The Top 100 Cited Articles in Essential Tremor

Introduction

The number of times an article is cited in scientific journals reflects its impact on a specific biomedical field or specialty and reflects the impact of the authors’ creativity.1–3 The Institute for Scientific Information (ISI) has been collecting citation and other academic impact information since 1945, and this information has been electronically available since 1979. ISI (now a subsidiary of Thomson Corp.) has named their newest journal citation system “Science Citation Index Expanded,” and it is one of the databases available under the banner of the Web of Science.4 Citation data from peer-reviewed articles are indexed for more than 10,000 high-impact journals in the sciences and social sciences, as well as in the arts and humanities.

In recent years, several citation analyses of the top-cited articles have become available for various areas of clinical neurology and neurosurgery, including epilepsy,5 multiple sclerosis,6 Parkinson’s disease,7 brain metastases research,8 and neurosurgery.9–11 However, no similar study has been performed in the field of essential tremor (ET). ET is one of the most common involuntary movement disorders, and it is the most common tremor disorder.12–14 The past few years have witnessed an exponential growth in the number of scientific publications on ET, mainly across the disciplines of epidemiology, clinical features, and pathology.15–18

We used the electronic version of the Science Citation Index Expanded to determine which published articles in ET have been cited most often by ranking the 100 top-cited works. The expectation was to provide references for future studies and clinical practices.

Methods

A search was performed using the bibliometric database ISI Web of Science for articles including the keyword “Essential tremor*” in the title (the asterisk was included in the search string as a wild card character). In early April 2013, we identified the 100 top-cited articles (Table 1) published in professional journals since 1960 by the Web of Science. All articles in each given journal were ranked by the number of citations listed on Web of Science. The full texts were mainly selected by PubMed and ScienceDirect.

Basic information concerning country of origin (based on the first author), institution, year of publication, publication name, and citations of the target articles were collected from Web of Science using the analyze tool. Additional analyses were then performed to determine authorship, article type, study design, and level of evidence. For each study, the level of evidence was graded according to Hadorn et al.19

Cited half-life is defined as the number of publication years from the current year, which account for 50% of current citations received.4 This index helps to evaluate the age of the majority of cited articles published in a journal.4 The h-index aims to measure the cumulative impact of a researcher’s output; the value of h is equal to the number of papers (n) that have n or more citations.20

Results

Journals and publication dates of the top-cited articles on ET

The 100 top-cited articles (Table 1) were published in 26 journals. Journal title, impact factor, number of articles, and cited half-life are listed in Table 2. Almost half of the articles were retrieved from two journals: Neurology (n = 31) and Movement Disorders (n = 18). The 100 top-cited articles on ET were published from 1960 to 2010 (Figure 1). The greatest number of top-cited articles (n = 10) were published in 2000, and eight of the top 10-cited articles were published in the 1990s (Table 1) (Figure 1).18,21–29 Ninety-one of the articles were published within the past 25 years (Figure 1).

FIGURE 1

TABLE 2

Journals that Published the 100 Top-Cited ET Articles

Impact Factor
Rank	Journal	2011	5-Year Impact Factor	Number of Articles	Cited Half-Life
1	Neurology	8.312	7.634	31	8.7
2	Movement Disorders	4.505	4.449	18	5.3
3	Annals of Neurology	11.089	10.644	7	9.0
4	Brain	9.457	10.545	7	7.8
5	Archives of Neurology	7.584	6.928	7	8.9
6	Journal of Neurology Neurosurgery and Psychiatry	4.764	4.953	6	>10.0
7	Lancet	38.278	33.797	3	8.9
8	Nature Genetics	35.532	33.096	2	6.8
9	Acta Neurochirurgica	1.520	1.839	2	9.8
10	New England Journal of Medicine	53.298	50.075	1	7.8
11	Lancet Neurology	23.462	20.598	1	3.8
12	Journal of Clinical Investigation	13.069	15.430	1	9.4
13	Proceedings of the National Academy of Sciences of the United States of America	9.681	10.472	1	7.8
14	Nature Clinical Practice Neurology	7.636	6.712	1	4.4
15	Acta Psychiatrica Scandinavica	4.220	4.299	1	>10.0
16	Parkinsonism & Related Disorders	3.795	3.173	1	3.9
17	European Journal of Neurology	3.692	3.239	1	4.5
18	Journal of Neurology	3.473	3.203	1	6.5
19	Journal of Neurosurgery	2.965	3.088	1	>10.0
20	Journal of Neural Transmission	2.730	2.662	1	6.1
21	Acta Neurologica Scandinavica	2.469	2.347	1	9.8
22	Journal Neurological Sciences	2.353	2.441	1	8.7
23	Neuroepidemiology	2.305	2.817	1	6.3
24	Neuroscience Letters	2.105	2.168	1	8.3
25	Stereotactic and Functional Neurosurgery	1.849	1.913	1	9.4
26	Canadian Journal of Neurological Sciences	0.968	1.177	1	112

Sixty-two authors contributed two or more articles to the list, and five contributed 10 or more (Table 3). The number of authors per article ranged from one to 30, with the most common figures being five (16 articles), three (15 articles), two (12 articles), and seven (12 articles). The most frequent first authors of the top-cited articles in ET were Elan D. Louis (n = 27) and William C. Koller (n = 15); when combined, they accounted for 42% of the articles on the list (Table 3).

TABLE 3

Authors with Two or More Top-Cited ET Articles

Rank	Author	Number of Citation Classics	Number of Articles		h-index
			As First Author	As Co-Author
1	Louis ED	27	17	10	48
2	Koller WC	15	4	11	58
3	Jankovic J	12	2	10	83
4	Pahwa R	11	3	8	46
5	Lyons K	11	0	11	21
6	Findley LJ	9	3	6	36
7	Rajput AH	8	5	4	40
8	Rajput A	8	2	6	37
9	Benito-León J	7	6	1	29
10	Elble R	6	2	4	19
11	Deuschl G	6	2	4	56
12	Bermejo-Pareja F	6	1	5	20
13	Busenbark K	5	1	4	16
14	Ford B	5	0	5	35
15	Vonsattel JP	5	0	5	34
16	Ondo W	4	3	1	20
17	Tröster AI	4	1	3	37
18	Honig LS	4	0	4	28
19	Wilkinson S	4	0	4	17
20	Wenzelburger R	4	0	4	16
21	Thompson PD	4	0	4	54
22	Ottman R	4	0	4	41
23	Hubble JP	4	0	4	27
24	Brooks DJ	4	0	4	87
25	Bain PG	3	2	1	25
26	Raethjen J	3	1	2	17
27	Speelman JD	3	0	3	32
28	Ross GW	3	0	3	39
29	Pullman S	3	0	3	12
30	Marsden CD	3	0	3	126
31	Jurewicz EC	3	0	3	10
32	Hariz MI	2	2	0	31
33	Connor GS	2	1	1	3
34	Colebatch JG	2	1	1	38
35	Benabid AL	2	1	1	60
36	Stolze H	2	1	1	19
37	Deng H	2	1	1	22
38	Wills AJ	2	1	1	15
39	Alesch F	2	1	1	18
40	Hauser W	2	0	2	23
41	Weiner WJ	2	0	2	45
42	Tang MX	2	0	2	50
43	Stefansson K	2	0	2	24
44	Schwartz K	2	0	2	19
45	Rozdilsky B	2	0	2	29
46	Robinson CA	2	0	2	8
47	Overman J	2	0	2	10
48	Oertel WH	2	0	2	55
49	Nash J	2	0	2	14
50	Moskowitz CB	2	0	2	14
51	Lee H	2	0	2	31
52	Kong A	2	0	2	6
53	Jenkins IH	2	0	2	16
54	Jakobsson F	2	0	2	6
55	Higgins JJ	2	0	2	18
56	Benedickz J	2	0	2	1
57	Frackowiak RSJ	2	0	2	76
58	Faust PL	2	0	2	12
59	Cote L	2	0	2	33
60	Borden S	2	0	2	4
61	Barnes LF	2	0	2	7
62	Andrews H	2	0	2	26

According to their countries of origin (based on the first author), more than half of the articles were generated in the United States (n = 54), which led the list (Table 4). The United Kingdom (n = 11) was the second-most productive country, followed by Spain (n = 8), Germany (n = 7), and Canada (n = 4). France and Sweden each contributed three articles to the list. The 13 leading institutions that provided two or more top-cited ET articles are listed in Table 5. Columbia University in the United States produced the largest number of top-cited ET articles (n = 19), followed by the University of Kansas (n = 18), and Baylor College of Medicine (n = 9), both in the United States, and Móstoles University Hospital in Spain (n = 6).

TABLE 4

Country of Origin of the 100 Top-Cited ET Articles

Rank	Country of Origin	Number of Articles
1	USA	54
2	United Kingdom	11
3	Spain	8
4	Germany	7
5	Canada	4
6	France	3
7	Sweden	3
8	Austria	2
9	Italy	2
10	Iceland	2
11	Finland	1
12	Belgium	1
13	Turkey	1
14	Papua New Guinea	1

TABLE 5

Institution of Origin of Authors with Two or More Top-Cited ET Articles

Rank	Country of Origin	Number of Articles
1	Columbia University, New York, New York, USA	19
2	University of Kansas Medical Center, Kansas City, Kansas, USA	18
3	Baylor College of Medicine, Houston, Texas, USA	9
4	Móstoles University Hospital, Móstoles, Madrid, Spain	6
5	University of Saskatchewan, Saskatoon, Saskatchewan, Canada	5
6	Hammersmith Hospital, London, United Kingdom	4
7	Christian-Albrechts-University of Kiel, Kiel, Germany	4
8	Oldchurch Hospital, Romford, United Kingdom	3
9	Institute of Neurology, London, United Kingdom	2
10	deCODE Genetics, Reykjavik, Iceland	2
11	Southern Illinois University School of Medicine, Springfield, Illinois, USA	2
12	University of Vienna, Austria	2
13	National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland, USA	2

Among the 100 top-cited ET articles, a large number had a case-series design (n = 34), and the next most common design was the case-control study (n = 24) (Table 6). Only six articles were clinical trials (Table 6). According to the study designs, the level of evidence of articles is presented in Table 7. Most articles were classified as level B evidence.

TABLE 6

Study Design of the Clinical Studies

Study Design	Number of Articles
Experimental study
Randomized, controlled trial, open-label trial	6
Observational study
Cohort study	4
Case-control study	24
Descriptive study
Cross-sectional study	10
Case series	34
Case report	2
Methods paper	1
Review articles or expert opinion	18
Animal research	1

TABLE 7

Level of Evidence of the Clinical Studies

Levels of Evidence	Number of Articles
Level A: Well-conducted RCT with 100 patients or more (including multi-center and meta-analyses); well-conducted RCT with fewer than 100 patients (one or more institutions and meta-analysis; well-conducted study)	1
Level B: Well-conducted case-control study, poorly controlled or uncontrolled (including RCT with one or more major or three or more minor methodological flaws), observations studies with high potential for bias (case series with comparison to historical controls), case series or case reports, conflicting evidence with more support	79
Level C: Expert opinion	18

We also performed an analysis in which we excluded any examples of self-citations (i.e., instances in which authors cited their own work), and the results of this analysis did not change the main findings of the study (e.g., country of origin, institution of origin, top cited authors, etc.).

Discussion

In medical literature, the study of the number of times authors reference an article is one measure of the influence of the publication, and this type of citation analysis is widespread.1–3 The evaluation of specialty-wide citation analysis has been reported in other areas of the neurosciences.5–11 By ranking the 100 most-cited works, we aimed to determine which published journal articles on ET have exerted the most citation influence. The top article was cited 293 times. This figure is far lower than that found in Parkinson’s disease, in which the top-cited article garnered 4,327 citations.7

In 2002, Callaham et al.30 found that the impact factor of the original publishing journal, rather than the methodology or quality of the research, was an effective predictor of the number of times an article would be cited per year. However, as shown in the present study (Table 2), the actual citation value of the individual article did not positively relate to the impact factor of the journal.

In general, as shown in Table 3, the 100 top-cited articles in ET were articles that have been available for 10 or more years, and only one target article was published more recently (in 2010). It has been reported that scientific articles begin to be cited 1 or 2 years after publication, and reach a maximum citation rate 7 to 10 years after publication.31 However, an interval of 10 to 20 years is needed for maximal recognition of prominent articles in a field.32,33 This may explain why recently published articles were cited rarely, and few appeared on the list.

Among the 100 top-cited articles on ET, the most common design was the case series design (n = 34), followed by the case-control study (n = 24) (Table 6), implying that descriptive and observational studies are most frequent for ET. This suggests the relative ease of carrying out simpler study designs in ET. It is known that different study designs could correspond to different levels of evidence. In the hierarchy of research study designs, systematic reviews, meta-analyses, and well-conducted randomized clinical trials (RCTs) provide the highest quality of evidence for most clinical or interventional questions, and the lowest grade is applied to expert opinions. Among the 100 target articles, there were only six RCTs. Moreover, only one RCT was found among the top 20 medical articles. This is a common finding among the different studies that have analyzed the 100 most-cited papers in different disciplines.7,31,34–36 There are several possible explanations for the low numbers of RCTs. First, RCTs are expensive and time consuming. Second, it is difficult to gather large sample and control groups. Third, it may be that RCTs were published relatively recently, so they had not yet reached a significant number of citations.

Compared to Parkinson’s disease, there are very few articles addressing more basic science issues in ET (e.g., laboratory studies including gene discovery, molecular biology, and cellular biology; neuroimaging studies; neuropathological studies; and animal model studies).7 Scientific understanding of ET is in its infancy.

This study had a number of potential limitations. First and foremost, this survey was restricted to journal articles with the term “essential tremor” in the title. In other words, some significant ET articles could have been excluded, and it is possible that these were high-quality articles with a large number of citations. Second, this type of citation analysis does not include citations in textbooks and lectures, and an author’s or authors’ potential preference to cite articles in the journal in which they seek to publish their work.37 Third, there is a clear time effect in citation analysis, with the most recent articles being at a disadvantage.30 The time from publication played an apparent role here, with late 1990s and early 2000s being the most prominent years of publication for the 100 most-cited articles. Recent articles are clearly disadvantaged in citation analysis. Fourth, the language of publication plays a major role, with an obvious bias for articles published in English-language journals. Fifth, there are biases inherent in the age of the database, as any articles published in the 19th or early 20th century would be excluded, which likely causes some true “classic” articles to be excluded. Finally, Kuhnian philosophy38 would tell us that in a scientific community there is a tendency for adherence to a paradigm. In this context, there is “snowball effect” for citations, as other authors are more likely to cite an article because of previous citations to that article rather than its content or quality.