The top cited clinical research articles on sepsis: a bibliometric analysis

Introduction The objective of this study was to identify and characterize the most highly cited clinical research articles published on sepsis. Methods A comprehensive list of citation classics in sepsis was generated by searching the database of Web of Science-Expanded (1970 to present) using keywords 'sepsis' or 'septic shock'. The top 50 cited clinical research papers were retrieved by reading the abstract or full text if needed. Each eligible article was reviewed for basic information, including country of origin, article type, journals, authors, and funding sources. Results A total of 2,151 articles were cited more than 100 times; the 50 top-cited clinical articles were published between 1974 and 2008. The number of citations ranged from 372 to 2,932, with a mean of 678 citations per article. These citation classics came from nine countries, of which 26 articles came from the United States. Rush University and the University of Pittsburgh lead the list of classics with six papers each. The 50 top-cited articles were published in 17 journals, with the New England Journal of Medicine and Journal of the American Medical Association topping the list. The top 50 articles consisted of 21 clinical trials and 29 observational studies. Conclusions Our bibliometric analysis provides a historical perspective on the progress of clinical research on sepsis. Articles originating from the United States and published in high-impact journals are most likely to be cited in the field of sepsis research.


Introduction
Sepsis is a systemic inflammatory response syndrome that occurs during severe infection. It remains a leading cause of death in critically ill patients [1]. Numerous critical care and infectious disease specialists and researchers have focused their efforts on sepsis in an attempt to gain a better understanding of the pathophysiological basis of sepsis or to develop new methods for the diagnosis and treatment of sepsis. Large numbers of articles have been published annually and have given new insights into the mechanism or treatment of sepsis [2].
It is generally accepted that publications represent the central part of a research process. Citation rating is a popular method for evaluating the impact of an investigator or a publication in the scientific community concerned. The frequency of citing has significant implications for authors, journals, institutions and even nations [3]. A remarkable citation history of an author often signifies great honor or recognition in a particular area of research. Although there are obvious disadvantages in assessing the quality of a study simply based on the citation rating, it is widely accepted that this is the best method currently available for judging the merit of a paper or a journal [4]. Citation analysis is also a feasible tool to comprehensively recognize the research advances in the past and future research trends in a specific field.
Clinical research refers to research conducted with human beings, including studies in patient-oriented research, epidemiological and behavioral studies, outcomes and health service research. Clinical study plays a special role in the fight against sepsis because it can provide overwhelming evidence for the treatment and diagnoses of a disease or disorder. Analysis of the most cited articles allows clinical investigators to identify the most popular field of research in sepsis and will give us insights into the characteristics and quality that are required for an article to become widely cited.
Recently, various specialties have attempted to summarize 'citation classics' or the most commonly cited articles in their fields [5][6][7][8]. In order to systematically review the citation classics dedicated to sepsis, we conducted the current study to focus exclusively on the 50 top cited clinical articles in an attempt to provide a bibliometric perspective of the progress in sepsis research. We also intended to identify factors contributing to the successful citation such as journals in which the articles were published and related countries.

Materials and methods
The database of the Institute for Scientific Information (ISI) Web of Science Expanded citation index (1970 to present) was searched using the keyword 'sepsis' or 'septic shock' to identify the citation classics cited more than 400 times. This database includes peer-reviewed publications indexed from more than 10,000 high impact journals worldwide. The 'document type' was applied to limit the format of publications and the type of articles. Papers published as 'article' were selected for further analysis. Each article on the list was reviewed by reading the abstract first and only studies dedicated to clinical research on sepsis were selected for further analysis. The following information was recorded: authors, the number of citations, year of publication, country of origin, institution, journal, funding source, and article type or subfield (for example, randomized controlled trials, observational research). We also calculated the publication output on sepsis by normalizing the total number of articles to the gross domestic product (GDP) per capita of the included countries. The information of GDP per capita was retrieved from the website of the World Bank. All electronic searches were conducted on 13 March 2012. Correlation analysis between GDP per capita and the total number of articles on sepsis was conducted using GraphPad Prism 5.0 software.

Results
A total of 67,558 papers were identified in the initial search for the period from 1970 to present, with 50,192 published as 'article' and 4,748 classified as 'review'. Among them, 2,151 articles were cited more than 100 times.
Of the top 50 cited clinical trials, the mean number of citations was 678 (range 372 to 2,932) and six papers were cited more than 1,000 times (Table 1). These articles were published between 1974 and 2008, of which about 50% were published after 1995 ( Table 2).
The 50 top cited articles originated from nine countries, with the United States (26) and France (8) being the most prolific (Table 3). Given some articles were authored with multiple sources of origin, especially those in the form of international research collaborations, the country of origin was defined by the address of the corresponding author. The leading institutions are shown in Table 4. Rush University and the University of Pittsburgh were found to be the most productive institutions, with six articles each. There was a weak correlation between the GDP per capita and the numbers of articles on sepsis related to the nine countries (r 2 = 0.02).
The 50 top cited clinical articles were published in 17 journals, predominantly in New England Journal of Medicine (n = 11) and Journal of the American Medical Association (n = 11), followed by Critical Care Medicine (n = 6) ( Table 5). Table 6 presents a list of the most productive authors, indicating that Sprung CL authored six articles, followed by Abraham E, Bone RC, Fisher CJ and Vincent JL.
Of the 50 top cited articles, 21 reported clinical trials, of which 19 were designed as randomized controlled trials (RCT) and 29 reported observational studies. Among these original research papers, 15 were funded by public foundations, 12 received support from commercial companies, six were supported by both, and for the remaining 17 the funding source was not specified (Table 7). Specifically, five studies received grants from the National Institutes of Health (NIH). Eight of the 19 RCTs reported supportive interventions and 11 studies showed detrimental or neutral effects [See Additional file 1].

Discussion
The present study summarizes several features of influential articles in sepsis research during the past 40 years by means of a literature review. It was found that 26 (52%) of the 50 articles were from the United States. The country dominance is also found in other clinical disciplines including Urology [5], Orthopedic Surgery [6], Critical Care Medicine [7], General Surgery [8], Emergency Medicine [9] and Anesthesia [10]. The underlying reasons might be due to its large population of senior researchers and adequate research budgets for scientific investigation. It is generally presumed that the correlation between the research output and GDP should be an important issue in bibliometric analysis. But we found that the correlation between the GDP per capita and the total number of articles published on Table 1 The top 50 cited clinical trials on sepsis.

Ratings Article
No   sepsis was weak in the related nine countries. This might be due to the limited number of retrieved articles and included countries. In addition, it was found that American authors tended to cite local papers [11] and US reviewers had a significant preference to accepting papers written by native researchers [12]. Our study found that most of the classics were published in highimpact journals, which is consistent with the result of other reviews, supporting the well known paradigm that top cited articles are often published in journals topping the impact factor list, which in turn maintains the high impact factor of these journals [13]. Financial support from public foundations or commercial companies has evolved over time in response to changes in professional codes, laws and markets [14]. Public funds have given a great push to the development of medical research and public health. In our study, 15 papers reported funding support from public institutions or national foundations. Industry-funded science has been widely debated because of the susceptibility to various kinds of biases. Nevertheless, it has played and will continue to play a critical role in the research process [15]. Our review shows that a total of 18 research projects received grants from ommercial companies. A remarkable thing should be considered that some papers, especially the old ones, might not have fully reported their financial conflicts.
Clinical research has bridged the gap between basic science and improved human health and is heavily weighted towards biomedical science. In the past 40 years, clinical studies have tested large numbers of therapeutic agents and provided insights into the pathophysiological basis of sepsis. The list of the top cited articles shows some interesting trends and pinpoints major advances in sepsis research. Sepsis has been defined as hyper-inflammation and excessive activation of the immune system characterized by a prolonged cytokine storm. Therefore, many studies have targeted certain cytokines highly expressed in sepsis for therapeutic or diagnostic application. IL-6 and TNF-α appeared to be good markers for predicting severity and prognosis of sepsis. However, clinical trials have failed to demonstrate promising results using antibody blockade to these cytokines, for sepsis initiates much more complex immunologic responses. Glucocorticoids have been another hot topic in the field of sepsis research. Owing to the positive effect on the sensitivity of vascular smooth muscle to catecholamines, steroid has been considered as a promising agent in septic shock patients. Currently, the use of glucocorticoid cannot be recommended as the standard of care, but it is feasible in the course for patients with septic shock that does not respond to conventional measures.
Of the 19 RCTs, eight demonstrated promising and supportive interventions [See Additional file 1]. Considering that RCT-based studies provide greater quality evidence than other study designs, trials with promising results often encourage clinicians to apply these interventions. Unfortunately, successful measures for sepsis treatment seem difficult to achieve; in the top cited 50 articles, 11 of the 19 RCTs have proved to be non-efficacious. Reasons related to the negative results include the reporting quality of these studies, small sample size or heterogeneity of the ICU patient population [16].  Our review has several limitations. First, we elected to limit our research to 'sepsis' or 'septic shock' in the topic field, which may miss some citations related to our analysis such as those indexed with 'LPS'. In addition, our search via the Web of Science expanded database from 1970 to present is quite sensitive and papers published before were excluded, so it is likely that some true 'classic' articles were missed in this review.
Another limitation is the inherent bias of the citation analysis [17]. Total cites of an article accumulate over time which means that older publications would definitely receive more citations than new ones. Recent publications with a short span of time to generate citation rates are possibly underestimated with respect to their impact. In addition, oriented or biased citing, including self-citation, in-house, or negative citation (bias towards potential negative credits) is also a problem that should not be ignored [18]. An important thing to be kept in mind is that impact factor or citation analysis is not an index to evaluate the quality of scientific research, but rather a measure of recognition. In other words, the number of citations of an article should not be considered equivalent to its importance.

Conclusions
Our bibliometric analysis provides a historical perspective on the progress in clinical sepsis research in the past 40 years. Papers originated from the US and published in high-impact journals are most likely to be cited in the field of sepsis research.

Key messages
• The mean number of citations per article was 678 and six papers were cited more than 1,000 times. • The US is responsible for the most contributions to clinical studies on sepsis.
• The top cited 50 articles on sepsis were published in 17 journals, led by the New England Journal of Medicine (n = 11) and the Journal of the American Medical Association (n = 11), followed by Critical Care Medicine (n = 6).
• About half of research papers received funding support from public institutions or national foundations.
• Nineteen articles were designed as RCTs, among which eight papers reported promising results.

Additional material
Additional file 1: Therapy or agents tested in RCTs on sepsis. Nineteen studies among the citation classics were designed as RCTs. Treatments tested in eight articles were found to be promising or supportive interventions.