Extending the Lund-Malmö creatinine-based GFR equation to cystatin C–validation results from the European Kidney Function Consortium (EKFC) cohort of children and adults

Abstract

The aim of the present study was to extend the creatinine-based Lund-Malmö GFR equation for use with rescaled cystatin C (r-LMR_Cys) and validate it against measured GFR (mGFR) in the EKFC cystatin C cohort of children (n = 2,293) and adults (n = 7,727). Rescaling was obtained by dividing each biomarker by a Q-value, representing the population-specific median biomarker level among healthy individuals. Validation included median bias/precision/accuracy (percent estimates within ±30% of mGFR, P₃₀). Performance was compared with the EKFC-equation (EKFC_Cys), the CAPA cystatin C equation, the corresponding equations based on rescaled creatinine (r-LMR_Cr and EKFC_Cr) and the arithmetic mean of r-LMR_Cr and CAPA (r-LMR_Cr+CAPA), r-LMR_Cr and r-LMR_Cys (r-LMR_Mean), and EKFC_Cr and EKFC_Cys (EKFC_Mean). The overall P30 of r-LMR_Cys in adults was 86.2% (95% CI 85.4%-86.9%), which was 6.6 percentage points (pp; 95% CI 5.8–7.4 pp) higher than for CAPA and similar to r-LMR_Cr (P₃₀ 87.4%, 95% CI 86.6%−88.1%). r-LMR_Cys and EKFC_Cys exhibited similar performance both overall and across subgroups of age, sex, GFR and BMI and in children. All three arithmetic mean equations had similar P₃₀-accuracy and generally performed better than the corresponding single-marker equations. Our results show that the Lund-Malmö GFR equation can be adapted for use with rescaled cystatin C with performance that is similar to the best-performing equations based on rescaled creatinine. The generality of the applied biomarker rescaling principle implies that the future demand for population- and biomarker-specific GFR estimating equations can be expected to decrease substantially.