All experimental conditions were in accordance with the Korean Animal Protection Act, including detailed provision for transport and slaughter. Treatments, handling conditions, and experimental procedures were controlled and/or approved by the Ministry of Agriculture, Food, and Rural Affairs of South Korea. Slaughter procedure was under the supervision of the Korea Institute for Animal Products Quality Evaluation.

A total of 111 Yorkshire (pure breed castrated male) pigs were evaluated. All pigs were non-carriers of the halothane gene and clinically healthy. The treatment conditions, both before and after slaughter, were same for all pigs. The pigs were divided randomly into different pens in a commercial farm, with a stocking density of 1.0 m² per pig. The temperature was maintained at 20±2°C, and artificial light was provided from 0900 to 2100. Water and food were available ad libitum at nipple drinkers and food dispensers. In addition, all pigs were fed the same commercial diet in accordance with the National Research Council (1998). When the pigs reached a live weight of 115±5 kg, they were transported to a commercial slaughter facility. Approximately 4 h before transport, feed was withdrawn. The loading and unloading took place as parallel to the ground as possible. Electric prods were not used during loading or unloading. The truck had two straight decks, and a hydraulic lift raised the floor of the deck to create the second deck. The truck was equipped with natural ventilation, and the stocking density was approximately 0.47 m² per pig. The pigs were transported for 1 h to the slaughter facility. The pigs were not mixed with unfamiliar pigs during transport or lairage. To reduce stress during lairage, the pigs were showered with water and provided with drinking water ad libitum. After being rested for approximately 12 h at the slaughter facility, the pigs were slaughtered. All the pigs were slaughtered during the winter period in one of 4 batches (30, 30, 30, and 21 pigs per batch). The pigs were immobilized through electrical stunning (2 to 4 s with over 1.25 A), shackled by the right leg, and exsanguinated while hanging. At exsanguination, blood samples were collected from each pig. A dehairing process was performed in a scalding-singeing combination by scalding for 3 min at 60°C, and then singeing via passing through gas burner for 10 s at 1,200°C, and any remaining hair was removed using a knife and flame. The carcasses were eviscerated, split into the left and right sides, and then cooled in the chilling room (4°C).

At 45 min postmortem, muscle samples for the analysis of glycogen and lactate contents were taken from the longissimus lumborum muscle at the 9th and 10th thoracic vertebrae. After 24 h of chilling, muscle samples for the measurement of muscle glycogen and lactate contents and pork quality traits were taken from the longissimus lumborum muscle at the 10th and 15th thoracic vertebrae. The collection of all muscle samples and pH measurements was made on the right side of each carcass.

Blood samples were collected from each pig at exsanguination after electrical stunning. The blood samples were collected using tubes treated with potassium oxalate/sodium fluoride (BD Vacutainer fluoride tube; Becton Dickinson, Franklin Lakes, New Jersey, USA) to inhibit further glycolysis. Blood glucose and lactate levels were then measured using hand-held devices (blood glucose: OneTouch Ultra, LifeScan, Inc., Milpitas, California, USA; blood lactate: Lactate Scout, EKF Diagnostics, Barleben, Magdeburg, Germany). All measurements were completed within 10 min of exsanguination at the slaughter facility. Blood glucose level was expressed in milligrams per deciliter (mg/dL), and blood lactate level was expressed in millimoles per liter (mmol/L).

Glycogen content of muscle was measured via the method described by Dreiling et al. (1987). Approximately 1.5 g of tissue was minced, suspended in 10 mL of 9% cold perchloric acid (PCA), and thoroughly homogenized for 30 to 45 s with a mechanical tissue disrupter. After centrifugation (15,000×g at 4°C), the supernatant was decanted and saved for glycogen determination. An iodine color reagent was prepared by combining 100 mL of saturated CaCl₂ with 1.3 mL of the solution (0.26 g of iodine, 2.6 g of potassium iodide in 10 mL of distilled water) (prepared freshly every day). The 2.6 mL of the iodine color reagent was added to 0.4 mL of the glycogen standards or tissue extracts. In this way, glycogen standard curves were developed for each set of samples so that linear regression equations could be used to determine the glycogen concentrations of the test samples.

Lactate content was determined spectrophotometrically at 340 nm using a commercial kit (Boehringer-Mannheim, Darmstadt, Germany). Approximately 500 mg of muscle was homogenized for 30 s in 2 mL of 1 M PCA. Potassium hydroxide (2 M) was added to neutralize the solution, and distilled water was added to make a final volume of 10 mL. After 20 min of refrigeration and centrifugation, lactic acid concentration was measured.

Muscle pH at 45 min postmortem (pH_{45 min}) and 24 h postmortem (pH_{24 h}) was measured in the chilling room directly at the 7th and 8th thoracic vertebrae of each carcass (right side) using a portable pH meter for meat (HI 99163, Hanna instruments, Woonsocket, Rhode Island, USA).

For analysis of meat color, WHC, and Warner-Bratzler shear force (WBS), samples were obtained from the pork loin at the 10th to 15th thoracic vertebrae from the right side of each cold carcass. Meat color was measured with a Minolta Chroma Meter (CR-400, Minolta Camera Co., Tokyo, Japan), following a slightly modified version of the method of Honikel (1998). Samples were cut from the pork loin at 24 h postmortem and placed on a table for 30 min, exposing their surfaces to air without any packaging (for bloom) before color measurement. Triplicate measurements were averaged, and the results were expressed as Commission Internationale de l’Eclairage lightness (L^*), redness (a^*), and yellowness (b^*).

To evaluate the WHC, drip loss, cooking loss (Honikel, 1998), and filter-paper fluid uptake (FFU) (Kauffman et al., 1986) were measured. To determine drip loss, meat samples were cut from the pork loin at 24 h postmortem and weighed immediately to obtain their initial weight of drip loss. The samples were then placed in netting and suspended in an inflated bag, ensuring that the sample was not in contact with the bag. After a 48 h storage period at 4°C, the samples were taken from the bag, gently blotted dry, and reweighed. Drip loss was expressed as a percentage of the sample’s initial weight. To measure cooking loss of water, different samples were freshly cut from the pork loin at 24 h postmortem and weighed to obtain their initial weight. The samples were then put in thin-walled polyethylene bags and placed in a continuously boiling water bath. The samples were cooked to 75°C internal temperature and, once this end-point temperature had been reached, the bags were removed from the water bath. Thereafter, the samples were cooled in an ice slurry and kept under chilled conditions (1°C to 5°C) until equilibration. Then, the samples were taken from the bag, blotted dry, and reweighed. Cooking loss was expressed as a percentage of the sample’s initial weight. For FFU, further samples were cut from the pork loin at 24 h postmortem. Filter paper (Whatman #2, 42.5 mm in diameter, GE Healthcare, Little Chalfont, Buckinghamshire, UK) was pre-weighed, placed on the surface of the sample for just under 2 s to absorb fluids, and weighed. The FFU was expressed as milligrams of exudate absorbed into the filter paper.

To measure WBS, the pork loin was cut into 20-mm-thick chops. Pork chops from each sample were cooked to a final core temperature of 75°C in a continuously boiling water bath. After cooking, six cores (1.27 cm in diameter), parallel to the longitudinal orientation of the muscle fibers, were taken from each pork chop for WBS measurement. WBS was determined using an Instron Universal Testing Machine (Model Series IX, Instron Co., Norwood, MA, USA) with a Warner-Bratzler shearing device. The samples were sheared perpendicular to the long axis of the core, and the WBS value was calculated taken from the peak of the force curve (Honikel, 1998).

All statistical analysis was performed using the Statistical Analysis System (2009). Pearson correlation coefficients between blood and muscle metabolites or between blood and muscle metabolites and pork quality traits were determined using the CORR procedure. To establish multiple regression models for pork quality traits, blood or muscle metabolites were used as independent variables in the REG procedure. Furthermore, the stepwise procedure was used to examine the percentage of variation in certain pork quality traits explained by blood and muscle metabolites. The qualitative variable (batch) was coded as dummy variables to eliminate their effect on the regression model.

Table 1 shows the correlation coefficients between the following variables: exsanguination blood glucose and lactate levels and postmortem muscle glycogen and lactate contents. Levels of blood glucose and lactate showed a positive relationship. Blood glucose level was negatively related to muscle glycogen content at postmortem (i.e. 45 min and 24 h postmortem) but positively correlated with muscle lactate content at postmortem. Blood lactate level was also negatively related to muscle glycogen content at 45 min postmortem and positively correlated with muscle lactate content at postmortem. These relationships mean that high levels of blood glucose and lactate are associated with not only rapid glycolytic rate at early postmortem (glycogen and lactate contents measured at 45 min postmortem) but also extended postmortem glycolysis (glycogen and lactate contents measured at 24 h postmortem). There are negative correlations between postmortem muscle glycogen and lactate contents.

There are significant relationships between metabolites in exsanguination blood and postmortem muscle and pH, lightness, and WHC (Table 2). Blood glucose level is negatively related to muscle pH_{45 min} and pH_{24 h} and positively correlated to lightness, FFU, and drip loss. Blood lactate level is also negatively related to postmortem muscle pH and positively correlated with FFU and drip loss. Higher levels of exsanguination blood metabolites were associated with paler color and decreased WHC. Muscle glycogen content at 45 min postmortem had a positive relationship with muscle pH_{45 min} and negative relationship with lightness, FFU, and drip loss.

Muscle glycogen content at 24 h postmortem exhibited similar relationships, except that it was significantly correlated with muscle pH_{24 h}, rather than muscle pH_{45 min}. On the other hand, muscle lactate content at both 45 min and 24 h postmortem were negatively related to postmortem muscle pH and positively correlated with lightness, FFU, and drip loss. Lower glycogen content and higher lactate content in postmortem muscle were correlated with deteriorative pork quality.

Multiple regression analysis including stepwise regression was performed to investigate the influence of exsanguination blood metabolites (Table 3), postmortem muscle metabolites (Table 4), or both the blood and muscle metabolites together (Table 5) on pork quality traits. Levels of blood glucose and lactate at exsanguination explained 21.8% and 24% of the variation in muscle pH_{45 min} and pH_24h, respectively. They also accounted for 21% and 16.2% of the variation in FFU and drip loss, respectively. On the other hand, compared to blood metabolite levels, contents of muscle glycogen and lactate at postmortem explained a higher proportion of the variation in pork quality traits. Over 40% of the variability in postmortem muscle pH was explained by muscle metabolite content. Metabolites contents in postmortem muscle accounted for over 40% of the variation in lightness and drip loss. The variability of FFU was also explained up to 31.2% by muscle metabolite content.

When a stepwise regression analysis using metabolites from both exsanguination blood and postmortem muscle was carried out, the variability of pork quality traits was explained slightly further. Blood metabolites and muscle metabolites at 45 min postmortem accounted for 54.4% of variation in muscle pH_{45 min}. In the case of muscle pH_{24 h}, 46.2% of the variation was explained by blood and muscle metabolites. In fact, muscle lactate content at 24 h postmortem alone accounted for 30.1% of the variation. Variability of lightness, FFU, and drip loss showed similar tendencies, with muscle lactate content at 24 h postmortem alone explaining a considerable portion of the variation once again.