Triglyceride–glucose index and atherogenic index of plasma as predictors of cardiovascular risk in rectal cancer survivors

Fangmin Shi, Chengnan Liu, Wei Liu, Hao Cai, Qiming Zhao, Chen Zhang, Qiyuan Liu, Nan Wu, You GuoFrontiers in Cardiovascular Medicine 2026 · DOI

TL;DR

Rectal cancer survivors face a significantly higher risk of cardiovascular disease (CVD), and this study reveals that two simple metabolic markers, the triglyceride-glucose index (TyG) and the atherogenic index of plasma (AIP), are independent predictors of this secondary CVD risk. These easily obtainable indices, especially when combined and considered alongside age, can aid in stratifying risk for these patients, with potential mechanistic insights provided by proteomic analysis.

What Problem Does This Paper Address?

Rectal cancer survivors have a substantially increased risk of developing cardiovascular disease (CVD) compared to the general population. This elevated risk is thought to stem from various factors, including the vascular damage from radiotherapy, metabolic toxicity of chemotherapy, and the underlying metabolic dysregulation associated with cancer itself, such as insulin resistance. While traditional risk factors for CVD are well-established, there's a need for simpler, readily available markers to help stratify risk specifically within this vulnerable patient population. The study aims to investigate whether two composite metabolic indices, the triglyceride-glucose index (TyG) and the atherogenic index of plasma (AIP), which are known indicators of insulin resistance and dyslipidemia respectively, can serve as effective predictors of secondary CVD in rectal cancer survivors.

Methods

The researchers conducted a retrospective cohort study using data from rectal cancer patients treated at the First Affiliated Hospital of Gannan Medical University. They analyzed clinical and laboratory data, including lipid profiles and glucose levels, to calculate TyG and AIP. To assess the association with secondary CVD, they employed univariate and multivariate logistic regression analyses. Interaction analyses were performed to explore how age and sex modify these associations. The predictive value of these indices was evaluated using receiver operating characteristic (ROC) curves and by developing nested prediction models. Additionally, a mediation analysis was conducted using proteomics data from the UK Biobank to identify potential biological pathways linking TyG and AIP to CVD.

Datasets

  1. Rectal Cancer Patient Cohort: Data from 1,431 rectal cancer patients treated at the First Affiliated Hospital of Gannan Medical University between January 2010 and December 2023. This cohort was used for the primary association and predictive value analyses.
  2. UK Biobank (UKB) Proteomics Data: Data from 53,026 participants in the UK Biobank, including proteomics data from 1,081 rectal cancer patients. This dataset was used for the mediation analysis to identify potential underlying biological mechanisms.

Key Insights and Ideas

  • Simple Markers, Significant Risk: The study highlights that even basic metabolic markers like TyG and AIP, which are easily measured in routine blood tests, hold significant predictive power for CVD risk in rectal cancer survivors. This is crucial for clinical practice where quick and accessible risk assessment is vital.
  • Age as a Critical Modifier: The association between TyG/AIP and CVD risk is not uniform across all ages. The study found that age significantly modifies this relationship, with stronger effects observed in older patients (≥62 years). This suggests that age-specific risk stratification is essential.
  • Synergistic Prediction: Combining TyG and AIP offered a modest but statistically significant improvement in predicting CVD risk compared to using clinical variables alone. This points towards the value of a composite risk assessment using these metabolic indices.
  • Unveiling Mechanistic Pathways: The proteomic analysis provided a glimpse into the biological underpinnings of these associations, identifying proteins involved in lipid metabolism, complement/coagulation, and tight junction pathways. This opens doors for understanding how these metabolic derangements lead to cardiovascular issues in this population.
  • Cancer-Specific CVD: The matched case-control analysis suggested that cardiovascular disease in rectal cancer survivors might have distinct metabolic characteristics compared to primary CVD in non-cancer patients, underscoring the need for tailored risk models.

Key Results

  • Independent Predictors: Both TyG (OR 1.102, p=0.005) and AIP (OR 1.062, p=0.018) were independently associated with secondary CVD risk in rectal cancer patients, even after multivariable adjustment.
  • Age Interaction: Age significantly modified the association between TyG/AIP and CVD risk (interaction p < 0.05), with stronger effects in patients aged 62 years and older.
  • Predictive Improvement: The combined TyG + AIP index modestly improved prediction of CVD risk, showing an Integrated Discrimination Improvement (IDI) of 1.2% (p=0.013) when added to clinical variables.
  • Proteomic Findings: Proteomic analysis identified 14 mediating proteins, with enrichment in pathways related to lipid metabolism, complement/coagulation, and tight junctions.

Limitations and Future Work

The authors acknowledge several limitations:

  • Retrospective Design: The study's retrospective nature prevents definitive causal inference.
  • Single Time-Point Measurements: The analysis relied on single measurements, limiting the ability to assess dynamic changes in biomarkers during cancer treatment and their evolving impact on CVD risk.
  • Treatment Regimen Variability: Individual treatment regimens (chemotherapy, radiation) could not be fully accounted for, and these may differentially impact metabolic parameters and cardiovascular risk.
  • Generalizability: The findings from a predominantly European ancestry cohort (UK Biobank) and a single tertiary care center may not be universally generalizable to other ethnic populations or healthcare settings.
  • External Validation: The predictive models developed have not been externally validated.
  • Complex Biological Interactions: The complex regulatory relationships among identified mediator proteins warrant further investigation.

Future work suggested includes expanding sample sizes, conducting multi-center and multi-ethnic studies, external validation of predictive models, and further exploration of the identified mediator proteins and their mechanistic roles in cardiovascular disease development in rectal cancer survivors.

Summary updated

Abstract

Background Rectal cancer survivors face elevated cardiovascular disease (CVD) risk, but simple metabolic indices for risk stratification remain underexplored. We investigated associations of triglyceride-glucose index (TyG) and atherogenic index of plasma (AIP) with secondary CVD in rectal cancer patients. Methods We analyzed rectal cancer patient data from the First Affiliated Hospital of Gannan Medical University (January 2010 to December 2023) to evaluate the associations between TyG, AIP and CVD. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed using UK Biobank proteomics data to identify mediating proteins. Results This study included 1,431 cases, among which 1,240 had no CVD and 191 had CVD. Both TyG (OR 1.102, 95% CI: 1.030–1.179, p = 0.005) and AIP (OR 1.062, 95% CI: 1.010–1.116, p = 0.018) showed independent associations with CVD risk in multivariable models. Age significantly modified these associations (interaction p &amp;lt; 0.05), with stronger effects in patients ≥62 years. Combined TyG + AIP modestly improved prediction (IDI: 1.2%, p = 0.013) compared to clinical variables alone. Proteomic analysis identified 14 mediating proteins enriched in lipid metabolism, complement/coagulation, and tight junction pathways. Conclusions TyG and AIP indices were independently associated with secondary CVD risk in rectal cancer patients, with age-dependent effects. While the predictive improvement was modest, these easily obtainable indices may aid risk stratification when combined with traditional factors. Proteomic findings suggest potential mechanistic pathways warranting further investigation.
BibTeX
@inproceedings{shi2026triglycerideglucose,
  author = {Fangmin Shi and Chengnan Liu and Wei Liu and Hao Cai and Qiming Zhao and Chen Zhang and Qiyuan Liu and Nan Wu and You Guo},
  title = {{Triglyceride–glucose index and atherogenic index of plasma as predictors of cardiovascular risk in rectal cancer survivors}},
  booktitle = {Frontiers in Cardiovascular Medicine},
  year = {2026},
  doi = {10.3389/fcvm.2026.1707899},
}
Figure