Abstract
Background
Prediabetes is a well-established risk factor for cardiovascular diseases (CVD), stroke, and cancer, but whether this risk is uniform across all individuals remains uncertain. Additionally, heterogeneity within the prediabetic population, driven by mechanisms like lipid metabolism dysfunction and inflammation, has not been fully explored.
Methods
This study analyzed data from 3,065 prediabetic participants (mean age: 58.8 ± 9.7, 53.8% women) from the China Health and Retirement Longitudinal Study. We used generalized linear models, Cox regression, and subgroup analyses to evaluate the risk of progression to CVD, stroke, and cancer. To delineate distinct subgroups among individuals with prediabetes, we devised a stratification framework predicated on the coexistence of dyslipidemia, the triglyceride-glucose (TyG) index, and the C-reactive protein-triglyceride glucose (CTI) index.
Results
Prediabetic participants were classified into four subgroups: metabolically sensitive, inflammation-driven, metabolic lipotoxicity, and systemic metabolic-inflammatory (a subtype defined by concurrent dyslipidemia and elevated CTI values indicating combined metabolic and inflammatory dysregulation). Compared with the metabolically sensitive subtype, the inflammation-driven, metabolic lipotoxicity, and systemic metabolic-inflammatory subtypes exhibited significantly heightened risks of cancer, CVD, and stroke, even after adjustment for confounding variables (all p < 0.05). Furthermore, dose–response analyses revealed a robust linear association between increasing levels of CTI, low-density lipoprotein cholesterol, and TyG index and the incidence of cancer, CVD, and stroke.
Conclusions
Prediabetic individuals with inflammation and/or lipid metabolism dysregulation are at a markedly higher risk of developing CVD, stroke, and cancer. Tailored prevention and intervention strategies are necessary to address the distinct risk profiles within the prediabetic population.
Supplementary Information
The online version contains supplementary material available at 10.1186/s13098-025-02065-0.
Keywords: Prediabetes, Inflammation, Lipid metabolism, Cardiovascular diseases, Stroke
Introduction
Prediabetes, an intermediate stage of glucose dysregulation that often precedes type 2 diabetes, affected approximately 720 million individuals globally in 2021, with projections indicating this number will rise to 1 billion by 2045 [1–3]. In the United States, an estimated 38% of individuals with prediabetes progress to type 2 diabetes annually, underscoring the urgent need for effective prevention strategies to address this growing public health challenge. Epidemiological data from China indicate a prediabetes prevalence of approximately 35.2%, with rates rising to nearly 50% among adults aged 50 years and older [4]. Although prediabetes is widely recognized as a precursor to type 2 diabetes, its significance extends beyond this progression. Mounting evidence indicates that prediabetes independently elevates the risk of cardiovascular diseases (CVD) [5, 6], stroke [7, 8], and cancer [9, 10], even in the absence of progression to diabetes. These findings suggest that prediabetes represents a broader systemic metabolic dysfunction with far-reaching health consequences. However, the extent to which these risks are distributed across the prediabetic population remains poorly understood. It is unclear whether all individuals with prediabetes experience uniformly elevated risks, as this population is likely heterogeneous, encompassing subgroups with varying biological drivers and clinical outcomes. This lack of clarity underscores the need to disentangle the underlying heterogeneity within prediabetes and its implications for disease risk.
Emerging evidence has identified lipid metabolism dysfunction and chronic low-grade inflammation as two key mechanisms contributing to the pathogenesis of metabolic disorders, including diabetes and its complications [11–13]. Dyslipidemia, characterized by elevated triglycerides (TG), low high-density lipoprotein cholesterol (HDL-C), and increased atherogenic lipid particles, plays a central role in promoting atherosclerosis, endothelial dysfunction, and insulin resistance, all of which are critical drivers of CVD [14, 15]. Preclinical studies have elucidated that dyslipidemia in prediabetes exacerbates vascular pathology through lipid-induced oxidative stress and mitochondrial dysfunction [16]. For instance, mouse models of prediabetes demonstrate that elevated TG and low-density lipoprotein cholesterol (LDL-C) trigger reactive oxygen species (ROS) production, impairing endothelial nitric oxide synthase activity and promoting atherogenesis [17]. Clinically, a meta-analysis by Cai et al. [6] confirmed that prediabetes is associated with a 1.5-fold increased risk of CVD, mediated by dyslipidemia-driven endothelial dysfunction. Similarly, chronic inflammation, mediated by pro-inflammatory cytokines and immune cell activation, has been implicated in the development of vascular damage, metabolic dysregulation, and tumorigenesis [18]. Clinical studies have suggested that chronic inflammation in prediabetes activates NLRP3 inflammasome pathways, promoting systemic low-grade inflammation and plaque formation, key precursors to cancer, CVD and stroke [19–21].
While abovementioned two pathways have been independently associated with adverse health outcomes, their potential interplay and combined impact remain underexplored in prediabetes, a condition where metabolic abnormalities often coexist and interact, potentially amplifying the risk of complications. Given this complexity, it is critical to examine whether distinct subgroups of prediabetes, defined by their lipid and inflammatory profiles, exhibit differential risks for adverse health outcomes. Identifying such subgroups could provide valuable insights into the biological underpinnings of prediabetes-related complications and inform targeted prevention strategies. In particular, individuals with a "mixed" phenotype, characterized by concurrent lipid metabolism dysfunction and inflammation, may represent a high-risk subgroup warranting specific attention. Thus, the primary objective of this study was to investigate the heterogeneity within the prediabetic population by stratifying individuals into distinct subgroups based on lipid metabolism dysfunction and inflammation, and to evaluate their differential risks for incident cancer, CVD and stroke. Specifically, we utilized the triglyceride-glucose (TyG) index, a validated surrogate for insulin resistance, the C-reactive protein-triglyceride glucose (CTI) index, which integrates metabolic and inflammatory markers, and LDL-C, a key atherogenic lipid, to define these subgroups. The TyG index captures insulin resistance, a central feature of prediabetes that drives macrovascular complications. The CTI extends the TyG index by incorporating C-reactive protein (CRP) to reflect combined metabolic-inflammatory dysregulation. LDL-C was included due to its established role in atherosclerosis and its contribution to CVD risk in prediabetes. By examining these markers, we aimed to elucidate how their interplay contributes to the elevated risks of CVD, stroke, and cancer, providing insights into tailored prevention strategies for high-risk prediabetic subgroups.
Methods
Study design and population
Our study utilizes data from the China Health and Retirement Longitudinal Study (CHARLS) (http://charls.pku.edu.cn/en), a nationally representative longitudinal survey targeting individuals aged 45 years and older in China [22]. CHARLS employs a complex, probability-proportional-to-size sampling method to gather comprehensive health-related data from middle-aged and elderly populations across the country. The baseline survey, conducted between 2011 and 2012 (Wave 1), was supported by multistage probability sampling, and to date, three waves of follow-up data have been released (Wave 2 in 2013, Wave 3 in 2015, and Wave 4 in 2018). The CHARLS protocol was approved by the Institutional Review Board of the National Development Institute of Peking University (IRB00001052-11015), and all participants provided written informed consent prior to enrollment (https://charls.pku.edu.cn/en/About/About_CHARLS.htm) [22].
The selection process for the study population is illustrated in Fig. 1. A total of 11,847 participants from Wave 1 who were followed up in Wave 4 were initially included in the analysis. We excluded individuals younger than 45 years (n = 487), those lacking key data on fasting plasma glucose (FPG), glycosylated hemoglobin (HbA1c), TG, LDL-C and HDL-C (n = 3,748), and those with incomplete socio-demographic, lifestyle-related, or anthropometric information (n = 1,232). Additionally, individuals with abnormal values (mean ± 3 standard deviations) and without prediabetes or outcome information were excluded (n = 1,377). After applying these exclusion criteria, 3,065 participants with prediabetes were included in the final analysis.
Fig. 1.
Flowchart of participant selection. FPG: fasting plasma glucose, HbA1c: glycosylated hemoglobin, TG: triglycerides, LDL-C: low density lipoprotein-cholesterol; HDL-C: high-density lipoprotein cholesterol
Diagnosis of prediabetes
Prediabetes and its progression were defined in accordance with the diagnostic criteria established by the American Diabetes Association (ADA). Progression of prediabetes was categorized as either progression to diabetes or regression to normal fasting glucose (NFG). In this study, blood glucose status was assessed primarily based on the ADA standards for impaired fasting glucose. NFG was defined as FPG < 5.6 mmol/L and HbA1c < 5.7%; prediabetes was defined as FPG between 5.6 and 6.9 mmol/L or HbA1c between 5.7% and 6.4%; and diabetes was defined as FPG ≥ 7.0 mmol/L, HbA1c ≥ 6.5%, or a self-reported history of diabetes [1]. Of note, The ADA criteria also include impaired glucose tolerance (IGT), defined as a 2-h post-load glucose of 7.8–11.0 mmol/L following an oral glucose tolerance test (OGTT). However, IGT was not included in our definition of prediabetes due to the absence of OGTT data in the CHARLS database, which provides only FPG and HbA1c measurements.
Assessment of covariates
Covariates included sociodemographic characteristics self-reported by participants, such as age, gender, education level (categorized as primary school or below, middle school, and high school or above), place of residence, marital status, smoking and drinking history, as well as body mass index (BMI).
Outcomes of interest
The primary outcome of the study was CVD, while stroke and cancer were designated as secondary outcomes. The presence of heart disease was determined based on the question, “Has a doctor ever told you that you have been diagnosed with a heart attack, angina pectoris, coronary heart disease, heart failure, or other heart problems?” CVD was defined as self-reported heart disease and/or stroke. Similarly, the occurrence of stroke was ascertained through the question, “Has a doctor ever told you that you have been diagnosed with a stroke?” The incidence of cancer was determined using the question, “Has a doctor ever told you that you have been diagnosed with cancer?”.
Stratification of prediabetes into metabolic and inflammatory clusters
To further categorize patients with prediabetes, a stratification model was developed based on the presence of dyslipidemia, TyG and the CTI. Dyslipidemia was characterized by the presence of total cholesterol concentrations ≥ 240 mg/dL, TG concentrations ≥ 150 mg/dL, LDL-C ≥ 160 mg/dL, HDL-C concentrations < 40 mg/dL, or a self-reported history of dyslipidemia. The TyG index was calculated as ln TG (mg/dl) × FPG (mg/dl)/2]. The CTI was defined as 0.412*ln (CRP [mg/L]) + ln [TG (mg/dL) × FPG (mg/dL)/2]. This approach categorized patients with prediabetes into four distinct subtypes based on specific metabolic and inflammatory profiles: systemic metabolic-inflammatory type, characterized by the coexistence of dyslipidemia and a CTI value above the mean, indicating concurrent metabolic dysregulation and systemic inflammation; metabolic lipotoxicity type, defined by the presence of dyslipidemia with a CTI value at or below the mean, reflecting isolated lipid metabolic disturbances in the absence of significant inflammation; inflammation-driven type, represented by the absence of dyslipidemia but with a CTI value above the mean, suggesting that systemic inflammation is the primary feature; and metabolically sensitive type, denoted by the absence of dyslipidemia and a CTI value at or below the mean, indicative of minimal metabolic or inflammatory dysfunction.
Statistical analysis
To mitigate potential biases arising from incomplete data, we employed multiple imputation using fully conditional specification to address the issue of missing values. Baseline variables were reported as means ± standard error or medians (interquartile range) for continuous variables, and as frequencies (percentages) for categorical variables. Differences in continuous variables were assessed using two-sample t-tests, ANOVA, or Mann–Whitney U tests, as appropriate [23]. The associations between exposure variables and incident outcomes were evaluated using generalized linear models and multivariable Cox proportional hazards models, with results presented as odds ratios (ORs) or hazard ratios (HRs) along with 95% confidence intervals (CIs). Three models were developed for the CHARLS cohort: Model 1 was unadjusted (crude model); Model 2 adjusted for age, sex, education, and marital status; and Model 3 further adjusted for age, sex, education, marital status, smoking, drinking, and BMI. The starting point was defined as the baseline assessment in Wave 1 conducted between 2011 and 2012. Endpoints were defined as the first occurrence of CVD, stroke, or cancer, recorded through Wave 2–4 (2013–2018). Participants were censored at the earliest of the last follow-up (Wave 2–4), or loss to follow-up. Subgroup analyses were conducted to examine the associations between prediabetes and the risks of CVD, stroke, and cancer across various subpopulations stratified by age (< 60 vs. ≥ 60 years), gender, and smoking status. Restricted cubic spline (RCS) Cox proportional hazard regression models were employed to investigate the dose–response relationships between the TyG index, CTI, and the risks of CVD, stroke, and cancer in prediabetic participants. These models were used to assess both the overall and non-linear associations, with results adjusted for key covariates including age, sex, education, marital status, geographic location, smoking status, alcohol consumption, and BMI. Sensitivity analyses were performed to ensure the robustness of the findings. Participants who developed CVD, stroke, or cancer during or prior to Wave 2 were excluded to minimize potential reverse causation bias. Besides, an additional comparison between the results with and without multiple imputation was conducted. To address potential competing risks among CVD, stroke, and cancer, we employed a multinomial logistic regression model. The outcome was categorized into five levels: (i) no event, (ii) CVD only, (iii) stroke only, (iv) cancer only, and (v) multiple events. This model estimates the associations of prediabetes subtypes (reference: metabolically sensitive) with each outcome category, adjusting for age, sex, education, marital status, smoking, drinking, and BMI. All statistical analyses were performed using R software (version 4.4.0), with two-sided p values < 0.05 considered statistically significant.
Result
Population characteristics
Of the original study participants, a final analytic sample of 3,065 individuals was included in the analysis. The selection process excluded individuals with incomplete data or extreme values (Fig. 1). A detailed description of the selection process can be found in the supplementary materials. A total of 3,065 adults (mean age: 58.8 ± 9.7) were included with similar distributions observed across the subtypes (p = 0.283). Notable heterogeneity emerged in sex distribution, with women comprising 53.8% (n = 1,648) of the cohort. Education, marital, location, and BMI was consistent across subtypes. However, significant differences were observed in diabetes outcomes after a median follow-up period of four years. The overall prevalence of diabetes was 8.8%, ranging from 5.9% in the metabolic lipotoxicity group to 11.5% in the inflammation-driven group (p < 0.001) (Table 1). Metabolic and inflammatory profiles provided robust evidence supporting the reliability and distinctiveness of the subtype classifications. Dyslipidemia, for instance, was exclusively present in the metabolic lipotoxicity group (Table 1). Similarly, systemic inflammation, as indicated by CRP or CTI levels, was markedly elevated in the inflammation-driven group compared to other subtypes, particularly the metabolically sensitive group (p < 0.001).
Table 1.
Baseline characteristics of participants with prediabetes in our study
| Characteristics | Overall | Metabolically sensitive type | Inflammation-driven type | Metabolic lipotoxicity type | Systemic Metabolic-Inflammatory Type | p value |
|---|---|---|---|---|---|---|
| Numbers | 3065 | 1316 | 1047 | 255 | 405 | |
| Age (mean, SD) | 58.8 (9.7) | 58.5 (9.7) | 59.2 (10.0) | 59.2 (9.0) | 58.7 (8.9) | 0.283 |
| Sex (%) | < 0.001 | |||||
| Female | 1648 (53.8) | 650 (49.4) | 585 (55.9) | 157 (61.6) | 232 (57.3) | |
| Male | 1417 (46.2) | 666 (50.6) | 462 (44.1) | 98 (38.4) | 173 (42.7) | |
| Marital (%) | 0.316 | |||||
| Married | 2696 (88.0) | 1172 (89.1) | 909 (86.8) | 220 (86.3) | 358 (88.4) | |
| Non-married | 369 (12.0) | 144 (10.9) | 138 (13.2) | 35 (13.7) | 47 (11.6) | |
| Education (%) | 0.03 | |||||
| College or above | 116 (3.8) | 37 (2.8) | 42 (4.0) | 11 (4.3) | 26 (6.4) | |
| High school | 913 (29.8) | 403 (30.6) | 299 (28.6) | 71 (27.8) | 128 (31.6) | |
| Primary school or below | 2036 (66.4) | 876 (66.6) | 706 (67.4) | 173 (67.8) | 251 (62.0) | |
| Location (%) | < 0.001 | |||||
| City/town | 637 (20.8) | 228 (17.3) | 232 (22.2) | 55 (21.6) | 116 (28.6) | |
| Village | 2426 (79.2) | 1088 (82.7) | 813 (77.8) | 200 (78.4) | 289 (71.4) | |
| Smoking (%) | 0.089 | |||||
| Current smoker | 834 (27.7) | 373 (29.0) | 297 (28.8) | 55 (21.9) | 96 (23.9) | |
| Ex-smoker | 283 (9.4) | 116 (9.0) | 97 (9.4) | 21 (8.4) | 46 (11.4) | |
| Non-smoker | 1896 (62.9) | 799 (62.0) | 639 (61.9) | 175 (69.7) | 260 (64.7) | |
| Drinking (%) | 0.007 | |||||
| Drink but less than once a month | 228 (7.9) | 118 (9.6) | 62 (6.3) | 20 (8.3) | 27 (7.0) | |
| Drink more than once a month | 591 (20.5) | 279 (22.6) | 192 (19.5) | 39 (16.2) | 75 (19.5) | |
| None of these | 2064 (71.6) | 836 (67.8) | 733 (74.3) | 181 (75.4) | 282 (73.4) | |
| BMI (mean, SD) | 24.7 (45.0) | 22.9 (3.8) | 26.7 (75.1) | 23.1 (3.7) | 26.7 (26.0) | 0.152 |
| Diabetes after 4 years (%) | < 0.001 | |||||
| Yes | 269 (8.8) | 83 (6.3) | 120 (11.5) | 15 (5.9) | 45 (11.1) | |
| No | 2796 (91.2) | 1233 (93.7) | 927 (88.5) | 240 (94.1) | 360 (88.9) | |
| Dyslipidemia (%) | < 0.001 | |||||
| Yes | 660 (21.8) | 0 (0.0) | 0 (0.0) | 255 (100.0) | 405 (100.0) | |
| No | 2363 (78.2) | 1316 (100.0) | 1047 (100.0) | 0 (0.0) | 0 (0.0) | |
| CRP (mean, SD) | 2.6 (6.0) | 0.9 (1.0) | 4.6 (8.4) | 0.9 (0.7) | 3.9 (7.9) | < 0.001 |
| CTI (mean, SD) | 8.8 (0.7) | 8.2 (0.4) | 9.4 (0.5) | 8.3 (0.4) | 9.5 (0.5) | < 0.001 |
| TyG (mean, SD) | 8.7 (0.5) | 8.4 (0.4) | 9.0 (0.5) | 8.5 (0.3) | 9.2 (0.5) | < 0.001 |
triglyceride-glucose index (TyG) and the C-reactive protein-triglyceride glucose index (CTI)
Association between prediabetes clusters and the development of CVD, stroke, and cancer
Table 2 summarizes the findings from the generalized linear models evaluating the associations between prediabetes clusters and the progression to CVD, stroke, and cancer. In the unadjusted model, all three prediabetes clusters demonstrated significantly increased risks for CVD (OR Inflammation-Driven Type = 1.36, 95%CI: 1.12–1.64, p = 0.0017; OR Metabolic Lipotoxicity Type = 2.01, 95%CI: 1.51–2.69, p < 0.0001; OR Systemic Metabolic-Inflammatory Type = 2.07, 95%CI: 1.62–2.64, p < 0.0001), stroke (OR Inflammation-Driven Type = 1.36, 95%CI: 1.02–1.81, p = 0.04; OR Metabolic Lipotoxicity Type = 1.58, 95%CI: 1.02–2.44, p = 0.04; OR Systemic Metabolic-Inflammatory Type = 2.39, 95%CI: 1.71–3.34, p < 0.0001), and cancer (OR Metabolic Lipotoxicity Type = 2.19, 95%CI: 1.1–4.35, p = 0.03; OR Systemic Metabolic-Inflammatory Type = 2.18, 95%CI: 1.21–3.94, p = 0.01) compared to the metabolically sensitive type. After adjusting for potential confounding factors (Models 2 and 3), inflammation-driven, metabolic lipotoxicity, and systemic metabolic-inflammatory types remained significantly associated with higher risks of CVD, stroke, and cancer when compared to the metabolically sensitive type (Table 2).
Table 2.
Generalized linear models of the role of prediabetes clusters in assessing the development of CVD, stroke, and cancer
| Outcomes | Clusters | Model 1 | Model 2 | Model 3 | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Estimate | SE | OR (95%CI) | p value | Estimate | SE | OR (95%CI) | p value | Estimate | SE | OR (95%CI) | p value | ||
| CVD | Metabolically Sensitive Type | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
| Inflammation-Driven Type | 0.3 | 0.1 | 1.36 (1.12–1.64) | 0.0017 | 0.23 | 0.1 | 1.26 (1.04–1.53) | 0.02 | 0.14 | 0.11 | 1.16 (0.94–1.42) | 0.17 | |
| Metabolic Lipotoxicity Type | 0.7 | 0.15 | 2.01 (1.51–2.69) | < 0.0001 | 0.62 | 0.15 | 1.86 (1.38–2.49) | < 0.0001 | 0.67 | 0.15 | 1.96 (1.45–2.66) | < 0.0001 | |
| Systemic Metabolic-Inflammatory Type | 0.73 | 0.12 | 2.07 (1.62–2.64) | < 0.0001 | 0.64 | 0.13 | 1.9 (1.49–2.44) | < 0.0001 | 0.55 | 0.13 | 1.73 (1.33–2.25) | < 0.0001 | |
| Stroke | Metabolically Sensitive Type | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
| Inflammation-Driven Type | 0.3 | 0.15 | 1.36 (1.02–1.81) | 0.04 | 0.31 | 0.15 | 1.36 (1.01–1.82) | 0.04 | 0.3 | 0.16 | 1.35 (1–1.84) | 0.05 | |
| Metabolic Lipotoxicity Type | 0.46 | 0.22 | 1.58 (1.02–2.44) | 0.04 | 0.49 | 0.23 | 1.63 (1.05–2.54) | 0.03 | 0.48 | 0.23 | 1.61 (1.02–2.55) | 0.042 | |
| Systemic Metabolic-Inflammatory Type | 0.87 | 0.17 | 2.39 (1.71–3.34) | < 0.0001 | 0.9 | 0.17 | 2.46 (1.75–3.46) | < 0.0001 | 0.84 | 0.18 | 2.31 (1.62–3.31) | < 0.0001 | |
| Cancer | Metabolically Sensitive Type | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
| Inflammation-Driven Type | 0.12 | 0.27 | 1.13 (0.66–1.93) | 0.65 | 0.09 | 0.27 | 1.09 (0.64–1.87) | 0.75 | 0.1 | 0.28 | 1.11 (0.64–1.92) | 0.71 | |
| Metabolic Lipotoxicity Type | 0.78 | 0.35 | 2.19 (1.1–4.35) | 0.03 | 0.74 | 0.35 | 2.09 (1.05–4.16) | 0.04 | 0.74 | 0.36 | 2.09 (1.04–4.19) | 0.037 | |
| Systemic Metabolic-Inflammatory Type | 0.78 | 0.3 | 2.18 (1.21–3.94) | 0.01 | 0.73 | 0.3 | 2.07 (1.14–3.75) | 0.02 | 0.66 | 0.31 | 1.94 (1.05–3.58) | 0.034 | |
OR: odd ratios; CI: confidence interval; SE: standard error. Model 1 was unadjusted (crude model); Model 2 adjusted for age, sex, education, and marital status; and Model 3 further adjusted for age, sex, education, marital status, smoking, drinking, and body mass index (BMI, continuous)
Bold means statistically significant
To further explore the relationship between prediabetes clusters and the incidence of CVD, stroke, and cancer, we employed multivariable Cox proportional hazards models. Similarly, regardless of whether the potential confounders were adjusted, a significant positive association between inflammation-driven, metabolic lipotoxicity, and systemic metabolic-inflammatory prediabetes clusters and the risks of CVD, stroke, and cancer was consistently observed across different models (Fig. 2). Notably, the elevated OR observed for the metabolic-inflammatory prediabetes cluster, in comparison to the inflammation-driven and metabolic lipotoxicity clusters, indicates that the coexistence of inflammation and lipid metabolic dysfunction confers a heightened risk for the development of CVD (OR Inflammation-Driven Type = 1.24, 95%CI: 1.04–1.48, p = 0.017; OR Metabolic Lipotoxicity Type = 1.84, 95%CI: 1.44–2.35, p = 1.06E-06; OR Systemic Metabolic-Inflammatory Type = 1.76, 95%CI: 1.43–2.17, p < 0.0001), stroke (OR Inflammation-Driven Type = 1.32, 95%CI: 0.99–1.77, p = 0.058; OR Metabolic Lipotoxicity Type = 1.57, 95%CI: 1.02–2.42, p = 0.04; OR Systemic Metabolic-Inflammatory Type = 2.21, 95%CI: 1.59–3.08, p < 0.0001), and cancer (OR Metabolic Lipotoxicity Type = 2.06, 95%CI: 1.04–4.06, p = 0.037; OR Systemic Metabolic-Inflammatory Type = 1.9, 95%CI: 1.04–3.47, p = 0.036) (Table 2, Fig. 2). The proportional hazards assumption for the Cox models was evaluated, with results presented in Supplementary Table S1. The P values for covariates and global models exceeded 0.05, confirming the validity of the proportional hazard’s assumption.
Fig. 2.
Multivariable Cox proportional hazards models of the role of prediabetes clusters in assessing the development of CVD, stroke, and cancer. HR: hazard ratios; CI: confidence interval; SE: standard error. Model 1 was unadjusted (crude model); Model 2 adjusted for age, sex, education, and marital status; and Model 3 further adjusted for age, sex, education, marital status, smoking, drinking, and body mass index (BMI, continuous)
Subgroup analysis
Subgroup analyses stratified by age, gender, and smoking status were performed to examine the associations of CTI and TyG with the risk of CVD and stroke in various subgroups (Fig. 3). The findings demonstrated significant positive associations between CTI and TyG and the risk of CVD, as well as between TyG and the risk of stroke (Fig. 3A) across all subgroups analyzed. For instance, significant positive associations between CTI and CVD risk were observed in subgroups including individuals aged < 65 (OR = 1.21, 95% CI: 1.08–1.35, p = 0.001), females (OR = 1.29, 95% CI: 1.14–1.45, p < 0.0001), and individuals without smoking history (OR = 1.25, 95% CI: 1.11–1.40, p < 0.001). Similarly, significant associations between TyG and CVD risk were observed in individuals aged < 65 (OR = 1.22, 95% CI: 1.05–1.41, p = 0.01), and non-smokers (OR = 1.29, 95% CI: 1.10–1.50, p = 0.001) (Fig. 3B).
Fig. 3.
Subgroup analyses. A CTI in the risk of CVD and stroke; B TyG in the risk of CVD and stroke
Dose–response relationship and sensitivity analyses
The dose–response relationships between the CTI, LDL-C, and TyG indices and the risk of CVD, stroke, and cancer are illustrated in Fig. 4. These RCS curves reveal a significant positive linear association between CTI, LDL-C, and TyG and the incidence of CVD, stroke, and cancer, after adjusting for key covariates such as age, sex, education, marital status, geographic location, smoking status, alcohol consumption, and BMI (p-values overall associations < 0.05; p-values non-linear associations > 0.05). Specifically, the risk of CVD, stroke, and cancer exhibited a consistent increase with elevating CTI levels, with a pronounced acceleration observed beyond the inflection point at 7.83 (Fig. 4A-C). A comparable trend was observed for CVD and stroke with respect to TyG, where the risk escalated markedly after a threshold of 8.05 (Fig. 4D-E). Furthermore, at LDL concentrations exceeding 78.093 mmol/L, a notable rise in the risk of both CVD and stroke was evident (Fig. 4G-H). In sensitivity analyses, consistent results were observed after excluding participants who had experienced CVD, stroke, or cancer at baseline or during the first wave of follow-up (Supplementary Tables S2–S3). To enhance the transparency and robustness of our findings, we conducted an additional analysis comparing the results using both the complete-case dataset and the dataset with multiple imputation. As a result, the HRs for the associations between prediabetes clusters and the risks of CVD, stroke, and cancer remained consistent in direction and significance between the complete-case and imputed models (Supplementary Tables S4–S5). Supplementary Table S6 presents the results of the multinomial logistic regression model. Relative to the metabolically sensitive subtype, the Metabolic Lipotoxicity subtype exhibited a significantly elevated risk of isolated CVD (HR = 2.26, 95% CI: 1.62–3.14, p < 0.0001) and isolated cancer (HR = 3.77, 95% CI: 1.62–8.76, p = 0.002). The Systemic Metabolic-Inflammatory subtype was strongly associated with isolated CVD (HR = 1.70, 95% CI: 1.26–2.28, p < 0.0001), isolated stroke (HR = 2.35, 95% CI: 1.45–3.81, p < 0.0001), and multiple events (HR = 2.61, 95% CI: 1.61–4.21, p < 0.0001). In contrast, the Inflammation-Driven subtype showed no significant associations with any outcome category (Supplementary Table S6).
Fig. 4.
Restricted cubic spline Cox proportional hazard regression models. A CTI (unitless, derived from C-reactive protein [CRP, mg/L], triglycerides [TG, mg/dL], and fasting plasma glucose [FPG, mg/dL]) in the risk of CVD; B CTI in the risk of stroke; C CTI in the risk of cancer; D TyG (unitless, derived from TG [mg/dL] and FPG [mg/dL]) in the risk of CVD; E TyG in the risk of stroke; F TyG in the risk of cancer; G LDL-C (mg/dL) in the risk of CVD; H LDL-C in the risk of stroke; I LDL-C in the risk of cancer
Discussion
To the best of our knowledge, this is the first study to provide critical insights into the heterogeneity of prediabetes and its relationship with CVD, stroke, and cancer, emphasizing the distinct contributions of lipid metabolism dysfunction and chronic inflammation. By stratifying prediabetic individuals into four distinct subtypes—metabolically sensitive, inflammation-driven, metabolic lipotoxicity, and systemic metabolic-inflammatory—we identified significant differences in the risks of adverse health outcomes. Notably, the inflammation-driven, metabolic lipotoxicity, and systemic metabolic-inflammatory subtypes were consistently associated with higher risks of CVD, stroke, and cancer compared to the metabolically sensitive cluster.
The practical clinical impact of this study on the routine management of prediabetes lies in its ability to refine screening, prevention, and intervention approaches. For screening, integrating biomarkers like the CTI and TyG indices into routine blood tests can help clinicians identify high-risk subgroups—such as the inflammation-driven or systemic metabolic-inflammatory types—prompting earlier cardiovascular or metabolic assessments. In terms of prevention and intervention, the inflammation-driven subtype may benefit from anti-inflammatory strategies (e.g., Mediterranean diet or exercise), while the metabolic lipotoxicity subtype could require lipid-lowering therapies (e.g., statins) and dietary adjustments. The systemic metabolic-inflammatory subtype, with the highest risk, warrants a combined approach targeting both inflammation and lipid metabolism. These strategies enhance the precision of prediabetes management, moving beyond a uniform approach to address the distinct needs of each subgroup.
The elevated risk observed in the inflammation-driven cluster is consistent with the central role of chronic low-grade inflammation in vascular damage [24] and tumorigenesis [18]. Systemic inflammation, characterized by elevated levels of pro-inflammatory cytokines such as interleukin-6 and tumor necrosis factor-alpha, disrupts endothelial function, accelerates atherogenesis, and promotes insulin resistance—key drivers of CVD and stroke [25, 26]. Moreover, inflammation is closely linked to cancer development through mechanisms such as enhanced angiogenesis, inhibition of apoptosis, and DNA damage. Similarly, the metabolic lipotoxicity cluster demonstrated a high risk of adverse outcomes, highlighting the pathological role of dyslipidemia in prediabetes. Elevated TG, LDL and low HDL-C exacerbate endothelial dysfunction, promote the accumulation of atherogenic lipoproteins, and drive insulin resistance via lipid-induced mitochondrial stress [27, 28]. These mechanisms are critical in the development of CVD and, to a lesser extent, cancer.
Interestingly, the systemic metabolic-inflammatory cluster, which combines features of both lipid metabolism dysfunction and inflammation, exhibited the highest risk across all outcomes. The underlying mechanisms of additive effect may lie in the synergistic interplay between metabolic dysregulation and inflammation, amplifying disease progression through mechanism. One key pathway is the oxidative stress-inflammation feedback loop. Dysregulated lipid metabolism, particularly the overproduction of oxidized low-density lipoproteins, leads to excessive generation of reactive oxygen species (ROS) [29]. ROS activate key inflammatory signaling pathways, such as nuclear factor-kappa B and the NLRP3 inflammasome, which promote the release of pro-inflammatory cytokines like IL-1β and TNF-α [30]. This creates a positive feedback loop, wherein inflammation amplifies oxidative stress, further damaging cellular components and impairing metabolic processes. Furthermore, inflammation disrupts glucose homeostasis through its impact on insulin signaling pathways [31, 32]. Pro-inflammatory cytokines such as IL-6 and TNF-α inhibit the insulin receptor substrate-1/PI3K/Akt signaling cascade, leading to insulin resistance. Chronic inflammation also induces beta-cell dysfunction and apoptosis, worsening glycemic control and establishing a vicious cycle between metabolic dysregulation and immune activation. In addition to systemic effects, immune cell reprogramming plays a critical role in the metabolic-inflammatory cluster. Dysregulated lipid metabolism skews macrophage polarization toward the pro-inflammatory M1 phenotype while suppressing the anti-inflammatory M2 phenotype [33, 34]. Elevated levels of free fatty acids activate Toll-like receptor 4 signaling, which exacerbates adaptive immune responses and drives further cytokine production [35]. Furthermore, lipid abnormalities impair vascular endothelial function by inducing ROS production and local inflammation, promoting atherogenesis and increasing thrombosis risk. Targeting these mechanisms—such as oxidative stress, insulin resistance, and immune cell polarization—offers a promising therapeutic strategy to mitigate the compounding effects of metabolic dysregulation and chronic inflammation.
Our study advances prior research by presenting a stratified model capable of delineating biologically distinct subtypes within the prediabetic population. Through restricted cubic spline analyses, we observed a linear dose–response relationship between the TyG index, CTI, and disease risk, even after adjusting for confounding factors. These findings provide critical insights into the interplay between metabolic and inflammatory dysfunctions and their collective impact on disease progression. Insulin resistance, characterized by diminished cellular sensitivity and responsiveness to insulin, manifests as a critical precursor to type 2 diabetes, often emerging years before its clinical onset [36]. Previous studies have examined the TyG index’s relationship with CVD morbidity and mortality, with Zhang et al. [36] reporting a U-shaped association in individuals with diabetes or prediabetes and pre-existing CVD, and others identifying a positive correlation in patients with stable CVD [37]. The TyG index has also been validated as a robust surrogate marker for insulin resistance, showing strong correlation with the glucose clamp technique [38]. Similarly, CRP, a non-specific marker of systemic inflammation, is strongly associated with cancer, CVD, and stroke risk and serves as a robust biomarker for its prediction. The CTI index, initially described by Ruan et al. [39], integrates systemic inflammation and insulin resistance, and has been widely adopted in clinical research. Studies have demonstrated CTI’s association with cancer mortality and CVD prevalence in the general population, as well as its predictive role for stroke in hypertensive individuals [40]. However, these studies often treated prediabetes as a homogeneous entity, failing to account for its heterogeneity. In contrast, our findings reveal that the coexistence of lipid metabolism dysfunction and inflammation confers a disproportionately higher risk of CVD, stroke, and cancer, suggesting a synergistic impact. This aligns with emerging evidence that metabolic inflammation—defined as the intersection of metabolic dysregulation and chronic inflammation—represents a distinct pathogenic pathway in metabolic disorders.
This study has several limitations that warrant consideration. First, the reliance on self-reported data for key outcomes, including cancer, cardiovascular disease (CVD), and stroke diagnoses, introduces significant potential for misclassification bias (e.g., under- or over-reporting of diagnoses due to diagnostic inaccuracies or varying medical access) and recall bias (e.g., participants may inaccurately recall or report past medical events, leading to systematic errors in outcome ascertainment). These biases may affect the accuracy of our risk estimates and could lead to underestimation or overestimation of the associations between prediabetes subtypes and adverse outcomes. Objective clinical assessments, such as medical record verification or diagnostic imaging, would improve outcome accuracy but were not available in the CHARLS database. Second, the study lacks granularity in biomarker measurements. The CHARLS database provides only single-timepoint measurements of C-reactive protein (CRP) and lipid profiles, without repeated assessments or more specific inflammatory markers (e.g., interleukin-6, tumor necrosis factor-alpha, or oxidized low-density lipoprotein). This limits our ability to capture dynamic changes in metabolic-inflammatory crosstalk over time, potentially underestimating the strength of associations with CVD, stroke, and cancer. Third, while participants with self-reported CVD or stroke at baseline were excluded to focus on incident events, we observed that a significant proportion of cardiovascular events in prediabetic participants occurred within the first two years of follow-up. This raises challenges in excluding participants with pre-existing or predisposing macrovascular conditions prior to enrollment. Furthermore, due to limitations in the CHARLS database, other macrovascular comorbidities (e.g., peripheral artery disease, subclinical atherosclerosis, or other vascular conditions not explicitly captured by self-reported heart disease or stroke) were not systematically assessed at baseline, as the database relies primarily on self-reported diagnoses for these conditions. This may introduce residual confounding or misclassification bias, as undiagnosed or unreported macrovascular conditions could have been present and influenced our findings. To address this, we performed sensitivity analyses to ensure the robustness of our findings. Participants who developed CVD, stroke, or cancer during or prior to Wave 2 were excluded to minimize potential reverse causation bias. The results remained consistent with our primary findings, providing further confidence in the robustness of our conclusions regarding the associations between prediabetes subtypes and adverse outcomes. Fourth, the observational nature of the study precludes causal inference, and residual confounding cannot be entirely ruled out despite rigorous statistical adjustments. Future studies should employ longitudinal designs with repeated biomarker measurements to establish causal relationships between prediabetes subtypes and adverse outcomes. Fifth, the CHARLS database lacks detailed data on potentially influential factors, including dietary habits, family history (particularly first-degree relatives with cancer), and specific medication use (e.g., statins, metformin, or other drugs with cardiovascular protective or cardiotoxic effects). Consequently, we were unable to assess the potential confounding or mediating effects of these factors, which may limit the depth of our findings. Sixth, the low number of mortality events in the CHARLS database, combined with stringent screening criteria for prediabetes and exclusions for incomplete data, resulted in insufficient sample sizes per cluster to generate reliable Kaplan–Meier survival curves or absolute incident rate curves. This data limitation precluded these analyses and may affect the generalizability of time-to-event findings. Seventh, while the study population is representative of middle-aged and elderly individuals in China, the findings may not be generalizable to other populations with different genetic, environmental, and lifestyle factors. Multinational studies are needed to validate the applicability of these findings in diverse settings.
Conclusion
The traditional ‘one-size-fits-all’ approach to prediabetes management may be inadequate, as our findings reveal significant heterogeneity among prediabetes subtypes. Specifically, the inflammation-driven, metabolic lipotoxicity, and systemic metabolic-inflammatory clusters demonstrated significantly elevated risks of CVD, stroke, and cancer compared to the metabolically sensitive group. This stratification framework provides a nuanced classification of prediabetic individuals, facilitating tailored therapeutic and preventative strategies targeting specific metabolic and inflammatory phenotypes.
Supplementary Information
Acknowledgements
We appreciate the work by the CHARLS Study for their contributions.
Author contributions
Wenjie Li: Conceptualization, methodology, data curation, software and writing-review & editing. Jian Chen: Data curation and software. Wei Wang & Yiming Fu & Qin Zeng: Funding acquisition, project administration, supervision and validation. The work reported in the paper has been performed by the authors, unless clearly specified in the text.
Funding
None.
Data availability
Data that support the results of this study are available from the corresponding author upon request.
Declarations
Ethics approval and consent to participate
None.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Wenjie Li and Jian Chen Contributed equally.
Contributor Information
Yiming Fu, Email: yimingfu_gd@163.com.
Wei Wang, Email: wwei9500@smu.edu.cn.
Qin Zeng, Email: 16613673@qq.com.
References
- 1.Echouffo-Tcheugui JB, Perreault L, Ji L, Dagogo-Jack S. Diagnosis and management of prediabetes: a review. JAMA. 2023;329(14):1206–16. [DOI] [PubMed] [Google Scholar]
- 2.Saeedi P, Petersohn I, Salpea P, Malanda B, Karuranga S, Unwin N, et al. Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: results from the International Diabetes Federation diabetes Atlas, 9(th) edition. Diabetes Res Clin Pract. 2019;157:107843. [DOI] [PubMed] [Google Scholar]
- 3.Tabák AG, Herder C, Rathmann W, Brunner EJ, Kivimäki M. Prediabetes: a high-risk state for diabetes development. Lancet. 2012;379(9833):2279–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Li Y, Teng D, Shi X, Qin G, Qin Y, Quan H, et al. Prevalence of diabetes recorded in mainland China using 2018 diagnostic criteria from the American Diabetes Association: national cross sectional study. BMJ. 2020;369:m997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Cosentino F, Grant PJ, Aboyans V, Bailey CJ, Ceriello A, Delgado V, et al. 2019 ESC guidelines on diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD. Eur Heart J. 2020;41(2):255–323. [DOI] [PubMed] [Google Scholar]
- 6.Cai X, Zhang Y, Li M, Wu JH, Mai L, Li J. Association between prediabetes and risk of all cause mortality and cardiovascular disease: updated meta-analysis. BMJ. 2020;370:m2297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Khan MM, Roberson S, Reid K, Jordan M, Odoi A. Prevalence and predictors of stroke among individuals with prediabetes and diabetes in Florida. BMC Public Health. 2022;22(1):243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Schlesinger S, Neuenschwander M, Barbaresko J, Lang A, Maalmi H, Rathmann W, et al. Prediabetes and risk of mortality, diabetes-related complications and comorbidities: umbrella review of meta-analyses of prospective studies. Diabetologia. 2022;65(2):275–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Shen E, Chen X. Prediabetes and the risk of lung cancer incidence and mortality: a meta-analysis. J Diabetes Investig. 2023;14(10):1209–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Keesari PR, Jain A, Ganampet NR, Subhasri GSD, Edusa S, Muslehuddin Z, et al. Association between prediabetes and breast cancer: a comprehensive meta-analysis. Breast Cancer Res Treat. 2024;204(1):1–13. [DOI] [PubMed] [Google Scholar]
- 11.Petrenko V, Sinturel F, Riezman H, Dibner C. Lipid metabolism around the body clocks. Prog Lipid Res. 2023;91:101235. [DOI] [PubMed] [Google Scholar]
- 12.Li W, Dong P, Wang W. Association of systemic inflammation markers in cancer mortality with diabetes: evidence from national health and nutrition examination survey. Acta Diabetol. 2024;61(11):1403–12. [DOI] [PubMed] [Google Scholar]
- 13.Huang Y, Qin L, Guangtao L, Dong P, Xiong X, Wan Y. Inflammatory responses of high mobility group protein B1 in disease: current trends, novel insights, and challenges. Life Conflux. 2024;1(1):e106. [Google Scholar]
- 14.Wu L, Parhofer KG. Diabetic dyslipidemia. Metabolism. 2014;63(12):1469–79. [DOI] [PubMed] [Google Scholar]
- 15.Mangione CM, Barry MJ, Nicholson WK, Cabana M, Chelmow D, Coker TR, et al. Statin use for the primary prevention of cardiovascular disease in adults: US preventive services task force recommendation statement. JAMA. 2022;328(8):746–53. [DOI] [PubMed] [Google Scholar]
- 16.de Wit-Verheggen VHW, van de Weijer T. Changes in cardiac metabolism in prediabetes. Biomolecules. 2021. 10.3390/biom11111680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Xu D, Jiang Z, Sun Z, Wang L, Zhao G, Hassan HM, et al. Mitochondrial dysfunction and inhibition of myoblast differentiation in mice with high-fat-diet-induced pre-diabetes. J Cell Physiol. 2019;234(5):7510–23. [DOI] [PubMed] [Google Scholar]
- 18.Greten FR, Grivennikov SI. Inflammation and cancer: triggers, mechanisms, and consequences. Immunity. 2019;51(1):27–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Alfadul H, Sabico S, Alnaami AM, Amer OE, Hussain SD, Wani K, et al. Acute glycemic control in prediabetes individuals favorably alters serum NLRP3 inflammasome and related interleukins. Int J Mol Sci. 2023. 10.3390/ijms241813837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Bäck M, Yurdagul A Jr., Tabas I, Öörni K, Kovanen PT. Inflammation and its resolution in atherosclerosis: mediators and therapeutic opportunities. Nat Rev Cardiol. 2019;16(7):389–406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sharma BR, Kanneganti TD. NLRP3 inflammasome in cancer and metabolic diseases. Nat Immunol. 2021;22(5):550–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zou Y, Lu S, Li D, Huang X, Wang C, Xie G, et al. Exposure of cumulative atherogenic index of plasma and the development of prediabetes in middle-aged and elderly individuals: evidence from the CHARLS cohort study. Cardiovasc Diabetol. 2024;23(1):355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhang Z, Zhao L, Lu Y, Xiao Y, Zhou X. Insulin resistance assessed by estimated glucose disposal rate and risk of incident cardiovascular diseases among individuals without diabetes: findings from a nationwide, population based, prospective cohort study. Cardiovasc Diabetol. 2024;23(1):194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Adrover JM, Del Fresno C, Crainiciuc G, Cuartero MI, Casanova-Acebes M, Weiss LA, et al. A neutrophil timer coordinates immune defense and vascular protection. Immunity. 2019;50(2):390-402.e310. [DOI] [PubMed] [Google Scholar]
- 25.Haybar H, Bandar B, Torfi E, Mohebbi A, Saki N. Cytokines and their role in cardiovascular diseases. Cytokine. 2023;169:156261. [DOI] [PubMed] [Google Scholar]
- 26.Georgakis MK, Gill D, Rannikmäe K, Traylor M, Anderson CD, Lee JM, et al. Genetically determined levels of circulating cytokines and risk of stroke. Circulation. 2019;139(2):256–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Li Q, Park K, Li C, Rask-Madsen C, Mima A, Qi W, et al. Induction of vascular insulin resistance and endothelin-1 expression and acceleration of atherosclerosis by the overexpression of protein kinase C-β isoform in the endothelium. Circ Res. 2013;113(4):418–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Schönfeld P, Wojtczak L. Short- and medium-chain fatty acids in energy metabolism: the cellular perspective. J Lipid Res. 2016;57(6):943–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Catapano AL, Pirillo A, Norata GD. Vascular inflammation and low-density lipoproteins: is cholesterol the link? A lesson from the clinical trials. Br J Pharmacol. 2017;174(22):3973–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Dominic A, Le NT, Takahashi M. Loop between NLRP3 inflammasome and reactive oxygen species. Antioxid Redox Signal. 2022;36(10–12):784–96. [DOI] [PubMed] [Google Scholar]
- 31.Ormazabal V, Nair S, Elfeky O, Aguayo C, Salomon C, Zuñiga FA. Association between insulin resistance and the development of cardiovascular disease. Cardiovasc Diabetol. 2018;17(1):122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.De Felice FG, Ferreira ST. Inflammation, defective insulin signaling, and mitochondrial dysfunction as common molecular denominators connecting type 2 diabetes to Alzheimer disease. Diabetes. 2014;63(7):2262–72. [DOI] [PubMed] [Google Scholar]
- 33.Batista-Gonzalez A, Vidal R, Criollo A, Carreño LJ. New insights on the role of lipid metabolism in the metabolic reprogramming of macrophages. Front Immunol. 2019;10:2993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Orecchioni M, Ghosheh Y, Pramod AB, Ley K. Macrophage polarization: different gene signatures in M1(LPS+) vs. classically and M2(LPS-) vs. alternatively activated macrophages. Front Immunol. 2019;10:1084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wu HM, Ni XX, Xu QY, Wang Q, Li XY, Hua J. Regulation of lipid-induced macrophage polarization through modulating peroxisome proliferator-activated receptor-gamma activity affects hepatic lipid metabolism via a Toll-like receptor 4/NF-κB signaling pathway. J Gastroenterol Hepatol. 2020;35(11):1998–2008. [DOI] [PubMed] [Google Scholar]
- 36.Zhang Q, Xiao S, Jiao X, Shen Y. The triglyceride-glucose index is a predictor for cardiovascular and all-cause mortality in CVD patients with diabetes or pre-diabetes: evidence from NHANES 2001–2018. Cardiovasc Diabetol. 2023;22(1):279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Cheng Y, Fang Z, Zhang X, Wen Y, Lu J, He S, et al. Association between triglyceride glucose-body mass index and cardiovascular outcomes in patients undergoing percutaneous coronary intervention: a retrospective study. Cardiovasc Diabetol. 2023;22(1):75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala MG, Hernández-González SO, et al. The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp. J Clin Endocrinol Metab. 2010;95(7):3347–51. [DOI] [PubMed] [Google Scholar]
- 39.Ruan GT, Xie HL, Zhang HY, Liu CA, Ge YZ, Zhang Q, et al. A novel inflammation and insulin resistance related indicator to predict the survival of patients with cancer. Front Endocrinol (Lausanne). 2022;13:905266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Tang S, Wang H, Li K, Chen Y, Zheng Q, Meng J, et al. C-reactive protein-triglyceride glucose index predicts stroke incidence in a hypertensive population: a national cohort study. Diabetol Metab Syndr. 2024;16(1):277. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data that support the results of this study are available from the corresponding author upon request.