*While the full article for this executive summary is currently under peer review, these initial findings are being made available on BeefResearch.org to enable the industry to act on the research, inform the scientific community of ongoing work, and help prevent duplication of research efforts. Once peer review is complete, a link to the published article will be added to this summary.
KEY TAKEAWAYS
- Intramuscular fat plays a clear role in palatability.
- When comparing cross studies significant differences were found due to grade among strip steaks but this relationship is less prominent in other cuts.
- Significant heterogeneity exists between studies lending to difficulty in predicting specific marbling thresholds for sensory outcomes.
BACKGROUND
Beef carcass quality is greatly dictated by intramuscular fat (IMF) content or marbling score. Frequently consumer liking of beef palatability is correlated with IMF. And, trained sensory evaluation indicates a greater probability for a positive eating experience as IMF increases. However, there are numerous examples in the literature where consumers fail to determine significant palatability differences among beef across all USDA quality grades. Specifically, USDA Select, Low Choice, and/or Upper 2/3 Choice are not rated to be different by consumers. Additionally, consumers have also failed to determine differences in overall liking, flavor liking, and juiciness of USDA Prime, Choice and Select beef. Despite these studies, it is generally recommended that increases in IMF positively influence beef palatability. However, the studies cited above and other studies in published literature are not entirely in agreement with this recommendation. There is no question lipids contribute to sensory response and that IMF contributes to beef palatability. However, it is unclear if the IMF contents observed among quality grades significantly influence consumer liking. Another current limitation is that no study has been conducted to determine a meaningful difference in IMF or marbling score to consumers. This project aims to utilize a meta-analysis approach to elucidate this answer.
The objective of this study was to evaluate the contribution of intramuscular fat to beef palatability using consumer sensory data among beef muscles from carcasses with variable marbling level reported in published literature. The objectives were met by evaluating screening and extracting reported consumer sensory data from published literature.
Methodology
This study utilized a systematic literature approach to identify scientific publications from peer-reviewed journals that evaluated consumer sensory attributes from ≥ 2 marbling or quality grade levels. Search queries were constructed by the research team and reviewed by a University Research Librarian. The final literature search used PubMed, AGRICOLA, Web of Science, Scopus, Academic Search Complete, and ScienceDirect to identify relevant articles. Final number of articles (n=2,396) from these databases were downloaded on October 19, 2023 for use in this study.
Databases were filtered to limit articles to available full text and English language publications. Furthermore, only studies published since 2005 were included because overall beef tenderness has consistently increased since the 1990s as indicated by the National Beef Tenderness Survey series. Additionally, the aim of this study was to capture modern production systems and consumer preferences. Two independent reviewers of the research team evaluated identified studies between the date of database inception through October 2023 publication dates. Inclusion criteria consisted of the following: 1) full text available in English, 2) use of consumers to evaluate beef palatability, and 3) comparison of at least 2 intramuscular fat and/or marbling levels. Primary exclusion criteria included the following: 1) beef palatability was not assessed, 2) did not compare at least 2 groups with differences in marbling score or intramuscular fat, 3) untrained consumer panel data not included, 4) the article was a narrative review or commentary, 5) the article was a systematic review or commentary, 6) the citation was an abstract (or conference proceedings or thesis), 7) included incomplete data or 8) was the same paper that was identified in one of the six databases. After removing duplicates and other articles that did not meet inclusion criteria, n=52 articles were included in the final analysis.
Output variables of interest extracted from final studies included cooked muscle, estimated marbling level (IMF or Quality Grade was used to estimate marbling level in some studies), postmortem aging duration, target degree of doneness, cookery method, country, number of panelists, breed or live animal characteristics, other relevant variables to characterize the sample, consumer ratings of acceptability/overall liking, flavor liking, juiciness, tenderness and reported variability of consumer ratings. Reported consumer sensory data extracted from final articles (n=52) were scaled to a continuous line scale (0 = no intensity for consumer sensory attributes; 100 = very high intensity for consumer sensory attributes). Calculations of standard deviation and differences between means within studies for all possible marbling levels reported were used to compute standardized mean differences (SMD; effect sizes). Since many studies included multiple levels of consumer sensory ratings for different quality grade or estimated marbling levels, the lower estimated marbling level was considered the control group within each analysis comparison.
R statistical software (R Core Team, 2024) were used to analyze consumer sensory ratings by standardized mean differences in which the difference between treatment (ex. estimated marbling level corresponding to USDA low choice) and control (ex. estimated marbling level corresponding to USDA premium choice) means were standardized using the computed standard deviations. Random effects models for each consumer sensory variable to estimate the effect size derived from weighted mean differences, 95% confidence intervals of SMD, pooled random effect significance, and estimates of heterogeneity and heterogeneity significance using the REML method. Plots of SMD between consumer sensory attribute means and estimated marbling differences were constructed in which the size of the data point reflects the weighted effect size to visualize the relationship between SMD and estimated marbling level difference between treatment and control data in step-up comparisons (i.e., reported data corresponding to select to low choice, low choice to premium choice, premium choice to prime). Plots and meta-regression models were utilized to evaluate the impact of finite marbling differences on consumer sensory ratings beyond generalized quality grade category comparisons. Linear regression models were constructed using estimated marbling level to predict consumer sensory rating means. All models included analysis of consumer sensory ratings from striploin steaks, top sirloin steaks and all steaks combined.
The meta-analysis models and boxplot assessment in this study provide preliminary insight into the impact of estimated marbling level on predicting consumer sensory traits of cooked beef muscles. While significant effects of broad quality grade categories were observed in forest plots, visual assessment of the variability that exists among reported consumer sensory attributes between studies suggests that more research is needed to determine what level of marbling increases may be associated with consumer sensory ratings. Estimated marbling levels constrained to reported means under specific quality grade categories limit the ability to test whether incremental levels of marbling or intramuscular fat directly predict consumer sensory traits. Since marbling level was not always reported within each quality grade category in each study, these data are limited in serving as a sole predictor of consumer sensory ratings. These results should be interpreted with caution since assessment of bias and high level of heterogeneity may impact direct comparison between studies and impact overall random effects estimates. While meta-analysis techniques used in this study provided valuable insights and serve as important tools to pool large datasets together, pooling raw data (i.e., prior to computing means and estimates of variance) should allow more direct estimates of incremental marbling level to be used in future statistical models evaluating the effect on consumer sensory attributes. In addition, future research focused on assessing consumer sensory attributes from more levels of marbling within a quality grade category could allow researchers to better predict the role of marbling among consumer responses.
Implications
The results of this study help to better understand the role that marbling level (or estimated marbling level derived from intramuscular fat) may play in predicting consumer sensory ratings. These pooled data from published studies will serve as foundational estimates to compare future research findings to. By pooling large datasets of raw data (i.e., individual data points prior to computing means and estimates of variance), additional research should be able to better estimate the incremental effect of marbling level increases on consumer palatability traits. These data suggest that future consumer sensory work should be conducted to better predict smaller increases in marbling scores within quality grade categories on consumer sensory responses.