#4506 ACCURACY OF NOVEL GFR ESTIMATING EQUATIONS BASED ON CREATININE, CYSTATIN C OR BOTH IN ROUTINE CARE

Abstract Background and Aims New equations to estimate GFR (eGFR) based on creatinine, cystatin C or both have been developed in the last two years. A comprehensive comparison of their accuracy is currently lacking, particularly in cohorts not involved in their development or validation and among people with comorbid conditions. Method We included 6174 adults from the Stockholm Creatinine Measurements (SCREAM) project referred for plasma clearance of iohexol during 2011-2021, in whom we observed 9579 concurrent measurements of creatinine, cystatin C and iohexol clearance. We assessed the performance against measured GFR (mGFR) of eGFR equations proposed by the CKD-EPI collaboration (CKD-EPI 2009, 2012 and 2021), European Kidney Function Consortium (EKFC 2021 and 2023), and the revised Lund-Malmö (2011) and CAPA (2014) equations, which are used in Sweden. Bias was expressed as the median difference in eGFR minus mGFR, with negative biases indicating underestimation of mGFR. P30 described the percentage of individuals with eGFR within 30% of mGFR. Correct classification was defined as agreement of eGFR and mGFR categories using the KDIGO GFR categories. Subgroup analyses were conducted according to age, sex, BMI, eGFR, cancer, cardiovascular disease, diabetes, heart failure and liver disease. Results Mean age was 57 years, 46% of participants were female, mean mGFR was 62 mL/min/1.73 m2 and mean BMI was 26 kg/m2. Cardiovascular disease was the most common comorbid condition (30%), followed by liver disease (28%), diabetes (26%) and cancer (26%). Equations that used both creatinine and cystatin C had better performance than eGFR using each marker alone, regardless of the equation used; all such equations had small bias and P30 close to 90%. Among creatinine-based equations, CKD-EPI 2009 and CKD-EPI 2021 showed larger overestimates of mGFR than EKFC 2021 and revised Lund-Malmö, with median biases of 5.6, 9.1, 2.7 and 0.2 mL/min/1.73 m2, respectively (Table 1). There were no meaningful differences in performance across eGFR equations based on cystatin C. Findings were consistent across subgroup analyses stratifying for comorbid conditions (Figure 1). Conclusion eGFR equations that combined information on creatinine and cystatin C performed better than equations based on creatinine or cystatin C alone in this Swedish cohort of routine referrals for plasma clearance of iohexol. There was larger variation in the performance of equations based on creatinine than cystatin C.