Back to search

The University of Edinburgh

Metabolomic profiles of cardiovascular diseases in type 2 diabetes and in the general population

Abstract

dc:description.abstract

INTRODUCTION: Cardiovascular disease (CVD) is the leading cause of death globally, especially in people with type 2 Diabetes (T2D). Metabolomic data has the potential to both shed light on the pathobiological underpinning of CVD and also facilitate the identification of potential novel predictors for better personalizing risk assessment. Metabolomic profiles of clinical CVD have been assessed previously, but little is known about metabolites associated with subclinical atherosclerosis in either general or T2D populations. In addition, existing cardiovascular risk scores perform inadequately in T2D populations, and any incremental value of adding metabolites to traditional risk factors for predicting CVD in people with T2D is currently uncertain. AIMS: To determine metabolomic profiles of subclinical and clinical CVD in both T2D and general populations. Furthermore, to investigate whether combining a small panel of metabolites with a current cardiovascular risk score (QRISK3) could facilitate better risk prediction for CVD in people with T2D. METHODS: Analyses were undertaken on existing datasets from the Edinburgh Type 2 Diabetes Study (ET2DS, a longitudinal cohort of 1,066 men and women with T2D aged 60-75 years at baseline) and the Orkney Complex Disease and VIKING Health – Shetland Studies (ORCADES, N=2,078; VIKING, N=2,105) for investigations in T2D and in the general population respectively. Blood metabolomic data were measured by nuclear magnetic resonance spectroscopy in the ET2DS (serum) and two general populations (plasma) from baseline. In ET2DS, markers, including ankle-brachial index (ABI) and carotid intima-media thickness (cIMT), and clinical CVD were available from baseline and during 10 years of follow-up. These cardiovascular phenotypes were available only at baseline in ORCADES and VIKING. Associations between metabolites and cardiovascular phenotypes were assessed using multivariable regression models and the least absolute shrinkage and selection operator (LASSO). Bonferroni correction was used to take into account multiple comparison issues. For the prediction of 10-year CVD, 100 repeats of Cox LASSO were used to select metabolites as components for a metabolites-based risk score (MRS). The predictive performance of the MRS was assessed in relation to a reference score which consisted of components of QRISK3 plus prevalent CVD and statin use at baseline. RESULTS: In ET2DS, increased levels of serum lactate, glycerol, creatinine, and glycoprotein acetyls were associated with decreased baseline ABI in both multivariate regressions [βs (95% Confidence interval, CI) ranged from -0.025 (-0.036, -0.015) to -0.023 (-0.034, -0.013), all p<0.0002] and LASSO analysis, after adjustment for age and sex. Following adjustment for other traditional cardiovascular risk factors, the associations remained nominally significant. Lactate showed nominal association with ABI measured at Year 4 and Year 10 after adjusting for baseline ABI. All four ABI-associated metabolites showed positive associations with prevalent CVD [odds ratios (ORs) (95%CI) ranged from 1.29 (1.13, 1.47) to 1.49 (1.29, 1.74), all p<0.0002] and incident CVD [ORs (95%CI) ranged from 1.19 (1.02, 1.39) to 1.35 (1.17, 1.56), all p<0.05]. In contrast, no metabolite was significantly associated with cIMT. In ORCADES and VIKING, no statistically significant association between plasma metabolites and atherosclerotic markers was found. However, for all of the four metabolites associated with ABI in the ET2DS which were also measured in ORCADES and VIKING (all except glycerol), the trend was for associations with ABI to be in the same direction as for the ET2DS (i.e., higher metabolite levels with lower ABI/greater levels of atherosclerosis). Eighty-six metabolites, mainly lipid ratios and fatty acids, were significantly associated with self-report prevalent CVD in ORCADES [ORs (95%CI) ranged from 0.57 (0.48, 0.67) to 1.60 (1.35, 1.88), all p<0.0002]. In the risk prediction analysis on the ET2DS dataset, twelve metabolites relating to amino acids, fatty acids, fluid balance, glycolysis, and lipoproteins were chosen as components of the MRS which showed positive associations with incident CVD over 10 years after adjustment for traditional cardiovascular risk factors [hazard ratio 2.67 (95%CI 1.96, 3.64)]. The c-statistic for the reference model alone was 0.709 (95%CI 0.679, 0.739) which increased modestly to 0.728 (95%CI 0.700, 0.757) for the reference model combining the MRS. Compared with the reference model, the net reclassification index and integrated discrimination improvement for the combination model were 0.362 (95%CI 0.179, 0.506) and 0.041 (95%CI 0.020, 0.071) respectively. CONCLUSIONS: In older people with T2D, serum metabolites relating to glycolysis, fluid balance and inflammation showed independent associations with both a subclinical atherosclerotic marker and with symptomatic CVD. Some lipid-related components and fatty acids in plasma were significantly associated with prevalent CVD in the general population, but associations with markers of subclinical atherosclerosis were not statistically significant. A small panel of metabolites might improve predictive performance for cardiovascular risk over-and-above traditional risk factors in people with T2D, but further investigation, including external validation, is required before firm recommendations can be made.

Degree

thesis:*
Grantor dc:publisher
The University of Edinburgh
Year dc:date.issued
2023

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Huang, Zhe
Advisors dc:contributor.advisor
  • Price, Jackie
  • Klaric, Lucija
  • Wilson, James

Subjects

dc:subject × 6

Rights

Language dc:language.iso
en

Identifiers

dc:identifier.*

Chain of custody

source
Harvested from
University of Edinburgh
Base URL
era.ed.ac.uk/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Huang, Zhe. Metabolomic profiles of cardiovascular diseases in type 2 diabetes and in the general population. The University of Edinburgh, 2023. https://hdl.handle.net/1842/40962