Introduction

Parkinson’s disease (PD) is the second-most prevalent neurodegenerative disease which characterized by the progressive loss of dopaminergic neurons in the substantia nigra pars compacta [1, 2]. It is a major contributor to worsened medical outcomes, poor quality of life, disability, and nursing home placement [3, 4]. Medical researchers have increasingly explored the mechanisms, early recognition, and prevention of PD. However, the PD research literature has not been analyzed to assess the quality of scientific insights in this area.

Citation analysis is an important method for determining the influence of an article on scientific progress as well as evaluating the impact factor (IF) of a scientific journal [5]. The study of citation analysis may help to identify articles, research topics, and authors of influence.[6]. Attempts have been made to identify the most cited articles in many fields, including neurosurgery,[7], traumatic brain injury [8], orthopedics [9], radiology [10], and surgery [11]. However, no citation analysis of PD has been published. Therefore, we analyzed and characterized the 100 most-cited articles on PD, to obtain an indication of the most impactful advances, developments, and discoveries in this field during the past century.

Methods

In September 2017, we performed a citation search on the bibliometric database ISI Web of Science (Philadelphia, PA, USA) from 1900 to 2017 for articles pertaining to PD. The search used the key term “Parkinson*” in the title (*as a wild card character used in search string). Articles on the list were reviewed by two independent reviewers by reading the abstracts. The full texts were acquired from PubMed, EMBASE, or ScienceDirect when necessary. Only studies focused on PD as the main topic and published in English were included. Articles were excluded if they were not pertinent.

We identified and selected the 100 most-cited articles related to PD, which were then manually reviewed by two independent investigators using the modified approach of the methods by Azer [12] and Lim et al. [13]. We compiled the information on the journal name, citation number, IF, title, number of authors, authorship (first, second, and corresponding authors), publication year, and country of origin of each article. If the authors were from multiple countries, the country of origin was deemed to be that of the corresponding author.

Statistical analysis

Data were described by the median or interquartile range. The Wilcoxon rank sum test and Spearman test were used to evaluate different indicators. All data were analyzed with SPSS V.17 software (SPSS, Chicago, IL, USA). All probability values were two-tailed, and statistical significance was defined as p < 0.05.

Results

Citation count and publication year

A total of 75,913 articles on PD were identified in the Web of Science core database after the initial search in the period from 1900 to present. We selected the 100 most frequently cited articles and ranked them according to the number of citations listed in Table 1. Citation numbers ranged from 669 to 6902, with a median number of 944. Among them, one article was cited more than 5000 times and 44 articles were cited more than 1000 times. The citation index (median 58, range 22–250) was correlated with number of the citations per article (r = 0.690, p = 0.000).

Table 1 The top 100 cited articles on Parkinson’s disease
Full size table

The 100 most-cited articles were published between 1967 and 2010, and 51% of these papers were published between 1996 and 2005. Three articles were published before 1970 and none of the articles was produced after 2010. The most productive period was 2001 to 2005, during which 27 articles were published. However, when calculating the mean citation number for each article, those published in 1966 to 1970 had the largest citation number (n = 3023) (Fig. 1). The single year with the most cited articles was 2003 (n = 10). The number of citations was highest for the period 2001 to 2005 (31,937). The Spearman test indicated an uptrend between the citation index and time (r = 0.649, p < 0.001). There is low correlation between time and number of citations (r = − 0.299, p = 0.03), but there was a positive correlation between time and citation index (r = 0.375, p < 0.001).

Fig. 1
figure 1

Numbers and number of citations of top 100 cited articles in 5-year periods

Full size image

Journals publishing the top 100 articles

The 100 most-cited articles on PD were published in 31 different journals. All the journals with more than one article are listed in Supplemental Material 1; these were predominantly comprehensive medical publications, led by the Science with 15 articles, followed by the New England Journal of Medicine (NEJM) with 13. In addition, Nature, Proceedings of the National Academy of Sciences of the United States of America, and the Annals of Neurology contributed seven, seven, and six of the most cited articles, respectively. The IFs of the 100 most-cited articles ranged from 4.083 to 72.406. Half of the top 100 articles (50 articles) were published in the high-IF journals (IF > 20). The journal IF was significantly correlated with the number of top 100 articles (r = 0.645, p = 0.005) and a low correlation with the number of citations (r = 0.455, p = 0.067).

Authorship, origins and institutions

A majority (84%) of the top 100 articles were produced by 3 or more authors. A list of the most frequently appearing authors is presented in Supplemental Material 2. With regard to individual contributions, A.J. Lees was the most frequently cited author, with listings on 8 of the top 100 articles (as first author, 1; as corresponding author, 3) and a total of 12,132 citations. Dr. Lees was followed by P. Jenner and A.E. Lang, both of whom authored 6 of the top 100 articles, with 5425 and 5340 citations, respectively (Supplemental Material 2). A total of 14 countries contributed to the top 100 articles. As expected, the USA was the most productive country with 44 publications, followed by England (17), France (9), Germany (7), and Canada (5 articles), with all other countries contributing less than 5 publications, as shown in Fig. 2. Articles originating from the USA also had the highest number of citations (total = 57,234). Of the top 100 articles, 8 institutions provided three or more articles. Among them, the leading institutions were University of London with 13 articles, followed by the U.S. National Institutes of Health (NIH) with 6 articles and Harvard University with 5 articles (Supplemental Material 3).

Fig. 2
figure 2

The top 100 articles were analyzed in terms of their country of origin. a Number of articles from each country. b Number of citations for the articles from each country

Full size image

Publication type

Among the 100 most-cited articles, there were 24 clinical studies, 54 laboratory studies, 20 reviews, and 2 clinical guidelines (Fig. 3). The number of total citations per article ranged from 688 to 6902 (median, 869) for clinical studies and from 670 to 4873 (median, 1007) for laboratory studies. Of the 24 clinical articles, surgical therapies were addressed by 10 articles, medical therapies by 5, clinical function by 4, clinical genetics by 2, case reports by 2 articles, and clinical staging by 1 article. In particular, the surgical studies focused on deep brain stimulation (five studies), cellular transplantation (three studies), pallidotomy (one study), and intraputaminal delivery of glial cell-derived neurotrophic factor (GDNF) (one study). All the reports of medical therapy were published in high-IF journals, four in the NEJM and one in Science. We found that 8 of the 10 surgical research papers were published in high-IF journals: 5 in the NEJM and 1 each in the Lancet, Science, and Nature Medicine. The earliest study included in the top 100 cited articles on medical therapy was published in 1967, while the earliest study on surgical therapy was published in 1990. Furthermore, the studies of medical therapy had higher median citations per article than the surgical studies (median 964 vs. 856). In the laboratory studies, 18 articles used animal models and 15 articles addressed brain pathology, 11 articles focused on identifying genetic mutations, and 10 articles evaluated the cellular and molecular biology of PD. We also found that 52% of the laboratory studies appeared in the NEJM, Lancet, Science, or one of the Nature journals (Nature, Nature Genetics, Nature Medicine, and Nature Neuroscience). Among them, the articles about brain pathology were most frequently cited (n = 22,475), followed by articles addressing genetic mutations (n = 21,200).

Fig. 3
figure 3

Distributions of research type of the top 100 cited articles on Parkinson’s disease

Full size image

Discussion

PD is the second most common neurodegenerative disease, affecting approximately 1% of the population over the age of 60 and 4% over 80 [14]. The current study is the first to assess the characteristics of the 100 most-cited articles in the field of PD, and it allowed us to recognize historical patterns and trends in PD research, which has undergone considerable change in recent years. The results may facilitate recognition of important advances and prevalent areas of research interest in PD and may help basic scientists and clinicians design future research. The current study helps to identify classical research and high-impact journals by providing information regarding authors, institutions, and journals.

Citation analysis is a useful bibliometric method, introduced in 1987, that has been widely used in various fields and has proven to be important for both authors and journals. For authors, citation analysis not only helps them to recognize important research progress but also helps add useful perspectives on historical developments in areas of academic interest. For journals, data from citation analyses may attract manuscripts with higher citation potentials. In addition, citation analysis may, to some extent, help researchers produce better work. However, we must recognize that since the citation number for any given paper is strongly influenced by the prominence of the journal of publication, it may not have a great relationship with the scientific merit of the manuscript [15, 16], and at the same time, the citation number could also be affected by factors such as the author’s geographical origin, language, and gender [17, 18]. Although there are some disadvantages with evaluating the quality of the article by its citation rating, it is still the most widely accepted current method to determine the merits of a paper or journal [19]. At present, more and more articles are labeled as “top cited” or “the most cited” in various medical disciplines, but there is a paucity of literature on citations of articles about PD.

PD, the incidence of which gradually increases with age, is a major global health problem that is associated with increased medical costs and thus places a heavy burden on some communities. In recent years, there have been significant changes in strategies for the prevention, diagnosis, and treatment of PD. Therefore, there is an urgent need to find appropriate directions for research as well as to better design future studies. This may be accomplished through analyzing classic articles to better understand the history and development of PD research. The current study is the first to assess the leading article citations in PD research and will contribute to the ability of authors or readers to recognize the quality of the research reports, identify the key discoveries that have been made as a result of past efforts, and illuminate developing trends in scientific research.

It is well established that the number of citations an article garners is affected by the date of publication [20]. The longer the time since the article was published, the greater the chance of being cited. However, unlike the majority of citation analyses which report peaks between 1980 and 1995, in this analysis, the most productive period was 1996 to 2005, which may be partially accounted for by an increase in numbers of articles and improvements in research quality. In order to overcome the impact of the publication time on the likelihood of citation, we also assessed the citation index as a measure of the true impact of an article independent of short-term trends. The results were consistent with an increase of articles and improvements in research quality.

High-IF journals attract submissions from authors because they not only potentially provide prominence for the research results but also help the author get more attention. The IF of a journal is the most important predicator for citations, and most top-cited articles are published in high-IF journals [21, 22]. This study also confirmed that the IF was positively related to the top 100 cited articles and the number of citations. When taking into account the 5-year IF of the journals, 52 articles were published in the high-IF journals (IF > 20), 23 articles were published in journals with IF between 10 and 20, and only 12 articles were published in the low-IF journals (IF < 5). In the clinical realm, most articles on medical and surgical therapies were published in the highest impact general medical journal (the NEJM). And most animal and basic laboratory studies were published in the high-IF science journals Nature and Science. These results further validate the hypothesis that researchers tend to cite papers from a few core journals in their specialty [23]. In addition, the Proceedings of the National Academy of Sciences of the United States of America and the Annals of Neurology were the sources for several articles in this study.

Fourteen countries contributed to the 100 most-cited articles. The USA ranked first, similar to the citation analyses in other specialties [24,25,26]. This finding confirms the important influence of the USA in the study of PD worldwide, which can be explained by the large scientific community and enormous financial resources available to it. Among the top 16 institutions, 11 (69%) are in the USA. Moreover, although authors usually prefer to publish in their local journals, authors in American and European countries tended to publish in American journals [27]. However, the leading institution for publications in PD is the University College of London, which published 13 of the 100 most-cited articles with 17,048 total citations. The most frequently cited authors, A.J. Lees and P. Jenner, were both from the University College of London and had 8 and 6 articles that were on the list, respectively, that reported results of laboratory studies. In addition, we found that no authors from Africa, Oceania, or South America contributed to the 100 most-cited works, which may be related to information access, difficulties in research, and language barriers in these areas, indicating a great disparity in scientific publications between the developing and developed regions of the world.

PD was first described in 1817, and although significant efforts have been made during the past several decades, the pathogenesis of PD remains unclear. Therefore, the study of the pathogenesis of PD is still an area of active interest, which is consistent with our findings that most of the top 100 cited articles (54%) in the present study report the results of basic research. The median citation number per basic research article was higher than that of clinical research articles (1007 vs. 869). Among laboratory studies, the most frequent types were descriptive animal models (n = 18), followed by neuropathological studies (n = 15), characterizing genetics (n = 11), and evaluations of cellular and molecular biology (n = 10).

Research directions constantly change with time. In the 1970s, research was mainly focused on motor fluctuations from l-dopa therapy, assessment of secondary PD, and attempts to further characterize PD. In the 1980s, in addition to the side effects of levodopa treatment, studies addressed cognitive dysfunction, surgical transplantation of stem cells, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) models, and the role of oxidative degeneration in the pathogenesis of PD. In the 1990s, attention focused on surgical treatments such as pallidotomy, deep brain stimulation, and stem cell transplantation, as well as the role of gene mutations and synuclein in the pathogenesis of disease. After 2000, the studies tended to focus on the efficacy of deep brain stimulation (DBS), the therapeutic effects of stem cells and neurotrophic factors, and other gene mutations such as leucine-rich repeat kinase 2 (LRRK2), DJ-1 (protein deglycase, also known as Parkinson disease protein 7), and PTEN-induced putative kinase 1 (PINK1) that can also lead to PD.

Limitations

Although we attempted to rule out all possible design flaws, there are still some limitations in the current study. First, we applied a direct and reproducible approach that clearly limits the references to those with “Parkinson’s disease” in the title of the article, which may lead to omission of some publications related to the disease. For instance, papers that referenced “Parkinson’s disease” in the abstract or keywords, but not in the title, were omitted from our study. Second, despite a meticulous search of the Web of Science, citation information can also be obtained from Google Scholar and Scopus, which may show different results [28, 29] and lead to some research reports being missed. Third, this kind of study of IFs favors earlier published articles, often excluding newly published high-quality studies that have not had the opportunity yet to gain sufficient citations. Fourth, the number of citations alone cannot fully quantify the value of the contribution to the field and may miss important, influential but less frequently cited papers. Fifth, the effect of self-citation and citations of irrelevant articles can also increase the overall citation rate, and this possibility was not addressed in the design of our study. Finally, the language of publication was restricted to English, which would also generate bias.

Conclusions

The current citation analysis dealt with most of the influential studies on PD and presented a detailed list that will change dynamically as the field moves forward. Our analysis has collected a number of highly influential articles from a variety of perspectives, including medical and surgical treatment, basic research, clinical research, and characterization and classification of the disease using pathological methods, and highlights the research trends and academic achievements. Our analysis also provides an insight into the frequency of citations on PD and reveals the quality of research, discoveries, and trends steering PD research worldwide.