lower protein content (P = 0.041) than normal fillets. The technological quality was only slightly impacted by BMM, while PSE-like fillets had higher cooking loss (P = 0.009) and a tougher texture after cooking (P < 0.0001) than DFD fillets. For the first time, this study confirmed the presence of multiple quality issues in the Quebec poultry supply chain, and provided valuable data to support future research efforts.


Introduction
Poultry accounted for 41% (135 MT) of the global meat consumption in 2020, and is expected to represent 52% of all produced meat worldwide by 2030 (OECD 2022).To meet this increasing demand, the poultry industry combines genetic selection, optimized nutritional strategies, and flock management practices to improve production.Genetic selection is the main tool used by the industry to develop strains of broiler chickens with higher growth rates, greater breast meat yields, and enhanced feed efficiency (Zuidhof et al. 2014;Petracci et al. 2015).Over time, increased selection pressure to further improve broiler's performance has led to unintended changes in muscle structure and metabolism.These changes include increased muscle fibre diameters (i.e., muscle hypertrophy), reduced perimysium spaces, reduced capillary density, and reduced glycogen reserves in the pectoralis major muscle (Abasht et al. 2016;Velleman 2019;Praud et al. 2020).Evidence suggests that hypoxia and low energy storage in breast muscles are major determinants in the inducement of histopathological changes characteristic of emerging meat quality issues such as breast muscle myopathies (BMM), including white striping (WS), wooden breast (WB), and spaghetti meat (SM) (Soglia et al. 2016(Soglia et al. , 2017;;Malila et al. 2019;Baldi et al. 2020;Mazzoni et al. 2020).Evidence also suggests that changes induced by hypoxia, including reduced mitochondrial activity and increased oxidative stress, are aggravating muscle damage and increasing the severity of these myopathies (Praud et al. 2020;Li et al. 2022).
The consequences of these quality defects are well documented for both the consumer and the processing industry.For instance, Kuttappan et al. (2012b) conducted an analysis of consumer acceptance of breast fillets categorized as normal (NORM), moderately affected, and severely affected (SEV) by white striping.The results of their study showed that 57% of consumers disliked SEV fillets and over 50% said that they would definitely not buy these fillets because they associated them with being fattier than NORM fillets.According to de Carvalho et al. (2020), the more consumers are informed about BMM, the less likely they are to accept and purchase BMM-affected raw chicken fillets.These studies highlight the negative impact of BMM on consumer acceptance and purchasing behaviour for chicken breast meat, which can lead to a negative economic impact on the poultry industry.
The consequences of these quality issues go beyond their impact on the visual aspect of raw breast meat and on the consumer's perception of breast meat quality.In fact, these quality defects modify both the technological and nutritional qualities of breast meat.For instance, breast fillets affected by severe WS and/or WB show significantly reduced cooking yield and increased hardness compared with normal fillets (Mudalal et al. 2015;Thanatsang et al. 2020;Wang et al. 2022).Given the large volumes of breast meat transformed by the global processing industry, these changes can result in considerable economic losses.BMM are associated with significantly increased intramuscular fat (up to +0.6%) content and reduced protein content (up to -2.0%) (Baldi et al. 2019;Tasoniero et al. 2020), which can damage consumer perception of broiler breast meat as a healthy alternative to red meat.
It is now well established that genetics account for 50%-65% of the variability of BMM (Alnahhas et al. 2016;Lake et al. 2021) and that fast-growing, high-yielding strains of broiler chickens are the most susceptible to develop BMM (Kuttappan et al. 2012a;Petracci et al. 2015;Santos et al. 2021).Although the same fast-growing broiler strains are used worldwide, the prevalence of BMM varies between countries, genotypes, and genders (Petracci et al. 2019; Caldas-Cueva and Owens 2020).In Canada, little is known about the state of BMM in the poultry industry.In a recent study, the prevalence of these myopathies was evaluated on 9250 breast fillets sampled from 37 flocks of broiler chickens processed in slaughterhouses in Ontario (Che et al. 2022a), the largest poultry meat-producing province in Canada.This study showed a prevalence of 36%, 12%, and 96% for SM, severe WB, and mild or moderate forms of WS, respectively.
In Quebec, the second-largest broiler meat production province, no data are yet available on the presence of myopathies or about other breast meat quality issues, to the best of our knowledge.The aim of the present study thus was to evaluate the presence of quality defects in the grocery stores in Quebec, and to analyze their impact on poultry meat's microbial, technological, and nutritional properties in the local context.

Sample collection
In Quebec, Ross 308 and Cobb 500 are the two most wildly used broiler strains.These fast-growing strains are used in the conventional and alternative (organic and free-range) systems because of the lack of medium-or slow-growing strains on the Canadian market.Additionally, the largest market segment in Quebec consists of broilers processed at 2.15-2.5 kg at the age of 35 and 45 days in the conventional and alternative systems, respectively.For the purposes of this study, we randomly purchased 206 broiler breast fillets from a variety of grocery stores between January and April in the Saint Lawrence River valley, which is the most demographically dense area in the province of Quebec both in terms of population and agricultural activities.At each store visit, samples (n = 20-23 fillets) representing all available brands at the store were randomly selected and transported on ice to the pilot plant at the Faculty of Agricultural and Food Sciences (Université Laval, Quebec, QC, Canada), where they were immediately transferred to a cold processing room (4 • C) for analysis.Samples were then given unique individual identifiers and information about the store, brand, purchase date, best before date, weight, and price per kg was recorded.In this study, we chose to sample breast fillets from grocery stores because our aim was to evaluate the presence of differ-ent quality defects at the final product level and to analyze their consequences for different meat quality attributes on the product as purchased by the final consumer.

Microbial analyses
Immediately after recording the relevant information, the packaging containing breast fillets was aseptically opened to remove the plastic film covering it.Next, the ventral (skin-side) surface of each fillet was swabbed using a sterile sponge (Whirl-Pak Speci-Sponge Environmental Surface Sampling Bags, Nasco Sampling/Whirl-Pak, Madison, WI, USA) moistened with 10 mL of sterile 2% buffered peptone water (77187, Sigma-Aldrich, Saint-Louis, MO, USA).The sponges were maintained on ice until the microbial analyses were performed within the following hours.The length and width of each fillet were measured and recorded to calculate the swabbed surface area.Total aerobic mesophilic (TAM) counts were performed on plate count agar medium (Becton, Dickinson and Company, NJ, USA) incubated at 35 • C for 48 h (Health Canada 2020).Enterobacteriaceae counts were evaluated on 3M Petrifilm™ and were incubated at 37 • C for 24 h (Health Canada 2007).Salmonella spp.were counted on selective CHROMagar™ Salmonella plus medium (CHROMagar Microbiology, Paris, France) and incubated at 37 • C for 24 h according to the manufacturer's recommendation.
All measurements were carried out in duplicate, and bacterial counts were transformed to a log 10 value of colony-forming units per 20 cm 2 (log 10 CFU/20 cm 2 ) prior to statistical analysis.

Meat quality
Immediately after the swabbing for microbial analyses, breast muscle pH and colour parameters (L * , a * and b * ) were measured as described previously (Alnahhas et al. 2014).Briefly, the pH was measured using a portable pH meter (Ross, Orion Star A221, Thermo Scientific, Beverly, CA, USA) combined with an Orion Kniphe electrode (Thermo Fisher, Nepean, ON, Canada) and a temperature compensation probe (928007MD, Micro ATC probes, Maryland, USA) by direct insertion in the thickest (cranial) part of the muscle from the dorsal (bone side) surface of the pectoralis major.As for meat colour, it was measured on the same spot as the pH using a Chromameter (Chromameter CR-400, Minolta Ltd., Osaka, Japan) equipped with a conical open port and an 8 mm aperture, a diffuse illumination with 0 • viewing angle geometry, and a D65 light source according to the CIE trichromatic colour system (Petracci and Baéza 2011).Next, fillets were individually weighed and evaluated for the presence of WS, WB and SM according to the scales described in the literature (Kuttappan et al. 2016;Baldi et al. 2021).Then, samples were taken for measurement of the proximal composition, lipid and protein oxidation, and cooking loss (CL; see relevant sections below and Fig. 1).
The CL was measured by taking a muscle sample (average weight of 60 ± 2 g) from the middle part of breast fillets, placed in a plastic bag (Whirl-Pak bag, Nasco Whirl-Pak , USA), cooked in a water bath at 85 • C until an internal temperature of 76 • C was reached (Petracci and Baéza 2011).After Fig. 1.Sampling schema: cores A, B, and C were sampled, after the pH and the colour were measured in the same zone, to measure primary lipid oxidation products (FOX), secondary lipid oxidation production (TBARS), and protein oxidation (DNPH), respectively.The blue vertical and horizontal discontinuous lines correspond to the length and width of breast fillets, respectively.The green rectangle corresponds to the part of the fillet that was sampled for cooking loss and Warner-Bratzler shear force of cooked meat.Parts 1 and 2 correspond to parts sampled for the proximal analyses.cooking, samples were cooled down in an iced water bath for 10 min, taken out of their bags, wiped gently with an absorbent paper and weighted again.The CL was expressed as a percentage from sample's initial weight before cooking.Cooked meat samples were then cut into stripes (1 cm × 1 cm × 3 cm) parallel to the direction of the muscle fibres to evaluate their shear force using a texture analyzer (ZwickiLine, Zwick/Roell, Germany) equipped with a Warner-Bratzler blade moving at crosshead speed of 150 mm/min.This test was performed in triplicates and the average of the maximum force (N/cm 2 ) required to shear the replicates was reported.

Proximate analyses
Proximate composition (moisture, ash, protein, and lipid contents) of breast fillets was measured in duplicates for each sample according to official methods (AOAC 1995).Samples were taken from the cranial and caudal parts of breast fillets (30 ± 1 g) and stored at −20 • C until they were analyzed.First, samples were lyophilized (model 50L Virtual EL-85, VirTis, Los Angeles, CA, USA) at −80 • C, under a 100 mTorr vacuum and a shelf temperature of 20 • C for 7 days.Then, the cranial and caudal parts of each sample were homogenized together using a grinder (Knife Mill Grindomix GM 300, Haan NRW, DE) at a speed of 2000 rpm for 30 s and a unique sample per fillet was taken for all analyses.

Lipid peroxidation and protein oxidation
Samples used to measure oxidation parameters were all stored at −80 • C until they were analyzed.Primary products of lipid oxidation, including lipid hydroperoxides, were determined on samples using the ferrous oxidation-xylenol orange (FOX) method as described by Grau et al. (2000).Briefly, the equivalent concentration of cumene hydroperoxide (CHP) was measured in the supernatant after homogenization (VDI 25, VWR, Radnor, PA, USA) of 3 g of sample in cold 100% HPLC grade methanol and centrifugation at 3000g for 10 min at 4 • C. In a new tube, these solutions were added in the following order and mixed: 250 μL of a 1 mmol/L aqueous solution of (NH 4 ) 2 Fe(SO 4 ) 2 , 100 μL of a methanol solution of 25 mmol/L of H 2 SO 4 , 100 μL of a methanol solution of 0.1 mmol/L of xylenol orange, 450 μL of MeOH, and 100 μL of the supernatant.The standards consisted of 100 μL of a CHP solution of known concentration.The hydroperoxide concentration (mg of CHP equivalent/kg) corresponded to the absorbance at 580 nm (Varioskan, Thermo Fisher Scientific, Waltham, MA, USA) after a 60 min incubation in the absence of light.
The secondary products of lipid peroxidation were quantified in duplicates of samples using the thiobarbituric acid reactive substances index (TBARS) according to Lynch and Frei (1993).Briefly, samples (5 g) were homogenized in 10 mL of PBS (Sigma-Aldrich, St. Louis, MO, USA).After centrifugation at 4 • C, 3000g for 15 min, 12.5 μL of butylated hydroxytoluene was added to 500 μL of the supernatant.After vortexing, 250 μL of 30% trichloroacetic acid (TCA) was added to the mixture followed by an incubation on ice for 30 min before being centrifuged at 4 • C, 3000g for 10 min.In new tubes, 500 μL of the obtained supernatant was added and mixed with 37.5 μL of EDTA and 125 μL of 1% thiobarbituric acid before incubation in boiling water (100 • C) for 15 min.Samples were then cooled at room temperature for 10 min and centrifuged at 4 • C, 2000g for 10 min.Absorbance was measured at 530 nm using a spectrophotometer (Varioskan, Thermo Fisher Scientific, Waltham, MA, USA).The results were expressed in mg of MDA/kg of meat.
Protein oxidation was determined in duplicate by the 2,4-dinitrophenyl hydrazine (DNPH) method based on the Cayman Chemical Company test kit protocol (Item No. 10005020, Ann Arbor, MI, USA).Briefly, 1.5 g of meat was homogenized with 10 mL of phosphate buffer (50 mmol/L phosphate buffer pH 6.7 containing 1 mmol/L EDTA) using an Ultraturrax (IKA, Wilmington, NC, USA).Samples were then centrifuged at 4 • C, 4700g for 10 min.To remove nucleic acids, 100 μL of 10% streptomycin sulfate (S6501-50G, Sigma-Aldrich, Ontario, Canada) was added to 900 μL of supernatant and incubated at room temperature for 15 min prior to centrifugation at 4 • C, 6000g for 10 min.After collecting 200 μL of supernatant in a new tube and adding 800 μL of DNPH, the samples were incubated in the dark while vortexing them for 15 min.Once the incubation was completed, 1 mL of 20% TCA was added and the samples were centrifuged at 4 • C, 10 000g for 15 min.The supernatant was discarded and 1 mL of 10% TCA was added to the pellet before being centrifuged at 4 • C, 10 000g for 15 min, and then the supernatant was discarded.Excess DNPH was removed by three washes with 1 mL of ethanol/ethyl acetate (1:1) followed by vortex mixing and centrifugation at 4 • C, 10 000g for 15 min.After each wash, the supernatant was discarded.Following the last wash, 500 μL of guanidine hydrochloride (6 mol/L) was added, vortexed to suspend the precipitated protein pellet, and then centrifuged at 4 • C, 10 000g for 15 min.Absorbance was measured on the supernatant using a spectrophotometer (Varioskan, Thermo Fisher Scientific, Waltham, MA, USA) at 370 nm to determine carbonyl concentration.The concentration of DNPH was expressed in nmol of carbonyls/mg of protein.

Statistical analysis
To achieve the objectives of the present study, sampled breast fillets were grouped into different categories before statistical analysis.First, samples were grouped according to their status with regard to BMM into unaffected (n = 106), WSaffected (n = 54), WB-affected (n = 19), SM-affected (n = 8), and fillets affected by more than one myopathy (BMM+, n = 19).Fillets were also grouped into a BMM-affected (A, n = 100) or unaffected (U, n = 106) group.Next, they were grouped based of their pH using cut-off values defined by Zhang and Barbut (2005) into PSE-like (pH ≤ 5.7, n = 40), DFD (pH ≥ 6.1, n = 14), and pH-NORM fillets (n = 152).Finally, samples were grouped according to their brand into three categories; fillets marketed under an economical or store brand (ECON, n = 78), an established commercial brand (COM, n = 52), or as being of superior quality (SUPQ, n = 76).This last category included breast fillets from broilers reared in small-scale farms under alternative (free-range and/or organic) conditions; they are marketed as premium brand.
The association between the classification criteria was tested using the Fisher's exact test as implemented in R version 4.2 (R Core Team 2020).The effect of all classification criteria on quantitative quality (technological, microbial, and nutritional) traits and oxidation parameters was then analyzed using a linear mixed-effects model as implemented in the lmerTest R package (Kuznetsova et al. 2017).The model included the classification criteria as fixed effects, and the stores (n = 9) as a random effect to account for interstore random variability.Results were reported as least squares means and their standard errors (SEs).To understand the effect of fillets' physical attributes on the development of BMM, a mixedeffects binomial logistic regression model was fitted to the data using the package lme4 of R (Bates et al. 2015).This model included the BMM status (U = unaffected, A = affected) as a dependent variable and the fillet cranial thickness (mm), width (cm), and fillet length (cm) as independent variables.To account for random variability between stores, store locations were also included in the model as a random effect.The results of this analysis were reported as the odds ratio (OR) and their 95% confidence intervals (CI).For all statistical tests, the significance was determined at P < 0.05.

Breast muscle myopathies
In the present study, 48.5% of all samples collected (n = 100/206) were classified as BMM affected (Fig. 2).In our population of samples, WS accounted for 26.2% of all myopathic fillets and was the most frequently encountered BMM followed by WB and BMM+, which were equally represented among myopathic fillets (9.2% for each of these categories).Finally, SM was the least observed BMM in this work (3.8%).In terms of severity, apart from a few severe cases of WS (1.46%) and cases of fillets with multiple myopathies (9.2%), most BMM-affected fillets had only mild to moderate degrees of severity.

pH-related quality defects
According to the cut-off values defined by Zhang and Barbut (2005), 26.2% of all sampled fillets showed extreme variations in muscle pH.More specifically, 19.4% (n = 40) of our population of samples were classified as PSE-like meat (pH ≤ 5.7), while 6.8% (n = 14) of these samples were classified as DFD meat (pH ≥ 6.1).

Relationship between product brand and meat quality
All three categories of product brands included myopathic fillets.As shown in Fig. 3A, the intrabrand distribution of unaffected and BMM-affected fillets were in the same range of values (P > 0.05) and between-brand differ-ences were not statistically significant (P = 0.26 for the brand-by-myopathy Fisher's test).Conversely, the distribution of brands in relationship to pH-related quality issues differed significantly (Fig. 3B).ECON fillets were significantly more represented in the PSE-like category (60%) than COM (7.2%) and SUPQ fillets (32.5%;P < 0.0001).There were also more SUPQ fillets in this category than COM fillets (P = 0.0001).Furthermore, COM brand was the least affected by PSE-like meat.Finally, no significant differences were found between brands in the pH-NORM and DFD meat categories.

Physical characteristics of sampled breast fillets
In this study, we did not find statistically significant interaction between BMM and pH-related quality defects for breast fillets' physical characteristics.
The mixed-effects binomial logistic regression model showed that there were significantly higher chances for a fillet to switch from unaffected to BMM-affected with increased cranial thickness (OR = 1.175, 95% CI = 1.094-1.263,P < 0.0001) and with increased width (OR = 1.372, 95% CI = 1.113-1.691,P = 0.003), but not with increased fillet length (OR = 1.073, 95% CI = 0.906-1.271,P = 0.41).Figure 4 illustrates the predicted probability from the abovementioned regression model for a breast fillet to develop a BMM in function of its cranial thickness (Fig. 4A) and width (Fig. 4B) in the range of values observed in the current study for these two attributes.With regard to pHrelated quality defects, our study did not reveal any marked differences in physical characteristics between breast fillets classified as PSE-like, pH-NORM, or DFD meat (data not shown).

Product brand
Fillets sold as SUPQ were significantly thicker at the cranial end than COM fillets (P = 0.009), but not thicker than ECON fillets (P = 0.17; Table 2).ECON fillets were longer than   SUPQ fillets (P = 0.0002), and slightly longer than COM fillets (P = 0.06).Both ECON and SUPQ fillets had a similar weight and were significantly heavier than COM fillets (P < 0.001 and P = 0.003, respectively).No significant differences were found in fillet width between the three classes.

Technological quality traits of breast meat
The two-way interaction between BMM and pH-related quality defects was not significant for the measured traits of meat technological quality.

pH-related quality defects
Measured technological quality traits varied significantly between pH-related classes (Table 3).As expected, compared with DFD fillets, PSE-like fillets were paler (L * , P < 0.0001), yellower (b * , P = 0.0002), had a greater CL (P = 0.009), and a less tender texture after cooking (WBSF, P < 0.0001).Fillets with a pH in the range of normally observed values (pH-NORM) were mostly at the midpoint between PSE-like and DFD meat for the above-reported traits.

Impact of product brand on technological quality traits
The differences between classes of product brands with respect to the technological quality of sampled breast fillets are shown in Table 4. ECON breast fillets had a slightly but significantly lower pH (−0.09 units, P = 0.001) than COM fillets and tended to have a lower pH (−0.05 units, P = 0.059) than SUPQ fillets.
ECON fillets had a higher CL than COM fillets (+2.1%,P = 0.0003) and SUPQ fillets (+1.1%,P = 0.048), though COM and SUPQ did not differ significantly.No significant differences in meat texture were found between breast fillets from different brands.
SUPQ fillets were less yellow than COM (−2.3 units, P = 0.0002) and ECON (−1.17 units, P = 0.04) fillets.ECON and COM fillets did not differ in yellowness.In terms of lightness, ECON fillets tended to be paler than SUPQ fillets lightness (L * , P = 0.054).We did not find significant differences between product brands for redness (a * , P = 0.51) of breast fillets.

Breast muscle myopathies and pH-related quality defects
Contrary to the technological quality traits, the two-way interaction between BMM and pH-related quality defects was significant for the microbial quality of evaluated breast fillets (P = 0.005, P = 0.03, P = 0.03 for the counts of total aerobic mesophiles, Enterobacteriaceae, and Salmonella, respectively).
An interesting trend was observed in TAM counts (Fig. 5A): the fillets unaffected by BMM showed decreasing counts of TAM as pH increased (PSE-like > pH-NORM = DFD).The greatest log reduction (1.32 log unit) was observed between PSElike and DFD in BMM-unaffected fillets.BMM-affected fillets showed an opposite trend, with TAM counts increasing with pH (PSE-like = pH-NORM < DFD).Interestingly, the log difference in TAM counts between DFD fillets and the other two pH-related classes was greater than one log unit in both BMM-affected and unaffected fillets.Additionally, the highest log difference of TAM counts (1.80 log unit) was observed between DFD fillets from unaffected and BMM-affected fillets.
BMM-affected DFD fillets also showed significantly higher counts of Enterobacteriaceae compared with all other groups except for PSE-like fillets that were not affected by BMM (Fig. 5B).This last group (PSE-like BMM-unaffected) had significantly higher counts of Enterobacteriaceae (2.46 ± 0.19 log 10 CFU/20 cm 2 ) than fillets that showed normal pH and were not affected by BMM (pH-NORM BMM-unaffected; 2.03 ± 0.11 log 10 CFU/20 cm 2 ).All log differences between pH-related classes were ≤0.97 with counts ranging from 2.99 ± 0.31 for DFD-BMM affected fillets to 2.02 ± 0.39 log 10 CFU/20 cm 2 for DFD-BMM unaffected fillets.
As can be seen on Fig. 5C, Salmonella counts were significantly higher in DFD-BMM affected fillets than the other pH-related classes irrespective of their status with regard to BMM.The greatest log difference for Salmonella counts (1.04

Product brand
The results of microbial analyses with respect to product brand are shown in Table 5. Breast fillets marketed under higher quality brands (i.e., SUPQ) had significantly lower counts for the three analyzed taxa (aerobic mesophilic, Enterobacteriaceae and Salmonella counts) with log differences varying from 0.62 (Salmonella) to 2.61 (Enterobacteriaceae) log units.The highest microbial counts were observed for the ECON brand and were intermediate for COM with the log difference ≤0.41 log unit between the two classes.

Breast muscle myopathies and pH-related quality defects
The statistical analysis revealed a significant two-way interaction between BMM and pH-related quality defects for fillet protein (P = 0.041) and ash content (P = 0.02), but not for dry matter, moisture, or lipid content.In BMM-unaffected breast fillets, protein content decreased significantly with increased pH values from PSE-like to DFD (PSE-like > pH-NORM > DFD), while in BMM-affected breast fillets, protein content was comparable between the three classes of pHrelated quality defects (Fig. 6A).A similar trend was found for fillets' ash contents (Fig. 6B).Ash contents decreased with increased pH values (PSE-like > pH-NORM = DFD) in BMM-unaffected fillets, while no significant differences were observed between pH classes in BMM-affected fillets.
The results of the other traits of proximate analyses with respect to BMM are presented in Table 6.Overall, myopathic fillets showed a greater lipid content (1.63% vs. 1.27%,P < 0.001) than unaffected fillets.More specifically, unaffected breast fillets had a lower lipid content than fillets affected by WS (P < 0.0001), SM (P = 0.002), and BMM+ (P < 0.0001), but they were not statistically different to WB-affected fillets (P = 0.58).WB-affected fillets had a lower lipid content than fillets with SM (P = 0.02), WS (P = 0.01), and BMM+ (P = 0.004).On the contrary, no significant differences in lipid content were found between WS, SM, and BMM+ affected fillets in the present study.
Table 7 summarizes the results of the proximate analyses with respect to pH classes.As can be seen in this table, PSE-like fillets had lower water content than pH-NORM (P = 0.001) and DFD fillets (P = 0.04), while no significant differences were observed between pH-NORM and DFD fillets.Consequently, PSE-like fillets had higher dry matter content than pH-NORM (P = 0.001) and DFD fillets (P = 0.03).No significant differences in lipid content were found between pH classes.

Oxidation parameters
Lipid oxidation measured by the FOX and the TBARS methods, and protein oxidation measured by the DNPH method, did not vary significantly between myopathies, pH classes, or product brands (data not shown).Only COM fillets tended to have a higher concentration of secondary products of lipid peroxidation than ECON fillets (0.51 ± 0.04 mg of MDA/kg vs. 0.41 ± 0.04 mg of MDA/kg of meat; P = 0.06).

Discussion
The aim of this study was to investigate the presence of chicken breast meat quality issues in grocery stores in the province of Quebec, and to analyze their impact on the technological, microbial, and nutritive qualities of this meat.Data gathered will help to identify areas for future research efforts to reduce meat quality defects, to improve meat quality for customers in the province of Quebec, and to improve the economic benefits for processors.
Our results confirmed multiple quality issues in fresh broiler breast fillets sampled from all retail grocery stores that were visited.Indeed, one out of two sampled breast fillets were affected by one or multiple BMM with mostly mild to moderate degrees of severity.The low frequency of severe cases was probably due to the fact that breast fillets in this study were sampled from retail markets, and severe cases of BMM are usually diverted towards further processing or other uses.Interestingly, Fisher's exact test did not reveal any significant association between BMM and product brands (ECON, COM, and SUPQ).This result suggests that broiler chickens reared in alternative (i.e., nonintensive) systems and fed organic feed or grain-based diets in Quebec are equally susceptible to the development of BMM as conventionally raised broiler chickens.In the province of Quebec, strains used in free-range systems are mostly the same fast-growing strains used in the conventional production system.This is because the Canadian standards relative to these systems do not require the use of specific strains (for example, see CAN/CGSB-32.310-2020regulating the organic production Table 6.Effect of breast muscle myopathies on the proximate composition of broiler breast meat. Lipid, % 1.27 ± 0.07 c 1.64 ± 0.08 a,b 1.34 ± 0.12 c 1.80 ± 0.17  system).This could, at least partly, explain our data.This explanation is supported by previous research showing that BMM are mainly determined by genetics (Alnahhas et al. 2016;Lake et al. 2021) and that fast-growing strains of broiler chickens are more susceptible to the development of these myopathies than slower growing strains (Petracci et al. 2015;Santos et al. 2021).
Another type of meat quality issue that appeared in our 206 samples was associated with variations in muscle energy metabolism and specifically, with extreme variations in breast muscle pH, including PSE-like and DFD meat, which are also influenced by genetics and production systems (Petracci et al. 2009;Gigaud et al. 2011).These two wellknown quality defects accounted for 26% of sampled breast fillets with PSE-like being almost three times more frequent than DFD meat.Interestingly, BMM were similarly distributed between pH-related classes of meat quality.Usually, severe or extremely severe cases of BMM are associated with lower glycogen contents (Abasht et al. 2016) and higher ultimate pH (pHu) (Baldi et al. 2020) but, this is not necessarily the case for mild or moderate BMM (Campo et al. 2020), such as the ones observed in our samples.
Our findings also confirmed that the physical attributes of fillets, including cranial thickness and width, are good predictors of the risk of BMM.According to our analyses, an increase of 1 mm in fillet thickness or of 1 cm of fillet width was associated with a respective increase of 17% and 37% of the odds for the fillets to develop a myopathy.These results are in line with reports from the literature, which show that increased width and thickness in the cranial part of the pectoralis major muscle were associated with increased likelihood that this muscle would switch from normal to myopathic, and have higher BMM severity scores (Griffin et al. 2018;Oliveira et al. 2021;Che et al. 2022b).
Various meat quality defects affected the technological quality of breast fillets.Most sampled breast fillets were only mildly to moderately affected by BMM, which could explain the limited effect of these myopathies on the measured technological quality traits; this is in agreement with previous studies (Mudalal et al. 2015;Campo et al. 2020).
Conversely, variations in muscle pH significantly altered technological quality traits.Our results agree with the large body of literature on pH-related quality issues showing that rapid and/or extensive decline of postmortem muscle pH leads to increased denaturation of muscle proteins coupled with decreased water-holding capacity, and subsequently with decreased meat tenderness (Huff-Lonergan and Lonergan 2005; Zhang and Barbut 2005;Sheard et al. 2012).Furthermore, our findings showed that breast fillets marketed under an economical brand did show lower technological quality than fillets marketed under established commercial or superior quality brands.These results could be explained by the fact that ECON (and to a lesser extent, SUPQ) fillets represented a greater proportion of the PSE-like group than COM fillets.
In the current work, breast fillets' microbial quality, as measured by TAM, Enterobacteriaceae, and Salmonella counts, was markedly altered by the combined effect of BMM and muscle pH.Fillets unaffected by BMM showed similar counts of Salmonella no matter their respective pH-related quality classes.In BMM-affected fillets, higher values of breast muscle pH were associated with significantly higher Salmonella counts compared with muscles with lower pH.In the case of Enterobacteriaceae and total aerobic mesophiles, microbial counts decreased or tended to decrease with increased pH values in BMM-unaffected fillets, but in BMM-affected fillets, the opposite was observed, with microbial counts increasing with increased pH values.Overall, these findings suggest that the presence of BMM in breast fillets with high pH values could potentially exacerbate microbial spoilage, which is in agreement with the literature.In their study, Dalgaard et al. (2018) investigated the effect of moderate and severe cases of WB on the microbial growth on breast fillets stored for 6 or 8 days at 4 • C under 70% nitrogen and 30% CO 2 .These authors found an increase of total cell counts and Enterobacteriaceae counts from unaffected, to moderately to severely affected fillets at the end of storage.Interestingly, these authors reported significantly higher pH values in WB-affected fillets than unaffected fillets (6.02 vs. 5.82, P < 0.05), which could have contributed to the increased counts of this taxa in their study.
Since only 2.9% out of the 206 fillets tested were classified as extreme BBM (severity scores of 3 for WS and of 2 for SM), this could have limited the extent of the observed effect of BMM on microbial counts in the present study as mild and moderate BMM were not associated with increased pHu.Breast muscle pHu is entirely determined by glycogen reserves present in the pectoralis major muscle at slaughter (Le Bihan-Duval et al. 2008) and was shown to explain up to 87% of the variation in aerobic plate counts in pork loin (Holmer et al. 2009).When these reserves are low, meat pHu remains high (>6.0),which allows microbial growth that would be otherwise inhibited under normal values of pHu (Newton and Gill 1981).Gratta et al. (2019) reported that normal fillets with no myopathy had a lower microbial shelf life than affected fillets.However, in their study, both normal and myopathic fillets had pH values ≥6.Allen et al. (1997) studied the relationship between the pH and shelf life in broiler breast fillets categorized as dark (pH of 6.08-6.22)or light (pH of 5.76-5.86)and reported a faster bacterial spoilage in dark fillets than in light-coloured fillets; this is also in line with our results for BMM-affected DFD fillets.
With respect to product brand, SUPQ breast fillets showed significantly lower counts for all three taxa counted, than COM or ECON fillets did; this indicated that SUPQ fillets (free-range and/or organically raised) were of better microbial quality.However, it is important to note that relatively to production conditions, hygiene level during processing and packaging is a major determinant of the microbial quality of poultry meat (Marmion et al. 2021).
In this study, data were not available about packaging type or atmospheric conditions, or about the number of days between slaughter and sampling.Differences in these parameters could have partly contributed to the variation in the reported microbial counts (Chmiel et al. 2018;Nauman et al. 2022).Furthermore, selection criteria adopted by different processing plants could be more stringent for premium products leading to improved microbial quality relative to standard commercial products.In addition, there were fewer SUPQ fillets in the DFD meat class of muscle pH, which could have also partly contributed to their lower microbial counts.Differences in packaging and atmospheric conditions could have also contributed to TAM counts unexpectedly decreasing instead of increasing with increased muscle pH in BMM-unaffected fillets.
Finally, it is important to note that sampled fillets were all safe for consumption from a microbiological standpoint according to the local regulations (MAPAQ 2019).
With regard to breast fillets' chemical composition, we found a significant interaction between BMM and muscle pH for protein content.In unaffected fillets, protein content decreased significantly with increased pH values, while in myopathic fillets, no significant differences were found between pH classes.In addition, BMM-unaffected PSE-like fillets had higher protein content than BMM-affected fillets despite their pH value.This finding is in agreement with previous studies showing decreased protein content and increased lipid content in myopathic fillets compared with unaffected fillets.For instance, Soglia et al. (2016) investigated the effect of WB and WB+WS on the proximate composition of breast fillets and showed that fillets' protein content decreased significantly and progressively from normal fillets to fillets affected by a combination of WB and WS (NORM > WB > WB+WS).In a more recent study, Baldi et al. (2018) showed that protein content in both superficial and deep layers of breast muscles was significantly higher in unaffected fillets compared with fillets affected by WS, SM, and a combination of WS and SM (NORM > WS > WS+SM).For the effect of muscle pH, the lower water content and higher dry matter content that we found in PSE-like fillet could partly explain their increased concentration of protein and ash compared with other pH classes.The lack of significant differences in weight between breast fillets of different pH categories (data not shown) suggests that the higher water loss in PSE-like fillets was mainly a result of their lower pH and not a difference in their weight.
The higher intramuscular lipid content reported here in myopathic fillets compared with unaffected fillets is also in line with reports from the literature (Soglia et al. 2016;Baldi et al. 2018;Gratta et al. 2019).The observed increase in lipid content and decrease in protein content in myopathic fillets are likely to be induced by the overwhelming degeneration of muscle tissue, and the subsequent replacement of this tissue by fat and connective tissue due to the lipidosis and fibrosis observed in myopathic fillets (Kuttappan et al. 2012a;Sihvo et al. 2014).
In their severe forms, BMM are associated with oxidative stress and increased oxidative damage in breast muscle tissue (Abasht et al. 2016;Malila et al. 2019).In our samples, despite the higher lipid content in myopathic fillets compared with unaffected fillets, concentrations of primary and secondary lipid peroxidation products remained at similar levels in both myopathic and unaffected fillets.This result could be explained, in part, by the fact that only a few severe myopathy cases were found in our samples.This explanation is supported by the literature where, for instance, Li et al. (2022) analyzed MDA concentrations in samples taken from the superficial and deep parts of breast fillets affected by increased degrees of WB.These authors showed that over mild to moderate degrees of WB, MDA concentration remained at similar levels in the samples taken from the deep part of the muscles, just like those we sampled.Similarly, these authors did not find any significant differences in the concentration of carbonyl groups (i.e., protein oxidation) between normal, mildly, and moderately affected fillets when samples were taken from the deep part of the pectoral muscles.Overall, these findings suggest that profound changes in muscle structure and metabolism, such as those observed in muscles with severe BMM, could be the trigger to expose lipids and proteins to oxidative processes and to induce the oxidative damage associated with these myopathies.
In this work, we analyzed breast fillets sampled from grocery stores during the period between January and April.The limitations of this study included the lack of information on storage time and conditions after slaughter, on packaging and its atmospheric conditions.We tried to account for variability induced by these factors by fitting mixed-effects statistical models to the data that captured some of the between-store variability.The remaining uncaptured random variability could have contributed to some of the unexpected findings such as the lack of differences between pH classes in terms of lipid and protein oxidation and the unexpected trend of some microbial counts decreasing or remaining stable with increasing pH in BMM-unaffected fillets.Further research is needed to elucidate the mechanisms underlying the effect of the interaction between BMM and muscle pH on meat quality.

Conclusions
To the best of our knowledge, this study is the first to investigate the presence of BMM in broiler chickens for sale in the province of Quebec.Our results showed that multiple quality issues related to muscle structure and metabolism were present in breast fillets sold in Quebec's grocery stores.In our samples, almost half of breast fillets were affected by BMM, and more than one-quarter of all fillets showed a pH-related quality defect.These quality defects had significant negative impact on the technological, microbial, and nutritional qualities of breast fillets.
Our study reiterated the importance of breast muscles' physical attributes, including cranial thickness and width, as risk factors for the development of BMM.It also provided original data on the relationship between breast myopathies and pH with respect to meat microbial quality.Additionally, this study highlighted the observation that breast fillets from the conventional, free-range, and organic production systems in Quebec were equally susceptible to the development of these myopathies, which is probably due to the use of the same fast-growing strains in all these production systems.
Overall, among fillets produced in the conventional system, those marketed as economical (average consumer price of $15.21/kg over the sampling period) had a lower technological, microbial, and nutritional value than fillets marketed under an established commercial brand ($18.63/kgaverage over the same period).These two categories had lower technological, microbial, and nutritional value compared with free-range and/or organically produced chicken fillets ($21.42/kg over the same period).The difference in the average price per kg of breast meat between the ECON and the COM categories, which are both produced in the conventional system, clearly demonstrates the economic loss that the industry can incur due to reduced meat quality, and the benefits of monitoring and controlling factors that affect meat quality.

Fig. 2 .
Fig. 2. Distribution of sampled breast fillets (n = 206) according to different degrees of severity of breast muscle myopathies.Norm = normal breast fillet; WS1 = mild, WS2 = moderate, and WS3 = severe degree white striping; WB1 = mild and WB2 = moderate degree of wooden breast; SM1 = moderate and SM2 = severe degree of spaghetti meat; BMM+ = breast fillet with multiple breast myopathies.WS and WB were evaluated according to Kuttappan et al.'s (2016) scale.SM was evaluated according to Baldi et al.'s (2018) scale.

Fig. 3 .
Fig. 3. Distribution of different product brands between different categories of breast muscle myopathies (A) and of pH-related defects (B).ECON = economic or store brand; COM = commercially established brand; SUPQ = brand of superior quality like organic or grain-fed broilers; Normal = normal breast fillets with regard to breast muscle myopathies (BMM); BMM-affected = breast fillets with one or more BMM; PSE-like = pale, soft, exudate-like meat; NORM = normal pH breast meat; DFD = dark, firm, dry meat.
Breast muscle myopathies: NORM = unaffected fillets, WS = white striping, WB = wooden breast, SM = spaghetti meat, BMM+ = fillets with more than one myopathy.BMM£ = P value of the effect of breast muscle myopathies categorized as NORM, WS, WB, SM, and BMM+ (categorical trait), BMM2 = P value of the effect of myopathies categorized as unaffected or BMM-affected (binary trait).Means with different letters in a row(a, b, c, d)  are significantly different at P < 0.05.

Fig. 4 .
Fig. 4. Predicted probability (and its 95% confidence interval) for a breast fillet to develop a breast muscle myopathy in function of the cranial thickness (A) or of the width of the fillet at the cranial end (B).Predictions were computed in the range of data observed in the current study.BMM = breast muscle myopathy.
11 0.20 § PM-pH = pH, L * = lightness, a * = redness, b * = yellowness of the pectoralis major muscle, CL = cooking loss, WBSF = Warner-Bratzler shear force of cooked breast meat (N/cm 2 ).Note: Classes of production brand: ECON = economic or store brand, COM = commercially established brand, SUPQ = brand of superior quality including organic or grain-fed broilers.Means with different letters in a row (a, b) are significantly different at P < 0.05.

Fig. 5 .
Fig. 5. Effect of the interaction between breast muscle myopathies and breast muscle pH on microbial counts of total aerobic mesophiles (A), Enterobacteriaceae (B), and Salmonella (C) in sampled breast fillets.Data are expressed as least squares mean ± SE.
a 1.39 ± 0.16 b < 0.0001 § Counts are expressed as log 10 colony-forming unit (CFU)/20 cm 2 .Means with different letters in a row (a, b, c) are significantly different at P < 0.05.Note: Classes of production brand: ECON = economic or store brand, COM = commercially established brand, SUPQ = brand of superior quality including organic or grain-fed broilers.

Fig. 6 .
Fig. 6.Effect of the interaction between breast muscle myopathies and breast muscle pH on protein (A) and ash (B) content of breast muscles.Data are expressed as least squares mean ± SE.
Breast muscle myopathies: NORM = unaffected fillets, WS = white striping, WB = wooden breast, SM = spaghetti meat, BMM+ = fillets with more than one myopathy.BMM£ = P value of the effect of breast muscle myopathies categorized as NORM, WS, WB, SM, and BMM+ (categorical traits), BMM2 = P value of the effect of myopathies categorized as unaffected or affected (binary traits).Means with different letters in a row (a, b, c) are significantly different at P < 0.05.

Table 1 .
Physical characteristics of normal and myopathic broiler breast fillets.

Table 2 .
Physical characteristics of broiler breast fillets from different product brands.

Table 3 .
Effect of pH-related quality issues on technological quality traits of broiler breast meat.

Table 4 .
Effect of product brand on the technological quality traits of broiler breast meat.

Table 5 .
Effect of product brand on microbial counts of broiler breast fillets.

Table 7 .
Effect of pH-related quality defects on the proximate composition of broiler breast meat.Note: pH classes: PSE-like = pale, soft, exudate-like meat (pH ≤ 5.7), NORM = breast meat of normal pH, DFD = dark, firm, dry meat (pH ≥ 6.1).Means with different letters in a row (a, b) are significantly different at P < 0.05.