LAB-135 Validation of Analytical Test Procedure

5.1 Overview

5.1.1 Analytical and biological test method validation / verification shall include consideration of the following validation parameters:

5.1.1.1 Specificity

5.1.1.2 Linearity

5.1.1.3 Range

5.1.1.4 Accuracy

5.1.1.5 Precision

5.1.1.5.1 Repeatability

5.1.1.5.2 Intermediate precision (intra laboratory)

5.1.1.5.3 Reproducibility (inter laboratory)

5.1.1.6 Limit of Detection

5.1.1.7 Limit of Quantitation

5.1.1.8 Robustness

5.1.1.9 Reproducibility (for biological test methods only)

5.1.2 Validation parameters applicable to different types of methods are indicated by an “X” in the tables below:

Table 1: For Non-Compendial Analytical and Biological Test Methods

            (a)        To be considered if critical to the test method

                        (b)       Reproducibility is not required for analytical test method validation, but its results may be used in place of intermediate precision.

Table 2: For Compendial Analytical and Biological Test Methods

(a)      An assessment should be performed by considering the impact of different excipients, formulations and impurity profiles, and only necessary parameters need be evaluated.

(b)      Reproducibility is not required for analytical test method validation, but its results may be used in place of intermediate precision.

(c)      To be considered if critical to the test method.

5.2 Validation Parameters

5.2.1 Specificity

5.2.1.1 For Identification Tests:  Specificity needs to discriminate between the compound(s) of interest and any other compounds that are likely to be in the sample such as excipients, impurities or degradation products.

5.2.1.2 For Assay and Impurity Tests:  Representative chromatograms should be used to demonstrate specificity for each component.

5.2.1.2.1 If impurities or degradation products are available, the API or drug product will be spiked with impurities.

5.2.1.2.2 For assay, specificity is established by demonstrating that the analytical response of the analyte is unaffected by the presence of other components in the sample.

5.2.1.2.3 For an impurity test, specificity is established by demonstrating  that the impurities are separated from each other and/or from other components in the same matrix.

5.2.1.2.4 If impurities or degradation products are not available, API or drug product will be subjected to relevant stress conditions (forced degradation), such as light, heat, humidity, acid / base hydrolysis and oxidation.  Chromatographic peak purity assessment will then be performed using diode array detector or mass spectrophotometer.  Alternatively specificity will be demonstrated by comparing the test results of samples containing impurities or degradation products to a second well-characterised method.

5.2.1.2.5 Absence of Interference from Excipients:  Prepare a placebo by mixing the excipients in the same ratio indicated in the drug product formulation.  A portion of the placebo is also stored at 105^oC for one day (previously:  60^oC for 2 weeks).  Analyse the placebo, stressed placebo and drug product concurrently with a standard (active(s) only) solution using the proposed test method, with a run time 2.5 times the specified run time.  Compare these chromatograms against each other.  Interference with the active peak(s) must not be greater than 0.5% by area with respect to the active peak(s).

5.2.1.2.6 Absence of Interference from Potential Known Degradation Products of the Active (if available):  Spike a solution of the active using a solution of the degradation product(s) at its (their) specification level(s) (if known) or at approximately 1.0% of the analytical concentration of the active.  If necessary, individual solutions of each degradant are prepared for the aid of identification.  The chromatograms are inspected for interference.  Photo-diode array detector or mass spectrometry can also be employed to assess peak purity.  The known potential degradation products should be separated from each other and/or from other components in the same matrix.

5.2.1.2.7 Forced Degradation for API or Finished Products:  If impurities or degradation products are not available, forced degradation can be carried out as per the following guidelines:

5.2.1.2.7.1 Light:  Prepare 50 mL of the sample solution as per the analytical method.  Store the sample in a light cabinet 2000Lux for one month (equivalent to 1.4 million Lux hours) or leave the sample under sunlight for 8 hours.  Compare the chromatogram to that of a control chromatogram.

5.2.1.2.7.2 Heat:  Prepare 50 mL of the sample solution as per the analytical method.  Store in a 105ºC oven for 7 days or boil it on a waterbath for 7 hours).  Compare the chromatogram to that of a control chromatogram.

5.2.1.2.7.3 Acid:  Add 2 mL of 1 M HCl to 50 mL of the sample solution prepared as per the analytical method (Proposed:  2 mL of conc. HCl in 50 mL sample and boil for 1 hours on waterbath).  Store for 7 days.  Compare the chromatogram to that of a control chromatogram.

5.2.1.2.7.4 Base:  Add 2 mL of 1 M NaOH to 50 mL of the sample solution prepared as per the analytical method (Proposed:  2 mL of 5 M NaOH in 50 mL sample and boil for 1 hours on waterbath).  Store for 7 days.  Compare the chromatogram to that of a control chromatogram.

5.2.1.2.7.5 Oxidation:  Add 5 mL of 1.0%w/v hydrogen peroxide to 50 mL of the sample solution.  Mix and store for 7 days.  (Proposed:  1 mL of 30%w/v hydrogen peroxide in 50 mL sample and boil for 1

hours on waterbath) Compare the chromatogram to that of a control chromatogram.

5.2.1.2.7.6  Stress conditions can be changed since the aim of forced degradation is to ensure that the product / API is definitely degraded so that the method can be used to demonstrate that separation of active peak from any possible break-down peaks is achieved.  Much stronger conditions, combined with heat, have better chances to achieve degradation and can also reduce waiting time and speed up the validation process.

5.2.1.3 For Dissolution Tests:  Specificity needs to discriminate between the compound(s) of interest and any other compounds that are likely to be in the sample solution such as excipients, dissolution media reagents.

5.2.1.3.1 For HPLC assay, the active peak(s) should be separated from all other peaks.  Any interference with the active peak(s) must not be greater than 0.5% by area with respect to the active peak(s).

5.2.1.3.2 For UV assay, absorbance contribution from placebo must be less than 2% with respect to the active absorbance.

5.2.1.4 For Biological Tests:

5.2.1.4.1 For compendial test methods, specificity must be demonstrated under actual use conditions in the laboratory, including any minor variations in the test method.

5.2.1.4.2 For non-compendial test methods, specificity should be determined for each active to demonstrate that:

5.2.1.4.2.1 The active or internal standard are not interfered by any other substances (e.g. impurities, metabolites);

5.2.1.4.2.2 The active concentration is within the linear range of the method.

5.2.2 Linearity

5.2.2.1 Linearity shall be determined using at least five active concentrations / levels to cover the nominal active concentration specified in the test method (e.g. 50%, 75%, 100%, 125% and 150%).  It shall be demonstrated directly on the analyte or with the analyte contained in the sample matrix.

5.2.2.2 Statistical methods are used to calculate a regression line using the method of least squares.  The correlation coefficient (R²), y-intercept, slope of the

regression line and the residual sum of squares must be reported.  A graph of the data shall be included in the method validation report.

5.2.2.3 The test method is considered to be linear and unbiased if the correlation coefficient (R²) is greater than or equal to 0.995.

5.2.3 Range

The specified range is established by confirming that the test method provides an acceptable degree of linearity, accuracy and precision when applied to samples containing amounts of analyte at the extreme ends of the specified range of the analytical method.

5.2.3.1 The following minimum specified range will be considered for API’s:

5.2.3.1.1 Assay:  80 to 120% of the test concentration.

5.2.3.1.2 Impurity determination:  from the reporting level (e.g. QL) of the impurity to 120% of the specification.

5.2.3.2 The following minimum specified range will be considered for Drug Products:

5.2.3.2.1 Assay:  80 to 120% of the test concentration.

5.2.3.2.2 Content Uniformity:  a minimum of 70 to 130% of the test concentration.

5.2.3.2.3 Dissolution:  ± 20% over the specified range (e.g. condition for S3 is “no result is less than Q – 25%”, hence Range would be from Q – 45% (25% + 20% = 45%) to 130% (upper limit for dissolution is 110% + 20% = 130%).

5.2.3.2.4 Impurity determination:  From the reporting level (e.g. QL) of the impurity to 120% of the specification.  For validation of impurity test methods where specification has not yet been set, the range around a suggested (probable) limit will be used.

5.2.3.2.5 If assay and purity are performed together as one test and only a single standard is used, linearity, accuracy and precision must be demonstrated from the reporting level of the impurities to 120% of the assay specification.

5.2.4 Accuracy

5.2.4.1 The accuracy of an analytical method is the closeness of test results obtained by that method to the true or most likely value.  Accuracy may often be expressed as percent recovery by the assay of known, added amounts of analyte.

5.2.4.2 Accuracy will be established across the specified range of the test method.  This applies to assay methods, including methods for dissolution and content uniformity, to quantitative impurity methods for API and drug product and to biological method.

5.2.4.3 For API, accuracy shall be demonstrated by one of the following methods:

5.2.4.3.1 Application of the analytical method to an analyte of known purity;

5.2.4.3.2 Comparison of the results of the proposed analytical method to a second, well characterised method, the accuracy of which is stated and/or defined; or

5.2.4.3.3 Accuracy may be inferred once precision, linearity and specificity have been established.

5.2.4.4 For Drug Products, accuracy shall be demonstrated by one of the following methods:

5.2.4.4.1 Known quantities of the API to be analysed are added to the product placebo mixture and the test method is applied.

5.2.4.4.2 In cases where all product excipients cannot be obtained, it is acceptable either to add known quantities of the analyte to the drug product or to compare the results obtained from a second, well characterised method, the accuracy of which is stated and/or defined.

Table 3: Accuracy Limits for Given Release Limits of Active

5.2.4.5 For Impurity (Quantitation), accuracy shall be assessed on samples (API or drug product) spiked with known amounts of impurities.

5.2.4.5.1 In case where it is impossible to obtain samples of certain impurities and/or degradation products, it is considered acceptable to compare results obtained by an independent procedure.

5.2.4.5.2 The response factor of the API can be used to quantitate the impurities and/or degradation products.  Alternatively a surrogate (e.g., compound whose structure and properties are closely related to the impurity of interest) may be used to demonstrate accuracy.

5.2.4.5.3 For API and drug product quantitative impurity methods, accuracy may be inferred when the acceptance criteria for specificity, linearity and precision are met.

 Table 4: Accuracy Limits for Given Release Limits of Impurities

5.2.4.6 The following accuracy requirements shall apply:

5.2.4.6.1 For assay methods (including methods for content uniformity), biological methods and quantitative impurity methods for APIs and drug products, accuracy shall be assessed using minimum of nine (9) determinations over a minimum of three (3) concentration levels covering the specified range of the test method [e.g., three (3) concentrations with three (3) replicate sample preparations each].  Results from each of the nine (9) determinations shall be reported as percent recovery by the assay of known added amount of analyte in the sample or as the difference between the mean and the accepted true value together with the confidence intervals.

5.2.4.6.2 For dissolution methods, accuracy shall be assessed using eighteen (18) determinations over a minimum of three (3) concentration levels covering the specified range of the test method [i.e., three (3) different concentrations, each consisting of six (6) replicate samples from six dissolution vessels].  Results from each of the eighteen (18) determinations must be reported.  Individual recovery results must be within 97% to 103% of the input.

5.2.4.6.3 Accuracy shall be reported for each concentration level as percent recovery by assay of a known added amount of analyte in the sample or as the difference between the mean and accepted true value together with the confidence intervals.

5.2.5 Precision

5.2.5.1 Validation of tests for assay (including dissolution and content uniformity), quantitative determination of impurities and physical properties should include precision.  Precision is expressed as the variance, standard deviation, or relative standard deviation (RSD).

5.2.5.2 Precision will be considered at three levels:  (1) Repeatability, (2) Intermediate precision, and (3) Reproducibility.

5.2.5.3 Either intermediate precision or reproducibility is evaluated.  It is not necessary to evaluate both.

5.2.5.4 Repeatability

5.2.5.4.1 For assay methods (including content uniformity), biological methods and quantitative impurity methods:

5.2.5.4.1.1 If the method requires composite sample to be ground, repeatability shall be assessed using a minimum of nine determinations over a minimum of three concentration levels over the specified range of the analytical method.  [e.g., three different sample weights used providing three different concentration levels (e.g. 80%, 100% and 120%), each consisting of three replicate sample preparations].

5.2.5.4.1.2 If the method requires whole tablets to be used, six replicate samples are to be prepared and analysed as per method.

Table 5: Method Precision Limits for Actives

Table 6: Method Precision Limits for Degradants and Impurities

5.2.5.4.2 For dissolution method:  Six tablets shall be used to perform the dissolution test.  The % RSD for six results should not be more than 5%.

5.2.5.5 Intermediate Precision

Intermediate precision is established depending on the circumstances under which the method is intended to be used.  The effects of operational and environmental variables on the precision of the method will be measured.  Typical variations to be measured include days, analysts and equipment.  It is not necessary to study these effects individually.

5.2.5.6 Reproducibility

Reproducibility is assessed by inter-laboratory studies.  It is only considered when the method will be used across multiple sites.

5.2.6 Limit of Detection (DL)

5.2.6.1 The detection limit shall be at or below the reporting limit for the compound under test.  DL shall be required for limit tests except where the Quantitation Limit (QL) is established.  DL is determined by the analysis of samples of known concentrations of analyte and by establishing the minimum level at which the analyte can be detected.  The DL and the method used for determining the DL shall be presented in the validation report.

5.2.6.2 If DL is based on visual evaluation or signal-to-noise ratio, the presentation of relevant instrument response is considered acceptable for justification.

5.2.6.3 In case where an estimated value for the DL is obtained by calculation or extrapolation, this estimate shall subsequently be validated by the independent analysis of a suitable number of samples known to be near or prepared at the DL.

5.2.6.4 DL based on signal-to-noise shall be applied to analytical procedures, which have a measurable baseline noise.  Determination of signal-to-noise ratio is performed by comparing measured signals from samples with known low concentrations of analytes and a baseline free of signals.  A signal-to-noise ratio of 3:1 is considered acceptable for determining the detection limit.

5.2.6.5 DL based on the standard deviation of the response and the slope is expressed as: 3.3 s DL = S

Where,             s = the standard deviation of the response

S = the slope of the calibration curve.

5.2.6.6 The slope S is estimated from the calibration curve of the analyte generated using a minimum of five points.  The value of s shall be estimated using one of the following approaches:

5.2.6.6.1 Based on the Standard Deviation of the Blank

Measurement of the magnitude of the analytical background response is performed by analysing a minimum of five blank samples and calculating the standard deviation of the responses.

5.2.6.6.2  Based on the Calibration Curve

A specific calibration curve/s shall be studied using samples containing the analyte in the range of DL.  The residual standard deviation (root mean square) of the regression line or the standard deviation of y-intercepts of multiple regression lines shall be used as the standard deviation.

5.2.7 Limit of Quantitation (QL)

5.2.7.1 The quantitation limit is established for quantitative tests for impurities.  The QL and the method used for determining the QL shall be presented in the validation report.  The QL shall be validated by analysis of a minimum of samples with known concentrations of analyte and by establishing the minimum level at which the analyte can be quantified with acceptable accuracy and precision.  At a minimum, the QL shall be at or below the reporting limit for the analyte under test.

5.2.7.2 Visual examination will be used for non-instrument methods.

5.2.7.3 QL based on signal-to-noise shall be applied to analytical procedures, which have a measurable baseline noise.  Determination of signal-to-noise ratio is performed by comparing measured signals from samples with known low

concentrations of analytes and a baseline free of signals.  A signal-to-noise ratio of 10:1 is considered acceptable for determining the QL.

5.2.7.4 QL based on the standard deviation of the response and the slope is expressed as: 10 s QL = S

Where,             s = the standard deviation of the response

S = the slope of the calibration curve.

5.2.7.5 The slope S is estimated from the calibration curve of the analyte generated using a minimum of five points.  The value of s shall be estimated using one of the following approaches:

5.2.7.5.1 Based on the Standard Deviation of the Blank

Measurement of the magnitude of the analytical background response is performed by analysing a minimum of five blank samples and calculating the standard deviation of the responses.

5.2.7.5.2 Based on the Calibration Curve

A specific calibration curve/s shall be studied using samples  containing the analyte in the range of QL.  The residual standard deviation (root mean square) of the regression line or the standard deviation of y-intercepts of multiple regression lines shall be used as the standard deviation.

5.2.7.6 The QL will be validated by the analysis of a minimum of three samples in the product matrix, known to be at the concentration of the QL, and demonstrating accuracy and precision.

5.2.8 Robustness

5.2.8.1 Robustness shall be considered during either the method development phase or validation on the type of procedure under study.

5.2.8.2 Robustness shall demonstrate the reliability of an analysis with respect to small deliberate variations in analytical parameters.  If measurements are susceptible to variations in analytical conditions, the analytical conditions must be controlled or a precautionary statement included in the procedure.

5.2.8.3 Data collected during robustness studies shall be used to set or verify system suitability requirements.

5.2.8.4 Examples of typical variations include:

5.2.8.4.1 Stability of analytical solutions;

5.2.8.4.2 Sample extraction time;

5.2.8.4.3 Influence of variations of pH in HPLC mobile phase;

5.2.8.4.4 Influence of variations in HPLC mobile phase composition; 5.2.8.4.5 Different HPLC columns;

5.2.8.4.6 Different HPLC flow rates;

5.2.8.4.7 Different column temperatures;

5.2.8.4.8 Different filters;

5.2.8.4.9 Influence of variation of reagent quantity in dissolution media; 5.2.8.4.10 Influence of variation of instrument operation parameters for biological test.

5.2.8.5 Stability of reference standard and sample solutions involves the storing of solutions in the laboratory environment and in a refrigerated environment is compared to freshly prepared standard solutions.  Suitable time points include 0, 6, 12, 24, 48, 72 and 120 hours.  For reference standard solutions degradation of 1% or more is unacceptable.  For sample solutions degradation of 1.5% or more is unacceptable.

5.2.8.6 Sample extraction time involves a sonication / extraction study on the sample and evaluating the results based on the amount of analyte quantitated by the analytical method.  Record observations for the different time points on a graph.  A plateau consisting of three data points should be observed when accepting the sonication / extraction time point.  The beginning of the plateau will be considered as the optimum sample extraction time.

5.2.8.7 With variations in the HPLC procedure the system suitability requirements of the analytical method should be met.  Also the sample result obtained for each condition should be within the precision acceptance limit, as detailed in Tables 5 and 6.

5.2.9 Reproducibility

5.2.9.1 Biological test methods must be tested for reproducibility and shown to be precise and accurate when used in two or more laboratories.

5.2.9.2 A reproducibility experiment should only include test samples prepared from a homogeneous or composite sample.

5.3 System Suitability

5.3.1 System suitability test parameters are to be established for a particular method depending on the type of method being validated.

5.3.2 Typical system suitability parameters for chromatographic systems to be considered during method validation are:

5.3.2.1 Resolution – Peak resolution (R) of the peaks of interest and/or other components.  Determine the resolution, if necessary, for the critical pair of peaks of interest by making one injection of the Resolution Solution.  The data collected during the validation will be used to establish the specification limit.

5.3.2.2 System Precision – Perform a minimum of 6 consecutive injections of standard solutions and calculate the %RSD of the peak response (area or height).  The %RSD for the peak response should not be greater than 2.0%, unless otherwise specified in the development protocol.  Also calculate the %RSD of the peak retention time.  The %RSD for the peak retention time should not be greater than 1.0%, unless otherwise specified in the development protocol.

5.3.2.3 Tailing Factor – Calculate the tailing factor(s) for each analyte(s) peak using the standard injections used for system suitability.  The data collected during the validation will be used to establish the specification limit.  Typically, the tailing factor(s) should not be outside 0.0 to 1.5 (British Pharmacopoeia) unless otherwise specified in the development protocol.

5.3.2.4 Column Efficiency (Number of Theoretical Plates) – Calculate the number of theoretical plates for the active(s) peak using the standard injections used for system precision.  The data collected during the validation will be used to establish the specification limit.