Skin autofluorescence, arterial stiffness and Framingham risk score as predictors of clinical outcome in chronic kidney disease patients: a cohort study

BACKGROUND: The risk of cardiovascular disease (CVD) is predicted by Framingham's CVD risk scores (FRS) but the high CVD-related mortality in patients with chronic kidney disease (CKD) is only partially explained by traditional CVD risk markers. Therefore, there is a need to explore whether other CVD risk markers may improve risk prediction. Although arterial stiffness measured by augmentation index (AIx) and tissue content of advanced glycation end-products (AGEs) measured by skin autofluorescence (SAF) are two biomarkers that associate with CVD and mortality in CKD, it is not known how they compare with FRS. We evaluated associations between SAF, AIx and FRS, and their associations with CVD and mortality in CKD patients. METHODS: SAF (AGE Reader) and AIx (SphygmoCor; adjusted for 75 heart beats per minute) were measured in 261 clinically stable and extensively phenotyped patients with CKD Stage 5 (median age 56 years, 66% male, 20% diabetes; 130 non-dialysed, 93 patients on peritoneal dialysis and 38 patients on haemodialysis). Multivariate receiver operator characteristics (ROC) curve analysis and multivariate Cox models followed by C-statistics were used to evaluate CVD-related and all-cause mortality risk associated with SAF, AIx and FRS during follow-up for median 25 months with 46 deaths. RESULTS: In multivariate regression analysis, SAF associated with FRS, haemoglobin, fat body mass index and CVD, and inversely with per cent handgrip strength (HGS). AIx associated with FRS, and inversely with per cent HGS. Associations of SAF and AIx with high-sensitivity C-reactive protein (hsCRP), serum albumin, statin therapy and renal replacement therapy were not statistically significant. In ROC analysis, area under the curve (AUC) for CVD mortality ranged from AUC = 0.72 (AIx and FRS, respectively) to AUC = 0.78 (FRS + AIx), and for all-cause mortality from AUC = 0.70 (AIx) to AUC = 0.79 (FRS + AIx). In multivariate Cox analysis, after adjusting for 1-standard deviation (1-SD) of FRS, 1-SD increase of SAF associated with all-cause mortality and 1-SD increase of AIx associated with CVD mortality and all-cause mortality. After further adjustments for hsCRP, albumin and presence of CVD, AIx (but not SAF) remained independently associated with CVD mortality, hazard ratio (HR) 2.14 [95% confidence interval (95% CI) 1.18-3.89] and all-cause mortality, HR 1.74 (95% CI 1.16-2.60). CONCLUSIONS: In patients with CKD Stage 5, SAF and aortic stiffness associated with mortality, independently of FRS. After adjusting for additional confounders including inflammation, aortic stiffness remained as an independent predictor of outcome. Since the contribution of SAF and aortic stiffness compared with FRS in ROC curve analysis was relatively modest, this underlines the importance of traditional CVD risk factors in CKD.