Scientific Validity Keys for Supplement GMPs

A core concept across GMPs for many industries is scientific validity, and this is also one of the necessary requirements of the dietary supplement GMPs. For example, the purpose of an ingredient specification is to disclose scientifically valid methods and results for the tests, and these methods and results are used to verify the quality and identity of the material being sold.   

Scientific validity means that tests must be suitable for what they are intended to measure. In a rapidly evolving industry, scientific validity is a core principle guiding our efforts to ascertain the identity, safety, and label claims of the material that millions of people take to support their health. 

To apply scientific principles to the measurement means that we develop a foundation of confidence in test results that accumulates only through repeated testing of viable hypotheses. During the process, we understand that like with many scientific measurements, sources of error exist which tend to increase with complexity. For example, complex samples containing thousands of chemical constituents (e.g., botanical extracts), and instrumentation methods that have a lot of variables all contribute to our bank of known unknown and unknown unknowns.

Todays analytical technology to measure analytes in complex mixtures is way ahead of the not-too-distant past, but now we understand a mitigating factor: that with greater power and resolution comes an increasing number of factors that may cause test results to be inaccurate or imprecise.

For example, it can be difficult to account for systematic error associated with dirty HPLC columns or non-optimal HPLC conditions. Inaccurate purity data on reference standards (due to either inaccurate standard purity values, or degradation during storage) is also a common source of error when we are trying to figure out the real result of a test. Another source of error arises from the calculation of the results--moisture can account for a certain amount of the measured weight of both samples and standardsbut in different amounts that can affect the end result.

Other sources of error in HPLC can be chalked up to incomplete dissolution during sample prep. For example, it is a common assumption that when a sample dissolves during HPLC sample prep, then it is fully ionized and thus is not strongly bonded to any solid particles (which of course would get caught on the filter and not pass into the detector).

If both standard and sample dissolve to the same degree, no problem!  But (unknown unknown) error creeps in when the reference standard dissolves to a different degree than the sample. 

Compound this with the high heat and pressure applied by an analytical instrument like HPLC, and chemical reactions can happen in the complex sample to degrade what you are measuring, all while your reference standard survives nicely to the detector. This scenario can also happen the other way around, where the matrix stabilizes the analyte better than the standard solution under the HPLC conditions. 

Generally speaking, it is difficult to predict analytical uncertainty, so it is important to develop and validate methods and standards to reduce as much as possible the impact of major sources of error. At the end, injecting a biological sample into a big, expensive box results in something of an educated guess, especially for labs that may only sporadically test a given sample with a given test.

Boring stuff, sure, and an estimated 6 percent of readers have made it this farso congratulations because here is the payoff: scientific validity in QC testing becomes a relevant subject only when the cost of mistakes is quantified. Some issues that can occur include the approval of out-of-spec materialand the rejection of good material. Both of these scenarios cause increased expenses for retesting and re-sourcing.  

Here is a short list of some practices that QC units can perform to achieve scientific validity as per GMPs:

--Review your labs methods for their suitability for the intended purpose. There are good independent labs out there that will share method information, and answer your questions. Always ask whether the sample is being tested in triplicate and request to receive the individual values.

--Review the documentation on the reference standard, specifically the methods and results of the testing used to determine its purity. When was the standard made, when was its purity last tested, and how was it stored in between? 

--Blind your sample so your lab does not know what value to expect.

--Test control samples (samples that do not contain the suspected analyte, OR samples that you previously sent to the same lab).

--Work with labs that can demonstrate having worked to some basic degree to optimize/validate the method.

Sounds like costly work, but not so much when put in perspective of the potential costs. With transparency among customer, supplier, and lab together, a little teamwork goes a long way to reduce the costs and maximize the benefits of quality systems.