Research hotspots of polycystic ovary syndrome and hyperandrogenism from 2008 to 2022: bibliometric analysis

Abstract

Background

Polycystic Ovary Syndrome (PCOS) is the most frequent endocrine disorder in female adults, and hyperandrogenism (HA) is the typical endocrine feature of PCOS. This study aims to investigate the trends and hotspots in the study of PCOS and HA.

Methods

Literature on Web of Science Core Collection (WoSCC) from 2008 to 2022 was retrieved, and bibliometric analysis was conducted using VOSviewer and CiteSpace software.

Results

A total of 2,404 papers were published in 575 journals by 10,121 authors from 2,434 institutions in 86 countries. The number of publications in this field is generally on the rise yearly. The US, China and Italy contributed almost half of the publications. Monash University had the highest number of publications, while the University of Adelaide had the highest average citations and the Karolinska Institute had the strongest cooperation with other institutions. Lergo RS contributed the most to the field of PCOS and HA. The research on PCOS and HA mainly focused on complications, adipose tissue, inflammation, granulosa cells, gene and receptor expression.

Conclusion

Different countries, institutions, and authors should facilitate cooperation and exchanges. This study will be helpful for better understanding the frontiers and hotspots in the areas of PCOS and HA.

Keywords:

• Polycystic ovary syndrome
• hyperandrogenism
• bibliometrics
• VOSviewer
• CiteSpace

Introduction

Polycystic Ovary Syndrome (PCOS) is one of the most common endocrine disorders in female adults, approximately 5% to 20% of global women of childbearing age have PCOS [1–3]. PCOS is characterized by hyperandrogenism (HA), ovulatory dysfunction, and polycystic ovarian morphologic features, and the most typical feature is HA, about 65% to 80% of PCOS patients have signs or symptoms of HA [4,5]. HA suppresses follicular development, leading to ovulatory dysfunction and fertility [6]. Moreover, HA increases the risk of adverse pregnancy outcomes [7], cardiovascular disease, metabolic disease, sleep disorder [8], anxiety, depression [9], and endometrial cancer [10], which seriously affects the quality of the patient’s life and aggravates social and economic burden [11]. Therefore, a lot of existing studies focus on PCOS and HA, while the rapid accumulation of knowledge makes it intricate to identify key information and keep up with the latest research hotspots, especially for those who are new to the field.

Bibliometric analysis was first proposed by Pritchard in 1969 [12], since then it has grown rapidly to appraise the overall trend of research activity. VOSviewer and CiteSpace are the two most common software used in bibliometric analysis. VOSviewer offers text mining functionalities to construct and visualize co-occurrence networks of significant terms such as authors, institutions, countries, and projects extracted from scientific literature. Its robust graphical display capabilities effectively illustrate collaborative relationships among projects. However, it comes with certain limitations, including the need for manual data updates, cumbersome settings, and poor compatibility. CiteSpace, developed within the realm of scientometrics and data visualization, is a visual analysis software with distinct keyword burst analysis capabilities. It excels in showcasing changes in research hotspots and trends, unveiling connections within and between disciplines, and exploring prevalent issues within specific disciplines. Notably, CiteSpace demands advanced technical skills and a robust computer configuration, offering high operability. It is worth noting that diagrams produced by different researchers using CiteSpace may exhibit significant variations [13–15]. Quantitative analysis of literature helps to better grasp the contributions of various institutions, countries, and authors, as well as frontiers and hot spots of studies related to PCOS and HA [16].

To our knowledge, this is the first bibliometric analysis to comprehensively evaluate and compile the research findings in the field of studies related to PCOS and HA. This study collected the scientific publications on PCOS and HA in Web of Science Core Collection (WoSCC) database for 15 years and made a bibliometric analysis based on the publications. Our study aims to systematically comb through the scientific publications, help researchers quickly acquire the current state of research, provide scholars and clinicians with significant information and help them understand the most active hotspots and trends in the field of PCOS and HA.

Materials and methods

Data sources and search methods

In order to ensure the comprehensiveness and accuracy of the search data, the search and screening were conducted by 2 researchers, and the search strategy was formulated in consultation. We used The Science Citation Index Expanded (SCI-E) and the Social Sciences Citation Index (SSCI) in the WoSCC database (the most adequate database source for bibliometrics) as citation indexes. We searched the subject words of “polycystic ovary syndrome” and “hyperandrogenism” in the MeSH Database in PubMed, thus, the search strategy of our study was TS= “Polycystic Ovary Syndrome” OR “Syndrome, Polycystic Ovary” OR “Ovary Syndrome, Polycystic” OR “Stein Leventhal Syndrome” OR “Stein-Leventhal Syndrome” OR “Syndrome, Stein-Leventhal” OR “Ovarian Degeneration, Sclerocystic” OR “Sclerocystic Ovarian Degeneration” OR “Sclerocystic Ovary Syndrome” OR “Ovarian Syndrome, Polycystic” OR “Polycystic Ovarian Syndrome” OR “Sclerocystic Ovaries” OR “Sclerocystic Ovary” OR “Ovary, Sclerocystic” AND “hyperandrogenism” OR “HAIR-AN syndrome”. The language was limited to English and the publication types were articles and reviews, the literature ranged from Jan 1st, 2008 to Dec 31st, 2022. The retrieved records were saved in “plain text file” format and “full record and cited references” was selected. We then manually read the titles and abstracts one by one and removed unrelated literature. The literature was screened back-to-back, and a third party was involved when encountered disagreements.

Informed consent was not required for this study because all data included in our paper were downloaded from public databases and our study did not involve human participants.

Bibliometric analysis

Publications on PCOS and HA included in the WoSCC database from 2008 to 2022 were visualized by bibliometric analysis using VOSviewer (version 1.6.18 Leiden University, Leiden, Netherlands) and CiteSpace (version 6.2.R2), which provided clear visual graphs and research hotspots and trends [13,17–19].

All valid data were imported to VOSviewer (version 1.6.18) for visualization of countries, institutions, authors, journals, co-citations and keywords, and to CiteSpace (version 6.2.R2) for burst detection of keywords and co-citations. Various tools such as EndNote, WPS Office and Microsoft Excel 2019 were also used in this study for data collection and graphing. Noteworthily, some data in our study were transformed during the visual analysis, for example, similar keywords were merged into one keyword, and data from Hong Kong or Taiwan were merged into that from China.

Results

General information and the trend of publication outputs

A total of 2,547 records were found in SCI-E and SSCI, after excluding meeting abstracts, proceedings, book chapters, and non-English, duplicated or retracted records, 2,404 articles were finally included in our study (Supplementary Figure 1). From 2008 to 2022, the number of papers about PCOS and HA has increased, which can be divided into two stages: (1) Initial stage: the number of papers published had a stable growth before 2020; (2) Rapid development stage: the number of publications had a significant increase from 2020 to 2022, more than 260 papers were published in 2022 (Supplementary Figure 2). The trends of publication outputs indicated that the field of PCOS and HA has developed rapidly in recent 3 years and has become a research hotspot.

Distribution of major countries

The 2,404 articles came from 86 countries. The top 10 countries published 1,996 papers, representing 83% of the total (Table 1). The top 3 countries in terms of number of papers are the USA (486, 20%), China (460, 19%), and Italy (219, 9%). However, regarding average citations, Australia (85 citations) ranked the highest, followed by Spain (52 citations) and Italy (48 citations). The network visualization map of top 30 countries formed 5 stable clusters (Figure 1A). The USA had the closest cooperation with China, and it also actively cooperated with other countries, such as Australia, Italy and Turkey. Apart from the USA, China closely cooperated with Sweden as well. Besides, European countries had a stable cooperative relationship (Figure 1B).

Figure 1. Contribution and global distribution of countries. A: The countries with a minimum number of 17 papers. B: Global map of countries with a minimum number of 17 papers. Nodes represent countries, and the larger the node, the higher the number of papers. The lines represent the association between countries, the thicker the line indicates the stronger association. Color represents clustering, and nodes with the same color belong to the same cluster.

Table 1. Top 10 countries with most publications.

Rank	Country	Publications	Citations	Average Citations
1	USA	486	20195	42
2	China	460	9010	20
3	Italy	219	10579	48
4	Turkey	186	3477	19
5	England	151	6316	42
6	India	115	2022	18
7	Australia	109	9260	85
8	Poland	96	1915	20
9	Spain	89	4591	52
10	Brazil	85	1819	21

Distribution of major institutions

A total of 2,434 institutions participated in these publications, and the top 10 institutions produced 16% of the articles (377/2404) (Table 2). The number of publications from Monash University, Shanghai Jiao Tong University and Karolinska Institute was 49, 45, and 40, respectively, ranking the top 3 institutions. Interestingly, the University of Adelaide published a small number of articles, it ranked the highest in average citations. The network visualization graph of the top 50 institutions formed 7 clusters (Figure 2). Among the key network hubs connecting different institutions, universities cooperated more compared with other institutions such as companies, and Karolinska Institute showed the best connections with other institutions. Additionally, several Chinese institutions such as Shanghai Jiao Tong University, Fudan University, Heilongjiang University of Chinese Medicine and Peking University contributed prominently.

Figure 2. Contribution of institutions with a minimum number of 17 papers. Nodes represent institutions, and the larger the node, the higher the number of papers. The lines represent the association between institutions, the thicker the line indicates the stronger association. Color represents clustering, and nodes with the same color belong to the same cluster.

Table 2. Top 10 institutions with most publications.

Rank	Organizations	Country	Publications	Citations	Average Citations
1	Monash Univ	Australia	49	3967	81
2	Shanghai Jiao Tong Univ	China	45	817	18
3	Karolinska Inst	Sweden	40	2271	57
4	Univ Adelaide	Australia	37	4385	119
5	Univ Chile	Chile	37	1339	36
6	Univ Calif Los Angeles	USA	36	4180	116
7	Univ Gothenburg	Sweden	35	1001	29
8	Heilongjiang Univ Chinese Med	China	34	930	27
9	Peking Univ	China	32	1001	31
10	Univ Sao Paulo	Btazil	32	570	18

1 Monash Univ-Monash University. 2 Shanghai Jiao Tong Univ-Shanghai Jiao Tong University. 3 Karolinska Inst-Karolinska Institute. 4 Univ Adelaide-The University of Adelaide. 5 Univ Chile-University of Chile. 6 Univ Calif Los Angeles-University of California, Los Angeles, UCLA. 7 Univ Gothenburg-University of Gothenburg. 8 Heilongjiang Univ Chinese Med - Heilongjiang University of Chinese Medicine. 9 Peking Univ–Peking University. 10 Univ Sao Paulo-University of Sao Paulo.

Distribution of main authors

A total of 10,121 authors contributed to these articles. Lergo RS was the most productive researcher with 34 articles, and he had the highest citations compared with other authors, while, Azziz R ranked first among co-cited authors (Table 3). 85 authors met the thresholds of a minimum number of 9 papers for an author, comprising 47% of all published papers (1,119/10,121), and their visual map formed 22 clusters (Figure 3A). Hu M et al. group and Walters KA et al. group were emerging researchers that had tight cooperation, but they had weak cooperation with other active authors (Figure 3B).

Figure 3. Contribution of authors with a minimum number of 9 papers. A: The network visualization map of these authors formed 22 clusters. B: Chart of average year of author publication. Nodes represent authors, and the larger the node, the higher the number of papers. The lines represent the association between authors, the thicker the line indicates the stronger association. Color represents clustering, and nodes with the same color belong to the same cluster.

Table 3. Top 10 authors with most publications and highest co-cited authors.

Rank	Author	Publications	Citations	Co-cited author	Co-citations
1	Legro RS	34	5088	Azziz R	1856
2	Escobar-Morreale HF	31	3101	Legro RS	1214
3	Azziz R	27	3779	Diamanti-Kandarakis E	1058
4	Stener-Victorin E	25	1780	Dunaif A	925
5	Qiao J	23	871	Carmina E	919
6	Dumesic DA	22	2283	Fauser BCJM	910
7	Carmina E	19	2716	Ibanez I	774
8	Teede HJ	19	1846	Escobar-Morreale HF	737
9	Abbott DH	19	878	Chang J	699
10	Luque-Ramirez M	19	662	Ehrmann DA	607

Distribution of major journals

In total, 575 academic journals published papers related to PCOS and HA, and Gynecological Endocrinology had the highest number of outputs (145, 6%) (Supplementary Table 1). Journal of Clinical Endocrinology & Metabolism (17,775 citations) ranked first among co-cited journals (Supplementary Table 2). The visualization mapping of the top 50 journals established 6 clusters (Figure 4A). Although Frontiers in Endocrinology (45 articles), International Journal of Molecular Science (19 articles) and Nutrients (12 articles) are relatively new journals (Figure 4B), they published a majority of articles and made certain contributions.

Figure 4. Contribution of journals with a minimum number of 10 papers. A: The network visualization map of these journals formed 6 clusters. B: Chart of average year of journal publication. Nodes represent journals, and the larger the node, the higher the number of papers. Color represents clustering, and nodes with the same color belong to the same cluster.

Highly cited and co-cited literature

Highly cited literature refers to the top 1% of papers by field and publication year, and co-citation represents that two or more articles are cited by one or more papers at the same time. Both highly cited and co-cited papers reflect certain research trends and hotspots. Among the top 20 cited publications, 11 described the definition or diagnostic criteria for PCOS (Supplementary Table 3). Azziz R et al. (1249 citations, ranked first) emphasized the significance of HA in the diagnosis of PCOS, and March WA et al. (1130 citations, ranked second) made a great contribution to the morbidity of PCOS under various diagnostic standards. Also, Escobar-Morreale HF published a review that summarized the definition, etiology, diagnosis and treatment of PCOS, which made a great flutter. Among the top 39 co-cited publications, 52,093 papers were cited by 2,404 articles, and the top 39 papers were cited at least 100 times, the top 1 paper was contributed by Chang J et al. which was cited 698 times, much higher than the second one (622 times). The paper of Azziz R et al. that demonstrated the epidemiology of PCOS ranked third (Supplementary Table 4). Visualization map of top 39 co-cited references formed 3 clusters (Figure 5A). Connectivity analysis of all references formed 12 domains (Figure 5B). Studies related to PCOS and HA formed a bridge at multiple aspects, and we should pay more attention to the association among inflammation, pregnancy, steroidogenesis and bisphenol a.

Figure 5. Co-cited papers. A: Visualization of co-cited papers with a minimum number of 100 citations. B: Connectivity analysis of papers based on Citespace formed 12 domains. The node represents the paper, and the larger the node, the higher the number of citations. A line indicates that two papers are cited at the same time, and a thicker line indicates more co-citations. Color represents clustering, nodes with the same color belong to the same cluster.

Keywords and hotspots

After merging similar keywords, 6,121 keywords were extracted from 2,404 papers. Polycystic Ovary Syndrome, Hyperandrogenism and Woman were the top 3 keywords with the number of occurrences of 1,876, 1,036, and 829 respectively (Table 4). According to the keywords given by the authors, we found that oxidative stress, inflammation, androgen excess, and insulin resistance were the main mechanisms that cause the diseases. The visualization map of the top 100 keywords formed 4 clusters (Figure 6A). The research hotspots were adipose tissue or obesity, oxidative stress and inflammatory response, granulosa cells, gene and receptor expression (Figure 6B).

Figure 6. Map of high-frequency keywords. A: Visualization of top 100 keywords formed 4 clusters. B: Average year of top 100 keywords occurrence. Nodes represent keywords, and the larger the node, the higher occurrence of the keyword. Color represents clustering, and nodes with the same color belong to the same cluster.

Table 4. Top 10 most frequently occurring keywords in each of the top 100 keywords cluster.

Cluster	Top 10 most frequent keywords
Red	Polycystic Ovary Syndrome(1876), Hyperandrogenism(1036), Women(829), Testosterone(276), Androgens(243), Expression(190), Anti Mullerian Hormone(187), Luteinizing Hormone(107), Infertility(104), Insulin(95).
Green	Metformin(228), Oxidative Stress(161), Obese Women(151), Inflammation(133), Sensitivity(108), Insulin Sensitivity(98), Glucose(94), Adipose Tissue(93), Disease(87), Double Blind(79).
Blue	Insulin Resistance(1028), Obesity(390), Metabolic Syndrome(300), Androgen Excess(292), Impaired Glucose Tolerance(126), Young Women(110), Cardiovascular Risk(88), Hormone Binding Globulin(76), Cardiovascular Disease(73), Body Mass Indx(69).
Yellow	Prevalence(508), Diagnostic Criteria(412), Risk(228), Hirsutism(187), Association(152), Consensus(96), Features(84), Management(81), Hyperandrogenemia(74), Quality of Life(74).

Discussion

Bibliometrics is a comprehensive framework that amalgamates statistical, mathematical, and philological methodologies to quantitatively assess academic literature within distinct domains [19]. Both androgen excess and insulin resistance (IR) play a key role in the progression of PCOS, and there are several bibliometric analyses of PCOS and IR [12, 20], however, no record pertaining to PCOS and HA was found. We analyzed 2,404 papers published in 575 journals by 10,121 authors from 2,434 institutions in 86 countries.

Based on our analysis, we found that most papers were published in developed countries, and the cooperation between developed countries and developing countries was limited. Moreover, a large disequilibrium of publications in institutions was detected due to the investment and reputation, the investment of research in PCOS and the reputation of the institutions are positively associated with the number of publications and academic cooperation. Despite publishing a relatively small number of articles, the University of Adelaide has achieved the highest average citation rate, signaling its impactful contributions. This success may be attributed to the university’s substantial early investments in the field of PCOS epidemiology. Notably, an article from the University of Adelaide holds the second position among all cited articles. Besides, the Karolinska Institute’s extensive collaborations with numerous institutions can be linked to its status as one of the world’s largest and premier medical schools, with a strong focus on medical teaching and research. Therefore, those interested in delving into the epidemiology and diagnostic criteria of PCOS may find fruitful collaborations with the University of Adelaide. Conversely, researchers focusing on the pathogenesis and management of PCOS could explore partnerships with the Karolinska Institute.

Price’s law can identify core authors in a subject area, which suggests that around 50% of the articles on the same subject belong to a pool of high-production authors. In our study, the number of papers published by the top 85 authors accounted for 46.5% of the total, which is roughly consistent with Price’s Law, thus the 85 authors were comprised of the core authors in research of PCOS. Furthermore, most publications were found in specific journals in the areas of reproductive medical science, endocrinology and metabolism, and publications in these journals were mainly categorized as clinical research [21–26]. Additionally, through analyzing the literature citation, keywords, and hotspots, we reviewed the diagnosis, risk factors, pathogenesis, and treatment of PCOS.

Currently, there are three globally accepted definitions or diagnostic criteria for PCOS. However, the choice of a specific definition may vary depending on the issue addressed by individual researchers, given the wide range of clinical features and syndrome outcomes. For instance, when the primary clinical or research concern is the long-term risk of metabolic or cardiovascular diseases in patients with PCOS, the National Institutes of Health (NIH) or the AE-PCOS Society is chosen. If the research interest is genetic factors, adopting stringent criteria such as those outlined by NIH may be essential to ensure the homogeneity of the study population. When the focus is on assessing the risk of anovulatory infertility and/or ovarian hyperstimulation during ovulation induction, the Rotterdam criteria may be more appropriate. In our study, eleven out of the top 20 most cited articles focus on defining or establishing diagnostic criteria for polycystic ovary syndrome (PCOS), highlighting a pressing need for internationally standardized diagnostic criteria for PCOS.

Thus, there are few reliable statistics on the epidemiology of PCOS due to the inconsistency of diverse standards in geographical and ethnic differences [27]. March, W.A. et al. conducted a comparative analysis of polycystic ovary syndrome (PCOS) prevalence using different diagnostic criteria within the same cohort [28]. The results demonstrated that the Rotterdam and AES criteria yielded prevalence estimates up to twice as high as those obtained with the NIH criteria. Another study focused on Han Chinese women aged 19-45 years, the overall prevalence of PCOS was found to be 5.6%. The phenotypic distribution within this group revealed hyperandrogenism and polycystic ovaries as the most prevalent (37%), while chronic anovulation and polycystic ovaries were the least prevalent (15%) [29]. However, the clinical features of HA and menstrual cycle disorder are more common than ultrasound detection of follicles [30], while the ultrasound detection of 23 or more follicles in ovaries is a strong diagnosis of PCOS [31].

Genetic predisposition contributes to the occurrence of PCOS, 3 genome-wide significant loci were identified in European female patients with NIH PCOS phenotype [32], including two novel genetic sites and another site formerly observed in Chinese females. PCOS candidate loci such as LHCGR, FSHR, and ZNF217 regulate androgen biosynthesis [33]. Moreover, overweight/obesity is another risk factor that aggravates metabolic function in PCOS, studies have found that obese patients with PCOS had a higher rate of metabolic syndrome (16% versus 48%), insulin resistance (7% versus 28%), hypertension (8% versus 30%) and hyperlipemia (48% versus 73%) compared with non-obese patients [29, 34]. Also, high Body Mass Index in patients with PCOS increases 3-fold risk of hypertension, preeclampsia, and diabetes during pregnancy [35].

The pathogenesis of PCOS is associated with folliculogenesis, steroid production and the biological effects of gonadotropin [36,37]. The majority of patients with PCOS have defective steroid secretion [38], and high levels of testosterone in follicular fluid impair oocyte development [39]. Adipose tissue plays a key role in secretion of androgen, and androgen signal transduction in white adipose tissue mediates the progression of PCOS and its metabolic features [40,41]. Moreover, dysfunction in gonadal adipose tissue is involved in fertility alterations [42]. The overexpression of Anti-Mullerian Hormone (AMH) and AMH receptors in follicular granulosa cells shows the disturbance of the endocrine system [43,44], and the serum AMH concentration serves as a surrogate of HA in PCOS [45]. Low-grade chronic inflammatory state together with hyperinsulinemia, and HA acts in a vicious cycle in the pathophysiology of PCOS. On the one hand, low-grade chronic inflammation worsens the metabolism in patients with PCOS [46], on the other hand, abnormal activation of High Mobility Group Box1 (HMGB1) pathway in PCOS aggravates uterine inflammation [47], leading to worse metabolic function, abortion and infertility.

Indeed, abortion and infertility are two severe consequences caused by PCOS. Excessive androgens induce dysmorphia of uterine and PCOS-related gene expression that regulates placentation, decidualization and angiogenesis [48]. Also, increased level of reactive oxygen species in PCOS deteriorates placental formation, impairs angiogenesis, and enhances glycogen, which leads to abortion and infertility [49], and oxidative stress regulates Sex Hormone-binding Globulin(SHBG) that is critical for HA development [50].

Studies have reported that metformin [51] and antioxidant N-acetylcysteine [52] can improve the function of uterine, and vitamin D [53], low-dose spironolactone [54], puerarin [55], and herbal medicine [56] ameliorate oxidative stress and apoptosis, which might be the therapeutic intervention for abortion and infertility caused by PCOS. Other management of abortion and infertility caused by PCOS include lifestyle modification, bariatric surgery, pharmacotherapy, and laparoscopic surgery [57]. Besides, the AMH concentration in a different stage of puberty in daughters of women with PCOS is higher than in daughters of control women [58,59]. The daughters of women with PCOS show more severe ovarian dysfunction and metabolic derangements [21], thus, early screening and intervention should be used in daughters of women with PCOS.

Long-term health problems caused by PCOS, such as insulin resistance (IR), cardiovascular diseases, obstetric complications, dyslipidemia, hepatic steatosis, thrombosis, endometrial cancer, and psychological disorders should also be paid attention to [37, 60–62]. Most patients with PCOS have IR [63] that can be predicted by serum calprotectin levels [64]. Modified Cangfu Daotan, quercetin and troxerutin have therapeutic effects on IR in patients with PCOS [65–67]. Furthermore, inflammation in PCOS stimulates atherosclerotic inflammation, and oxidative stress increases the risk of cardiovascular diseases [68,69]. Inflammation-related biomarkers and microRNA (miRNA) can potentially reflect cardiovascular risk [70]. Besides, HA affects hepatic gene expression and metabolism, leading to hepatic steatosis [71]. Therefore, adipose tissue, obesity, oxidative stress, granulosa cells, gene and receptor expression are valuable research areas in the future study of PCOS and HA.

The keywords and hotspots in our study underscore the significance of oxidative stress and metabolic diseases associated with hyperandrogenism (HA-PCOS). Notably, oxidative stress plays a crucial role in inhibiting SHBG expression and secretion, representing a pivotal factor in the development of hyperandrogenism. Furthermore, oxidative stress is implicated in the development of insulin resistance (IR) in the skeletal muscle of PCOS patients [72]. Elevated oxidative stress levels and a deficiency in antioxidant status substantially heighten the risk of cardiovascular disease in women with PCOS [73]. The exploration of pathways related to oxidative stress in future research holds promising potential as a potent avenue for PCOS treatment. Studies have found that a combination therapy involving metformin and resveratrol induces antioxidant and anti-inflammatory systems in PCOS through the activation of SIRT1 and AMPK [74]. Additionally, interventions such as Vitamin D [53], low-dose spironolactone [54], puerarin [55], and herbal medicine [56] have demonstrated efficacy in improving oxidative stress and apoptosis in the context of PCOS.

Deficiencies in insulin action or the insulin signaling pathway are central to PCOS development [63]. While IR is prevalent in most women with PCOS, variations exist between phenotypes [75], and serum calprotectin levels provide predictive insights into IR [64]. The presence of visceral obesity in women with PCOS exacerbates functional ovarian androgen excess, contributing to elevated levels of inflammatory adipokines, insulin resistance, and lipogenesis [76].

Our study has several limitations. First, our research has a certain degree of selection bias due to limited databases and types of articles, however, WoSCC database is the most adequate database source for bibliometrics. Second, visual bias exists because of the inconsistency of parameter settings of VOSviewer and CiteSpace, while we reviewed the previous literature and discussed with the experts to minimize the bias. Finally, considering the multiplicity of keywords and the possibility of authors with eponymous names, it is difficult to exactly find all publications. However, the number of papers is relatively large and can reflect the certain reality of research field.

Conclusion

Our extensive bibliometric analysis provides a comprehensive overview of studies of PCOS and HA over 15 years. The number of publications in this field is generally on the rise yearly. The US and China are the leading countries in this area. Regarding the research institutions, Monash University is the institution with the highest number of publications. Lergo RS contributes the most to the field of PCOS and HA. Different countries, institutions, and authors should facilitate cooperation and exchanges. Currently, the research on PCOS and HA mainly focuses on complications, adipose tissue, inflammation, granulosa cells, gene and receptor expression, which represent the current research and developmental trends in future research.

Acknowledgement

Thanks to the hard work of all countries, institutions, and authors in this field, it is because of their efforts that this manuscript can be successfully completed.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, [WGH], upon reasonable request.

Additional information

Funding

Our study is jointly funded by the hospital-level cultivation project of the Third Affiliated Hospital of Beijing University of Chinese Medicine (BZYSY-2022-PYMS-13) and the special scientific research project of doctoral students of Beijing University of Chinese Medicine. None of them participated in this study.