Bibliometric Analysis of the Scientific Literature on Rheumatoid Arthritis-Associated Interstitial Lung Disease

Background In recent years, the number of studies on rheumatoid arthritis-related interstitial lung disease (RA-ILD) has been increasing, which has led to many publications on this topic. Our purpose is to identify research trends in RA-ILD and analyze the most-cited RA-ILD-related high-quality scientific publications. Methods All publications on RA-ILD in the Core Collection database of Web of Science were searched. The publication year, country, institution, total citations, and journal were extracted and analyzed. We used VOSviewer software or an online bibliometric analysis platform for cooccurrence analysis of the keywords, institutions, and countries involved. The 100 most frequently cited RA-ILD publications were analyzed. Results In total, 596 publications related to RA-ILD were obtained. Over time, the frequency of RA-ILD publications has increased. Globally, the United States provides the most publications on RA-ILD (n = 195). The institution with the highest publication output was the Mayo Clinic (n = 43). The journal “Annals of the Rheumatic Diseases” published most with 93 articles and received 338 citations. A clinical description was the most common research topic in RA-ILD-related publications. Conclusions In recent years, there has been an increasing number of studies on RA-ILD, and related publications have increased rapidly. This study is the first bibliometric study of RA-ILD-related publications. It can be used as a guide for clinicians and can help researchers choose research directions of interest in this field.


Introduction
Rheumatoid arthritis (RA) is a systemic autoimmune disease characterized by articular and extra-articular manifestations affecting approximately 1-2% of the general population [1]. Interstitial lung disease (ILD) is an extra-articular manifestation of RA, which occurs frequently in up to 80% of patients with RA. This may be the result of chronic immune activation and inflammation in RA, or the pulmonary toxicity caused by immunomodulatory drugs used to treat RA [2][3][4][5][6]. The prevalence of RA-ILD ranges from 1% to 58%, depending on the diagnostic means used and the severity in the RA population studied [6,7]. Currently, ILD is the second leading cause of death in patients with RA after car-diovascular disease [8]. Research on RA-ILD has increased to include its natural history, pathogenesis, radiological evaluation, clinical manifestations, and treatment [9][10][11][12]. However, the trend of RA-ILD research is unclear, and the most influential research in this field has not been systematically determined. Therefore, our purpose was to provide a bibliometric study of publications on RA-ILD.
Bibliometric analysis is a convenient and reliable statistical method that can quantitatively and qualitatively evaluate research trends in the research field. This analysis has long been used in the field of medical research and has been widely accepted by scientific researchers [13][14][15]. To the best of our knowledge, no bibliometric studies on RA-ILD have been published to date. Therefore, in this study, we use bibliometric statistical methods to identify the most influential publications and analyze the research status and trends in the RA-ILD research field.

Materials and Methods
2.1. Datasource. All the data of this study were obtained from articles retrieved from the core collection database of Web of Science on July 1, 2021.
2.2. Search Strategy. The retrieval steps and strategies were as follows: Title = rheumatoid arthritis AND Title = (interstitial lung disease OR interstitial pneumonia) AND Language = English AND Document type = (review OR article) AND Time span =1980 to 2021.

Statistical Tools.
VOSviewer, an online bibliometric analysis platform (https://bibliometric.com), and Excel software were used to extract and analyze all data. VOSviewer is a software that is usually used to visually analyze the collaborative network between countries, institutions, and authors and cocitation of keyword clusters to analyze research trends and hotspots. The role of the online bibliometric analysis platform is similar to that of VOSviewer. Excel software was used to extract and analyze various details of the publication, including author, title, journal, year of publication, institution, country, journal impact factors, and number of total citations.

Data
Extraction. According to the retrieval steps and strategies, the two authors independently fetched the article information and discussed the differences until they reached a consensus. Data were obtained from the core collection database of Web of Science, and the publication information was extracted and analyzed using Excel, online bibliometric analysis, and VOSviewer software.

Publication Analysis.
A total of 596 RA-ILD research articles were found in the core collection database of the Web of Science. The number of articles increased from 1981 to 2021 (Figure 1(a)). Quantitative analysis shows that in the past 10 years, global research on RA-ILD has increased rapidly, from four articles from 1981 to 1985 to 326 articles from 2016 to 2020. This result shows that RA-ILD has attracted increasing attention, and the research process of RA-ILD continues to accelerate.
The online bibliometric analysis platform was used to analyze cooperative relations between countries. The visual analysis shows that the USA has always been the center of RA-ILD research in the world, and Japan, France, China, and South Korea have been found to be potential research powers. (Figure 1(d)).

Institutions Analysis.
All the publications involve 1000 institutions. The results show that many institutions in the United States actively participate in RA-ILD research. The 10 most productive institutions internationally were Mayo Clinic (n = 43), National Jewish Health (n = 41), Brigham and Women's Hospital (n = 31), Colorado State University (n = 31), University of California, San Francisco (n = 19), University of Ulsan (n = 19), University of Modena and Reggio Emilia (n = 17), Harvard Medical School (n = 15), Queen Elizabeth's Hospital (n = 15), and University of Miami (n = 15), respectively (Figure 2(a)). According to the citation report, Mayo Clinic's articles were cited the most, namely, 1408 times, followed by the National Jewish Health, which was cited 858 times, and the University of California, San Francisco, 851 ( Figure 2(b)).
VOSviewer software was used to analyze the extent of cooperative relations between institutions. The institution with the most links, i.e., the highest link strength was recorded by the National Jewish Health Organization (n = 131), followed by the University of Colorado (n = 123 ), Mayo Clinic (n = 119), and Brigham and Women's Hospital (n = 117). In the VOSviewer software, the width of the line reflects the close relationship of interinstitution cooperation. The National Jewish Health had close collaborations with the University of Colorado and Mayo Clinic. Mayo Clinic had large collaborations with National Jewish Health, University of Colorado, Brigham and Women's Hospital, and Harvard Medical School (Figure 2(c)).

Journals Analysis.
All 596 articles in this study were published in 123 journals. Among these, the journals which had published at least 20 articles on the topic accounted for 73.8% of the total ( Table 1). The five journals with the most articles on the topic were Annals of the Rheumatic Diseases, Arthritis & Rheumatology, American Journal of Respiratory and Critical Care Medicine, Rheumatology, and Arthritis and Rheumatism. Moreover, articles in the Annals of the Rheumatic Diseases have been cited the most. More than 20 journals published on RA-ILD, and the average impact factor was 11.8, indicating a high level of reliability of the included studies.

Research Status and Analysis.
VOSviewer software was used to analyze the cooccurrence analysis of keywords in the RA-ILD research articles. When the minimum number of keywords appearing in the publication was set to five, 72 keywords were selected and divided into four clusters: "Clinical-Features," "Pathological-Features," "Treatment," and "Prevalence and mortality." In the "Clinical-Features" cluster, the most common keywords were "pneumonia," "idiopathic pulmonary-fibrosis," and "prognosis." In the "Pathological-Features" cluster, the most frequent keywords were "rheumatoid arthritis," "interstitial lung disease," and "fibrosis." In the "treatment" cluster, the most frequent keywords were "classification," "criteria," and "safety." In the "prevalence and mortality" cluster, the most     5 BioMed Research International frequent keywords were "prevalence," "mortality," and "risk" (Figure 3(a)).
To better understand the dynamic process of the RA-ILD research trends, we evaluated the evolution of the keywords (Figure 3(b)). We assigned colors based on the year the keyword appears in the article. For example, the yellow keyword appears later than the purple keyword. In the early stages, "idiopathic pulmonary fibrosis," "alveolitis," and "systemic sclerosis" were the main topics. Trends in recent years show that the terms "management," "predictors," "inflammation," and "progress" are becoming more and more popular.
3.6. The 100 Most-Cited Publications. The 100 most-cited publications on RA-ILD were published between 1984 and 2020 ( Table 2). The analysis indicated that 2001-2005 was the period when most of these studies were published, with 41 publications, followed by 2016-2020, with 29 publications (Figure 4(a)).
Of these 100 articles, the Mayo Clinic Medicine and National Jewish Health each generated seven publications, resulting in their being the most represented institutions on this topic, followed by the University of California in San Francisco (n = 5) and Queen Elizabeth Hospital (n = 4) (Figure 4(c)).
Overall, there were 52 different journals which published the 100 articles. "Rheumatology" was the most productive journal, with 8 articles and 632 citations, followed by "Arthritis and Rheumatism," with five articles and 529 citations (Table 3).
When considering the individual authors' academic contributions, Jay H Ryu, provided 11 publications, followed by Joyce C Lee and Eric L Matteson, each with 8 publications ( Table 4).

Discussion
ILD is one of the most common complications of RA and poses a great challenge to clinicians and researchers [16]. The prevalence of RA-ILD ranged from 1% to 58% in the different studies, which was related to the diagnostic techniques used and the study population that was included [17][18][19]. According to the literature, there are many risk factors for RA-ILD, including male sex, smoking, older age, high disease activity of RA, characteristics of extra-articular diseases (subcutaneous nodules), and seropositive RA autoantibodies (rheumatoid factor and anticitrulline protein antibody) [2,[20][21][22][23]. The most common presenting symptoms include exertional dyspnea, tachypnea, and bibasilar inspiratory crackles. In the advanced stages of the disease, symptoms of cyanosis, edema, and pulmonary hypertension may occur, leading to a reduced quality of life [24].      In addition to its impact on the quality of life, RA-ILD places a huge burden on the medical system, with an average total medical cost of more than $170,000 per patient over five years [8]. Our statistical and quantitative analysis shows a gradual increase in RA-ILD research results from 2011 to 2020, with more researchers and physicians focusing on this area of research. Despite the wide range of RA-ILD research, an analysis of the current status and trends in RA-ILD research is not clear. In this study, we analyzed, discussed, and described the current status, priorities, and trends of RA-ILD research. At the same time, our study will help RA-ILD researchers gain a more comprehensive understanding of the current state of RA-ILD research and thus guide the direction of future research.  Italy (5) Spain (2) Argentina (1) Finland (1) Japan (20) Korea (5) France (2) Australia (1) Ireland (1) UK (11) Canada (3) Gemany (2) Bangladesh (1) China (7) Austria (3) Mexico (1) Denmark (1  The number of articles related to RA-ILD has increased rapidly over the last 10 years. Globally, the USA ranks first in terms of the number of publications and citations, indicating that the USA has led to research on RA-ILD in the past few years. In terms of institutional contributions, the institution with the highest publication output is the Mayo Clinic (USA) and ranked first in the total citations. This reflects the institution's leadership in the field of RA-ILD research. Analysis of cooperation between countries and institutions shows that regional clusters are usually geographically specific. As a leader in the world economy and science, the USA has the most frequent cooperation with Japan, France, China, and South Korea. Researchers working on RA-ILD should pay close attention to them and collaborate with these institutions and countries. Annals of the Rheumatic Diseases, Arthritis & Rheumatology, American Journal of Respiratory and Critical Care Medicine, Rheumatology, and Arthritis and Rheumatism are the five most prolific journals in RA-ILD.

Research Foci.
Keyword analysis results showed that RA-ILD, rheumatoid arthritis, interstitial lung disease, and pneumonia were keyword cluster centers. In the early stages, "idiopathic pulmonary fibrosis," "alveolitis," and "systemic sclerosis" were the main topics. In recent years, more common keywords have included "management," "predictors," "inflammation," and "progression.  "Usual interstitial pneumonia in rheumatoid arthritisassociated interstitial lung disease" by Kim et al. in 2010 was the second most-cited article with 283 citations [16]. The authors determined that the pattern of common interstitial pneumonia (IP) found on high-resolution computed tomography (HRCT) is important for the prognosis of RA-ILD. Eighty-two patients with RA-ILD were identified retrospectively. "We determined the relationship between survival and the pattern of IP common on HRCT and compared it with patients diagnosed radiologically with idiopathic pulmonary fibrosis. Twenty (24%) of the 82 patients with RA-ILD had definite common IP. Survival in patients with RA-ILD was lower than that in patients without this pattern, similar to the survival of patients with idiopathic pulmonary fibrosis. In addition, a clear pattern of common IP on HRCT was associated with poor survival. Analysis of feature-specific HRCTs showed that traction bronchiectasis and cellular fibrosis were associated with poor survival. Women and a higher baseline carbon monoxide lung diffusing capacity were associated with better survival." "Histopathological and clinical features of interstitial lung disease associated with rheumatoid arthritis" by Lee et al. was the third most-cited article with 245 citations [25]. The authors studied the histopathological patterns and clinical characteristics of patients with RA-ILD according to the American Thoracic Society/European Respiratory Society consensus classification of idiopathic IP. "Eighteen patients with RA who underwent surgical lung biopsy for suspected ILD were included in this study. This study revealed diverse histopathological findings. Ten patients had a common interstitial pneumonia (UIP) pattern, six patients had a nonspecific interstitial pneumonia (NSIP) pattern, and two patients had inflammatory airway disease with tissue-type pneumonia.
Thus, the UIP pattern appears to be more common than the NSIP pattern in our study population."

Limitations.
Our study had several limitations. First, we extracted information related to RA-ILD from the Core Collection database of the Web of Science. It is possible that some influential publications were not included in this database and were therefore excluded from our study. Second, the date of our retrieval and extraction of data was July 1, 2021. Part of the data correspond to dynamic changes, but the trend of changes will not be extensive. Third, we retained only English articles in our search strategy.