Asia-Pacific journal of clinical oncology

Comparative performances of the new chronic kidney disease epidemiology equations incorporating cystatin C for use in cancer patients.

PMID 25471594


In cancer patients receiving chemotherapy treatment, accurate assessment of kidney function is required. The aim of our study was to investigate whether the inclusion of cystatin C together with creatinine in prediction equations would improve the prediction of glomerular filtration rate (GFR). Plasma creatinine and cystatin C were analyzed in 155 patients (cancer, n = 80, kidney donors, n = 75) undergoing (99m) Technetium diethylenepentaacetic (Tc-DTPA) GFR clearance. Equations by the CKD-EPI (chronic kidney disease epidemiology) group (creatinine-, creatinine + cystatin C-, cystatin C-based, respectively) and Cockcroft-Gault were compared with Tc-DTPA GFR by difference plots, receiver operator characteristics curve analysis, root mean square error, chi-squared analysis and percentage concordance according to carboplatin dosage. Comparisons between two creatinine methodologies (enzymatic vs Jaffe) were also performed. In the overall group, the combination creatinine and cystatin C equation had 69% of results within 20% of GFR (P20), a sensitivity of 86.3% and a specificity of 73.1% to detect reduced GFR at <90 mL/min/1.73 m(2), and a concordance of 78%. In contrast, the traditional Cockcroft-Gault equation had a P20 of 38.0%, with a large underestimation to predict GFR, thereby accounting for approximately 45% of dosing discordance. No obvious differences were obtained when comparing the performance of equations using the two creatinine methodologies. The inclusion of cystatin C in the CKD-EPI equations improved the prediction of kidney function in the overall population, although probably not sufficiently for it to be favored over radioisotopic GFR for guiding chemotherapy. More research is warranted to further improve estimated GFR equations for these purposes.