Date : June 23-24, 2005,
Location: Salt Lake City, Utah
Sponsor: Utah Natural Products Alliance (UNPA)
Presented as one review considering overlap between these two areas, presenters and attendees
This two day session represented continuation of a program held several weeks earlier, sponsored by UNPA in collaboration with AOAC to discuss analytical methods for dietary supplements. All three issues (Methods, Certificates of Analysis (CoA's) and Contract analytical Laboratories (CAL's)) are critical to managing quality in this industry, and the introduction to the current seminars certainly pointed out the connection and context.
Seminar host Loren Israelsen also pointed out that if GMP's are in fact, to be presented to industry by the FDA in September, as is currently expected, then a discussion of Certificate of Analysis expectations and the role and capabilities of contract analytical labs is crucial to GMP implementation, whether the final rule is exactly like the proposed rule from a few years ago, or not. (The proposed rule indicated that relying on Certificates of analysis were not an alternative to testing, and did not specify absolute testing responsibilities for raw material suppliers. As Israelsen would point out in this seminar, with higher C of A expectations (perhaps, in fact, a series of templates), then testing can become confirmatory and even a tool of commerce and marketing. In the current environment, with such mixed expectations and standards, this is pretty much impossible.
Adequate and complete C of A's can be a practical tool and lead to trust between buyer and seller. In the forthcoming 'GMP environment', with a high third party testing burden, managing the CoA process becomes a collaboraborative process. Finished products must meet 100 percent of their label claim for the entire shelflife and if any label claim is tied to efficacious or specified amounts of an ingredient, it quickly becomes apparent how critical a reliable CoA, and ultimately how reliable any subsequent testing is.
Israelsen predicted that some companies will move away from quantification and towards qualitative and more traditional claims, while others will respond to the standards demanded and continue to provide actual amounts on labels. He observed that while FDA enforcement activity will be uncertain, the label will likely be a key target, with companies being required to unequivocally prove the product contains what is on the label, a process which necessarily starts with the C of A.
Israelsen also noted that incomplete or misrepresentative information on a C of A might, even now, be considered a violation of the terms of the commercial contract, in other words, a violation of the UCC code to and for which buyers might be able to take action against offending companies. Failure to verify or establish the reliability of a supplier Certificate of Analysis should be considered an extremely high business risk for buyers and manufacturers.
With this as a general background, the formal presentations were set to begin with a discussion and example of one company's approach to the preparation of reference materials.
Dr. Volker Christoffel, of Germany-based Bionorica AG presented, "Reference Materials for Certificates of Analysis: Assuring Authenticity and Reliability', and outlined the company's approach and rationale for the preparation and complete qualification of reference standards to be used for identification and then quantification. Dr. Christoffel went through the rigorous analytical process including the use of quantitative NMR (nuclear magnetic resonance, outlined some aspects of the preparation, a total scope of work which includes thousands of dollars of internal resources over approximately 4 months, and another many thousands of dollars in external resources and testing. The company's re-qualification guideline for quantitative reference materials is every three years, and five years for identification. The chief difference between the two materials (reference standard for identification and for quantification), Christoffel explained, is the testing for purity. The standard itself is developed to the highest level of purity, verified by rigorous determination using two independent methods. Impurities are also examined, including organic, inorganic, residual solvents and water. All the methods used for this qualification are validated and a monograph is compiled, complete with all specifications and tolerances.
The next presentation on the agenda was made by Bill Popin from USP (U.S. Pharmacopeia), who spoke about USP programs including the use of the USP monograph on Certificates of Analysis. Popin pointed out that the monograph is a suite of properties and requirements that can't be cherry-picked in order for a product to cite the monograph for a particular product. He also noted that with regular revisions to the USP, Certificates of analysis should always cite a consistent edition (be wary if the edition number is inconsistent) and really should only cite the most recent USP edition.
In developing its programs and monographs, USP utilizes a participative, collaborative process, meaning there is opportunity for industry / expert participation at all stages of development, refinement and review. Each type of material (botanical, non-botanical, ingredient, processed, dosage form inc.) has its own monograph.
Popin went on to describe the scope of committees including expert committees, botanicals, non-botanicals, information and general chapters. (General Chapters is not product specific.) His presentation also included a brief overview of the Dietary Supplements Validation Program (DSVP) which can apply to both ingredients and finished products. For finished products, the program includes testing, audits and a detailed examination of a company's documentation and paper-based systems.
Popin noted that the organization (USP) is gearing up for its scientific meeting in September.
Jim Kababick of Flora Research Laboratories provided an excellent high energy practical guide to analytical terms and processes and what to look for as companies interpreted the Certificate of Analysis. This well-rounded overview clarified the types of analysis, some of the instrumentation used, and could probably have been the template for the 'Analytical Testing for Dummies' providing attendees not only with some excellent questions to ask, but also a logical approach to evaluating the testing being performed.
The presentation included examining Certificates of Analysis to identity those that are simply photocopies. Dealing with the CoA, the point that California's Proposition 65 level for lead was based on a daily exposure (ug/day Pb) rather than an analytical value, and so manufacturers needed to keep this point in mind as they formulated product.
As he worked his way through the classes of tests, Kababick identified physiochemical properties as a 'first line of defense', a relatively cheap set of tests that can help pinpoint problems and irregularities early. This set includes density and numerous other properties, some of which impact the suitability of a product for its intended applications, such as 'tap density' which might be quite important for contract manufacturing. Another physiochemical analysis, a measurement of ash, as an example, could lead to a determination that a product was adulterated.
As a general observation, Kababick noted that the CoA must state the method used, a fact significant for water analysis and various purity and impurity measurements. He also noted that for certain methods, such as microbial tests, there are various ways of reporting results and each type of method again gives a different result, and depending on the circumstances and requirements, some might be more appropriate than others. In many cases, merely stating pass' is not good enough. Kababick also noted that the method, in most cases, is a critical component of detection. for instance, flame ionization atomic absorption analysis will not measure lead, and various means of measuring inorganic and organic materials have different limits and sensitivities. The method selected must be appropriate for the material and the objective.
In closing, Kababick pointed out that suppliers can, in fact, use CoA's as a marketing tool, and in the relationship between suppliers and manufacturers, can be the criteria for material acceptance. A subject briefly touched upon was whether the converse would be effective - a supplier suggesting that accepting a deficient CoA when a better one was available, would put themselves in a situation of heightened business risk. One fact was clear though. In order to protect themselves, manufacturers must raise their expectation for supplier-provided Certificates of Analysis, and in many cases, should be providing and defining the methods to be used by the testing laboratory. It seems a bit ironic too, that as companies move to qualify vendors, efforts to qualify labs are lagging far behind.
Paula Brown from the British Columbia Institute of Technology (BCIT) was next up to speak about 'What Does This Certificate of analysis Mean?' including problems and issues found on the CoA. In what was a perfect shift from the previous speaker, Brown went through some common deficiencies and warning signs that even those inexperienced with the reading of a CoA could pick up by the conclusion of her presentation.
Brown reaffirmed that the analytical method is critical and ideally, it will be identified on the actual CoA. She advised those in attendance to look for dates, lot numbers and the signatures of those performing the analysis or assuming responsibility for the CoA, and other techniques and tools to help determine whether the CoA provided was a reproduction and copy of someone else's lab work. Above all, she advised companies to develop a direct relationship with whoever conducted the analysis, and to support ongoing efforts by AOAC to increase commitment to method development and validation to improve overall quality of analysis and ultimately products.
In her presentation, frequently observed CoA errors were pointed out including inconsistencies that make it obvious that 'dry-labbing' or another practice was going on.
Dr. Sidney Sudberg of Alkemists Pharmaceuticals, Inc. next presented issues dealing specifically with botanical identification and speciation, using recent examples such as Coleus forskohlii contamination with belladonna in Italy and hoodia patent implications, as areas where identity confirmation and qualitative analysis could have a significant commercial impact. Dr, Sudberg also introduced the concept of HPTLC (High Performance Thin Layer Chromatography) as a identification tool and qualitative technique.
This was an effective introduction to the next presentation, that of Eike Reich, from Germany-based CAMAG Laboratory, who further described the effectiveness and practicality of the analytical technique, especially for botanical identification. Methods developed for this technique can be GMP compliant and fully validateable, and its versatility allows it to be used for raw plants, extracts and finished products. The technique is rapid, specific, reasonably cost-effective (certainly on a per sample basis), and users can do pooled samples and can track changes over time such as one would likely expect during natural or forced degradation.
CAMAG, it turns out, has prepared a TLC atlas as a reference tool, and the technology allows for quick run time, multiple samples and is specific enough to differentiate between plants of different geographical origin, species type, etc. The software itself is qualified to facilitate compliance.
Dr. Bill Obermeyer from Consumerlab.com next described his experience in investigating adulteration of saw palmetto, using analytical chemistry to identify materials that were adulterated. He noted that the trigger for this concern was significant crop destruction in Florida that created an issue of decreased supply and increased demand, leading to commercial pressure. When product began appearing on the market and selling for a significantly reduced price or under 'shady' circumstances (mysteriously appearing stock), or with slightly different characteristics (color etc.), it was obvious how important it was to identify those products that were adulterated, as quickly and cost-effectively as possible.
The first point Dr. Obermeyer made was that material from China was suspect since the product did not grow there. From there markers could be used to examine the profile of the product, specifically fatty acid and sterol ratios. As well, a complete organoleptic (sensory) evaluation and microscopic analysis would assist too, and the results would be compared with reference materials. Obermeyer went on to describe the use of iodine and saponification values and the use of many tests to pinpoint material that was in fact adulterated. He also pointed out the need for GMP procedures, and specifically method validation to more easily identify other adulteration problems.
This presentation pointed out a critical fact that industry has been wrestling with for many years. As a product becomes popular and supply of highest quality material begins to dwindle, replacement materials can get incorporated into ingredients and therefore finished products as soon as this economic incentive exists. With loose supply chain control, poor sourcing practices, and poor supplier-manufacturer relationships, this adulterated material can make it quite deep into the supply chain before there's an issue. Industry has no first-alert notification system to communicate that this might, or is, in fact occurring.
Warren Majerus from Pharmore Ingredients presented his experiences dealing with offshore ingredient supply, specifically as it related to glucosamine. Taking form Dr. Obermeyer's presentation immediately before, Majerus identified the causes of shifting quality, specifically increased demand, the loss of manufacturing identification in more convoluted supply chains with numerous intermediaries and brokers, and finally, variable testing methods combined with the practice of 'lab shopping', moving from analytical lab to lab until you find one willing to pass and issue a CoA on your product.
Majerus suggested a vendor audit, with third party auditors who know what to look for, if possible. He told of personal experiences with poor quality materials and vendors, 'bait and switch' practices, cross supply agreements and collaboration between offshore companies to 'exchange materials' in certain circumstances. The key, said Majerus, is to develop meaningful specifications including testing protocols and CoA standards. He advised buyers to look behind the picture and not to take at face value everything you are told about business practices. He specifically pointed out that careful audit and selection will identify competent suppliers, but you will need to assess meticulously, question, and then verify. Will a company reveal its source/? Will they accept an audit? Do they have third party specifications if they cannot meet your material demand?
One of the general observations for the day was that if a product is dramatically cheaper than its counterparts, question the CoA and other company practices even more closely than you otherwise must.
Day Two : Contract Analytical Laboratories
In Introducing Day Two of the seminar series, Loren Israelsen again drew out the connection between to two days' presentation - both aspects, Certificates of analysis and the performance and engagement of Contract Analytical Laboratories are critical to quality assurance programs, and manufacturers need to take ask questions of both the CoA providers and the analytical labs, to assure themselves the appropriate level of service is being performed.
Early in Day Two, an unresolved question emerged. If there is a problem with the analytical lab, who does one blow the whistle to?
First up on Day Two was Libby Harvey Fitzgerald, from Fitzgerald Consultants, who provided an overview of the contract analytical lab environment including the types of lab, their resources, and their capabilities, in an attempt to open the perceived ‘black box of mystery surrounding these essential services. She noted where they were located and pointed out hat almost without fail, there was a qualified analytical lab close to most manufacturing locations and so auditing of these facilities was actually a practical exercise.
Harvey Fitzgerald also identified who regulated the labs, noting that it was quite dependent on the program involved, with FDA regulation a nominal issue at best. There is typically no accreditation and few performance standards. Many of these labs, in fact, are registered with the US FDA but unless they have actually been audited by that agency, then the value of the registration is extremely limited. Both NSF International and USP do perform some lab audits; typically this is to allow the labs to participate in the programs offered by these organizations. Harvey Fitzgerald identified at least 25 labs with supplement experience and capabilities in addition to R & D and university labs, that are often called upon to offer contract services. Some of these too, participate in the method development and validation processes for the AOAC.
Harvey Fitzgerald predicted that the forthcoming GMP"s will force companies into tighter relationships with contract labs, and she urged those reevaluating their practices currently to consider reviewing documentation provided by the labs, auditing them, engaging them in round robin testing programs where numerous labs perform the same analyses, and not to 'lab shop'. In closing she reiterated a message from the previous day's presentation in suggesting that manufacturers specify the appropriate methods for the lab to use.
Darryl Sullivan from Covance Laboratories was next to present, identifying attributes of contract labs that clients should seek out and helping these groups better set expectations and tools to assess lab competency. In his presentation, he identified resource capabilities (personnel and qualifications), experience, depth of Standard Operating Procedures (SOP's) as core elements. Acknowledging that accreditation opportunities are limited, where they exist, they identify those labs that are proficient in certain programs and analyses.
Sullivan raised a key question for consideration: How do you know they're experts in what you need them to provide? He urged those considering lab services to push the labs, asking questions and seeking clarification on unclear points and rationale. He suggested auditing the labs, examining documents for method validation, questioning their method selection criteria, and examining specifically their instrument operating procedures. He also suggested that you, as the client, write the 'OOS' (out of specification) protocol that you want the lab to follow. Request an analytical instrument list, he advised, to ensure yourself there would be adequate instrument redundancy and to establish the age of the instruments. He urged clients to clearly identify the matrices a particular lab is capable of analyzing, the types of analytes, limits of quantification, types of customers they work with, whether they deal primarily with raw materials or finished products (and can adjust accordingly), and the uncertainty inherent in methods and measurement. For reference standards and samples, are they purchased or in house? Is the range of method use appropriate? Does the lab participate in any proficiency testing programs, AOAC collaborative studies?
Sullivan suggested an audit and assessment, requesting an organizational chart and CV's of critical personnel, as well as a list of relevant SOP's, and then shared with the seminar audience the AOAC questionnaire dealing with lab proficiency. In closing he urged clients to ask even more questions at the end of all testing.
Jana Hildreth, Scientific Liaison from AOAC International, presented 'Battling Conflicting Results: Is your lab up to the test?', including identifying reasons why one might experience conflicting analytical results. These reasons include method discrepancies, inadequate or inconsistent reference materials, lab error and outright fraud. If a method is the issue, it is conceivable that the method is not rugged enough to extend to a different lab, different instrument, or slightly different operating conditions or matrix. In this case, the method likely is not validated - or at least adequately validated. Common material issues include improper storage, and inadequate stability. Errors can occur during sample preparation and calculations, obviously impacting final results. And in the case of fraud, in some cases, the analysis is never even performed, or there is no calibration.
Examples Hildreth cited where an inappropriate method might be used included St. John's wort, tribulus and chondroitin sulfate. In certain cases, labs will use a 'best available' non-specific method, but the results can be misleading unless you understand the method's limitations. both quality and availability of reference materials is frequently an issue, sampling procedures, linearity over the range of analyses, inappropriate use of official methods, misuse of monographs (ie. the method is for a single ingredient and not a finished product) and the effect of a complex matrix can also lead to inconsistent or misleading lab results.
Hildreth reiterated the message that the client needs to ask questions and that lab shopping is not productive. She urged involvement in AOAC method development and validation programs, noting that methods are posted on the AOAC website and that the process is consultative. In closing, she urged clients to question anything that looks suspicious and observed that it is the client's responsibility to tell the lab what to test for, what the specifications are, and to invest in quality and quality partners.
Eike Reich from CAMAG again took the podium, to discuss further the implications o using HPTLC for botanical analysis. In this presentation, he noted the enhanced reproducibility compared to TLC as well as its advantage in being both time-effective and validateable, and capable of qualitative, semi-quantitative and quantitative work. For quantitative analysis Reich noted recent studies with aristolochic acids and St. John's wort. In conclusion Reich noted the cost for HPTLC analysis at a sample cost of about $30K and an additional $25K for quantitative capabilities.
Next up was Mary Hardy from the UCLA Botanical research Center, Cedar-Sinai Integrative Medicine Program. In a presentation entitled 'Safety & Efficacy of Herbal Medicine: The Research Base', Dr. Hardy effectively illustrated the connection between research requirements and full qualification and analysis of the materials going into research programs. A critical requirement for research is the ability to produce multi-batches that are identical, so the question of reproduction, and assuring this reproduction (and inherent safety) is paramount. Examples Dr. Hardy noted included variations in St. John's wort products and the adulteration and variability of PC-SPES product.
Dr. Hardy specifically noted the NCCAM Interim Policy (Interim rule notice #NOT-AT-05-004 NCCAM) which states, "Provide evidence to convince NCCAM that biologically active test agents and their placebos proposed for investigation are of sufficient (known) quality to ensure that investigations can be reproduced".
In her presentation, Dr. Hardy observed that the description of herbal interventions is typically inadequate, (ie no species or product identification including lot number etc.) and there is no standardization of activity. In most cases, the method for material preparation is not described or inadequately described, all of which decreases the likelihood of reproducing the material and hence the research. all these factors including managing traceability, identifying geographical region, harvest time etc. allows complete documentation of all the parameters and variables which might affect the product.
When Dr. Hardy further commented on the concept of 'known quality', she specified complete characterization using markers and other principles as 'thoroughly as the state of science allows' including purity and assurance of batch to batch consistency if study requires multi-batches. If the study used a single batch, the batch must be stable during its period of use, indicating that an examination of its stability must have been undertaken. Dr. Hardy suggested that certification helps provide some of these assurances, as does following good agricultural practices to generate the material. Complete control along the entire research trail including product collection and management also gives assurance that the research will be meaningful and can be reproduced.
Paula Brown from BCIT next presented her observations on contract Analytical Laboratories under Canada's Natural Health Products (NHP) system, beginning with a backgrounder on the Canadian regulatory environment. She noted that if a product specification was on the label, it must declare potency (strength), and that in a company's natural product license number application, a company must provide specifications for both raw materials and finished products including test methods and tolerances. She observed that the 'real' situation currently in Canada is one where regulations exist 'on paper', and enforcement is reactive rather than proactive, with the most significant implications being for products which cross the border into the country. According to Brown, there is mounting pressure on CAL's, and evolving need for assessing purity and potency of materials. She specifically noted the problem with limited validated methods, and in many cases, poor lab practices and reference materials.
Adulteration, according to Brown, is a significant issue, compounded by the fact that the NHP Directorate has not specified any types of tests or provided any guidance on how to prove potency. Some other issues that yet need to be resolved include how to handle matrix complexity, parameters for validation, and training of operators and compliance personnel. With labs generating non-reproducible results, it is impossible for a CoA to be generated and for companies and products to comply with the regulations.
Brown told of her own recent experiences in examining lab practices in a recent multi-lab study. She observed a lack of consistency between labs, problems with sample handling, improper use of standards, including in-house reference standards, calculation errors, non-validated methods, as well as improper adoption of methods. Upon further examination, she pinpointed the chemical standard as a major issue causing variation, and so she herself supplied the standard for participating labs to use in subsequent analyses. In this case, the RSD's were much better, although a few participating labs still did not make the mark.
Brown's current activities include an analytical proficiency program for British Columbia laboratories for goldenseal.
In closing, Brown commented on the major gaps and deficiencies in information and methods. She suggested key including validation of in-house methods, SOP's for reference materials, ensure all CoA's have adequate information. For manufacturers, select the lab carefully, audit the lab and its SOP's, and understand that generally, you get what you pay for.
Last up for Day Two was Dr. Lori Bestervelt from NSF International, who spoke about qualification of analytical laboratories. She observed that NSF performs lab audits for those labs who would participate in its own certification or analytical programs, but that no real accreditation programs, or consolidated qualification criteria exist for analytical labs and their capabilities. a significant general problem, according to Dr. Bestervelt, is the lack of (trained) auditors, and that few individuals really know to ask the right questions. She confirmed that a formal lab accreditation program, as some have suggested for this sector, might raise lab testing costs and ultimately lead to higher retail costs.
Confirming statements form pervious presenters, Bestervelt observed that the industry still needed methods and urged those present to get behind AOAC in its efforts and method development and validation. When asked about an existing model that might be relevant for certification in our industry, Bestervelt told about the EPA drinking water model as a possibility, and noted that many of the certification programs are performed at state level. She added that the ISO 17025 standard for Quality Assurance programs is another possible model, but once again observed the deficiency in qualified auditors. Currently most certifications that are seen are self-declarations and it is extremely difficult to compare lab quality, most of the information about lab performance is limited to the purchaser alone, is inconsistent and there are very few lab audits performed. In general, Bestervelt noted, there is extremely limited CoA criteria and very limited lab expectations.
In providing some guidance on lab assessment, she would typically like to see a written Quality Assurance) QA plan including an organizational chart, SOP's, sampling procedures, calibration documents, validation information and QC checks, chain of custody protocols and examples, record keeping, and training program records. She would assure that the lab has determined appropriate detection limits and whether they are open to audit.
Bestervelt went on to talk about the NSF approach to lab qualification including a QA questionnaire to cover all of above including whether the software is current. This paper based assessment is essentially a desk audit, and then the organization would participate in yearly auditing as well as review of all raw data and QC samples, blind testing programs etc.
In concluding and speaking specifically about the practicality of industry establishing a qualification or accreditation program, Dr. Bestervelt noted that a system would ideally be mo0nitored by a neutral third party, possibly with ISO 17025 as a benchmark, and would involve document review plus on site audits (experienced auditors would be needed) and would also include a proficiency testing program (on a product by product basis?) and round robin analyses with pass/fail criteria.
Participants in the audience appeared mixed regarding the merits and costs of such a system.
Overall Conclusions from the Two Day Program:
1) Industry needs to establish, define and administer (higher) expectations for contract analytical labs
2) Industry needs to understand analytical methods and to have a better understanding of what questions to ask
3) There needs to be more transparency across these two critical areas of the value chain - CoA's and CAL's
4) A template identifying the minimum expectations on CoA's would be helpful
5) A first alert system should be developed when potential for economic adulteration exists
6) Sourcing practices need to evolve to better discern who is actually providing and testing raw materials
7) Methods need to be developed and standardized