Performance of Creatinine and Cystatin-Based Equations on Estimating Measured Glomerular Filtration Rate in People with Hematological and Solid Cancers.

Abstract

BACKGROUND Glomerular filtration rate (GFR) assessment is important in clinical practice, with implications for diagnosis, prognostication, and drug dosing. People with cancer are at risk of imprecision in GFR estimation. This cross-sectional study evaluated the performance of various creatinine and cystatin C-based equations in comparison to measured GFR (mGFR) in people with cancer. METHODS We retrieved data for all adult patients who had mGFR by urinary iothalamate clearance between 2011 and 2023 at Mayo Clinic and use of an electronic health record diagnosis code for cancer within two years prior to mGFR. The Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), European Kidney Function Consortium (EKFC), and Cockcroft-Gault (CG) equations were computed, along with performance metrics (bias, precision, and root mean square error, RMSE). Confidence intervals were generated by bootstrapping and analysis were stratified by solid and hematological cancer. RESULTS From all adults with cancer and mGFR, 1145 had both creatinine and cystatin C available within seven days of mGFR. Among all equations, the creatinine- cystatin C CKDEPI equation provided the best performance, with small bias (median 3.0, 95%CI 2.3-3.8) and higher precision (RMSE 14.5) compared to creatinine-only or cystatin-only equations (RMSE varying from 16.6 to 20), and this was also true in solid and hematological cancers. The creatinine-cystatin EKFC equation had a similar performance to CKDEPI, CG showed worst precision (30% of people with errors above 30%), and cystatin C CKDEPI equation was the most biased, prone to underestimation of mGFR. CONCLUSIONS In our cohort of patients with mGFR and cancer, the CKDEPI creatinine-cystatin C equation performed best for GFR assessment, and this was true for both solid and hematological cancers. Our findings give support for the preferential use of creatinine and cystatin C-based equations instead of creatinine-only or cystatin C-only equations in people with cancer.