Research Trends in Advanced Glycation End Products and Obesity: Bibliometric Analysis

The aim of this study was to conduct a bibliometric analysis of the scientific articles on advanced glycation end products (AGEs) and obesity. English-language journal articles about AGEs and obesity were retrieved from the Scopus database. The OpenRefine application was used for data cleaning, the VOSviewer software program for analysis of the trends of year of publication, country, institution, journal, authors, references, and keywords. Microsoft Excel and Tableau Public were applied for the visualizing of the publication trends. Data collection was performed on 3 February 2022, from a total of 1170 documents. The Mann–Whitney test and Spearman test with software SPSS ver.28.0.1.1. were used to assess the relation between open access journal statuses, years of publications, and CiteScore. The results of the study showed that there was an increase in studies on processed foods, including AGEs and obesity. The United States was the country with the largest contribution in this field, with the highest number of citations. The Nutrients journal published the largest number of articles on this topic, particularly in the last two years. The present focus of the studies is on ultra-processed foods. The open access journals have younger medians of the year of publication and higher medians for number of citations than do closed access journals (p < 0.001 and p < 0.05, respectively). A strong negative association was seen between CiteScore and the year of publication (r = −0.64 [95% CI: −0.67, −0.60]), p < 0.001. We present this bibliometric analysis to furnish the most recent data on the description, visualization, and analysis of AGEs and obesity.


Introduction
Obesity is a global problem that is still unsolved. This may also be seen in Indonesia from the yearly increases in the prevalence of obesity [1]. The shift in lifestyle as a result of globalization and modernization leads to an increase in sedentary activities and the consumption of unhealthy foods such as processed foods, and in overnutrition [2,3]. This lifestyle, in conjunction with smoking [4] and hyperglycaemia [5], is known to increase the level of advanced glycation end products (AGEs) in the body, which subsequently cause disease complications [6].
Publications about AGEs started to develop with the discovery of AGEs in the body in connection with diabetes [7]. The increase in publications became more noticeable after the breakthrough discovery of AGEs in foods [8] that are consumed daily [9][10][11][12][13]. There were also other discoveries, such as the effect of heating [14], pH [15], and the duration of food storage [16], all of which may be changed to reduce the level of AGEs in foods. In spite of this, in general the studies demonstrated that processed foods and high fat

Materials and Methods
The methodology of this study may be divided into three phases, i.e., the identification of the sources and criteria for the data search, software selection and data extraction, and finally data analysis and interpretation ( Figure 1).

Phase I-Identification of Data Sources and Search Criteria
The investigators tried to identify publications on AGEs in an objective manner: to analyze the distribution of scientific productions with regard to researchers, institution, or country, and to evaluate the structure and evolution of the field of AGE research. Our study used documents obtained from the Scopus database, which is a database of peer-reviewed journals that allows bibliometric studies in analysis and visualization, supporting the scope of journals on AGEs.
The inclusion criteria were then confined to the final publications of articles in Englishlanguage journals and excluding review articles, book chapters, conference articles, books, notes, letters, errata, short surveys, and others. However, our study did not place limits on the year of publication. The final number of documents in this phase was 1170 documents.
The inclusion criteria were then confined to the final publications of articles in English-language journals and excluding review articles, book chapters, conference articles, books, notes, letters, errata, short surveys, and others. However, our study did not place limits on the year of publication. The final number of documents in this phase was 1170 documents.

Phase 2-Software Selection and Data Extraction
The mined data from Scopus were exported as comma-separated values in Microsoft Excel Home and Student 2019 edition version 16.57. The information that was downloaded for each publication was data on authors, abstract, keywords, document titles, year, number of citations, bibliographic information, funding details, open access journal label, and references. The values of Source Normalized Impact per Paper (SNIP) were taken from the article of Bajjou M.S. [26]. The statistics of Q-ranking, SCImago Journal & Country Rank (SJR), and h-index were taken from the website https://www.scimagojr.com/journalrank.php (accessed on 29 March 2022). The

Phase 2-Software Selection and Data Extraction
The mined data from Scopus were exported as comma-separated values in Microsoft Excel Home and Student 2019 edition version 16.57. The information that was downloaded for each publication was data on authors, abstract, keywords, document titles, year, number of citations, bibliographic information, funding details, open access journal label, and references. The values of Source Normalized Impact per Paper (SNIP) were taken from the article of Bajjou M.S. [26]. The statistics of Q-ranking, SCImago Journal & Country Rank (SJR), and h-index were taken from the website https://www.scimagojr.com/journalrank. php (accessed on 29 March 2022). The VOSviewer software version 1.6.17, from Leiden University, Leiden, Netherlands, was used for bibliometric construction and graphics. In data cleaning, the spelling of words with the same meaning was uniformly formatted using the OpenRefine software version 3. 5

Phase 3-Data Analysis and Interpretation
Data were analyzed and interpreted objectively and subjectively, and subsequently underwent structured and documented evaluation. Performance analysis was performed objectively, i.e., by evaluating the trends for year of publication, country, journal, and greatest number of documents and citations. Scientific mapping was performed subjectively, namely by evaluating keywords, author networks, and scientific journal co-citation networks. We also analyzed the relation of open access journal status, including years of publications and CiteScore.

Distribution of Publication by Year
There were 1170 publications of studies on AGEs in processed foods and obesity, in the form of articles in English-language journals from 1988 to 2022 in the Scopus database (Table S1). Figure 2 shows that there were yearly increases in the number of research publications. There was a considerably sharp increase in 2003 that triggered a surge in the number of publications in the following years.
year of publication and number of citations.

Phase 3-Data Analysis and Interpretation
Data were analyzed and interpreted objectively and subjectively, and subsequently underwent structured and documented evaluation. Performance analysis was performed objectively, i.e., by evaluating the trends for year of publication, country, journal, and greatest number of documents and citations. Scientific mapping was performed subjectively, namely by evaluating keywords, author networks, and scientific journal co-citation networks. We also analyzed the relation of open access journal status, including years of publications and CiteScore.

Distribution of Publication by Year
There were 1170 publications of studies on AGEs in processed foods and obesity, in the form of articles in English-language journals from 1988 to 2022 in the Scopus database (Table S1). Figure 2 shows that there were yearly increases in the number of research publications. There was a considerably sharp increase in 2003 that triggered a surge in the number of publications in the following years.
The publication data in the graph for 2022 seem lower than 2021. This is because data are taken until 3 January 2022, so they do not reflect the publication data in one whole year. However, if it is simulated with the same publication rate per month, the estimated publication figure reaches 200 articles in 2022, or exceeds the number of publications in 2021.  The publication data in the graph for 2022 seem lower than 2021. This is because data are taken until 3 January 2022, so they do not reflect the publication data in one whole year. However, if it is simulated with the same publication rate per month, the estimated publication figure reaches 200 articles in 2022, or exceeds the number of publications in 2021.

Contributions by Country and Region
Bibliometric coupling analysis was used to evaluate the number of documents from the researchers' country of origin. There were 112 countries that met the minimum criterion of one document. The distribution of the countries involved in the publications may be seen in Figure 3. The United States predominates in the number of publications on AGEs and obesity with 382 documents. Bibliometric coupling analysis was used to evaluate the number of documents from the researchers' country of origin. There were 112 countries that met the minimum criterion of one document. The distribution of the countries involved in the publications may be seen in Figure 3. The United States predominates in the number of publications on AGEs and obesity with 382 documents.  Table 1 shows the top 15 countries based on the number of documents in the field of AGEs and obesity using VOSviewer. Brazil, the UK, and Australia are the countries contributing the largest number of publications after the United States. Regarding the number of cited publications, the US also occupies the highest rank with 19,362 citations, while the UK and Italy occupy the second and third ranks with 4203 and 3988 citations, respectively. The developed countries that are equated with a high consumption of fast foods and processed foods dominate the number of documents in this field.   Table 1 shows the top 15 countries based on the number of documents in the field of AGEs and obesity using VOSviewer. Brazil, the UK, and Australia are the countries contributing the largest number of publications after the United States. Regarding the number of cited publications, the US also occupies the highest rank with 19,362 citations, while the UK and Italy occupy the second and third ranks with 4203 and 3988 citations, respectively. The developed countries that are equated with a high consumption of fast foods and processed foods dominate the number of documents in this field.  Table 2 shows the three topmost ranking countries per cluster based on the number of documents. In Cluster 2 (coded green), the three highest ranking countries from Asia among the fifteen countries with the highest number of documents are Japan, China, and India with 57, 53, and 37 documents, respectively.  Oman and South Sudan collaborate with the UK and the US, having links of 80 and 78, respectively, which are the lowest collaboration levels if compared with the other countries in this field. Lontchi-Yimagou, E., from Cameroon in Cluster 5, published an article on AGE receptors as inflammatory biomarkers in the body that may be associated with insulin resistance and the complications of obesity. This publication contributes much to the knowledge of the probable inflammatory pathway, which may have increased the number of citations for this document. The collaboration of the authors from Cameroon with other authors from the United States, the Netherlands, Australia, and South Africa increase the number of links to 26, if compared with Oman (4) and South Sudan (4). This is also seen in Cluster 7, where Cyprus contributes one document on dietary quality, particularly in connection with ultra-processed foods, in subjects participating in the European I.Family Study. In the latter study, there are 17 authors involved in writing documents with authors from Cyprus with 12 affiliations and 9 countries with links to Cyprus (47 links), which is higher than for authors from Cameroon, with 4 authors (26 links). The difference in year of publication may have contributed to the difference in the number of citations for Cyprus (for 2021) in comparison with Cameroon (for 2013).   Open access journals may be defined as those containing documents that may be freely downloaded and read online without having subscriptions or filling their affiliations [27]. Our study shows that the majority or 60.9% (713/1170) of journals are open  Open access journals may be defined as those containing documents that may be freely downloaded and read online without having subscriptions or filling their affiliations [27]. Our study shows that the majority or 60.9% (713/1170) of journals are open access. Table 3 shows that the median of the year of publication in open access journals is lower than in non-open access journals (p < 0.001); in other words, the articles are more recent. Our study shows that open access journals have higher medians for the number of citations than do non-open access journals (p < 0.05). The CiteScore has a negative association with the year of publication (r = −0.64 [95% CI: −0.67, −0.60], p <0.001).  Table 4, below, shows the top 15 journals based on the number of documents. The journal Nutrients contributed 5.6% of all documents on AGEs and obesity. The US is predominant in being the country with journal domains that publish the greatest number of articles on AGEs and obesity (seven journals), followed by the UK (five journals) and Europe (three journals).   Table 5 showed the 15 most-cited documents, representing 27.6% of the total number of citations.

Collaboration between Authors
The collaboration between authors was evaluated using co-authorship by means of VOSviewer. There were 98 authors who were co-authors, among the 5954 authors having a minimum number of four documents. Figure 6 shows the 28 authors who collaborated with other authors in the writing of articles.
The collaboration of Cai, W., Vlassara, H., Uribarri, J., Striker, G.E., Chen, X., and Zhu, L. is most apparent as indicated by the proximity of the circles and the width of the lines connecting the circles. These authors have the same affiliation, namely Icahn School of Medicine at Mount Sinai, New York, NY, USA. Vlassara, H. and Cai, W. have the highest total link strengths of 19, followed by Uribarri, J. (18), Chen, X. (18), and Striker, G.E. (15). Chen, X. and Zhu, L. also collaborated with researchers outside their institution, as shown by the lines connecting the green and red clusters. The other authors in the red cluster collaborated more frequently with researchers in other clusters. This may be seen from the large number of lines connecting the red circles with circles of other colors, such as green, dark blue, light blue, and violet. Li, Y. is the topmost co-author (10 documents), followed by Uribarri, J.

Analysis of Keywords
Bibliometric analysis was used to evaluate the correlation of the most frequent keywords used by the authors of scientific articles on associated topics. This may facilitate the search for concepts and topics in the concerned field. This analysis used the VOSviewer application that can build a multidimensional representation of various topics. There are 53 keywords that appear minimally 10 times out of a total of 2505 keywords on this topic. The keywords are grouped into four clusters with a total of 440 links and a total link strength of 1434.
Obesity is the most popular keyword used by authors on this topic, and is found in Cluster 3 with 310 co-occurrences and has links with 50 other words. The next most frequently used keywords in Cluster 3 are physical activity, childhood obesity, food addiction, and community health problems. The keyword AGEs is the second most-used by several researchers in connection with this topic and lies in the red-colored Cluster 1. The

Analysis of Keywords
Bibliometric analysis was used to evaluate the correlation of the most frequent keywords used by the authors of scientific articles on associated topics. This may facilitate the search for concepts and topics in the concerned field. This analysis used the VOSviewer application that can build a multidimensional representation of various topics. There are 53 keywords that appear minimally 10 times out of a total of 2505 keywords on this topic. The keywords are grouped into four clusters with a total of 440 links and a total link strength of 1434.
Obesity is the most popular keyword used by authors on this topic, and is found in Cluster 3 with 310 co-occurrences and has links with 50 other words. The next most frequently used keywords in Cluster 3 are physical activity, childhood obesity, food addiction, and community health problems. The keyword AGEs is the second most-used by several researchers in connection with this topic and lies in the red-colored Cluster 1. The green-colored Cluster 2 contains 18 nodes that are dominated by ultra-processed foods, whereas the yellow-colored Cluster 4 is focused on cardiovascular disease. (Table 6). Figure 7 shows that obesity is at the center of the keywords used by researchers and is closely associated with AGEs. The strength of the association between the keywords AGEs and obesity is shown by the thickness of the lines connecting the two circles (34 links). The use of AGEs as a keyword occurs together with the keywords oxidative stress (20 links), inflammation (15 links), insulin resistance (10 links), and type 2 diabetes mellitus (18 links). The topic of obesity is also frequently associated with ultra-processed foods (29 links). The circles of obesity and AGEs are seen to be closer to each other than to ultra-processed foods. The proximity of AGEs and AGE receptors is shown by the shorter distance between their circles (14 links). The topic of obesity is close to diet (34 links), which is associated with nutrition transition (12 links), quality (12 links), and dietary pattern, such as processed foods (19 links), ultra-processed foods (31 links), and fast foods (12 links).  (18 links). The topic of obesity is also frequently associated with ultra-processed foods (29 links). The circles of obesity and AGEs are seen to be closer to each other than to ultra-processed foods. The proximity of AGEs and AGE receptors is shown by the shorter distance between their circles (14 links). The topic of obesity is close to diet (34 links), which is associated with nutrition transition (12 links), quality (12 links), and dietary pattern, such as processed foods (19 links), ultra-processed foods (31 links), and fast foods (12 links).

Co-Citation Networks of Cited Authors
There were 94,188 authors who became references in minimally two studies on AGEs and obesity, obtained through bibliographic analysis using VOSviewer, among whom 802 authors had 20 minimal citations. Table 7 below shows 15 of the most cited authors.

Co-Citation Source Networks
This analysis refers to direct observations on journals that have been repeatedly cited in AGE studies. The most active and influential scientific sources were identified using VOSviewer. There were 15,214 scientific sources in this study, and with the determined minimum threshold of 20 citations, we obtained 425 scientific sources that met the criteria for further discussion. The map below shows the results of the VOSviewer analysis, indicating the presence of six clusters in the network (Figure 9).

Co-Citation Source Networks
This analysis refers to direct observations on journals that have been repeatedly cited in AGE studies. The most active and influential scientific sources were identified using VOSviewer. There were 15,214 scientific sources in this study, and with the determined minimum threshold of 20 citations, we obtained 425 scientific sources that met the criteria for further discussion. The map below shows the results of the VOSviewer analysis, indicating the presence of six clusters in the network (Figure 9).

Co-Citation Source Networks
This analysis refers to direct observations on journals that have been repeatedly cited in AGE studies. The most active and influential scientific sources were identified using VOSviewer. There were 15,214 scientific sources in this study, and with the determined minimum threshold of 20 citations, we obtained 425 scientific sources that met the criteria for further discussion. The map below shows the results of the VOSviewer analysis, indicating the presence of six clusters in the network (Figure 9).

Discussion
Our research showed there is an increase in references to AGEs in processed food publications especially from 2003 onwards. This increase in the number of publications may have been affected by several factors. The first is the increased prevalence of obesity in nearly all countries, both developed and developing countries [43,44]. Obesity is no longer limited to a certain socio-economic group [45]. Globalization, with changes in dietary patterns and sedentary lifestyles, both in urban and rural areas, also contributes to the occurrence of obesity. The second factor is the discovery of AGE pathways [29,[46][47][48][49][50] and of databases of AGEs in foods from several countries [10,11,13,51,52]. The third factor is the open information system, where an increasing number of journals offer open access, [53,54] and the availability of big data to facilitate access to the most recent knowledge [55,56].
Nutrients is the journal that accepts the greatest number of articles in this field. The number of articles in Nutrients has increased sharply since 2018 with 7 documents, and continued to increase in 2019 and 2020 with, respectively, 16 and 22 documents per year. Although in 2021 the number of documents was 18, a decrease of 18.2% from the previous year, nevertheless this journal is still the largest contributor to the topics of AGEs and obesity. The decrease in the number of articles may have been caused by the current COVID-19 pandemic. The results of this study are in line with the study of Riccaboni, M. and Verginer, R., who showed a decrease in the number of non-COVID-19 publications by 10-12% during the pandemic in 2019 and 2020 [27].
In Figure 5, a thick line may be seen that connects PLoS ONE with Public Health Nutrition (link strength 12), showing the sufficiently high number of researchers preferring to publish articles in both the aforementioned journals. The node for Nutrients is adjacent to that of PLoS ONE, although differing in color. This may have been due to the fact that the accepted articles covered the same topics, although in general there is a difference in coverage between the two journals. The acceptance criteria in Nutrients are nutrition that is associated with human health and the methods for evaluating nutritional status, whereas PLoS ONE covers physics, medicine, engineering, social sciences, and the humanities. The most frequently cited article is 'Global nutrition transition and the pandemic of obesity in developing countries', issued in 2012 by Popkin from the University of North Carolina, USA [28]. His article exposes the changes in lifestyle, i.e., an increased consumption of processed foods and decreased physical activity, which have attracted attention in low-and middle-income countries, both in urban and rural areas.
The hypothesis on the AGE pathways and the complications of type 2 diabetes was published by Evans, J. from the Medical Research Institute (J.L.E.), San Bruno, California, USA, in an article titled "Oxidative stress and stress-activated signaling pathways: A unifying hypothesis of type 2 diabetes" [29]. This document is the second most-cited one (1680 citations). The article 'Dietary emulsifiers impact the mouse gut microbiota promoting colitis and metabolic syndrome' is the third most-cited document [30]. Chassaing, B. discusses the metabolic syndrome induced by the emulsifiers carboxymethylcellulose and polysorbate-80, which were subsequently associated with changes in the species composition of the microbiota and the increase in pro-inflammatory potential [30].
Upon further analysis of these top 15 cited documents, it turns out that the most frequent topic is AGEs with noncommunicable disease (NCDs) complications, such as polycystic ovary syndrome (PCOS), diabetes mellitus, metabolic syndrome, or diabetic nephropathy [29,[34][35][36][38][39][40]. The next most-cited topics are the increased consumption of processed foods and the effect of ultra-processed foods on obesity [28,31,32,37,42], and the association of AGEs with diet or of endogenous AGEs with inflammation [30,33,41].
Obesity is an abnormal condition due to an excess of fat in the body, which can be the cause of various pathological conditions. [1]. The fundamentally strong association between obesity and inflammatory mediators is caused by the chronic exposure to a high AGE diet, resulting in chronic inflammation and insulin resistance [61,62]. It was shown that binding between AGEs and receptor AGEs (membrane-bound receptors of AGE (mRAGE)) could cause cellular dysfunction interfering with cellular communication, changes in protein structure and function, and mitochondrial dysfunction, leading to cell death [63]. Binding to RAGE can also increase reactive oxygen species (ROS), activate inflammatory signaling such as the nuclear factor-k-light-chain-enhancer of activated β cells (NF-kβ), tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and induce insulin resistance, which is often associated with obesity. Meanwhile, binding with other receptors such as AGE-R1, AGE-R2, and AGE-R3 plays a role in AGE clearance [64]. The close relationship between AGEs and the AGE receptors is also shown in several articles [29,39].
It has been observed that serum AGEs ( N-CML, methylglyoxal (MG)) are higher in patients with obesity and one other criterion of the metabolic syndrome than in patients without the metabolic syndrome criteria. This shows that obesity is correlated with a high AGE diet and not with a high calorie intake. Higher TNFα levels have also been reported in the same subjects, consistent with evidence that a high AGE level increases inflammation and insulin resistance [65]. Therefore, circulating AGEs have been suggested as biomarkers for the identification of the transition from obesity to metabolic syndrome, for detection of obesity at risk of metabolic syndrome, and for monitoring the efficacy of dietary interventions [66].
The close positive relationship between AGEs in foods and the level of calories consumed was reported by Baye et al., such that it is difficult to differentiate the individual contributions of each of these risk factors. Based on the tables and the database analysis included in the study paper, fats, meats, cheese, and pulses (if processed, canned, or grilled at high temperatures) have the highest AGE contents. The groups of fats and meats that have a high lipid and protein content tend to have higher AGE contents [9,67].
The study conducted by Uribarri also reported the same, in that although AGEs as a whole are not associated with markers of fat mass, in the presence of obesity and one or more criteria of metabolic syndrome, serum AGE levels are actually higher. Therefore, high AGE levels in obese adults may function as a signal for the transition from obesity to metabolic syndrome [68]. Bearing in mind that AGEs are modifiable non-traditional risk factors, these findings may assist in the timely and targeted prevention of the risk of metabolic syndrome [68,69].
Advanced glycation end products and inflammatory factors were not increased in subjects with obesity who did not have the other criteria of metabolic syndrome and were designated as being metabolically healthy, consistent with the view that an increase in body fat does not always indicate a risk for metabolic syndrome. In this context, the level of AGEs in healthy subjects with obesity may indicate a transition category of metabolic syndrome, such as high blood pressure, dyslipidemia, or subclinical cardiovascular disease, and determine the initial risk toward subsequent metabolic syndrome. Furthermore, outside of excessive nutrition, the exposure to AGE-rich foods may disclose important pointers and a temporal relationship between healthy and unhealthy obesity or obesity at risk for metabolic syndrome [68].
McGlashan, S.R. showed that a high fat diet has a positive correlation with AGEs in serum, tissues, and an increase in the total percentage of fat. High concentrations of AGEs in serum and tissues can be decreased with optimal physical activity intervention. In addition, reducing the intake of AGEs can also reduce the circulating levels of AGEs. From this study, it was found that a high-fat diet and physical activity affected the total body fat percentage, systemic inflammation, accumulation of AGEs, and increased the expression of proinflammatory cytokines. The effect of physical activity can increase the intra-hepatic breakdown process of AGEs that bind to proteins and peptides so that it will increase the elimination and reduction of AGEs in the kidneys [70,71].
Our study tried to identify the highly cited articles, journals, authors, countries, and the latest themes related to AGEs and obesity. Our study highlighted that there is an incremental increase in studies concerning ultra-processed food as part of a poor quality diet and relating to NCDs and obesity. There do not seem to be many data relating to ultra-processed food and the elderly. Therefore, there seems a need for research in this field. In the study of AGEs, procedure standardization in the laboratory between researchers is still needed.
One limitation of the present study is that the downloaded data may have been incomplete, because the extracted data were limited to titles and keywords, the Scopus database, published journal articles, and articles in English. There are other AGE structures that were not specifically written in the search terms such as N -(carboxyethyl)lysine (CEL), pentosidine, glyoxal (GO), and alpha-dicarbonyl compounds. Another limitation of this study is the possibility that other topics that are not related to this study were entered in the analysis and were visualized.

Conclusions
We present this bibliometric analysis to provide the most recent data on the description, visualization, and analysis of AGEs and obesity. There has been an increase in the study of processed foods that includes studies on AGEs and obesity. The United States are the largest contributors of studies in this field, with the highest total of citations. The Nutrients journal has published the largest number of articles on this topic, particularly in the last two years. The focus of the studies is currently centered on ultra-processed foods.