A comparison of creatinine measurement by the Jaffe and enzymatic methods in an outpatient population

Serum creatinine concentrations and estimated glomerular filtration rates (eGFR) are widely used for the evaluation of renal function. The Jaffe and enzymatic methods are the most common methods for creatinine measurement. The Jaffe method is commonly less expensive than enzymatic methods but may be more susceptible to interferences. Significant savings could be obtained if populations could be identified where the interference rate of the Jaffe method is acceptably low. The study objective was to compare creatinine and corresponding eGFR results from representative Jaffe and enzymatic creatinine methods in an outpatient population and determine the prevalence and magnitude of differences. This study analyzed 543 unique, randomly selected, outpatient samples. Samples were analyzed using both the Jaffe and enzymatic creatinine methods using an Abbott Architect c8000. eGFRs were calculated using the CKD-EPI and MDRD equations. Orthogonal (Deming) regression showed no significant difference between the two assays. The slope was 1.006 (95% CI: 0.998, 1.103) and the intercept was -0.005 (95% CI: -0.015, 0.006). The average difference (bias) was -0.007 mg/dL. The Bland- Altman limits of agreement (LOA) for the creatinine differences were -0.139 and 0.136 mg/dL. Of the CKD-EPI eGFR discrepancies, 3.1% resulted in a reclassification at the 60 mL/min/1.73m2 decision limit. The difference between the predicted and actual CKDEPI discordance at the 60 mL/min/1.73m2 decision limit was not statistically significant (χ! ! = -0.13, p=0.89). Of the MDRD eGFR discrepancies, 4.8% resulted in a iv reclassification at the 60 mL/min/1.73m2 decision limit. The difference between the predicted and actual MDRD discordance at the 60 mL/min/1.73m2 decision limit was not statistically significant (χ! ! = 0.31, p=0.76). Discrepancies in the CKD-EPI and MDRD eGFRs based on the Jaffe method did not result in a statistically significant increase in disease reclassifications at the 60 mL/min/1.73m2 decision limit in an outpatient population. The number of discordant eGFR results slightly varied by what discordance criteria was used. An equal number of discordances were observed with the MDRD and CKD-EPI equations when eGFR were based upon measurement error (217 of 543). When discordance criteria was based upon observed differences (Bland-Altman), the MDRD equation showed slightly more discordances (26 of 543) compared to the CKD-EPI (21 of 543). Studies are needed to characterize the relative rate of interference in additional populations.