Global trends of Pseudomonas aeruginosa biofilm research in the past two decades: A bibliometric study

Abstract Pseudomonas aeruginosa biofilm formation is a primary cause of chronic infections. This has been a highly active area of research over the past two decades due to causing high mortality risks in immunocompromised patients. This study evaluates global trends in the dynamic and rapidly evolving field of P. aeruginosa biofilm research through bibliometric and visualized analyses. Publications from 1994 to 2018 on P. aeruginosa biofilm research were retrieved from Web of Science, Scopus, and PubMed, and their bibliometric data were systematically studied. The VOSviewer software was used to conduct global analyses of bibliographic coupling, coauthorship, cocitation, and co‐occurrence. A total of 9,527 publications were included in this study. The overall number of publications and research interest in the field displayed a strongly rising trend. The USA made the greatest contributions to the field, with the highest h‐index and number of citations compared with other countries, while Denmark had the highest average citation per publication. The Journal of Bacteriology had the highest number of publications in the field, while the University of Copenhagen was the institution with the highest contribution influence. Co‐occurrence network maps revealed that the most prominent topics in P. aeruginosa biofilm research were mechanistic studies, in vitro/in vivo studies, and biofilm formation studies. Pseudomonas aeruginosa biofilms constitute a dynamic research area in microbiology with increasing global research interest. Future studies will likely focus on investigating the mechanisms of biofilm formation to solve infection‐associated clinical problems.


| INTRODUC TI ON
Pseudomonas aeruginosa is a gram-negative bacterium commonly found in soil and water (Rahme et al., 1995). It is an opportunistic pathogen capable of causing various infections, particularly in immunocompromised or critically ill patients, including bacteremia, sepsis, pneumonia, and skin and wound infections (Kipnis, Sawa, & Wiener-Kronish, 2006). Infections caused by P. aeruginosa can be very serious and life-threatening, particularly with the emergence of drug-resistant strains (Zowawi et al., 2015). Antimicrobial resistance is one of the biggest threats to human and animal health today. This has caused widespread concern since P. aeruginosa is one of the most prevalent nosocomial pathogens, responsible for 57% of total hospital-acquired infections (Appanna, Sarabhai, Sharma, & Capalash, 2013).
Biofilm formation is a primary characteristic of P. aeruginosa chronic infections (Bjarnsholt, 2013). A biofilm is a community of bacteria attached to a substratum or surface (Chen, Yu, & Sun, 2013), typically consisting of densely packed, multi-species populations of cells. The main feature of biofilms is the presence of highly hydrated extracellular polymeric substance (EPS), including polysaccharides, proteins, and extracellular DNA (eDNA) (Billings, Birjiniuk, Samad, Doyle, & Ribbeck, 2015). Pseudomonas aeruginosa biofilms are generated through a process whereby the bacterial cells are surrounded to form an aggregated structure, which exhibits increased resistance to antibiotics and other anti-infection agents (Head & Yu, 2003;Mah et al., 2003;Rybtke, Hultqvist, Givskov, & Tolker-Nielsen, 2015;Whiteley et al., 2001). Therefore, infections caused by biofilm-forming P. aeruginosa, such as in cystic fibrosis of the lung, are almost impossible to eradicate (Guo et al., 2014), and additional challenges are encountered when treating infections caused by multi-drug-resistant strains. These complications lead to increased patient morbidity and mortality, higher costs of treatment, and greater rates and time of hospitalization (Costerton, Stewart, & Greenberg, 1999;Lister, Wolter, & Hanson, 2009). Diverse strategies are being pursued to develop novel agents that can kill (new antibiotics) or disarm (antivirulence) the pathogen (Bassetti, Vena, Croxatto, Righi, & Guery, 2018).
In the context of dynamic changes in the field of P. aeruginosa biofilm research, it becomes particularly interesting to understand the most important contributions to the field and to predict future research trends. In this study, a bibliometric analysis was conducted on publications in the field of P. aeruginosa biofilm research to reflect the global state of the field over the past two decades.
This type of analysis can qualitatively and quantitatively evaluate the trends within a particular research community over a defined time frame. It also provides a neat avenue for predicting the main direction of future research in the field, drawing on the relative contributions of different journals, research groups, institutes, and countries (Bassetti et al., 2018). In recent years, bibliometric analysis has been successfully applied in a range of research areas to increase the transparency of published studies (Jia et al., 2015; Wang, Xing, Zhu, Dong, & Zhao, 2019), as well as to assist the development of clinical policies and guidelines (Tacconelli et al., 2018).
Bibliometric analysis has not been previously applied to the field of P. aeruginosa biofilm research to investigate the quality and quantity of published studies. The importance of this field is reflected by the recent listing of carbapenem-resistant P. aeruginosa by the World Health Organization (WHO) as one of three bacterial species for which there is a critical need for the development of new antibiotics to treat infections (Tacconelli et al., 2018). Our study, therefore, provides a timely analysis of the overall status of P. aeruginosa biofilm research, revealing trends that would be useful for gaining a broad understanding of global developments in the field and future directions.

| Data source
The Web of Science (WoS) was used for retrieving publication data due to its ability to provide a comprehensive citation search across multiple databases containing cross-disciplinary research studies.
Bibliometric analysis was applied to the publication data. Scopus and PubMed databases were also searched to complement the findings.

| Search strategy
WoS was used to retrieve the publication data, with the date range

| Data collection
The full records of each publication retrieved in the search, including author names, nationalities and affiliations, article title, year of publication, name of publishing journal, keywords, and abstract, were downloaded as a.txt file from the WoS, Scopus, and PubMed databases and subsequently imported into Excel 2016. Two authors (YZ and BW) independently extracted relevant data from these publications. Any disagreement was resolved by discussion or seeking assistance from an external expert to reach consensus. The data were then analyzed independently by the two authors using GraphPad Prism 7.

| Bibliometric analysis
The basic characteristics of retrieved publications, including the total number and number by country, total and average citation frequency by country, and H-index, were described using the intrinsic functions of WoS. The H-index was used to measure the impact of scientific research arising from different countries within the chosen field of study. The H-index assigns a value based on an H number of published papers, each of which has been cited by other published papers at least H times. The H-index for a particular country is, therefore, a reflection of research impact, incorporating both the total number of publications and the average number of citations per publication.

| Visualized analysis
VOSviewer (Leiden University, Leiden, The Netherlands) was used for constructing and visualizing the bibliometric networks in this study. Networks were constructed based on bibliographic coupling, coauthorship, and cocitation analysis, using the information on authors, journals, institutions, and countries associated with the publications. Co-occurrence networks were also constructed to allow visualization of important scientific terms arising from the included publications.  Table 1.

| Global publications by country
The total number of articles in the field from Scopus and PubMed was 7,280 and 5,206, respectively, all of which were covered by the WoS search. The WoS search results were used for all subsequent bibliographic coupling, coauthorship, cocitation, and co-occurrence analyses using VOSviewer.

| Global publications by year
The highest number of articles in the field was published in 2017 (1,110; 11.65%). From 1994 to 2018, global publications in the field exhibited a strong trend for exponential growth (Figure 1c).

| Prediction of global publication trends
The logistic regression model was used to create the time curve of the number of publications from which future trends could be predicted ( Figure 1d). This curve suggests that the field is currently in a phase of rapid growth in global publication outputs, but the rate of growth is expected to gradually decline within the next decade.

| Bibliographic coupling analysis
Bibliographic coupling measures links between publications that cite the same document. Two publications are "coupled" if they cite the same reference in their bibliographies, indicating the probability that these two publications share a common subject matter. The strength of a link between two publications indicates the number of cited references these two publications have in common. This can be scaled up to analyze the link strength of journals, institutions, and countries which have published papers within a particular subject matter, reflecting the degree of bibliographic coupling.

| Journals
VOSviewer was used to analyze the journal names of all publications included in this study.

| Institutions
VOSviewer was used to analyze the institutions of all publications included in this study. Total link strength was shown for 820 in-

| Countries
VOSviewer was used to analyze the countries of origin for all pub-

| Authors
VOSviewer was used to analyze a total of 399 authors with more than 10 publications in the field, identified from all of the publications in-

| Institutions
VOSviewer was used to analyze a total of 960 institutions with more than 10 publications in the field, identified from the publica-

| Countries
VOSviewer was used to analyze a total of 57 countries with more than 10 publications in the field, identified from the publications included in this study (Figure 4c). The five countries with the highest total link strength were USA (total link strength = 1,247 times), England (530), Germany (401), Australia (354), and Canada (352).

| Cocitation analysis
Cocitation analysis measures the link between two items that are both cited by the same document. Within a cocitation map of publications, the total link strength of a publication indicates the total number of cocitations the publication has with other publications included in the map. This can be scaled up to reflect the total cocitation link strength of authors and journals.

| Authors
VOSviewer was used to analyze a total of 4,002 authors that were

| References
VOSviewer was used to analyze a total of 3,103 references that were cocited in more than 20 publications, identified from all of the publications included in this study (Figure 5b). The five references with the highest total link strength were Costerton JW, 1996

| Journals
VOSviewer was used to analyze the journal names of the references for all publications included in this study, for which the journals

| Co-occurrence analysis
Co-occurrence analysis measures the co-occurrence links between terms. The co-occurrence link strength indicates the number of publications in which two terms occur together. Co-occurrence analyses have been performed in scientific and other disciplines to monitor past developments and have proven useful in identifying future research directions and popular topics within a specific discipline. In TA B L E 1 Institutions with the highest number of publications on Pseudomonas aeruginosa biofilms research within each of the top 10 contributing countries this study, keywords were identified as words that have been used more than 5 times in titles and abstracts among all included publications, which were analyzed using VOSviewer. There were 2,825 keywords identified that were classified broadly into three clusters: "mechanistic studies," "in vitro/in vivo studies," and "biofilm formation studies" (Figure 6a). These clusters reflected the most prominent topics of research on P. aeruginosa biofilms to date.
The identified keywords were then defined by specific colors using VOSviewer based on the average time they appeared in all included publications ( Figure 6b). Keywords that on average appeared during earlier time periods were denoted by colors toward the blue end of the spectrum, while keywords that on average appeared later were denoted by colors toward the red end. For the publications included in this study, most publications prior to 2014 appeared to focus on "mechanistic studies" and "in vitro/in vivo studies," while "biofilm formation studies" appeared to be a topic attracting more attention in recent years and may become an important topic of focus in the near future. The same identified keywords were also mapped by frequency of appearance (Figure 6c).

| Trends in the research field
In this study, we used a combination of bibliometric and visualized analyses to generate a representation of the current state of the field on P. aeruginosa biofilms research. We have analyzed the relative contributions of individual authors, journals, institutions, and countries to this well-established but still rapidly progressing research field, as well as existing and evolving topic areas that are likely to attract continued research interest in the coming years.
Since the field first picked up in 1994, there has been an exponential increase in publication output in the field leading up to 2018, with particularly dramatic increases seen over the past 3-4 years.
This is due to the pronounced ability of P. aeruginosa to induce biofilm-related disease that is causing significant and ongoing concern in the clinical setting, which has spurred increasing research effort in developing strategies that target their control and eradication  (Qu et al., 2016). In our study, significant bibliographic coupling was found for publications arising from 820 institutions over 57 countries worldwide, among the almost ten thousand publications in the field that have appeared since 1994. These numbers reflect the strength of the field and the overwhelming knowledge that has been generated over the past 20 years. From the trends observed in this study, it is likely that P. aeruginosa biofilms will continue to be an epicenter of intense research interest in the future, possibly with more studies focusing on biofilm formation as indicated by our co-occurrence analysis.

| Countries
The USA is the world leader in the field of P. aeruginosa biofilm research, which had the highest total citation frequency and H-index for its publications, as well as the top rank for bibliographic coupling and coauthorship analyses conducted by country. These trends suggest that publications from the USA collectively have the highest quality and academic impact in this field and also encompass the strongest collaborations worldwide. On the other hand, a number of European countries such as Denmark, England, Germany, and Switzerland have seen rapid growth in publication number, quality, and impact in this field over the past decade, as well as increasing collaboration with other countries.
The number of publications on P. aeruginosa biofilm research arising from China has increased rapidly over the past decade, as reflected by its total number of publications ranking second among countries worldwide. However, the total citation frequency, average citation frequency, and H-index of these publications from China only ranked 8th, 19th, and 9th, respectively. This large discrepancy between the number of Chinese publications and their perceived quality and academic impact may be due to the inherent differences in the Chinese academic evaluation system compared with most Western countries, which focuses primarily on publication quantity rather than quality (Zhai, Wang, & Li, 2016). With

| Journals
The link strengths of journals determined through bibliographic coupling and cocitation analyses in our study reflected the degree of relatedness of specific journals to the field of P. aeruginosa biofilm research and also gave an indirect indication of journal rankings in this field.
In addition, cocitation analysis revealed that the landmark studies in Information on topic areas in the field organized by publication date gives further interesting insights into future developments ( Figure 6b). Several terms colored toward the red end of the spectrum, indicating more recent publication dates, belong under the "biofilm formation study" cluster, suggesting a shift in research direction to related topics in the coming years. More interestingly, some small areas of emerging research in the field also colored within the red spectrum due to their recent appearance in the literature, overlapped with the "remote" terms identified using the earlier co-occurrence map grouped by topic area. These may represent rapidly evolving topics in the field of P. aeruginosa biofilm research where substantial research activity has already begun and will see significant developments in the near future, such as coatings, wounds, and multi-drug resistance.

| Strengths and limitations
We have conducted extensive bibliometric and visualized analyses to reflect the current state of P. aeruginosa biofilm research through analyses of citation frequency and quality, bibliographic coupling, coauthorship, cocitation, and co-occurrence. This gives a large volume of information that reveals important trends through many different dimensions, which to the best of our knowledge is the first study of its kind in the field. Nevertheless, there were some limitations to the accuracy of our analyses. First, only English language studies from the

| CON CLUS IONS
The present study shows the global trends in P. aeruginosa biofilm research based on bibliometric and visualized analyses of publication data. The USA and Denmark made the greatest contributions to this field, while the Journal of Bacteriology was the journal that produced the greatest number of relevant articles.
The research effort in the field and publication outputs are likely to continue with an upward trend, particularly with heightened interest in recent years on expansion into multidisciplinary research areas. Continued research in the field of P. aeruginosa biofilms, particularly targeting multi-drug-resistant strains, will hopefully provide new solutions in the near future to address the longstanding clinical concerns they have caused.

ACK N OWLED G M ENTS
This study was supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
None declared.

F I G U R E 6
Co-occurrence analysis of global research on Pseudomonas aeruginosa biofilms. (a) Mapping of keywords in the research area. The size of the points represents the frequency of appearance, and the keywords are divided into three clusters: in vitro/in vivo studies (red), biofilm formation studies (green), and mechanistic studies (blue). (b) Distribution of keywords according to the chronological order of appearance. Keywords in blue appeared earlier than those in yellow, and keywords in red appeared the latest. (c) Distribution of keywords according to the mean frequency of appearance. Keywords in red occurred with the highest frequency, followed by yellow, green, and cyan