Title: Determining Testing Patterns and Acceptance Criteria for Analytical Method Transfers | |||||
| Guidance Number: 88 | |||||
| Prepared by: | Date: | Supersedes: | |||
| Checked by: | Date: | Date Issued: | |||
| Approved by: | Date: | Review Date: | |||
Determining Testing Patterns and Acceptance Criteria for Analytical Method Transfers
Introduction
This document provides guidance for setting experimental testing patterns and acceptance criteria for Analytical Method Transfer Exercises (ATME).
This document provides guidance to GLP sites in identifying lots and number of samples for testing, setting appropriate acceptance criteria for conducting transfers.
The extent of AMTE testing should be commensurate with the method capability and its intended use. The specification levels, validation data and historical performance for each method should be reviewed (where available) to determine the appropriate transfer criteria/limits. Based on the intended use of the method, appropriate acceptance criteria may include evaluation of inter laboratory differences, system suitability, reproducibility, selectivity, sensitivity, recovery and/or comparison to the appropriate specification or method requirements (e.g. spectroscopic or chromatographic identification).
Appendices I and II provide decision trees for conducting assay and impurity method transfers.
Identification of Lots to be used for Transfer
When an API or a single strength of a drug product is being transferred, a minimum of one lot may be used for the AMTE. If a single lot is to be used, it is recommended that the testing pattern include multiple analysts, multiple days and /or multiple instruments to assess the Receiving Laboratory’s (RL) ability to generate consistent, reproducible results. Where multiple lots or batches are used for the transfer, the Transferring Laboratory (TL) should evaluate the necessity for including multiple analysts, multiple days and /or multiple instruments in the transfer testing pattern.
If multiple drug product strengths are manufactured from a common or similar blend, only the highest and lowest bracketing strengths need to be included in the AMTE.
Successful transfer of the high and low dosage strengths will qualify the RL to test all of the strengths within the bracketed range, provided that the sample preparations for the different strengths are similar.
When identifying the materials to be used for the transfer, the TL will determine if historical data will be used or if comparative data will be generated specifically for the transfer. Historical data is defined as data generated by a qualified laboratory outside the transfer process. Some sources of historical data are stability, certificate of analysis, and validation data. If commercial lots are used for the transfer, it is recommended to avoid using lots whose most recent results lie near the specification limit. This is not a concern for expired, purposely degraded, or development lots.
Establishing Acceptance Criteria
Tables 1 and 2, in conjunction with the recommendations given herein, provide initial guidance for establishing acceptance criteria. When setting the acceptance criteria for any method transfer, it is recommended that the validation documents, the specification limits, and any available stability data be reviewed. The review of stability data, if available, from the lot to be used in the transfer as well as from other lots is important as it may indicate sources of variability other than the method (e.g. product variability) that must be accounted for when establishing the replicate pattern and acceptance criteria.
At the conclusion of the testing, the TL should review the data to ensure that the results not only meet the criteria set in the transfer plan, but they should also ensure that there is not a significant bias in the data (e.g. the RL passes but their results are consistently higher, or lower, and at the criteria limit). If a bias does present itself, it is strongly recommended that an investigation be conducted to ensure that potential long-term issues are minimized.
Replicate and Criteria Setting When Transferring Assay Methods
1. Replicates
For assay testing of a composite sample of a drug product, assay testing of an API, or unit dose testing of reconstituted lyophiles and liquids, a minimum of 6 individual sample preparations/lot should be analysed. For unit dose testing of POS formulations, solid oral drug products, transdermal patches and inhalations packaged in pre-metered dosage units, an analysis of 10 individual sample preparations/lot is recommended as a minimum.
If the content uniformity method utilizes the same procedure as described in the assay, then a site may be qualified for content uniformity testing by successfully completing the transfer of the assay method. This rationale should be captured in the transfer documentation. If the Content Uniformity test is different from the assay, then 10 samples/lot should be analysed.
2. Inter-and Intra-Laboratory Acceptance Criteria
When transferring methods that determine the strength of an Active Pharmaceutical Ingredient (API) or Drug Product, the inter-laboratory agreement may be assessed by setting a criterion for determining the absolute or relative difference of the overall means between the laboratories, or by using statistical analysis.
If an absolute or relative difference approach is to be employed, use the guidance in Table 1 as a starting point. In addition to Table 1, review the available validation and stability data for different lots. If the time-point to time point variability for the selected lot and/or other stability lots is greater than the guidance provided herein or the results presented in the validation, then the inter-laboratory acceptance criteria may be adjusted accordingly.
When setting acceptance criteria for counter ions and/or preservative content, less stringent acceptance criteria may be applied.
Table 1 provides suggested %RSD criteria to be used depending on the specification limits. Review historical and validation data where possible to confirm the method has the capability to achieve this level of precision. Adjust the %RSD if the validation and historical data support it and justify the criteria in the Transfer Plan.
If the same analytical procedure is used to determine assay and content uniformity, then the transfer plan should reflect that if the RL meets the criteria for assay, then the RL will also be qualified to perform Content Uniformity. If two separate methods are used, then both the assay method and Content Uniformity methods should be transferred. For Content Uniformity testing, the results should be evaluated vs. the requirements in the appropriate compendia and compared to the results generated by the TL.
Replicate and Criteria Setting When Transferring Impurity Methods
Where lots are not available with impurities at a level suitable for meaningful interlaboratory comparison, strategies will be considered to adequately assess the ability of the RL to detect and quantify impurities. This may include spiking/recovery experiments, where appropriate. If a spiking experiment is conducted, and at least one specified impurity is available, recovery should be assessed at or between the Quantification Limit (QL) and Specification limit. For multiple strength drug products made from a common or similar blend, only one dosage strength needs to be spiked.
If the historical data shows levels of an impurity in the lot identified for the transfer below the QL but above the Limit of Detection (LOD), then it is recommended that the experimental plan include a requirement to perform multiple analyses (e.g. 3) of an un-spiked sample from the same lot and then use the average results from the un-spiked sample in the recovery calculations to correct for the presence of the impurity in the sample. In instances where impurities are not available, it may be acceptable to dilute a standard of the main component down to the QL and perform a study to ensure that the RL can achieve the appropriate levels of precision and sensitivity. Other strategies may be used as long as they are in the Transfer Plan.
1. Replicates
If the samples contain impurities at levels suitable for inter-laboratory comparison, a minimum of 3 sample preparations/lot should be analysed when conducting the method transfer. If the impurity method evaluates main band assay and impurities from the same sample injection or from a diluted sample, then the assay replicate pattern should be used for both. For TLC methods, only one sample replicate is required.
2. Inter-and Intra-Laboratory Acceptance Criteria
If the lot or lots contain impurities at levels suitable for meaningful inter-laboratory comparison, then the guidance provided in Table 2 should be reviewed and applied as appropriate. Where spiking experiments are required, the following guidance, in conjunction with available validation data, may be used to establish recovery ranges and precision requirements. For spiking levels between 0.1% and 1.0%, the recommended mean recovery should be 100% ± 25%. For levels at or below 0.1%, the mean recovery should be 100% ± 40%. Typically, precision should be set accordingly using either %RSD or the absolute difference between the highest and lowest recovery results. Where the difference between the highest and lowest recovery results is used, set the criteria at 1/4th of the recovery range (e.g. ≤ 20% High -Low for a recovery range of 140% -60%).
For the transfer of TLC methods, there is no intra-laboratory precision requirement. When conducting an inter-laboratory comparison, the criteria should be set such that the reported level of each impurity is consistent with the historical TL result. If a spiking experiment is to be conducted for a drug product, the impurity should be spiked at a level below and above the specification limit. For Active Pharmaceutical Ingredients, it may be possible to qualify a RL based on its ability to observe the lowest level standard band. It is also important to obtain a copy of the RL’s TLC plate for verification of the reported result.
Replicate and Criteria Setting When Transferring Residual Solvent Methods
Even when lots are available with residual solvent levels at or above the QL, it is recommended that a spiking experiment be conducted to accurately assess the RL’s ability to perform the testing. When conducting the experiment, a minimum of one solvent needs to be assessed to qualify the RL. Ideally, the solvent with the lowest specification limit should be used. If the historical data show levels of that solvent in the lot identified for the transfer, then it is recommended that the experimental plan
includes a requirement to perform multiple analyses (e.g. 3) of an un-spiked sample from the same lot and then use the average results from the un-spiked sample in the recovery calculations to correct for the presence of the solvent in the sample.
1. Replicates
At a minimum, three spiked samples should be analysed when transferring a residual solvent method. 2. Inter-and Intra-Laboratory Acceptance Criteria The following guidance, in conjunction with available validation data, may be used to establish recovery ranges and precision requirements. For spiking levels between 0.1% and 1.0%, the recommended mean recovery should be 100% ±
25%. Typically, precision should be set accordingly using either %RSD or the absolute difference between the highest and lowest recovery results. Where the difference between the highest and lowest recovery results is used, set the criteria at 1/4th of the recovery range (e.g. ≤ 20% High-Low for a recovery range of 140-60%).
Replicate and Criteria Setting When Transferring Identification Methods
When transferring an identity method, a single sample preparation may be used for the assessment; however, it is also possible to include multiple samples, including those that would fail the test. Typically, identification methods have acceptance criteria within the method itself; therefore, it is acceptable to qualify the RL by meeting this requirement. No inter-or intra-laboratory acceptance criteria are required; however, if the identity method involves an intricate technique (e.g. proteolytic map identities) it may be advisable to include an intra-laboratory criteria to ensure the RL can perform the test consistently.
Replicate and Criteria Setting When Transferring Dissolution Methods
1. Replicates
At a minimum, 12 units/lot are required to transfer a dissolution method. If review of historical data for the lots to be used in the transfer, as well as other lots, show product variability > 10% RSD, then it is recommended that the RL test 2x the minimum requirement for the specific dosage.
2. Inter-and Intra-Laboratory Acceptance Criteria
In general, each individual result must meet compendia requirements. Regardless of formulation, the interlaboratory acceptance criteria should be set using the average results at the final test time indicated in the product’s specifications. A typical inter-laboratory acceptance criterion is ± 6% (absolute) based on a long term precision (RSD) of 10%; however, historical data, such as stability data from multiple time points, should be reviewed. If the data supports another criterion (e.g. time point to time point differences are >6%), then the more appropriate criterion should be used with suitable justification provided in the Transfer Plan.
An extended release formulation will have multiple test time points and the drug release ranges are listed in the specification for the earlier test times. In these cases, it is difficult to set inter-and intra-laboratory acceptance criteria. It is possible to set a criterion that the RL results must be within the ranges provided in the specification.
It is strongly recommended that the data be reviewed and compared to the TL data to ensure there are no biases evident in dissolution testing. If there is a suspected bias in the data, further investigation is warranted.
Replicate and Criteria Setting When Transferring Particle Size (non-Sieve) Methods
1. Replicates
It is recommended that at least 6 samples be analyzed for a given lot.
2. Inter-and Intra-Laboratory Acceptance Criteria
The following guidance, in conjunction with available validation data, may be used to establish appropriate inter-and intra-laboratory requirements. Typically, the inter-laboratory acceptance criteria should be _10% relative difference between the TL and RL at each specification point. RSDs should also be _10%.