Serum metabolomic profiles associated with subclinical and clinical cardiovascular phenotypes in people with type 2 diabetes

Abstract

Background Atherosclerotic cardiovascular diseases (CVD) is the leading cause of death in diabetes, but the full range of biomarkers reflecting atherosclerotic burden and CVD risk in people with diabetes is unknown. Metabolomics may help identify novel biomarkers potentially involved in development of atherosclerosis. We investigated the serum metabolomic profile of subclinical atherosclerosis, measured using ankle brachial index (ABI), in people with type 2 diabetes, compared with the profile for symptomatic CVD in the same population. Methods The Edinburgh Type 2 Diabetes Study is a cohort of 1,066 individuals with type 2 diabetes. ABI was measured at baseline, years 4 and 10, with cardiovascular events assessed at baseline and during 10 years of follow-up. A panel of 228 metabolites was measured at baseline using nuclear magnetic resonance spectrometry, and their association with both ABI and prevalent CVD was explored using univariate regression models and least absolute shrinkage and selection operator (LASSO). Metabolites associated with baseline ABI were further explored for association with follow-up ABI and incident CVD. Results Mean (standard deviation, SD) ABI at baseline was 0.97 (0.18, N = 1025), and prevalence of CVD was 35.0%. During 10-year follow-up, mean (SD) change in ABI was + 0.006 (0.178, n = 436), and 257 CVD events occurred. Lactate, glycerol, creatinine and glycoprotein acetyls levels were associated with baseline ABI in both univariate regression [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. The associations remained nominally significant after adjustment for major vascular risk factors. In prospective analyses, lactate was nominally associated with ABI measured at years 4 and 10 after adjustment for baseline ABI. The four ABI-associated metabolites were all positively associated with prevalent CVD [odds ratios (ORs) ranged from 1.29 (1.13, 1.47) to 1.49 (1.29, 1.74), all p < 0.0002], and they were also positively associated with incident CVD [ORs (95% CI) ranged from 1.19 (1.02, 1.39) to 1.35 (1.17, 1.56), all p < 0.05]. Conclusions Serum metabolites relating to glycolysis, fluid balance and inflammation were independently associated with both a marker of subclinical atherosclerosis and with symptomatic CVD in people with type 2 diabetes. Additional investigation is warranted to determine their roles as possible etiological and/or predictive biomarkers for atherosclerotic CVD.

Chewing the fat: How lipidomics is changing our understanding of human health and disease in 2022

“…Due to the elevated risk of mortality with these diseases, there is a need for strategies in its identification, prevention, and treatment within healthcare. In 2022, as evidenced in Figure 5, the bulk of research articles focussed on related diseases: CVD, 232–255 diabetes, 166,233,249,250,256–278 liver disease, 51,279–293 MetS 277,294,295 and obesity 277,295–305 …”

“…Bibliometric analyses of clinical lipidomics publications in 2022 using Scopus search terms “lipidomics” AND “clinic*” on 15 th December 2022 …”

“…Bibliometric analyses of clinical lipidomics publications in 2022 according to separation techniques, detectors, and RPLC columns used …”

Metabolomics data improve 10-year cardiovascular risk prediction with the SCORE2 algorithm for the general population without cardiovascular disease or diabetes

“…The threshold was adopted from a previous study to enhance model generalization and mitigate overfitting. 24 The selected metabolites were subsequently incorporated into the SCORE2 model to construct new sex-specific competing risk algorithms using the r-package cmprsk (version 2.2-11).…”

Improving 10-year cardiovascular risk prediction in patients with type 2 diabetes with metabolomics

“…Metabolites selected by LASSO in at least 95% of instances were designated as our metabolites of interest, a threshold adopted from a previous study (23) to enhance model generalization and mitigate overfitting. These selected metabolites were subsequently incorporated into the SCORE2-Diabetes model to construct new sex-specific competing risk models using the r-package cmprsk (version 2.2-11).…”