process validation Approaches

3. Approaches

Two basic approaches to the validation of the process itself exist (apart from the qualification of equipment used in production, the calibration of control and measurement instruments, the evaluation of environmental factors, etc.), namely the experimental approach and the approach based on the analysis of historical data.
The experimental approach, which is applicable to both prospective and concurrent validation, may involve:

• Extensive product testing.
• Simulation process trials.
• Challenge/worst case trials.
• Controls of process parameters (mostly physical).

One of the most practical forms of process validation, mainly for non-sterile products, is the final testing of the product to an extent greater than that required in routine quality control. It may involve extensive sampling, far beyond that called for in routine quality control and testing to normal quality control specifications, and often for certain parameters only. Thus, for instance, several hundred tablets per batch may be weighed to determine unit dose uniformity. The results are then treated statistically to verify the "normality" of the distribution, and to determine the standard deviation from the average weight. Confidence limits for individual results and for batch homogeneity are also estimated. Strong assurance is provided that samples taken at random will meet regulatory requirements if the confidence limits are well within compendial specifications.
Similarly, extensive sampling and testing may be performed with regard to any quality requirements. In addition, intermediate stages may be validated in the same way, e.g. dozens of samples may be assayed individually to validate mixing or granulation stages of low-dose tablet production by using the content uniformity test. Products (intermediate or final) may occasionally be tested for non-routine characteristics. Thus, subvisual particulate matter in parenteral preparations may be determined by means of electronic devices, or tablets/capsules tested for dissolution profile if such tests are not performed on every batch.
Simulation process trials are used mainly to validate the aseptic filling of parenteral products that cannot be terminally sterilized. This involves filling ampoules with culture media under normal conditions, followed by incubation and control of microbial growth. In the past, a level of contamination of less than 0.3% was considered to be acceptable; however, the current target level should not exceed 0.1%.
Challenge experiments are performed to determine the robustness of the process, i.e. its capacity to operate smoothly when parameters approach acceptable limits. The use of ranges of parameters for the quality of the starting materials in experimental batches may make it possible to estimate the extent to which the process is still capable of producing an end-product that meets the specifications.
The physical parameters of the process are monitored in normal production runs to obtain additional information on the process and its reliability. Extra temperature-sensitive devices installed in an autoclave or dry-heat sterilizer (in addition to probes used routinely) will permit an in-depth study of the heat distribution for several loads. Heat-penetration measurements are recommended for injectable products of higher viscosity or with volumes larger than 5 ml. A tableting press equipped with pressure-sensitive cells will be helpful in collecting statistical data on the uniformity of die-fill and therefore on mass uniformity.
In the approach based on the analysis of historical data, no experiments are performed in retrospective validation, but instead all available historical data concerning a number of batches are combined and jointly analysed. If production is proceeding smoothly during the period preceding validation, the data from in-process inspection and final testing of the product are combined and treated statistically. The results, including the outcome of process capability studies, trend analysis, etc., will indicate whether the process is under control or not.
Quality control charts may be used for retrospective validation. A total of 10-25 batches or more are used for this purpose, preferably processed over a period of no longer than 12 months, and reviewed together. (Batches rejected during routine quality control are not included in this review since they belong to a different "population", but failure investigations are performed separately.) A critical quality parameter of the end-product is selected, e.g. the assay value or potency, unit dose uniformity, disintegration time, or extent of dissolution. The analytical results for this parameter for the batches under review are extracted from past batch release documentation and pooled together, while the results from each batch are treated as subgroups. The grand average ("process average") and control limits are calculated and plotted on graphs or charts in accordance with the instructions given in numerous publications on control charts (see Bibliography, page 91).
A careful review of the charts will enable the reliability of the process to be estimated. A process may be considered reliable if the plotted data are within the control limits and the variability of individual results is stable or tends to decrease. Otherwise, an investigation and possibly an improvement are needed.¹

¹ It may be noted that, once control charts for past batches have been prepared, they become a powerful tool for prospective quality management. Data for new batches are plotted on the same charts and, for every result outside control limits, a reason, that is a new factor affecting the process, is sought and, when found, eliminated. By consistently applying this approach over a period of time the process may be considerably improved.

In addition, information on product-related problems is also analysed. The reliability of the process is demonstrated if, for a considerable time, there are no rejections, complaints, returns, unaccountable adverse reactions, etc. The process may be certified as retrospectively validated if the results of statistical analysis are positive and the absence of serious problems is documented. However, it should be emphasized that this approach is not applicable to the manufacture of sterile products.
Table 1
Example of priorities for a process validation programme

Type of process	Validation requirements
New	Every new process must be validated before approval for routine production
Existing: Processes designed to render a product sterile	All processes affecting sterility and manufacturing environment must be validated; the most important is the sterilization stage
Non-sterile production	Low-dose tablets and capsules containing highly active substances: validation of mixing and granulation in relation to content uniformity
	Other tablets and capsules: validation of tablet compressing and capsule filling in relation to uniformity of mass