lecture1_assay validation

《lecture1_assay validation》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、Lecture1AssayValidationOUTLINEIntroduction-Background-RegulatoryrequirementsCalibration-Standardcurve-ClassicalmethodAssayvalidation-Keyparameters-Statisticalinference-Validationdesign-AnexampleStatisticalconcerns-Intervalhypothesis-Weightselection-ModelselectionIntroductionBackgroundNewcompo

2、undItisnecessarilytodevelopassaymethodsortestproceduresfordeterminingtheactiveingredientofthecompoundincompliancewiththeUSP/NFstandardsfortheidentity,strength,quality,andpurityofthecompound.Oldcompound-Newtechnology-Timeand/orcostconsiderationIntroductionObjective-Toproducebestanalyticalresult

3、s-Accuracy-ReliabilityRegulatoryrequirements-GoodValidationPractice(Diding,1980)-GuidelineforsubmittingSamplesandAlyticalDataforMethodsValidation(FDA,1987)-GuidanceNoteonAnalyticalValidation(CommitteeforProprietaryMedicinalProducts,1989)-USP/NF(1990)-CGMPIntroductionTheCGMPrequiresthatanassaym

4、ethodforassessingcomplianceofpharmaceuticalproductswithestablishedspecificationsmustmeetproperstandardofaccuracyandreliabilityCalibrationAnumberofknownstandardconcentrationpreparationsInstrument:e.g.,HPLCResponses:e.g.,AbsorbanceorpeakresponseLinearRegressiony=a+bxTABLE2.3.2CalibrationDataDisp

5、layConcentration,XiAbsorbance,Yi17.65(L)97.48517.65(L)95.40622.06(M)121.20022.06(M)121.96826.47(H)142.34626.47(H)145.464S90.04426.47(H)141.83526.47(H)135.62522.06(M)113.81422.06(M)112.89017.65(L)89.87217.65(L)90.964Source:ChowandShao(1990b)CalibrationTheclassicalmethodModelLSestimatorsCalibrat

6、ionStandardcurvey=a+bxForagivenunknownsamplex0,itsconcentrationCanbeestimatedasCalibrationStandardcurveY=-2.363+5.429XCoefficientofdeterminationTABLE2.3.3ANOVATableforDatainTable2.3.2SourceofvariationDfSumofSquaresMeansquareFvaluepvalueRegression14586.0904586.090289.302<0.0001Residual10158.523

7、15.852Total114744.613AssayValidationSingle-systemvalidation-Todemonstratethemethodisaccurateandreliable-NoreferencemethodsinvolvedMultiple-systemvalidation-Tocomparethestandard,compendial,and/oralternativemethods-Tocomparetheoriginalcon