5

of 14
14 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Clinical Chemistry 52:1 5–18 (2006) Special Report Recommendations for Improving Serum Creatinine Measurement: A Report from the Laboratory Working Group of the National Kidney Disease Education Program Gary L. Myers,1* W. Greg Miller,2 Josef Coresh,3 James Fleming,4 Neil Greenberg,5 Tom Greene,6 Thomas Hostetter,7 Andrew S. Levey,8 Mauro Panteghini,9 Michael Welch,10 and John H. Eckfeldt11 for the National Kidney Disease Education Program Laboratory Working Group Background: Reliable serum cr
Document Share
Document Tags
Document Transcript
  Recommendations for Improving Serum CreatinineMeasurement: A Report from the LaboratoryWorking Group of the National Kidney DiseaseEducation Program Gary L. Myers, 1* W. Greg Miller, 2  Josef Coresh, 3  James Fleming, 4 Neil Greenberg, 5 Tom Greene, 6 Thomas Hostetter, 7 Andrew S. Levey, 8 Mauro Panteghini, 9 Michael Welch, 10 and John H. Eckfeldt 11 for the National Kidney Disease Education Program Laboratory Working Group Background: Reliable serum creatinine measurementsin glomerular filtration rate (GFR) estimation are criticalto ongoing global public health efforts to increase thediagnosis and treatment of chronic kidney disease(CKD). We present an overview of the commonly usedmethods for the determination of serum creatinine,method limitations, and method performance in con-junction with the development of analytical performancecriteria. Available resources for standardization of serumcreatinine measurement are discussed, and recommenda-tions for measurement improvement are given.  Methods: The National Kidney Disease Education Pro-gram (NKDEP) Laboratory Working Group reviewedproblems related to serum creatinine measurement forestimating GFR and prepared recommendations to stan-dardize and improve creatinine measurement.  Results: The NKDEP Laboratory Working Group, in col-laboration with international professional organizations,has developed a plan that enables standardization andimprovedaccuracy(trueness)ofserumcreatininemeasure-ments in clinical laboratories worldwide that includes theuse of the estimating equation for GFR based on serumcreatinine concentration that was developed from theModification of Diet in Renal Disease (MDRD) study. Conclusions: The current variability in serum creatininemeasurements renders all estimating equations for GFR,including the MDRD Study equation, less accurate in thenormal and slightly increased range of serum creatinineconcentrations [ < 133  mol/L (1.5 mg/dL)], which is therelevant range for detecting CKD [ < 60 mL  min  1  (1.73m 2 )  1 ]. Many automated routine methods for serum creat-inine measurement meet or exceed the required precision;therefore, reduction of analytical bias in creatinine assaysis needed. Standardization of calibration does not correctfor analytical interferences (nonspecificity bias). The biasand nonspecificity problems associated with some of theroutine methods must be addressed. © 2006 American Association for Clinical Chemistry Chronic kidney disease (CKD) 12 is a major public healthproblem in the United States. The incidence and preva-lence of end-stage renal disease, kidney failure treated by 1 Division of Laboratory Sciences, National Center for EnvironmentalHealth, Centers for Disease Control and Prevention, Atlanta, GA. 2 Department of Pathology, Virginia Commonwealth University, Rich-mond, VA. 3 Department of Epidemiology, Johns Hopkins University, Baltimore, MD. 4 Department of Science and Technology, Laboratory Corporation ofAmerica, Elon, NC. 5 Ortho Clinical Diagnostics, Rochester, NY. 6 Department of Quantitative Health Science, Cleveland Clinic Foundation,Cleveland, OH. 7 National Kidney Disease Education Program, National Institute of Dia- betes and Digestive and Kidney Diseases, Bethesda, MD. 8 Division of Nephrology, Tufts New England Medical Center, Boston, MA. 9 Department of Clinical Sciences “Luigi Sacco”, University of Milan,Milan, Italy. 10 Chemical Science and Technology Laboratory, National Institute ofStandards and Technology, Gaithersburg, MD. 11 Department of Laboratory Medicine and Pathology, University of Min-nesota, Minneapolis, MN.* Address correspondence to this author at: Division of Laboratory Sci-ences, National Center for Environmental Health, Centers for Disease Controland Prevention, Atlanta, GA 30341. Fax 770-488-4192; e-mail GMyers@cdc.gov.Received April 26, 2005; accepted September 2, 2005.Previously published online at DOI: 10.1373/clinchem.2005.0525144 12 Nonstandard abbreviations: CKD, chronic kidney disease; GFR, glomer-ular filtration rate; K/DOQI, Kidney Disease Outcomes Quality Initiative;NKDEP, National Kidney Disease Education Program; MDRD, Modification ofDiet and Renal Disease; IDMS, isotope dilution mass spectrometry; GC, gaschromatography; LC, liquid chromatography; PT, proficiency testing; EQAS,external quality assurance scheme; CAP, College of American Pathologists;IVD, in vitro diagnostics; ISO, International Organization for Standardization; JCTLM, Joint Committee on Traceability in Laboratory Medicine; and CLSI,Clinical and Laboratory Standards Institute. Clinical Chemistry 52:15–18 (2006) Special Report 5  dialysis, and transplantation have more than quadrupledover the last 2 decades (1) . The estimated number ofpeople with earlier stages of CKD is  19 million, includ-ing  8 million people with a reduced glomerular filtra-tion rate (GFR)  60 mL  min  1  (1.73 m 2 )  1 and another  11 million with a GFR  60 mL  min  1  (1.73 m 2 )  1  butan abnormally high albumin excretion (urine albumin-to-creatinine ratio  30 mg/g) (2) .The National Kidney Foundation Kidney Disease Out-comes Quality Initiative (K/DOQI) and the NationalKidney Disease Education Program (NKDEP) within theNational Institute of Diabetes and Digestive and KidneyDiseases recently defined CKD as either kidney damageor a GFR  60 mL  min  1  (1.73 m 2 )  1 for 3 months ormore, irrespective of cause, and classified stages of CKDseverity based on GFR (3) . GFR is traditionally consideredthe best overall index of kidney function (4) . The thresh-old of GFR  60 mL  min  1  (1.73 m 2 )  1 was selected asthe definition of CKD because at this value approximatelyone half of an adult’s normal kidney function is lost,leading to several possible complications (3) .Understanding by laboratorians worldwide of the im-portance of reliable serum creatinine measurements inGFR estimation and of factors that may affect creatininemeasurement is critical to ongoing global public healthefforts to increase the diagnosis and treatment of patientswith CKD. The NKDEP Laboratory Working Group, incollaboration with international professional organiza-tions, has developed a plan that enables standardizationand improved accuracy (trueness) of serum creatininemeasurements in clinical laboratories worldwide. Materials and Methods for Estimating GFR GFR cannot be measured by direct means, but it can beassessed by measuring the urinary clearance of exogenousfiltration markers such as inulin, iohexol, or iothalamate (5–7) . However, because of difficulty in use, expense,radiation exposure, and radionuclide regulatory require-ments, these methods have limited use in the routinelaboratory and are typically confined to the researchsetting.GFR is often estimated clinically from serum concen-trations of endogenous creatinine (8) or cystatin C (9 , 10) .Serum cystatin C has not yet been adequately evaluatedas an index of GFR (11) , however, and serum creatininealone should not be used to assess the GFR or to detect thepresence of CKD because it is affected by the GFR and byfactors independent of GFR, including age, sex, race, bodysize, diet, certain drugs, and laboratory analytical meth-ods (12 , 13) . More accurate and precise estimations ofGFR can be obtained with equations that empiricallycombine all of the average effects from factors that affectserum creatinine other than GFR (14) . The currentlyrecommended estimating equation was developed fromthe Modification of Diet in Renal Disease (MDRD) Study (15) and is based on GFR values measured by iothalamateclearance in 1628 adults and subsequently validated inanother 1775 adults in the African American Study ofKidney Disease (AASK) (16) . The “four-variable” MDRDStudy equation (Eq. 1) uses age, sex, race (African-Amer-ican vs non–African-American), and serum creatinine(sCr) (17) :GFR [mL  min  1  (1.73 m 2 )  1 ]  186  (sCr)  1.154  (age)  0.203  (0.742 if female)  (1.210 if African-American) (1)The MDRD Study equation does not require a bodyweight variable because it normalizes GFR for a standard body surface area of 1.73 m 2 . The MDRD Study equationhas been demonstrated to be useful for CKD patients andperforms similarly in diabetic vs nondiabetic individuals (18) , but its use is unclear for healthy individuals and isnot recommended for hospitalized patients (19) .Because of the current variability in calibration ofserum creatinine assays, assays not calibrated in agree-ment with the kinetic alkaline picrate assay used in theMDRD Study introduce a source of error into GFR esti-mates. This calibration error is relatively greater andcontributes to larger uncertainty in GFR estimates atlower creatinine values near the upper limit of the refer-ence interval (20) . The progressively larger effect onestimated GFR of different calibration biases of creatininemethods is shown in Fig. 1 (21) , and the progressivelylarger effect of measurement imprecision at lower creati-nine values is shown in Fig. 2. Thus, calibration bias andmeasurement imprecision for serum creatinine have amuch larger impact on the uncertainty in estimated GFRwhen serum creatinine is close to the reference value, Fig. 1. Effect of creatinine calibration bias on estimated GFR. Lines represent estimated GFR values with no bias and with the indicatedamount of bias in serum creatinine measurements for a 60-year-old non–African-American female for whom the estimated GFR is 60 mL  min  1  (1.73 m 2 )  1 ata creatinine of 88.4  mol/L (1.00 mg/dL). The biases shown represent theminimum, maximum, and frequently observed values for 50 different methodgroups assaying a fresh-frozen serum specimen in the 2003 CAP survey (126)  .For an estimated GFR of 60 mL  min  1  (1.73 m 2 )  1 , a calibration difference of 11  mol/L (0.12 mg/dL) is associated with an error in GFR estimate of   12%.The error in GFR estimates over the range of biases observed is from  7.5% to  27%. Fig. 1 was derived from Murthy et al. (21)  , but is updated here to thebiases observed in the 2003 CAP survey. 6 Myers et al.: Improving Serum Creatinine Measurement to Estimate GFR  which is the relevant range for detecting early CKD [GFR  60 mL  min  1  (1.73 m 2 )  1 ]. This limitation applies toall estimating equations based on serum creatinine, not just the MDRD Study equation (22) . For this reason, theNKDEP currently recommends that GFR estimates above60 mL  min  1  (1.73 m 2 )  1  be reported simply as “  60mL  min  1  (1.73 m 2 )  1 ” rather than as a discrete numericvalue (3 , 18) . Variability in creatinine calibration andmeasurement imprecision also contributes to substantialuncertainty in estimating GFR in children, who usuallyhave lower serum creatinine concentrations than doadults. For estimating GFR in children, the Schwartz andthe Counahan–Barratt equations are recommended (23–26) . Both provide GFR estimates based on a constantmultiplied by the child’s height divided by the measuredserum creatinine concentration.  Sources of Variability in Estimating GFR Sources of variability in GFR estimates include underly-ing biological variability in GFR, biological variability inserum creatinine, and errors in the measurement of serumcreatinine and in the estimating equation.GFR may vary in response to meals, exercise, posture,changes in blood pressure, and other conditions. GFR isalso affected by pregnancy, glucose control in diabetes,extracellular fluid volume, antihypertensive medications,and acute and chronic kidney disease (27) . Error mayoccur in measurement of serum and urine filtration mark-ers or of urine flow rate, or in techniques for urinecollection. Variability among clearance periods duringGFR measurement may also lead to error (28) . Medianintraindividual CVs reported for measured GFR rangedfrom 6.3% to 7.5% (6 , 29) . These GFR measurements weremade in controlled studies; consequently, the intraindi-vidual variability was likely lower than would be ob-served in a typical clinical setting.In 2 published studies, the mean intraindividual CVsfor serum creatinine were 4.1% and 4.3%, respectively (30 , 31) . Neither of these studies included patients withCKD.  Analytical Performance Specifications for GFR Estimates Percentile distribution of the differences between esti-mated and measured GFR is a useful measure to assessthe accuracy of GFR estimates. The K/DOQI reported thatfor an independent sample of 1070 participants evaluatedin the GFR range  90 mL  min  1  (1.73 m 2 )  1 ,  90% ofGFR estimates calculated by use of the MDRD Studyequation were within 30% of the measured GFR (3) . Thisoverall measure of clinical performance included errorcomponents from several sources: measurement of serumcreatinine, including specimen nonspecificity effects andthe effects on determinants of serum creatinine other thanGFR, including generation, secretion, and elimination;and from measurement of GFR as iothalamate clearance,including physiologic differences in renal function andvarious comorbid conditions.Considering the various types of error, an estimatedGFR within 30% of a measured GFR was consideredacceptable by K/DOQI for clinical interpretation to iden-tify individuals with CKD as defined by GFR  60mL  min  1  (1.73 m 2 )  1 for 3 months or more and tofollow subsequent progression of the disease. For exam-ple, at a GFR of 60 mL  min  1  (1.73 m 2 )  1 , the range ofGFR values would be 42–78 mL  min  1  (1.73 m 2 )  1 .  Analytical Performance Specifications for Serum Creatinine Measurement Serum creatinine measurements must have a smallenough total error that the impact on the total uncertaintyof estimated GFR remains within clinically acceptablelimits. The critical serum creatinine concentration corre-sponding to a GFR of 60 mL  min  1  (1.73 m 2 )  1 varieswith the age, sex, and race of the patient (3) . Typicalvalues for serum creatinine at this critical GFR are 88.4  mol/L (1.00 mg/dL) for a 60-year-old non–African-American female, 99  mol/L (1.18 mg/dL) for a 60-year-old African-American female, 114  mol/L (1.30 mg/dL)for a 60-year-old non–African-American male, and 135  mol/L (1.53 mg/dL) for a 60-year-old African-Americanmale. Thus, creatinine values within or very close to manypublished reference intervals are consistent with substan-tial reduction in GFR in some patients. For the samedemographic groups at an estimated GFR of 30mL  min  1  (1.73 m 2 )  1 , the serum creatinine values are162, 190, 209, and 247  mol/L (1.83, 2.15, 2.37, and 2.79mg/dL), respectively. Because of the dramatic increase inthe impact of creatinine bias and imprecision on the errorof an estimated GFR as the serum creatinine value getssmaller (Figs. 1 and 2), the laboratory measurementperformance goal is currently targeted at a creatinine Fig. 2. Effect of creatinine measurement imprecision on estimatedGFR. Solid lines represent the upper and lower limits of the 95% confidence interval forestimated GFR for a 60-year-old non–African-American female for whom theestimated GFR is 60 mL  min  1  (1.73 m 2 )  1 at a creatinine of 88.4  mol/L(1.00 mg/dL), using a value of 5.3  mol/L (0.06 mg/dL) as the measurementSD. This SD was the median SD observed for 50 different method groupsassaying a fresh-frozen serum specimen with a creatinine value of 80  mol/L(0.90 mg/dL) in the 2003 CAP survey (126)  . The dashed lines represent theupper and lower limits of the 95% confidence interval for estimated GFR basedon the largest peer-group SD, 12  mol/L (0.13 mg/dL), observed in the survey. Clinical Chemistry 52, No. 1, 2006 7  concentration of 88.4  mol/L (1.00 mg/dL), which isconsistent with a GFR of 60 mL  min  1  (1.73 m 2 )  1 forsome adult demographic groups and is at the lower rangeof measurement, where the impact of bias and impreci-sion will be greater than at higher values.The 2 primary components of measurement error inserum creatinine are systematic bias, a consistent errortypically resulting from calibration differences betweenmeasurement procedures, and random measurement er-ror, including within-laboratory effects, between-labora-tory random variability in day-to-day calibration, andspecimen-specific effects. In a simulation study, gaussian-distributed random errors and increasing systematic bi-ases were added to the baseline serum creatinine mea-surements of 491 patients in the MDRD Study validationsubset who had serum creatinine measurements between88.4 and 132  mol/L (1.00 and 1.50 mg/dL). The increasein root mean square error in the estimated GFR, comparedwith an iothalamate-measured GFR, was calculated foreach increment in added bias and imprecision.The upper bounds for combinations of systematic biasand imprecision in a serum creatinine measurement thatwould increase the root mean square error in estimatingGFR by no more than 10% [an arbitrary modest increaseconsistent with previous recommendations for the impactof measurement error on clinical utility of laboratoryresults (32) ] are shown in Fig. 3. The serum creatininemeasurements obtained in the MDRD Study were as-sumed to have zero bias; thus, the bias increments should be interpreted as a difference from a zero bias condition.The SD increments were added to the underlying SD inthe MDRD Study [2.65  mol/L (0.03 mg/dL) for creati-nine in the 88.4–133  mol/L (1.00–1.50 mg/dL) range].The line in Fig. 3 represents combinations of added biasand SD at which the root mean square error was  12.22mL  min  1  (1.73 m 2 )  1 (a 10% increase). Under the con-ditions of the simulation analysis, the method perfor-mance parameters in Fig. 3 should generalize to othersettings. A more detailed description of the simulationanalysis can be found in the online Data Supplement thataccompanies this report at http://www.clinchem.org/content/vol52/issue1/.The required laboratory measurement performance forserum creatinine can also be estimated based solely on theunderlying biological variability (33) . Any approach mustconsider both imprecision and bias in making an estimateof analytical performance required to meet a clinicalinterpretation goal. A desirable imprecision goal has beenproposed as one-half the intraindividual biological vari-ability because this will not increase the total error morethan 12% (32) . A more recent recommendation for ana-lytical performance goals based on intra- and interindi-vidual biological variability has included both impreci-sion and bias and has empirically proposed tieredrecommendations in categories consistent with minimum,desirable, and optimal method performance to supportclinical interpretation of a result (33) . The desirable im-precision is consistent with the previous maximum 12%increase in total error, and the other categories for impre-cision and bias are arbitrary extensions to allow categori-zation of method performance to identify those that mayneed improvement. The analytical goals for serum creat-inine measurement using this approach are summarizedin Table 1. The minimum total error goal is estimated at11.4% and the desirable total error goal at 7.6%. Clinical Laboratory–Based Analytical Systems forMeasuring Serum Creatinine to Assess GFR The methods most widely used to measure serum creat-inine are alkaline picrate methods, enzymatic or partiallyenzymatic assays, and HPLC methods. Isotope-dilutionmass spectrometry (IDMS) high-order reference methodshave been developed for assignment of reference materi-als but are available in only a few highly specializedlaboratories worldwide.Because no systematic differences between serum andplasma measurements have been reported, we considerserum and plasma results as equivalent (34) . Serumcreatinine has been found to remain stable during long-term storage and after repeated thawing and refreezing (35) and for up to 24 h in clotted whole blood at roomtemperature (36) . alkaline picrate methods The method of Jaffe (37) is commonly used to measureserum creatinine in routine laboratories. The presence ofinterfering substances, particularly proteins, in serum canlead to the overestimation of serum creatinine by as muchas 15%–25% by various Jaffe methodologic applications.Many endogenous and exogenous interfering substancescontribute to the lack of analytical specificity in the Jaffemethod (38–76) . Interferences from glucose (65–67) andacetoacetate (58) are particularly important because dia- betic persons are a high-risk population to develop CKD.Several modifications, including optimization of kinetic Fig. 3. Total error budget for creatinine measurements in the range88.4–133  mol/L (1.00–1.50 mg/dL). The line  represents the limit of systematic biases and random imprecisions thatproduce a relative increase of   10% in the root mean square error whenestimating GFR using the MDRD Study equation. 8 Myers et al.: Improving Serum Creatinine Measurement to Estimate GFR
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x