Putting Standards to the Test
With FDA's recent emphasis on validating testing methods, industry is keen on developing standards that could become adopted by the agency.
by Erik Swain, Senior Editor
Medical device package testing, both for seal strength and overall integrity, has been until recently a very uneven practice. Device companies had their internal procedures, but there was very little industrywide guidance on how to do the tests properly, and different labs could produce different results for the same test.
The 1997 publication of ISO 11607, "Packaging for Terminally Sterilized Medical Devices," identified a number of appropriate tests, but it did not show how to conduct them. As a result, coming up with standards has fallen to other organizations, primarily the American Society for Testing and Materials (ASTM) F02.60 subcommittee on medical device packaging.
Peel testing of packaging can involve extremely low forces, often just a few grams, so accuracy of the force measurement is a key consideration. Photo courtesy of instron corp.
Those efforts are particularly significant now because FDA has become more aggressive in asking device manufacturers for proof that a given test method has been validated in their own laboratories, industry sources say.
ISO 11607 has been recognized by FDA as a consensus standard. If the test methods referred to in that document can also be standardized, chances are good that they too can achieve consensus-standard recognition. And that would mean test method validation could become a whole lot less complicated.
"It is the intent of this subcommittee to submit to FDA these test methods that are referred to in ISO 11607 for consideration to be recognized as consensus standards," says Hal Miller, director of packaging technology at Johnson & Johnson (New Brunswick, NJ) and chairman of the ASTM F02.60 subcommittee. "Many companies are now asked for test method validation data, but if these are recognized as consensus standards, that might not be the case anymore. Then companies could basically self-certify that they comply with the test methods because they follow the ASTM standards."
The process to achieve consensus will not be simple. "We still need to clarify that if we did have standardized test methods, what steps would have to be repeated for each manufacturer that uses them," says John Spitzley, associate fellow of packaging at Medtronic, Inc. (Minneapolis) and a member of the subcommittee. "For example, does sensitivity have to be repeated once it's established? These are the kinds of issues we're trying to nail down."
"Customers are keenly interested in seeing us get this done," says Stephen H. Franks, executive vice president and principal at T. M. Electronics (Worcester, MA). "They don't want to have to validate test methods along with a product."
Standards for visual inspection (ASTM F1886) and dye penetration (F1929) have already been published. One for accelerated aging (F1980) just came out a few months ago (see this month's Regulatory Focus). With most of the integrity test methods published or almost done, the subcommittee's focus this year has been on seal-strength tests, Miller says.
One such test is the peel-adhesion test. Patrick Nolan, vice president and technical director of testing lab DDL Inc. (Minneapolis), has written a draft, and now feasibility studies are being conducted. This method would be an alternative to ASTM F88, "Standard Test Method for Seal Strength of Flexible Barrier Materials."
"This standard would evaluate a much longer section of a package," Nolan says. "The F88 method is looking at only a 1-in. cross section of an entire seal on a package. With a peel-adhesion test, you are looking at the full width of a package and can see the strength and continuity of the full length of the seal. That is a good indicator, especially for consistency. The method can determine within 0.1 in. where the weak spots are. Then you can make corrections. There could be a cool spot on the platen or a warp in the material. Now you can be more specific when you look for weaknesses."
The F88 standard is undergoing revisions to further address lab-to-lab variations, says Marie Tkacik, engineering manager at Tolas Health Care Packaging (Feasterville, PA). She chairs a task group of the Sterile Packaging Manufacturers' Council (SPMC), a subset of the Flexible Packaging Association, that is working with ASTM as members of the F02.60 and F02.30 subcommittees.
"There is an abundance of information that sheds light on the result of using different methods and techniques," she says. "A big source of variation is caused by the handling of the tail, either unsupported or supported at 180 and 90 degrees. It is not unusual for two labs to get different results in their testing, and it often boils down to the method or technique being used. Drawings of different methods for handling tail angle of peel are now a part of the current F88. Other differences are the cut size of the sample, the crosshead speed, and reporting of data at either peak values or some average over the length of the peel. This last one, calculating average values, is an important issue to address since there can be variations of strength within a seal. This subject is currently being reviewed for addition to F88 through the F02.30 Test Method subcommittee."
Another project is a standard for a burst test with restraining plates. There is already an unrestrained burst test standard, ASTM F1140.
"Unrestrained burst testing is not a particularly good way to detect the weakest point in a peelable pouch," explains Neil Lorimer, senior quality engineer at Rexam Medical Packaging (Mundelein, IL), who is working on the standard, which could be ready in the spring. "The stresses are not imparted evenly. By putting the pouch within restraining plates, you can equalize stresses around the bag, and you have a higher probability of detecting the weakest point and getting a value for the seal strength at that point."
The standard, Lorimer says, will be similar to F1140 but will have its own number. "It describes how to do the test and provides more information on differences between test methods. The standard is functionally similar, with a higher probability of determining the weakest point of the pouch. It shows how the restraining plates are used. There is further information in the appendices for plate gap setting recommendations and considerations, factors influencing test results between laboratories, and calculations of stress functions for safety when designing restraining plates."
F1140 itself is undergoing revisions, overseen by Dave Bohn, quality systems–package engineering group, Medtronic. Actually, it is a re-revision, reinstating guidance for blister packages that was taken out in 1996 with the idea of becoming its own standard.
"This revision standardizes wording and pulls blister packages back in," Bohn says. "Blister packaging testing was removed to develop a separate standard for unrestrained and restrained tests; however, an analysis by Michigan State University showed that the restrained test did not reduce variation from the test and did not minimize the effects caused by packaging size, shape, and geometry. With work proceeding on the restrained test for pouches in a separate standard, it was determined to leave F1140 an unrestrained test and to add blisters back in."
Further harmonization could be achieved if a correlation were established between results from the F88 and F1140 methods. This was not believed possible until Stephen H. Franks and Donald S. Barcan, president of Donbar Industries Inc. (DBI; Long Valley, NJ), published a study in the August 1999 issue of Medical Device & Diagnostic Industry, a sister publication to PMP News. They found that a correlation could be established when the process sealing parameters were varied, but the results held only for a specific material and test configuration and could not be universally applied.
Some companies, such as Rexam, have also developed similar correlation models between the F88 and the restrained burst testing methods. However, these have to be applied to specific package types and test methods as they have not found that a universal correlation model can be applied.
There is still some work going on with integrity test methods. Franks is working on a standard for pressure-decay testing, in which the engineer inflates a nonporous package and measures its pressure loss. Precision and bias testing for the standard will be done in the next few months. "The pressure-decay test method is more sensitive than visual tests available at the moment," he says.
Also in the works is a standard for helium leak testing, to be used with True Technology Inc.'s (Newton, MA) SealCheck 210, a nondestructive testing system for porous packages that quantifies leaks. The method, which introduces helium into the package and uses a sampling probe to scan the seal to determine the concentration of helium that has escaped, has received increased interest from industry, but not for the reasons expected, says Kenneth Andersen, True Technology's vice president of sales and marketing.
"The original reason for a nondestructive test was that you could test every part," he says. "Some companies want to use it that way, but not many. Others want to use it to validate their production. Their reason is more that the measurement is so good, not that it's nondestructive. The dye test destroyed your ability to reproduce failures. There was no evidence to say half the lot was good or bad. With our method, you can retest the same package multiple times to weed out false positives. You can also have sample packages held for evaluation. You can take them to a production lab. If there is a suspect lot, you can run all 1000 through the test. Another advantage is the validation process. You can run 400 samples and get quantifiable information. That's almost more important than detecting."
And Curt Larsen, packaging consultant for DuPont Medical Packaging Systems (Wilmington, DE), is revising ASTM F1585, "Guide for Integrity Testing of Porous Barrier Medical Packages." It is essentially "a description of what is available for integrity testing of medical device packages," he says. "We have to update and include methods that have been balloted and passed, eliminate references that aren't needed anymore, and add pending methods."
One of the original factors behind standardizing physical tests for package integrity was to persuade FDA to stop requiring microbial challenge tests, which many in industry believe are inaccurate and too costly. That transformation appears to have taken place.
Microbial challenge tests, says Medtronic's Spitzley, are "more the exception than the rule. The last few conversations I had [with authorities] indicate it's not an issue any more."
That is also the conclusion that the Burron OEM Division of B. Braun Medical Inc. (Allentown, PA) has reached, says John Grimm, director of OEM product development. "We do not do microbial bombardment as a rule," he says. "Some of our customers still do require microbial bombardment as a function of validation. Some prefer burst testing, others peel testing. We have to be willing to work with customers to serve their needs. If you show consistency in the way you approach these tests in your submissions, FDA seems to be fine with that."
Validation is also an area where progress is being made. M. Louis Arin, materials engineer at FDA's Winchester Engineering and Analytical Center (Boston), is working on an ASTM standard about test method validation. Suppliers say they realize the importance of the issue and are trying to help with validation as much as they can.
"One thing FDA seems to be concerned about is the software used in the system," says Dick McManuis, product manager, Instron Corp. (Canton, MA). "If software is custom or low-volume, that seems to be of more concern than if it is an off-the-shelf system. We provide our customers with information about our software QA procedures and our procedures for tracking revisions of the software. Our policy is to maintain old copies of the software. If there is a problem, we can recreate it because we have older versions safely stored. That seems to go a long way toward satisfying FDA."
"You have to show you can verify your technique," adds David Morris, president of ITI Qualitek (Billerica, MA), which makes nondestructive leak testing equipment. "We ask for samples of a customer's product, cut a hole that we know to be a certain size, and verify our equipment can pick it up."
CHOOSING TEST METHODS
Then there is the issue of what tests are the most appropriate to use for particular applications. There is no current guidance from FDA and only F1585 from ASTM. Companies have to trust the judgment of their packaging engineers or consultants. "If a customer I'm working with is interested in determining only the leaks in a seal, I direct them to dye penetration," Nolan notes. "If they want more of a whole package leak test, we are more likely to use a vacuum leak test or a bubble emission test. Those evaluate for pinholes and seal channel leaks."
SPMC is trying to address ways to choose test methods. "We're putting a decision-tree kind of matrix together for the design and evaluation of the package to provide guidance when choosing which tests to perform in R&D and at the time of manufacture," says Jackie Daly Johnson, vice president and chief operating officer, Beacon Converters Inc. (Saddle Brook, NJ), a member of the task force. "It is not primarily for closed-package integrity, but it will reference ASTM standards on testing finished packages. It will show which test methods are appropriate at what time of development of the package. There's not really one place where you can get a list of what is most appropriately employed."
With all of the different options available, one thing that can help, says William Fridman, president of Mark-10 Corp. (Hicksville, NY), which makes components for force tests, is "a universal type of a setup. You set up several different tests on the same piece of equipment rather than have dedicated testers—often dedicated testers are not as cost-effective."
NEW TECHNOLOGY, NEW STANDARDS
As ASTM subcommittee members finish writing standards for existing test methods, the next step will be to identify new testing technologies and come up with standards for them. For example, Cincinnati Test Systems Inc. (Village of Cleves, OH) recently introduced a leak tester for liquid-filled packages, and it has found a use in the medical industry for contact lenses.
"You put the package in a chamber and evacuate the chamber around the package," says Gary Grebe, marketing manager. "That creates a force inside the package, and it would force the liquid through the leak to the outside. Liquid in a vacuum will boil off. As it boils off, it becomes gas. The gas lowers the vacuum level. It monitors how much vacuum is lost and the smoothness of the pressure curve to see if there was a burst of pressure."
There is no standard for this method yet. Maybe someday there will be.
"We'll work on anything that anyone wants us to do," J&J's Miller says. "We are driven by whatever ideas come up from the membership."