The CKD-EPI and MDRD equations to estimate GFR. Validation in the Swedish Lund-Malmö Study cohort

To compare the recently developed CKD-EPI equation to estimate GFR in adult Swedish-Caucasians with the MDRD equation.

Swedish-Caucasians (N = 850, 376 females; median age 60, range 5-95 years) referred for plasma iohexol-clearance (median 55, range 5-223 mL/min/1.73 m²) constituted the Lund-Malmö Study cohort. Bias, precision (interquartile range, IQR, of the differences between estimated and measured GFR), accuracy expressed as percentage of estimates ±10% (P₁₀) and ±30% (P₃₀) of measured GFR, and classification ability for five GFR stages <15, 15-29, 30-59, 60-89 and ≥90 mL/min/1.73 m² were compared.

Overall there were no important differences between the equations; CKD-EPI/MDRD median values of bias +5.4%/+3.4%, IQR both 14 mL/min/1.73 m², P₁₀ 36%/34%, P₃₀ both 80%, and correctly classified GFR stages 68%/67%. P₃₀ for the CKD-EPI equation was substantially higher than for MDRD at GFR ≥90 mL/min/1.73 m² (93% versus 79%). The MDRD equation performed better in the GFR interval 30-89 mL/min/1.73 m², while accuracy was limited for both equations at GFR <30 mL/min/1.73 m² (P(30) <75% in both females and males). The CKD-EPI/MDRD equations caused a +22%/+14% bias in the 18-39 year interval and the MDRD equation a +18% bias ≥80 years. Both equations performed poorly in males with BMI <20 kg/m² (CKD-EPI/MDRD median bias +36%/46%).

Overall the recently developed CKD-EPI equation performed well but was not superior to the MDRD equation. The CKD-EPI equation may be preferred in screenings of general populations and in the elderly. None of the equations appeared reliable among patients with markedly decreased GFR, young adults and underweight males.