Paired strip loins (M. longissimus lumborum, Institutional Meat Purchase Specifications [IMPS] #180), top sirloin butts (M. gluteus medius, IMPS #184B), and eye of round (M. semitendinosus, IMPS #171C) were collected from 15 (n = 15) A-Maturity beef carcasses with no defects (USDA Low Choice, Small^00–100 marbling scores) from a commercial facility at 3 d postmortem, vacuum sealed and transported to the Red Meat Laboratory of the University of Arkansas for processing.

Upon arrival, the muscle pairs from each carcass were aged at 2°C until 28 d postmortem before fabrication. Figure 1 shows a schematic of the sample allocation and processing used in the current study. Each side of the paired muscles was assigned to either steak freezing (STK) or section freezing (SEC). Each muscle was then equally portioned into 3 sections and was randomly assigned to 3 different subsequent secondary consumer freezing methods: 1) blast freezer (BLS) (Master-Bilt, New Albany, MS, USA; −20°C), 2) chest freezer (CST) (Whirlpool Corporation, Benton Harbor, MI, USA; −20°C), and 3) regular refrigerator freezer (FRI) (Midea Group, China; −18°C). For muscles assigned to the STK freezing methods, each section was cut into at least 5 steaks (2.54 cm), whereas muscles assigned to SEC remained as sections. All samples were then weighed and individually vacuum packaged. The samples were randomly distributed based on their freezing size assignment (SEC versus STK) to closed cardboard boxes following typical industry practice. Each box contained an equal number of samples and similar weights (∼19 kg/box). The temperature changes during freezing and thawing were monitored using Thermo Button data loggers (Model 21G, ProgesPlus - PlugAndTrack, France) and a T-type thermocouple (Omega Engineering, Stamford, CT, USA) connected to an OctTemp 2000 data logger (Madge Tech, Inc., Warner, NH, USA), which was embedded in additional samples specifically designated for temperature monitoring.

All samples were subject to the first freezing cycle in a −20°C BLS and stored for 35 d. At the end of the first freezing storage period, samples were thawed at 3°C until an internal temperature of 2°C was reached (approximately 5 d for STK and 7 d for SEC). Following the completion of the thawing process, samples were separated and subjected to a second freezing cycle based on the previously assigned freezer types. The samples assigned to BLS freezing were frozen in closed boxes as previously mentioned. CST and FRI samples were frozen in the freezers without any boxes to simulate consumer freezing practices. Samples were then stored frozen in the assigned freezers for 35 d and rethawed using the previous method after storage duration. After the second thawing process, all samples were blotted dry and reweighed for purge loss, and SEC samples were cut into at least 5 steaks (2.54 cm). The samples were then collected for water-holding capacity (WHC), color stability, instrumental tenderness, consumer sensory evaluation, and biochemical analyses. Samples not immediately used for analyses were individually vacuum-packed and stored in a −80°C freezer until further analysis.

The pH of each sample was measured using an insert-type meat pH probe (Testo 205, Testo North America, West Chester, PA, USA) in duplicate before and after all the freezing treatments. The probe was calibrated to pH 4 and 7 standards following the manufacturer’s instructions before any measurements.

Several assays were conducted to measure the meat WHC at the end of the repeated freeze-thaw cycles. The percentage purge loss from the freeze-thaw process was determined by measuring the weight before and after freezing treatments. Drip loss was determined using the Honikel drip loss method (Honikel, 1998) with modification (Setyabrata and Kim, 2019) and expressed by calculating the difference in the initial and final mass of the sample in percentage. The cook loss was determined by weighing and cooking the sample with a clamshell griddle (Griddler GR-150, Cuisinart, Glendale, AZ, USA) until the internal steak temperature reached 71°C, and samples were reweighed after cooking. Cooking loss was determined by calculating the difference between the sample weights before and after cooking.

The surface color stability of the steak samples was assessed daily using both instrumental and trained panelists during a simulated color display. Samples were placed on foam trays with soaker pads and overwrapped with oxygen-permeable Polyvinyl chloride (PVC; 15,500–16,275 cm³ O²/m² /24 h at 23°C, E-Z Wrap Crystal Clear Polyvinyl Chloride Wrapping Film; Koch Supplies; Kansas City, MO, USA). Samples were displayed for 5 d with daily rotations in an open-front display case (5 shelves; each shelf has 2 LED light bars [P105998A, Lux 1628 – 2040, 3500 K, 48”, 4.80 Watt., Hillphoenix Inc., Conyers, GA, USA]). The surface color of the muscles was measured daily using a Hunter MiniScan EZ spectrophotometer (Hunter, Reston, VA, USA) set with a 25-mm aperture, illuminant A, and 10° standard observer angle. CIE L*, a*, and b* values were recorded from 3 random locations on the surface of the sampled beef muscles. Obtained CIE L*a*b* values were used to calculate the saturation index [(a*2 + b*2)^1/2] and hue angle [(b*/a*) tan⁻¹] (King et al., 2023). Total color change (ΔE) between day 1 and day 5 was also calculated using [(ΔL*)² + (Δa*)² + (Δb*)²]^0.5 to observe color differences during the display (King et al., 2023).

The visual color assessment of the beef muscles was carried out by a group of 12 trained color panelists who had previously passed the Farnsworth-Munsell 100 Hue Test. The sensory panelists were trained in alignment with the American Meat Science Association Meat Color Guidelines (King et al., 2023). Lean color was scored on an 8-point scale [1 = extremely dark red, 2 = dark red, 3 = moderately dark red, 4 = slightly dark red, 5 = slightly bright red, 6 = moderately bright red, 7 = bright red, 8 = extremely bright red]. Conversely, discoloration was scored on a 7-point scale [1 = no discoloration (0%), 2 = slight discoloration (1–19%), 3 = small discoloration (20–39%), 4 = modest discoloration (40–59%), 5 = moderate discoloration (60–79%), 6 = extensive discoloration (80–99%), 7 = total discoloration (100%)].

The lipid oxidation was measured using the 2-thiobarbituric acid reactive substances (TBARS) assay, following the method described by Setyabrata and Kim (2019) on samples before and after the simulated retail display. The TBARS samples were individually powdered after being snap-frozen in liquid nitrogen. Five grams (5 g) of the samples was homogenized with 15 mL deionized water (DW) and 50 μL of butylated hydroxyanisole (BHA) (10% butylated hydroxyl anisole solution in 90% ethanol) for 15 seconds at 8,000 rpm (T-25 Ultra Turrax, IKA Works Inc., Wilmington, NC, USA). One mL of the homogenate was combined with 2 mL of 20 mM 2-thiobarbituric acid solution in 15% trichloroacetic acid (TBA/TCA; w/v) solution. The mixture was heated in a water bath (80°C) for 15 min and cooled in ice water for 10 min. After cooling, the mixture was centrifuged at 2,000 g for 10 min, and the supernatant was filtered with filter paper (Whatman No. 4, Cytivia, Marlborough, MA, USA). The supernatant was aliquoted into 96-well plates (200 μL), and the absorbance was read at 531 nm using the microplate spectrophotometer against blank (2 mL of DW and 4 mL of TBA/TCA solution). TBARS value was expressed as mg malondialdehyde (MDA) per kilogram of meat by multiplying the absorbance at 531 nm by the MDA molecular extinction coefficient of 5.54.

Warner-Bratzler shear force (WBSF) values were measured on samples assigned for cook loss measurement using the TA-XT Plus Texture Analyzer (Stable Micro System Ltd., Godalming, UK) coupled with a V-shaped blade attachment. The cooked steaks were stored at 2°C overnight before the shear force measurement. A total of 6 cores (1.27-cm diameter each), parallel to the direction of the muscle fibers, were collected per steak using a manual coring device while avoiding any obvious connective tissues and fat. The cores were sheared perpendicular to the muscle fiber orientation. The crosshead speed was set to 3 mm/second, and a 5-kg load cell was utilized during the measurement. The average peak shear force (N) of the cores from each sample was calculated.

Sensory evaluation of the beef samples by consumers (N = 120 panelists) was conducted at Texas Tech University (institutional review board #2017-598) following the method by Hernandez et al. (2023). Meat samples were cooked with a clamshell griddle (Griddler GR-150, Cuisinart, Glendale, AZ, USA) until the internal steak temperature reached 71°C. After cooking, the steak samples were cut into 1-cm³ cubes and placed in an individual sample cup labeled with 3-digit codes in a balanced and randomized design. The samples were then held at 55°C in a food service warmer (Cambro Manufacturing, Huntington Beach, CA, USA) and served within 10 min.

Consumer panelists were provided with an electronic ballot (Qualtrics, Provo, UT, USA), a plastic fork, a toothpick, a napkin, an expectorant cup, a cup of water, and palate cleansers (unsalted crackers and diluted apple juice) to use between samples. Each ballot contained an information sheet, demographic questionnaire, beef steak purchasing behavior sheet, and sample ballots. The samples were evaluated for flavor, tenderness, juiciness, and overall liking on unstructured 100-point line scales. Scales were verbally anchored at each endpoint and midpoint (0 = extremely dislike/extremely tough/extremely dry; 50 = neither dislike nor like/neither tough nor tender/neither dry nor juicy; 100 = extremely like/extremely tender/extremely juicy). Following scoring, each panelist was asked to rate each sample as either acceptable or unacceptable for each palatability trait.

The study was a complete block design with a split-split plot arrangement for treatment allocation. Each carcass served as a block. Data from each muscle were analyzed separately. The muscle size effect (SEC versus STK) was analyzed as the whole plot, and the freezer type was considered as the subplot. The data were analyzed using PROC GLIMMIX procedure of SAS 9.4 (SAS Institute, Cary, NC, USA) with size, freezer types, and their interaction as fixed effects. The carcass was considered a random effect. Day was added as a fixed variable to analyze the lipid oxidation data. Day was a repeated measure for instrumental and color analysis. The panelist, session, and final peak temperature were added as covariates for consumer sensory evaluation analysis. The significance was considered at P < 0.05, and a tendency was declared at 0.05 ≤ P < 0.10. Significant least square means were separated using the F-test.

The impact of different freezing sizes and freezer-type treatments on the temperature decline rate of the different muscles is presented in Figure 2. As expected, STK samples had a faster freezing rate compared with SEC during the first freezing process in a −20°C BLS (Figure 2A). Kiani and Sun (2011) reported that −1°C to −7°C is a critical freezing temperature range in meat, dictating the characteristics of ice crystals formed within the product. The present study showed that STK samples took about 20 h to pass through the critical freezing temperature zone. In comparison, the SEC samples took about 30 h, which could have resulted in larger ice crystal formation within the SEC samples compared with STK samples (Lee et al., 2022).

During the subsequent second freezing using different consumer freezer types, BLS had a faster and more uniform freezing rate, attaining a lower temperature of −20°C within 40 h of freezing, compared with other freezer types, regardless of muscles and size of the products (Figure 2B). Generally, the STK samples were found to have a faster freezing rate compared with SEC when considering the impact of size (Supplementary Figure 1). However, the rate of temperature decline was not consistent among the freezer-type treatments. Although further confirmation is still required, it could be speculated that the rate of temperature decline would be impacted by product location within the freezer. Depending on the location in the freezer, airflow surrounding the samples might be limited, which then reduces the rate of heat transfer and explains the differences observed in the current study.

Tables 1, 2, and 3 show the pH and WHC results for strip loins (LL), top sirloins (GM), and eye of rounds (ST), respectively. The pH of the product was impacted by the different treatments applied. A significant size effect was observed on the pH of the LL and GM muscles, reporting a greater value in STK samples compared with the SEC samples (P < 0.05) for both muscles, regardless of the freezer type. The pH of the ST samples was impacted by both size and freezer main effects. Unlike LL and GM, the STK samples from ST had decreased pH compared with ST SEC samples (P < 0.05). Additionally, a significant freezer effect was observed in ST samples. The FRI and BLS samples had the greatest and least pH values, respectively (P < 0.05), while CST samples were not different from the other treatments in ST samples (P > 0.05). Although significantly different, the pH differences were minuscule among the treatments and potentially not practically meaningful, and all the pH values remained within what would be considered the normal ultimate pH range (Page et al., 2001).

The WHC of the different muscles was analyzed using purge loss, drip loss, and cook loss. A size × freezer interaction effect was observed on the purge loss of LL, in which SEC-BLS from LL had a greater purge loss compared with STK-BLS and STK-CST (P < 0.05). Additionally, SEC-BLS had a similar purge loss with SEC-CST, SEC-FRI, and STK-FRI in LL(P > 0.05). Drip loss was influenced by size in the LL muscle, exhibiting greater drip loss in STK samples compared with SEC samples (P < 0.05). No significant impact was observed for cook loss of LL samples.

In the GM muscle, purge loss and cook loss were not impacted by the treatments (P > 0.05). The drip loss, however, was impacted by size×freezer interaction effect in GM samples. Greater drip loss was observed in the STK-CST sample compared with the SEC-CST, SEC-FRI, and STK-BLS samples (P < 0.05). Those treatments, however, were not different from SEC-BLS and STK-FRI (P > 0.05).

Size and freezer main effects were observed for purge loss in ST, revealing greater loss in SEC samples than in STK samples (P < 0.05). The CST samples from ST had the highest purge loss, whereas BLS samples had the lowest (P < 0.05), although both freezer treatments had similar purge loss compared with FRI samples (P > 0.05). No significant impact was observed on drip loss in ST muscle. Conversely, a significant size × freezer interaction effect was reported for the cook loss of ST. The SEC-BLS had greater cook loss compared with SEC-FRI (P < 0.05). Both treatments were not significantly different from the SEC-CST, STK-BLS, STK-CST, and STK-FRI (P > 0.05).

Purge loss in LL was impacted by a size × freezer interaction effect (P < 0.05), whereas purge loss in ST was impacted by both size effects (P < 0.05) and freezer effects (P < 0.05). Greater purge loss was observed in SEC samples compared with STK samples from both LL and ST. The greater purge loss from the SEC samples corresponds to their slower freezing rates during both the first and second freezing in this study. The amount of meat exudate continues to increase after freezing and thawing, following the structural damage induced by ice crystals. It was suggested that greater loss is often observed as the freezing time increases beyond 19.5 min (Leygonie, 2012). Hence, the higher purge loss observed from SEC samples in this study is attributable to delayed freezing, which may have retarded ice crystal nucleation (slower freezing rate) and promoted the formation of larger and irregular extra-myofibrillar ice crystals (Lee et al., 2022; Grujić et al., 1993; Rahelić et al., 1985), resulting in greater cryo-damage to the product structure. Similar results were previously reported by Setyabrata et al., (2019), in which a faster freezing rate generated less purge loss compared with a slower freezing rate.

Conversely, drip loss was significantly lower in SEC samples compared with STK samples. The reduced loss might be due to the higher purge loss in SEC samples, and thus, relatively less drip could be released compared with STK samples (H. W. Kim et al., 2018). There was a muscle-specific impact of the freezing treatments on cook loss, showing a significant difference only within the ST samples, where the SEC-BLS treatment had the highest cook loss (P < 0.05). The difference in muscle responses could potentially be attributed to the greater connective tissue commonly observed in ST muscle. The presence of high connective tissue was previously reported to cause greater cook loss due to the shrinkage of the connective tissue during the cooking process (Purslow, 2005). Additionally, due to its functionality as a locomotive muscle, ST muscles often have greater type IIb muscle fibers (Mashima et al., 2019). A recent report suggested that type IIb fiber-based muscles expelled more water during cooking compared with those with more type I fiber (Vaskoska et al., 2020), which explains the difference observed only in the ST muscle in this current study. Further study, however, will be needed to confirm the speculation.

A 5-d simulated retail display was conducted to identify the impact of different freezing treatments on the meat color quality and stability of the different beef muscles. Regardless of the muscle, no three-way interactions were observed for all color traits (P > 0.05). With the exception of CIE L* for GM muscle, significant size × day interactions were observed in all color traits, regardless of specific muscle types (Figure 3). The lightness (CIE L*), redness (CIE a*), yellowness (CIE b*), and chroma values were higher in SEC samples on day 1 of display compared with other samples (P < 0.05). The SEC samples were observed to generally maintain a higher color quality value until day 3 or day 4 of display, where the SEC sample had similar color quality compared with STK samples on day 1 of display (P > 0.05). Within the same display day, SEC samples consistently maintained greater color value compared with STK throughout the simulated display (P < 0.05). Similarly, instrumental discoloration (hue angle) was also significantly lower in SEC D1 samples compared with all samples during the color display (Figure 3E). The SEC samples maintained lower discoloration until day 3 of the display in which it had similar discoloration to STK D1 or D2 samples (P > 0.05). The results of the trained color panel evaluation were similar to the instrumental measurements, indicating a greater lean color (redness) score and lower discoloration score for SEC samples compared with STK samples from the beginning until the end of the simulated retail display (P < 0.05) (Figure 4). Hence, the SEC samples indicated greater color quality compared with the STK samples, regardless of muscle type. Although greater SEC had greater color quality, a greater ΔE value was observed in these samples compared with STK samples in all muscles (P < 0.05, Tables 1, 2, and 3), indicating lower color stability and more color alteration in SEC samples during the retail display. No interaction and main freezer effect were observed on ΔE value for all the samples (P > 0.05).

Generally, freezing leads to a reduction in color and color stability observed in meat products (Hergenreder et al., 2013; Aroeira et al., 2016; Kim and Kim, 2017). Furthermore, repeated freeze-thawing practice is known to reduce color integrity in meat (Qi et al., 2012) resulting from its extensive cryo-damage to the muscle structure. In the current study, SEC samples exhibited better color quality, following the 2 freezing treatments, despite their slower freezing rates, which could have caused more cryo-damage due to the formation of larger ice crystals within the products. The lower color quality observed in STK could potentially be due to the exposure of the product surface to the purge accumulated within the packaging. Meat exudate is well-known to contain high heme-iron, which could potentially lead to greater oxidation in meat (Liu et al., 2023), thus resulting in lower color quality in the STK samples. While SEC samples were also exposed to the exudate during the process, a fresh new product surface was generated during fabrication into steaks before display. The larger size of the SEC samples likely limited the interaction with purge to the outer portion of the product, shielding the interior portion from the environment and hence limiting the extent of color degradation. This could potentially explain the improved color quality observed in SEC compared with STK samples.

Additionally, a greater extent of cryo-damage from larger ice crystals to the muscle cell will also lead to mitochondrial disruption (Lee et al., 2022). The loss of mitochondrial activity might contribute to the improved color quality initially observed in SEC samples. Mitochondria are known to compete with myoglobin for oxygen and could hinder the development of oxymyoglobin from presenting a bright cherry-red color in beef (Ramanathan et al., 2020). The disruption of mitochondria also leads to the release of lysosomal enzymes and disruption in metmyoglobin-reducing ability, impairing oxidative stability and causing discoloration (Jia et al., 2021). This could potentially explain the currently observed color changes in SEC samples, in which greater color quality was observed in the beginning, because myoglobin was able to react with oxygen to generate oxymyoglobin without competition from mitochondria. However, the color deteriorates quickly during the display period as the reducing ability of the sample was impaired due to cryo-damage.

The lipid oxidation extent was measured using TBARS and was found to vary between the different muscles used in the current study (Figure 5). A significant size × freezer × day interaction was identified in LL, demonstrating the greatest TBARS value in SEC-FRI samples compared with other treatments at day 1 of display (P < 0.05). On day 5 of display, SEC-CST had higher lipid oxidation compared with other samples (P < 0.05), followed by SEC-BLS, SEC-FRI, and STK-FRI, which were not different from each other (P > 0.05). The STK-BLS and STK-CST had the lowest level of oxidation (P < 0.05) at the end of the 5-d display, with both treatments sharing similar values to samples at the beginning of the display (P > 0.05). The GM muscle was the sole muscle impacted by the display period effect, with all samples having greater lipid oxidation at the end of the 5-d display period (P < 0.01). A size × day effect was found in the ST muscle, indicating no significant difference in lipid oxidation in all treatments, initially. However, SEC samples had higher TBARS values compared with STK samples at the end of the display, regardless of freezer treatments (P < 0.05).

Multiple freezing-thaw cycles have been reported to induce greater oxidation of the final meat products due to the release of pro-oxidants resulting from muscle cell disruption (Xia et al., 2009; Leygonie et al., 2012). Similar to the current study, the application of blast freezing followed by different freezing methods increased the extent of lipid oxidation in pork patties, regardless of the subsequent freezing methods (Tuell et al., 2020). Although the oxidation was increased, in general, freezing the product in a smaller size reduced the extent of the lipid oxidation. In STK products, the faster freezing rate and smaller ice crystal formation could limit the degree of cryo-damage to the muscle structure. This could reduce the extent of pro-oxidant release, thus increasing the oxidative stability of the product compared with SEC samples. Supporting the currently observed oxidation results, SEC samples were also observed to have greater color changes and discoloration rates compared with STK samples. This potentially indicated that the extensive damage from the larger ice crystals greatly reduced the antioxidative capacity of SEC samples.

The WBSF value was only significantly affected by the size effect in LL and GM, demonstrating greater tenderness in SEC compared with STK in both muscles (P < 0.05, Tables 1 and 2). No impact was observed in ST muscles with similar WBSF values regardless of the treatments (P > 0.05, Table 3). A decrease in shear force following freezing has often been reported in the current literature and is attributed to ice crystal formation damaging muscle structure (Grayson et al., 2014; Aroeira et al., 2016; Setyabrata et al., 2019). Although both treatments underwent a freezing process, the current results indicated greater tenderness in SEC samples compared with STK samples. This could be attributed to the nucleation of larger ice crystals resulting from a slower freezing rate in the SEC samples.

Additionally, the greater cell disruption by the ice crystals could also lead to early release and activation of protease enzymes previously contained within the cell. Previous reports by Lee et al. (2021) and Stafford et al. (2024) observed greater activation of calpain 1 and cathepsin B on previously frozen samples compared with nonfrozen samples. Those authors attributed the changes to increased ice crystal damage, disrupting muscle cells and organelle structures. This could potentially contribute to improved tenderness observed in SEC samples compared with STK samples, although further confirmation of these protease activities will need to be made because both SEC and STK samples were frozen in the current study. The release of these enzymes and the freezing damage, however, did not impact the tenderness of the ST muscle, potentially due to the high connective tissue present in the muscle compared with LL and GM. Although there might be some connective tissue structural damage from the ice crystals, it is possible that the extent of structural damage was minimal and did not immediately translate to greater improvement in tenderness for ST samples.

The demographic characteristics of the 120 participants of the sensory evaluation are presented in Supplementary Table 1. The majority (51.7%) of the participants indicated that flavor was the most important palatability trait when consuming beef, followed by tenderness (27.8%) and juiciness (20.6%). The participants mainly preferred their beef cooked to medium-rare (37.8%) or medium (30%), and 81.1% consumed beef 1–5 times/wk.

Consumer panels rated flavor, tenderness, juiciness, and overall quality similarly for LL and ST muscles, regardless of product size and freezer type (P > 0.05, Tables 1 and 3). Likewise, the consumer panels also showed similar acceptability of those palatability attributes in LL and ST muscle (P > 0.05, Tables 1 and 3). Only tenderness in GM muscle was significantly impacted by the size × freezer interaction effect (Table 2). The SEC-FRI had a lower tenderness-liking score compared with SEC-BLS and STK-FRI (P < 0.05), while they were not different compared with all other treatments (P > 0.05). Correspondingly, the consumer acceptability evaluation for GM also showed similar patterns, where SEC-BLS and STK-FRI had the highest tenderness and overall acceptability ratings compared with other treatments (P < 0.05). Overall acceptability for GM muscle was also impacted by the size × freezer effect, showing SEC-FRI to have the lowest acceptability rating (P < 0.05).

The current survey results highlight the changing consumer preferences and focus when selecting beef products. Although tenderness has been the most important palatability trait, recent studies have reported a shift to flavor as the most important palatability trait (McKillip et al., 2017; Hernandez et al., 2023). This change in preferences has been attributed to the improvement and implementation of aging programs to promote consistent production of more tender products (Martinez et al., 2017; Gonzalez et al., 2024).

Although freezing is generally perceived to negatively impact meat palatability, consumer ratings in the current study suggested that different freezing conditions only minimally impacted LL and ST muscle. Eckhardt et al. (2024) reported similar results, noting minimal differences in palatability traits for striploin steaks subjected to double freezing using both blast and consumer freezing methods. Differences in sensory rating were observed only in the GM muscle in the current study. Contradicting the WBSF results, SEC-FRI was rated as the least tender by consumers, although the same treatment was found to have the lowest WBSF value. This inconsistency could potentially be attributed to the high tenderness variation in GM muscle compared with LL and ST (Rhee et al., 2004). Those authors reported that GM had greater variation in connective tissue and sarcomere length, which were suggested to be responsible for the tenderness variation observed in the muscle. It is of interest to note that although not significantly different, STK samples consistently received lower flavor liking and acceptability compared with SEC in all muscles, warranting further investigation.