Portal de Periódicos da CAPES

How Valid Are GFR Estimation Results From the CKD-EPI Databases?

2017; Elsevier BV; Volume: 71; Issue: 3 Linguagem: Inglês

10.1053/j.ajkd.2017.11.003

1523-6838

Jonas Björk, Sten-Erik Bäck, Gunnar Nordin, Ulf Nyman,

Renal cell carcinoma treatment

Levey et al1Levey A.S. Tighiouart H. Simon A.L. Inker L.A. Comparing newer GFR estimating equations using creatinine and cystatin C to the CKD-EPI equations in adults.Am J Kidney Dis. 2017; 70: 587-589Abstract Full Text Full Text PDF PubMed Scopus (26) Google Scholar concluded that current evidence does not support reporting estimated glomerular filtration rate using any other creatinine equation than CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). However, their results do not provide evidence of external validity; the original authors of the equation,2Levey A.S. Stevens L.A. Schmid C.H. et al.A new equation to estimate glomerular filtration rate.Ann Intern Med. 2009; 150: 604-612Crossref PubMed Scopus (16096) Google Scholar after analyzing multiple models in the CKD-EPI database (n = 3,896), stated “We selected a single model as the best equation for general use.”2Levey A.S. Stevens L.A. Schmid C.H. et al.A new equation to estimate glomerular filtration rate.Ann Intern Med. 2009; 150: 604-612Crossref PubMed Scopus (16096) Google Scholar A true external validation should not influence the equation coefficients or which equation is finally selected. By contrast, results from several postdevelopment validation studies suggest that the Lund-Malmö-revised (LMR) equation is equally or more accurate than CKD-EPI.3Björk J, Bäck SE, Ebert N, et al. GFR estimation based on standardized creatinine and cystatin C: a European multicenter analysis in older adults [published online ahead of print October 6, 2017]. Clin Chem Lab Med. https://doi.org/10.1515/cclm-2017-0563.Google Scholar, 4Evans M. van Stralen K.J. Schon S. et al.Glomerular filtration rate-estimating equations for patients with advanced chronic kidney disease.Nephrol Dial Transplant. 2013; 28: 2518-2526Crossref PubMed Scopus (48) Google Scholar The utilized indirect creatinine calibration may also jeopardize the conclusion by Levey et al.1Levey A.S. Tighiouart H. Simon A.L. Inker L.A. Comparing newer GFR estimating equations using creatinine and cystatin C to the CKD-EPI equations in adults.Am J Kidney Dis. 2017; 70: 587-589Abstract Full Text Full Text PDF PubMed Scopus (26) Google Scholar Creatinine values from various data sets2Levey A.S. Stevens L.A. Schmid C.H. et al.A new equation to estimate glomerular filtration rate.Ann Intern Med. 2009; 150: 604-612Crossref PubMed Scopus (16096) Google Scholar were first aligned to the Beckman Synchron-CX3 Jaffé assay and then to Roche enzymatic using a simplistic adjustment equation: creatinineRoche = 0.906 × creatinineBeckman. This adjustment was in sharp contrast to that of Vickery et al5Vickery S. Stevens P.E. Dalton R.N. van Lente F. Lamb E.J. Does the ID-MS traceable MDRD equation work and is it suitable for use with compensated Jaffe and enzymatic creatinine assays?.Nephrol Dial Transplant. 2006; 21: 2439-2445Crossref PubMed Scopus (149) Google Scholar: creatinineIDMS = −0.184 mg/dL (−16.3 μmol/L) + 0.998 × creatinineBeckman. Similar results were seen when creatinine samples from our group (Lund, Sweden), measured with an externally quality assured (Equalis, Uppsala, Sweden) isotope-dilution mass spectrometry (IDMS)-traceable Roche enzymatic assay, were reanalyzed on the CX3 before inclusion in the CKD-EPI database2Levey A.S. Stevens L.A. Schmid C.H. et al.A new equation to estimate glomerular filtration rate.Ann Intern Med. 2009; 150: 604-612Crossref PubMed Scopus (16096) Google Scholar: creatinineLund = −0.185 mg/dL (−16.3 μmol/L) + 0.954 × creatinineBeckman (rearranged equation based on personal communication, CKD-EPI Study Group, 2011). Our tentative reanalysis for the CKD-EPI database, using the potentially more accurate creatinine adjustment equation by Vickery et al, eliminates the superiority of the CKD-EPI equation vis-à-vis LMR (Table 1). Findings from the CKD-EPI databases regarding the merits of any creatinine equation should therefore be interpreted with caution until IDMS traceability can be established.Table 1Effects of the Creatinine Adjustment on the Performance of GFR Estimation Equations in the CKD-EPI Validation DatabaseCreatinine AdjustmentEquationBias, Median, mL/min/1.73 m2Precision, IQR, mL/min/1.73 m2Accuracy, P30, %Original1Levey A.S. Tighiouart H. Simon A.L. Inker L.A. Comparing newer GFR estimating equations using creatinine and cystatin C to the CKD-EPI equations in adults.Am J Kidney Dis. 2017; 70: 587-589Abstract Full Text Full Text PDF PubMed Scopus (26) Google ScholarCKD-EPI−2.2 (−2.6 to −1.8)16.6 (15.8 to 17.2)84.2 (83.0 to 85.3)Original1Levey A.S. Tighiouart H. Simon A.L. Inker L.A. Comparing newer GFR estimating equations using creatinine and cystatin C to the CKD-EPI equations in adults.Am J Kidney Dis. 2017; 70: 587-589Abstract Full Text Full Text PDF PubMed Scopus (26) Google ScholarLMR−7.4 (−7.8 to −6.8)18.2 (17.6 to 19.1)79.7 (78.4 to 81.0)ReanalysisCKD-EPI0.5 (0.1 to 1.0)17.0 (16.2 to 17.9)80.9 (79.7 to 82.1)ReanalysisLMR−4.3 (−4.8 to −4.0)16.8 (15.9 to 17.5)81.9 (80.7 to 83.1)Note: n = 3,896. In the original CKD-EPI Study, slope 0.906 and intercept 0.0 were used to obtain IDMS-traceable creatinine values from the Beckman Synchron-CX3 Jaffé assay. Our reanalysis, made possible through cooperation between the Swedish Council on Health Technology Assessment (www.sbu.se/214) and the CKD-EPI Coordinating Center, instead used slope 0.998 and intercept −0.184 mg/dL.5Vickery S. Stevens P.E. Dalton R.N. van Lente F. Lamb E.J. Does the ID-MS traceable MDRD equation work and is it suitable for use with compensated Jaffe and enzymatic creatinine assays?.Nephrol Dial Transplant. 2006; 21: 2439-2445Crossref PubMed Scopus (149) Google Scholar Bias was assessed as the median difference between estimated and measured glomerular filtration rate (eGFR − mGFR) and presented with nonparametric confidence intervals. Precision was assessed as the interquartile range (IQR) of the difference between eGFR and mGFR and presented with bootstrap confidence intervals (1,000 replications). Accuracy, the percentage of participants for whom the difference between eGFR and mGFR is at most 30% of mGFR (P30), was presented with asymptotic confidence intervals. Open table in a new tab Note: n = 3,896. In the original CKD-EPI Study, slope 0.906 and intercept 0.0 were used to obtain IDMS-traceable creatinine values from the Beckman Synchron-CX3 Jaffé assay. Our reanalysis, made possible through cooperation between the Swedish Council on Health Technology Assessment (www.sbu.se/214) and the CKD-EPI Coordinating Center, instead used slope 0.998 and intercept −0.184 mg/dL.5Vickery S. Stevens P.E. Dalton R.N. van Lente F. Lamb E.J. Does the ID-MS traceable MDRD equation work and is it suitable for use with compensated Jaffe and enzymatic creatinine assays?.Nephrol Dial Transplant. 2006; 21: 2439-2445Crossref PubMed Scopus (149) Google Scholar Bias was assessed as the median difference between estimated and measured glomerular filtration rate (eGFR − mGFR) and presented with nonparametric confidence intervals. Precision was assessed as the interquartile range (IQR) of the difference between eGFR and mGFR and presented with bootstrap confidence intervals (1,000 replications). Accuracy, the percentage of participants for whom the difference between eGFR and mGFR is at most 30% of mGFR (P30), was presented with asymptotic confidence intervals.