Fig 1.
Overview of experimental design.
We randomly selected 529 patient samples from outpatient samples that were submitted to the laboratory for SCr measurements (Population 1). This group of samples was used for the method comparison study and to evaluate the analytical precision of the Jaffe and enzymatic methods. We estimated biological variability in the outpatient population by identifying all outpatients who had at least two SCr measurements in 2013 (Population 2). This group contained 13,243 patients with a total of 42,195 SCr measurements. We evaluated the magnitude of discordances (standardized discordances) and the frequency of discordances (expected discordance vs observed discordance).
Fig 2.
The solid lines indicate zones centered at 15, 30, 45, and 60 ml/min/1.73 m2. Lower: Relationship between serum creatinine and eGFR. eGFR was estimated using the Chronic Kidney Disease Epidemiology (CKD-EPI) equation.
Table 1.
Comparison of precision of the Jaffe and enzymatic methods for serum creatinine (SCr).
Fig 3.
Precision Profile for eGFR Measurements.
These Figs show the standard deviation of the eGFR as a function of eGFR for the CKD-EPI method for the enzymatic (left panel) and Jaffe method (right panel). All data (age, sex, and race) are combined.
Fig 4.
Bland-Altman plots for creatinine and eGFR.
Upper panel: Creatinine. Lower panel: eGFR.
Fig 5.
Biological Variability of eGFR.
Each point represents the difference between consecutive eGFR measurements in one patient. The dashed line represents the estimated limit of biological deviation (95% confidence limits).
Fig 6.
Comparison of analytical variation, biological variation and unconditional discordance.
The inner lines (maroon) are the limits of analytical variation and the outer (black) lines are the limits of biological variation. Forty-two percent of the discordances exceed the limits of analytical variation. None of the discordances exceed the limits of biological variation.
Fig 7.
Comparison of observed discordances to analytical variation.
Each line represents the difference between the Jaffe result and enzymatic result. The vertical lines indicate the decision limit and two standard deviations of the difference (Jaffe-enzymatic) due to measurement imprecision at the decision limit. The arrows are directed from the enzymatic result toward the Jaffe result (the Jaffe result is greater than enzymatic when arrows point from left to right). Heavy lines indicate statistically significant discordances (i.e., greater than two standard deviations of the analytical variation) and light lines indicate nonsignificant discordances.
Table 2.
Discordance rate by eGFR decision point.
Table 3.
Analysis of expected discordance rates.
Table 4.
Risk assessment for Jaffe method.