Abstract
Study Design:
Bibliometric review of the literature.
Objective:
This bibliometric analysis aims to give an overview of the most influential academic literature written on spinal tuberculosis.
Methods:
All databases included in the Thomson Reuters Web of Knowledge were searched for the most influential publications in spinal tuberculosis. The most cited articles published between 1950 and 2015, with the main focus on orthopedic surgery, were identified using a multistep approach, and a total of 100 articles were included. The publications were then analyzed in this bibliometric analysis.
Results:
The number of citations ranged from 243 to 36, with an average of 77.11. The articles were published in 34 different journals, and the studies were conducted in 20 different countries. The top 3 countries, India, the United States, and China, published a total of 51% (n = 51) of all articles. Low-burden countries produced 60% (n = 60) of all articles in our list. African centers produced only 4% (n = 4) of all included articles.
Conclusions:
Indian and Chinese researchers dominate evidence in spinal tuberculosis. Other areas with high disease burden, such as Africa, do not feature. Most publications are retrospective studies with a low level of evidence.
Introduction
For the past 50 years, an increase in the incidence of tuberculosis (TB) has been reported globally, 1 but the disease is most common in in low- to middle-income countries. 2 Through immigration and increased mobility of travelers, TB remains a global challenge and is still the leading cause of death caused by a single pathogen. 2 Spinal disease is the most common form of skeletal TB, 3,4 and is one of the main pathologies seen in spinal and general orthopedic units in the developing world. Growing economies in countries like China and India have led to an increase in medical research, but most low-income countries cannot generate scientific evidence as resources are already stretched with clinical work. 5 Yet, especially for these countries, it is challenging to address the clinical burden of endemic diseases without indigenous evidence and solutions.
Bibliometric analysis can be used to analyze the influence of articles by quantitative appraisal of its impact 6 in TB research 7 as well as in spinal surgery. 8 The objective of this study is to identify the 100 most cited and therefore most influential articles on spinal TB.
Methods
From all available journals of the Thomson Reuters Web of Knowledge, we selected the 100 most influential articles on spinal TB between 1950 and 2015 by means of a bibliometric analysis.
Inclusion Criteria
All articles with focus on the clinical course or treatment of spinal TB were included. “Spinal tuberculosis” was defined as an infection of the spine caused by Mycobacterium tuberculosis. The “spine” was defined as the cervical, lumbar, thoracic, and sacral vertebrae and all intervertebral discs. All age groups were included.
Exclusion Criteria
All articles not predominantly focusing on spinal TB were excluded. This was the case in articles on spinal infections caused by various pathogens with only a minority of TB infections. All studies with a nonorthopedic focus were also excluded, that is, an article on the mechanism of TB drugs. Articles focusing on the infection of the spinal cord were excluded.
Selection Process
The selection process was started using a topic and title search in the “all database” of the Thomson Reuters Web of Knowledge. Seven keywords were used: spinal, spine, vertebral, Potts, spondylodiscitis, tuberculosis, and spondylitis. An asterisk was used to extend the search. For example, vertebr* will search for vertebral and vertebrae. The operator “AND” was used to narrow the search; by using it, the results will contain all Boolean queries. For example, vertebr* AND tubercu* will contain both keywords, whereas vertebr* OR tubercu* will contain either vertebr* or tubercu*. The 150 most cited articles were then explored for more keywords. One additional keyword was added, resulting in a total of 8 keywords.
With these final 8 keywords, we conducted the literature search, again in the “all database” of the Thomson Reuters Web of Knowledge, which led to a result of 11 539 articles. The first 100 publications were chosen for further analysis based on the consecutive ranking of their citation rate.
For each study included, we extracted the following parameters: title, year of publication, total citations, citations in 2014, citation density (citation number divided by the article age), article age, journal, first author, senior author, geographic origin, institution, number of patients, and age group of patients (adult, children, or both).
In case of only a single author, the author was considered both first and senior author.
Analysis
All articles were systematically analyzed for the following information: area of research, type of study, and level of evidence (LoE).
The level of evidence was established using the Journal of Bone and Joint Surgery American Volume 9 ranking, with level I being the strongest and level V being the weakest level of evidence. Articles were first divided into 4 categories: diagnostic, prognostic, therapeutic, and economic.
The area of research was also assigned according to the primary focus of the article and subdivided into 4 categories (Table 1). The same was done for study type, which we divided into 5 categories (Table 2).
Area of Research.
Type of Research.
We ranked all articles according to their total citation number, number 1 having the highest number of citations and number 100 having the lowest. In case of an equal number, the articles were ranked according to citations in 2014 and citation density respectively.
Results
The number of citations ranged from 36 to 243, with an average of 77.11. All articles were published between 1950 and 2015 in 34 different journals (Table 3). The research originated from 19 countries (Table 4). Most evidence was level IV.
Publishing Journals.
Geographic Origin.
All articles were published in the English language.
Top Ten (Table 5)
The citation number in our top 10 articles ranged from 137 to 243. The articles were published between 1956 and 1999 and had an average age of 28.3 years (CI). The top journals in our top 10 were Spine (n = 3; 30%) and Journal of Bone and Joint Surgery, American Volume (n = 3; 30%). Both A. R. Hodgson and M. S. Moon have authored 2 articles (20%); all other authors in our top 10 are responsible for one publication. A total of 6 articles (60%) originate from 3 different countries: China (Hong Kong; n = 2; 20%), South Korea (n = 2; 20%), and the United States (n = 2; 20%). The remaining articles were from India, Japan, Spain, and Turkey (Table 4).
Top 10 Publications.
The most cited article was written by H. G. Watts et al in 1996, in Journal of Bone and Joint Surgery, American Volume, with 243 citations. It reviews the clinical course of spinal TB with a LoE III (Table 6).
The Most Influential Publications: Total Citations (n), Citations in 2014 (n), Citation Density (Number/Year).
Geographic Origin
A total of 20 different countries was identified, located in 4 different continents: North America, Europe, Africa, and Asia. The geographic origin responsible for most publications was India (n = 23; 23%), followed by the United States (n = 15; 15%) and China (n = 13; 13%). See Figure 1. Together these countries published a total of 51 articles (51%) of our selection. Of the 13 articles originating from China, 10 (17.92%) were from Hong Kong.
Geographic origin. Countries are highlighted with light gray (least publications) to dark gray (most publications).
Asia published most articles with 57% (n = 57), followed by Europe (n = 22; 22%), North America (n = 17; 17%), and Africa (n = 4; 4%; Table 4).
Journals
Most articles were published by Journal of Bone and Joint Surgery, British Volume, with 20 publications (20%), followed by Spine (n = 14; 14%), Journal of Bone and Joint Surgery, American Volume (n = 9; 9%), Clinical Orthopaedics and Related Research (n = 8; 8%), and American Journal of Roentgenology (n = 4; 4%). These 5 of the 34 journals in our selection produced 55% (n = 55) of the included articles.
Authorship
A total of 11 authors in our list published 2 or more articles. S. Rajasekaran, from India, is the most published author with a total of 5 publications (5%), 5 as first author and 2 as first or senior author. S. M. Tuli has published most articles (n = 3; 3%) as senior author and was listed as the second most published author (Table 7).
Authorship.
Article Age
The oldest article included in our list was published in 1951 and written by J. Dobson. It was cited 46 times and has a citation density of 0.71, with zero citations in 2014. The most recent article included, published in 2011, is by R. K. Garg et al, with a citation number of 41 and a citation density of 8.20.
Most articles were published between 1990 and 2000 (n = 38; 38%), and a total of 32 articles (32%) were published before 1990.
Level of Evidence
Our top 10 consists of only level III evidence (n = 5; 50%) and level IV evidence (n = 5; 50%). Only 4% (n = 4) of our list were Level I studies. The large majority (n = 50; 50%) of all studies was assigned a level IV or lower. Rajasekaran of India published 2 articles with level I evidence. Of the 4 articles, 3 originated from India and 1 article is from the United Kingdom. Their area of research varied from clinical course (n = 2), imaging (n = 1), and surgical technique (n = 1). Most of the remaining articles were level IV evidence (n = 42).
The distribution of area of research and study type of all included articles can be found in Tables 1 and 2.
Discussion
Article age
The oldest article in our list was published 65 years ago, written by J. Dobson. It is ranked 75 in our list with 46 citations and might be outdated when compared to the mean age of all articles of 23.2 years. This is supported by the fact that it has not received citations in 2014.
Most articles were published between 1980 and 1990 (n = 56; 56%). The 5 oldest publications are all published before 1997 but are still of significance in current literature. Of these articles, 3 have received more than 100 citations and 2 are in our top 10, both written by A. R. Hodgson et al in places 2 and 3.
Geographic Origin and Journals
Most publications (n = 23; 23%) in our list are from India, a country listed by the World Health Organization (WHO) as a high-burden country in the incidence and prevalence of TB as well as multidrug resistance TB and TB-HIV coinfection. India is followed by the United States, responsible for 15 articles (15%; Figure 1). Although the United States faces severe TB outbreaks, 10 it is not listed as a high-burden country by the WHO (Figure 2). China is among the top 3 countries with 13 publications (13%). Hong Kong published 10 of these 13 articles (76.9%), attributed to the work of Hodgson, promoting early and radical anterior debridement as the “gold standard” for spinal TB. China is listed both as a high-burden TB country with a high multidrug resistance TB and HIV-TB coinfection. 5
Estimated TB burden. Incidence of TB per country, as taken from the most recent WHO TB report. Countries are highlighted with light gray (lowest incidence) to dark gray (highest incidence).
On the other hand, only 4 publications (4%) originate from the continent of Africa, with 3 of them being conducted in South Africa: E. B. Hoffman et al; J. A. Louw and S. Govender; P. G. Konstam et al conducted their research in Nigeria.
Africa has one of the highest TB burden and highest TB-HIV coinfection rate worldwide 2,5 (Figure 2), yet there is a discrepancy between clinical experience and published research. This might point toward insufficient resources available to African institutions to conduct research or publish their experience.
Study Category, Area of Research, and Study Type
Most studies were retrospective studies (n = 70; 70%), with only 6 (6%) prospective studies, of which 4 were randomized controlled trials (4%).
This indicates that retrospective research is still seen as important evidence in global literature on spinal TB and is of value to conduct, and on the other hand, it shows that prospective research and randomized controlled trials are very scare and future research should fill this gap.
The most popular area of research was the clinical course of spinal TB (n = 30; 30%), focusing on the overall clinical picture of the disease, diagnosis, and treatment. The second most popular area was surgical technique (n = 26; 26%) followed by clinical outcome (n = 23; 23%) and imaging (n = 21; 21%). This shows that most centers in high-burden areas have no access to funding and extensive laboratory support to conduct basic science research or large randomized controlled trials.
Quality
Most studies were retrospective and therefore granted level IV evidence. This could be due to the fact most studies were conducted in resource restrained areas were little funding and equipment is available and randomized controlled trials are hard to set up. Indian researchers seem to be the most resourceful as 3 of the 4 prospective studies were conducted in India.
On the other hand, relatively simple studies are contributing to our knowledge on spinal TB, showing that centers with limited resources can contribute to the body of evidence without the need of major research resources.
Limitations
Bibliometric analysis provides an overview of literature at one specific time point. Due to the dynamic field in which academic literature is used, citation numbers change every day. For instance, the most recently published article is by R. K. Garg et al, published in 2011. It received 41 citations, was cited 12 times in 2014 and it has a citation rate of 8.2 per year. Looking at the citation density and number of publications in 2014, this article is likely to climb into the top 100 most cited articles in the future. Therefore, it has the possibility to become more important in the current literature. This publication highlights the fact that we might need to focus on citation density and citations in recent years to keep up with the dynamic effects of younger but influential articles.
One problem of our study might be that it cannot answer the question, whether the low citation rate of countries with high disease burden is due a low productivity itself or due to a low citation rate per se. Future research could address this by assessing the global research output of these specific countries.
Article age is of importance in a bibliometric analysis; the older an article is the longer it can gather citations. P. G. Konstamp et al published their article in 1962 and it has been granted 104 citations since, making this the 20th article in our list. In 2014, the article did not receive new citations, which could question its significance in the current literature.
Conclusion
This bibliometric study provides the 100 most influential articles on spinal TB. Most studies focused on clinical outcome and were LoE IV. India contributed most articles, followed by the United States and China. Almost half of the articles in our selection originated from developed countries. Only 4 articles were produced from African centers. Our study serves to highlight the low impact of publications in the field of spinal TB from endemic areas, especially the African continent. There is a lack of prospective studies with a higher level of evidence. Collaborations of centers in low-income countries and high burden of disease with centers with high research output in countries with sufficient resources may be one solution to overcome this research gap.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.