There are many areas where GMP compliance is directly related to analytical testing. The aspects of quality that relate to testing should be reflected in the specifications set by the manufacturer. 21 CFR Part 111.70 of the U.S. Dietary Supplement cGMP states that the manufacturer is responsible for establishing specifications for identity, purity, strength and composition related to incoming components, as well as limits on those types of contaminants that may adulterate or lead to adulteration of the finished product.
For a botanical-based dietary supplement, specifications would include, but are not limited to, identifying the correct plant material, the plant part, adherence to established limits for non-pathogens and assurance that pesticides and toxic elements are within an acceptable range. 21 CFR Part 111.75 compels the manufacturer to then verify through appropriate scientifically valid testing or examination that these specifications have been met. So keep in mind that the methods employed to ensure the specification is met are actually part of the specification (Constructing Product Specs, 2009-06).
The Good, the Bad, & the Ugly
If the law says you must have specifications and those specification include tests or examinations and those tests or examinations were not actually conducted (regardless of what you spent) you have broken the law. Since the Dateline NBC report that profiled a “Dry Lab” and the last “Quality Focus” column, “Lab Shopping by Price: A False Economy” many discussions have focused on the cost and necessity of analytical testing. Questions such as, “Why does testing cost so much?” and “Do we really need to do all these tests?” and “Which are the good labs?” have been floating around the dietary supplement community. The short answers are: “It doesn’t.” and “Yes.” and “No comment.”—although these answers probably won’t satisfy most manufacturers, so let me explain...
Focusing on Identity Testing
Focusing on just one specification—identity—let’s take a look at what is a reasonable expectation of cost and effort. The different approaches and challenges related to identity testing of plant-based dietary supplements have been addressed in many past columns, see “Identity Out of Focus” (April 2008), “A Riddle Wrapped in an Enigma, Shrouded in Mystery” (January 2010) and “The Perils of Paula Episode 3: Interview with a Botanist” (September 2009).
To recap, the simplest and most definitive identity tests are macroscopic (organoleptic) and rely on defined anatomical characteristics, color, smell, fracture and even taste. To use these tests the material being tested must have retained those diagnostic qualities, and the person conducting the tests must be trained to perform those examinations. Even microscopy requires that diagnostic features of the plant material
remain intact, and that the person conducting the examinations have experience (documented) and skill (also documented). When materials are highly processed these features disappear, thereby forcing the manufacturer to rely on chemical tests—which is where things get tricky.
Nature of Ingredient Dictates Test
It has been said repeatedly that the dietary supplement industry doesn’t need “fancy chemical tests,” which is often in reference to HPLC analysis, but if the material does not retain those diagnostic characteristics required for appropriate use of macroscopic/microscopic tests for identity, then manufacturers are left with little recourse other than to use a chemical-based analytical test. This is the place where many manufacturers find themselves and this is where the suggestion that identity testing should cost on average $60 (Federal Register (2007) 72(121):34931-34932) is simply not accurate.
The fact is the nature of the ingredient dictates the nature of the test(s) that must be employed to ensure the specification for identity is met. Just looking at (or smelling, tasting, etc.) a tea cut herb might be enough of an identity test if the person doing the looking is an expert (documented) with experience (also documented). If you discount the cost of having such a person in your company in the first place, then perhaps such a test would be considered way cheaper than the UPLC/MS/MS used in an attempt to identify a spray-dried alcohol extract of the same herb.
Economic Impact of GMP Compliance
If you managed to get through all of the preamble and the actual final GMP regulations themselves, you’re excused if you didn’t read XXIII (Paperwork Reduction Act Compliance) or Section XXIV. However, Section XXIV of the Federal Register notice in which the GMP final rule was published is a mandatory Economic Impact Analysis, which evaluates the costs/benefits of the new GMP rules. It was written by economists (with much economic modeling) and so is not an easy read, but there are several interesting parts worth exploring.
There are sections about labor costs and costs associated with implementing and maintaining GMP. For more information about the anticipated cost to implement GMP, by FDA’s reckoning, see: www.gpo.gov/fdsys/pkg/FR-2007-06-25/html/07-3039.htm. The analysis features a lot of tables and cost breakdowns, so selectively plucking small pieces of information risks obscuring the complexity of the economic impact evaluation. So I leave it to you readers to peruse the entirety of section XXIV at your leisure.
I will focus on one section, though—the discussion about the cost of testing. XXIV. “Analysis of Impacts” [Federal Register (2007) 72(121):34931-34932] states: “Testing costs vary according to frequency and complexity. The more frequently technicians perform tests, the lower are the costs per test. Many tests require sophisticated equipment… Using sophisticated equipment requires trained personnel. Even simple physical or organoleptic testing requires training or experienced personnel. The type of ingredient, compound or product can also affect the cost ...”.
Overall this is an accurate portrayal of the current state of affairs for dietary supplement testing. However, as FDA also acknowledges on page 34932, “We do not have sufficient information on the range of testing costs for botanical ingredients to determine if the average cost of testing is higher or lower than for other ingredients.”
With my own understanding of what is involved in conducting a chemical HPLC assay for a series of marker compounds in a botanical, I can state with confidence that the average cost per test is higher than $60. If for no other reason than the costs of chemical calibration standards, equipment and training required to run such an assay are more than $60.
The Numbers Don’t Always = Quality
The trend of focusing on a number to put on a product label, which comes from quantifying numerous markers as a measure of “strength” is a very expensive approach to testing, which may not actually have anything to do with quality. Testing the level of select marker(s) is one way of monitoring quality of incoming materials, and quality control in finished product manufacture. That being said, measuring a whole bunch of compounds in a botanical ingredient (with attendant costs) may not actually tell you anything about the identity, safety or effectiveness of the dietary supplement. If markers are used as a minimal definition of “good enough” for marketing purposes, then the costs of that possibly unnecessary testing fall outside this column, which is intended to focus on quality.
It is easy to fixate on costs, but it is important not to forget that the justification for Congress (which authorized GMP in DSHEA) and FDA in promulgating the GMP regulations in the first place, was firmly based on historical experience. As such, I draw attention to a different part of Section XXIV in which FDA produced a table (see page 29) of adverse outcomes attributable to manufacturing processes between 1990 and 1999, along with their associated negative health costs [Federal Register (2007) 72(121):34922-34923]. See Table 1.
Now I know it is not popular to re-hash unpleasant history, but while you can argue the appropriateness of including several items in that table (e.g., L-tryptophan/EMS), the fact is some events have resurfaced recently. One of the elements chosen by Chris Hansen (Dateline NBC) was lead. Why pick on lead? Well, there have been lead poisonings before. Selenium, too. The past is a very good place to start when considering what specifications should be included for a product and where to focus those testing dollars.
I admit that Table 1 isn’t comprehensive, in some cases it’s too specific (Do you sell plantain?), and yeah botulism is a stretch…but anywhere that there is a risk associated with NOT testing (such as pathogens or heavy metals) is a good place to spend dollars. While I understand the marketing push to engage in costly analysis of marker compounds, FDA has repeatedly said its focus is safety. So it is best to keep that in mind when creating a product specification and making decisions on the tests used to ensure that those specifications are met.
I’ll close by referring one last time to the topic of my last column and the Dateline NBC show that aired in March. Gently stated, they were about unbelievably low contract lab prices. So why is testing so expensive? It really isn’t, considering the inherent challenges and costs involved in testing and the potential costs of testing poorly or not testing at all.
Paula Brown has supported the natural health and food industry
for more than a decade by conducting applied research on product quality, safety and efficacy. Dr. Brown was appointed Fellow of the AOAC in 2009 after serving five years as General Referee, six years on the Dietary Supplement Task Force, participating on seven Expert Review Panels, and directing three collaborative studies. She sits on the American Botanical Council Advisory Committee, the Natural Health Products Program Advisory Committee for Health Canada, grant review committees for NCCAM, and Chairs NSF’s Joint Committee for Dietary Supplements. She is the director of the Natural Health & Food Products Research Group at the British Columbia Institute of Technology. She can be reached at email@example.com; Website: www.bcit.ca/appliedresearch/nrg.