Validation of Chromatographic Methods-FDA

8/4/2019 Validation of Chromatographic Methods-FDA

1/33

Center for Drug Evaluationand Research (CDER)

Reviewer Guidance'Validation ofChromatographic Methods

November 1994CMC 3


2/33

TABLE OF CONTENTS

I INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1II TYPES OF CHROMATOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

A . High Performance Liquid Chromatography (HPLC ) . . . . . . . . . . . . . . . 21 Chiral Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Ion-exchange Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . 33 . Ion-pair1Affinity Chrom atography . . . . . . . . . . . . . . . . . . . . . . . . 34 . Normal Phase Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . 35 . Reversed Phase Chromatography . . . . . . . . . . . . . . . . . . . . . . . 36 Size Exclusion Chromatography . . . . . . . . . . . . . . . . . . . . . . . . 4

B Gas Chromatography (GC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4C . Thin-Layer Chromatography (TLC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Ill REFERENCE STANDARDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5IV PARAMETE RS FOR VALIDATION OF HPL CHROM ATOGRAPH ICMETHODS FOR DRUG SUBSTANCE AND DRUG PRODUCT . . . . . . . . . . 7

A . Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8B Detection Limit and Quan titation Limit . . . . . . . . . . . . . . . . . . . . . . . . . 8C . Linearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11D Precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1 Repeatability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13. . . . . . . . . . . . . . . . . . . . . . . . . . . .Injection Repeatability 13

b Analysis Repe atability . . . . . . . . . . . . . . . . . . . . . . . . . . . 15


3/33

2 . Intermediate Precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153. Reproducibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

E Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16F Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16G . Robustness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16H Samp le Solution Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17I Specificitylselectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17J System Suitability Spec ifications and Tests . . . . . . . . . . . . . . . . . . . . . 21

1 Capacity factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .222 . Precisionllnjection repea tability . . . . . . . . . . . . . . . . . . . . . . . . . 223 . Relative retention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 . Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225. Tailing factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.36 . Theoretical plate num ber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

K General Points to Consider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28V . COMMENTS AND CO NCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29VI . ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29VII . REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29


4/33

REVIEWER GUIDANCE'VALIDATION OF CHROMATOGRAPHIC METHODS

1. INTRODUCTION

The purpose of this technical review g uide is to presen t the issues to cons ider whenevaluating chromatog raphic test methods from a regulatory perspective. Thedocument discusses the points to note and weaknesses of chromatography so thatCDER reviewers can ensure that the m ethod's performance claims are properlyevaluated, and that sufficient information is available for the field chem ist to assess themethod. Ana lytical terms, as defined by the International Con ference ofHarmonization (ICH), 1993, have been incorporated in this guide.Chromatographic methods are commo nly used for the quantitative and qualitativeanalysis of raw materials, drug substances, drug products and com pounds in biologicalfluids. The components monitored include chiral or achiral drug, process impurities,residual solvents, excipients such as preservatives, degradation products, extractablesand leachables from container and closure or man ufacturing process, pesticide in drugproduct from plant origin, and m etabolites.The objective of a test method is to generate reliable and accurate data regardless ofwhe ther it is for acceptance, release, stability or pharmacokinetics study. Data aregenera ted for the qualitative and quantitative testing during developm ent and post-approval of the drug products. The testing includes the acceptance of raw materials,release of the drug substances an d produ cts, in-process testing for quality assurance ,and estab lishment of the exp iration dating period.Validation of a m ethod is the process by which a method is tested by the developer oruser for reliability, accuracy and preciseness of its intended purpose. Data thus

'This guidance has been prepared by the Analytical Methods TechnicalCom mittee of the Ch emistry Manufacturing Controls Coordinating Com mittee (CMCCC) of the C enter for D rug Evaluation and R esearch at the Food and DrugAdm inistration. Although this guidance does not create or confer any rights for or onany person and doe s n ot operate to bind FDA or the industry, it does represent theagency's current thinking on the validation of chromatographic metho ds. For additionalcopies of this guidance, contact the Division of Comm unications Manag ement, HFD-210, CDE R, FD A, 5600 Fishers Lane, Rockville, MD 2085 7 (Phone: 301-594-1012).Send one se lf-addressed adhesive label to assist the offices in processing yourrequest. An electro r~ic ersion of this guidance is also ava ~lable ia Internet the WorldWide Web (WWW ) ( connect to the FDA Home P age at WW W.FDA.GOV/CDER andgo to the "Regulatory Guidance" section).


5/33

generated become part of the methods validation package submitted to C DER .Metho ds validation should not be a one-time situation to fulfil Agen cy filingrequirements, but the methods should be validated and also designed by the developeror user to ensure ruggedness or robustness. Methods should be reproducible whenused by other ana lysts, on other equivalent equipmen t, on other da ys or locations, andthroughout the life of the drug product. Data that are generated for acceptance ,release, stability, or pharm acokinetics will only be trustwo rthy if the m ethods u sed togene rate the data are reliable. The process of validation and method design alsoshould be early in the de velopment cycle before important data are ge nerated.Validation should be on-going in the form of re-validation with m ethod cha nges.II. TYPES OF CHROM ATOGRAPHYChromatography is a technique by which the components in a sample, carried by theliquid or gaseous pha se, are resolved by sorption-desorp tion steps on the stationaryphase.

A. High Performance Liquid Chrom atography (HPLC)HPL chromatograph ic separation is based on interaction and differentialpartition of the sample be tween the mobile liquid phase and the stationaryphase. The comm only used chromatographic methods can be roughlydivided into the following groups , not necessa rily in order of importance:

1. Chiral2. Ion--exchange3. Ion--pair/affinity4. Normal phase5. Reversed phase6. Size exclusion

1. Chiral ChromatographySeparation of the enantiomers can be achieved o n chiral stationaryphases by forma tion of diastereom ers via derivatizing agents ormobile phase additives on achiral stationary phases. When usedas an im purity test m ethod, the sensitivity is enhanced if theenantiom eric impurity elutes before the enan tiomeric drug.


6/33

2. Ion-exchange ChromatographySeparation is based on the charge-bearing functional groups,anion exchange for sample negative ion ( X I, or cation exchangefor sample positive ion (X'). Gradient elution by pH is comm on.

3. Ion-pair1Affinity ChromatographySeparation is based on a chemical interaction specific to the targetspecies. The m ore popular reversed phase mode uses a bufferand an added counter-ion of opposite charge to the sample withseparation being influenced by pH, ionic strength, temperature,concen tration of and type of organic co-solvent(s). Affinitychromatography, commo n for macromolecules, employs a ligand(biologically active molecule bonded cova lently to the solid matrix)which interacts with its homologous antigen (analyte) as areversible complex that can be eluted by changing bufferconditions.

4. Normal Phase ChromatographyNormal phase chromatography is a chromatographic techniquethat uses organic solvents for the m obile phase and a polarstationary phase. Here, the less polar comp onents elute fasterthan the more polar components.

5. Reversed Phase ChromatographyThe test method most comm or~ly ubmitted to CDER is thereversed phase HPLC m ethod. UV detection is the most commondetection technique.Reversed phase chromatography, a b onded phasechromatographic technique, uses water as the base solvent.Separation based on solvent strength and selectivity also may beaffected by column temperature and pH. In general, the m orepolar components elute faster than the less polar components.UV detection can be used with all chromatographic techniques.The co ncern for this type o f detector is the loss of sensitivity withlamp aging, and varying sens itivity at the low level depending ondesign and/or manu facturer. A point to note is that observationson the HPL chromatograms, by UV detection in combination withreversed-phase HPLC , may n ot be a true indication of the facts for


7/33

the following reasons:Com pounds much m ore polar than the compound of interestmay be m asked (elute together) in the solvent FronVvoidvolume.Com pounds very less polar than the analyte may eluteeither late during the chromatograph ic run or are retained inthe column.Com pounds with lower UV extinction coefficients or differentwavelength maxima m ay not be detectable at the low levelrelative to the v isibility of the an alyte since only onewavelength is no rmally monitored.

6. Size E xclusion ChromatographyAlso know n as gel permeation or filtration, separation is based onthe mo lecular size or hydrodynamic volume of the com ponents.Molecu les that are too large for the pores of the po rous packingmaterial on the colum n elute first, small molecules that en ter thepores elute last, and the e lution rates of the res t depend on theirrelative sizes.

6. Gas C hromatography (GC)Gas chromatography is based on the volatilized sample transported bythe carrier gas as the mo ving phase through the stationary phase of thecolumn where separation takes place by the sorption/desorption process.Samp les for gas chromatographic analysis are normally low m olecularweight com pounds that are volatile and stable at high temperature. Inthis respect, residual solvents in drug substances and drug products aresuitable for gas chroma tographic analysis. Chemical derivatives can alsobe formed to achieve volatility and therm al stability.Com mon detectors are flame ionization (FID) for carbon -containingcompounds, electron capture (ECD ) for halogenated compounds, flamephotom etric (FPD ) for corrlpounds containing sulphur or phosp horousand nitrogen-phosphorous (NPD) for compounds containing nitrogen orphosphorous. Chiral separation also can be achieved by gaschromatography. Separation by the packed column is rapidly beingreplaced by the capillary colurrln that p rovides improved resolution andanalysis speed. The location of the analyte on the gas chromatogram is


8/33

described by retention time (R,) which is similar to HP LC.C . 'Thin-Layer Chrom atograp hy ('TLC)

Thin-layer chromatography is the simplest of the m ore comm onchromatographic techniques. Separation is based on migration of thesample spotted on a coated (stationary phase) plate with one edgedipped in a m ixture of solvents (mobile phase). The w hole system iscontained in an enclosed tank.Detection techniques include fluorescence, U V and sprays (universal andspecific) for compounds that are not naturally colored. The location of theanalyte on the TLC plate is described by the R, value wh ich is the ra tio ofthe m igration distance of the com pound of interest to the mobile phasefront.

Of the three techniques, gas, liquid and thin-layer, TLC is the m ost universal testmethod as all components are present on the plate and with appropriate detectiontechniques, all components can be observed. However, it normally is not as accurateor sensitive as HPLC . TLC has a higher analytical variation than HP LC, although onesees the "whole picture" when appropriate detection schemes are selected.Ill. REFERENCE STANDARDSA re ference standard is a highly purified compound that is well characterized.Chromatographic methods rely heavily on a reference standard to provide accuratedata. Therefore the quality and purity of the reference standard is very importan t. Twotypes of reference standards, chem ical and nuclidic, exist. W ith the latter, the rad io-label purity should also be considered as well as the chemical purity.As described in the Guideline for S ubmitting Samples and Analytical D ata for MethodsValidation, the two ca tegories of chemical reference standards are as follows:

USP INF reference standard that does not need characterization, andnon-compendia1 standard that should b e of the highest purity that can beobtained by reasonab le effort and shou ld be thoroughly characterized toassure its identity, strength, quality and pu rity.

The points to note are:Most U SPINF reference standards do no t state the purity of thecompound.


9/33

The purity correction factor for non-US P reference standards isrecommended to be included in the calculation o f the test method.In addition to structurally-related impurities from the synthesis process,other process impurities like heavy metals, residual solvents, moisture(bound and unbound), pesticides for p roducts of plant origin, anddegrada tion products can also contribute to the lack of purity in thereference standard.The drying of the reference standard before use , if stated in the method,will eliminate residual solvent(s), unbound mo isture and sometimesbound mo isture (depending on the drying conditions). The drying step isalways included for hygroscopic compounds. On the other hand , dryingcan resu lt in the loss of a hydrate or cause degradation in heat-sens itivecompounds.

Chrom atographic test methods use either external or internal standards forquantitation.A. An external standard method is used when the standard is analyzed on aseparate chromatogram from the sample. Quantitation is based on acomparison of the peak arealheight (HPLC or GC) or spot intensity (TLC)of the sample to that of a reference standard of the analyte of interest.

The external standard method is more approp riate for samples as follows:1. Sample with a single target concentration and narrowconcen tration range, e.g., acceptance and release tests.2. Simple sam ple preparation procedure.3. Increased base line time for detection of potential extraneouspeaks, e.g., impurities test.

B. With an internal standard m ethod, compound of kn own purity that doesnot cause interference in the analysis is added to the samp le mixture.Quantitation is based on the response ratio of comp ound of interest to theinternal standard vs the response ratio of a sim ilar preparation of thereference standard (HPLC or GC). This technique is rarely used for TLCmethods.The internal standard method is m ore appropriate for samples as follows:


10/33

1. Com plex sample prepa ration procedu res, e.g., multipleextractions.2. Low concentration samp le (sensitivity being an issue), e.g.,pharmacokinetics studies.3. Wide range of concentrations expected in the sample for analysis,e.g., pharmacokinetics studies.

Although CDER does not specify whether the m ethod must use an internal orexternal standard for quantitation, it is commonly observed that HPLC methodsfor release and stability and TLC methods use external standards; and methodsfor biological fluids and GC methods use internal standards .The workina concentration is the target concentration of the compound o f interest asdescribed in the m ethod. Keeping the concentrations of the sample and the standardclose to each other for the external standa rd method iniproves tl ie accuracy of themethod.Recommendations:

1. Include the purity correction factor, if knowti, of the reference standard inthe calculation.2. State .the working concen trations of the standard and sa mple in themethod.

IV. PARAMETERS FOR VALIDATION OF HPL CHROMATOGRAPHICMETHODS FOR DRUG SUBSTANCE AND DRUG PRODUCTThough many types of H PL chromatographic techniques are available; the m ostcomm only submitted method, the reversed-phase HP LC with UV detection, is selectedto illustrate the pa rame ters for validation. The criteria for the validation of thistechnique can be extrapolated to other detection m ethods and chromatographictechniques . For acceptan ce, release or stability testing, accuracy should be optimizedsince the need to show de viation from the a ctual or true value is of the grea testconcern.

A. AccuracyAccuracy is the m easure of how close the experimental value is to thetrue value.


11/33

Accuracy studies for drug substance and drug product are recomm endedto be perform ed at the 80, 100 and 120% levels of label claim as stated inthe Gu ideline for Submitting Sam ples and Ana lytical Data for M ethodsValidation.For the drug product, this is performed frequently by the addition ofknown amounts of drug by weight or volume (dissolved in diluent) to theplacebo formulation working in the linear range of detection of tlieanalyte. This would be a true recovery for liquid formulations. Forformulations such as tablet, suppository, transderm al patch , this couldmean evaluating potential interaction of the active drug with theexcipients in the diluent. From a practical standpo int, it is difficult tomanufacture a single unit with know n amount of active drug to evaluaterecovery. This test evaluates the specificity of the method in thepresence of the excipients un der the chromatographic conditions used forthe analysis of the drug product. It will pick up recovery problems thatcould be encoun tered during the sam ple preparation and thechromatographic procedures. However, it does n ot count the effect of themanufacturing process.At each recorr~mendedevel studied, replicate samples are eva luated.The RSD of the replicates will provide the analysis variation or howprecise the test method is. The mean of the replicates, expressed as %label claim, indicates how accurate the test me thod is.Recommendations:Recovery data, at least in triplicate, at each level (80, 100 and 120% oflabel claim) is recommended. The m ean is an estimate of accuracy andthe RSD is an estimate o f sample analysis precision.

B. Detection Lim it and Quan titation Lim itThese limits are no rmally applied to related substance s in the drugsubstance or drug product. specifications on these limits are subm ittedwith the regulatory impurities method relating to release and stability ofboth drug substance and drug product.Detection limit is the lowest concentration of analyte in a sam ple that canbe detected, but not necessarily quantitated, under the stated


12/33

experimental conditions. Quan titation limit is the lowest concentration ofanalyte in a sarr~ plehat can be determined with acceptable precision andaccuracy under the stated experimental conditions.With U V detectors, it is difficult to assure the de tection precision o f lowlevel compounds due to po tential gradual loss of sensitivity of detectorlamps with age, or noise level variation by detector manufacturer. At lowlevels, assurance is needed that the de tection and quantitation limits areachievable with the test method each time. With n o reference standardfor a given impurity or means to assure detectab ility, extraneous peak(s)could "disappearlappear." A crude method to evaluate the feasibility ofthe extraneous peak detection is to use the percentage claimed fordetection limit from the area counts of the analyte. For example,detection limit claim of 0.01% for the analyte integrated area count of50,000 will give an area count of 5 that is not detectable.Though USP expresses detection limit and quan titation limit in terms of 2 .or 3, and 10 times noise level respectively, this concep t is not verypractical. Noise level on a detector during the method developmentphase m ay be d ifferent when samples are assayed on different detectors,etc. The use of standard(s) in the test method at the quantitation limitlevel (proposed by the applicant) is assurance that the impurity can beobserved and quantitated.Detector sensitivity can vary with the m odel number andlor m anufactureras illustrated in Table 1 for the analysis of a compound b y twocommercial detectors. The data should not be taken as the expectedratio of sensitivity of the two detectors. It is not known if other parameterswhich can also play a part, e.g., age of lamp , column , were consideredwhen setting these limits.


13/33

Table I. Comparison of Detector Sensitivity Limits in TwoCommercial Detectors.

Quantitatio 0.21% I 0.07%II n ~ i m i t IIIDetector 1 I Detector 2

I I

1 Detection I 0.16% II

I Limit I

One also should be cautious that baseline noise is not interpreted asextraneous peaks. Undulations may be observed at the void volume ifthe diluent for the sample is different from the so lvents (proportion andtype) used in the m obile phase.If a referen ce standard for the compo und of interest is available, astandard close to the quantitation limit or the specification could be used.For monitoring peak(s) with no reference standard for the impurity, adiluted reference standard of the drug substance is recomm ended. Themethod should then check that the high and low concentrations areoperating in the linear range of detection of the drug substance .Otherw ise the inform ation that is expressed as % area or height of thedrug substance peak from the same HPL chromatogram will be biased. Itshould also be noted that the extraneous peak using area count does notconsider the detection response which depends on the UV extinctioncoefficient or absorptivity of the com pound .Recommendations:1. Ana lysis repea tability and injection repeatability data at thequantitation limit.2. Use of an add itional reference standard a t the quan titation limitlevel in the test me thod.


14/33

C. LinearityThe l inear range o f detectabil i ty tha t obeys Beer's La w is dependentonzthe c o m p o ~ ~ n dnalyzed and detector used. The wo rking sampleconcentration and samples tested fo r accuracy should be in the l inearrange.Figures 1 and 2 i l lustrate the behavior of UV response vs.concentration o f a (a) l inear and (b ) non-linear relationship. A point tonote is tha t w he n mo nitoring impurity peaks expressed as percentarea of the parent drug substance, the impurity observed may n ot bea true reflection o f the theore tical amo unt if the non-l inear section o fthe concen tration curve is employed. In addition, the actual amountwil l be obtained only if the extinction coefficient or absorptivity valuesare the same for bo th impu rity and parent compound, Imp urityreference standards are o ften needed.

Figure 1. Con centrations vs. Peak Areas of Standards t o Il lustrateLinearity.

regression coefficient = 0.999998intercept = 0.103slope = 0.000011

Concentrat o nrwa&-- ,. ..wwow-

. ..4owwo--

wwooo--

~ww o - -

. .I-k--+.t I-(-.- - .--. I 4 I--70 Y) - ,- ,lo t ; . k L


15/33

Figure 2. Concentrations vs. Peak Areas of Standards Outside the Linear Range.

Concen t r a t i o n ug/mL

Recommendations:The linearity range for examination depends on -the purpose of the testmethod. For example, the recorr~mended ange for an assay methodfor content would be NLT * 20% and the range for anassaylimpurities combination method based on area % (for impurities)would be +20% of target concentration down to the limit ofquantitation of the drug substance or impurity. Under mostcircumstances, regression coefficient (r) is 2 0.999. Intercept andslope should be indicated.


16/33

D. PrecisionPrecision is the m easure of how close the data values are to each otherfor a number of m easurements under the same analytical conditions.ICH has defined precision to contain three com ponen ts: repeatability,intermediate precision and reproducibility. Ruggedness as defined inUSP XXll < I225>, 1990 incorporates the concepts described under theterms "intermediate precision", "reproduc ibility" and "robustness" of thisguide.I. Repeatability

a. Injection RepeatabilitySens itivity is the ability to detect sma ll changes in theconcen tration of the analyte in the sample. Sens itivity canbe partially controlled by m onitoriug the specification forinjection reproducibility (system suitability testing).The sensitivity or precision as measured by multipleinjections of a homogeneous sample (prepared solution)indicates the performance of the HPLC instrument under thechromatographic conditions and day tested. Theinformation is provided as part of the validation data and asa systeni suitability test. The specification, as thecoefficient of variation in % or relative standard deviation(RSD), set here will de te~ min ehe variation limit of theanalysis. The tighter the value, the mo re precise orsensitive to variation one can expect the results. Thisassumes that the chromatograph does not ma lfunction afterthe system suitability testing has been performed . Keep inmind, however, that it does n ot consider variations due tothe drug product manufacturing and laboratory samplepreparation procedures. As a n illustration for injection andR, variation, Table 2 provides representative data collectedwhen a leak developed in the chromatographic systemduring sampling. The set of four duplicate samples wereinjected sequen tially. Variations in peak area and drift ofretention times are noted. Sets of typical data from a well-behaved system for comparison are shown in Table 3.


17/33

Table 2. Represe ntative Injection Repeatabili ty Data for an HPLChroma tographic System that Dev eloped a Leak DuringSampling.

Table 3. Rep rese ntative Injection Repea ta,bility Data for Se lectFormulations from a Norma lly Functional HPLChromatographic System.

SampleA1A2

B182

C1C2

D lD2

R,5.625.66

5.876.13

6.216.48

6.736.99

Dosage Form

InhalationSolution

Solution forInhalation

Capsule

Peak Area21 5569921 20466

22056592288355

22270662265279

25818882602016

n

10

10

10

A R,0.04

0.26

0.27

0.26

Mean & SD

1993162 +5029

17222536288

1744320 +3133

A Peak Area35233

82696

38213

20128

RSD

0.25%

0.37%

0.18%


18/33

Recommendations:As part of methods validation, a minimum of 10 injectionswith an RSD of 11% is recomm ended. With the methodsfor release and stability studies, an RS D of I % RSD forprecision of the system suitability tests for at least fiveinjections (n 2 5) for the active drug either in drug substanceor drug product is desirable. For low level impurities, highervariations may be acceptable.

b. Analysis RepeatabilityDeterm ination, expresse d as the RSD , consists of mu ltiplemeasurements of a sam ple by the same a nalyst under thesame analytical conditions. For practical purpose , it is oftencomb ined with accuracy and carried out as a single study.See section 1V.A und er Accuracy.

lntermediate Precisionlntermediate precision was previously known as part ofruggedness. The a ttribute evaluates the reliability of the method ina d ifferent environment other than that used during d evelopment ofthe m ethod. The objective is to ensure that the method will providethe same results when similar samples are analyzed once themethod development phase is over.Depending on time and resources, the m ethod can be tested onmu ltiple days, analysts, instruments, etc.lntermediate precision in the test m ethod can be' partly assu red bygood sys tem su itability specifications. 'Thus, it is importan t to settight, bu t realistic, system suitability specifications.Recommendations:As a m inimum, data gene rated as described under section 1V.AAccuracy, for two separate occasions, is recommended to indicatethe intermed iate precision of the test method.


19/33

3. ReproducibilityAs de fined by ICH, reproduc ibility expresses the p recisionbetween laboratories as in collaborative studies. Multiplelaboratories are desirable but not always attainable be cause of thesize of the firm.Recommendations:It is not normally expected if intermediate precision isaccomplished.

E. RangeRange is the interval betwee n the h igh and low levels of analyte studied.See a lso sections 1V.A and C under Accuracy and Linearity respectively.The ranges recom mended in sections 1V.A and C under Accuracy andLinearity can b e applied to other ana lytes, e.g., prese rvatives.

F. RecoveryRecovery is expressed as the amountJweight of the compound of interestanalyzed as a percentage to the theoretical amoun t present in themedium.Full recovery should be obtained for the compoun d(s) of interest. Duringthe sam ple preparation procedure, the com pound of interest is recoveredfrom excipients in the formulation matrix ranging from a sim ple aqueoussolution to comp lex cream formulation, and from potential adhesion tocontainer/closure compone nts, e.g., glass vial, metered valve. Ingenera l, a simpler sample preparation procedure will result in a lowervariation of recovery. Data collection for recovery are discussed insection 1V.A under Accuracy .

G. RobustnessICH defines robustness as a measure of the method's capability toremain unaffected by small, but deliberate variations in methodparam eters. Robustness car1 be partly assured by good systemsuitability specifications. Thus, it is important to set tight, but realistic,system suitability specifications.


20/33

Testing varying some or all conditions, e.g., age of columns, colurr~nype,column temperature, pH of buffer in mobile phase, reagents, is normallyperformed.Recommendations:Data obtained from stud ies for robustness, though not usually submitted,are recomm ended to be included as pa rt of method validation.

H. Sam ple Solution StabilitySolution stability of the drug substance or drug product after preparationaccording to the test me thod should be evaluated acco rding to the testmethod. Mos t laboratories utilize autosamplers with overnight runs andthe samp le will be in solution for hours in the laboratory environmen tbefore the test procedure is completed. This is of concern especially fordrugs that can undergo degradation by hydrolysis, photolysis or adhesionto glassware.Recommendations:Data to suppo rt the sample solution stability under norm al laboratoryconditions for the duration of the test procedure, e.g., twenty-four hours,sho uld be generated. In exceptional cases where m ultiple days areneeded for sample preparation or solution storage, an appropriatestability time should be selected.

I. SpecificitylselectivityThe analyte should have no interference from other extraneouscomponents and b e well resolved from them . A representative HPLchromatogram or profile should be generated and submitted to show thatthe extraneous peaks either by addition of known com pounds or samplesfrom stress testing are baseline resolved from the parent analyte.Examples of the extraneous peaks are a s follows:

For the drug substance or raw material, the related substances toconsider are proce ss impurities (which include isom eric impurities)from the synthesis process, residual pesticides, solvents, andother extraneous components from extracts of natural origin.


21/33

For the drug product, the related substances may be impuritiespresent in the active drug, degrada tion products, interaction of theactive drug with excipients, extraneou s components, e.g., residualsolvents from the excipients or manufacturing process, leachables .or extractables from the container and closure system or from themanufacturing process.Submission of data from stress testing of the drug substance using acidand ba se hydro lysis, temperature, photolysis and oxidation accord ing tothe G uideline for S ubmitting Sam ples and Analytical D ata for MethodsValidation is recommended. Representative HPL chromatograms arerecommended for stressed and non-stressed samples that include testmethods for impurities, preservatives, etc. and placebo sample. With theimpurities test m ethod, the HPL chromatogram should indicate thepresence of impurities at the leve l of detectionlquantitation claimed. Thechromatograms should be legible, labeled, and the time or time scale an dattenuation should be indicated.Points to note are as follows:

1. The parent peak may be expan ded, e.g., by increasing theconcen tration, attenuation change , so that extraneouspeaks can be observed at a reasonable size to evaluatestability-indicating capab ility. See comm ents in section 1V.Bunder Lim its of Detec tionIQuantitation.2. The baseline should be on-sca le as off-scale baseline(observed as a flat straight line) can h ide minor peaks.

Peak purity can be determined by the photo-diode array detector. Lowlevel extraneous components present under the comp ound of interest,however, may not interfere or influence the U V spectrum of the ana lyte.Figures 3 and 4 illustrate the com bination of UV spectroscopy and H PLchromatography by photo-diode array detection using (a) 3-dimensionalplot and (b) conventional HPL chromatogram. The analyte elutes at 4.7minutes. It should be n oted that the quality of the U V spectra for the lowlevel components is poor.


22/33

Figure 3. A Representative 3-Dimensional Plot of the HPL Chromatogramwith the UV Spectra.

T i mi . rnnnrI1 - - - 1 4 . 1 m i r r


23/33

Figure 4. A Representative Conven tional HPL Chromatogram w i t h UVSpectra.

Wh en stressed samples are used, an appropriate dete ctort integ ratorsett ing should be selected. For example, to be able to d etect lo wlevels, e.g., 0.1% degradation products, the parent peak should be ofa size tha t a t least a 0.1% detectabi l i ty o r area c ou nt is feasible.Recommendations:Representative HPL chroma tograms should be sub mitted fo; stressedand non-stressed samples tha t include impuri t ies tes t m ethod,preservat ive(s), etc, wi th the related placebo sample. RepresentativeHPL chromatogram(s) to show se lec ti vi ty by the add i tion o f kno wnextraneous compounds also should be subm it ted.


24/33

J. System Suitability Specifications and TestsThe accuracy and precision of HPLC data collected begin with a well-behaved chromatographic system. The system suitabilityspecifications and tests are parameters that provide assistance in.achieving this purpose. This section explains the terms as indicated inFigure 5, and provides recommendations and illustrations.

Figure 5. Definition of Terms for the System Suitability Parameters.

WhereW = width of the peak determined at either 5% (0.05) or 10% (0.10) rom thebaseline of the peak heightf = distance between peak maximum and peak front at Wto = elution time of the void volume or non-retained componentst, = retention time of the analytet = peak width measured at baseline of the extrapolated straight sides to baseline


25/33

t = retention time of the analytet = peak width measured at baseline of the extrapolated straight sides to baseline1. Capacity factor (k')

k' = (t, - to)IoThe capacity factor is a m easure of where the peak of interest islocated with respect to the void volume, i.e., elution time of thenon-retained components.Recommendations:The peak should be well-resolved from other pea ks and the voidvolume . Gene rally the value of k' is > 2.

2. Precisionllnjection repeatability (RSD)Injection precision expressed as RSD (relative standard dev iation)indicates the performance of the HPL chromatograph whichincludes the plumbing, column, and environmental conditions,the time the samples are analyzed. It should be noted that samplepreparation and manufacturing variations are not considered.Recommendations:RSD of s 1% for n 2 5 is desirable.

3. Relative retention (a )

Relative retention is a me asure of the relative location of twopeaks. This is not an essential parameter as long as the resolution(R,) is stated .4. Resolution (R,)

R, is a measu re of how well two peaks are separated. For reliablequantitation, well-separated peaks are essential for quantitation.This is a very useful param eter if potential interference peak(s)may be of concern . The closest potential eluting peak to the


26/33

R, i s min imal ly in fluenced by the ra t io o f the t w o com poundsbeing measured. The reso lu t ion of peaks as ind icated by the R,values is sho wn in Figure 6.Figure 6. Separation of Peaks as Indicated by R, Values.

Recommendat ions:R, o f > 2 between the peak of in terest and the c losestpoten t ial interfer ing peak ( impu ri ty, excipient, degrad at ionprod uct, interna l standard, etc.) is desirable.

5. Tai l ing factor (T)

The accuracy of quant i tat ion decreases with increase in peaktai l ing because of the di f f icul t ies e ncountered by3the integratorin determin ing wheretwh en th e peal< ends and hence thecalculat ion of the area under the peal


27/33

peak of interest. Figures 7 and 8 i l lustrate the tai l ing factorsand the effe ct on quanti tat ion. I f the integrator is unable todetermine exactly whe n an upslope or downslope occurs,' accuracy drops.

Figure 7. HPLC Peak with Various Tailing Factors.

Tailing factor = 1.3 I\Tailing factor = 3.7 I


28/33

Figure 8. Effect of Peak Tailing on Ouantitation.

Recommendations:T o f 5 2


29/33

6. Theoretical plate num ber (N)

Theo retical plate num ber is a measure of column efficiency, that is,how m any peaks can be located per unit run-time of thechromatogram.N is fairly constant for each peak on a chrom atogram with a fixedset of operating conditions. H, or HETP, the he ight equivalent of atheoretical plate, measures the colum n efficiency per unit length(L) of the column. Param eters which can affect N or H includepeak position, particle size in column, flow-rate of mobile phase,column temperature, viscosity of mobile phase, and molecularweight of the analyte. Figure 9 shows one set of compounds A, B ,and C under two different chromatographic conditions resulting in,e.g., R,s for B to be 3 and 8.5 minutes, respectively. Anexamination of peak B indicates that the theoretical plate valuesare different even though the peaks appea r similar visually.


30/33

Figure 9. Effect of Retention Times on Theoretical Plates.


31/33

Recommendations:The theoretical plate number depends on elution time but ingeneral.should be > 2000.

General Recommendation:System suitability testing is essential for the assurance of the qualityperformance of the chromatographic system. The amount of testingrequired will depend on the purpose of the test method. For dissolution orrelease profile test methods using an external standard m ethod, k', T andRSD are minimum recommende d system suitability tests. Foracceptance, release, stability, or impuritiesldegradation me thods usingexternal or internal standards , kt, T, R, and RSD are recom mended asminimum system suitability testing parameters. In practice, each me thodsubm itted for validation should include an appropriate number of systemsuitability tests defining the necessary characteristics of that system.Additional tests may be selected at the discretion of the applicant or thereviewer.

K. General Points to ConsiderSome basic points to note in the test m ethod are:1. The sam ple and standard should be dissolved in the mobile phase.If that is not possible, then avoid using too high a leve l of theorganic solvent as com pared to the level in the mobile phase.2. The sample and standard concentrations should be close if not thesame.3 . The samples should be bracketed by standards during theanalytical procedure.4 . Filtration of the sam ples be fore injection is occasionally observed.Filtration will remove particulates (centrifugation performs thesame function) that may clog columns. Adhesion o f the analyte to

the filter can also happen. This will be of irr~portance specially forlow level impurities. Data to validate this aspect should besubm itted by the applicant.


32/33

V. COMMEN TS AND CONCLUSIONSHPL Chrom atoaraphic M ethods for Drua Substance and Drua Product.Methods should not b e validated as a one-time situation, but methods should bevalidated and designed by the develope r or user to ensure ruggedness orrobustness throughout the life of the m ethod.The variations due to the drug product manu facturing process, the laboratorysamp le preparation procedure and the instrument performance contribute to theaccuracy of the data obtained 'from the analysis. With prope r validation and tightchromatographic performance (system suitability) criteria, an improvem ent in thereliability of the data ca n be obtained. Variations, except from the drug productmanufacturing process, will be minimized. Only with good reliable validatedmethods, can data that are gene rated for release, stability, pharmacokinetics betrust-worthy.

VI. ACKNOWLEDGEMENTSCom ments froni Drs. Hoiberg, Poochikian, Blumenstein, Schroeder, Look,Tolgyesi (HFD-150); Dr. Layoff (HFH -300); Drs. Zimmerman and P iechocki(HFD-110); and the A nalytical Methods Technical Com mittee: Dr. Sheinin, Mr.Shostak, Ms. Cunningham, M s. Jongedyk, Mr. Leutzinger, Dr. Seggel, Ms.Sharkey and M r. Smela are appreciated.

VII. REFERENCES1. Guideline for Submitting Sa mples and Analytical D ata for MethodsValidation, February 1987.2. United States Pharmacopeia, XXII, 1990. < I225>.3. Text on Va lidation of Analytical Procedures, International Con ference onHarmonization, September 1993.


33/33

Subm itted by:

Linda L. Ng, Ph.D.

Approved by CMC CC:

Charles Kumkum ian, Ph.D. Roger L. Williams, M.D.

Documents

Validation of Chromatographic Methods-FDA