Research Trends and Foci in Osteoarthritis Pain from 2012 to 2022: Bibliometric and Visualization Study

Introduction

Osteoarthritis (OA) is a common disease which leads to disability by causing multiple joint injury including knee, hand and hip.¹ Systematic analysis has demonstrated a growing number of patients diagnosed with OA, with its incidence increasing over the past few decades.² In certain regions, OA affects more than 10% of the population,^3,4 causing significant suffering for patients and imposing a substantial economic burden on families and governments.

OA pain, the most common symptom of OA, has attracted researchers’ attention in recent years. OA is considered as a whole-joint⁵ disease, cartilage, subchondral bone, and synovium probably all have key roles in disease pathogenesis.⁶ Pain in OA may be due to local and central sensitization of pain pathways causing normal stimuli to be perceived as painful.⁷ Periosteum, subchondral bone, periosteum, ligaments, synovium, and the joint capsule are all densely innervated and contain nerve endings that could be the potential source of nociceptive stimuli in OA.⁸

Risk factors of OA pain have been emphasized. A multiple logistic regression suggests that body mass index (BMI), increasing age, suffering years of OA and many other factors can be used to predict the flare of knee OA pain.⁹ Also, some of the risk factors mentioned above are crisis of OA itself,¹⁰ which indicates the pain can hardly be avoided when OA occurs. These risk factors generated OA pain by different mechanisms. Studies have shown that sources of OA pain can be nociceptive, inflammatory and neurogenic,¹¹ and concrete mechanisms include the sensitization of afferent neurons in OA joint by nerve growth factor, cytokine and neuronal injury.^11,12 Clinical phenotypes of OA pain also gained increasing attraction as it can result in different health outcomes.¹³ Different subtypes of OA pain have been classified by various clinical indicators, as well as diverse respond to treatment,^14,15 while exploration of phenotypes and their clinical significance is still not enough.

In order to alleviate the pain and disability of joint function caused by OA, diverse therapies have been carried out. Pharmacotherapy occupies the most important position, demonstrated by the extensive utilization of nonsteroidal anti-inflammatory drugs (NSAIDs), acetaminophen and opioids.¹⁶ However, traditional drugs gradually show their shortages, such as gastrointestinal complications, reduced liver function and headache, which activates the generation of novel pharmacotherapy.¹⁷ Antibodies against neuron growth factor and inflammatory cytokines prevent the activation of OA pain, while ion channels appear on the stage as the driving effect of hyperexcitability of neuron.^18,19 Besides, surgical treatment such as total knee replacement shows positive effects, and more desirable outcomes are seen when it is combined with non-surgical therapies.²⁰ Osteotomy and cooled ratio frequency ablation are also indicated to perform expectable efficacy.^21,22 Furthermore, clinical trials tell us that physical activity and massage can lead to recovery of body function and anesis of OA pain.^23,24

In consideration of the current status of OA pain research, a growing number of scientists have found that bibliometrics can be applied in this field. However, most of present bibliometric analysis lay emphasis on the pain management and recovery instead of the overall research status of OA pain.^25–27 In order to offer practical guidance to researchers who focused on OA pain research, this study conducted a bibliometric and visualization analysis in this field.

Methods

Data Source and Search Strategy

Existing literature data were collected from the Science Citation Index Expanded (SCI-expanded) of Web of Science (WOS) Core Collection database on 31 December 2022. The publications time of the published articles ranged from 2012 to 2022. The search strategy was as follow: [TI = (Osteoarthritis OR OA) AND (pain OR painful)] AND [PY= (2012-01-01 to 2022-12-31)] (TI: title, PY: Publication Year).

Inclusion Criteria and Exclusion Criteria

Inclusion Criteria and Exclusion Criteria is represented in Figure 1. English is the chosen language, and four types of publications were included: articles, review articles, editorial materials and letters. A total of 4054 articles were identified from the Web of Science Core Collection. Meeting abstracts, corrections, proceedings papers and new items were excluded. Finally, after excluding 1561 articles, 2493 articles were included.

Figure 1 Flow diagram of search strategy.

Data Extraction

We obtained bibliometric indicators by calculating the number of publications, journals, references citations, extracted H-index and keywords. The H-index means that a journal or researcher has at least H published papers that are cited at least H times per paper. Former indicators were produced by for kinds of software in this study:

1. Microsoft Excel 2019 for calculating the frequencies and percentage of the published materials;
2. VOSviewer (version 1.6.17) for the bibliometric networks;
3. R4.2.1 and bibliometrix package was used to calculate the citations metrics;²⁸
4. CiteSpace (version 5.1.R6) for Keywords citation burst map.

Results

Description of Retrieved Published Articles

A total of 2493 articles were identified from the WOS database, including 2053 Article-type articles, 258 Review-type articles, along with 97 editorial materials, 85 letters (Table 1).

Table 1 Types of Retrieved Documents (2012-01-01 to 2022-12-31)

Temporal Trends of Citations and Publications

According to Figure 2, the overall trend of the number of productions has continued to increase. The number of publications which peaked in 2021 (343) has become nearly three times more than the number of them in 2012 (119). The number of citations kept reducing during the investigated years, except for an uptrend in 2018, and average citations showed the largest number was 51 in 2012, while Figure 3 shows that the total average citations per article is 19. The specific number of annual scientific productions and citations are shown in Table 2. The results of Pearson correlation analysis from RStudio indicated that the number of publications showed positive correlation with year, with a Pearson correlation coefficient of 0.483, and the number of citations showed passive correlation with year, with a Pearson correlation coefficient of −0.922 (p < 0.01). Average citations per year showed obvious passive correlation with year (−0.989), number of publications (−0.537) and positive correlation with number of citations (0.911).

Table 2 Annual Scientific Productions and Citations

Figure 2 Correlation analysis between citation and publication metric. ***p<0.01.

Figure 3 Articles–citations histogram.

Productive Authors and Co-Authorship

A total of 11,041 authors were contained in the documents included in this study. Table 3 describes the Top 10 authors distributed by publications and citations. Fillingim, Roger B. (University of Florida, USA) and Bennell, Kim L. (The University of Melbourne, Australia) both published 32 articles, and they tied for the first place ranked by publications in OA pain field. Fillingim, Roger B. completed 1 of 32 articles as the first author, while Bennell, Kim L. completed 9 of the 30 articles as the first author. Actually, Bennell, Kim L had the highest Dominance Factor (DF) among the top 10 authors. Authors distributed by citations showed difference compared with authors distributed by publications, Neogi, T. with a total of 1049 citations were the first in the rank. Malfait, Anne-Marie (Rush University, USA) with a total of 984 citations was in the second place but had higher H-index than Neogi, T. Bennell, Kim L. with a total of 699 citations shared the highest H-index (13) with Malfait, Anne-Marie. VOSviewer software was used to analyze the co-authorship and construct a visual map to show influential researchers and potential collaborators (Figure 4), the larger the author’s tag, the more articles the author has published. The thickness of the lines between each author is proportional to the number of times they appear together in the same document. The thicker the lines between tags, the more times they appear in the same item.

Table 3 Top 10 Authors Distributed by Publications and Citations

Figure 4 Network visualization map of co-authorship in osteoarthritis pain research.

Geographical Distribution of Publications and Citations

A total of 82 countries was included in this study, research output was mainly concentrated in the United States, England, China, Australia and a few other countries. The top 10 countries in terms of research output were respectively the USA (797 articles, 20,727 citations), England (310 articles, 9412 citations), China (281 articles, 3547 citations), Australia (254 articles, 5884 citations), Canada (186 articles, 4664 citations), Japan (140 articles, 1952 citations), Netherlands (128 articles, 2560 citations), Denmark (105 articles, 3288 citations), Spain (98 articles, 2494 citations), Italy (97 articles, 1691 citations). All these results are available in Table 4. Countries such as Denmark (31 average citations) and England (30 average citations) seemed more popular in OA pain research field. The USA, as the leading authority in this field, was the country that conducts the most scientific cooperation with other countries (Figure 5). Countries of corresponding authors are ranked by RStudio, which identified articles as SCP as long as their corresponding authors are from one country. Following this principle, Korea had almost no cooperative relationship with other countries (Figure 6). Although Japan ranked eighth in publication, Japanese articles receive fewer citations and do not seem to receive much attention from other countries worldwide.

Table 4 Top 10 Countries Distributed by Publications and Citations

Figure 5 Network visualization map of international collaboration among countries.

Figure 6 Publication distributed by single-country publications (SCP) and multiple-country publications (MCP). Abbreviations: SCP, single country publications; MCP, multiple country publications.

Keywords Analysis

Research Foci Summarized from Keywords

The frequency of a keyword’s occurrence makes a positive correlation with the degree which the keyword is of interest to researchers. The more frequently a keyword appears, the more likely it is to be a research hotspot in the field. VOSviewer was used to calculate keywords from 2493 retrieved documents, the total number of Author Keywords (DE) and Author Keywords-Plus (ID) was 3343 and 3560, respectively. In Author Keywords (DE) group, the keywords used at high frequency were: “Osteoarthritis”, “Pain”, “Knee Osteoarthritis”, “Knee”, “Chronic Pain”, “Knee Pain”, “Exercise”, “Meta-analysis”, “Inflammation”, “Quality of Life”. But in Author Keywords-Plus (ID) group, “Hip”, “Management”, “Knee Osteoarthritis”, “Prevalence”, “Arthritis”, “Efficacy”, “Older-adults”, “Double-blind”, “Association”, “Disability” occurred more frequently. The frequency results are presented in Table 5. To further demonstrate the hot keywords for each year from January 1, 2012 to December 31, 2022, the top 50 keywords citation burst map (Figure 7) was built to exhibit the annual hotspots.

Table 5 Keywords by Frequency

Figure 7 Top 50 keywords with the strongest citation bursts.

Subdisciplines Summarized from Keywords Co-Occurrence Map

Mapping with the VOSviewer technique of author keywords with minimum occurrences of 20 divided keywords into 4 clusters, which was expressed as circles in four different colors in Figure 8. Specifically, the red cluster concentrated on “Clinical feature, Diagnosis, Therapy and Rehabilitation”, which can be concluded by knee osteoarthritis, hip osteoarthritis, knee pain, hip pain, chronic pain, low back pain, gait analysis, quantitative sensory testing, physiotherapy. Keywords in yellow cluster, such as magnetic resonance imaging, mri, ultrasonography, ultrasound were related to “Imageological Examination”. Keywords in green cluster can be concluded as “Prognosis”, proved by keywords like quality of life, sleep, disability, obesity and depression. Keywords in blue clusters, such as monosodium iodoacetate, inflammation, nerve growth factor and hyaluronic acid were related to the topic “Pathogenesis”. Pain, Knee osteoarthritis, and osteoarthritis were the first three keywords which frequently appeared in early 2012–2021, but with the development of OA pain field, their subdisciplines became more and more (Figure 9).

Figure 8 Network visualization map of the author keywords.

Figure 9 Co-citation map (timeline view) of keywords from publications on osteoarthritis pain research field.

Preferred Journals

The top 10 most cited and productive journals in OA pain research field are listed in Table 6. Osteoarthritis and Cartilage with a total of 178 publications obtained Publication Rank 1, followed by Arthritis Care & Research with a total of 87 articles. The third was BMC Musculoskeletal Disorders with a total of 85 articles. All other journals on the list include fewer than 80 articles, the 2021 Journal Citation Reports indicated that only Osteoarthritis and Cartilage, Pain, Annals of the Rheumatic Disease, Arthritis Research Therapy were listed in Q1 Journals. In the top 10 most cited journals list, Osteoarthritis and Cartilage with a total of 5964 citations won the Rank 1, followed by Pain with a total 2367 citations, most of the journals on this list were from the Q1–Q2. It was necessary to point out that the impact factor (IF) of Arthritis and Rheumatism was from 2015 JCR, because the IF in subsequent years was unavailable. During the investigated period, Osteoarthritis and Cartilage was the first journal to include the largest number of articles in the field according to Figure 10, and continues to be a leader in subsequent years. Journal of Pain Research was relatively a newcomer to the field, building its way to the forefront of the list based on the number of articles in the field in recent years.

Table 6 Top 10 Journals Distributed by Publications and Citations

Figure 10 Visualization of top 10 prolific journals publications distribution.

Analysis of Top 10 Most Cited Articles

The top 10 most cited articles in OA pain research field are listed in Table 7. With a total of 810 citations, the article entitled “The epidemiology and impact of pain in osteoarthritis” released by Neogi, T in 2013 was most cited articles in OA pain research field, the 2021 Journal Citation Reports indicate that Osteoarthritis and Cartilage was a high-quality journal with 7.507 IF. The following article was “What proportion of patients report long-term pain after total hip or knee replacement for osteoarthritis? A systematic review of prospective studies in unselected patients” created by Beswick, AD in 2012 and obtained a total of 751 citations. The Rank 3 in the list was “Incidence, prevalence, costs, and impact on disability of common conditions requiring rehabilitation in the United States: stroke, spinal cord injury, traumatic brain injury, multiple sclerosis, osteoarthritis, rheumatoid arthritis, limb loss, and back pain” with a total of 533 citations. In this list, most of the documents were article-type articles except two review-type articles. The last five on the list were respectively “Impact of exercise type and dose on pain and disability in knee osteoarthritis”, “Effectiveness of non-steroidal anti-inflammatory drugs for the treatment of pain in knee and hip osteoarthritis: a network meta-analysis”, “Efficacy and safety of paracetamol for spinal pain and osteoarthritis: systematic review and meta-analysis of randomised placebo controlled trials”, “An efficient randomised, placebo-controlled clinical trial with the irreversible fatty acid amide hydrolase-1 inhibitor PF-04457845, which modulates endocannabinoids but fails to induce effective analgesia in patients with pain due to osteoarthritis of the knee”, “2016 Patellofemoral pain consensus statement from the 4th International Patellofemoral Pain Research Retreat, Manchester. Part 1: Terminology, definitions, clinical examination, natural history, patellofemoral osteoarthritis and patient-reported outcome measures”. The average number of citations for the top 10 most cited articles was 449.

Table 7 Top 10 Cited Literatures

Most Influential Institutions

As illustrated in Table 8, the top 10 most productive organizations distributed by publication were respectively Boston University (91 articles; 4050 citations), University of Sydney (87 articles; 2223 citations), The University of Melbourne (71 articles; 1710 citations), University of Nottingham (68 articles; 2330 citations), University of Florida (59 articles; 1153 citations), The University of Alabama at Birmingham (50 articles; 1356 citations), Duke University (55 articles; 1206 citations), University of Oxford (50 articles; 1034 citations), University of California San Francisco (44 articles; 1123 citations), Monash University (43 articles; 901 citations). Among them, Boston University, University of Nottingham and the University of Sydney were the top three institutions ranked by citations. Tufts University (23 publications, 1254 citations) had relatively low total citations among the listed 10 organizations, however, they had the highest two average citation (55). Organizations including Boston University, University of Sydney, University of Nottingham were the leading authority in this field, in addition, articles born from Boston Univ were preferred by more other institutions in this field (

Table 8 Top 10 Institutions Distributed by Publications and Citations

Figure 11).

Figure 11 Network visualization of institutions collaboration.

Discussion

Global Trends of the Research on OA Pain

This study presents a comprehensive overview by using bibliometric analysis to access the published articles on OA pain in the last decades. From January 1, 2012 to December 31, 2022, the overall trend of the number of published articles has continued to increase, and the fastest growth rate of investigated publications appeared from 2016 to 2021. The results showed that research in OA pain attracted wide attention from researchers. The growing number of existing literatures will provide a solid basis for the following exploration. The downtrend of number of citations can be attributed to the citation lag, the earlier the articles were published, the more cited chance they will get.

The majority of the top 10 most productive authors were from The United States, and only four authors were from other countries, reflecting that it was a dominated player in this field. Fillingim, Roger B. and Bennell, Kim L. were two most productive authors, while the latter had higher DF. In order to describe the most influential authors in this research field, the authors’ citations ranking was also taken in consideration. Despite the fact that Fillingim, Roger B. was the most productive author, his name was not seen in the citations rank, while Neogi, T. became the top candidate of the list with 1049 total citations. Besides, half of the top 10 productive authors distributed by citations were from the USA, suggesting that it was also a main contributor in OA pain research.

The United States had a dominative contribution to both publications and citations, and it was the country with the most cooperation with other others. Besides, England, China and Australia have published more than 200 articles in this research field. But these productive countries had fewer average citations compared with Denmark.

Osteoarthritis and Cartilage seemed to be the most influential journal in OA pain research field with both the largest number of publications and citations. Consistently, the most highly cited article was from Osteoarthritis and Cartilage. Arthritis Care & Research was the second productive journal and the third most cited journal. BMC Musculoskeletal Disorders also showed its advantage in publication, while Pain released less paper but was cited more frequently. All in all, after considering 2021 JCR, publications and citations, Osteoarthritis and Cartilage was still the authoritative journal in this field.

Boston University ranked first in both publication list and citation list, indicating it had outstanding contribution to this field. The University of Sydney also showed its advance by ranking second in publication list and third in citation list. University of Nottingham was also influential because of its fourth place in publication rank and second in citation rank. In addition, The University of Melbourne and Duke University were also seen in both lists. Compared to the study of Cheng et al, Monash University and University of California San Francisco have dropped out the top five, demonstrating the rising two institutions had greater efforts in more recent years.²⁵

Hotspots and Frontiers

Based on the top keywords and literature, we ascribe the research hotspots as follows:

Affected body part of OA pain: “Osteoarthritis” and “Pain” were the top two author keywords mainly because of the search term [(TI = [(Osteoarthritis OR OA)] AND [TI = (pain OR painful)]. They were followed by “Knee Osteoarthritis” and “Knee”, and these two words were also the only words whose frequency was beyond 100 after “Osteoarthritis” and “Pain”. In addition to give a general accession of keywords, Author Keywords-Plus were also referred. Both “Hip” ranked first and “Knee Osteoarthritis” ranked third in this list revealed these two body parts suffered from osteoarthritis most frequently. Hand is also a main pathogenic position but gets much less attention than knee and hip,³⁹ possibly because osteoarthritis in larger joints tends to cause more severe pain.

Rehabilitation of OA pain: According to the top 10 keywords, “Exercise” and “Rehabilitation” were active keywords in Author Keywords list, and “Management” ranked second in Keywords-Plus list, indicating that rehabilitation of OA pain had shown its importance. Keywords which related to rehabilitation formed green cluster in Figure 9. This opinion was also approved by the facts that one of the top 10 cited literature focused on this field. This article written by Juhl et al compared several exercise types and got the conclusion that single-type exercise was more effective in reducing pain and disability than mixing types of exercise.³⁴ In addition, there are also a large number of studies concentrating on the efficacy of exercise in total knee and hip arthroplasty.^40,41

Treatment of OA pain: It can also be seen from Table 7 that five of the publications were related to OA pain’s treatment. Kerbs et al verified opioids showed no special effect than nonopioid so that it was not an appropriate therapy for knee OA pain.³² In a network meta-analysis, diclofenac was considered to be the most effective non-steroid anti-inflammatory drug to deal with pain and disability in OA patients.³⁵ Another research indicated that intra-articular injection of triamcinolone acetonide was not suitable for patients with symptomatic knee Osteoarthritis.³³ Huggins, J. P. found that the irreversible fatty acid amide hydrolase-1 inhibitor PF-04457845 was invalid for knee osteoarthritis pain by clinical trials, which was opposed to the results obtained by animal models.³⁷ Teatment of paracetamol showed negative therapeutic effect as well.³⁶ According to the mentioned four most cited papers, we can draw a conclusion that clinical trial may be the main research method in treatment of OA pain, and the correction of traditional therapies occupy a major part. Physiotherapy for osteoarthritis pain is also receiving increasing attention, many researchers engaged in exploring new physiotherapy methods and evaluating the clinical efficacy of physiotherapy.^42,43

Strengths and Limitations

Various artificial intelligence software such as VOSviewer and CiteSpace were used for visualization of co-authorship, country and institutional collaboration, keywords co-occurrence and keywords citation burst map, it provides a practical method for global data analysis in international medical systems. Besides, we employed systematic searching and quantitative statistical analysis, our bibliometric study was significantly more intuitive and comprehensive than a literature review would have been. There are some defects in this study. VOSviewer cannot visualize keywords along with relevant time data in the same figure, resulting in deflective predictions in hotspots. The research term limited relevant words like “Osteoarthritis” and “pain” to title, which may lead to incomplete searching results to this study, and it also can be redundant if authors did not put the search terms in the article title. Our study only included literature indexed in WOS, which may result in potential inaccuracies when predicting research hotspots due to the absence of literature from other databases.

Conclusion

This study identified OA pain-related publications from January 1, 2012 to December 31, 2022 and presented their global trends and current status. The total publications of OA pain generally increasing over the time in the last decade, and the escalating rate remained a high level. Many researchers contributed to OA pain field, while Bennell, Kim L. and Neogi, Tuhina were two main contributors. The United States ranked first in total publications, total citations and international collaboration. Analysis of keywords showed rehabilitation and treatment of OA pain were the hotspots in this area, and the main force concentrated on hip osteoarthritis and knee osteoarthritis, which means hand osteoarthritis needs to be further studied. Osteoarthritis and Cartilage was the major publishing journal for OA pain exploration. “The epidemiology and impact of pain in osteoarthritis” and “What proportion of patients report long-term pain after total hip or knee replacement for osteoarthritis? A systematic review of prospective studies in unselected patients” were the most influential article and review in this field, respectively. Boston University made greatest contribution to OA pain research. Reference learning showed tight correlation between OA pain and central sensitization, hip and neuromuscular exercise, indicating that relative work might have a bright future.