Integration in Chromatography - BEBAC • Consultancy Services for ...

Integration in ChromatographyChromatography•Substances (analytes & interferences) continuouslyexchange between Mobile andStationary Phases•Different Mechanisms in parallel (solubility, lipophilicity,ionization,…)•Retention influenced by type of Stationary Phase,column lenggth, composition of Mobile Phase (typeand % organic modifier, gradient, pH, buffer),temperature, pressure, flow rate,…Workshop | Bucarest, 19 March 20132• 32

Integration in ChromatographyRetention Time and Tailing3.94’ / 82164.24’ / 75365.30’ / 60675.90’ / 53026.50’ / 49558.46’ / 36763.96’ / 53464.26’ / 68095.34’ / 39155.94’ / 43906.56’ / 32128.50’ / 2779Workshop | Bucarest, 19 March 20138• 32

Integration in ChromatographyCharacteristics of PeaksGaussian PeakTangentsPeak Width at Tangentsprojected to Baseline: wInflection Points of GP:Width = 2σat 60.65% of Peak HeightPeak Width at 50%of Peak Height: w 0.5Peak Width at 10%of Peak Height: w 0.1Workshop | Bucarest, 19 March 20139• 32

Integration in ChromatographyCharacteristics of PeaksGaussian PeakInflection Point1 st DerivativePeak Time of GP atRoot of 1 st Derivative2 nd DerivativeInflection Points of GP atMax/Min of 1 st DerivativeInflection Points of GP atRoots of 2 nd DerivativeWorkshop | Bucarest, 19 March 201310 • 32

Integration in ChromatographyCharacteristics of PeaksEMG PeaksT f(0.31/0.18=1.722)T f(0.70/0.32=2.188)Retention time t R(6.58’)Retention time t R(8.46’)Capacity factor k’ (t R-t 0)/t 0=3.387Capacity factor k’ (t R-t 0)/t 0=4.640Tailing factor at 50% heightT f=b/a (0.21/0.19=1.105)w 0.5=0.40’w 0.5=0.49’Elution time ofvoid t 0(1.50’)w 0.1=1.05’IUPAC at 10% heightT f=b/a (0.72/0.33=2.182)w 0.1=1.02’w=~1.699×w 0.5=0.68w=0.830 1 2 3 4 5 6 7 8 9Workshop | Bucarest, 19 March 201311 • 32

Integration in ChromatographyCharacteristics of Peaks1 st Derivative (slope)0 1 2 3 4 5 6 7 8 9 12 nd Derivative0 1 2 3 4 5 6 7 8 9 1Workshop | Bucarest, 19 March 201312 • 32

Integration in ChromatographyRecommendations•Capacity factor k’ for analytes >2•Example:(6.58–1.50)/1.50=3.39 (8.46–1.50)/1.50=4.64 •Resolution between two adjacent peaks•R s = 2 × (t R2–t R1) / (w 1 + w 2 )Baseline width w not accessible; for a Gaussian [sic]peak w ~ 1.699 × w 0.5 holds.•Desirable >2•Example: 2×(8.46–6.58)/(0.68+0.83)=5.69 Workshop | Bucarest, 19 March 201313 • 32

Integration in ChromatographyRecommendations•Run times•The longer, the better the separation – but•Peak heights will decrease(band broadening → worse LLOQ)•Run times are decreased by• Type of stationary phase C18 → C8• ↓ Column length• ↑ Particle size 3 µm → 5 µm• ↑ Flow rate• Type of organic modifier in mobile phase CH 3 OH → CH 3 CN• ↑ % of organic modifier in MP• ↑ TemperatureWorkshop | Bucarest, 19 March 201315 • 32

Integration in ChromatographyHurry up!k’ 3.39T f2.18R s5.69k’ 4.64T f2.190 1 2 3 4 5 6 7 8 9k’ >2, R s >2, T f

Integration in ChromatographyIntegration•Peak ‘recognition’•Automatic vs. manual•Chromatography Data System (CDS)Workshop | Bucarest, 19 March 201317 • 32

Integration in ChromatographyIntegration•Peak ‘recognition’•Peak start and end triggered by:•Upward-/downward slope detection:For a Gaussian peak upward- / downward thresholdsare the same, but in chromatography peaks arealways asymmetrical.Some data systems correct for that by using moreslices if the slope is negative or even change to adifferent fitting algorithm.Workshop | Bucarest, 19 March 201320 • 32

Integration in Chromatography•Peak ‘recognition’IntegrationBaselineChromatogramSlope ThresholdSlopeFour consecutive values abovethreshold: integration of peaktriggered at data point –3.5.50 5.75 6.00 6.25 6.50Workshop | Bucarest, 19 March 201321 • 32

Integration in ChromatographyIntegration•Automatic vs. manual•Integration parameters are saved in the CDS’method and work in the background•The automatic integration may fail:•Mainly for small peaks close to the LLOQ•But also (rarely) for high peaks,if a series of positive random noise triggers an ‘end ofpeak’ too early ornegative random noise draws the baseline too late.•There is no ‘correct’ integration for any given peak!Identical raw data most likely will result in differentvalues if evaluated by another CDS.Workshop | Bucarest, 19 March 201322 • 32

Integration in ChromatographyIntegration•Automatic vs. manual•All chromatograms should be reviewed and theintegration corrected if necessary•The review has to be done before concentrationsare calculated.Changing integration of a peak in order to force acalibrator / QC towards the expected value (e.g.,make a batch valid which would be rejectedotherwise) or a pre-dose concentration

Integration in ChromatographyIntegration•Automatic vs. manual•Review and manual correction acceptableaccording to current GLs (FDA 2001, EMA 2011)•SOP in place•Consistently across the study’s chromatograms•Report which chromatograms were reintegrated(why, by whom, when – all the usual dataneeded for an audit trail).Workshop | Bucarest, 19 March 201325 • 32

Integration in ChromatographyIntegration•Automatic vs. manual•Example: LC/MS-MS, risperidone, protein precipitation,dilution factor 8, API 4000, software Analyst1.4.1; 1 ng/mL and 0.1 ng/mL (at LLOQ)Integration methodautomated (smoothing 1, bunching 2)1 ng/mL 0.1 ng/mLCV (n=10)6.5%15.1%manual correction (one analyst)manual correction (ten analysts)6.3%5.2%(3.8% – 6.8%)11.1%12.8%(6.9% – 16.0%)H Kirchherr, Data Evaluation in LC-MSIn: H-J Kuss and S Kromidas (eds.), Quantification in LC and GC, Wiley, p243-259 (2009)Workshop | Bucarest, 19 March 201326 • 32

Integration in Chromatography•Automatic vs. manualIntegration•Some analyst are afraid of getting problems in aninspection – believing automatic integration is the‘gold standard’ and manual integration some kind ofdata manipulation.•Example:Fairly recent (06/2010) BE study, active l-enantiomervs. racemate, LC/MS-MS; chromatograms of• high calibration standard• low QC sampleWorkshop | Bucarest, 19 March 201327 • 32

Integration in ChromatographyIntegration•Automatic vs. manual•It would be possible to calculate peak areas by deconvolution.Not available in current CDS!Only supportedby Merck / Hitachi’smid-90sD-7000 HPLCSystem Manager(HSM v4.1) orexternal software(PeakFITfrom Systat).4.0e+53.5e+53.0e+52.5e+52.0e+51.5e+51.0e+55.0e+4Peak 1 (EMG fitted)Peak 2 (EMG fitted)Chromatogram (measured)Chromatogram (EMG fitted)0.0e+05.5 6 6.5 7 7.5 8 8.5 9 9.5 10 10.5Workshop | Bucarest, 19 March 201329 • 32

Integration in ChromatographyIntegration•Chromatography Data System (CDS)•Bundled with chromatograph / MS• Xcalibur ® (Thermo Scientific)• Analyst ® (Applied Biosystems/MDS Sciex)• EZChrome Elite (Agilent Technologies)• Empower (Waters)• Chromeleon ® (Dionex)• LabSolutions (Shimadzu)•Commercial, vendor independent• PowerChrom ® (eDAQ)•Cross-platform freeware• ezDataPowerChrom ® (chemilab.net)•Deconvolution• PeakFIT ® (Systat)Workshop | Bucarest, 19 March 201330 • 32

Integration in ChromatographyIntegration•Chromatography Data System (CDS)•Important points• Audit Trail?• Data transfer to LIMS?• Data format: Preferable not only the integrationparameters, but the raw peak slices are stored.• ANDI/netCDF (AIA) Chromatography Data InterchangeFormat (ASTM standard E1947-98)• Last resort: CSV (Character Separated Variables)• FDA 21 CFR Part 11 compliant (rarely; ask!)• If possible data should not be stored only at the instrument’sPC, but copied to a central location for securedbackup.• Provide the sponsor a DVD with raw data files.Workshop | Bucarest, 19 March 201331 • 32

Integration in ChromatographyThank You!Integration inChromatographyOpen Questions?Helmut SchützBEBACConsultancy Services forBioequivalence and Bioavailability Studies1070 Vienna, Austriahelmut.schuetz@bebac.atWorkshop | Bucarest, 19 March 201332 • 32