Objective: To compare the bias, absolute bias, precision and accuracies between the equations, viz., CKD-EPI (Scr), CKD-EPI (Scys) and MDRD in Indian patients with type 2 diabetes.
Methods: 198 patients who underwent 24 h urinary collection for assessing kidney function between November 2014-January 2015 were included. Cohen's κ coefficient, Bland-Altman plot were calculated between estimated kidney function equations, and bias, precision, accuracies was calculated between the formulae.
Results: The mean eGFR based on MDRD, CKD-EPI (Scr) and CKD-EPI (Scys) equations were 64.5±21.9, 70.2±25.1 and 74.7±31.0 ml/min/ 1.73m2 respectively. The overall mean absolute bias was smallest for MDRD vs CKD EPI (Scr). The precision was also least for MDRD vs CKD EPI (Scr) indicating that the agreement between these equations is consistent for the range of values. MDRD vs CKD EPI (Scr) had the highest accuracy in comparison to other compared formula. The performance between MDRD versus CKD EPI (Scys) was different. There was a good agreement between MDRD and CKD EPI (Scr).in both stage 3 and stage 4 CKD. The MDRD vs CKD EPI (Scr) classified 72.2% of the patients correctly.
Conclusions: In conclusion, there was a good agreement between CKD-EPI (Scr) and MDRD equations. CKD-EPI equation based on creatinine estimation is widely accepted method and clinicians may use this equation in routine clinical practice to assess kidney function among patients with type 2 diabetes.