Assessment of plasma ceramides as predictor for subclinical atherosclerosis

Background and aims: Ceramides have been identified as novel biomarkers for cardiovascular disease (CVD) related events and mortality but their role in etiology of subclinical atherosclerosis is unknown. We aimed to assess association between plasma ceramides and carotid intima-media thickness (CIMT) and evaluate predictive value of the ceramides for high CIMT over traditional CVD risk factors. Methods: Association between plasma ceramides and CIMT in the Young Finns Study participants was analyzed with CIMT as outcome and ceramides along with traditional risk factors as predictors with regression model. Predictive value of the ceramides and related coronary event risk test (CERT) score for high CIMT as surrogate marker of subclinical atherosclerosis was assessed by comparing logistic regression-based prediction models including, i) traditional risk factors and ceramides, ii) traditional risk factors and CERT score, iii) age, sex and ceramides or alternatively CERT score with a reference model including only traditional risk factors. The prediction models were fitted to training data (70% data) and tested on test data (30% data). The predictive models were assessed with area under the receiver operating curve (AUC). The variance of AUC was estimated by repeating the model fitting and testing for 1000 bootstraps of the original data. Results: Predictive models with plasma ceramides or alternatively with CERT score in addition to age and sex variables were able to predict high CIMT with AUC 0.726 and 0.720 respectively. However, the ceramides and CERT score did not have statistically significant added predictive value for high CIMT over traditional risk factors. Conclusions: The new systemic biomarkers, high-risk plasma ceramides and CERT score, showed promising predictive performance for high CIMT with only age and sex as additional variables. This may help in predicting subclinical atherosclerosis for primary prevention. (c) 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).