实验室质控改善方案

一、前言

6月18日我公司送水样（南玻项目水样）外检，外检项目包括铁、铜、锌、总磷和氨氮共计5个项目。根据广州工业微生物检测中心的检测报告，铁、铜、锌和总磷的分析结果和我公司实验室分析结果相差非常大，基于第三方检测公司的权威和可靠性，为了提高我公司实验室检测结果的准确度，特提出以下实验室质控改善方案。

二、质控改善方案

1、实验室现状

1.1 铁、铜、锌、总磷等分析项目制作工作曲线时使用分析纯试剂配制标准溶液，而国家标准分析方法的要求标准溶液配制需用优级纯以上的试剂；

1.2 工作曲线没有定期校准和重新制作，有的工作曲线使用时间已达1年以上。分光光度计长时间使用后，其工作性能会出现一定的波动，这对被测溶液的吸光度会产生一定的影响。另外，使用不同批次试剂也会对同一浓度被测样品的吸光度产生影响。

1.3 样品抽检频率，每2个月一次，偏差率在13.9%~25%之间波动，远大于实验室质控不大于5%的偏差要求。

2、改善方案

2.1 购买铜、锌标准溶液（规格：1000ug/ml，50mg/瓶）、磷酸氢二钾（优级纯）和硫酸铁铵（优级纯）配制铜、锌、磷酸根和铁离子的标准溶液，重新制作工作曲线；

2.2 定期（2月/次）配制已知浓度的质控样进行分析，及时检验分析结果的准确度；控制分析结果偏差在5%以内；

2.3 在排除其他因素干扰的情况下（如加入的试剂过期、样品预处理不当、分光光度计工作是否正常等），根据质控样的偏差值大小确定是否需要重新制作工作曲线：

偏差值＜5%，OK！偏差值＞5% ，重新制作工作曲线。

2.4 工作曲线使用时间最长要求不超过6个月，即每使用6个月需重新制作一次。

2.5 实验室配制试剂需标注配制日期和有效期等，使用前先确定试剂是否在保质期内。若试剂过期，则需重新配制。

2.6固定一些比色管或者容量瓶用于分析同一的指标，可避免其他试剂带来的干扰

2.7 定期对实验人员进行理论考试，加强实验员的理论知识。

3、实验室常规项目质控要求一览表

技术部

20##年7月3日

第二篇：实验室质控

CCA-13276;NoofPages9

ClinicaChimicaActaxxx(2013)xxx–xxx

ContentslistsavailableatScienceDirect

ClinicaChimicaActa

journalhomepage:/locate/clinchim

Acategory1EQAschemeforcomparisonoflaboratoryperformanceandmethodperformance:AninternationalpilotstudyintheframeworkoftheCalibration2000project

RobJansena,?,NutharJassamb,AnnetteThomasc,CarmenPerichd,PilarFernandez-Called,AnaPaulaFariae,HelenaCorreiae,JulianH.Barthf,CasWeykampa,ChristaCobbaertg,MarcThelenh,CarmenRicósd

DutchFoundationforQualityAssessmentinMedicalLaboratories(SKML),Nijmegen,TheNetherlandsDepartmentofClinicalBiochemistry,HarrogateDistrictFoundationTrust,Harrogate,UKc

WEQASQualityLaboratory,CardiffandValeUniversityHealthBoard,Cardiff,UKd

SpanishSocietyofClinicalChemistryandMolecularPathology(SEQC),AnalyticalQualityCommission,Spaine

InstitutoNacionaldeSaude,Portugalf

BloodSciences,OldMedicalSchool,LeedsTeachingHospitalsTrust,Leeds,UKg

DepartmentofClinicalChemistry,LeidenUniversityMedicalCenter,Leiden,TheNetherlandsh

DepartmentofClinicalChemistryandHematology,AmphiaHospital,Breda,TheNetherlands

articleinfoabstract

Introduction:Inthemodernhealthcareservice,patientsreceivecareinmultiplehospitalsandhealthcaresettings.Therefore,harmonizationofresultsfromdifferentmethodsandinstruments,bothbetweenandwithinlaborato-ries,isoftheutmostimportance.ThepresentpilotstudyaimstotesttheuseofaCategory1EQAschemeacrossfourEuropeancountriesbyassessingthecurrentlevelofequivalenceoftestresults.

Method:ThisworkwasledbytheDutchExternalQualityAssuranceSchemeSKMLandinvolved28laboratoriesfromthreeregionsintheUK,SpainandPortugal,and120laboratoriesfromTheNetherlands.Asetofsixcom-mutablesamples,targetedwithreferencemethods,werecirculatedand18biochemistryanalytesweretested.Resultsandconclusions:TheTotalError(TE)score,de?nedastheprobability(%)thatresultsarewithintheTotalErrorAcceptable(TEA)limits,fortheeighteenanalyteswascalculated.Ourdatashowthatthereisaneedforfurtherharmonizationoflaboratorydata,inparticularforelectrolytes(calcium,chloride,magnesium,sodium),enzymes(ALT,amylase,AST,LDH),lipids(HDL-cholesterol),andforsubstrates(creatinine,totalprotein).Lackofperformanceconsistencybetweeninstrumentswasseenformostanalytes.Thelackofharmonizationisstillpresentdespitemanufacturerclaimsofestablishedtraceability.

?2013ElsevierB.V.Allrightsreserved.

Articlehistory:

Received25May2013

Receivedinrevisedform5November2013Accepted5November2013AvailableonlinexxxxKeywords:

ExternalqualityassessmentCommutabilityBiologicalvariationReferencemethodTraceabilityHarmonization

1.Introduction

Mosteffortsinthemanagementofanalyticalqualityinclinicalchemistryandlaboratorymedicinehavefocusedonthereductionofwithin-laboratoryvariationandtheassessmentofbetween-laboratoryvariation.Inrecentyearstheimportanceofminimizingbias,bothbetweenlaboratoriesandwithinalaboratory,hasbecomeparamount.Patientsarefrequentlytreatedbyateamofphysiciansratherthanone,oftenextendingacrossseveralhealthcaresettingsandmakinguseofinformationfromseverallaboratories.Inmonitoringpatientsduringtreatment,theabsenceofbiasfromonemeasurementtothenext,togetherwithminimumimprecisionisessential.Calibrationandharmonizationofresultsfromdifferentanalyzers,bothbetweenand

Abbreviations:EQA,externalqualityassessment;TE,totalerror;IVDD,invitrodiagnosticmedicaldevicesdirective.?Correspondingauthor.

E-mailaddress:of?ce@skml.nl(R.Jansen).

实验室质控

0009-8981/$–seefrontmatter?2013ElsevierB.V.Allrightsreserved./10.1016/j.cca.2013.11.003

withinlaboratories,andthecontinuityofsuchharmonizationintimeare,therefore,oftheutmostimportance.Smallassaybiasesmayhavealargeimpactonpatientclassi?cationandonthenumberofpatientstobetreated,particularlyforassaysforwhichcut-offvaluesareused.Thisistrue,forexample,inlipidandlipoproteinanalyses,inwhichstringentcut-offvaluesareusedthroughouttheworldforthepreven-tionandtreatmentofcardiovasculardiseases.Itisalsotrueforcreati-nineintheestimationofrenalfunctionandforhumangrowthhormoneinhGHde?ciency.

TheAmericanAssociationforClinicalChemistry(AACC)conferenceinOctober2010focusedontheroadmap[1]toreachharmonizationforanalytesforwhichnoreferencesystemisde?ned.However,evenforanalytesforwhichsuchsystemsexist,standardizationisoftenlacking.Theprocessisde?nedasstandardizationiftheanalyteisclearlyde?nedandreferencemethodandstandardsexist.Harmonizationiscon?nedtodescribeprocesseswhereoneormoreoftheseelementsaremissing.ExternalQualityAssessment(EQA)schemesshouldplayacentralroleinachievingharmonizationandintruenessveri?cation.Itiswidely

Pleasecitethisarticleas:JansenR,etal,Acategory1EQAschemeforcomparisonoflaboratoryperformanceandmethodperformance:Aninter-nationalpilotstudyintheframeworkoftheCalibration2000project,ClinChimActa(2013),/10.1016/j.cca.2013.11.003

2R.Jansenetal./ClinicaChimicaActaxxx(2013)xxx–xxx

Creatinine

30252015Difference μmol/L

1050-5-10-15-20-2560

100

120

140160μmol/l

180

200

μmol/l

SampleYouA120B191C154D146E207F78Total149Precision1.8%Count6Outliers0

Ref126.3196.0157.0149.0219.478.7154

Cons12619xxxxxxxxxxxx5675442

CV5.1%3.6%5.2%3.8%3.3%9.1%4.2%2.7%

Meth1271961621622188215626114

Anal12619213914921580151602

220240

Regression line Y=7 + 0.920XStatistics valid for your

Method:Alk. Picrate, kinetic with compensation-group:Jaffe

Analyser:Olympus AU2700

Fig.1.ExampleofadifferenceplotoftheCombischeme.ThegreenareaisTEAtoleranceareaaroundthereferencemethodtarget.Theblueareaisstateofthearttolerancelimitaroundtheconsensusmeanvalue.ThegreensquaresareyourresultsforsamplesA–F.Theblacklineisyourregressionline.Verticalbarsindicate±3SDbetweenlabs.Precisioniswithin-labCV.Anal.isyouranalyzer.

acceptedthatweneedcommutablematerials[2,3],referencemethodtargetvaluesandtolerancelimitsbasedonthebiologicalvariationconcept[4–6].EQAschemeshavingthesecharacteristicshavebeendenotedasCategory1schemes[7].TheimportanceofusingthisconceptinEQAschemeswasstressedrecentlyinseveralsessionsduringtheBio-RadConvocationofExpertsonLaboratoryQuality2010inBardolino,Italy[8]againin2011inSalzburg,Austria.IntheCalibration2000projectoftheDutchNEQASorganizerSKML,thiswasachievedforseveralanalytes[9–15].TheInVitroDiagnosticMedicalDevicesDirective(IVDD)requirestraceabilitytoreferencesystems[16].Thesesystemsarede?nedforanumberofanalytes.Fortheseanalytestruenessveri?cationispossibleandharmonizationiswithinreach.

TheCalibration2000projectinTheNetherlandsproducesmaterials[9–15]forgeneralclinicalchemistry,proventobecommutableinconformitywiththeClinicalandLaboratoryStandardsInstitute(CLSI)C53A[17].Thesamplesaretargetedwithreferencemethods,undertakenineitherTheJointCommitteeforTraceabilityinLaboratoryMedicine(JCTLM)listedReferencelaboratoriesorinInternationalFederationofClinicalChemistryandLaboratorymedicine(IFCC)networklaboratories,ifavailable,andresultsareprocessedintheCombiEQAscheme[11–14]inwhichparticipatinglaboratoriesassayseveralsamplescoveringtheclinicallyrelevantconcentrationrange.TheschemeusesthebiologicalvariationbasedTotalErrorallowable(TEA)atthedesirablelevelastolerancelimit.Harmoniza-tionofminimalacceptableperformancecriteriaamongEQAorga-nizersisdesirable[18].

Thepresentstudyisapilotstudy.ItaimstotesttheuseofaCategory1EQAschemeacrossthecountries,UK,Spain,PortugalandTheNetherlands,andtocompareinapilotstudytheperformanceoftheparticipatinglaboratoriesandthemethodsused.Theresultsofthepilotstudyareseenasapreliminaryviewoftheroleofcategory1EQAtoimproveharmonizationinEurope.2.Methods

IntheSKMLCombischeme,24samplesareanalyzedforgeneralchemistryparametersinthecourseofayear,i.e.atafrequencyofonesampleper2weeks.Forlipids,aseparatededicatedbatchof24samplesisused.Thesamplesarepreparedaccordingtoexactlythesameproto-colaspreviouslypreparedsampleswhichwereproventobecommut-able[9–15].Inshort,twomastersamplesareprepared,onefrompoolednormalhumanleftoverseraandonefrompoolednormalhumansera,spikedwithabnormalpools,minerals,recombinanthumanenzymesandhumanalbumin.Themasterpoolsaremixedintenproportionsthusobtaining12concentrationlevels.Afterdispens-ing,vialsarefrozenat?84°C.Previouslypreparedsamplesaccordingtothisprocedurewererepeatedlyproventobecommutable,whethermaster,spikedormixedsamples.Throughouttheyearscommutabilityhasbeenmonitoredbyincludinganative,singledonationspy-samplethatispreparedaccordingtoCLSIC37-A2.Concentrationscovertherangeofclinicalinterest.ThesamplesaretargetedbyJCTLMlistedlaboratoriesforelectrolytesandsubstrates,andbyIFCCorCDCnetworklaboratoriesforenzymesandlipids.Informationonrefer-encemethodsandlaboratoriesusedisprovidedassupplementarydata.Biologicalvariationbasedtolerancelimitsareused(TEAdesirable).

ThirtylaboratoriesfromthreeEuropeancountriesparticipatedinthisstudyinadditionto120regularlyparticipatingDutchSKMLEQACombischeme.TenlaboratorieseachparticipatedfromtheUK,Spain(onelabwithtwoproceduresforallanalytes,exceptforlipids)andPortugal.TheUKlaboratories'inclusioninthisstudywassolelybasedonexpressionofinterestfromlaboratoriesthathadreceivedaninvitationtoparticipate.Theauthorshavehadnopreviousknowledgeoftheanalyticalperformancefortheparticipatinglaboratories.

TheSpanishandthePortugueselaboratorieswereselectedfromthoselaboratoriesfallingwithinthe20thpercentileofthetargetdevia-tionoftheirnationalEQAschemes.However,theparticipatinglaborato-riesrangefromsmallindependenthealthcarelaboratoriestolargelaboratoriesservingteachinghospitals,amixofsizeandanalyticalplat-forms,whichre?ectsthesamedistributionineachcountry.

Asetofsixsamplesforgeneralchemistryandasetofsixsamplesforlipidsfrozenat?80°C,weretransportedondryicetoacentrallabora-toryineachofthethreecountries,andstoredat?80°C.Thefrozensamplesweredistributedondryicefromthecentrallaboratorytotheparticipatinglaboratories.SamplesarrivedthawedintwoPortugueselaboratoriesandthesewerediscarded.Sincemostofthelaboratories

R.Jansenetal./ClinicaChimicaActaxxx(2013)xxx–xxx3

Fig.2.TEscoresperanalyticalplatformfor28laboratories.TEscoresvaryconsiderablywithinusersofinstrumentsfromthesamemanufacturer.ThebluelinerepresentstheaverageTEscoreforTheNetherlandslaboratories.

didnothavea?80°Cfreezeravailable,thelaboratorieswereaskedtoanalyzethesamplesassoonaspossibleafterreceiptortostorethesam-plesat?20°Candanalyzethemwithin1week(theperiodofstability,asdeterminedbythesampleprovider).TheDutchparticipantsreceivedtheirsetsofsamplesatthestartoftheyearandkeptthemat?80°Cuntilanalysis.

Laboratorieswereaskedtoanalyze18analytesforwhichtargetvalueswereobtainedfromreferencelaboratoriesusinginternationallyrecognizedreferencemethodsandreferencematerials.Theycomprised:

5electrolytes:calcium,chloride,magnesium,potassium,sodium;6enzymes:ALT,amylase,AST,CK,Gamma-GT,LDH;2lipids:cholesterol,HDL-cholesterol;and

4substratesandaformula:creatinine,eGFR(F,55y,Caucasian),glucose,totalprotein,uricacid.

Thelaboratorieswereaskedtousetheirroutinemethodswithnoadaptationscomparedtoroutinepractice.Thelaboratoriesreportedtheirresults,methods,andtheinstrumentsused.ThelaboratoriesintheUKandSpainmostlyreportedSIunits,whilethelaboratoriesinPortugalmostlyreportedconventionalunits.Conventionalunitswere

convertedtoSIunitsbytheorganizerofthepilotstudy.Inafewcasesthereportedresultswerenotinagreementwiththereportedunitsandcorrectionsweremade.OneSpanishlaboratoryreportedcreatinineinSIunitsafterconvertingfromconventionalunits(mg/dL),butusingawrongconvertingfactor.ThismistakedidnotaffecttheeGFRresults,becauseaformulaforcreatininevaluesinmg/dLwasused.Resultsforcreatinineofthislaboratorywerediscarded.FourSpanishlaboratoriesusedapancreaticamylaseassayinsteadoftotalamylaseandtheirre-sultswerediscarded.2.1.Statisticalmethods

IntheCombischemereport,eachroundandforeachanalytetheindividuallaboratorydataisdisplayedasadifferenceplotofthesixresultscomparedwiththereferencemethodvalues(Fig.1).AtoleranceareaisconstructedaroundthereferencevaluesbasedontheTEA(desirable)limit[4,5].IntheCombischemethedesirableTEAlimit(TEA=1.65×0.5×CVw+0.25√(CVw2+CVb2))isusedratherthantheminimaloroptimallimitsasalternativeapproachessuggestedbyFraser[6].Linearregressioniscalculatedfromthelaboratoryresultsagainsttheconsensusmethodgroupmeanvalue.Asthesamples

实验室质控

are

Table1

TEA,averageTEscores,and%TEscores≥95%.Analyte

TEA%

CalciumChlorideMagnesiumPotassiumSodiumALT

AmylaseASTCK

Gamma-GTLDH

Cholesterol

HDL-cholesterolCreatinine

eGFR(F,55y,Caucasian)Glucose

TotalproteinUricacidOverall

2.41.54.85.80.914.626.315.230.322.211.48.511.18.98.97.23.412.4

NLAvTEscore(%)64646194479385949997849783766693589881

R.Jansenetal./ClinicaChimicaActaxxx(2013)xxx–xxx

NL%TEscN95%18162877584778296937687554147672896

PTAvTEscore(%)653957892680537683xxxxxxxxxxxx88539367

PT%TEscN95%001363063433863xxxxxxxxxxxx631363

ESAvTEscore(%)64816797428359889890639074336492779975

ES%TEscN95%2730228294540649191559060027733691

UKAvTEscore(%)7372799747xxxxxxxxxxxx882655796649977

UK%TEscN95%1030307020xxxxxxxxxxxx06020259030100

measuredondifferentdays,theresidualSDoftheregressionlinerepre-sentsthewithin-laboratorySDWL.Thedifferencebetweenthemeanofthesixlaboratoryresultsandtheaverageofthesixreferencevaluesisthebias.UsingSDWLattheaverageconcentrationofthesixsamplestheprobabilityisestimatedthatthelaboratoryresultswillbewithintheTEAtolerancearea.Thisprobabilityisthepercentageofthedensityfunction(thebroadnessofwhichisde?nedbySDWL)aroundthelaboratorybiasthatiswithintheTEAarea.TheTEscoreequalsthispercentage.Byde?nitionTEincludesbiasandimprecision.CausesoflowerTEscorescouldbesigni?cantpositiveornegativebias,oralargewithin-laboratorySD.Increasingly,minimalacceptableperformancecriteriabasedonthebiologicalTEAconceptarebeingutilizedwithinlaboratories.Thelevelofacceptanceisde?nedbyFraser[6]asminimal,desirableoroptimal.IntheSKMLscheme,performanceisconsideredtobeacceptableiftheresultsofalaboratoryarewithinthedesirableTEAareawithaprobabilityof95%.

ForeachanalytetheTEscoresoftheindividuallaboratoriesoftheUK,SpainandPortugalwereplottedagainstandcomparedwiththeaverageTEscoreoftheDutchlaboratories.Foreachanalyte,theindivid-uallaboratoryresultssortedbyinstrumentwerealsoplotted.AverageTE-scoresandthepercentageTEscoresN95%werecalculatedforthefourcountries.3.Results

Fig.1showsanexampleofadifferenceplotofthesixresultsofasinglelaboratoryforcreatinine.ThegreenarearepresentstheTEAareaaroundthereferencevalues.Theblueareaisthestateofthearttoleranceareaforthemethodgroupconsensusvalues.TheblueareainthiscaseshowsapositivedeviationfromthereferencevaluesinthelowerconcentrationrangeasisexpectedfortheJaffemethodgroup.TheplotshowsyourregressionversustheTEAtolerancearea,versusyourmethodgroupstateofthearttolerancearea,andthemethodgroupstateoftheart(bluearea)versustheTEAgreenarea.Thewithin-labCViscalculatedastheresidualCVoftheregressionline.TheTEscoreforthislaboratoryequals97%.IntheSKMLCombicon-ceptaTEscoreof95%isconsideredacceptable.Nexttothedifferenceplotatableisreportedtotheparticipantssummarizingtheresults.

Thereferencevaluesofthesixgeneralchemistryandthesixlipidsamplesfortheselected18analytes,arrangedbyanalytegroup(electrolytes,enzymes,lipids,substrates),thestandarduncertainties,thereferencemethodsandthereferencelaboratories,areprovidedassupplementarydata.Fig.2showsanexampleofTEscoresofindividuallaboratoriesforfouranalytessortedbyinstrument.Withtheexceptionofcholesterol,theotherthreeanalytes(AST,calciumandcreatinine)showinconsis-tencyofTEscoresevenwithinasingleanalyticalplatform.Thislackofperformanceconsistencybetweeninstrumentsandwithininstrumentwasseenforallelectrolytes,andenzymes(datanotshown),andisinagreementwithpreviouslyreportedresults[19].

Table1presentstheTEAvalues[5],averageTEscoresofthefourcountriesandthepercentageoflaboratoriesthathadaTEscore≥95%.TheNetherlands'TEscorewasthehighestat81%,followedbytheUK'sat77%,Spain'sat75%andPortugal'sat67%.

TEscoresforallelectrolytes,exceptpotassium,inallofthefourcountriesarelow(Table1,Fig.3).Urgentimprovementinharmoniza-tionisneededparticularlyforcalcium,chloride,magnesiumandsodiumwherelessthan30%ofthetrueTEscoreswereabovethe95%criterion.Thesameobservationwasmaintainedfortheenzymes.ThehighestTEscorewasseenforCKandGGT.However,awidevariationinTEscorewithinandbetweencountrieshasbeenrecordedforALT,ASTandamylase.Foramylase,laboratoriesshowtwotypesofTEscores,eitherTEabove95%oraverylowscoreoftenzero.CalibrationtoamethoddifferentfromthereferencemethodisthemaincauseoflowTEscores.FourSpanishlaboratoriesanalyzedpancreaticamylaseinsteadoftotalamylase.Theresultsoftheselaboratorieswereremovedastheyweretestingadifferentanalyte.

WiththeexceptionofTheNetherlands,Portugal,SpainandtheUKshowedpoorTEscoresforLDHwithmanyscoresofzeroobtainedand19outof28laboratorieshavingscoresbelow10%(Fig.4).ThisisduetothefactthatanumberoflaboratoriesareusingapyruvatetolactatemethodratherthantheIFCCreferencemethodutilizinglactateasasubstrate.Thesemethodsvarybyafactorof2andwillthereforehaveaprofoundeffectonbias,whichexplainsthepoorTEscorefortheselaboratories.

IngeneraltheTEscoresforenzymesinTheNetherlandsarehigher,andalargerpercentageofthelaboratoriesscoreabovethe95%limit,ascomparedtotheothercountries.

ForcholesteroltheaverageTEscoreswereabove90%andover85%ofthelaboratoriessatis?edthecriterionofTEscore≥95%(Table1andFig.5).However,theHDLmethodshavenotmatchedtheconsis-tentlyhighperformanceseenwithcholesterol.WhilethePortugueseachievedaTEscoreof100%foralltheparticipatinglaboratories,othercountriesdemonstratedawidervariationinperformance(Fig.5).

Forcreatinine(Table1andFig.6)lowaveragescoreswereobtainedaswellaslowpercentagesofTEscores≥95%,indicatingthatmany

R.Jansenetal./ClinicaChimicaActaxxx(2013)xxx–xxx5

1009080706050

Calcium

100908070

Mean TE scoreDutch labs

Chloride

6050

Mean TE scoreDutch labs

TE-score

40302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT080

40302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT080

Magnesium

1009080706050

TE-scoreMean TE scoreDutch labs

Potassium

1009080706050

TE-scoreMean TE scoreDutch labs

40302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT080

40302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT080

Sodium

10xxxxxxxxxxxx302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT080

TE-scoreMean TE scoreDutch labs

Fig.3.Electrolytes,TEscoresofindividuallaboratoriespercountry(region).ThebluelinerepresentstheaverageTEscoreforTheNetherlandslaboratories.

laboratoriesfailedtoachieveminimalacceptableperformance.Instru-mentsshowedwidelyvaryingresults.TheJaffémethodshadthelowestscore(datanotshown),whichisinagreementwithpreviouslyreportedresults[20].

ThishasaconsequenceforeGFR,whichwascalculatedusingdifferentformulaethe23participatinglaboratories.AverageTEscoreswerebelow70%andlessthanhalfofthelaboratoriesattainedaTEscoreof95%.GlucoseanduricacidmettheacceptableperformancecriterionofaTEscoreN95%forthemajorityofparticipatinglaboratoriesinthefourcountries.

ThedataforTotalProteinindicatedunsatisfactoryperformance,withaverageTEscoreswellbelow95%andmorethan70%ofthelaboratoriesfailingthe95%criterion.Fig.6illustratesthewidelyvaryingindividualscores.

6R.Jansenetal./ClinicaChimicaActaxxx(2013)xxx–xxx

ALT

1009080706050

TE scoreMean TE score Dutch labs

AST

TE scoreDutch labs

40302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT080

UK0UK0UK0UK0UK0UK0UK0UK0UK0UK1ES0ES0ES0ES0ES0ES0ES0ES0ES0ES0ES1PT0PT0PT0PT0PT0PT0PT0PT0Amylase

UK0UK0UK0UK0UK0UK0UK0UK0UK0UK1ES0ES0ES0ES0ES0ES0ES1PT0PT0PT0PT0PT0PT0PT0CK

10090

Dutch labs

80706050

TE-score

40302010

UK0UK0UK0UK0UK0UK0UK0UK0UK0UK1ES0ES0ES0ES0ES0ES0ES0ES0ES0ES0ES1PT0PT0PT0PT0PT0PT0PT0PT00

Gamma-GT

UK0UK0UK0UK0UK0UK0UK0UK0UK0UK1ES0ES0ES0ES0ES0ES0ES0ES0ES0ES0ES1PT0PT0PT0PT0PT0PT0PT0PT0TE-scoreDutch labs

LDH

1009080706050

TE-scoreMean TE score Dutch labs

40302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT080

Fig.4.EnzymeTEscoresofindividuallaboratoriespercountry(region).ThebluelinerepresentstheaverageTEscoreforTheNetherlandslaboratories.

4.Discussion

IntheEuropeanUnion,theIVDD98/79/EC[21]demandstraceabilityoftestresultstoahigherorderreferencematerial.Thismeansthattheresultsforeachinstrumenttypeshouldbecomparablewithreferencemethodresults.However,thispilotstudyshowsconsiderablewithininstrumentandbetweenlaboratoryvariationsinTEscores.AlthoughthenumberofparticipatinglaboratoriesfromoutsideTheNetherlandsissmall,theymaybeconsideredasrepresentativeofcountriesbecausetheyarepositionedwithinthe20thpercentileofthetargetdeviationdistributionintheirnationalEQAschemes(Spain,Portugal)orarerepresentativeforawholeregion(Yorkshire,UK).

R.Jansenetal./ClinicaChimicaActaxxx(2013)xxx–xxx7

Cholesterol

UK0UK0UK0UK0UK0UK0UK0UK0UK0UK1ES0ES0ES0ES0ES0ES0ES0ES0ES0ES1PT0PT0PT0PT0PT0PT0PT0PT0

实验室质控

HDL-Cholesterol

Dutch labs

Reeks2

UK0UK0UK0UK0UK0UK0UK0UK0UK0UK1ES0ES0ES0ES0ES0ES0ES0ES0ES0ES1PT0PT0PT0PT0PT0PT0PT0PT0Fig.5.LipidTEscoresofindividuallaboratoriespercountry(region).ThebluelinerepresentstheaverageTEscoreforTheNetherlandslaboratories.

Jansenetal.showedin2006[19]thatlargevariationbetweenmethodsandanalyticalplatformsexistfortheenzymesandthatinmanycasesthereisalackoftraceabilityandharmonizationdespitetheIVDrequirements.Thisstudyshowslittleimprovementwithen-zymeassays,especiallyforamylase[22]andLDH[22,23].LaboratoriesarestillusingmethodsthatdonotcomplywiththeIFCCrecommendedmethods(e.g.triosidesubstrateforamylaseorpyruvatesubstrateforLDH)andthisshouldbediscouraged.Furthermore,insomeenzymemethodse.g.ALT/AST,variationinTEscorehasbeenseenwithintheusersofthesameinstrument.Inourview,this?ndingmaybeattributedtotheuseofALT/ASTmethodslackingtheadditionoftheco-enzymepyridoxalphosphateinthereagentpack.Theco-enzymehasavariableandmarkedeffectontransaminaseactivity,especiallywithAST,whichcannotbecorrectedbycalibration.DespitetheIFCCrecommendations[24,25],manufacturersstillmarketmethodversionslackingpyridoxalphosphate.Methodsthatdonotcontaintheco-enzymecannotbecon-sideredtraceable.Anothersourceofdiscrepant(biased)resultshasbeenobservedforaSpanishlaboratory(datanotshown)forASTandgamma-GTwhenaroutinecalibratorwastraceabletoanon-commutablereferencematerial,whereasresultswerecorrectwhencor-rectlycalibratedandtraceabletoareferencemethod.Intheseexamples,themanufacturerscanplayapivotalroleinpavingtheroadtoharmoni-zation,simplybyremovingundesirablemethodsfromthemarket.

ThevariationintheTEscoresinUK,SpainandPortugalcannotbeexplainedbythedifferentanalyticalplatforms.Fig.2showsexamplesofTErangesfortheinstrumentsused.WithinthesameinstrumentTEscoresvarygreatly,insomecasesfrom0%to100%.Oneexplanationforthisistheproductionbymanufacturersofmorethanoneassayonthesameplatformforsomeanalytes,e.g.LDHlactatetopyruvateandpyruvatetolactate,andAST/ALTwithandwithoutpyridoxalphosphateP5P,whilsttraceabilitydemandstheIFCCrecommendations.Otherrea-sonscouldbebiasduetotheuseofdifferentfactors,differentcalibra-tors,andvaryingwithin-laboratorySD.Thebiascouldbeproportional,constantormixedi.e.varyingacrosstheconcentrationspan.Inspectionofthedatashowsthatinmanycasesalloftheseerrorsarepresent.E.g.forcreatininemanylaboratoriesusethenon-compensatedJaffékineticmethod,givingapositivebiasatlowconcentrationlevel.LaboratoriesneedtoshowacceptableprecisionaswellasbiastoattainaTEscoreof95%.LackofcommutabilityofthereferencematerialusedforroutinecalibratortraceabilityhasbeenseenasamajorreasonforbiasedresultsintheSpanishgroup.Thesamehappensformagnesiumandsodium.Ourdatashowsthaturgentimprovementinharmonizationisneededparticularlyforcalcium,chloride,magnesiumandsodiumwhereless

than30%oftheTEscoreswereabovethe95%criterion.Harmonizationofanalytesthathaveanarrowbiologicalvariabilitycanbeimprovedbysharingacommonbutclinicallyrelevantanalyticalgoal[26].Examplesofdifferentkindsoferrorsmadeareprovidedassupplementarydata.

Alltheanalytesinthisstudyhaveawell-de?nedreferencemeasure-mentprocedureandtraceabilitychain,yetconsiderableanalyticalvariationhasbeenseen.Thissuggeststhatstandardizationaloneisnotsuf?cienttoguaranteeproductionofcomparableresults.Traceabil-ityofamethodtohigherorderreferencemeasurementmethodsdoesnotnecessarilymeanthatthe?eldmethodresultsareidenticaltothereferencemethodresults.Itrequiresafunctionalrelationshipbetweenthemethodandthereferencemethodandreferencematerial.Fromapatient'sperspective,resultsfromdifferentlaboratoriesshouldnotonlybetraceabletothereferencemethod,i.e.showade?nedfunctionalrelationshiptothereferencemethod,butshouldinadditionbestan-dardized,i.e.giveequivalentresultstothereferencemethod.TheCali-bration2000project[9–11,15]andthepresentresultsshowthatharmonizationisachievableforsomeanalytesasshownintheCategory1EQAscheme.TheCombischemeinitspresentform,usingcommut-ablesamples,valueassignedwithreferencemethods,andhavingbiologicalvariation-basedtolerancelimits,hasbeenoperationalinTheNetherlandsforover7years.InanattempttoreplicateTheNetherlandsexperiencewithlargernumbersoflaboratories,thePortuguese,theSpanishandtheUKEQAschemeorganizersarecon-sideringcollaborationinatleastoneroundperyearintheSKMLCombischeme.

Since2005,theSpanishSocietyofClinicalChemistry(SEQC)haveundertakenaneducationaltaskinrecommendingtheuseofbiologicalvariationbasedastolerancelimitsandthesecriteriaareincludedintheparticipants'reports.Despitethis,thegroup'sresultsarenotassatisfactoryastheyshouldbe.Thisismainlyduetothelackofmethodharmonizationandtraceabilityandnottoadifferentcultureinqualitymonitoringpractices.5.Conclusion

TheIVDD98/79/ECdemandstraceabilityoftestresultstoareferencesystem,ifavailable.Ourdatashowthatthereisaneedforfurtherharmonizationoflaboratorydata,inparticularforelectrolytes(calcium,chloridemagnesium,sodium),enzymes(ALT,amylase,AST,LD),lipids(HDL-cholesterol),andforsubstrates(creatinine,totalprotein).Lackofperformanceconsistencybetweeninstrumentswasseenformostanalytes.ThevariationintheTEscorescannotbeexplainedbythe

实验室质控

8R.Jansenetal./ClinicaChimicaActaxxx(2013)xxx–xxx

1009080706050

Creatinine

10090

Dutch labs

80706050

Dutch labs

TE-score

40302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT08

实验室质控

403020100

UK01UK02UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT05

实验室质控

Dutch labs

TE-score

1009080706050

Glucose

1009080706050

Total Protein

Dutch labs

TE-score

403020100

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT08

实验室质控

1009080706050

Urate

Dutch labs

TE-score

40302010

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT08

实验室质控

Fig.6.SubstrateTEscoresofindividuallaboratoriespercountry(region).ThebluelinerepresentstheaverageTEscoreforTheNetherlandslaboratories.

differentanalyticalplatforms.WithinthesameinstrumentTEscoresvarygreatly,insomecasesfrom0%to100%.Lackofharmonizationisstillpresent,despitemanufacturers'claimsofestablishedtraceability.Currentdatashowsthatthestandardizationofmethodsisinsuf?cienttoresultincompleteconsistencyinreportingoflaboratoryresultsandneedstobefollowedbyharmonizationofthemethodsandpractices.

Declarations

Competinginterests:noneFunding:none

Ethicalapproval:notrequiredGuarantor:DrRobJansen

UK01UK02UK03UK04UK05UK06UK07UK08UK09UK10ES01ES02ES03ES04ES04ES05ES06ES07ES08ES09ES10PT01PT02PT03PT04PT05PT06PT07PT08

实验室质控

R.Jansenetal./ClinicaChimicaActaxxx(2013)xxx–xxx9

Acknowledgments

Thecontributionofthe30laboratoriesinPortugal,SpainandtheUKinperformingtheanalysisofthesixsamplesandinthesubmissionoftheresultsarekindlyacknowledged.AppendixA.Supplementarydata

[16][14]

[15]

Supplementarydatatothisarticlecanbefoundonlineat/10.1016/j.cca.2013.11.003.References

[1]MillerWG,MyersGL,GantzerML,KahnSE,Sch?nbrunnerER,ThienpontLM,etal.

Roadmapforharmonizationofclinicallaboratorymeasurementprocedures.ClinChem2011;57:1108–17.

[2]ThienpontLM,St?cklD,FriedeckyB,KratochvílaJ,BudinaM.Truenessveri?cationin

Europeanexternalqualityassessmentschemes:timetocareaboutthequalityofthesamples.ScandJClinLabInvest2003;63:195–201.

[3]MillerWG,MyersGL,RejR.Whycommutabilitymatters.ClinChem2006;52:553–4.[4]FraserCG.Generalstrategiestosetqualityspeci?cationsforreliabilityperformance

characteristics.ScandJClinLabInvest1999;59:487–90.

[5]MinchinelaJ,RicósC,PerichC,Fernández-CalleP,AlvarezV,DomenechM,etal.Bi-ologicalvariationdatabase,andqualityspeci?cationsforimprecision,biasandtotalerror(desirableandminimum).The2012update/biodatabase-2012-update.htm.[accessed08-07-2013].

[6]FraserCG,HyltoftPetersenP,LibeerJC,RicósC.Proposalsforsettinggenerally

applicablequalitygoalssolelybasedonbiology.AnnClinBiochem1997;34:8–12.[7]MillerGW,JonesGRD,HorowitzGL,WeykampC.Pro?ciencytesting/externalquality

assessment:currentchallengesandfuturedirections.ClinChem2011;57:1670–80.[8]AdamsO,CooperG,FraserC,HubmannM,JonesG,PlebaniM,etal.Collective

opinionpaperon?ndingsofthe2011convocationofexpertsonlaboratoryquality.ClinChemLabMed2012;50:1547–58.

[9]JansenRTP.Thequestforcomparability:Calibration2000.AccredQualAssur

2000;5:363–6.

[10]JansenRTP,KuypersAWHM,BaadenhuijsenH,BesselaarAMHPvanden,Cobbaert

CM,GratamaJW,etal.Kalibratie2000.NedTijdschrKlinChem2000;25:153–8.[11]BaadenhuijsenH,ScholtenR,WillemsHL,WeykampCW,JansenRTP.Amodelfor

harmonizationofroutineclinicalchemistryresultsbetweenclinicallaboratories.AnnClinBiochem2000;37:330–7.

[12]BaadenhuijsenH,SteigstraH,CobbaertC,KuypersA,WeykampC,JansenR.

Commutabilityassessmentofpotentialreferencematerialsusingamulticentersplit-patient-samplebetween-?eld-methods(twin-study)design:studywithintheframe-workoftheDutchproject“Calibration2000”.ClinChem2002;48:1520–5.

[13]CobbaertCM,WeykampC,BaadenhuijsenH,KuypersA,LindemansJ,JansenR.

Selection,preparation,andcharacterizationofcommutablefrozenhumanserumpoolsaspotentialsecondaryreferencematerialsforlipidandapolipoprotein

[17]

[18]

[19]

[20]

[21][22]

[23]

[24]

[25]

[26]

measurements:studywithintheframeworkoftheDutchProject“Calibration2000”.ClinChem2002;48:1526–38.

BaadenhuijsenH,KuypersA,WeykampC,CobbaertCM,JansenR.ExternalqualityassessmentinTheNetherlands:timetointroducecommutablesurveyspecimens.LessonsfromtheDutch“Calibration2000”project.ClinChemLabMed2005;43:304–7.

CobbaertCM,WeykampC,FranckP,deJongeR,KuypersA,SteigstraH,etal.Sys-tematicmonitoringofstandardizationandharmonizationstatuswithcommutableEQA-samples—?veyearexperiencefromTheNetherlands.ClinChimActa2012;414:234–40.

MüllerMM.Implementationofreferencesystemsinlaboratorymedicine.ClinChem2000;46:1907–9.

ClinicalandLaboratoryStandardsInstitute(CLSI).Characterizationandquali?cationofcommutablereferencematerialsforlaboratorymedicine;approvedguideline.CLSIC53-A940WestValleyRoad,Suite1400,Wayne,Pennsylvania19087-1898USA:ClinicalandLaboratoryStandardsInstitute1-56238-726-X;2010.

RicósC,RamónF,SalasA,BunoA,CalafellR,MoranchoJ,etal.Minimumanalyticalqualityspeci?cationsofinter-laboratorycomparisons:agreementamongSpanishEQAPorganizers.ClinChemLabMed2012;50(3):455–61.

JansenR,SchumannG,BaadenhuijsenH,FranckP,FranziniC,KruseR,etal.Truenessveri?cationandtraceabilityassessmentofresultsfromcommercialsystemsformeasurementofsixenzymeactivitiesinserum.AninternationalstudyintheEC4frameworkoftheCalibration2000project.ClinChimActa2006;368:160–7.

DelangheJR,CobbaertC,HarmoinenA,JansenR,LaitinenP,PanteghiniM.Focusingontheclinicalimpactofstandardizationofcreatininemeasurements:areportbytheEFCCWorkingGrouponCreatinineStandardization.ClinChemLabMed2011;49:977–82.

LexEU.Directive98/79EConinvitromedicaldevices.OffJL1998;331:1–37.

SchumannG,CanaliasF,JoergensenPJ,KangD,LessingerJM,KlaukeR.IFCCreferenceproceduresformeasurementofthecatalyticconcentrationsofenzymes:corrigendum,notesandusefuladvice.InternationalFederationofClinicalChemistryandLaboratoryMedicine(IFCC)—IFCCScienti?cDivision.ClinChemLabMed2010;48:615–21.

SchumannG,BonoraR,CeriottiF,Clerc-RenaudP,FerreroCA,FérardG,etal.IFCCprimaryreferenceproceduresforthemeasurementofcatalyticactivityconcentra-tionsofenzymesat37°C:Part3.Referenceprocedureforthemeasurementofcat-alyticconcentrationoflactatedehydrogenase.ClinChemLabMed2002;40:643–8.SchumannG,BonoraR,CeriottiF,FérardG,FerreroCA,FranckPF,etal.InternationalFederationofClinicalChemistryandLaboratoryMedicine.IFCCprimaryreferenceproceduresforthemeasurementofcatalyticactivityconcentrationsofenzymesat37°C.InternationalFederationofClinicalChemistryandLaboratoryMedicine:Part4.Referenceprocedureforthemeasurementofcatalyticconcentrationofalanineaminotransferase.ClinChemLabMed2002;40:718–24.

SchumannG,BonoraR,CeriottiF,FérardG,FerreroCA,FranckPF,etal.InternationalFederationofClinicalChemistryandLaboratoryMedicine.IFCCprimaryreferenceproceduresforthemeasurementofcatalyticactivityconcentrationsofenzymesat37°C.InternationalFederationofClinicalChemistryandLaboratoryMedicine:Part5.Referenceprocedureforthemeasurementofcatalyticconcentrationofaspartateaminotransferase.ClinChemLabMed2002;40:725–33.

JassamN,LindsayC,HarrisonK,ThompsonD,BosomworthMP,BarthJH.Theimple-mentationofasystemformanaginganalyticalqualityinnetworkedlaboratories.AnnClinBiochem2011;48:136–46.

相关推荐

实验室质控改善方案

第二篇：实验室质控

专栏推荐