A bibliometric and Altmetric analysis of the 100 top most cited articles on dentin adhesives

Bibliometrics is one of the few subfields involved in the measurement of science outputs [1]. Bibliometric indicators are useful tools for evaluating research performance, provided they are precise, advanced, up-to-date, combined with expert knowledge, and interpreted and applied with care [2]. Citation analysis is a principal bibliometric approach [2]. Citations may not fully reflect the quality of a work, but highly cited articles often present new ideas or address important problems, so they are valuable in the scientific world. Additionally, the frequent citation of an article could be a strong indication of its reliability as a source for researchers to substantiate their methods or arguments [3].

Since 1945, the Institute of Scientific Information (ISI) has been collecting bibliometric data from published scientific papers, but their collection was not launched until the Science Citation Index (SCI), a special tool for measuring citations, was first published in 1962 [4]. Today, the most widely used databases for bibliometric studies are the citation indexes produced by Thomson Reuters, especially Web of Science (WoS) and its predecessor, the SCI [2]. Google Scholar, a tool sponsored by the Internet search company Google, was created to provide users with a simple way of searching a broad range of scientific literature. Google Scholar employs a matching algorithm to search for keyword search terms in the title, summary, or full text of an article from various publishers and websites [5]. Around the same time Google Scholar was announced to the public, Elsevier introduced Scopus, an indexing and abstraction service that includes its own citation-tracking tool. Scopus has reportedly indexed more journals than WoS has and included more international and open-access journals [5].

Altmetric (https://​www.​altmetric.​com) is powered by Digital Science, a Macmillan company that focuses on technology to aid scientific research. It collects data from three primary sources: social media (e.g., Twitter, Facebook, Google, Pinterest, and blogs); traditional media, both mainstream (e.g., The Guardian and New York Times) and science-specific (e.g., New Scientist and Scientific American); and online reference managers (e.g., Mendeley and CiteULike). It also calculates the score of an article on the basis of its wager on those sources. This is an algorithm-calculated quantitative measure of the article's quality and amount of attention [6].

In early 2018, Digital Science & Research Solutions launched Dimensions, a novel online academic platform designed to provide a distinct viewpoint on research outcomes. Grant awards, journal and book publications, mentions of social media, academic citations, clinical trials, and commercial patents are considered research outputs. The publication and citation contents at Dimensions are created and constantly updated by integrating data from multiple sources, including multiple clinical trial records, open-access articles, indexes covering many scientific journals, databases with content licenses, and open-access databases [7].

Numerous citation analyses and the most cited articles have become available in dentistry, including areas such as caries [8, 9], bulk-fill composites [10], endodontics [11‐17], implants [18‐20], pediatric dentistry [21, 22], periodontology [23, 24], oral medicine and radiology [25‐27], and orthodontics [28], and in topics such as dental traumatology [12, 29], tooth wear [30], minimally invasive dentistry [31], orofacial pain [32], and dental education [33]. Some citation analysis studies have included articles published in multiple dentistry journals [4, 34‐38] or in a single dentistry journal [13, 39].

Dentin adhesives appear to have made tremendous progress over the years since adhesives were first introduced in 1955 by Buonocore in a study on the bonding of resins to etched enamel surfaces and later after the introduction of resin bondings to adhere to etched dentin by Fusuyama et al. [40‐42]. Dental adhesive technology is constantly evolving with the rapid changes in commercial adhesives. These developments are the result of numerous laboratory and clinical studies, and the data obtained are highly important in showing the potential success of these materials and in guiding future research [43].

The basic mechanism of bonding to enamel and dentin involves the replacement of resin monomers with the minerals removed from the dental hard tissues, which cause porosity, and upon setting, micromechanical interlocking occurs in the formed porosities [44]. Adhesives can be classified as “etch and rinse” or “self-etching” depending on the underlying adhesion strategy, and the degree of substance exchange varies significantly among these adhesives [44]. Nevertheless, the success of both adhesion strategies has been reported in both laboratory and clinical research. However, it’s important to note that their effectiveness may depend on the specific product being used [45].

To date, no bibliometric analysis has been carried out to provide a more comprehensive perspective to evaluate research on various topics in the field of dentin adhesives, enabling us to anticipate future advancements and direct research efforts in this area. Thus, the purposes of this study were to gain insight into the scientific interests, research trends, and development within the field of dental adhesives by using WoS, Scopus, Google Scholar, Altmetric, and Dimensions.

To identify the most cited articles on dentin adhesives, our study was conducted in two stages, in which bibliometric and altmetric analysis data were collected. Institutional ethics committee approval was not necessary because the data used in this study were obtained from publications.

Initially, the WoS database (http://​www.​webofknowledge.​com) was used for the bibliometric analysis. On February 12, 2023, a search was conducted in the "Web of Science Core Collection (WoSCC)" using the search terms listed in Table S1, starting from the year 1945. The most commonly used free and Medical Subject Headings (MeSH) terms in the published literature on dentin adhesives were combined to create keywords. The field tags as “Topic” were selected, and the search resulted in 142,494 articles ranked according to the first option with the highest number of citations. Then, respectively, the search was restricted to articles written in the English language (n = 137,996), and ‘Science Citation Index Expanded (SCI-E)’ and ‘Emerging Science Citation Index (ESCI)’ index limitations were applied, resulting in 123,086 articles. The document types “article” and “review article” were selected (n = 115,845). After screening the articles, all studies were exported into the Excel program as a full record.

After ranking the articles according to their numbers of citations in the WoS database, two independent researchers (F.K. and M.D.) reviewed the titles and abstracts of the articles to identify the candidates for full-text review. Apart from the restrictions set, the eligibility criteria consisted mainly of studies that had data or topics that directly included dentin adhesives. The first 100 articles with the highest number of citations according to the criteria were identified independently by the two researchers (F.K. and M.D.). All results were cross-checked, and inconsistencies were resolved after reading the full texts of the articles and reviewing the relevant literature. The inter-examiner agreement was quantified using the kappa coefficient.

After the top 100 most cited articles were identified, the citation counts were manually retrieved for each article from the Scopus (https://​www.​scopus.​com/​), Google Scholar (https://​scholar.​google.​com), and Dimensions databases (https://​app.​dimensions.​ai) on the same date to provide a more comprehensive view, as the citation count of the same article may vary on different dates (date of access: March 3, 2023).

For the altmetric analysis, the Altmetric Attention Score (AAS; a metric that automatically calculates the weighted count of social media attention received by a research output) was used. The 100 most cited articles were accessed by manually scanning the Altmetric Explorer database (https://​www.​altmetric.​com) through the “Advanced Search” option using “publication title” or “DOI” simultaneously (date of access: March 3, 2023). A donut graph with different colors representing the amount of attention given to the different types of output was constructed with the AASs. Articles that were found in the database but were not cited in other articles and those that were added to the database either by institutional implementation or through a non-scoring source were displayed in the donut with a question mark. If the article was not mentioned at all in any article or if this output did have a score at one point but had been removed/reduced because of changes in the number of mentions, it was represented with “0” in the altmetric donut. At this point, there would be no difference in that both cases would indicate having no tracked attention or altmetric score assigned to the research output (help.altmetric.com).

The top 100 most cited articles are shown in Table 1 according to their numbers of citations as indicated in the WoSCC database, from highest to lowest, including results from all databases searched. As the numbers of citations were the same, our top 100 list consisted of 101 articles. After the final list was confirmed, the top 100 most cited articles were analyzed by the researchers, who recorded the number of citations, publication name (title), year of publication, journal name and impact factor, author(s) (name, number, and authorship position), country, institution, and type and field of study. When the article analysis results were discrepant between the two independent researchers, a consensus decision was reached through a discussion.

More recent articles were listed with priority for articles with the same numbers of citations. The list of journal names was arranged in order of their numbers of top-cited articles, and the Journal Impact Factor (JIF) 2021 from the Journal of Citation Reports (https://​jcr.​clarivate.​com) was used to rate journals with the same numbers of articles (Table S2). The institute of origin was based on the address of the first author's affiliation. If the first author worked at more than one institution that belonged to more than one country, each institution and country were counted. The type of study was classified as clinical, basic, review, systematic review, meta-analysis, or lecture based on the article type. To determine the area of study, the full text of each article was carefully examined by identifying concepts based on MeSH terms from PubMed.

The Visualization of Similarities (VOS) Viewer software program (version 1.6.15; Centre for Science and Technology Studies, Leiden University) was used to analyze the co-authorship network and journals. SPSS version 21 (IBM Corporation, USA) was used for the statistical analysis of the frequencies of the descriptive measures.

The top 100 most-cited articles are listed in Table 1 according to the number of citations. The most cited article, published in 2003 by Van Meerbeek et al. in Operative Dentistry, had 1288 citations and was a lecture on adhesion to enamel and dentin (Table 1). The least-cited article had 198 citations. The top 100 most cited articles had a total of 34,526 citations, and the mean number of citations per article was 342.

In our study, the citation count of the top 100 articles was between 1288 and 198 on WoS, between 1464 and 208 on Scopus, between 3118 and 276 on Google Scholar, and between 1100 and 184 on Dimensions. The total number of citations was highest on Google Scholar, followed by Scopus, WoS, and Dimensions. However, the number of citations for the same article differed between the databases. In addition to scientific articles, Google Scholar includes citations from books, theses, and other works, so the results from the database should be interpreted with caution. Currently, Scopus only counts citations from 1996 onwards, which is a major shortcoming for identifying the most cited journal articles, but expansion of the citation count to before 1996 has been planned for the near future [29]. On the other hand, the database indexes more international and open-access journals than the WoS [5]. In accordance with another study, the Dimensions database was assessed using a free application that does not provide entry to the website's functionalities, including grants, patents, clinical trials data, and analytical tools [7]. By contrast, citations were collected using the complete versions of WoS and Scopus [46]. Moreover, while the number of citations on WoS, Scopus, and Dimensions showed no correlation with the AASs, the number of citations on Dimensions strongly correlated with those on WoS and Scopus. Both Dimensions and Altmetric can provide a more comprehensive assessment of research effects [46]. In parallel with the main logic of our study, the “all databases” section of the ISI Web of Knowledge database was selected as the main database in other studies because it can count citations in scientific articles over a wide period from 1945 to the present [12, 22, 29]. In our study, the number of citations was lower than those in studies conducted in different dentistry areas such as endodontics (between 2115 and 246 citations) [17] and implant dentistry (between 2229 and 199 citations) [19], but higher than those in other studies on dentistry areas such as pediatric dentistry (between 182 and 42 citations) [22], oral medicine and radiology (between 624 and 86 citations) [26], and orthodontics (between 545 and 89 citations) [28]. In fact, the citation rates differed for each specialization depending on the number of researchers working in a specific field [15]. In addition, the wide variety of subdisciplines in specific fields may be another influencing factor in the citation rate.

Of the most cited articles in our study, 89.1% (90 articles) were published before 2010. Our findings are consistent with those of other studies [15, 17, 19, 28, 29]. Contrary to our findings, the most cited articles in some studies were published in the past decade [11, 12, 22]. The oldest articles have more time to be cited than recent articles, regardless of their scientific significance, hence the risk of exemption from the recent influential articles [25]. As supported by our findings and previous studies, it can be considered that an article needs a publication period of at least 6 to 15 years to receive sufficient citations and become a citation classic [19]. This may explain why none of the 100 most cited articles in our study were published in the last 5 years. According to Kuhn's philosophy, the scientific community has a tendency to stick to a paradigm[47]. In this context, this means that citations have a “snowball effect” because other authors are more inclined to cite articles on the basis of their numbers of earlier citations and not their content or quality [48]. On the other hand, a publication with more than 400 citations should be considered a classic, but in some areas where researchers have fewer, 100 citations may merit a study [15, 49]. The first 13 most cited articles in our study were cited more than 400 times, whereas the 100th article was cited 198 times. Therefore, in our study, the attribution rate was influenced not just by the snowball effect but also by the article's content or quality. Moreover, when the AASs were analyzed, the rate of mentioning articles published after 2010 on social media was 23.3% (8 articles), which is higher than the citation rates. One study found a high correlation between the citation count in Dimensions and those in WOS and Scopus but found no correlation between the citation counts in WOS, Scopus, Dimensions, and Altmetric [46]. Another study reported a weak but positive correlation between the AAS and the number of citation [50]. In Altmetric, behavior is completely different from the classic citation system, allowing recently published works to achieve more recognition and visibility quickly. Thus, Altmetric can highlight newly published research articles with higher prevalence rather than top-cited articles, which are usually at least 1 or 2 decades old [46].

In our study, 70% of the first 100 most cited articles (71 articles) were published in journals with an impact factor greater than 5 and a high impact factor for the field of dentistry. Except for one, the other 29 most cited articles were published in journals with impact factors higher than 2, of which 15 were in journals with impact factors greater than 3, which indicates a relatively high impact. This result was consistent with those of other studies [4, 24, 35]. It is well known that researchers choose high-impact journals for their article submissions and that journals with high impact factors attract high-quality articles [26]. However, no correlation was found between the journal impact factor and the number of articles that received the most citations [17, 26]. On the contrary, the number of citations and the relevant impact factor have been found to be closely correlated in a limited number of journals, especially in areas with high citation intensity [4, 24]. This can be attributable to the fact that articles with high citation rates tend to be published in journals with high impact factors [35]. In addition, more than a third of the articles have been published in specialty journals, including the subjects of our study, and this result may justify why fewer journals have attracted more attention [26]. Therefore, this conforms to Bradford's law, which explains why only a few journals in a subject area are most frequently cited and consequently most likely to be of interest to researchers in the discipline [22, 50, 51]. In line with our findings, similar results have been observed in other studies [17, 25, 26].

This study shows that 25 of the 100 most cited articles originated in Japan. The introduction of resin bonding to etched dentin by Fusuyama et al. [41], along with extensive research conducted in the following decade, and later, the definition of hybrid layer by Nakbayashi [52], had a significant influence on most of the highly cited articles, all of which had Japanese origins. In our study, 20 of the 100 most cited articles were affiliated with the Catholic University of Leuven in Belgium and were published between 1992 and 2012. This was followed by 10 articles from Tokyo Medical and Dental University, spanning the years 1979 to 1999, and 7 articles from the University of Hong Kong, Prince Philip Dental Hospital, covering the period between 1996 and 2005. These universities are particularly focused on the subspecialty of dental adhesion. Remarkably, although nearly one-fifth of the 100 most cited articles were produced by institutions in Japan, the most cited articles were from Belgium (Catholic University of Leuven), particularly considering that Japanese articles were among the earliest and most pioneering contributions to the field. Despite Belgium's modest population, researchers from this country have been comparatively prolific in operative dentistry-related publications during the study period [29, 53], aligning with our finding that researchers affiliated with this center had two or more highly referenced articles (Fig. 3). Also, the reasons for the high citation rates of Belgian articles could be attributed to factors such as international collaboration, research infrastructure, and visibility within the global scientific community. In addition, in line with the results of other studies [15, 17], approximately one-third of the most cited articles (28 articles) in our study were produced by independent institutions. It's essential to consider the extent of international collaboration in dentin adhesive research. Articles resulting from collaborative efforts between researchers from various countries might have received more citations due to their diverse perspectives and broad relevance.

In our study, most of the top-cited articles were in the field of basic research (69 articles), followed by reviews (25 articles) and systematic review and/or meta-analysis (5 articles). Only two of the top cited articles reported clinical experiences. Consistent with our findings, other studies have reported that most of the top-cited articles were in the field of basic science [15, 17, 39]. On the other hand, other studies found that most top-cited articles reported clinical experiences [4, 19, 25, 28]. However, one study found that the most top-cited articles were reviews [13]. These differences may be due to differences in subspecialties in the field of dentistry. Most of the top-cited articles in our study were in the field of basic science. In the early stages of dentin adhesive development, the papers that formed the foundation of the field generally focused on basic research, investigating the principles of adhesion, the composition of adhesives, and their interactions with dentin. Some of the pioneering articles from this period, while groundbreaking, may have been more cited because of their age. Basic research in dentin adhesives, a subspecialty of operative dentistry, is crucial to investigating the efficacy of new materials or modified techniques [15]. In vitro studies play an important role in enhancing methods and providing early data on which later research with greater evidence can be based [11]. In our study, most topics in basic science were on the ultramorphological structures of dentin and adhesive interfaces (39 articles), followed by bond strength to dentin (34 articles) and hybrid layers (25 articles). The integration of knowledge from new basic science research into the subspecialty practices of operative dentistry provides the opportunity to address major clinical issues [15]. However, the fact that our study detected very few systematic reviews, meta-analyses, and RCTSs among the most cited papers suggests that more such studies on dentin adhesives are needed.

As with other citation analyses, this study has some limitations. By including many databases, the differences in the number of citations between databases were tried to be eliminated. The current study excluded several articles, as indicated by the title, due to its focus on including only the top 100 most-cited articles. In addition, articles written in languages other than English and books or conference proceedings as document type were not included in the study.

Most top-cited articles (89.1%) were published before 2010. In our study, the most frequently cited articles were concentrated in a few journals. As first author, Van Meerbeek B. has the highest number of articles with nine articles and a total number of 4650 citations. The highest top-cited 100 articles originated from Japan. The most top-cited articles originated from the Catholic University of Leuven in Belgium. Basic science research had the highest number of articles, followed by reviews. The primary foci of basic research were the ultramorphological structures of dentin and adhesive interfaces. The major topic of the reviews was hybrid layers. Only 2 RCTs and a few systematic reviews and meta-analyses were published. Thus, in the future, studies with high levels of evidence, such as systematic reviews, meta-analyses, and RCTs, are required.