GMPs and Method Validation: Setting the Record Straight

With the issuing of the final rule regarding cGMPs for dietary supplements by the FDA, there now appears to be considerable confusion about the analytical methods that companies (or their contract labs) are permitted to use in their QA operations. This seems to stem from the mention of "validated" methods in the 2003 proposed rule.

Both the proposed and the current rule state that analytical methods MUST be 'scientifically valid.' Use of "validated methods" is not explicitly mandated. However, Section 111.320(a) in the final rule states “You must verify that the lab examinations and testing methodologies are appropriate for their intended use” This can only be done by performing a series of experiments that demonstrates that the method "works" on your product(s) and in your hands. The name given to the series of experiments that you must perform to prove that your method is "fit for its intended purpose" is a validation study.

The lowest level of validation study is an in-house or single laboratory validation, but other levels are available (multi-laboratory study, etc.). FDA seems to have removed the term 'validated method' from the final rule because there are different levels of validation available and they did not wish to endorse any particular source of such methods. Use of a term (validated) for which there are several definitions would also have served to make the final rule ambiguous in this area. As in the section that mandates that companies establish their own specifications and then come up with methods for testing whether or not the specifications are met, FDA has apparently chosen to lay out the principle (Section 111.320(a)) and then leave it to the regulated company to figure out how to verify appropriateness for intended use.

In addition to the section mentioned above, other sections also essentially mandate that the lab validates the method in order to be compliant with the rule. In section 111.75, the rule requires companies to set specifications and to test to see if they are met. 111.95(b)(4) requires a company to make and keep documentation for why the results of appropriate tests or examinations for the product specifications selected under final Sec. 111.75(c)(1) ensure that the dietary supplement meets all product specifications.

These sections essentially mandate that you justify the rationale used to determine that your tests for specifications are “fit for their intended purpose.” The only way to demonstrate this is to: 1) define the purpose of the test(s) and 2) have undertaken a series of experiments to demonstrate that the test(s) are adequate for the purpose(s). This is a de facto ‘validation’ (a single lab validation will work).

As a colleague has pointed out, it would be ”convenient to be able to get your methods from a "Big Book O' Methods" and to require everyone to use the same method that you use, but this is not specifically necessary for GMP purposes….Simply pulling a method from a book on a shelf (even an "Official Methods" book) has never been enough, and will continue to not be enough.” Labs must be able to demonstrate that even an Official method is “Fit for Purpose” (i.e. that it works on their product and in their hands) which is accomplished by performing an in-house validation or performance verification.

From a review of the preamble to the final rule, it is apparent that FDA was trying to respond to the industry’s comments that products in our industry are not like drugs. Our products often consist of multiple components brought together in various forms that are entirely unique. For that reason, we are often not able to use “boiler plate” methods for quality control. There is a great difference between a protocol for testing acetaminophen in an OTC tablet and testing for ginsenosides in a 27-component formulation. FDA has left it up to the manufacturer to determine what specifications are needed and to find suitable testing methodology so that companies would have the freedom to use the best science for their particular product. A method may be great for one product and completely inappropriate for the same marker in another formulation.

This “gift” from the FDA is a two-edged sword. It gives manufacturers the freedom to develop specifications and test methods that meet the unique needs of their product matrix. But there is a cost associated with this freedom. Manufacturers must be able to justify their test methods in the context of raw material and finished product specifications, and document to FDA’s satisfaction that the method performs as advertised. There is no real ambiguity. “Process validation” is a term with which FDA is familiar, and there are a number of recognized protocols (including one published by FDA) for demonstrating that a method is adequate for its intended use, whatever that use may be. If you want to play it safe, you will have conducted a validation study compliant with AOAC single-laboratory validation guidelines or USP <1225> method validation guidelines or the validation guidelines in USP’s General Chapter on method validation. Another good idea is to publish your method in a peer-reviewed scientific journal where it will undergo a peer review process for comment and, if published, will add support for the choice of test method.

About ARC (Analytical Research Collective)
ARC is a group of volunteers whose mission is to collect, review, evaluate and disseminate scientific information on analytical methods. ARC is composed of Paula Brown from BCIT, Jana Hildreth from Blaze Science Industries LLC, James Neal-Kababick from Flora Research Laboratories, David Ji of Analytical Laboratories in Anaheim Inc., Mark C. Roman from Tampa Bay Analytical Research, Inc and Jeff Varelman of KML Laboratories Inc.

For more information on validation and ARC, please contact Jana Hildreth at 310-920-4517.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.