BACKGROUND: The prevalence of chronic kidney disease is high in the elderly, but the effects of a decrease in the eGFR on mortality and functioning are still unclear. The aim of this study was to determine whether the combination of the eGFR and the eGFR slope is a predictor of mortality and functional decline. METHODS: The eGFR (MDRD equation) and the eGFR slope were calculated. The slope was calculated using four annual eGFR measurements taken from 85 to 88 years of age. Mortality and changes in the Mini-Mental State Examination (MMSE), the Geriatric Depression Scale (GDS-15) and the Activities of Daily Living (ADL) scores were analysed as outcomes from 88 years onwards. RESULTS: 378 persons aged 88 years participating to the Leiden 85 plus study, an observational prospective cohort study in the general population, were included. A combined analysis of the baseline eGFR and the eGFR slope showed that an eGFR of >60 ml/min combined with an eGFR decrease of >= 3 ml/min/year and an eGFR of <60 ml/min combined with an of eGFR decrease >=5 ml/min/year were independent predictors of increased mortality. The baseline eGFR, the eGFR slope and a combination of both factors did not predict changes in the MMSE, GDS or ADL scores between 88 and 90 years. CONCLUSION: The combination of the eGFR and the eGFR decrease allows the identification of subgroups of very elderly with increased mortality risks and of subgroups of very elderly with an eGFR of <60 ml/min without an increased risk of mortality.