|Year : 2017 | Volume
| Issue : 3 | Page : 229-238
A neglected infection in literature: Childhood musculoskeletal tuberculosis – A bibliometric analysis of the most influential papers
Michael Held1, Marie-Fien Bruins1, Sophie Castelein1, Maritz Laubscher1, Robert Dunn1, Sven Hoppe2
1 Department of Orthopaedic Surgery, Orthopedic Research Unit, Groote Schuur Hospital, University of Cape Town, Cape Town, South Africa
2 Department of Orthopaedic, Inselsspital, University of Bern, Bern, Switzerland
|Date of Web Publication||31-Jul-2017|
Orthopedic Research Unit, Groote Schuur Hospital, University of Cape Town, 7925 Observatory, Cape Town
Source of Support: None, Conflict of Interest: None
Pediatric tuberculosis (TB) is known to have a wide range of presentations, and if left untreated, primary TB may lead to bone and joint involvement. The literature on this topic is very scarce, and no comprehensive systematic review or meta-analysis of the current knowledge is available to date. The aim of this study is to identify and analyze the literature with highest impact based on citation rate analysis. All databases of the Thomson and Reuters “Web of Knowledge” were used to conduct our search of the 100 most cited articles on this topic published between 1950 and 2014. The included articles were analyzed in terms of citation rate, age, study type, area of research, level of evidence (LOE), and more. All 100 articles were published between 1967 and 2011 in 51 different journals. The average citation rate was 74.26, all articles were on average 23.1 years, and most studies were originated from India (n = 22), followed by the USA (n = 21). The majority of publications were review articles (42%), described clinical course (n = 48), and assigned an LOE IV (44%). TB infection is a high burden disease in low-income countries but widely studied in a fi rst world setup. This research gap between the geographic distribution of disease burden and origin of publications could initiate possibilities for high-burden countries to share their opinion. Their experience is of a high level of importance and relevance which furthermore is necessary to create a more accurate picture of pediatric musculoskeletal TB burden in literature.
Keywords: Bone, childhood, joint, musculoskeletal, pediatric, tuberculosis
|How to cite this article:|
Held M, Bruins MF, Castelein S, Laubscher M, Dunn R, Hoppe S. A neglected infection in literature: Childhood musculoskeletal tuberculosis – A bibliometric analysis of the most influential papers. Int J Mycobacteriol 2017;6:229-38
|How to cite this URL:|
Held M, Bruins MF, Castelein S, Laubscher M, Dunn R, Hoppe S. A neglected infection in literature: Childhood musculoskeletal tuberculosis – A bibliometric analysis of the most influential papers. Int J Mycobacteriol [serial online] 2017 [cited 2018 Feb 17];6:229-38. Available from: http://www.ijmyco.org/text.asp?2017/6/3/229/211941
| Introduction|| |
Tuberculosis (TB) is only second to HIV regarding mortality due to a single pathogen. In 2014, TB infected a total of 9.6 million people, which included an estimated one million children.
Extrapulmonary involvement accounts for approximately 20% of all TB-infected patients, and musculoskeletal manifestations may account for 10%–15% of this specific group , but can reach 35%.
For all cases of TB infection, bone and joint involvement may account for 4%–5% of the total patient population.,
Pediatric TB is known to have a wide range of presentations, with a variety from limited pulmonary or nodal disease to severe extrapulmonary or disseminated disease. In children, untreated primary TB may lead to bone and joint involvement in up to 7%.
Musculoskeletal TB is a serious condition that may have severe consequences for the affected individual, especially in childhood. Known to be regularly misdiagnosed as a pyogenic bone or joint infection, accurate treatment is frequently delayed which influences the clinical outcome negatively.,,,,,
Bibliometric analysis was originally designed to measure and compare the impact of scientific journals  and is now recognized as a valid technique to analyze the scientific value and impact of articles by means of a quantitative appraisal of citations, articles, and journals.
So far, there is no systematic analysis of the scientific literature about this specific topic. This study aims to address this research gap with a bibliometric analysis of the most influential publications in musculoskeletal childhood TB.
| Materials and Methods|| |
We performed a bibliometric analysis of the 100 most cited articles on musculoskeletal TB infections in children published in any medical or nonmedical journal between 1950 and 2014.
All studies on TB of bones and joints were screened for the patients' age of the study population. Only studies addressing TB in children were included.
Studies focusing only on adults were excluded. Furthermore, studies that did not have any focus on bone or joint involvement due to TB infection were excluded from our analysis. For example, a study describing the clinical course of TB which only focused on the general epidemiology with no records of musculoskeletal involvement was excluded. All studies describing adults with a background of childhood TB were excluded as well.
We performed an advanced title and topic search in “all databases” of Thomson and Reuters “Web of Knowledge,” using the following Boolean query as a basis: Tubercul* AND (child OR paediatric*). The asterisk (*) implying that the specific word may have different endings to cover all possible meanings (such as tubercular, tuberculous, or tuberculosis). In addition to this, our basic query as mentioned above was followed by any of the terms: AND bone OR joint OR musculoskeletal OR orthopaedic OR osteoarticular OR Potts-disease OR hip OR knee OR ankle OR elbow OR shoulder OR foot OR hand OR wrist OR spinal.
This first search produced 646 articles. We systematically analyzed the titles and abstracts of the top 50 most cited articles of this search to determine further keywords which might have been missed in our first search.
A second search was performed with 22 new keywords (List I) with the following Boolean query as a basis: Tubercul* AND (child* OR pediatric* OR paediatric* OR infant*).
This second combined search yielded 4465 results in total, which were sorted based on the number of citations, from highest to lowest. Of the top 338 articles, we chose the 100 most cited studies based on our inclusion and exclusion criteria.
For each single study included, the following parameters were documented: title, journal, year of publication, age, first author, senior author, total citations, citation density, citations in 2014, geographic origin, institution, number of patients, and age group of patients (adults and/or children).
The selected articles were analyzed according to subspecialty, type of study, and a level of evidence (LOE). For the LOE, the guidelines of the Journal of Bone and Joint Surgery, American Volume were used to assign level I (highest LOE) to V (lowest LOE) for every clinical study. Nonclinical (i.e., basic research) studies were not assigned to an LOE. A consensus approach was used in cases of disagreement between the researchers.
For studies with only one reported author, both first and senior authorship were assigned.
The ranking of included studies was given according to their total number of citations, from highest to lowest citation rate. Studies with an equal number of citations were ranked according to their average citations per year and citations in 2014.
For statistical analysis, the program SPSS (PYTHON Version 3.4.3, SPSS IBM, NY, USA) was used.
| Results|| |
Age – Oldest versus newest
All articles included were published between 1967 and 2011, with an average publication age of 23.1 years [Table 1]. The citation rate per publication of the total of 100 studies ranged from 690 to 30, with an average citation rate of 74.26.
The two oldest articles of this selection are from 1967 (Hodgson et al. The pathogenesis of Pott's Paraplegia and Robins' TB of the wrist and hand) and both have a citation rate of over 30. The most recent article included is on spinal TB by Garg et al. and hits 41 citations in total.
Most articles included studies were published in the first decade of 2000 (n = 30). There were no publications included from the 1950s although this decade was eligible to be recorded in this bibliometric analysis.
From 51 journals in which these articles were published [Table 2], the most influential journal was The Journal of Bone and Joint Surgery-British Volume (J Bone Joint Surg British), which accounted for 23% of articles. It was followed by the Journal of Bone and Joint Surgery American Volume (J Bone Joint Surg Am)with seven articles and Spine, International Orthopaedics, and Clinical Orthopaedics and Related Research, with four articles each (Appendix, List I. Terms [Additional file 1]). All together, these five journals published for 42% of the chosen articles. Four nonorthopedic journals with main focus on pediatric medicine had two publications or more. These were Pediatric Infectious Disease Journal (n = 3), Pediatrics (n = 2), Pediatric Radiology (n = 2), and Archives of Disease in Childhood (n = 2).
The most influential author was Rajasekaran from India marked as first author of six publications and two as senior author [Table 3]. Four publications were published by Moon, Jain, and Leong, all with at least one senior authorship.
Of a total of 27 different countries identified [Table 4], the geographic origin with most published studies was India (n = 22), followed by the USA with 21% of included publications. Countries with <10 published articles were Hong Kong (n = 9), England (n = 7), South Africa (n = 7), South Korea (n = 6), and Canada (n = 4). From most countries, only one article was published, except for Turkey, Sweden, Germany, and France, with each two articles.
Asia as a continent produced most publications (40%) with India in the first position (n = 22), Hong Kong second (n = 9), and South Korea third (n = 6). Taiwan, Singapore, and China all published one. This was followed by North America with 25%, of which the USA brought up 21 and Canada four.
Europe produced 24% of all articles, with most from England (n = 7). All responsible for two publications were Turkey, Sweden, Germany, and France. Switzerland, Scotland, Portugal, The Netherlands, Italy, Ireland, Greece, Finland, and Belgium all published one each.
The continent of Africa only produced eight percent of all publications, seven of which from South Africa and one from Algeria. The Middle East published just two from Saudi Arabia and Iran. Papua New Guinea was the only country representing the continent of Oceania with just one publication.
Although there were no language restrictions recorded in our inclusion criteria, all studies included were published in English.
The majority of publications were identified as review articles (42%) when looking at study type [Figure 1]. In our selection, 32 studies were listed as therapeutic studies. In total, 19 studies were diagnostic studies, focusing on medical imaging and tests. Seven articles were prognostic studies.
Category, area of research
With respect to the area of research [Figure 2], almost half of all publications focused on the clinical course (n = 48). Clinical outcome was the focus of 22 studies; surgical technique was listed third with 14% and medical imaging with 11%, respectively. Only a limited number of publications focused on medical treatment (n = 3) or laboratory tests (n = 2).
Anatomy and patients
Half of the articles (53%) focused on TB of the spine. The latter half (n = 47) focused on different body parts or specific joints such as the knee, hip, and ankle.
Only 25% of all the included publications focused exclusively on children [Figure 3]. The majority of which (n = 22) provided specified ages of all included patients, whereas three studies did not. The remaining 75% focused on both adults and children, which was listed as one of our inclusion criteria. Similarly, the majority specified for age of the included patients (n = 46) versus 29 studies which provided unspecified age groups.
Forty-four articles were retrospective level IV studies [Figure 4]. Only two studies were level I studies, seven studies level II, 20% level III, and six studies level V. There were 14 publications that could not be assigned an LOE due to being a “basic” study type, providing a general overview and therefore do not go into detail with regard to our studied topic.
Of the 10 most cited publications [Table 5], 50% originated from the United States (n = 5).
J Bone Joint Surg Am, Medicine and Spine all published two papers in this group.
These top 10 most influential articles were published on average 20.9 years (31–6), with an average citation rate of 245.1 (690–139) and an average citation density of 11,976 (31.41–5.15). The year of 2014 alone accounted for an average of 8.1 (28–0) citations.
The most cited publication was written by Basset al. in 1994 and listed with 690 citations in total, a citation rate of 10 in the year 2014 and a citation density of 31.41 per year. This is the only article published in our selection by the American Journal of Respiratory and Critical Care Medicine. The second most cited article was a review article from 1984 of Alvarez and Mccabe with a total of 304 citations, focusing on extrapulmonary TB.
Listed as third is the J Bone Joint Surg Am article by Watts and Lifeso on TB of bones and joints published in 1996.
The author with most publications in the top 10 was Moon from South Korea with two publications.
| Discussion|| |
Age – Oldest versus newest
Although our selection of articles has an average age of 23.11 years, most were published between the years of 2000 and 2010 (n = 30). The 1990s were the second most productive decade with 28 studies. This means that most high impact articles have been written in just 20 years around the turn of the century.
The most recently published article in our selection is from the year 2011. With a total citation number of 41, this merely 5-year-old article was cited 12 times in the year of 2014 alone, exceeding our top three publications. This publication focuses on spinal TB and is once more a review article. It is written by Garg and Somvanshi and originates from India.
The oldest articles included are almost 50 years, and although listed further down in this ranking, the number of citation was still above 30.
The most cited article in 2014 is another recent article from 2010, listed in the 13th place. It has the highest citation number of the preceding year with 30 total citations and a citation density of 18.17 citations per year, the third highest in our selection. It is titled Spinal TB: A Fresh Look at an Old Disease and has the potential to become one of the most influential publications in childhood musculoskeletal TB.
The J Bone Joint Surg British has so far published the highest number of articles in our selection (n = 23), followed by J Bone and Joint Surg Am (n = 7). Yet, the top articles in terms of citation have been published by American journals (J Bone Joint Surg Am, Medicine and Spine).
Most publications (49%) were found in journals focusing on orthopedics. Of these, 10 articles were published in journals focusing especially on spine. Interestingly, 13 publications could be found in 11 journals focusing on medical imaging or radiology. In total, only five journals could be identified as specialized in pediatrics, bringing up just ten publications (n = 10) in total, but seven (n = 7) ended up in the top 10.
TB remains a global health issue, especially in the developing countries., Although cases of TB are seen all around the world, the true burden of TB disease lies in the developing world, where this pathogen is still responsible for infecting around ten million people each year. South Africa is known to have the third highest incidence rate of TB worldwide, after India and China.
Most publications originated from India, which also is the country with the highest burden of TB in the world. With only one publication less, the USA-based research was also the main contribution to our selection of articles. Besides India, most research originated from first world countries indicating a major lack of indigenous research from countries of high disease burden.
This might be due not only to resource limitations but also to distance between medical centers and collaborators, especially in Africa.
For instance, India has many orthopedic centers with pediatric subspecialists, whereas these services are very scarce in African countries.
The “overview article” was the most frequently identified study type in the included publications. This shows the distinct characteristics of an infrequently studied topic. As the treatment of musculoskeletal TB in children is still a highly complex component, therapeutic studies were listed as a second most identified study type, demonstrating the comprehensiveness of its therapy and outcome.
Overall, there was a low level of scientific evidence as most publications were based on retrospective, level IV studies, conducted over long periods of time. Especially in studies that focused exclusively on children, the LOE was low.
This might be due to the fact that this disease is rare in children or that countries with high TB burden are often resource restrained and do not have the infrastructure to conduct large prospective trials. Spinal TB is most frequently seen in musculoskeletal TB infection which might explain that this was also the most studied anatomical body part.
Anatomy and patients
According to our inclusion criteria, all included studies had to involve children, which did not mean to exclude adults. The majority of publications also included adults. It shows that highly influential publications in TB research, specifically focusing on children, are very rare.
All top 10 publications were written and published in the decades before 2000, making it more likely to gain a higher number of citations mainly due to the age of the article.
These articles were still cited regularly in 2014 and stay relevant to the research field.
The publications with the highest citations in this group are review articles and might be used by many authors as a basis for their literature review.
Bibliometric analyses have possible limitations., They are based on citation analysis which is time dependent. Therefore, it has to been taken into account that such a study only reflects the findings of one moment in time and provides a snapshot of most influential literature. Nonetheless, they remain to be the current gold standard with regard to establishing the scientific impact factor.
Yet, the citation number of an article is merely a display of the overall interest in that specific publication and does not reflect its quality per se. It only shows the recorded time in the cited references. This may even mean it was only cited by other papers to refute the research, to demonstrate its faulty methods, fraudulent data, or other limitations.
| Conclusion|| |
Although most publications originate from India, there is a lack of publications from other countries with a high burden of disease. Furthermore, the quality of studies is rather low, musculoskeletal TB is less frequently studied in children than in adults and spinal TB is mostly addressed. Therefore, possibilities are initiated and opportunities are raised for high burden countries to share their opinion. Since they should be considered experts on the subject, their limited-resource experience is of a high level of importance and relevant to the rest of the affected third world countries.
Future studies should focus on prospective research and consist of larger patient populations to improve the LOE. Finally, musculoskeletal TB in childhood specifically should be addressed more often and include affected anatomical areas other than the spine.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Westall J. Tuberculosis levelling off worldwide. BMJ 1997;314:921.
Rieder HL, Snider DE Jr., Cauthen GM. Extrapulmonary tuberculosis in the United States. Am Rev Respir Dis 1990;141:347-51.
Houshian S, Poulsen S, Riegels-Nielsen P. Bone and joint tuberculosis in Denmark: Increase due to immigration. Acta Orthop Scand 2000;71:312-5.
Girling DJ, Darbyshire JH, Humphries MJ, O'Mahoney G. Extra-pulmonary tuberculosis. Br Med Bull 1988;44:738-56.
Berman S, Kibel MA, Fourie PB, Strebel PM. Childhood tuberculosis and tuberculous meningitis: High incidence rates in the Western Cape of South Africa. Tuber Lung Dis 1992;73:349-55.
Carrol ED, Clark JE, Cant AJ. Non-pulmonary tuberculosis. Paediatr Respir Rev 2001;2:113-9.
Petersdorf R, Harrison T. Harrison's Principles of Internal Medicine. New York: McGraw-Hill; 1983.
Garfield E. Citation analysis as a tool in journal evaluation. Science 1972;178:471-9.
Ioannidis JP, Boyack KW, Small H, Sorensen AA, Klavans R. Bibliometrics: Is your most cited work your best? Nature 2014;514:561-2.
Boudoulas H. Essays on citation classics in Cardiology. Cardiology 2014;127:263-4.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]