Combined creatinine-cystatin C CKD-EPI equation significantly underestimates measured glomerular filtration rate in people with type 2 diabetes mellitus

To evaluate the accuracy of creatinine and cystatin C (cysC) equations to estimate glomerular filtration rate (GFR) in type 2 diabetes mellitus (DM) patients and healthy adults. Case-control study including 84 patients with type 2 DM and 100 healthy adults with measured GFR (mGFR)≥60mL/min/1.73m2. GFR was measured by 51Cr-EDTA and estimated (eGFR) by the following equations using creatinine, cysC or both markers: Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), Caucasian Asian Pediatrics and Adults (CAPA), CKD-EPI creatinine-cystatin C (CKDEPI-CC), and CKD-EPI cystatin C (CKDEPIcysC). Agreement was evaluated by Bland & Altman analysis. Healthy individuals were 66% females, aged 38±14years; they presented mGFR 112±19mL/min/1.73m2 and eGFR by CKD-EPI, CKDEPI-CC, CKDEPIcysC and CAPA equations, respectively, 108±17, 102±15, 97±16 and 93±16mL/min/1.73m2. DM group were 50% females, aged 59±19years and presented mGFR 104±27 and eGFR 87±19, 80±18, 74±20 and 73±18mL/min/1.73m2, respectively. All equations significantly underestimated mGFR, excepting creatinine-based CKD-EPI in the healthy group. The performance was considerably worse for GFRs above 120mL/min/1.73m2. In both healthy and type 2 DM patients, cystatin C-based equations, including the combined CKD-EPI creatinine-cystatin equation, failed to improve the accuracy of GFR estimation, especially for normal and high normal GFR values.