BACKGROUND: A combined liver and kidney transplantation (CLKT) is advocated for selected individuals with chronic kidney disease undergoing liver transplantation (LT). The aim was to develop a risk score to identify the patients whose estimated glomerular filtration rate (eGFR) would decline during the year post-LT to aid future patient selection for CLKT. METHODS: A training dataset of LT recipients was identified retrospectively from a prospectively compiled database (2000-2007). The eGFR was calculated at 1 year and those with an eGFR less than 30 mL/min were identified. Variables determined at the LT assessment were analyzed by logistic regression, discriminant function, and area under the receiver operating characteristic curve (AUC) analysis to develop the score. The score was validated in a prospective patient cohort. RESULTS: Three hundred sixty-eight LT recipients were followed up for 1 year (training dataset). The mean eGFR declined by 11.2+/-23.5 mL/min during that time (P<0.001). A pre-LT risk score to predict an eGFR less than 30 mL/min at 1-year post-LT was generated: -1.8+(1.5 if a history of hypertension)+(0.65 x proteinuria in g/24 hr)+(0.013 x serum creatinine in micromol/L)+(0.001 x duration of acute kidney injury or eGFR <60 mL/min in days). Reversible renal impairment should first be excluded. Progression was likely with a score more than 2.16. Sensitivity, specificity, and AUC were 99.2%, 100%, and 0.99, respectively. All, but one patient, in the validation cohort (n=149) were correctly classified. CONCLUSION: This information will complement previously published criteria for CLKT patient selection.
BACKGROUND: A combined liver and kidney transplantation (CLKT) is advocated for selected individuals with chronic kidney disease undergoing liver transplantation (LT). The aim was to develop a risk score to identify the patients whose estimated glomerular filtration rate (eGFR) would decline during the year post-LT to aid future patient selection for CLKT. METHODS: A training dataset of LT recipients was identified retrospectively from a prospectively compiled database (2000-2007). The eGFR was calculated at 1 year and those with an eGFR less than 30 mL/min were identified. Variables determined at the LT assessment were analyzed by logistic regression, discriminant function, and area under the receiver operating characteristic curve (AUC) analysis to develop the score. The score was validated in a prospective patient cohort. RESULTS: Three hundred sixty-eight LT recipients were followed up for 1 year (training dataset). The mean eGFR declined by 11.2+/-23.5 mL/min during that time (P<0.001). A pre-LT risk score to predict an eGFR less than 30 mL/min at 1-year post-LT was generated: -1.8+(1.5 if a history of hypertension)+(0.65 x proteinuria in g/24 hr)+(0.013 x serum creatinine in micromol/L)+(0.001 x duration of acute kidney injury or eGFR <60 mL/min in days). Reversible renal impairment should first be excluded. Progression was likely with a score more than 2.16. Sensitivity, specificity, and AUC were 99.2%, 100%, and 0.99, respectively. All, but one patient, in the validation cohort (n=149) were correctly classified. CONCLUSION: This information will complement previously published criteria for CLKT patient selection.