Bibliometric Review on the Volatile Organic Compounds in Meat

Meat flavor is an important aspect of meat quality that also influences consumer demand, and is therefore very important for the meat industry. Volatile organic compounds (VOCs) contribute in large part to the flavor of meat, and while increasing numbers of articles are published on this topic, reviews of these articles are very scarce. Therefore, our aim was to perform a bibliometric analysis of the scientific publications on VOCs in meat over the period 2000–2020. We selected 611 scientific sources from the Scopus database related to VOCs in meat (seafood excluded). The bibliometric information retrieved included journals, authors, countries, institutions, keywords, and citations. From this analysis, we drew up a list of the most important journals, authors, countries, and institutions, and the trends in VOC research on meat. We conducted a social network analysis (SNA) to identify the collaborations among the many authors and countries, and a keyword analysis to generate a network map of the authors’ keywords. We also determined which meat species were most frequently chosen as research subjects, traced the evolution of the various methods/instruments used, and explored the research tendencies. Finally, we point out the need for further research in defining meat quality, improving meat flavor, identifying adulterants, and certifying the authenticity of meat.


Introduction
Meat is a very important source for human nutrition and health. Ensuring meat quality and safety are, therefore, issues of increasing importance in today's meat industry.
Volatile organic compounds (VOCs) are small molecules deriving from several chemical precursors that are vaporized into the air. They are emitted from the surface of meat, so they have a strong relationship with certain properties of it. VOCs therefore have the potential to be useful tools for assessing various meat quality and safety traits, and for this reason they have recently been attracting a great deal of attention. Overall, they contribute to meat odor and affect consumer satisfaction [1]. Several studies have shown that specific individual VOCs can be biomarkers, providing important information on meat, such as differentiating species [2,3], breeds [4], and length of aging [5]. Moreover, VOC profiles can be used to assess the quality of meat with Protected Designation of Origin (PDO), Protected Geographical Indication (PGI) and Traditional Specialty Guaranteed (TSG) [6,7] certifications [6,7], and have the potential to be a reliable tool for identifying food adulteration [8].
Nowadays, many papers on meat VOCs are the result of multidisciplinary collaborations in areas such as meat safety from the microbiological point of view [9][10][11], changes in meat freshness during storage [12,13], changes in meat flavor following processing [14], the effects on meat aroma of the animals' diet [15,16] and environmental conditions [17,18].
• conduct a search of publications in the Scopus database using the keywords "Volatile organic compounds & Meat"; • restrict the timespan to 2000-2020; • select only English language articles; • screen the 1482 articles retrieved to identify and exclude those not related to the aims of this work, such as those on the VOCs of fish meat or meat broth, or atmospheric pollution related to meat; • compile a dataset of the 611 selected articles for bibliometric analysis; • extract bibliometric information from Scopus by CVS file download, including year, journal, author(s), institution(s), country/ies, keywords, citations, etc.; • manually complete the database to include two columns for information on "meat species" and "instrument/analytical method".

Data Statistics
A quantitative statistical approach was adopted for evaluating the performances of journals, authors, institutions and countries, including the number of articles published per year, total number of citations, the most cited articles, and the most productive authors, institutions and countries. We also considered the scientific influence of journals and authors using metrics such as H-index scores, impact factors, and quartile of journal, and other indicators that were added to each entry in the dataset.
A summary of the information extracted and retained in the dataset is given in Table 1.

Social Network Analysis (SNA)
Social network analysis (SNA) was carried out with the "bibliometric" package in R studio, which revealed the patterns of the relationships among the data, and allowed us to make qualitative and quantitative assessments of the contributions of the variables (authors, institutions, countries) shown on the nodes and the edges [39]. For our study, the objects we picked were authors, institutions, and countries. The network mapping results were visualized with the VOSViewer software.
We used a keywords network analysis to explore the knowledge structures among the research fields; our original bibliometric dataset included "author keywords" and "index keywords", but for this study we used the authors' keywords as they give a more direct indication of the research topic despite being more subjective. Analysis of the authors' keywords reveals the patterns of topic relationships at the micro level, which can help researchers find the hot topics and trends in the field. More simple classification statistics and trends were used for the studied species and instrument information.

Results and Discussion
The earliest research into the VOCs of meat was published at the beginning of the 20th century, when [40] reported finding VOCs in beef for the first time. Over the following decades, only a few papers were published, but in the mid-20th century VOCs in meat started to gain greater attention, and during the second half of the last century about 10 articles per year were published, including some on related topics, such as meat broth, meat extract, and fish meat. By the end of the 20th century, there had been a moderate increase in published research on VOCs, but it has only been in the present century that research on the VOC profiles of meat really took off. The last five years in particular have seen an exponential increase, which seems to be due to breakthroughs in the available instruments/technologies over this period, but also due to increasing interest from consumers and the meat industry, who are now paying greater attention to meat flavor and quality, and food safety.
In this study, we focus only on this century, from the years 2000 to 2020, and exclude research on related topics.

Journals and Publications
As we have seen, Table 1 summarizes the information retrieved from Scopus, and some of the statistics derived from it give an initial overall picture of the data collected. Bibliometric analysis of that database allowed us to extract further information, starting with the trends in the numbers of articles published and the citations they received over the period of time studied. Figure 1 shows the number of articles by year of publication and the number of citations the articles received in the years following their publication (sum of total citations per year of publication). The general trend is increasing and can be divided into three periods: the first is 2000-2007, when the number of published articles varied between 5 and  ; the second period is 2008-2015, which saw an increase from 25 to 35 papers  per year; the third period is the last 5 years (2016-2020), when the number of publications  increased from 38 to 66 per year. The total number of citations is obviously related to the number of articles published. In the first period there were about 50 citations per article, in the second period about 30 citations per article, and during the last period, as expected, the number of citations per paper dropped from about 20 per article to just a few for the most recently published, although the number of citations of these articles is, of course, expected to increase greatly over the coming years. It is also clear that the large number of citations received by the older articles is related to the large increase in articles published in subsequent years. Figure 1 shows the number of articles by year of publication and the number of citations the articles received in the years following their publication (sum of total citations per year of publication). The general trend is increasing and can be divided into three periods: the first is 2000-2007, when the number of published articles varied between 5 and 20 per year; the second period is 2008-2015, which saw an increase from 25 to 35 papers per year; the third period is the last 5 years (2016-2020), when the number of publications increased from 38 to 66 per year. The total number of citations is obviously related to the number of articles published. In the first period there were about 50 citations per article, in the second period about 30 citations per article, and during the last period, as expected, the number of citations per paper dropped from about 20 per article to just a few for the most recently published, although the number of citations of these articles is, of course, expected to increase greatly over the coming years. It is also clear that the large number of citations received by the older articles is related to the large increase in articles published in subsequent years. The conclusion that can be drawn is that there is a large and growing interest on the part of the scientific community in the VOC profile of meat, and this interest is expected to increase further in the near future.

The Most Prolific Journals and Their Evaluation and Evolution
From 2000 to 2020, 98 different journals published articles on meat VOCs. Of these, 48% are ranked in the first quartile (Q1) of their highest subject category, 31% in Q2, 13% in Q3, and 8% in Q4 or unknown. The average H-index of the journals is 83, and the average impact factor 2.5. From our quantitative analysis, we obtained the ranking of the journals according to the number of articles they had published; the top 25 are listed in Table 2. The conclusion that can be drawn is that there is a large and growing interest on the part of the scientific community in the VOC profile of meat, and this interest is expected to increase further in the near future.

The Most Prolific Journals and Their Evaluation and Evolution
From 2000 to 2020, 98 different journals published articles on meat VOCs. Of these, 48% are ranked in the first quartile (Q1) of their highest subject category, 31% in Q2, 13% in Q3, and 8% in Q4 or unknown. The average H-index of the journals is 83, and the average impact factor 2.5. From our quantitative analysis, we obtained the ranking of the journals according to the number of articles they had published; the top 25 are listed in Table 2.
The top journal with articles (i.e., one-quarter of the total) is Meat Science, which published studies on a large variety of topics. The second is Food Chemistry, with sixty articles (one tenth of the total), whose main topics concern meat composition, metabolism pathways and comparison of methods/instruments. The next is Journal of Food Science, with 36 articles, mainly concerned with describing VOC profiles rather than with the complexities of VOC pathways. Another aspect is the scientific influence of the journals; so, we re-ranked the journals by impact factor and H-index. The journals with the highest impact factors (10.3 to 7.1) are in the engineering fields (Biosensors and Bioelectronics, Ultrasonics-Sonochemistry, Sensors and Actuators, B: Chemical), although the number of articles published is very low. The journals with the highest H-indices are those publishing a large number of articles per year combined with a high impact factor. In first place is Analytical Chemistry (310), followed by It is worth noting that out of the 25 top journals, 10 are published in the United States, while about one-third are published by a single publishing company (Elsevier, Amsterdam, The Netherlands).

Most Cited Articles
The top 25 most cited articles dealing with meat VOC profiles are listed in Table 3. Both of the first two papers discuss changes in the VOC profiles of beef due to spoilage by bacteria, and each had close to 200 citations at the time of retrieval. They were published in 2009 and 2015, with corresponding authors from the University of Naples Federico II, Italy, and the Agricultural University of Athens, Greece, respectively. Both research groups have strong collaborations with other groups in this field. The third article, from the Matforsk AS-Norwegian Food Research Institute, Norway, deals with the antioxidant activity of grape seed extract to prevent meat spoilage, and puts forward new ideas on non-destructive methods for monitoring meat quality (hexanal and pentanal as the VOC markers). If we look at the total citations per year, we can see that the top four articles are among the six articles with more than ten citations per year (Table 3).

The Most Productive Authors and Their Collaborations
The 611 selected articles on meat VOC profiles were authored by 1862 researchers, the 20 most productive of whom are listed in Table 4. In terms of number of articles, the most important research group is that of D.U. Ahn at Iowa State University (USA), who published 38 articles during this 20-year period, receiving more than 1000 citations altogether, while 17 of these articles received more than 17 citations each (H-index). Ahn and his group worked on various research topics, such as meat composition, meat spoilage caused by bacteria, and VOC pathways during processing, and mainly on pork, chicken, sausages and ham.
The second most productive author in terms of articles published (37), but top for the number of citations received, is J.M. Lorenzo of the Centro Tecnológico de la Carne de Galicia in Spain, whose H-index was 21 (the highest) for articles on VOCs. His main research topic is the use of antioxidants in meat.
The third most productive author is M. Flores of the Instituto de Agroquímica y Tecnología de Alimentos (IATA-CSIC) Valencia, Spain (24 articles, 651 citations and Hindex 16), who works in particular on the VOC profiles of fermented meat.
It is worth noting that half of the authors listed in Table 4 are Spanish, almost a quarter are American, and a quarter are Chinese.
Having examined the authors by country, we now turn to our analysis of the networks of authors' collaborations, represented in Figure 2.
H-index 16), who works in particular on the VOC profiles of fermented meat.
It is worth noting that half of the authors listed in Table 4 are Spanish, almost quarter are American, and a quarter are Chinese.
Having examined the authors by country, we now turn to our analysis of the ne works of authors' collaborations, represented in Figure 2. The fourth author listed in Table 4, E.J. Lee, (15 articles, 264 citations and H-index 10 is one of D.U. Ahn's close co-workers. Other authors in the leading group (Figure 3), wh are involved in the most collaborations are Wang Y, Zhang Y, and Feng X, all from China an indication of the rapid growth in this research topic in East Asia (not only China) an their broad collaborative approach. In contrast, the major European groups working o meat VOCs are clearly "stand alone" groups that do not enter into many collaboration with other groups in the same or other countries. The fourth author listed in Table 4, E.J. Lee, (15 articles, 264 citations and H-index 10) is one of D.U. Ahn's close co-workers. Other authors in the leading group (Figure 3), who are involved in the most collaborations are Wang Y, Zhang Y, and Feng X, all from China, an indication of the rapid growth in this research topic in East Asia (not only China) and their broad collaborative approach. In contrast, the major European groups working on meat VOCs are clearly "stand alone" groups that do not enter into many collaborations with other groups in the same or other countries.  Table 5 shows the 10 most productive institutions in the research field of meat VOC profiles. As can be seen, the leading institution globally is Iowa State University in the USA, the only American institution in the top 10 research centers.

The Most Productive Countries and Institutions and Their Collaborations
Of the other nine institutions, four are in Southern Europe (Spain and Italy), and 5 are in East Asia (China, South Korea and Japan). The top Spanish institutions are two public institutes and one university, the top Italian institution is a university, and four of the five East Asian institutions are universities. Spain, China, USA, Italy and South Korea also hold leading positions among the most productive countries in terms of number of articles published and citations received (Figure 3). Between them, they account for about three quarters of all articles and citations from a total of 50 countries worldwide. The Asian countries have fewer citations per publication than the other countries, mainly because most of their articles were published in the last few years.    Table 5 shows the 10 most productive institutions in the research field of meat VOC profiles. As can be seen, the leading institution globally is Iowa State University in the USA, the only American institution in the top 10 research centers.

The Most Productive Countries and Institutions and Their Collaborations
Of the other nine institutions, four are in Southern Europe (Spain and Italy), and 5 are in East Asia (China, South Korea and Japan). The top Spanish institutions are two public institutes and one university, the top Italian institution is a university, and four of the five East Asian institutions are universities. Spain, China, USA, Italy and South Korea also hold leading positions among the most productive countries in terms of number of articles published and citations received (Figure 3). Between them, they account for about three quarters of all articles and citations from a total of 50 countries worldwide. The Asian countries have fewer citations per publication than the other countries, mainly because most of their articles were published in the last few years.
It is worth noting that all these leading countries are conducting basic and applied research, although a larger proportion of American research is focused on meat processing and the needs of the meat industry, while Southern European countries are more concerned with food safety, traditional products, and meat from local breeds, and Asian countries with technological issues and product characterization, including traditional foods.
Regarding inter-country collaborations, researchers from the USA collaborate in particular with researchers from East Asia (mainly China and South Korea), whilst Southern European countries collaborate with each other, Northern European countries (especially the UK, France, the Netherlands and Poland) and Latin America (mainly Brazil and Mexico).

Keyword Analysis and Networks
There are two types of keywords in the bibliometric dataset, authors' keywords and index keywords. For our study, we used authors' keywords in order to capture the central topics of their research [65]. The original dataset contained 1447 keywords, although of course a large number of these had similar meaning, such as "Volatile compounds" and "Volatile organic compounds", "meat" and "meats", "flavour" and "flavor", and so on, so we standardized the keywords to avoid meaningless relationships. We were then left with 1354 keywords for the next step of the analysis. Table 6 shows the top 20 authors' keywords listed in order of the number of occurrences, and classified according to the object of the study, the method and instrument used, the research topic, and the type of meat analyzed. "Volatile organic compounds" is the core concept of this review, so it is not surprising that it was also the keyword most frequently used by the authors. "Odor", "Flavor" and "Sensory" were the most frequent keywords related to VOCs, and of these, odor depends directly on the individual VOC in question, while flavor is a comprehensive indicator combining "odor", taste and other sensory characteristics [66]. The authors' keywords also included "color" and the general term "meat quality".
The table also shows the major research topics of the articles on VOC profiles: "lipid oxidation" is the most frequent keyword, followed by "irradiation", "fatty acids", "and "spoilage". Lastly, regarding the product analyzed, aside from the general term "meat", the keywords in order of frequency were "Beef" > "Pork" > "Dry-cured ham" > "Lamb" > "Chicken", whereas "sausage" and "burger" were found much less often.
In addition to the number of occurrences, we also analyzed the co-occurrences. Figure 4 illustrates the networks of authors' keyword co-occurrences and shows some interesting associations.

Types of Meat and Meat Products
Analysis of the authors' keywords revealed that VOC profiling of meat was mainly focused on two groups of meat products: one was muscle cuts, raw or cooked, obtained from different species and categories of meat animals, and the other was meat products obtained from industrial/artisanal processing. The first group mainly comprised meat cooked by different methods, such as grilling [69], frying [9], microwaving [70], boiling [71], etc. The second group mainly included fermented meat, ham, sausages (pork, horsemeat, poultry) [7], and burgers/patties (from different meats) [31,72]. It is worth noting that little research has been conducted on the effects of different cooking techniques on VOC profiles, so a future research priority should be the interaction between cooking method and the type of meat cooked. There is also little or no research comparing the VOC profiles of raw and cooked meat.
The table also shows the major research topics of the articles on VOC profiles: "lip oxidation" is the most frequent keyword, followed by "irradiation", "fatty acids", "an "spoilage". Lastly, regarding the product analyzed, aside from the general term "meat the keywords in order of frequency were "Beef" > "Pork" > "Dry-cured ham" > "Lamb" "Chicken", whereas "sausage" and "burger" were found much less often.
In addition to the number of occurrences, we also analyzed the co-occurrence Figure 4 illustrates the networks of authors' keyword co-occurrences and shows som interesting associations.

Types of Meat and Meat Products
Analysis of the authors' keywords revealed that VOC profiling of meat was main focused on two groups of meat products: one was muscle cuts, raw or cooked, obtaine from different species and categories of meat animals, and the other was meat produc obtained from industrial/artisanal processing. The first group mainly comprised me cooked by different methods, such as grilling [69], frying [9], microwaving [70], boilin [71], etc. The second group mainly included fermented meat, ham, sausages (por horsemeat, poultry) [7], and burgers/patties (from different meats) [31,72]. It is wor noting that little research has been conducted on the effects of different cooking tec niques on VOC profiles, so a future research priority should be the interaction betwee cooking method and the type of meat cooked. There is also little or no research compa ing the VOC profiles of raw and cooked meat. We also classified the articles according to the type of meat studied, and the results were similar but not identical to the ranking of authors' keywords ( Table 6). The highest proportion of all studies were carried out on beef (20%), especially in Europe and America, followed by ham (19%), especially the high-value dry-cured hams typical of Spain and Italy, pork (18%), the most frequently studied meat in China, and sausages (17%). Taking together ham, sausages and pork, the porcine species was the subject of more than half of all studies. The meat from small ruminants (lamb, mutton and goats) was the subject of 10% of all studies, chicken 8%, and other species/products (chicken, turkey, duck, goose, rabbit meat, etc.) the remaining 8%.
It is worth noting that very few articles deal with more than one type of meat, so the comparison and authentication of species is another research priority. Only one recent article [3] compared the VOC profiles of meat patties obtained from five species/categories (beef, veal, pork, chicken and turkey).

The Evolution of VOC Analytical Methods and Instruments
Aside from meat sampling and preparation, VOC profiling follows three steps: first, detecting the different molecules or fragments of them; second, identifying their chemical nature; and, lastly, associating the identified substance with a corresponding odor through the scientific literature or sensory evaluation by an expert or an olfactory instrument.
The 611 articles selected used several different analytical methods/instruments to detect VOCs. The aim of this review paper is not to provide an analytical description or compare the pros and cons of the different methods, but rather to trace the evolution of the various methods over time and consider the prospects in this field. As the same methods/instruments are often referred to in different articles by different names and/or acronyms, we first had to classify the methods according to analytical principle, and then group them into four major categories. This classification is summarized in Figure 5.
compare the pros and cons of the different methods, but rather to trace the evolution of the various methods over time and consider the prospects in this field. As the same methods/instruments are often referred to in different articles by different names and/or acronyms, we first had to classify the methods according to analytical principle, and then group them into four major categories. This classification is summarized in Figure 5.  The four categories were gas-chromatographic methods (group A: GC); methods combining gas-chromatography with mass-spectrometry (group B: GC-MS); direct massspectrometry methods (group C: MS); and, lastly, sensory methods (group D: sensory analysis), including E-nose and E-tongue. Table 7 shows the classification of the various methods/instruments and the evolution of their use over the period 2000-2020. Other sensory methods D-sensory ---2 2 The first category (A-GC) comprises various gas chromatography (GC) instruments/methods, a technique that has very good separation power [73], and can be combined with various detection techniques for qualitative and quantitative analysis of chemicals. Its major limitation is mainly that it is a very tedious, time-consuming process [74]. Another limitation is that the number of VOCs characterizing meat products can be very high and standards are not always available, so it is not easy to identify the compounds producing the chromato-graphic peaks. This explains why, with time, these methods/instruments have failed to attract much interest from researchers.
An example of a technique that uses GC but that can be very useful in research on the odor profiles of meat products is GC-O (Olfactory). This was a very popular technique in the early days of meat VOC research because it was a useful aid to experts in identifying odor VOCs, and is often used in combination with other techniques, such as GC-MS, to investigate the relationship between a chemical substance and a certain odor [75]. The main limitation of GC-O/MS is that it cannot provide quantitative measurements. GC-FID is another technique for identifying VOCs [76], which is usually combined with GC-MS and can be optimized for specific volatile compounds [77].
The second category (B-GC-MS) includes methods that combine gas-chromatography for separating the VOCs with mass-spectrometry for characterizing their chemical composition. This category of methods is very successful because it greatly improves the ability to precisely identify VOCs compared with simple GC methods (but is no less time-consuming). The great success of these methods is due in particular to the possibility of combining them with solid-phase-microextraction (SPME-GC-MS), which uses a fiber coated with an extracting phase to capture the VOCs in the headspace of a vial containing the meat sample, avoiding direct contact with the material. This method was used by more than half the researchers and can even now be considered the reference method for VOC analysis. Other interesting methods in this category are GC-IMS [78] and ultrafast GC-MS [79][80][81].
The third group of instruments makes direct use of mass spectrometry (MS) detection methods without using gas-chromatography to separate the VOCs. Some of these methods have been attracting increasing interest in recent years due to their sensitivity and high throughput. PTR-MS [24,82] and PTR-ToF-MS are of particular interest in this regard [18,33,83]. The latter technique was recently combined with SPME-GC-MS to improve the identification of chemical species from mass peak data [84].
The fourth group comprises sensory instruments, such as E-nose and E-tongue, that are already commonly used in the field of food safety to detect food-borne pathogens [85]. E-nose is a fast, non-invasive, real-time monitoring tool, consisting of non-selective sensors and a computer program that can discriminate patterns of VOCs, that has been used to analyze several meat samples [86]. Recently, several studies have combined the Enose instrument with GC-MS to evaluate meat VOCs; these have shown that while Enose has an excellent ability to discriminate between species, it cannot identify specific chemical substances, so combining it with GC-MS can give better qualitative results [87]. E-tongue is an instrument for evaluating the taste of food, such as sourness, bitterness, and astringency [88], and can be used as a complement to other methods specifically intended for VOC analysis. We found a few papers that used E-tongue in combination with GC-MS to evaluate beef flavor [89,90].

Conclusions
This bibliometric analysis of articles published in scientific journals on the VOC profiles of meat shed interesting light on the state of research, both past and present. First of all, we found a continuously growing interest in research in this field, and identified the leading journals, authors, institutions and countries involved in it. Most of the research on VOCs is carried out in the USA, particularly on meat products and industry needs, in Southern Europe (Spain and Italy), mainly on food safety and traditional meat products, and in East Asia (China and South Korea), mainly on technological issues and traditional foods. Collaboration network mapping revealed important partnerships, especially between the USA and East Asia, and those of Southern Europe with Northern Europe and Latin America. More than half of the studies were carried out on pork (especially in China), and pork products, such as ham (especially in Spain and Italy), and sausages. Regarding ruminants, almost a quarter of the published research was on beef (especially in the USA and Europe), whereas only about a tenth of all articles dealt with meat and meat products from small ruminants. Avian species and rabbits were much less frequently studied.
Numerous different techniques and instruments were also used, and these have evolved over time. Research on VOCs is broadening its objectives, increasingly focusing on food safety, adulteration, and food technology on one side, and authentication of traditional local products, characterization of local breeds and populations, and validating artisanal products, on the other side. Further research which compares meat products from different species and different cooking methods, and the interactions between them, is needed; moreover, it should focus on better characterizing the link between individual VOCs and their perceived odors.