Cystatin C is not superior to creatinine-based models in estimating glomerular filtration rate in former kidney donors

BACKGROUND: The long-term renal consequences of kidney donation need to be accurately quantitated. Cystatin C is a freely filtered glycoprotein that may not have the limitations of creatinine as a measure of glomerular filtration rate (GFR). Whether cystatin C is superior to creatinine-based estimates of GFR in those who have donated a kidney in the past has not been tested. METHODS: We assessed the performance of seven cystatin C and two creatinine-based GFR prediction equations in 187 former kidney donors against iohexol GFR for measuring GFR. We calculated bias, precision, and relative accuracy of these models. RESULTS: The majority of former donors had a GFR >60 mL/min/1.73 m(2). All cystatin C models, except the Rule model, overestimated GFR (range 5.3-31.4 mL/min/1.73 m(2)). Among the cystatin C models, the Hoek and Rule formulas were least biased at 5.3 and -3.8 mL/min/1.73 m(2), most precise at 0.41, and were within 30% of iohexol GFR, 89.3 and 96% of the time, respectively. The Modification of Diet in Renal Disease (MDRD) formula underestimated GFR by 7.2 mL/min/1.73 m(2), was most precise (R(2)=0.47) and fell within 30% of measured GFR at the highest frequency of 96%. When all models were given a rank based on their performance in the bias, precision and accuracy domains, the MDRD model was clearly superior. CONCLUSION: The MDRD equation is superior to cystatin C-based equations for estimating GFR in former kidney donors. Creatinine measurement is cheaper and the MDRD GFR is given out by most laboratories and therefore it should be the preferred model in this population.