Statistical analyses

Categorical variables were compared with Chi-square test, whilst differences between continuous variables were assessed by T-Student’s/Mann–Whitney or ANOVA/Kruskal–Wallis, depending on the normality of the data and the numbers of groups compared. The effect of SNPs was evaluated by regression analyses adjusting by confounding variables that were selected based on univariate tests or by clinical criteria. Significant covariates used included demographics (age, sex, body-mass index) and clinical parameters, namely hyperlipidaemia, hypertension, and CKD stage. A dominant model of inheritance was chosen for the genetic analyses, as the resulting comparison groups were the most balanced in terms of size, as conducted in previous studies of our group with CKD patients [37,38]. The CV impact of the SNPs was analysed by Kaplan–Meier curves, comparing the different genotypes with the log-rank test. SNPs with p-values lower than 0.1 were included in Cox regression models to establish their effect after controlling for traditional CV risk factors. Patients were followed up until the earliest of CV event, death or end of study.

Gene-gene interaction (epistasis) analyses were performed by using log-likelihood ratio tests adjusted by relevant covariates in a dominant model. In the resulting plots, the diagonal line contains the P values from likelihood ratio test for the crude effect of each SNP, which are sorted by their genomic position. The upper triangle in the matrix contains the P values for the interaction (epistasis) log-likelihood ratio test. The lower triangle contains the P values from likelihood ratio test comparing the two-SNP additive likelihood to the best of the single-SNP models. The network visualisation of the gene-gene interactions was generated by Cytoscape software package v. 3.9.1.

Statistical analyses were performed with the SNPassoc (2.0-11) in the R environment v. 4.1.3 and with IBM SPSS software (SPSS Inc., Chicago, IL, USA, v. 22.0).