This investigation is a systematic scoping review and bibliometric analysis conducted in accordance with the PRISMA Extension for Scoping Reviews (PRISMA-ScR) (16).

Population: Publications with at least one mainland-China institutional address that investigate the Mediterranean diet.

Intervention: Conceptualization or application of Mediterranean-diet principles, including Jiangnan-diet localization.

Comparators: Not applicable; field-level scoping review without comparative arms.

Outcomes: Bibliometric indicators (publication volume, citation impact, thematic clusters) and reported health domains (cardiometabolic, microbiome, sustainability).

Study designs: Original research articles, reviews, guidelines and clinical trials indexed in Web of Science Core Collection (WoS) and China National Knowledge Infrastructure Core Collection (CNKI) between 1 January 2006 and 2 February 2025 (17). Editorials, letters, basic laboratory studies without human subjects, animal experiments and single-case reports were excluded.

A bilingual search query was used for each database to maximize inclusivity. For Web of Science, the following syntax was applied: (TS = (Mediterranean diet) AND CU=CHINA) AND (DT==(“ARTICLE” OR “REVIEW”)), Timespan: 2006-01-01 to 2025-02-02 (Publication Date) | Exact search. For CNKI, the equivalent Chinese query was zhuti = “Dizhonghaiyingshi.” The search results were exported in UTF-8 encoding as .txt files (WoS) and .nbk files (CNKI).

Duplicate removal was conducted using Bibliometrix (R, version 4.2), which employs an automated author–title matching algorithm to identify duplicate records across databases. All flagged duplicates were manually verified by two independent reviewers to ensure accuracy.

Eligibility screening was performed in two stages, following PRISMA-2020 guidelines (18). In the first stage, database filters were used to exclude records that were not original research articles, such as reviews, editorials, letters, or commentaries. In the second stage, two independent reviewers assessed the titles and abstracts (and full texts where necessary) according to a predefined PEO framework: Population = humans in China; Exposure = Mediterranean diet; Outcomes = health, sustainability, or adherence-related measures.

Exclusion criteria were: (i) editorial content or commentaries; (ii) basic laboratory studies without human subjects; (iii) animal experiments; and (iv) single-case reports. Any disagreements were resolved by consultation with a third reviewer.

We extracted the following variables from each included record: (i) Bibliographic metadata: author(s), year of publication, title, journal, country/region of affiliation, institutional address, and document type; (ii) Study characteristics: health domain (e.g., cardiometabolic, microbiome, sustainability), study design (original research, clinical trial, review, guideline), and population focus; and (iii) Keyword and thematic variables: author keywords, MeSH-standardized keywords, and thematic cluster assignment.

Data charting was performed using a standardized extraction form that was pre-tested on a 5% sample of records to ensure consistency. Two reviewers (JD and SX) independently charted all data in duplicate. Any discrepancies were resolved through discussion until consensus was reached. No additional contact with study authors was required to obtain or confirm data.

Formal risk-of-bias tools (e.g., RoB 2, AMSTAR-2) are designed for clinical outcomes and are inapplicable to bibliometric metadata. Instead, potential biases—language restriction, database coverage and affiliation misclassification—are detailed in the Limitations subsection.

This scoping review was conducted using publicly available data from previously published studies and did not involve human participants, identifiable personal data, or interventions. Therefore, approval from an ethics committee was not required. All procedures followed the principles of responsible research and publication ethics in accordance with the guidelines of the Committee on Publication Ethics (COPE).

The authors affirm that this work is original, free from plagiarism or duplicate publication, and that all sources have been properly acknowledged. All listed authors meet established authorship criteria. Any conflicts of interest have been fully disclosed. The authors also commit to correcting or retracting the work if any major errors or misconduct are identified. All datasets and analysis scripts are publicly archived and available.

A total of 970 records were identified through database searching (Web of Science = 866; CNKI = 104). After automated and manual duplicate removal (n = 19) and automated ineligibility screening (n = 17), as well as removal of unrelated or incomplete records (n = 418), 516 records were retained for screening. Of these, 59 were excluded based on title/abstract assessment, leaving 457 full-text articles for eligibility assessment.

During full-text eligibility screening, the following were excluded: Commentaries (n = 15); Letters (n = 6); Animal model testing (n = 28); Molecular biology research without human participants (n = 22); No relevant content (n = 2). This yielded a final set of 384 articles for inclusion. Inter-rater reliability for screening was high (Cohen’s κ = 0.87). The complete selection process is presented in Figure 1.

Between 2006 and the search cut-off on 2 February 2025, a total of 384 eligible articles on Mediterranean diet research in China were identified. The annual publication volume increased substantially, from just one article in 2006 to a peak of 76 in 2022, corresponding to a CAGR of 23%. Output declined modestly to 55 articles in 2023 and stabilized at 57 in 2024, suggesting a pandemic-related surge followed by a post-pandemic plateau in scholarly production (Figure 2).

Citation metrics followed a similar upward trend. By 2023, the cumulative citation count surpassed 4,000, and the FWCI for the corpus reached 1.34, indicating that Chinese MedDiet articles are cited 34% more frequently than the global medicine baseline (data available in Supplementary Table S1). These Figures underscore both the quantitative expansion and qualitative impact of MedDiet scholarship in China over the past two decades.

Chinese Mediterranean diet research is primarily concentrated in coastal academic clusters. Table 1 highlights the three leading regional hubs—Beijing, Hong Kong, and the Zhejiang–Shanghai axis—which together account for 33% of the total publications and 38% of all citations. Peking University leads in output with 49 articles, while the Zhejiang–Shanghai collaboration (Zhejiang University and Fudan University) achieves the highest citation count (559) and the greatest TLS (412), indicating a central position in the national collaboration network.

At the author level, trans-Pacific collaborators rank highest in citation density. Willett W.C. (Cit/Doc = 73.2) and Pan A. (63.2) exemplify this trend through their extensive U.S.–China cohort studies. Network analysis identifies Li Y. (betweenness centrality = 0.19) and Woo J. (0.17) as strategic “boundary spanners” who connect the Beijing and Hong Kong clusters, respectively, enhancing national knowledge diffusion (Supplementary Table S1).

Geographic collaboration patterns show an intensifying east–west knowledge flow. Figure 3 presents the inter-provincial co-authorship network, with node size proportional to publication volume and edge width indicating co-authorship frequency. Notably, the proportion of edges linking inland provinces increased from 31% during Phase I (2006–2013) to 44% during Phase III (2020–2025), suggesting a gradual diffusion of research capacity beyond coastal strongholds—likely reflecting policy initiatives such as Healthy China 2030 that incentivize regional research partnerships.

Author-keyword co-occurrence analysis (threshold:≥5 appearances) identified 80 high-frequency terms, which were grouped into seven coherent thematic clusters using association-strength normalization in VOSviewer. These clusters account for 81% of TLS in the corpus. Clustering quality was deemed acceptable (Louvain modularity = 0.46; average silhouette = 0.71). The network map is shown in Figure 4, and quantitative details appear in Table 2.

Cluster 1 (purple) centers on the term “Mediterranean diet” and encompasses prevention-oriented research addressing chronic disease risk reduction. Cluster 2 (red) focuses on cardiovascular endpoints (e.g., coronary heart disease, risk), often linked with adherence—indicating a shift from epidemiological replication to behavioral implementation.

Microbiome research forms two distinct clusters: Cluster 3 (orange) links dietary patterns to gut microbiota and inflammation pathways, while Cluster 7 (teal) extends to the gut–brain axis, obesity, and Alzheimer-related outcomes. Cluster 4 (green) captures localized adaptations, with terms such as “Chinese,” “diet quality,” and “diabetes,” reflecting attempts to recalibrate MedDiet metrics to native food patterns. Cluster 5 (blue) emphasizes hypertension and blood pressure, whereas Cluster 6 (yellow) features mortality-focused cohort studies. Together, these clusters demonstrate a thematic diversification from cardiometabolic endpoints to digital adherence, microbiomics, and contextual adaptation.

To trace thematic evolution, we integrated results from LDA (K = 7, C_V = 0.53) and a VOSviewer-generated keyword overlay map (Figure 5). The literature was segmented into three distinct phases based on dominant topics and term frequency distributions (Table 3).

Phase I (2006–2013) is characterized by replication of Western cardiometabolic trials, with key terms such as “cholesterol” and “blood pressure.” Phase II (2014–2019) marks a pivot toward obesity and diabetes in the general population, accompanied by early mentions of the Jiangnan Diet. Phase III (2020–2025) introduces new thematic domains—gut microbiota, smartphone-based coaching, and life-cycle assessment (LCA)—reflecting broader trends in microbiomics, digital health, and sustainability. Notably, the share of publications focused on sustainability rose from under 2% in 2014 to 11% in 2024. This evolution signals a paradigm shift from biomedical replication toward culturally and technologically integrated systems-level research.

Bibliographic coupling (threshold: ≥20 shared references) revealed five major citation clusters (Figure 6). Cluster 1 (red) forms the central intellectual axis, anchored by Pan A., Chen Guochong, and van Dam R.M., and now includes Jiangnan Diet papers by Wang J., indicating convergence between localized dietary research and classical cardiometabolic replication.

Cluster 2 (green), led by Hu F.B. and China Kadoorie Biobank collaborators, includes large-scale epidemiological studies validating MedDiet scoring systems. Cluster 3 (blue) aggregates pediatric and community-intervention research by Xiao Xiong and Tang Dan. Cluster 4 (yellow) includes policy-oriented studies by Chen Junshi and Lv Jun that connect scientific literature to national nutrition guidelines. Cluster 5 (purple) focuses on cognitive and neuro-nutritional research but lacks Jiangnan-related content.

Network modularity (Louvain score = 0.41) indicates moderately distinct yet interconnected knowledge domains. Cross-cluster citation strength increased from 0.18 in 2006–2013 to 0.33 in 2020–2025, reflecting increasing interdisciplinarity and convergence of research aims.

To explore finer-grained shifts in vocabulary, term co-occurrence analysis (TCA) was performed across four time slices (2006–2010, 2011–2015, 2016–2020, 2021–2025). Key trends emerged in Figure 7.

Ingredient localization is evident in the lexical replacement of “olive oil + phenolics” with “rapeseed-oil phenolics” and “tea catechins,” indicating substitution of Mediterranean staples with local alternatives. Digital health terminology emerges post-2020, with frequent co-occurrences of “smartphone + adherence,” “m-health,” and “precision nutrition,” reflecting integration of mobile technologies into dietary research.

Sustainability discourse gains traction through the emergence of term-pairs such as “carbon footprint + life-cycle assessment” and “planetary-health diet + greenhouse gas,” which appear predominantly after 2020. Finally, microbiome research has evolved beyond general terms like “dysbiosis” to more mechanistic language, including “short-chain fatty acids” and “bile-acid metabolism.”

Collectively, TCA illustrates the conceptual maturation of the field, with vocabulary shifting from replication-oriented jargon to sustainability and localization-focused terminology.

Chinese MedDiet scholarship demonstrates a structured and multidimensional maturation trajectory. Quantitatively, publication output increased dramatically from a single article in 2006 to 76 in 2022, with FWCI (1.34) indicating above-average global influence. The sharp increase in Mediterranean diet publications after 2020 likely reflects heightened attention to dietary quality and immune resilience during the COVID-19 pandemic, when evidence-based dietary models attracted growing interest among clinicians and researchers. The subsequent decline after 2022 may be explained by the normalization of research priorities in the post-pandemic period, a shift in funding toward urgent domestic nutrition challenges (e.g., obesity, diabetes, food security), and increased scholarly engagement with alternative paradigms such as plant-based diets and the Jiangnan dietary model. These dynamics underscore that Mediterranean diet scholarship in China is shaped by evolving public health agendas and broader scientific trends, rather than following a simple linear trajectory.

Geographically, research capacity remains coastal-centric, led by Beijing, Hong Kong, and the Zhejiang–Shanghai corridor, yet cross-provincial collaboration has risen from 31 to 44%, reflecting broader dissemination of expertise.

Qualitatively, the field is no longer defined solely by cardiometabolic endpoints. Seven keyword clusters highlight the emergence of new domains such as microbiome–diet interactions, mobile-app–based adherence strategies, and environmental co-benefits of dietary reform. Temporal analyses confirm a three-phase evolution: initial cardiometabolic replication (2006–2013), expansion to public-health priorities and localization (2014–2019), and cross-disciplinary innovation (2020–2025). Bibliographic-coupling data reveal that Jiangnan-diet papers have transitioned from the periphery to the central clinical-citation school, signaling epistemic convergence.

Lexical drift analysis further corroborates this transformation: “olive oil + phenolics” pairs give way to “rapeseed-oil phenolics” and “tea catechins”; “smartphone + adherence” enters the lexicon; and “life-cycle assessment” and “carbon footprint” gain prominence after 2020. These findings collectively point to the emergence of a distinctive, China-adapted MedDiet paradigm that is digitally enabled, ecologically aligned, and culturally embedded.

Beyond publication trajectories, the health benefits of the Mediterranean diet are grounded in a synergistic matrix of nutrients and bioactive compounds. Polyphenols and carotenoids from fruits, vegetables, and wine exert potent antioxidant and anti-inflammatory effects (23); omega-3 fatty acids from fish and nuts reduce cardiometabolic risk and support cognitive function (24); monounsaturated fats from olive oil improve lipid metabolism (25); and dietary fiber from whole grains and legumes enhances glycemic control and fosters microbiome diversity (26). Collectively, these mechanisms underpin the Mediterranean diet’s cardiometabolic and neuroprotective effects and help explain the intensifying research attention in China (8). Their alignment with pressing national health challenges—such as obesity, diabetes, and healthy aging—further underscores the translational relevance of this dietary model.

Prior global bibliometric surveys reported CAGR estimates of 11–14% for MedDiet research up to 2020 (27, 28). The higher growth rate observed in China (23%) is largely attributable to post-2020 surges in microbiome and sustainability topics. Our analysis extends and deepens findings by Wang et al. (13) who documented initial trials of rapeseed-oil substitution but lacked network-level evidence of its mainstream adoption. Internationally, microbiome-oriented MedDiet studies began rising in 2018 (29), while China’s trajectory intensified after 2020, coinciding with the Precision Medicine Initiative and national investments in omics infrastructure (30).

Three macro-level forces appear to be shaping the field’s trajectory. First, policy alignment: the “Healthy China 2030” initiative endorses salt–oil ratio reform and plant-forward dietary models, explicitly referencing the Mediterranean diet as a template for chronic disease prevention (31). Second, digital infrastructure: with over 1.1 billion WeChat users, China offers a uniquely scalable platform for mobile health (mHealth) interventions (32). The rise of “smartphone + adherence” term-pairs in our corpus reflects the academic uptake of such tools (33). Third, ecological imperatives: the national dual-carbon pledge (peak carbon by 2030, neutrality by 2060) has stimulated interest in life-cycle metrics and carbon-footprint accounting in the food system (34). This aligns with recent findings that nearly half of food-system emissions originate during primary production (35), reinforcing the relevance of dietary shifts to climate goals.

This study adheres to the PRISMA-2020 extension for scoping reviews and bibliometric analyses, providing a transparent workflow and open-access data pipeline (40). The dual-language strategy (WoS and CNKI) enhances comprehensiveness relative to English-only reviews (41). Methodologically, triangulation across three analytical lenses—keyword co-occurrence, bibliographic coupling, and term co-occurrence—allowed for robust cross-validation of thematic, intellectual, and lexical patterns.

Limitations include: (i) the exclusion of 4% of articles indexed only in local journals outside Scopus, which may marginally underestimate citation metrics; (ii) temporal lag in citation-based metrics, which may underrepresent the coupling strength of recent publications (2024–2025); (iii) removal of very recent papers in the bibliographic coupling analysis due to short reference lists; and (iv) restriction of semantic analyses to titles and abstracts, which may miss full-text nuance, particularly in digital implementation studies.

First, future studies should integrate traditional epidemiological endpoints with life-cycle assessment to quantify the dual health-climate benefits of a Jiangnan-adapted Mediterranean dietary pattern. Nested cohort designs could assess correlations between diet-related carbon footprints and clinical outcomes, as recently piloted in large-scale European studies (42).

Second, mobile-health research should extend beyond short-term pilots. Cluster-randomized trials of ≥24 months are needed to evaluate whether app-based coaching can sustain behavioral change and reduce major cardiovascular events. Adaptive interventions incorporating gamification, push notifications, and peer support would offer insights for optimizing China’s billion-user mHealth ecosystem (38).

Finally, a multi-omics Jiangnan cohort—incorporating metagenomics, untargeted metabolomics, and food-omics (e.g., phenolic signatures of regional oils)—should be established to elucidate mechanistic pathways linking dietary polyphenols with inflammation, bile-acid metabolism, and gut–brain signaling. Recent Chinese studies involving > 900 participants demonstrate feasibility (43). Such a cohort could accelerate biomarker discovery, support personalized nutrition, and align with national digital health and climate-neutrality strategies.