Methods Validation: More Than Just Science

Methods Validation: More Than Just Science

by Mark Lange, Ph.D.

Methods arise as a result of applied research that involves a high degree of ingenuity and innovation. Companies are willing to financially invest in developing a test method because they understand that setting specifications is predicated upon the use of analytical procedures to monitor changes in product chemistry. Frequently, methods gain widespread use and begin to take a Darwinian path. Soon many variants of the same method are in use and validation and standardization are needed.

Validation of analytical methods for botanicals is a subject of considerable interest. Validation is the process of determining the suitability of methodology for providing useful analytical data. A method that is valid in one situation could be invalid in another.

The validation process verifies that the methodology is based on sound technical principles and that it has been written in a manner that is useable and practical. When little up front thought goes into selecting a test method, it can be unnecessarily expensive if providing more accuracy than necessary or completely useless if the accuracy of the method is unknown (Taylor JK, Anal Chem 55:600A-605A, 1983).

One group in this field, the Institute for Nutraceutical Advancement (INA), makes regular calls for methods in its Methods Validation Program (MVP). Method submissions are blinded as to their source and then examined by a scientific review panel. A single method is selected based upon availability of standards, supporting validation data, and ease of use in the lab.

A validation protocol is written for contractual agreement purposes and to bring uniformity to the process. The FDA Web site (www.fda.gov/cder/guidance) gives guidance to anyone wishing to draft a validation protocol. A minimum of three laboratories are contracted to validate a method. One lab is designated as a Primary Lab and it performs a full validation on the method, including linearity, limit of detection, precision, and recovery. The other labs are Confirming Labs and they gather inter-laboratory agreement data and perform ruggedness and robustness experiments. It has been the experience of INA that contracting labs and financially compensating them for their work is good policy. Contractual agreements allow a certain degree of control in matters of turn-around-time, expenditure limitation, and rejection of data due to improperly maintained instrumentation or not following the protocol.

Unique Problems

Several companies specializing in the area of natural products supply purified reference standards. Botanical reference standards generally are very expensive--several hundred dollars for 10 mg--and limited quantities are available. Often classes of compounds are used as markers and this multiplies the expense of a validation.

It has come to the attention of many botanical chemists that assay results can be dependent upon the source of the standard. Some of the most troublesome compounds are hyperforin, hypericin, eleutherosides and isoflavones. It is difficult to ascertain wherein the problem lies, but a degradation pathway leading to nearly indistinguishable homologues is a possibility. Chromatographic techniques are not detecting these degradants, yet changes in spectral properties from the mother compound lead to assay variance. I encourage suppliers of botanical reference standards to begin opening dialogue with their competitors so that these issues can be addressed.

The classical validation process necessitates confirmation of method accuracy. This can be accomplished in one of three ways:

Unfortunately, the unavailability of reference samples, the high cost of reference standards, and the lack of standard test methods precludes taking the classical approach to validating method accuracy. Therefore, it is important to prove that extraction efficiency has been maximized using the specified technique (sonication, refluxing, soxhlet) and solvent system. Once the preliminary experiment has been performed that explores the variables of technique vs. commonly used solvents, there is little more that can be done to prove accuracy of a botanical method. Spiking [on top of naturally occurring levels] with marker compounds can be done if the marker compounds are available and not cost prohibitive. Spiking will not mimic extraction from plant tissue, but it may give an indication of nonspecific binding of the analyte to the matrix.

A Case Study--Hypericin

Most St. John's wort (SJW) product is marketed with a 0.3% hypericins content. (However, recent developments indicate that hyperforin is linked to the plant's antidepressant activity in humans (Schellendberg R, et al., Pharmacopsychiatry, 31:44-53, 1998).) From an analytical chemist's point of view, these markers are challenging because of their instability when exposed to light and heat. Many times, I have seen perfect reproducibility of hypericin assays within a laboratory, yet when the same sample and method is transferred to a second lab, very different results arise. It soon became an inside joke with chemists working with SJW that virtually any hypericin assay result can be generated depending on how the sample is treated in the lab.

The INA MVP method for hypericins was taken with minor modification directly from literature (Gaedcke F, Deutsche Apotheker Zeitung, 137, 42.16:117-121, 1997). A brief exposure period to soft, white light is part of the sample preparation procedure. This step is necessary to achieve consistent inter-laboratory results, as conversion of protohypericin and protopseudohypericin to hypericin and pseudohypericin would occur to an undefined degree under typical laboratory lighting conditions. Although good method precision can be achieved within a single lab without monitoring lighting conditions, careful control of light exposure assures reproducibility between labs.

Many botanical chemists initially felt uneasy with photoconversion of compounds into the analytes. It was argued that we should not be "creating" hypericins in the lab and then assaying for them--it would give an artificially high result. On the other hand--short of specifying that chemists work in total darkness--the only way of achieving good interlab agreement is by carefully defining light exposure conditions. It took many phone calls and e-mails to gain agreement that this would be the best approach. The point being no matter how good a method is, it will not be accepted until the scientific community buys into it. An important step before validation work begins is making certain that all opinions are heard and that everyone is aware of what you are proposing to do.

Mark Lange, Ph.D., is the science director of the Institute for Nutraceutical Advancement (INA). INA's Methods Validation Program (MVP) is in its fourth year of work. INA is online at www.inanetwork.com.