Introduction
This guidance addresses the application of matrices and bracketing strategies to process validation (PV).
Bracketing and matrixing allow a ‘most appropriate challenge’ condition to be defined for a process or drug product family (the same drug product with different dosage strengths). This risk-based approach can allow the validation to be focused on the most challenging circumstances, or “worst cases.” Use of this approach can provide a significant benefit to reduce the overall validation effort. Bracketing or matrixing may be used for validation of Drug Product, Active Pharmaceutical Ingredient and Packaging processes when this approach can be justified.
Recommendations & Rationale
Bracketing is the assessment of a single parameter or variable by identifying the edge(s) of the range of conditions for the parameter or variable and assessing these during validation to span the possible range of that parameter/variable. Bracketing can be applied to process parameters, multiple pieces of identical equipment, and/or different size considerations for the same product, for instance.
Matrixing involves the assessment of the effect of more than one parameter or variable by using a multi-dimensional matrix to identify the “worst case” conditions for a combination of parameters/variables. These conditions are used during validation of the process, rather than validating all possible combinations.
Matrixing is typically used when there are significant similarities between products in a drug product family (e.g., same product different strengths in the manufacturing stage or different products with similar container closure in the packaging stage).
Examples of variables that might be assessed by bracketing and matrixing include, but are not limited, to:
- Batch size;
- DP dosage strength;
- Identical equipment (e.g. where setup and operating conditions are the same);
- Product packaging, such as where only a minor adjustment in packaging parameters is required to accommodate different bottle heights or dosage counts.
Matrixing across different products may be applied to the packaging validation of the final dosage form, for example to evaluate the packaging of different products in a common packaging presentation. As with other uses of bracketing and matrixing, the risk of using this strategy for the potential products encompassed by the matrixing plan should be considered, documented and approved.
The use of bracketing/matrixing for the validation of a manufacturing process across different products should be approached with caution because of the risk of overlooking other possible affects of the change. Use of this type of bracketing/matrixing requires a good understanding of the processes involved and the risks being assumed. For example, in the evaluation of a change of a critical material for different products, the excipient interactions, critical process parameters and critical quality attributes (CQAs) are not necessarily the same for each product. The effect of the change in the CQAs may be different for each product. A product sensitive to the change may experience a failure in a CQA (e.g. dissolution) while in a case of a product not sensitive to the change, it may experience no effect at all in its CQAs.
To obtain the maximum benefit with minimum risk from bracketing and matrices, it is necessary to have a well-developed understanding of the impact of critical process parameters on critical quality attributes. There should be a documented and justified rationale that explains why one set of test conditions (e.g., manufacturing process, product presentation, etc.) is representative of one or more related test conditions. Typically, the rationale is addressed by selecting parameters and/or products that represent the edges of a range or “worst case” of allowable conditions. The rationale and justification for the bracketing/matrixing strategy to be used in validating a process should be provided in the validation protocol, or in another document referenced in the protocol.
Depending on the circumstances, prospective and concurrent validation approaches may be used for validating a process using bracketing or matrixing. If a concurrent approach is used, an interim report provides a summary of the results obtained for a product batch, in order to justify the validation and release of one of the product presentations within the bracket/matrix. This approach may also assist in approving the manufacturing and/or release of additional batches of a particular presentation. At the completion of the validation, the validation report will address all batches.
At present, some regulatory authorities may not accept the use of bracketing or matrixing for validation. Japan, for example, currently requires that all combinations be validated.
The following examples include possible matrixing/bracketing approaches. There may be other acceptable approaches.
Example 1: Bracketing Batch size ranges for API
A range of batch sizes may be produced for an API in a particular piece of equipment (such as a dryer or a mill). The worst case situation(s) could be either smallest or largest expected batch size, or both of these, so bracketing the variability in batch size by validating the extremes in batch size (smallest and largest) could be used for validation of a range of batch sizes. Including examination of product homogeneity may often be essential in supporting this approach. Depending on the individual circumstances, it may be possible to extend the validated range after the initial three-batch validation is completed by validating a batch size that is not worst case with only one batch, if this is appropriately justified.
Example 2: Matrixing of blend ratios (Product name used are common in the market)
A new product, Amlodipine \ Statin Film-Coated Tablets, is to be transferred from its current manufacturing location to another site. Comparison of the four Amlodipine \ Statin ratios (see first column) and six dosage strength combinations (second column) being transferred is provided in Table 1.
ANSWERS for Example 2:
Here are two options that could be used (others are possible):
Option 1: Matrixing based on the blend ratios.
“Worse case” selections are the most extreme blend ratios (1:1 and 1:8).
It is common to include three lots at each extreme, and one lot each of intermediate dosages. For this example, this would mean validating three lots each of the 1:1 and 1:8 blend ratios, and one lot each of the other four intermediate strengths.
This option is preferred to show reproducibility for the blend. A disadvantage is that it will not as strongly show reproducibility of compressibility of the mixture.
Option 2: Matrixing based on dosage strength.
“Worst case” selections are the lowest and highest dosage strengths (5/10 and 10/40). This would typically mean validating three lots each of the 5/10 and 10/40 dosage strengths, and one lot each of the other dosage strengths (10/10, 10/20, 5/20, and 5/40). This option will better show reproducibility in compressibility of the mixtures since it includes the smallest tablet size with the lowest weight in the tablet of both active ingredients (which has the greatest chance of segregation on the press due to longest run and low active weight) and then the largest tablet with the highest weight of both active ingredients in the tablet (which has the greatest chance of poor press performance due to presence of active).
However, this option does not show reproducibility directly on the blend as well as Option 1.
Factors that might affect the matrixing include, but are not limited to:
- Nature of the formulation;
- Tablet shape and dimensions (because of potential impact on machine settings and tablet hardness)
- Compressibility
- Dissolution characteristics
- Batch size
- Equipment used; and
- Similarity of process parameters across the various blends.
If two different models of compressing machines will be used during the initial validation, the extent of validation will depend on the history of these machines.
If similar parameters can be used and experience with other products show the machines perform similarly, it may be enough to make lots on one machine without requiring additional validation lots on the other machine. If the history is that the machines perform differently, then it may be necessary to run a three-lot validation study on one machine and run the worst case (for compression) in triplicate on the second machine together with one lot each of all other strengths to confirm the parameters. This assumes that there has been development or qualification work on the machine to establish the parameters.
If machines of the same design are used, it is good practice for the validation study to include at least one lot on each of the machines, but it is not a requirement.
Evaluating any change should include assessment of the risk to product quality arising from the proposed change. Revalidation may not be necessary if the proposed change poses little risk to product quality. For this example, it is assumed that there is an appreciable risk to product quality because magnesium stearate is a critical component that functions in the mixture as a lubricant and prevents material from sticking in the press. It also impacts the dissolution properties of the formulation because of its lipophilicity. Validation of the source change is therefore considered necessary in this case because of its critical nature to the formulation. Given experience with the product’s blend ratios and dosage strengths and the manufacturing process, the matrixing strategy described for validation (see answer A in this example) can be used.
Example 3: Bracketing and Matrixing for a DP packaging process
A new product will be transferred from one manufacturing site to another. The product is a capsule dosage form. A common blend is used to prepare five different dosage strengths. The packaging presentations and capsule sizes are shown in Table 2:
ANSWERS for Example 3
Process Validation Strategy:
This example is one of matrixing because we have a solid dosage form compounded in various dosages with similar overall shape and size using a common blend. The challenge in the encapsulation stage is to be able to fill consistently the different capsule sizes at the specified weight.
The “worst case” extremes are 120 mg (lowest weight) and 420 mg (highest weight). A typical strategy would be to include three lots each of the worst case dosage weights, and one lot each of the dosages that fall between these extremes. Capsule size and capsule fill weight is dependent on dosage and thus assessment of this variable is already included in the matrixing of the range of dosage weights.
Packaging Validation Strategy:
This example is one of matrixing of the different strengths and capsule sizes. We have solid dosage units with similar capsule size packaged with the same tools in the same type of container/closure (for those that use the same components). It is also an example of bracketing of packaging components for those dosages that have similar components (bottle/closure).
All the bottles are filled with 30 capsules. In terms of capsules sizes, the worst cases are the smallest (size 2) and largest (00), but the intermediate size (0) should also be evaluated. Similarly for bottle sizes, the 80 cc, 90 cc and 120 cc bottles should be considered. Therefore, the packaging validation strategy will include:
Dosages to be Bracketed/Matrix in combination with packaging:
- 80 cc bottle: 3 validation runs of 120 mg
- 90 cc bottle: 3 validation runs of 240 mg
- 120 cc: 3 validation runs of 420 mg
If the process will be new to the receiving site, a prospective approach is typically required for each of the “worst case” dosage weights included in the validation study.