Applicability of the Framingham risk score in predicting mortality among kidney failure patients: an international analysis

Background The usefulness of Framingham risk score (FRS) for estimation of mortality risk in kidney failure patients was explored in three cohorts of end-stage renal disease (ESRD) patients. Methods Associations of FRS with mortality risk were analyzed among 392 Swedish non-dialyzed (median age 56 years, males 62.8%, median FRS 19.5%) and 109 hemodialyzed (median age 66 years, males 57.8%, median FRS 17.4%) ESRD patients and 1276 Chinese peritoneal dialysis (PD) patients (median age 50 years, males 55.6%, FRS 12.3%). We used Fine-Gray models to estimate relative risk associated with FRS and its components, restricted mean survival time (RMST) to quantitate survival differences associated with FRS, and receiver-operator characteristics (ROC) curve analysis to estimate prognostic value of adding non-traditional risk factors to FRS. Results Despite different baseline characteristics, 1-SD increase of FRS associated with similar increases (sub-hazard ratios 1.39–1.63) of both all-cause and cardiovascular mortality risk after adjusting for BMI, serum albumin, hemoglobin and C-reactive protein. RMST and ∆RMST showed shorter survival time to all-cause and cardiovascular mortality in middle and high vs. low FRS tertile. High vs. low FRS tertile showed 2- to 4-fold higher cumulative incident risk of all-cause and cardiovascular mortality. Adding non-traditional risk factors such as serum albumin and C-reactive protein to FRS improved risk prediction marginally. Conclusions Higher FRS associated with similar increases of mortality risk in three distinctly different cohorts of ESRD patients whereas addition of non-traditional risk factors had negligible impact. Framingham risk score is a valid tool for estimation of mortality risk in ESRD patients.