Originally published July, 1996
Ashweni Sahni and Curtis L. Larsen
Process validation is required to meet the current good manufacturing practices (GMP) regulation for medical devices. Since 1987, when FDA published its Guidelines on General Principles of Process Validation, publications and conferences on
the subject have helped manufacturers to better understand FDA's expectations.1 Yet there is some confusion about what will satisfy FDA inspectors.
FDA defines process validation as "establishing documented evidence which provides a high degree of assurance that a specific process will consistently produce a product meeting its predetermined specifications and quality characteristics." To meet this definition, then, pass/fail criteria should be established before the testing, the test results should show that the tested process will yield acceptable output in a consistent manner (not just once or twice, but whenever it is run), and documented evidence should prove that the test results and conclusions are correct. In other words, a process validation plan and a validation report, proving that the process is validated, are necessary.
While manufacturers need to ensure that all process outputs are of acceptable quality, not all processes need to be validated--only special ones whose results cannot be completely verified by inspection or testing. Typically a department or company will develop a list of special processes.
At SIMS Deltec (St. Paul, MN), such a list was developed. The following case study describes the validation of one special process, the sterile packaging of single-piece remote reservoir adapters (RRAs; see Figure 1).
THE CASE STUDY
For this packaging process, the process validation plan had sections describing the process, objective, reference documents, test samples, sample preparation, test or inspection before validation tests, test or inspection required for validation, pass/fail criteria, and test report.
The validation objective was to provide a test protocol for validating the assembly process and the performance of the package design for the new line of RRAs. Not only was the process to be validated, but also a design qualification was to be performed. This qualification would evaluate how the unit-for-sale packaging design, or the shelf container and shipping configuration, would protect the device from damage that could occur during customary conditions of processing, handling, storage, and movement through the distribution environment.
Before validation could begin, some preliminary tests had to be made. A process capability study was performed to establish the window within which the process would have to operate to produce acceptable results. Acceptable results in this case were seals visually acceptable in both width and continuity, and sufficient seal strength at all operating conditions. This established the minimum, nominal, and maximum process variables for the validation protocol and the operating conditions under which the design qualification samples were produced.
Each device that was to go into a sample package was tested before validation to make sure it was functionally acceptable. It was visually inspected to make sure there were no abnormalities such as cracks, abrasions, or other cosmetic defects. Then each package was visually inspected to make sure it met the requirements of referenced documents, which included the external and internal standards of the device and its packaging. Results were recorded on a data collection sheet like the one shown in Table I.
In this study, three groups of 60 samples were tested. One group had been produced at the low, one at the nominal, and one at the high process setting to ensure that packages built at the extremes of the process window would be acceptable. The samples, RRAs in sterile blister packages, were tested after undergoing three sterilization cycles. All groups were subjected to identical physical and climatic stress testing.
Each group of 60 samples comprised four subgroups, A, B, C, and D, which represented the approximate distribution of the various types of products (see Table II). The samples were prepared according to a standard bill of materials (BOM) and manufacturing process. All labels were applied, including the instructions for use. The device samples were functional and accepted after the final inspection. (If the actual labeling is not available, equivalent labels can be used with prior approval from the responsible engineer. Any label substitution should be recorded in the report.)
Each shelf box of 12 blister samples was marked with a lot number, and the samples were prepared with the low, normal, and high seal specifications, challenging the parameters of the process. Three trials were conducted, but the results of only one trial will be described here.
The shelf boxes were run through test conditions according to ASTM test method D 4169-90, "Performance Testing of Shipping Containers and Systems," distribution cycle 13, assurance level 1. This test method lists the needed equipment and tells how to perform the several tests that represent this particular distribution cycle. The ASTM test includes product mass, which is used to calculate the compression forces to which the samples will be subjected. The test also covers package orientation and stacking height.
After they were subjected to the stresses of ASTM D 4169-90, the shelf containers and sterile barrier packages were visually examined, and the device was examined for visual as well as functional requirements. The results were recorded for generating the process validation report (see Table III).
The test report section of the validation plan not only described who would write the validation report, and when, it also specified that the report include observations by test operators, lists of nonconforming devices, details of packaging samples, and the final conclusion as to whether the pass/fail criteria were met. Any deviations or discrepancies in testing were also to be reported.
PACKAGE AND DEVICE EXAMINATION
The pass/fail criterion for the testing was that there would be 95% confidence that the samples would exhibit a minimum of 95% probability of acceptance. This meant that there must be no failures among the 60
The visual examination revealed that all boxes had crushed corners from the shock and vibration testing and dirty outside surfaces from the vibration test equipment. This was expected and acceptable. Seven had creases from compression testing. This was also expected and acceptable. Six sterile blister samples had dents, but the dents did not compromise sterile barrier integrity. Two samples had wrinkled lid stock, which also did not compromise the barrier. All samples passed the seal width, dye, and seal strength test. In addition, there were no test-induced anomalies in the device components, and all the devices in the packages met functional requirements.
Based on statistical analysis of the test results, the packaging process met the criterion of 95% confidence of a 95% minimum probability of acceptance of products from the process. The process was thus considered validated and the design qualified.
The experience of validating the RRA packaging process showed that some
precautions are necessary for successful
The finished product that is sold to the customer should be used to determine the configuration of the sample for testing. In the case study, the unit tested was not only a blister package, but 12 blister packages housed in a corrugated, die-cut shelf box that was qualified to be shipped without an overpack. If the finished product to be tested comes in a more fragile shelf container such as a folding carton, the engineer should configure the test samples as though they were to be introduced into interstate commerce. This leads to additional concerns, such as whether the shipping department will package the product in the same way that it was packagrequire resterilization.
The design must be qualified at the extremes of the process window. Making duplicate samples at the low, nominal, and high process parameters will ensure this.
Samples must be built using released components. Tooling used in the build and sample analysis must also be documented. Equipment used in the build and sample analysis must be calibrated and in proper working order, and operators must be trained.
Strength and package integrity should not be co unit tested was not only a blister package, but 12 blister packages housed in a corrugated, die-cut shelf box that was qualified to be shipped without an overpack. If the finished product to be tested comes in a more fragile shelf container such as a foldnfused; they are very different. Sterile package strength generally refers to the results of tensile, burst, or creep tests. Only through analyses using such tools as visual inspection or leak testing can one evaluate integrity.
An environmental stress screening should be performed on the device before package design. Such a screening will establish the levels of physical and climatic stress that will cause damage to the package contents, which will help the engineer design a successful package.
The customer's requirements must be considered when the pass/fail criteria are established. In this case study, there were two basic ones: the device was to exhibit no posttest damage, and the sterile barrier package was to maintain its integrity. Posttest stress damage to shelf boxes and blister packages was expected and allowed as long as the devices met these two requirements.
Process validation not only ensures compliance to GMPs, but helps to improve processes and ensure consistently high-quality output. Validation, which also reduces costs, is good for business.
1. Guidelines on General Principles of Process Validation, Rockville, MD, FDA, May 1987.
Ashweni Sahni is director of quality systems and Curtis L. Larsen is principal packaging engineer for SIMS Deltec (St. Paul, MN).