The requirements for analytical method validation which are listed in “ANVISA RDC 166/2017-Validation of Analytical Methods” and “ANVISA RDC 53/2015- Guide on notification, identification, and qualification of degradation products in medicinal products,” and are going beyond those defined in ICH Q2(R1), FDA guidance for Industry and relevant chapters of USP, EP and JP are summarized below.

Selection of material for method validation

• The samples used to perform the analytical method validation experiments must be within the expiry date.
• Chemical Reference Substances provided by official pharmacopoeias or a well-characterized reference material must be used for method validation purposes. It is not acceptable to use working standards.
• If characterized reference standards are used, the full characterization report needs to be attached to the validation report.

Verification of compendial test methods

• For compendial methods, a partial validation study must be performed including an assessment of specificity, precision, and accuracy, if applicable.
• For quantitative impurity test methods, the partial validation must include the limit of quantitation.
• In the case of limit tests, specificity and limit of detection have to be assessed.

Validation waiver

• The approach of validation waiver is not accepted by ANVISA.

Specificity testing of stability-indicating test methods

• For quantitative test methods or limit tests which are applied to stability testing of drug substances and drug products, forced decomposition tests have to be performed to provide evidence of their stability-indicating properties.
• Forced decomposition testing must include the following stress conditions: heat, humidity, acid solution, alkaline solution, oxidizing solution, photolytic exposure, and metallic ions.
• The conditions should be varied so that a decrease in the main peak area of at least 10% is reached, optimally without occurring generation of secondary degradation products. If it is not possible to obtain a degradation level of at least 10%, a scientific rationale/justification has to be provided.
• In the case of drug product test methods, the forced decomposition study should be carried out on the formulation, on the excipient matrix, as well as on the isolated and, in the case of fixed-dose combinations, the combined active pharmaceutical ingredient(s).
• The forced decomposition study should be done on all dosage strengths of the medicinal product. Appropriate bracketing may be adopted in case of forced decomposition studies for a drug product with multiple dosage strengths. If the composition is dose-linear and thus the dosage strengths have the same qualitative and quantitative composition and are analyzed according to the same analytical procedure with the same working concentration, forced decomposition testing on one dosage strength might be sufficient. If the dosage strengths have the same qualitative composition, but differ in the quantitative composition and/or the working concentration applied for analytical testing, forced decomposition testing may be performed for the dosage strength with the lowest drug load (potentially the most conservative / highest potential for interactions with excipients) or on more than one dosage strength.

Linearity

• For establishing linearity, a minimum of five concentration levels distributed over the relevant range of the analytical procedure has to be used with triplicate solution preparation per level. The solutions per level have to be independently prepared.
• Linearity testing has to be performed with pure reference materials in the absence of the matrix. A combined approach of linearity and accuracy testing in the case of drug product test methods is not accepted by ANVISA.
• To apply unweighted linear regression, homoscedasticity (homogeneity of variances: constant variability of y-values over the whole concentration range) should be evaluated by a suitable model.
• Statistical tests should be performed at a significance level of 5% (95% confidence level).
• The correlation coefficient should be more than 0.990.
• The confidence interval of the slope (at 95% confidence level) has to be reported to provide evidence that the slope is significantly different from zero.

Range

• For the dissolution test method, if dissolution profile testing is performed, the validated range of the dissolution test method should include the lowest individual dissolution values obtained at the first testing time point. If the individual dissolution values at the first testing time point are below 5%, it might be required to also determine the limit of quantitation of the analytical procedure.

Precision

• It is not accepted by ANVISA to obtain the method variability data from linearity or accuracy testing. A separate precision experiment is required.
• For impurity methods, ANVISA requires performing precision testing at the specification limit concentration of the corresponding impurities. In the case of known impurities absent or present in the sample in a concentration smaller than the specified limit, the sample should be spiked with known concentrations of the corresponding impurity standard in order to reach the specification limit concentration.
• In the case of unknown impurities, precision testing should be performed on the active ingredient spiked to the matrix in a concentration corresponding to the specified limit for unknown impurities.
• Stressed samples or unpurified drug substance samples may be used for precision testing only in the exceptional case, that reference materials of the corresponding impurities are not available, e.g. due to non-stability in its isolated form.
• Intermediate precision testing should involve the analysis of the same sample at the same laboratory on at least two different days performed by at least two different analysts. The same concentrations and the same number of determinations as defined for repeatability testing should be applied (i.e. at least 3 concentration levels with at least 3 replicates per level or 6 replicates at 100% if the individually prepared test concentration).

Accuracy

• For drug substance methods, the approach that accuracy is inferred from specificity, linearity, and precision is not accepted. Accuracy has to be investigated by either applying the analytical procedure to a reference substance with known purity or by comparing the results with those from a second validated test method.
• For unknown impurities, accuracy should be assessed by spiking the matrix with the active ingredient or a known impurity, in concentrations considering the working range of the test method for unknown impurities.

Limit of detection

• For instrumental methods, the detection limit may be calculated from the signal-to-noise ratio or derived from the regression line. The signal-to-noise ratio should be equal to or greater than 2:1.
• In the case that LOD is calculated based on the signal-to-noise ratio or linear regression parameters, the limit has to be verified by the analysis of a suitable sample prepared at the detection limit concentration.

Limit of quantitation

• For instrumental methods, the quantitation limit may be calculated from the signal-to-noise ratio or derived from the regression line. The signal-to-noise ratio should be equal to or greater than 10:1.
• In the case that LOQ is calculated based on the signal-to-noise ratio or linear regression parameters, the limit has to be verified by the analysis of a suitable sample prepared at the quantitation limit concentration.
• Precision and accuracy should be additionally investigated for concentrations corresponding to the limit of quantitation.

Robustness

• Robustness testing should include the following parameters:
  • Sample preparation: Stability of solutions, extraction time, filter compatibility
  • Spectrophotometry: pH of the solution, different batches/manufacturers of solvent
  • Liquid chromatography: pH of mobile phase buffers, the quantitative composition of mobile phase, different batches/manufactures of columns, column temperature, flow rate
  • Gas chromatography: carrier gas velocity, temperature, different batches/manufacturers of columns.
• If the column variability (different batches/manufacturers) is evaluated as part of intermediate precision, it needs to be documented accordingly in the validation report.
• The absence of the evaluation of any of the proposed variations should be justified. In the case of the quantitative test method, the impact of the proposed variations on the results should be assessed by applying the same acceptance criteria as for accuracy.
• During robustness testing, it should be demonstrated that the system suitability criteria are fulfilled.

Author: Mohammed Raihan Chowdhury

Quality Assurance and Data Integrity Expert

Click for LinkedIn Profile