Fig. 3 depicts measured and estimated GFR for each group. In diabetic
patients, both creatinine methods largely underestimated measured
GFR. In this group, eGFR using enzymatic creatinine was higher,
and therefore closer to measured GFR than was Jaffe creatinine based
equation. Accuracy (the number of cases in which eGFR differed less
than 30% from measured GFR) was 87% for CKD-EPIenz and 86% for
CKD-EPIJaffe in healthy subjects. In the diabetic patients, accuracy
was significantly lower: 64% for CKD-EPIenz and CKD-EPIJaffe alike
(P b 0.01). Fig. 4 shows Bland–Altman plots for the agreement between
eGFR and measured GFR in the sample as a whole and in the
healthy and diabetic groups.