Objective This study sought to characterise the main dyslipidaemic phenotypes present in chronic kidney disease (CKD) and their association with coronary heart disease (CHD) risk.

Methods Analyses included 6612 individuals in the multiethnic study of atherosclerosis free of CHD at baseline. CKD was defined as an estimated glomerular filtration rate (eGFR) of 15 to <60 mL/min/1.73 m² (stages 3–4). Principal component analyses were used to characterise the main dyslipidaemic phenotypes of CKD accounting for the correlation among different lipoproteins and lipoprotein particles. CHD was defined as incident myocardial infarction, angina followed by revascularisation, resuscitated cardiac arrest or CHD death.

Results CHD developed in 303 individuals (5%) with eGFR ≥60 and in 72 individuals (12%) with CKD (p for difference <0.001). A dyslipidaemic phenotype (principal component 1 (PC1)) consisting of elevations in triglycerides, triglyceride-rich lipoproteins (VLDL particles), small LDL particles and reductions in HDL particles, was more common in those with CKD, compared with those without CKD (p for difference <0.001). This phenotype was also more strongly associated with CHD in those with CKD: adjusted HRs (95% CIs) per SD increase in PC1 1.13 (95% CI 1.00 to 1.27; P=0.05) and 1.51 (95% CI 1.17 to 1.94; P<0.001) in eGFR ≥60 and CKD, respectively (P for interaction=0.05).

Conclusion In individuals with mainly stage 3 CKD, a dominant lipid phenotype consisting of triglyceride-rich lipoproteins and other closely correlated lipoproteins is strongly associated with CHD risk. Future studies should investigate whether modification of the components of this phenotype leads to a reduction in the CHD burden in individuals with CKD.