Global Research Trends of Diabetes Mellitus and Metabolic Reprogramming: A Bibliometric and Visualization Analysis

Summary

This bibliometric analysis aimed to map the research landscape of diabetes mellitus and metabolic reprogramming (2016–2025) using 6,864 papers. It reveals rapid growth, China’s leading publication output, but moderate academic influence, and a shift toward gut microbiota, machine learning, and multi-omics integration.

Abstract

Diabetes mellitus (DM) is a global public health crisis. Metabolic reprogramming alters cellular energy production and plays a key role in DM pathogenesis, yet a systematic overview of research trends, collaboration networks, and emerging frontiers in this field is lacking. This study performed a bibliometric analysis of 6,864 English-language articles and reviews published between 2016 and 2025, retrieved from Web of Science, PubMed, Scopus, and Embase. The analytical workflow included: (1) annual publication trend analysis, (2) country/region and institutional collaboration mapping, (3) author and journal co-authorship and co-citation analysis, (4) reference co-citation and burst detection, and (5) keyword and topic trend analysis, using three bibliometric analysis tools. Annual publications remained below 800 from 2016 to 2024 but exceeded 1,000 in 2025, indicating rapid growth. China contributed the largest share of papers (50%) and accounted for many of the top institutions, yet centrality analysis showed that the United States and European countries still lead in academic influence and knowledge generation. The knowledge structure is built on three core foundations: insulin resistance, oxidative stress, and mitochondrial function. Emerging research focuses on gut microbiota, machine learning, and metabolic reprogramming. Keyword bursts and topic trend maps reveal a paradigm shift from static phenotype observation to dynamic, multi-omics-integrated systems medicine. To our knowledge, no prior bibliometric study has specifically focused on the intersection of DM and metabolic reprogramming; this study thus provides the first dedicated mapping of this field. While China dominates in publication volume, centrality metrics indicate room for stronger international collaboration and academic leadership. Future efforts should promote interdisciplinary cooperation and support basic-to-clinical translation. This study provides an objective reference for research management and strategic planning.

Introduction

Rapid societal changes over the past few decades have significantly altered nutritional patterns and lifestyles. These changes have contributed to a steady rise in the global prevalence of diabetes mellitus (DM). As a growing public health crisis, DM and its associated complications pose a serious threat to population health. These complications include cardiovascular diseases and end-stage renal disease, among others. They substantially reduce the quality of life for those affected¹. Currently ranked as the ninth leading cause of mortality worldwide, DM affects approximately 1 in 11 adults globally². DM is a chronic metabolic disease primarily caused by either insufficient insulin production or the body's impaired response to insulin, leading to persistently high blood sugar levels³. The pathogenesis of DM and its complications is highly complex, with metabolism-related mechanisms remaining a major focus of current research.

Metabolic reprogramming refers to the process by which cells alter their metabolic pathways to enhance energy production, primarily through mitochondrial oxidative phosphorylation (OXPHOS) and glycolysis⁴. The classic example of metabolic reprogramming is the “Warburg effect,” first proposed by Professor Otto Warburg in 1927. He observed that even under adequate oxygen conditions, tumor cells can generate ATP through glycolysis⁵. Beyond cancer, metabolic reprogramming through these processes contributes to the pathogenesis of numerous diseases, including gestational DM, diabetic nephropathy, sepsis, atherosclerosis, stroke, and hepatic fibrosis^4,6,7,8,9,10. DM is a chronic metabolic disorder characterized by hyperglycemia, resulting from varying degrees of insulin resistance and impaired insulin secretion¹¹. The link between metabolic reprogramming and DM is well-established: nutrient excess leads to abnormal lipid metabolism in the liver and muscle, promoting insulin resistance, while pancreatic β-cells suffer metabolic disruption and secretory dysfunction due to lipotoxicity, endoplasmic reticulum stress, and amyloid deposition, collectively driving the onset of DM¹².

In recent years, the rapid advancement of single-cell technologies and multi-omics analyses has catalyzed a major paradigm shift in metabolism research. The field has moved beyond static observation toward dynamic, mechanistic inquiry, opening new opportunities to clarify the pathogenesis of DM and identify potential therapeutic targets. Despite this progress, however, no systematic synthesis of global research trends, key contributors, and collaborative networks currently exists. A preliminary search of major databases (PubMed, Web of Science, Scopus) using terms related to “diabetes mellitus” and “metabolic reprogramming” identified no prior bibliometric study on this topic. More importantly, existing narrative reviews, systematic reviews, and meta‑analyses have focused on summarizing specific molecular mechanisms or clinical outcomes but have not provided a holistic, data‑driven mapping of the field’s evolution, knowledge structure, or emerging frontiers. Consequently, the current knowledge structure in this area remains largely unmapped. Unlike a narrative review, which may be subjective and dependent on the author’s expertise, bibliometric analysis offers a reproducible, quantitative, and unbiased method to capture large‑scale publication patterns, collaboration networks, and research trends. Instead, it reveals the intellectual, social, and conceptual structure of a research field over time. As such, bibliometric analysis can show what a field has studied, who has contributed, how knowledge flows, and where it is heading. However, it is important to recognize what this method cannot reveal. Bibliometric analysis does not evaluate the quality or validity of individual studies, nor can it assess clinical efficacy or mechanistic causality. It also depends on the coverage and accuracy of the underlying databases and may miss non‑English literature, preprints, or emerging journals. Thus, bibliometric analysis serves as a macro‑level mapping tool that complements, rather than replaces, traditional evidence synthesis methods.

To our knowledge, no prior bibliometric study has specifically focused on the intersection of DM and metabolic reprogramming. The study aims to leverage scientific literature data at scale to systematically uncover the evolution, intellectual foundation, hotspot dynamics, and emerging frontiers in this research area. The findings are expected to provide an objective basis and strategic insights for future scientific advancement and research planning.

Protocol

Data source and search strategy
To improve coverage and reduce source bias of the literature, four electronic databases were searched: Web of Science (including the Web of Science Core Collection, WoSCC), PubMed, Scopus, and Embase. The publication period was limited to January 1, 2016, to December 31, 2025. Only original research articles and review articles written in English were included. The search terms focused on two concepts: “diabetes mellitus” (DM) and “metabolic reprogramming”. The search strategies and all exclusion criteria used during the screening process are summarized in Figure 1.

The exact search strings used for each database are provided below:
Web of Science (WoSCC):
TS=((diabetes OR “diabetes mellitus” OR “type 1 diabetes” OR “type 2 diabetes” ) AND (“metabolic reprogramming” OR “metabolic remodeling” OR “glucose metabolism” OR “Warburg effect” OR “cellular metabolism” OR “mitochondrial dysfunction”))
*Filters: Publication years: 2016-2025; Document types: Article, Review; Language: English.*

PubMed:
(“diabetes mellitus”[Title/Abstract] OR “type 2 diabetes”[Title/Abstract] OR “type 1 diabetes”[Title/Abstract] OR “T2DM”[Title/Abstract] OR “T1DM”[Title/Abstract]) AND (“metabolic reprogramming”[Title/Abstract] OR “metabolic reprogram”[Title/Abstract] OR “ metabolic remodeling”[Title/Abstract] OR “Warburg effect”[Title/Abstract] OR “cellular metabolism”[Title/Abstract] OR “cellular metabolism”[Title/Abstract] OR “mitochondrial dysfunction“[Title/Abstract])
*Filters: Publication date: 2016/01/01 – 2025/12/31; Article type: Article, Review; Language: English.*

Scopus:
TITLE-ABS-KEY(“diabetes mellitus” OR “diabetes” OR “type 1 diabetes” OR “type 2 diabetes” OR “T1DM” OR “T2DM”) AND TITLE-ABS-KEY(“metabolic reprogramming” OR “metabolic reprogram” OR “mitochondrial dysfunction“ OR “metabolic remodeling“ OR “Warburg effect”)
*Filters: Publication year: 2016-2025; Document type: Article, Review; Language: English.*

Embase:
('diabetes mellitus'/exp OR 'diabetes mellitus':ab,ti OR 'type 2 diabetes':ab,ti OR 'type 1 diabetes':ab,ti) AND ('metabolic reprogramming':ab,ti OR 'metabolic reprogram':ab,ti OR 'Warburg effect':ab,ti OR 'metabolic remodeling':ab,ti)
*Filters: Publication year: 2016-2025; Document types: Article, Review; Language: English.*
All exclusion criteria used during the screening process are summarized in Figure 1. Records retrieved from each database were exported in plain text format, including full bibliographic information (title, authors, journal, publication year, abstract, keywords, and cited references).

Deduplication and Manual Screening Process
To ensure reproducibility, a stepwise deduplication and manual screening process was adopted: first, all plain text files exported from the four databases were imported into reference management software, and automatic deduplication was performed using the “References > Find Duplicates” function; then, the deduplicated records were exported to Excel format using the “Show All Fields” export style and manually screened based on titles and abstracts. The inclusion criteria were original research or review articles that clearly addressed diabetes (any type) and metabolic reprogramming (including glycolysis, oxidative phosphorylation, the Warburg effect, or mitochondrial metabolic adaptation), written in English. Exclusion criteria included editorials, letters to the editor, conference abstracts, case reports, animal studies unrelated to humans, and studies focusing solely on metabolic syndrome without addressing diabetes. Ultimately, 6,864 records were included in the bibliometric analysis (Supplementary File 1).

Software and methodologies for bibliometric analysis
This study employed three bibliometric analysis tools (Table of Materials) to systematically examine authors, institutions, sources, titles, keywords, cited references, and other detailed information from the articles. The raw data of the 6,864 studies (Supplementary File 1) served as the input for all three software tools. Bibliometric software #1 is a Java-based free software tool and assists researchers in analyzing, visualizing, and interpreting scholarly literature, knowledge networks, collaborative relationships, and research hotspots¹³. Bibliometric software #2 is a visualization tool for scientific bibliometrics and knowledge mapping, primarily employed to reveal development trends, research hotspots, and collaborative networks in academic fields¹⁴. Bibliometric software #3 is an R package for scientometric analysis and knowledge mapping, used to import, analyze, and visualize bibliographic data¹⁵. The detailed operating procedures and specific parameter settings for each software are provided in the Table of Materials.

Results

Analysis of publication trends (2016-2025)
From 2016 to 2025, a total of 6,864 publications focusing on DM and metabolic reprogramming were identified, spanning a 10-year period. Figure 2A illustrates the annual publication trends in this research field. Between 2016 and 2024, the yearly output remained below 800 articles. In 2025, the number exceeded 1,000 for the first time. As of the current date, the cumulative publication count has reached 6,864. This indicates that...

Discussion

This study systematically maps the intellectual, social, and conceptual structure of this emerging field over the past decade. Based on 6,864 publications retrieved from four databases, the findings reveal rapid growth in annual output, a dominant yet moderately centralized role for China in publication volume, and a knowledge structure centered on insulin resistance, oxidative stress, and mitochondrial function. Emerging frontiers include gut microbiota, machine learning, and metabolic reprogramming itself, signaling a ...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
VOSviewer (version 1.6.20)	Centre for Science and Technology Studies (CWTS), Leiden University, Netherlands	https://www.vosviewer.com/	Menu path for co-authorship analysis: Create > Create a map based on bibliographic data > Read data from bibliographic database files >  Select TXT files > Type of analysis: Co-authorship  > Unit of analysis: Authors (or Organizations) >  Counting method: Full counting > Minimum number of documents: 5 > Number of items to select: up to 1000 > Click “Next” > Verify selected items > Click “Create map” > Choose visualization parameters (attraction=2, repulsion=0, cluster resolution=1.00) > Save as .png or .txt. Menu path for co-citation analysis: Create > Create a map based on bibliographic data > Type of analysis: Co-citation > Unit of analysis: Cited references > Minimum number of citations: 10 > Remaining steps same as above.
CiteSpace (version 6.3.R1)	Chaomei Chen, Drexel University, USA	https://citespace.podia.com/	After running the analysis by selecting Keyword or Cited Reference in the Node Types, click Burstness and set the Top N parameter to 20 to detect the top 20 burst keywords or burst references with the highest frequency or citation counts.
bibliometrix (version 4.1.2)	Massimo Aria and Corrado Cuccurullo; R package	https://www.bibliometrix.org/	In RStudio, run library(bibliometrix) and biblioshiny(); in the opened biblioshiny web interface, perform the following operations sequentially: import data via “Data > Import” by selecting the file, setting the format to “RIS” or “BibTeX”, then clicking “Convert” followed by “Start”; generate annual scientific production via “Descriptive Analysis > Annual Scientific Production > Generate”; generate the country collaboration network via "Collaboration > Country Collaboration Network > Generate"; generate a thematic map via "Conceptual > Thematic Map" by setting the minimum word frequency to 5 and the number of clusters to 3, then clicking "Generate"; generate a keyword burst graph via “Conceptual > Keyword Burst > ReferencePublicationYearSpectra > Generate”