Report of the American Diabetes Association.docx

Report of the American Diabetes Association.docx

《Report of the American Diabetes Association.docx》由会员分享，可在线阅读，更多相关《Report of the American Diabetes Association.docx（71页珍藏版）》请在冰豆网上搜索。

ReportoftheAmericanDiabetesAssociation

ReportoftheAmericanDiabetesAssociation'sTaskForceonStandardizationoftheInsulinAssay

Robbins,DavidC.（chair）;Andersen,Lennart;Bowsher,Ron;Chance,Ron;Dinesen,Bo;Frank,Bruce;Gingerich,Ron;Goldstein,David;Widemeyer,Hsaio-Mei;Haffner,Steven;Hales,C.Nick;Jarett,Leonard;Polonsky,Kenneth;Porte,Daniel;Skyler,Jay;Webb,George;Gallagher,Kathy

AuthorInformation

FromtheNovoNordiskLaboratories（L.Andersen）,NovoNordiskA/S,Bagsvaerd.

andtheStenoDiabetesCenter（B.Dinesen）,Gentofte,Denmark

TheLillyResearchLaboratories（R.Bowsher,R.Chance,B.Frank）,EliLilly,Indianapolis,Indiana

TheDepartmentofPediatrics（D.Goldstein,H.-M.Widemeyer）,UniversityofMissouri,Columbia

andLincoResearch（R.Gingerich）,St.Louis,Missouri

TheUniversityofTexasHealthSciencesCenter（S.Haffner）,SanAntonio,Texas

CambridgeUniversity（C.N.Hales）,Cambridge,U.K.

TheDepartmentofPathology（L.Jarett）,UniversityofPennsylvania,Philadelphia,Pennsylvania

TheDepartmentofMedicine（K.Polonsky）,UniversityofChicagoHospitalandClinics,Chicago,Illinois

TheDepartmentofMedicine（D.Porte）,UniversityofWashington,Seattle,Washington

TheDepartmentofMedicine（J.Skyler）,UniversityofMiami,Miami,Florida

andthePennMedicalLaboratory（D.C.Robbins,G.Webb,K.Gallagher）,MedlanticResearchInstitute,Washington,DC.

AddresscorrespondenceandreprintrequeststoDr.DavidC.Robbins,MedlanticResearchInstitute,108IrvingSt.NW,Washington,DC20010.

BacktoTop

Abstract

Recentlarge-scaleepidemiologicalstudiesdemonstratethatbloodconcentrationsofimmunoreactiveinsulinpredictthedevelopmentofNIDDMandIDDMandareassociatedwiththeriskofseveraldegenerativediseases,suchascoronaryandperipheralvesselatherosclerosis,hypertension,anddyslipidemia.Thereliabilityofthesemeasurementsisdependentonabiologicalassaythathasnotbeenwellstandardizedbetweenlaboratories.Recognizingthis,theAmericanDiabetesAssociationorganizedataskforcetoassesscomparabilityofbloodinsulinmeasurementsbetweenlaboratoriesandtosuggesttechniquestoimprovecomparability.Thetaskforcefoundthatidenticalserumandplasmasamplesmeasuredindifferentlaboratoriesproducedwidelydisparatevaluesthatwereunacceptableforpopulationcomparisons.Useofasinglereferencestandarddidlittletoimprovecomparability.Assaycharacteristicssuchaslinearity,recovery,accuracy,andcross-reactivitytoproinsulinanditsprimaryconversionintermediatesvariedamongthelaboratories,andtheydidnotreadilyexplaindifferencesinthemeasurementsmadefromassaytoassay.Useofthesameassaykitindifferentlaboratoriesdidnotalwaysensurecomparablemeasurements.Linearregressionofassayresultsfromonelaboratorytoanarbitrarilychosenreferenceassaygreatlyimprovedcomparabilityanddemonstratedthepotentialvalueincomparingeachassaytoareferencemethod.Thetaskforcereportdefinesacceptableassaycharacteristicsandproposesathree-stepprocessofinsulinassayproficiencyandcomparability.Acentralreferenceassayandongoingsampleexchangewillbeneededtoallowreliablecomparisonsofinsulinmeasurementsmadeindifferentlaboratories.Rigorousqualitycontrolandcontinuousqualityimprovementareneededtomaintainreliabilityoftheinsulinmeasurement.Diabetes45:

242-256,1996

ADA,AmericanDiabetesAssociation;CV,coefficientofvariation;ELISA,enzyme-linkedimmunosorbentassay;IRMA,immunoradiometricassay;LOD,limitofdetection;LOQ,limitofquantitation;RIA,radioimmunoassay;WHO,WorldHealthOrganization.

Theinsulinassayhasbeenusedforover30yearstoestimatetheconcentrationofinsulin-relatedmoleculesinplasmaandotherbiologicalfluids[1].Initially,thedevelopmentoftheassayfocusedontheeaseofperformanceandthereproducibilityofmethodinasinglelaboratory.Becauseofimmunochemicaldifferencesbetweenbovine,porcine,andhumaninsulin,themixedbovine-porcine-bioassayedpreparationsusedasbiologicalstandardsthatwereavailableatthetimewerenotsuitableforstandardizingthehumaninsulintobeassayed.Therefore,whileimpurehumaninsulinwasavailableforuseasanin-housestandard,eachlaboratorywasresponsibleforitsowncalibrationandqualitycontrolandnointerlaboratorycomparisonwaspossible.Sincetheapplicationoftheassaywaslargelyrelatedtothestudyofthepathophysiologyofdiabetesandrelateddisordersinsmallnumbersofsubjectsortotheunderstandingofthephysiologicalregulationofinsulinsecretioninanimalsorinvitro,therewasperceivedtobealimitedneedforstandardizationbetweenlaboratories.Nevertheless,sincetherewasoftengreatvarianceinreportedvaluesfromdifferentlaboratoriesaroundtheworld,attemptsweremadetoprovideforinterlaboratorycomparisonsusingthevariousmethodologiesavailablefromtimetotime[2-4].Whilelimited,thesestudiesdemonstratedaratherlargevariation.Inpart,thismayhavebeenduetothelackofacommonstandardofdefinedcontent.Recognizingthisneed,theWorldHealthOrganization（WHO）providedaninternationallyrecognizedpreparationforeachspeciesofpurifiedinsulin--bovine,porcine,andhuman--thatcouldbeusedforanexternalstandard[5].TheavailabilityoflargequantitiesofpurehumaninsulinthroughrecombinantDNAtechnologyintheensuingyearsfurtherenhancedthepossibilityofsuchstandardization.Furthermore,thistechnologymeansthatapurechemicalfortheassayisnowavailablethatdoesnotdependonbioassay.Thepotencyoftherecombinantpurenativeinsulinmaterialswas6.0nmol/Uin1991[6]（thoughasrecentlyas1993itwasnecessaryforcommentstobepublishedtoclarifythebiopotencyofpurehumaninsulin[7]）.Thistranslatesto28.7U/mgpureinsulin.

Nevertheless,theuseoftheassayinclinicalmedicinehaslargelybeenconfinedtothediagnosisofinsulinoma.Therefore,whileitmightbepossibleandevendesirabletostandardizetheassay,therehasnotbeenaconcertedeffortbyanyorganizationtodososeriously.Large-scaleepidemiologicalstudiesdemonstratethatbloodconcentrationsofimmunoreactiveinsulin（thetermimmunoreactiveinsulinisusedinreferencetoassaysthatmayrecognizeothersubstances,inadditiontoinsulin,thatshareantigenicepitopeswithinsulin,suchasproinsulin,proinsulinconversionintermediates,andinsulinderivativesproducedbydeterioration,glycosylation,dimerization,degradation,etc.）predictthedevelopmentofNIDDM[8-10],andthereistheincreasingpossibilitythatimmunoreactiveinsulinlevelsmaybeusedintheclinicalassessmentofpatientstopredictdiabetessusceptibility[9,11,12].

However,thereliabilityoftheseclinicalandepidemiologicalconclusionsisdependentonthecharacteristicsoftheassays.Itisreported,forexample,thatproinsulinisdisproportionatelyelevatedinNIDDM[13,14]andtheabsolutelevelof"true"or"specific"insulinmayactuallybelowinNIDDM[15].Thus,aninsulinassaythatcross-reactswithproinsulinanditsdegradationproductsmaycontributetodifferentconclusionsthanonethatisinsulinspecific.Thecirculatingamountsofproinsulinmayalsobeelevatedamongindividualswithimpairedglucosetolerance[16].Furthermore,atleasttworeportsfindthatelevatedproinsulin（butnotspecificinsulin）predictsthedevelopmentofNIDDM[17,18].Hyperinsulinemiaand/orinsulinresistanceisalsoassociatedwithcoronaryheartdiseaseandcardiovascularriskfactors[19,20].Withsomeimmunoassays,proinsulinismorecloselyassociatedwithriskfactorsthanisinsulininbothdiabetic[21]andnondiabeticsubjects[22].

Inordertoresolvethisproblem,theAmericanDiabetesAssociation（ADA）appointedataskforcetoexplorethepotentialofstandardizingtheinsulinassay.Theworkofthetaskforceproceededinthreephases:

anassessmentofthestateoftheart,thedevelopmentofastandardizedprotocol,andanassessmentofitsapplicabilityinasmallgroupofexperiencedlaboratories.Fromthisexperience,aprotocolforstandardizationhasbeendeveloped.

BacktoTop

RESULTS

Participatinglaboratoriesandtheassayusedinthestandardizationprogram.LaboratoriesinvitedtoparticipateintheADAinsulinassaystandardizationprogramwerechosenonthebasisoftheirinterestinlaboratorymedicineappliedtodiabetesorbecauseoftheirparticipationinstudiesinwhichinsulinhasbeenacentraloutcomemeasure.ThetypesofassaysusedaresummarizedinTable1.Noteespeciallythatthreeoftheassayswereenzyme-linkedimmunosorbentassays（ELISAs）（laboratoriesC,E,andG）.AllthreeELISAsandtworadioimmunoassays（RIAs）werereportedtobeinsulinspecific.Oneinsulinassay（laboratoryD）wasknowntocross-reactnearlyequallywithproinsulin,proinsulinconversionintermediates,andinsulin.Theremainderoftheinsulinassayswereknowntocross-reactwithproinsulinanditsconversionintermediatestovariousdegrees.ThispointisillustratedinTable4andisaddressedmorefullybelow.

Table1

Table4

Unknownsampleexchangetodetermineinterlaboratoryvariance.Comparabilitybetweenlaboratorieswasunknown.Tomeasurevariabilitybetweenlaboratoriesandprecisionwithinlaboratories,thetaskforceobtainedbloodfrom16volunteers.Ofthese,14werehealthydonorswhotooknoroutinemedicationsanddidnothavediabetes.Onedonor（identifiedasdonor11inFigure1）hadatypeIVhyperlipidemiaandobesity,anddonor4hadIDDMandlowtitersofanti-insulinantibodies.Bloodwaswithdrawninasingledona