The top 100 most cited articles on total hip arthroplasty: a bibliometric analysis

Background Over the past few decades, more and more articles about total hip arthroplasty have been published. We noticed, however, little is known about the characteristics and qualities of these studies. Methods The databases of Web of Science Core Collection, BIOSIS Citation Index, MEDLINE, etc. were utilized for the identification of articles published from 1990 to May 2019. Total hip arthroplasty–related articles were identified, and the 100 most cited articles were selected for subsequent analysis of citation count, citation density (citations/article age), authorship, theme, geographic distribution, time-related flux, level of evidence, and network analysis. Results The selected 100 articles were published mainly in the 1990s (46%) and 2000s (47%) with almost equal amount. Citations per article ranged from 994 to 191. Leading countries were the USA followed by Canada, England, and Sweden, all located in North America and Western Europe. The most highlighted study themes were postoperative thrombosis and surgical methods and materials. The most common level of evidence was level III (35%). The network analysis connoted that radiography, acetabulum, reoperation, and bone cement had a high degree of centrality in the 1990s, while cement had a high degree of centrality in the 2000s and 2010s. Conclusions The time, area, and theme distribution of the top 100 most cited articles in the total hip arthroplasty have been thoroughly analyzed. It is noticeable that postoperative thromboembolism currently plays a major role in the field of total hip arthroplasty researches. However, most of them focus on the effectiveness of different treatments and drugs; little is known about its underlying mechanisms and influencing factors.


Introduction
Total hip arthroplasty has evolved greatly in the last century. Since the early attempts to replace the hip joint by Sir John Charnley during the late nineteenth century, there have been many advances over the years. Over the years, much research has been done to explore and advance this therapeutic method. It has been demonstrated that total hip arthroplasty is an effective therapy for several hip diseases such as necrosis of the femoral head [1-3], hip osteoarthritis [4][5][6], and femoral neck fracture [7][8][9]. Amid hundreds of research topics in total hip arthroplasty, postoperative outcomes and surgical methods and materials remain the major focus.
There is a unique tool, called bibliometric analysis, for analyzing the qualities and characteristics of published articles. It was first published in the JAMA in 1987, and it has been widely used across diverse fields to evaluate and estimate the importance of published articles or trends in the research spotlight [10][11][12].
The aim of this study was to investigate the 100 most cited publications in the field of total hip arthroplasty, highlighting intellectual milestones in the field, and analyzing the qualities and characteristics of the most cited original papers over the past 30 years.
Material and methods related to total hip arthroplasty. The search was performed by two independent researchers at the same time to enhance the search sensitivity. The search terms used were the following: "total hip arthroplast*" OR "total hip prosthesis implantation*" OR "total hip replacement*".
The search was conducted in May 2019 and yielded a total of 24,146 results. Then, filtering the search results by "journal articles," the results were 22,832. Only original articles were included. Thus, review articles, systematic reviews, meta-analyses, and guidelines were excluded while registry data were accepted. All articles cited less than 120 times were excluded to reduce the number of articles necessitating subsequent screening. This resulted in 628 articles included for analysis. Two independent investigators reviewed the title and the abstract of all included articles. Articles that met the following criteria were included: (1) basic study, animal study, and clinical trials related to any aspect of total hip arthroplasty; (2) the clinical therapeutic, prognostic, diagnostic, epidemiological studies of total hip arthroplasty; (3) the registry data in relevant institutions; (4) articles investigating materials or properties related to total hip arthroplasty. Disagreement between the two reviewers was discussed to reach an agreement. After the title and abstract review, there were 262 articles remaining. These articles were ranked in descending order of citations and the first 100 most cited articles were included in this analysis (Fig. 1).

Data extraction
All articles were reviewed by two independent welltrained investigators. The following information was listed for all articles: the journal name, publication date, first author, year of publication, geographic origin, total number of citations of the article, overall citation rate (total citations/article age), research theme, and level of evidence (methodology has been described elsewhere [13]).

Statistical analysis
The Shapiro-Wilk test was used to test the distribution of individual variables for normality. Normally distributed data are presented as mean ± standard deviation. Comparison between means was performed using oneway analysis of variance (ANOVA), and post-hoc testing was undertaken when necessary. Time-dependent trends were tested using the Mann-Kendall trend test. Correlation between variables was performed using the Spearman rank or Pearson tests. A P < 0.05 was considered to     be statistically significant. Analysis was performed using IBM SPSS Statistics, Version 20.0. The Ucinet for windows, version 6.212 was used to perform the degree of centrality analysis [14].

Results
We retrieved the 100 most cited articles on total hip arthroplasty and listed them in (Table 1). The number of citations ranges from 191 to 994, and a majority of articles were published in the 1990s (46%) and 2000s (47%). However, articles published since 2010 only accounted for 7% (Fig. 2). The year with the highest number of journals was 2003 (n = 8). The mean number of citations was 312 overall, 315 in the 1990s, 316 in the 2000s, and 269 in the 2010s. The Mann-Kendall trend test showed no time-dependent trend in the publication time of articles (P = 0.4162) but an increasing trend between the citation density and the time (P = 5.854E−14) (Fig. 3). The Spearman rank revealed a positive correlation between time and citation density (r = 0.695, P < 0.01). The Shapiro-Wilk test and the Kolmogorov-Smirnov test both indicated an abnormal distribution of the citation data. These articles were distributed in 15 countries, led by the USA (n = 37) followed by Sweden (n = 15), England (n = 14), Canada (n = 11), and so on. The distribution is demonstrated on the world map (Fig. 4). Almost all of the articles came from two regions: North America and Western Europe. Only 3 articles scattered in other areas: Australia and South Korea. The articles from the USA   (Table 2).
The first author with the most articles and their basic information is listed in Table 3. Eriksson BI, from Sweden, had 7 fist authorships mainly in the field of thrombogenesis. The total citations of his selected articles were 3209.
The largest number of articles were level III evidence (n = 35), with a mean number of 304 ± 116 citations per article, followed by the number of articles which represented level II (n = 25) and level IV (n = 25), with a mean number of 269 ± 216 and 281 ± 97 respectively. No significant difference in citations per article among the various levels of evidence has been found by using the one-way ANOVA (Fig. 7).
Network analysis of the author's key words or subject terms has been done in two periods of article published time: in the 1990s (46 articles) and in the 2000s and 2010s (54 articles). The result indicated that "follow-up study, radiography, acetabulum, reoperation, and bone cements" had a high degree of centrality in the 1990s ( Fig. 8), while "cement" had a high degree of centrality in the 2000s and 2010s (Fig. 9).

Discussion
This is the first bibliometric analysis of total hip arthroplasty papers. Several significant findings can be drawn from this analysis of the top 100 most cited papers published on total hip arthroplasty. The maximum number of citations only reached 994, and this was a paper on anticoagulation. One explanation is that the current analysis only focuses on articles published from 1990 to 2019. Therefore, there may not be enough time for the articles to be thoroughly cited. Majority of the top-cited articles were published in the 1990s with a similar number in the 2000s and only 5 articles published since 2010 made the top 100 most cited. A phenomenon known as "obliteration by incorporation" [15] was not demonstrated in our analysis. The regional distribution of articles has also been demonstrated. Majority of the articles originated from two major regions: North America and Western Europe. There are several reasons for this phenomenon: (1) total hip arthroplasty surgery originated in the USA in the 1960s; hence, they started to study it earlier and in greater detail; (2) Sweden was the first country to establish an artificial joint registry and they continue to analyze the data annually, Finland, Norway, and Denmark also established a joint registry system increasing their publications; (3) countries such as the USA, England, and Sweden have a fairly developed economy to support medical research; it has previously been demonstrated that there is a weak correlation  Table 2 Journal in which the top-cited 100 articles were published between the gross domestic product (GDP) per capita of a country and their research achievements [16]. It is of great importance that the achievements in these countries are shared worldwide so that everyone can benefit from their research advances.
This article also examined the theme distribution of the most cited articles. Surgical methods and materials, and postoperative thrombosis were the top two. The former has always been a research hotspot since the artificial joint replacement was invented [17][18][19]. Researchers share advances in how they improve joint stability and surgical prognosis through modifying the artificial joint or changing the materials. The degree of centrality analysis of the author's key words also indicated that "cement" has a high degree of centrality both in the 1990s and again in the 2000s. Using or not using cement during operation is still a controversial issue [20][21][22], which needs to be further explored in the future. Postoperative thrombosis has also been at the center of research interest [23]. Postoperative thrombosis is one of the important complications following total hip arthroplasty, which has a great impact on the prognosis of patients. In fact, blood clots that form after surgery have been of interest to researchers since very early years [24][25][26][27]. Up to now, thrombus formation after surgery is still widely investigated in different research fields including hip arthroplasty [28][29][30].
The level of evidence analysis demonstrated that the level of evidence ranged between I and IV, mainly level II, III, and IV. These are relatively high and evenly distributed, which is different from other bibliometric studies [31,32]. This may be because researchers have  invested more in total hip arthroplasty study in the past 30 years because it is a significant advancement in the treatment of pain and morbidity and has become the focus of people's attention. In the future, with the increasing demand for quality of life and the increasing problem of an aging population in some countries, especially in China, it is expected that total hip arthroplasty surgery will increase more rapidly, as will the related research articles. Older people are more likely to have hip problems; therefore, the aging population issue will The results showed that in total hip replacement patients, once-daily oral administration of 10 mg rivaroxaban was more effective in preventing thrombosis than once-daily subcutaneous injection of 40 mg enoxaparin [33]. The article with the second most citations is another research article by Eriksson, BI. et al., published in the Lancet. They compared the efficacy of dabigatran etexilate and enoxaparin in preventing thrombosis after total hip surgery. It has been proved that oral dabigatran etexilate is as effective and safe as enoxaparin in reducing the risk of venous thromboembolism after total hip replacement [34].
Actually, in the preliminary analysis, articles published on total hip arthroplasty were included from 1950 to 2019. Not surprisingly, the result showed that most of the top-cited articles were about surgical materials and prognosis. The most cited article was a study by DeLee, J G et al. in 1976, which reached 2289 citations. They explored the radiological separation of acetabular cement junctions after total hip arthroplasty and the result showed that the Fig. 8 Degree centrality analysis in the 1990s (46 articles). It showed that "follow-up study, radiography, acetabulum, reoperation, and bone cements" had a high degree of centrality in the 1990s Fig. 9 Degree centrality analysis in the 2000s and 2010s (54 articles). It showed that "cement" had a high degree of centrality in the 2000s and 2010s application of cement was not the explanation for the demarcation of the cement-bone junction; therefore, they concluded that there is no fundamental defect in the principle of using cement for acetabular components [35]. This research was of great importance in promoting the use of cement in total hip arthroplasty. Although these articles are of great significance, the current bibliographic analysis aimed to focus on the most important advances in total hip arthroplasty over the last 30 years. From the 1990s onwards, there were more and more articles published on postoperative thromboembolism prevention, and they were becoming more topical and influential.
This analysis yielded some valuable information, but there were limitations. While articles were filtered by how much they were cited, which overlooks the newly published articles that are significant in this field but have yet to reach high citation levels. Hence, it is biased towards historical articles. At the same time, only published articles were included while meeting records, textbooks, meta-analysis and reviews were excluded, which may result in omission bias.

Conclusion
This article highlights the top 100 most cited articles in the total hip arthroplasty over the last 30 years, including their time and area of distribution, their research topic, their authorship as well as their level of evidence. This analysis sheds light on the spotlights and characteristics of total hip arthroplasty research since 1990 to 2019. Furthermore, we have noticed that, currently, postoperative thromboembolism plays a major role in the field of total hip arthroplasty. However, no one has really solved this tough problem. Meanwhile, most of them were focus on the clinical trial while the basic laboratory research was insufficient, which calls for more researches based on this topic. Thus, in the future, the prevention and management of thrombosis after total hip arthroplasty should be further investigated, especially the basic mechanisms.
Abbreviations ANOVA: Analysis of variance; GDP: Gross domestic product