RT Journal Article SR Electronic T1 End-stage renal disease versus death in a Portuguese cohort of elderly patients: an approach using competing event analysis JF Journal of Investigative Medicine JO J Investig Med FD BMJ Publishing Group Ltd SP 1041 OP 1048 DO 10.1136/jim-2017-000480 VO 65 IS 7 A1 Josefina Santos A1 Isabel Fonseca A1 Jorge Malheiro A1 Idalina Beirao A1 Luisa Lobato A1 Pedro Oliveira A1 Antonio Cabrita YR 2017 UL http://hw-f5-jim.highwire.org/content/65/7/1041.abstract AB Chronic kidney disease (CKD) is higher in elderly, but mortality outweighs the risk of end-stage renal disease (ESRD). Our aim was to identify prognostic markers for ESRD or death in elderly CKD, within a competing-risk analysis. This is a longitudinal study of consecutive newly referred patients with CKD ages 65 years, followed until the time of the first event (ESRD or death), using a competing-risk analysis. A modified Charlson Comorbidity Index (mCCI) was subdivided into subgroups (0-2, 3-4, ≥5). Patients were followed for hospitalizations that occurred prior to the outcomes. Among 416 patients, age 76±8 years, 52% male, median estimated glomerular filtration rate of 32 mL/min per 1.73 m2, 50% had diabetes, and 67% cardiovascular disease. Over a median follow-up of 3.6 years, 36 patients progressed to ESRD (8.7%) and 103 died (24.8%). Older age (subdistribution HR (sHR)=1. 06; p<0.001), creatinine≥1.6 mg/dL (sHR=2.03, p=0.004), hemoglobin <11 g/dL (sHR=1.91, p=0.003), mCCI score≥5 (sHR=3.01, p<0.001) and having one or more hospitalizations (sHR=1.73, p<0.001) were associated with death before ESRD. The independent predictors for ESRD with competing risk of death were: lower age (sHR=0.94; p=0.009), creatinine≥1.6 mg/dL (sHR=3.26, p=0.006), hemoglobin <11 g/dL (sHR=2.15, p=0.027), peripheral vascular disease (sHR=3.45, p=0.001) and having one or more hospitalizations (sHR=1.56, p=0.031). Elderly referred patients with CKD are near threefold more likely to die than progress to ESRD. A competing-risk framework based on available clinical and laboratory data may discriminate between those outcomes and could be used as a decision-making tool.