Web of Science Core Collection (WoSCC) and Scopus are the main sources of citation data, with Scopus offering a larger journal coverage in all fields. The interdisciplinary coverage provided by these databases is highly valuable for the study and comparison of different scientific fields especially in the Natural Sciences, Engineering, and Biomedical Research domains (16). Therefore, the data for this study were obtained from both databases.

We searched WoSCC and Scopus from 2004 to June 23, 2022. The data search strategy included the topic “autophag* or mitophagy or autolysosome or er-phagy” and the topic “osteoarthrit* or osteoarthrosis”. Early access, meeting abstracts, letters, corrections, retractions, editorial materials, proceedings, publication with expression of concern, and retracted publications were excluded. Original and review articles reporting autophagy in OA were qualified for inclusion regardless of language. Our search identified 1142 records fulfilling the inclusion criteria: 581 papers from WoSCC and 561 from Scopus. After further screening the titles, abstracts, and full texts, 410 duplicates were excluded. Finally, 732 papers were included in this study. The search and selection process is shown in Figure 1 .

All included papers retrieved from WoSCC and Scopus were imported into Microsoft Excel, VOSviewer, and CiteSpace software for analyzing and visualizing. Microsoft Excel 2019 (Microsoft Corp, Redmond, Washington USA) was used to evaluate the number of papers published in each year and journal. CiteSpace is an interactive visualization tool, which combines information visualization methods, bibliometrics, and data mining algorithms to extract patterns from citation data (17). CiteSpace (V.6.1.R2) software (Chaomei Chen, Philadelphia, Pennsylvania, USA) was used to conduct visual analysis of the distribution of authors, institutions, countries/regions, keyword clusters, co-cited references, co-cited authors, co-cited journals, and timelines. VOSviewer (V.1.6.18) software (Leiden University, Leiden, the Netherlands) was used to provide density visualization.

A total of 732 papers (626 original and 87 review articles) published between May 2004 and June 2022 were included in this study. General information relating to the data used in this study is presented in Table 1 . From 2004 to 2022, the cumulative number of publications on autophagy in OA increased year by year ( Figure 2 ). From 2009 to 2021, the annual number publications increased steadily and reached its maximum in 2021, with a total of 172 outputs (23.50%). Between January and June 2022, the number of outputs reached 78 (10.66%), which was similar to the whole outputs in 2019. These trends indicate that autophagy in OA research is flourishing.

732 outputs published by 329 institutions from 55 countries/regions were included in this study. As shown in Table 2 , China produced the most publications (n=456, 62.30%), prior to the USA (n=115, 15.71%), South Korea (n=33, 4.51%), Japan (n=27, 3.69%), and Spain (n=25, 3.42%). Scripps Research Institute was the most productive publishing institution (n=26, 3.55%), followed by Zhejiang University (n=23, 3.14%), and Shanghai Jiao Tong University (n=22, 3.01%). Wenzhou Medical University, Nanjing Medical University, and China Medical University each had the same number of outputs (n=19, 2.60%).

Figure 3 shows the visual map of publications related to autophagy in the field of OA originating from different countries/regions and institutions. Each node represents a country or institution, and the size of the node is proportional to the number of outputs. Similarly, the number of citations is related to the size of each node’s label. The lines between the circles denote collaboration between countries or institutions; the wider the line, the stronger the collaboration. The centrality of nodes in the knowledge map is considered as a measure of the node significance and represents the interaction between nodes. Nodes with high centrality are outlined with a purple circle. In addition, the circles inside the nodes represent the number of publications produced by a country or institution in a given year, with different colors representing different years.

As shown in Figure 3A , the deeper the ring color, the lower the publication year; and the lighter the ring color, the more recent the publication. China ranked first in terms of the volume of recent publications on autophagy in OA. Papers published by the USA originated from the 2008–2022 period, while China published most of its outputs after 2012. Five nodes with high centrality, namely China (centrality=0.53), the USA (centrality=0.35), Spain (centrality=0.30), England (centrality=0.18), and Saudi Arabia (centrality=0.18) were identified. Thus, these countries are the major collaborators in the field of OA autophagy.

As shown in Figure 3B , colors vary from grey to red as time goes from 2010 to 2022. Scripps Research Institute was the most productive institution, but its centrality was relatively low (centrality=0.09). All the top 10 publishing institutions had low centrality.

In total, 444 authors have made contributions to this field. As shown in Table S1 , Martin Lotz had the most significant number of publications on autophagy in OA (n=30, 4.10%), ahead of Beatriz Caramés (n=13, 1.78%), and Bai Lunhao (n=10, 1.37%). It is worth noting that the centrality of authors was relatively low (≤0.01), suggesting a lack of influential authors in the field of OA autophagy. Co-cited authors are two or more authors simultaneously cited by another paper (18). Among 673 co-cited authors in this study, six had a citation frequency of over 100, and Caramés B was the most co-cited author (n=302), followed by Zhang Y (n=156) and Loeser RF (n=151). Among the co-cited authors, Aigner T had the highest centrality (n=45, centrality=0.16), followed by Blanco FJ (n=78, centrality=0.10) and Almonte-Becerril M (n=62, centrality=0.10).

The included 732 articles were published in 160 different journals. As shown in Table S2 , the highest number of outputs on autophagy in OA were published in Osteoarthritis and Cartilage (n=45, 6.15%), followed by International Journal Of Molecular Sciences (n=25, 3.42%) and Cell Death & Disease (n=15, 2.05%). Co-cited journals are two or more journals that are cited at the same time. Among the 539 co-cited journals, five were cited more than 300 times. Osteoarthritis and Cartilage was the top co-cited journal (n=487), followed by Arthritis Rheum-us (n=414), and Annals of Rheumatic Diseases (n=373).

Keywords are often used in publications to summarize the research topic, and their analysis can reveal research hotspots and directions in a specific field. The keywords (n≥70) related to autophagy in OA are shown in Table S3 . Among these keywords, autophagy occurred most frequently (n=296), followed by OA (n=263), apoptosis (n=189), cartilage (n=169), expression (n=153), and chondrocyte (n=123). Animal (n=66, centrality=0.21), aging (n=26, centrality=0.12), and non-human (n=86, centrality=0.10) had a centrality above 0.10, indicating the importance of these keywords in the field of autophagy in OA. Figure 4 presents the high-frequency keywords on a density map; the intensity of the color is proportional to how frequently the keyword appears in the publication.

Keywords with a strong citation burst are keywords that were cited frequently over a given period. Thus, keyword citation bursts can help track the rise or fall of research hotspots. The blue line indicates the time interval, and the red line shows the time of keywords outbreak from start to completion. As shown in Figure 5 , keywords with citation bursts first appeared in 2004 (growth plate) and persisted for a decade. “Cartilage cell” had a strong burst (strength=6.20), followed by “receptor” (strength=5.61) and “cell death” (strength=5.16). The most recent keywords with citation burst were “stress” and “mitophagy”, which appeared in 2020. The changes in keyword citation bursts over time reflect changes in the research trends.

The timeline mapping of keywords is used to present the high-frequency keywords in each cluster over time. As shown in Figure 6 , keywords on autophagy in OA were divided into the following eight clusters: OA, transforming growth factor beta 1 (TGF-β1), cardiovascular disease, skeletal muscle, chondrocyte, multiple sclerosis, growth plate, and polymorphism. Between 2004 and 2008, research on the role of autophagy in articular cartilage maintenance was in its infancy. Research during this period focused on autophagic pathways in growth plate cartilage. Since 2008, keywords such as OA, cardiovascular disease, skeletal muscle, and chondrocyte have entered the spotlight. Of these, “cardiovascular disease” has been a major research focus for some time. “Multiple sclerosis” and “TGF-β1” appeared in 2010 and 2012, respectively, and “polymorphism” began to gather attention in 2013; all three research topics have continued to be studied in recent years.

Co-cited papers refer to two papers that are cited as references by another identical paper, and co-citation analysis can be used to track the evolution of a particular field. Of the total 680 articles cited in papers on autophagy in OA, nine were referenced more than 40 times ( Table S4 ). The most cited paper was “Cartilage-specific deletion of mTOR upregulates autophagy and protects mice from osteoarthritis” (19). Strong reference citation bursts can be indicative of changes in research hotspots over a given time period. Figure S1 shows the top 25 references with the strongest citation bursts involved in OA autophagy. The first burst of co-cited references appeared in 2010. Most of the co-cited references appeared in the last 10 years, implying that research related to OA autophagy may continue to advance in the future.

According to the co-citation analysis, we used CiteSpace to perform a cluster analysis to extract cluster labels from the published papers and identify research frontiers in the field of OA autophagy. Credibility of clustering can be reflected by the silhouette score. A silhouette score >0.7 indicates a robust clustering result. As shown in Table S5 and Figure 7 , the reference network in the field of OA autophagy was extracted into eight co-citation clusters. The largest cluster was labeled “AMPK” and had 109 members with a silhouette value of 0.797. The most relevant citation for this cluster was “osteoarthritis” (n=42), published in The Lancet by Hunter DJ in 2019 (6). The next cluster labeled “macrophagy” had 76 members with a silhouette value of 0.698. The third cluster was labeled “tougu xiaotong capsule (TXC)” and had 61 members, with a silhouette value of 0.925. This cluster contained many papers, such as those by Caramés B (2012) and Sasaki H (2012), with strong citation bursts. Thus, AMPK, macrophage, TXC are the emerging research trends in the field of OA autophagy.

To the best of our knowledge, this is the first bibliometric analysis of autophagy in OA. In this study, we have identified the global hotspots and trends in this field of research. Our results show a steady increase in the cumulative number of publications in the field of OA autophagy from 2004 to 2022. The number of annual publications grew gradually between 2004 and 2017. However, since 2017, this growth has accelerated, reaching its current peak in 2021. These findings suggest that research on autophagy in OA is flourishing.

We found that China ranked top in terms of the output volume on OA autophagy, prior to the USA, South Korea, Japan, and Spain. Meanwhile, the centrality of China and the USA was far ahead of that of other countries, indicating that these countries were leading the international research efforts in the field of OA autophagy. Nine of the top 10 publishing institutions were from China, but showed a low level of collaboration. Therefore, collaboration between research institutions studying OA autophagy should be strengthened in the future. Martin Lotz and Osteoarthritis and Cartilage were the highest output author and journal, respectively. Meanwhile, the record for the most commonly cited papers by an author and journal were held by Caramés B and Osteoarthritis and Cartilage, respectively. Martin Lotz and Beatriz Caramés were among the top 10 authors and co-cited authors, reflecting their significant contributions to the field of OA autophagy. Osteoarthritis and Cartilage ranked first in both publications and citations, demonstrating its authority in the field. The manuscript entitled “Cartilage-specific deletion of mTOR upregulates autophagy and protects mice from osteoarthritis” (19), published in 2015, was most cited, and the paper entitled “Targeted deletion of Atg5 in chondrocytes promotes age-related osteoarthritis”, published in 2016, has the highest centrality (>0.10) (20).

Autophagy has already been awarded the Nobel prize. Prior studies of OA autophagy mainly focused on mechanisms underlying OA and autophagy, usually involving the characterization and localization of specific subcellular forms (e.g., autophagosome and autolysosome) at the cellular level, and the quantification of autophagy-specific proteins (e.g., LC3-II) at the molecular level (70). Our findings revealed that selective autophagy in OA has received increasing attention, with a focus on mitophagy and er-phagy, and substantial progress has been made in elucidating its regulatory mechanism. The reason for the selectivity is that certain autophagy-related proteins or other effectors to specifically recognize receptor molecules on mitochondria or endoplasmic reticulum, and then initiate autophagosome formation on their surfaces (71). Therefore, in autophagy research, it is more important to focus on its selective features, such as the observation and analysis of co-localization of autophagosomes with specific organelles, and the detection and analysis of expression levels and modification levels of characteristic linker molecules related to the mechanism (72). With the advancement of modern biological technology, the integration of traditional detection methods and emerging technologies can further improve the autophagy research system.

Given that autophagy is a molecular-level regulatory mechanism with subcellular morphological characteristics in the cellular microenvironment, the study of autophagy is easier to achieve at the cellular level in vitro, while the technical requirements for the in vivo study of autophagy at the holistic level are relatively high. Atgs are closely related to cell survival. Knockdown of Atgs such as Beclin1, often leads to the death of experimental animals during the embryonic or neonatal stages, so in vivo studies of autophagy at the overall level are currently only possible through partial knockdown of genes or knockdown of autophagy genes in specific organs (73, 74). The limited in vivo experiments have to some extent prevented autophagy target drugs from entering the clinical research stage. In future studies, there is a need to search for more reliable biomarkers of OA autophagy including new selective autophagy receptors, kinases and other effectors, develop new in vivo research techniques, and make new progress in autophagy study in OA.

The pathogenesis of OA is complex and closely related to autophagy, senescence, and apoptosis. Autophagy as an adaptive response can reduce chondrocyte death in the early stages of OA, but excessive autophagy may also lead to chondrocyte death as OA progresses (75). There are still many unproven mechanisms in OA autophagy research, such as the mechanism of synergistic regulation and intrinsic connection between mitophagy and er-phagy in OA, and the regulation of autophagy levels in pathological states. All these need to be further explored in future.

Autophagy has a broad clinical prospect in the treatment of OA, and targeting autophagic pathways provides a new direction for OA treatment. Particularly, locally targeted drug delivery, such as intra-articular injection of autophagy-regulated drugs, is expected to become a new hotspot in the treatment of OA. In the future, clinical studies on the treatment of OA with autophagy-regulated drugs can be carried out gradually. With the support of high quality clinical evidence, autophagy-activating drugs may be used to enhance or restore autophagic activity, thereby to stop disease progression for patients with early-stage OA; for those with intermediate and advanced stages, autophagy-inhibiting drugs may be applied to reduce excessive autophagy, so as to improve pain and joint function, and delay disease progression. In addition, some natural active substances with autophagy-regulating effects may be used as adjuvant therapy to further improve the clinical efficacy of OA, which also provides a new idea for the treatment of OA.

This study has some shortcomings. Firstly, we used CiteSpace software for data analysis, which may eliminate some data components from the analysis. Thus, we may have missed some detailed data. Secondly, our study was based on literature published between 2004 and 2022, meaning that our findings are only relevant to this research period and will need to be updated to keep pace with the volume of new research outputs.