Knowledge mapping of Cushing syndrome: A bibliometric analysis from 2000 to 2023

A literature research was conducted on November 11, 2023, using the Web of Science Core Collection database. The search strategy was as follows: TS = (“Cushing syndrome”) OR TS = (“Cushing syndrome”). The search period was restricted to publications from January 2000 to October 2023. A total of 5831 records were initially retrieved. After limiting the language to English, 5448 publications remained. Conference proceedings, theses, book chapters, news articles, and records with incomplete or duplicated information were excluded. Only original research articles and review articles were retained, generating 4661 records for bibliometric analysis (Fig. 1). This study did not involve human participants and animal subjects and therefore did not require institutional ethical approval.

Core bibliometric methods were applied in this study. These included: co-occurrence analysis, used to identify relationships among entities such as keywords, countries, and institutions; co-citation analysis, to examine the intellectual structure of the field; citation burst detection, based on Kleinberg algorithm, to identify emerging research trends; timeline visualization, to trace thematic evolution using average publication year; and cluster analysis, to classify network nodes into coherent research sub-domains.

Primary analyses were performed using VOSviewer (version 1.6.19; Leiden University, Leiden, The Netherlands), including country and institution mapping, journal and co-cited journal analysis, author and co-cited author mapping, reference co-citation analysis, and keyword clustering to identify hotspots and research frontiers. In visualizations generated by VOSviewer, nodes represent entities such as countries, institutions, journals, or authors. Node size and color denote quantity and category, respectively, while line thickness reflects the strength of collaboration or co-citation relationships.

CiteSpace software (version 6.1.R6; Drexel University, Philadelphia) was employed for journal dual-map overlays, and citation burst detection of references and keywords. Bibliometrix, an R-based package (version 3.2.1; https://www.bibliometrix.org↗), was used for trend and performance analysis. Journal quartile rankings and impact factors were sourced from Journal Citation Reports 2023. Descriptive statistics were performed using Microsoft Excel 2019 (Microsoft Corporation, Redmond).^[7,8]

The annual number of publications reflects the pace and direction of research in a given field. As shown in Figure 2, from 2000 to 2009, the number of publications on CS exhibited a gradual upward trend, ranging from 110 to 200 articles per year. Between 2010 and 2019, the annual output stabilized, fluctuating around 200 articles. From January 2020 to October 2023, annual publications exceeded 280. Overall, both the number of publications and total citations have shown a consistent increase, indicating growing scholarly attention and sustained research activity in this field.

A total of 4661 papers on CS were published by 3972 institutions across 86 countries. The 10 most productive countries and institutions are listed in Table 1. The United States ranked first, contributing 1205 publications and receiving 52,833 citations. Notably, 6 of the top 10 institutions are based in the United States and Italy, highlighting these countries’ central role in CS research. Leading institutions, such as the Eunice Kennedy Shriver National Institute of Child Health and Human Development and the University of Padua, are frequently involved in collaborative research initiatives, both domestically and internationally.

A co-occurrence network was generated using countries with at least 5 publications and institutions with at least 24 publications (Fig. 3A and B). The network maps show that the United States, England, Italy, Japan, and France are represented by larger nodes and thicker connecting lines, indicating higher publication output and stronger collaborative ties. The Eunice Kennedy Shriver National Institute of Child Health and Human Development, based in the United States, produced the highest number of publications and citations across all institutions. The international co-authorship patterns reveal that researchers from these leading countries frequently collaborate, sharing expertise, data, and resources to advance the field.

Research on CS was published across 1021 journals. The Journal of Clinical Endocrinology & Metabolism was the most relevant journal, publishing 321 articles (31.44% of total publications), followed by the European Journal of Endocrinology with 234 articles (22.92%). Clinical Endocrinology and Endocrine also contributed significantly, publishing 155 and 129 papers, respectively. Among the top 10 journals (Table 2), the Journal of Clinical Endocrinology & Metabolism and European Journal of Endocrinology held the highest impact factors (IF = 5.8). In terms of quartile rankings, 5 journals (50%) were classified as quartile 1 (Q1), 2 (20%) as quartile 2, 1 (10%) as quartile 3, and 2 (20%) as quartile 4, indicating that the majority of CS-related research is disseminated in high-impact journals.

A co-authorship threshold of at least 13 publications was applied, leading to 56 journals used to construct the journal co-occurrence network (Fig. 4A). The map reveals strong citation relationships between the Journal of Clinical Endocrinology & Metabolism and journals such as Clinical Endocrinology, Frontiers in Endocrinology, and Endocrinology.

Co-citation analysis (Table 2) showed the Journal of Clinical Endocrinology & Metabolism ranks first with 29,096 co-citations, followed by Clinical Endocrinology (8850), European Journal of Endocrinology (8635), New England Journal of Medicine (4345), and The Lancet (2316). Notably, The Lancet had the highest impact factor (IF = 168.9), followed by the New England Journal of Medicine (IF = 158.5). Among the top 10 co-cited journals, 7 (70%) were ranked as Q1, 2 (20%) as quartile 2, and 1 (10%) as quartile 3. Journals with a minimum of 500 co-citations were selected for the co-citation network (Fig. 4B). This visualization highlights the strong co-citation relationships of journals such as The Journal of Immunology with Blood, PLOS ONE, and Frontiers in Immunology.

The dual-map overlay (Fig. 4C) presents journal citation trajectories. Three major citation paths were identified. The orange path indicates that research published in Molecular/Biology/Genetics journals was primarily cited by Molecular/Biology/Immunology journals. The green paths show broader interdisciplinary connections, where studies published in Molecular/Biology/Genetics/Health/Nursing/Medicine journals were predominantly cited in Medicine/Medical/Clinical domains.

A total of 19,621 authors contributed to CS research. The top 10 most profilic authors and their citation counts are listed in Table 3. A collaborative network was constructed for authors with at least 18 publications (Fig. 5A). Authors with the largest nodes, such as Constantine A. Stratakis, Martin Reincke, Jérôme Bertherat, Susan M. Webb, and Felix Beuschlein, had the highest publication output. Strong collaborative relationships were identified: Jerome Bertherat frequently coauthored with Constantine A. Stratakis; Susan M. Webb with Alicia Santos, Eugenia Resmini, and Elena Valassi; and Martin Reincke with Felix Beuschlein. These findings reflect a high degree of international cooperation and knowledge exchange among key contributors.

Table 3 also presents the top 10 co-cited authors. Lynnette K. Nieman was the most co-cited author (n = 1552), followed by André Lacroix (n = 1197) and Rosario Pivonello (n = 963). Authors who had a minimum of 220 co-citations were included in the author co-citation network (Fig. 5B), reflecting a dense and interconnected knowledge foundation shaped by repeated referencing of authoritative sources.

The top 10 co-cited references, listed in Table 4, each received at least 226 citations, with the most highly cited reference exceeding 700 citations. The most co-cited paper was authored by Nieman Lk in 2008, published in The Journal of Clinical Endocrinology & Metabolism (Vol. 93, p. 1526), which provided clinical guidelines for testing and diagnosing endogenous CS.^[9] The guideline recommends that patients undergo an initial high-sensitivity screening test. If both test results are abnormal, an etiological workup is necessary. If discordant results occur, periodic hypercortisolism should be considered. References with at least 90 co-citations were used to generate the co-citation network (Fig. 6A).

Citation bursts identify publications that have received a sharp increase in attention over a specific period, often indicating influential or emerging work. Sixteen references with significant citation bursts were detected using CiteSpace (Fig. 7A). Burst periods ranged from 2002 to 2021. The reference with the strongest citation burst (strength = 61.94) was “The diagnosis of Cushing syndrome: an Endocrine Society Clinical Practice Guideline,” authored by Lynnette K. Nieman et al, with the burst lasting from 2009 to 2013.^[9] The second strongest burst (strength = 50.99) occurred for “Treatment of Cushing Syndrome: An Endocrine Society Clinical Practice Guideline,” published in The Lancet by Lynnette K. Nieman et al, from 2016 to 2020.^[10] This reference provided detailed guidance on restoring normal cortisol levels through effective clinical intervention. Overall, the burst strengths ranged from 19.58 to 61.94, with durations of 5 to 6 years, indicating sustained influence over time.

To distinguish foundational works from emerging studies, publications were classified as “citation classics” (≥500 citations) and “non-classics” (<500 citations), as presented in Table 5. Among the 16 citation classics, clinical practice guidelines (n = 5) and review articles (n = 9) dominated, highlighting their central role in conslidating knowledge and establishing standards of care. The remaining 2 publications were original research articles (n = 2).

The Journal of Clinical Endocrinology & Metabolism served as the principal publishing platform for these citation classics, accounting for 8 of the top 16 (50%). Notably, 12 of the top 16 classics (75%) appeared in journals classified as Q1 according to Journal Citation Reports 2023, reflecting their high impact and scholarly influence. Identifying citation classics provides insight into the field’s foundational literature and underlines key areas that continue to attract sustained research interest.

Keywords serve as condensed representations of a study’s core concepts and thematic focus. To refine the keyword dataset, generic disease-related terms such as Cushing syndrome, Cushing disease, and primary aldosteronism were excluded. Additionally, nonspecific terms such as society, European society, outcome, and masse were removed. A total of 5610 author keywords were identified. Of these, 59 appeared more than 25 times (Fig. 8A). The most frequent keywords were “cortisol” (n = 308), “adrenocorticotropic hormone” (n = 196), and “glucocorticoids” (n = 180). Table 6 lists the 20 most frequently used keywords in CS research.

Co-occurrence analysis grouped these keywords into 5 clusters. The red cluster relates to diagnostic studies and includes cortisol, adrenocorticotropic hormone, adrenal, ectopic adrenocorticotropic hormone secretion, and dexamethasone. The blue cluster centers on surgical approaches, including bilateral adrenalectomy, pituitary surgery, transsphenoidal surgery, and quality of life. The green cluster addresses etiology and includes adrenal incidentalomas, adrenal tumor, adrenalectomy, and laparoscopic adrenalectomy. The yellow cluster focuses on complications and comorbidities, with terms such as glucocorticoids, obesity, diabetes mellitus, insulin resistance, and metabolic syndrome. The purple cluster relates to pharmacotherapy and includes agents such as mifepristone, metyrapone, and ketoconazole. Overall, research in CS has concentrated on surgical treatment, pharmacologic management, and disease mechanisms.

Keyword overlay visualization (Fig. 8B) revealed that terms such as Armadillo repeat-containing protein 5 (ARMC5), pituitary surgery, and metyrapone have shown increasing frequency in recent years. Trend topic analysis (Fig. 8C) indicated that recent research (2022–2023) has focused on minimally invasive adrenalectomy, osilodrostat, primary bilateral macronodular adrenal hyperplasia, and the desmopressin test. Burst keyword analysis (Fig. 8D) identified corticotropin-releasing hormone and differential diagnosis as having the highest citation burst intensities. Currently, burst keywords include 11 beta-hydroxylase inhibitors, ARMC5 mutation, and somatic mutation.

Although keywords such as pituitary surgery, desmopressin test, and minimally invasive adrenalectomy have relatively low frequencies, they may represent emerging directions for future research. Additionally, primary bilateral macronodular adrenal hyperplasia and ARMC5 mutation both associated with somatic mutation, along with therapeutic agents like metyrapone, osilodrostat, and 11 beta-hydroxylase inhibitors, are likely to remain highly relevant research targets.

A comprehensive analysis of keyword co-occurrence, overlay, bursts, and trend reajectories suggests that somatic mutation and 11 beta-hydroxylase inhibitors currently represent the most active research frontiers in the field of CS.

This study presents a comprehensive bibliometroc analysis of global research output on CS using data from the Web of Science Core Collection. A total of 4661 articles were retrieved, published across 1021 journals by 3972 institutions from 86 countries. Scientometric tools, including VOSviewer and CiteSpace, were applied to map the intellectual structure and emerging trends in the field.

The publication trend from January 2020 to October 2023 revealed a notable increase in the volume of CS-related articles, reflecting the field’s growing importance and research momentum. The Eunice Kennedy Shriver National Institute of Child Health and Human Development (USA) and the University of Padua (Italy) emerged as the leading institutions, contributing significantly to global CS research. Based on citation metrics and impact factor rankings, The Journal of Clinical Endocrinology & Metabolism, European Journal of Endocrinology, and Clinical Endocrinology were identified as the most influential publication venues, with a strong focus on clinical research. These journals maintain robust citation relationships with high-impact titles such as New England Journal of Medicine and The Lancet, both of which are ranked in the Q1 quartile of Journal Citation Reports. Notably, The Lancet has the highest impact factor (IF = 168.90). Collectively, these core journals play a central role in disseminating research related to diagnosis, treatment, and patient management in CS, thereby shaping the direction and quality of scholarly contributions in the field.

Analysis of highly cited references and author contributions demonstrates that CS research over the past 2 decades has emphasized the development of diagnostic criteria and clinical practice guidelines. Diagnostic strategies have included measurement of urinary free cortisol, late-night salivary cortisol, and dexamethasone suppression testing, along with more recent approaches such as hair cortisol. Treatment guidelines have focused on surgical interventions and adjunctive therapies, particularly steroidogenesis inhibitors.^[9–16] The prominence of clinical practice guidelines among the most cited references highlights the field’s commitment to translating evidence into standardized care pathways. Core research themes have included pathophysiological mechanisms (especially genetics), diagnostic accuracy, treatment efficacy, and ong-term outcomes including quality of life. Keywords and trend analyses (Fig. 8A–D) identified 2 major and evolving research frontiers.