DRUG TESTING ADVISORY BOARD

Scientific Meeting on:
Drug Testing of Alternative Specimens and Technologies 
(Part II)
Transcript - Day 2

September 10, 1997

TABLE OF CONTENTS

 

FDA Clearance Process - Steve Gutman, PhD

Analytical Procedures for Hair Testing

- Ray Kelly, PhD

Analytical Procedures for Hair Testing

- Christine Moore, PhD

Implementation of a PT Program for Hair Testing in

Florida - Bruce Goldberger, PhD

Public Comment and Discussion

PROCEEDINGS (8:33 a.m.)

DR. BUSH: Good morning. I would like to introduce to you this morning our first speaker, Dr. Steven Gutman.

Dr. Gutman is the director of the division of clinical laboratory devices within the Center for Devices and Radiological Health at the Food and Drug Administration. He is a certified pathologist, with a BS from Ohio State, an MD from the Cornell University Medical College, and an MBA from the State University of New York, Buffalo. His residency training in anatomical pathology was completed at The New York Hospital and in clinical pathology at the Mayo Clinic. After 10 years as a clinical pathologist and chief of the laboratory service at the Buffalo VA Medical Center, he joined the division of clinical laboratory devices in the Office of Device Evaluation as a medical officer in 1992.

I present to you Dr. Gutman.

Agenda Item: FDA Clearance Process.

DR. GUTMAN: Good morning. The Division of Clinical Laboratory Devices, referred to as DCLD by those of us who know and love it, is currently composed of 50 scientists, ranging from medical technologists to clinical pathologists, who are actively involved in the premarket review of in vitro diagnostic devices.

We process two kinds of submissions. New versions of old devices are cleared, the specific and parochial word we use is cleared as premarket notifications. Because the part of the law that is cited is the 5(10)(k) section of the law, we commonly refer to these at 510K submissions. Fundamentally new devices are approved -- the specific and parochial and different word from cleared is approved -- as premarket approval applications, and these are commonly referred to as PMAs. In the semantic framework of FDA, whether a device is old or new depends on whether it was in the commercial marketplace at the time that the medical device amendments of 1976 were passed. If it was in the marketplace at that time, it is considered a valid scientific predicate. An in vitro diagnostic device is nothing more than a laboratory test. By far and away, the largest volume of work that we see falls in the category of 510Ks. We currently process about 1,000 510Ks a year. The operative term in our scientific review of a 510K is to evaluate or to establish substantial equivalents.

The notion is that the new version of the old product that can be traced to a marketed product in 1976 is substantially equivalent to that old product. So, the good news is that a new product will not be worse than an old product that was on the market in 1976. Of course, the bad news is that it may be substantially equivalent to a product that wasn't anything to write home about that was on the marketplace in 1976.

There is nothing arcane or opaque about our process. We are concerned with the heart and soul of laboratory tests, with the issues that most manufacturers, certainly all laboratorians know and love, and that is the accuracy of the tests, the precision of the tests, the analytical sensitivity of the test, the analytical specificity of the test. It is a quantitative test, perhaps, also of the linearity of the test. There is also nothing mysterious about the limitations of our review process. Our process involves no laboratory component.

We have no wet lab review. We do an entirely paper review of data submitted by a sponsor. As one of our senior medical officer often points out, we are somewhat akin to the editor of a medical journal in terms of evaluating products, the outcome being marketing rather than publication.

There is also the obvious challenge to us in our review process in trying to establish standards for equivalency. Those of you who read or obsess over the clinical pathology or the laboratory medicine literature appreciate the fact that, in fact, there are not terrific standards in place to establish equivalency when you look at new versions of old products. We process a much smaller complement of PMAs in the course of the year, usually somewhere in the neighborhood of one to two dozen PMAs in various states of activity during the course of the year. For a PMA there is no predicate. It is a fundamentally new product so that substantial equivalence is not a review issue. We are instead concerned fundamentally with demonstrating that this new product is safe and effective. In fact, in 1990 the safe medical device act expanded our interest in safety and effectiveness to encompass 510Ks as well.

When we look at substantial equivalents in 1997, we are concerned with making sure that the equivalency is based on a safety and effectiveness measurement. The impact of that has been that for 510Ks in 1997 we require as part of the submission either a summary of safety and effectiveness or a statement that such a summary will be made publicly available. The review consequences of that interest in safety and effectiveness is that for all PMAs and for at least a subset of 510Ks, we are not satisfied with looking at analytical data alone. We will push the limits and start to ask clinical questions and be interested in clinical or diagnostic sensitivity from the Basian, Galen and Gambino model, clinical or diagnostic specificity, increasingly interested in reference ranges and increasingly interested in the expected value or the likely biological or clinical activity that a particular analytical system might produce.

Again, there are no -- there is nothing mysterious about the review limitations for PMA. We are constantly bedeviled by a lack of gold standards in trying to evaluate new products, and often have to turn to silver or bronze or lead standards in trying to evaluate a new product.

We constantly are challenged by both overt and latent bias in data collection. If you look at medical literature and find difficulty in standards for equivalency in looking at old products, you can imagine how challenging it is when you get a fundamentally new product, to try to establish an appropriate threshold for safety and effectiveness. When you go to look at the medical literature, it is often not there. In vitro diagnostic devices are unique in the world of medical devices in that we have our own labeling regulations, which are spelled out in the Code of Federal Regulations in a section called 809.10.

In my biased view as a clinical pathologist -- and I had nothing to do with the drafting of these -- I think they are brilliant. I think they are so good that I intend to share all 15 of them with you this morning, albeit with brevity.

The labeling requirements in that regulation are:

that a product have a proprietary and established name for the product; a clearly stated intended use and indications for use of the product; a summary and explanation of how the product is likely to behave; and information about the principles of the procedure to be employed in the use of the product.

It requires, when appropriate, information on reagents, information on instruments, and information on specimen collection, handling and preparation. It requires an outlined, detailed procedures. It requires information on how results are to be generated or calculated. It requires that limitations be clearly elucidated. Last but certainly not least, it includes information on expected values, information on the heart and soul of what is going to happen when you run that test, the performance characteristics. It requires a bibliography, a name and place of business, and the date on which that labeling was produced, so you know how current it is. All of these components are not created equal. First and foremost in the labeling review is an evaluation of the intended use and the indications of use a sponsor for a particular product gets to make the claims for that product.

That intended use will, in fact, tremendously impact and drive the review, and the type of review, and in fact the type of submission -- the 510K or PMA. The questions likely to be raised in the review, and the data requirements likely to be required before the product reaches market are all, in large part, driven by the intended use, by the claims being made for a particular product.

A central preoccupation in the life of the Office of Device Evaluation, certainly in the life of the Division of Clinical Laboratory Devices over the last several years has been an effort to develop a simple and a standardized review model.

It is, again, my view as a biased clinical pathologist, that there is no excuse for poor evaluation of laboratory testing products in 1997, that there is a wealth of information to guide manufacturers and to guide the FDA in terms of appropriate method evaluation and review. There is a rich design and statistical literature. There are numerous voluntary standards to draw from. Our personal favorites are NCCLS, but there are European and international standards as well. There are several dozen FDA guidances which can be used to provide insight on appropriate techniques to use to establish performance for a wide variety of in vitro diagnostic products.

At the heart of the scientific model is a triad, the notion that you need a clear, up-front design or protocol to support your intended use or claims, that you need a careful and controlled collection of data, and that you then need to interpret that data with reasonable statistics. If you decide to use avant garde statistics, you need to explain to the FDA what you are doing and why. At a minimum for quantitative tests, we are concerned with accuracy or at least bias. We would prefer to see that expressed in some form of linear regression analysis. We are concerned with precision or repeatability. If the test is quantitative, we prefer to see a de novo analysis so that we can look at components of variation. We require, when appropriate, experiments to detail analytical specificity and analytical sensitivity.

For qualitative we require the same, but become more preoccupied with cutoff points and equivocal zones and trying to understand what a test means. In light of the product line being discussed today, there has been a great deal of effort over time for FDA to interact with SAMHSA and NIDA scientists in trying to establish appropriate cutoff points and trying to understand the appropriate cutoff points, and trying to make sure that when they do apply to a particular product, they are properly communicated and labeled.

There are at least two things that FDA review is not. In general, FDA review is not outcome oriented. We are usually looking at surrogate end points. We don't usually do long-term prospective studies to demonstrate a particular new test will actually, biologically, over 10 years save lives or not save lives. We are interested in what the test measures, what insight it provides in disease processes, but we don't require outcome studies. That may be good, that may be bad, but that is the reality of the review. Our studies are not usually prospective. Unlike most medical devices, we don't usually require longitudinal data. For most of our studies we require concurrent studies, where you take a sample and you compare the new versus the old method or the new versus a reference method or standard method, or the new method versus a clinical data set. But with rare exceptions we don't require data to be generated over a lengthy period of time. What we do require is attention to detail in your protocol, in your collection of data and in your statistical interpretation of those data sets.

So, at the risk of being redundant, the maxims in DCLD are, up front, reasonable design, good data collection, valid statistics, and last, but certainly not least, last and maybe most, is truth in labeling, so that you clearly demonstrate in the label what your performance is, what your limitations are and what the person who purchases this product might expect to encounter. We view ourselves as having a dual mission. We want to keep bad products off the market, but we want to foster good products and get them on the market in a rapid manner. We have been looking at a number of reengineering and reinvention processes to make that happen. But the core of our review process is an intense interest in good science. We think that good science makes for good regulation, good science makes for good laboratory practice. Good science makes for good business sense, because good products should be easier, over long periods of time, to market. Good science is what makes being regulator at least a little bit of fun. That concludes my remarks and I would be happy to answer questions.

Thank you, Donna. I would like to invite folks here, we are having a public panel meeting on September 25. Do you know the location?

DR. BUSH: (Comments off microphone.)

DR. GOTMAN: The Holiday Inn Bethesda, from 9:00 to 4:00. There should be information on the internet for people who are internet happy. The subject of that discussion will be one germane to you folks, and that is drugs of abuse testing kits for possible sale over the counter. We will be addressing scientific and labeling issues, lots of the issues that I have just talked about in the context of that particular issue. We have a draft guidance document on the internet now. We plan to issue a preliminary revision of that document in the next couple of days. We will be having our own panel of toxicologists and chemists and clinical experts comment on those issues. Folks who are sitting here who have a passion for those issues are welcome to come and listen or are welcome to call Sharon Lapolinin, the executive secretary of that committee, to come and talk. You are welcome to comment before. There is a little bit more muted ability to comment during or after, in terms of your opinion on what kinds of review criteria should be applied for that product line.

PARTICIPANT: (Question off microphone.)

DR. GUTMAN: If you go to the FDA home page, you will see a little bit more information; not a lot but a little more.

PARTICIPANT: (Question off microphone.)

DR. GUTMAN: For those of you who can't hear, the handout which didn't exactly resemble the talk was truth and labeling and has Clara's number on it. That number will probably give you a lot of information or at least some information on that upcoming panel meeting.

PARTICIPANT: (Comments off microphone.)

MR. CAPLAN: Can you comment on any of the FDA's prior experiences or problems or concerns encountered when you were looking at either general immunoassay products or on-site immunoassay products? Was there anything special developed or problems that might help us in looking at some of these products as we go forward?

DR. GUTMAN: That is pretty broad. We have 20 years of review experience and we have probably seen it all. I would say that the most challenging issues that we are currently dealing with -- this isn't necessarily historical -- but the most challenging issues we are currently dealing with are the issues which press the limit in terms of point of care and physician office lab. Now the potential for over the counter is that the technology has become more and more user friendly and simpler and simpler. There are a lot of issues about how to deal with that friendly technology, how to characterize that technology in the correct hands, to make sure that the label reflects the right use and, frankly, how to quality control those systems in places. We are being challenged constantly by the manufacturing community, which is a very creative community in terms of its scientific expertise, and a very creative community in terms of its marketing expertise, to try to keep the science and the marketing and the science and the submissions as a match.

I don't know that I have one particular theme, but the themes that we deal with are the ones that -- at least for this product lines -- are the ones that you obsess over, I am quite certain. There are the analytical failures, which are inevitable, the biologic failures which are inevitable. There are the issues of quality control, the issues of collection, potential for adulteration. All the issues you deal with we deal with. We have some limits in terms of what we can do. Our response is that when we are uncomfortable with those limits, we try to label so that whoever buys a product understands what they are getting.

DR. CAPLAN: Have you ever considered some sort of uniform response of immunoassays, kind of a standard response phenomenon as opposed to equivalence of one product to a previous product?

DR. GUTMAN: Yes, there is actually a lot of talk and a lot of discussion now about international harmonization. I don't know how familiar folks here are with the European standards but, in fact, the European approach, which is quite a bit different than ours, tends to be standard based rather than based on this equivalency determination. In fact, in the chemistry branch, the chemistry toxicology branch, there is some interest in developing a more standards based approach, using NCCLS documents as a sort of template or core, using those kinds of things as a template or core.

There are really two issues. One is having a standards base for the generation of the data. The second issue is deciding if there are certain thresholds you must meet before the product clears the market. They both challenge the regulatory and statutory paradigms under which we work. But we are working on it. If anybody has any great suggestions or ideas, there is a lot of discussion in the center right now about doing things better. We would be very receptive to working with anyone in the public sector or the professional sector or the private sector, the manufacturing sector, to do things better.

DR. KIPPENBERGER: My name is Don Kippenberger. I have a question or two. It seems like 1976 was a very important date for what you have been speaking about in part. Could you please explain about the grandfathering process that occurred, especially with urine kits, and exactly what process was done?

DR. GUTMAN: No, actually that is sort of a legal issue that dates back to the 1970s. It precedes me and is actually probably beyond my expertise. I need to get somebody from general counsel to talk. The only thing I know about it was that it was a very deliberate effort on the part of the congressional folks to allow existing technology to not be thwarted. Also, in the regulatory response by the agency, it was the reality that you had a fledgling new agency that took time to get on board. So, I don't have any great insights into either the history or the statutory basis for that, but I understand the rationale for it.

DR. KIPPENBERGER: But urine kits were grandfathered in 1976.

DR. GUTMAN: Urine kits were predicates. So, we are looking at equivalency determination, yes.

DR. WALSH: I would like to thank you, Dr. Gutman, for taking the time and having the courage to come before this group. It is important for the FDA to be involved. One of the issues that a lot of us have been concerned about over the last 10 years is that there are a lot of drug testing kits on the market being marketed for workplace testing that are being distributed under the forensic use exemption. Often I will see in a press release on a new product that they are putting together a package for submission to the FDA. Somehow those never seem to get submitted. Let me make it clear that I think most of the manufacturers have played by the rules and have gone through the FDA process and gotten that clearance, but others haven't. I would really appreciate if you could elucidate us, what is the policy in terms of forensic use exemption and what avenues does the FDA have in terms of forcing compliance with the Food Drug and Cosmetic Act.

DR. GUTMAN: The agency, I believe, would probably view certainly workplace tests -- I am not so certain about forensic tests -- but I think the agency would probably view both of those as medical devices, in that they could theoretically fall under the purview of the statute. The decision has very clearly been made, and it has been made in several centers, so it covers a wide variety of product lines. The decision has very clearly been made not to regulate forensic testing in the past. I would be surprised to see that fundamentally change in the future. There has been a little bit more interest or thought or discussion, particularly following congressional hearings last year, about the consistency of regulating and not regulating workplace testing.

The history has been that we have not regulated workplace testing. Those were policy or discretionary issues that were simply made on the basis of existing resources. If people have either passionate views that that is a good idea or passionate views that that is a bad idea, there is a lot of discussion in the agency now about our appropriate role in the regulation of virtually every product you can think of. You might want to express that to someone more important than me; Dr. Alpert, the office director.

Right now forensic testing and workplace testing are not being actively regulated. I don't know, if somebody were to send one in, if we would actually reject it. We might. The only time we would regulate a forensic or workplace test would be if it came in with a dual claim, if they want it for forensic and medical use. Our regulation would then be based on the fact that it is planned to be used in an ER or a hospital lab or a physician office lab or something like that.

DR. WALSH: I think the fine line in workplace testing is if you go to the AACC meeting and you ask the sales guys around, and you ask the question is this product cleared or approved by the FDA, they say, no, it doesn't need to be. It is for forensic use only. You follow on, what do you mean. Well, we recommend this for pre-employment drug testing. So, basically it is being used to make a hiring decision. However, if that same employer uses it on an already existing worker and then, on the basis of that result, requires them to go into an EAP or a treatment program, it seems to me that is a medical decision and then it comes under the act. It is a fine line.

DR. GUTMAN: I think you can make the argument. I don't have any trouble buying the argument. As I said, it is a policy decision and it is a resource decision. It is not statutory and it is a hard scientific call where to exactly make the line.

DR. JONES: Skip Jones from the DTAB board. I appreciate your comments to Dr. Walsh's question. As a follow up, we have been going through a lot of deliberations about whether FDA approval should be required for these devices. I wonder if you could respond to this position. If we took the position as a board to recommend that all devices would require FDA approval, could you respond to what the FDA's position would be to that? It is a what if proposition.

DR. GUTMAN: We wouldn't turn the submissions away. The factor that would probably drive them in is your requirement for a particular setting. I don't think we would reject them. If there was a policy problem, the easy game to play is for the company to say, for workplace and medical uses. Then we surely wouldn't reject them. If you folks as a separate government agency thought it was important that we looked at again, it is something that would probably have to go through Dr. Alpert or Dr. Burlington. My guess is that there is a mix. I don't know what percentage is reviewed versus what percentage isn't reviewed. We are pretty good at this kind of review now. We have had a lot of experience. I wouldn't reject it out of hand as something we wouldn't want to cooperate with.

DR. WALSH: (Question off microphone.)

DR. GUTMAN: The question is about the nature of the 9-25 meeting. The September 25 meeting is the chemistry toxicology advisory panel.

Our panels are not preselected, so one never knows what direction they will take or what interesting ideas they will come out with. The notion is to focus on a guidance document to deal with the scientific and labeling issues about an interesting new product line which would be over the counter drugs of abuse tests. Thank you.

(Applause.)

DR. BUSH: For the record, I just want to enter the FDA's web site address. It is www.fda.gov. Thank you, Dr. Gutman.

Next up as our next speaker is Dr. Ray Kelly. Ray Kelly received his bachelor's degree in biochemistry from Washington State University, a PhD in chemistry with a specialty in biochemistry and molecular biology from the University of Oregon. He had a post-doctoral appointment in forensic toxicology at Case Western Reserve University and Cuyahoga County Coroner's Office in Cleveland, Ohio. Previous jobs that he has held have given him an awful lot of experience in clinical and industrial toxicology. He was the laboratory director of MEDTOX Bio-analytical Laboratory in Chatsworth, California and currently he is the directory of toxicology at Associated Pathologists Laboratories in Las Vegas, Nevada. Dr. Kelly is a diplomate of the American Board of Forensic Toxicology, the past president of the California Association of Toxicologists. He is on the editorial review board for the Journal of Analytical Toxicology, and he is a fellow of the American Academy of Forensic Sciences, Toxicology Section. He is also a member of the American Association for Clinical Chemistry. He has authored more than 20 peer reviewed scientific papers in analytical, clinical and forensic toxicology, physical biochemistry and molecular biology. With that, I would like to introduce Dr. Ray Kelly.

Agenda Item: Analytical Procedures for Hair Testing.

DR. KELLY: Thanks, Donna. Dr. Autry and members of the Drug Testing Advisory Board, I appreciate the opportunity to speak today, and your continuing interest in the subject of emerging technologies. Our lab is certified by SAMHSA and accredited by CAP for forensic urine testing. We have attempted to structure our hair testing process using the same sound forensic principles. Like the board's guidelines, however, I will freely admit that it is a work in progress, and not every problem has yet been solved. In fact, I agree with something Yale said yesterday, that we have more questions than answers at the moment. In fact, we have received help and input from several people in this room in designing our program, for which I am profoundly grateful. By saying this, of course, I am not implying that all of them are unqualified supporters of hair testing. That is the usual disclaimer.

By way of introduction, let me tell you a little bit about our laboratory. About 20 percent of our work is clinical toxicology and about 80 percent forensic. The latter includes not only employment testing -- about 80 percent of the total -- but some medical examiner and law enforcement work. Of the employment testing, about 25 percent is hair and 75 percent urine. This slide shows a breakdown by industry group of our hair and urine clients.

Hair testing is a lot of times associated with the gaming industry. In the early years, that was certainly true. These recent statistics indicate that there is really not any more incidence of gaming clients among our hair testing clients than there is among our urine testing clients. In fact, the retail sales in health care, you see the percentage that the gaming constitutes. Government agencies here primarily represent the courts, and especially child custody actions and child protective services. The service industries include attorneys. The main discrepancies that you will see between urine and hair clients here is with regard to construction and transportation, which is perhaps what you would expect. This kind of shows how our clients are using hair testing. At APL we consider pre-employment testing a very appropriate use for hair.

You have heard presented in the past various advantages that hair testing has. The one that seems to be the most important to employers, in my judgement, is the ability to cover a longer time frame and, thus, to detect more users. Hair testing, obviously, isn't suitable for most types of testing of employed workers because it doesn't address the time frame that is of interest, which in the case of post accident and most reasonable suspicion testing is the few hours just following an incident. Some clients have tried to use hair testing to cover a situation where an employee fails to report an injury for several days. Then when the employee returns they will run a hair test. That really hasn't met with much success. Obviously when I testify in those kinds of situations I have to say that the result of the hair test has little or nothing to do with a particular incidence. It may have something to do with previous drug use. Some clients have employed random hair testing of current workers to help maintain a drug free work force. APL has a network of 20 owned collection sites, primarily in Nevada, which collect urine, hair and blood. We deal with about 200 collection sites around the country that are able to do hair collection for us that are not owned, obviously, by ourselves.

For our own collection sites and our own collectors we have a training course and on occasion make that available upon request to collectors from other organizations as well. All of our collectors undergo a rigorous training program before being allowed to collect forensic samples. Stage one involves just observing collections and attending a basic class on collection techniques, and collecting under direct supervision, one on one. Stage two, the individual attends a one-day certification course, which features an exam. They have to receive a passing grade on the exam and then they receive certificate. At that point they are allowed to collect forensic samples, as long as they keep their proficiency up. There is a feedback mechanism for that involving the toxicology client service department. If someone is working in one of our collection sites, they are not allowed to collect forensic samples unless they are either in supervised training or are supervised. If they happen to have failed the exam, which has happened a few percentage of the time, they are returned to the training phase until they can take the certifying course again.

Some of the guidelines for hair collection, hair and urine are collected at the same locations, but they are collected at separate times and using somewhat separate procedures. In particular, before hair samples are collected -- and there is a separate area set aside for that -- the work area and the collector's hands are clean before making a collection.

Here are some of the materials that are provided to the collector. Hairs collected from the top of the back of the head, the occipital crown, by clipping close to the skin. That is one or two millimeters above the skin. Despite what was mentioned yesterday, we do not pluck hair forcibly out of the scalp. I think that is done only in research situations and I think they have to pay for that or in some way talk people into participating in that. Since this is kind of a voluntary process, it wouldn't work well in that regard. We do advise collection from several adjacent sites in order to assure representative sampling, to minimize the cosmetic impact, and that can be significant in some cases. Don't look at me, please. I still have plenty of hair on the top of the back of my head.

We collect approximately 120 strands and try to make a due allowance for shorter, curlier samples by collecting a larger amount. If possible, the hair is then aligned by the root ends and the hair is then placed in a tapered foil sheet. It is folded lengthwise and inserted into a cardboard envelope. That is sometimes called a sample acquisition card, which we don't call it that because I am not sure the name isn't patented. It is then sealed with a tamper evidence security seal, and a bar code label from the external chain of custody test request form is transferred to the envelope in order to tie the two documents together. Next, the donor reads and initials the appropriate area on the envelope and then prints and signs his or her name on the test request form.

The collector does the same and this initiates the chain of custody. In the donor's presence, the collector places the top copy of the test request form into the outer pocket of the plastic security bag, seals it, places the sealed sample into the main section of the bag and seals and initials the entire bag.

The other copies of the form can be kept by the collection site or donor or can be forwarded to the employer.

We also provide the collectors -- all of them, but especially ones that aren't part of our organization -- with directions with pictures, indicating how the samples should be collected. Samples are transported to the lab by a lab courier or the U.S. Mail. One advantage that hair has over urine, is that it is lighter and occupies a lot less space and permits a larger number of hair samples to be sent at a time than urine. Further, there is no danger of leakage.

Chain of custody processes begin, of course, at the time of collection and document all the processes experienced by samples and aliquots on appropriate chain of custody forms.

Laboratory security is going to be familiar to many of those of you who are involved in federally regulated testing. The laboratory is in a secured, limited access facility. Each operational area within that facility is individually secured and access is limited to people who work there and appropriate supervisory personnel. These areas include accessioning or sample processing, the screening area, the confirmation area, certifying and both long and short-term storage. All visitors have to sign in and be escorted. Both samples and data are maintained in limited access areas.

We retain negative samples for approximately one month and positive samples for one year.

When samples are received in the laboratory, there is a process that takes place in which the seals are checked. One sample at a time, of course, is processed. The two documents, the test request form and the sample shipping card or envelope are compared to verify that the same information occurs on both. If these checks are met, then the laboratory accession member, which is a unique internal member, is assigned and the sample is entered into the computer.

The process of aliquoting the sample also involves cleanliness. The area is vacuumed initially and then washed and all the utensils used in the aliquoting process are washed. One sample at a time is processed and the clippers are washed in between samples. The process involves, for each sample, aligning the hair from the root end and clipping 3.9 centimeters to encompass a time frame of approximately 90 days. Then that hair is subdivided into short segments and 20 milligrams is laid out for each test.

We use two different immunoassays for screening. RIA is used for cocaine, opiates and PCP. We use ELISA for amphetamines and cannabinoids.

We were not able to locate a commercial RIA product for those two types of analytes that had the appropriate specificities. Because parent drug predominates in hair, we want to have significant cross reactivity with the parent compound. The sample is washed briefly with solvent at room temperature and then it is extracted for two hours at 75 degrees with a second solvent mixture.

This mixture is then evaporated to dryness and reconstituted in buffer to conduct the immunoassay. The calibrators and controls are expressed in terms of picograms per milligram.

This slide here shows a typical RIA procedure. This is Diagnostics Products Corporation RIA for cocaine. Here is a typical ELISA procedure. We are using Diagnostics Limited. Again, the response in ELISA, you will recall, is a colormetric response which declines in proportion to the concentration of drug in the sample. It is kind of hard to see the line here because it is black and it doesn't show up against the screen. This is a cannabinoid ELISA calibration curve showing the cutoff up there at the top. It is sort of hard to see, but in the upper left-hand corner is the negative calibrator. So, there is a substantial decline in going from zero to five picograms per milligram. Here is an amphetamine ELISA calibration curve. On this curve the cutoff occurs at a somewhat lower point, a cutoff of 300 picograms per milligram. The normal way that we apply this does not include a complete calibration curve. We simply employ single point calibration at the cutoff and controls that bracket the cutoff.

The calibration curve is used to validate new lots of reagent and to validate the cutoff itself. Cutoff concentrations are, for amphetamines, 300 -- remember, in picograms per milligram -- cannabinoids is THC at 5, cocaine 300, phencyclidine at 500 and opiates at 500.

Here are some of the cross reactivity data. This is the THC ELISA. The compound typically at the top on these slides is the one that is sort of the target compound. So, THC at 100 percent. These others are all at less than one percent. Non-reactive compounds are shown on this list.

The cocaine, RIA, one thing to note is that cocaine in this assay reacts much more strongly than benzoylecgonine and about the same as cocaethylene. That is extremely favorable because cocaine is the main species that is found in hair and cocaethylene is commonly found, too. We find that in about 30 percent of the samples.

Again, some non-reactive compounds.

PCP is, of course, the target compound. PCP, which as far as I know we have never found in hair or even in urine, has virtually the same cross reactivity. Then some of the PCP metabolites are at a lower level.

Again, the non-reactive compounds.

This is the opiate RIA. Here you see a couple of compounds, ethyl morphine and codeine that have higher cross reactivity than morphine. Most of the common opiates will show some level of reactivity. In our procedure we don't confirm anything other than codeine and morphine and 6 acetylmorphine, however. A series of non-reactive compounds.

Most of the cross reactivity data I just presented came from manufacturers. For the amphetamine ELISA procedure, however, we have conducted some experiments to check cross reactivities. In the experiments we equated the inverse of the ligand concentration at 50 percent binding to the antibody with the affinity constant for that compound. We then calculated the ratio of each affinity constant to that for D methamphetamine under the same conditions in order to derive a figure for percent cross reactivity. That is what you see plotted here, is some of these data. It is all in pretty colors, but I don't think you can see that either. This is what that looks like. You can see, because we did the work ourselves, we get results to the nearest 10th of a percent instead of just percent. That is not really true.

It is interesting that a couple of the compounds that are widely used, or were until very, very recently because of the recent bad news on them, the components of the fen fen diet plan, fenfluoramine and fentramine, show a certain degree of cross reactivity. Also in this assay, L. meth has a dramatically reduced reactivity in the assay versus D. meth. So, that is helpful. So far, we have never seen a hair sample that had an amount of L. meth in it so that it tested positive through our entire process. I suspect the main reason for that is because of the seven percent cross reactivity in the screening test.

It is fortunate, too, that ephedrine and phenylpropanolamine have fairly low levels. We do see these sometimes, but not in the kind of concentration ratios to the other amphetamines that you would see in urine.

The last slide showed all the compounds with cross reactivities above .1 percent. These compounds here didn't have any measurable immunochemical response, as shown on this slide. That was defined as a compound whose concentrations, even at 1,000 times the methamphetamine concentrations didn't approach 50 percent binding.

Again, on this slide it is kind of hard to see the individual points there in red. This is a validation of our cutoff for amphetamines using ROC analysis. The break point appears to lie between 200 and 250 picograms per milligram. So, our value of 300 is conservative and favors false negatives. I am sure that Vina will be thrilled to see this slide.

This slide shows confirmation methods. The same cutoff values are used for confirmation as for screening. So, we have -- basically they are all GC/MS methods with the exception of the cannabinoids which are MSMS.

The derivatizing agents are shown in the right column. Again, in most cases we are using TMS derivatives. In the case of THC itself, we use the TMS derivatives and positive CI for MSMS. In the case of carboxy THC, the PFPA derivative with negative CI. Right now those are being run in two separate chromatographic assays, because there is a problem combining those two derivatives for us. The kind of confirmation criteria that we use are relatively standard. The instrument, of course, is auto-tuned on a daily basis. In the case of the MSMS it is an operator tune, not an auto-tune, but it does use the same compound. Then routine maintenance is documented for the instruments.

The analyses are conducted with sensitivity toward the current LOD, LOQ and linearity figures, which I will provide in a moment. The qualitative criteria include retention times, ion ratios which, in deference to the lower drug concentrations in hair, we have had to make plus or minus 30 percent instead of 20 percent, which is standard for urine. Three ions for GC/MS and two ions which are, of course, daughter ions by MSMS. Quantitation is done by the internal standard method, with internal and external QC and single point calibration at the cutoff. We have controls that bracket the cutoff.

LOD is defined in the conventional way as the lowest concentration where the analyte is detected and fulfills ion ratio criteria but not necessarily the quantitative criteria of being within 20 percent of the target value. LOQ would be the same thing except it is within 20 percent. Upper limit of linearity is the highest concentration of the analyte that fulfills ion ratios and is, as well, within 20 percent of the target value.

Here are some of the parameters. For compounds like amphetamine and methamphetamine, in fact everything except for marijuana, I have listed here the GC/MS, LOD and LOQ values, although these analytes can be performed by MSMS and on occasion we do that.

We have a reporting rule that we don't report methamphetamine unless amphetamine is also present in the sample. It has to be present at a concentration of 100 picograms per milligram or greater. Sometimes to achieve that with ion ratios requires that it be done with MSMS.

Upper limits of linearity, in most cases, these are just the highest concentrations that we looked at. These concentrations are very, very low compared to the kind you are going to get in a urine extraction. There is not really a problem of lack of linearity. These are very low amounts of analyte. But they are, in general, higher than the kinds of concentrations that we actually see in samples. This is totally invisible. I guess I can get to my joke sooner rather than later. I was going to present a few of these slides.

This is a cocaine calibrator. Those of you in the front row, I think if you move forward five or ten feet, I think you will be able to see it. They are in the handout also. In fact, there is a better version of it in the handout. What happened is, I took these slides to a photo lab.

Donna graciously gave me three weeks notice, but I promptly went on vacation, so that blew a week right there. So, I was a little pressed for time.

I wound up taking these things to a photo lab. I think the guy processing them said, gosh, these things all look alike. That is maybe the explanation for why he gave me eight copies of one slide. He didn't do seven of the slides. Basically I am trying to deal with what I have got here.

This is cocaine in the calibrator and it shows three ions and the combined plot on the right-hand side, internal standard at the top.

This is cocaethylene. You know it is bad when you can't even read your own slides. This is benzoylecgonine.

Here is a sample, about 4400 picograms of cocaine, which is a reasonable level, certainly not the highest we have seen, but a high level. No cocaethylene to speak of in this sample and about 400 picograms of benzoylecgonine.

Here is methamphetamine and amphetamine. This the calibrator. Again, some of the ones that I was going to present that were actual samples were the ones that the guy, hey, I know the drill here. I am just going to make eight copies of the one slide. He will never know the difference. The way this projects, I have to agree with him. Codeine, morphine, phencyclidine. This is 6 acetyl morphine.

This is MS/MS data. If I can remember what the hell it is here, this is the 5 picogram calibrator. I am going to try this real high tech device here. There are two daughter ions here and this is the THC calibrator, 5 picograms per milligram. This is the corresponding internal standard. The way the data is presented, the internal standard is on a separate page. This is the low control, 3 picograms per milligram. This is carboxy THC calibrator, again, the two ions. This is a reconstructed ion chromatogram at the bottom, and the internal standard. It is kind of interesting to compare confirmation rates between urine and hair. You may notice that our confirmation rates, our positive rates, are quite high for urine as well as for hair.

It is significant, I think, to note the differences. These data come from people who were tested using a panel that involved both hair and urine, which many of our employers prefer. The reason they like it so well is that you are looking at kind of two different time intervals. If you find positives in both sample types then you have clearly defined two different episodes of use. The general feeling is that in order for hair test to be positive, somebody has to use a drug perhaps more than once, perhaps somewhat regularly. Perhaps that is not true in the case of cocaine, but for the others it seems to be true. That is helpful, but looking at a few day window in urine and a number of weeks in the hair sample. You will note, for example, there is a higher proportion of amphetamines and of cocaine that are found in hair. With cocaine it is really quite dramatic. That is consistent with a lot of the data in the literature about the tight specific binding that cocaine undergoes with hair. Typically we see about four times as many positives for cocaine in hair as we do in urine.

With opiates there are fewer positives seen in hair than in urine. Part of that is a function of the fact that in urine we are looking for a larger number of opiates. We look for hydrocodone and hydromorphone and oxycodone and dihydrocodeine as well, which we don't look for in hair. Perhaps that is a partial explanation of that.

PCP, we get very few, but the percentage rate is fairly comparable.

Then with the cannabinoids, that seems to fluctuate somewhat from month to month. It is generally in the same ballpark with hair and urine. We certainly have months where the hair positives are greater than the urines.

Overall, there is about 40 to 50 percent greater frequency of drugs detected in hair, as total analytes rather than total samples, than what you see in urine.

Addressing the issue of quality control, as I mentioned earlier, controls for both the initial and the confirmation tests are selected so that they bracket the cutoffs, giving you some assurance of linearity around those cutoffs. The quality control program is based on the forensic urine approach that we use. It is intended to make sure of the correct results. The difference with hair, of course, is the fact that we are using two different immunoassays. Our experience has been that ELISA is somewhat more difficult to QC than RIA is. Both are probably more difficult than urine.

You have certain challenges with hair. I am not looking for sympathy but the drug concentrations are low, there are definite matrix effects which are seen sometimes which are variable from sample to sample and create kind of a challenging situation. There is a limited amount of specimen available. This isn't true for the QC samples but it is true for the patient sample. As I indicated earlier, the parent drug predominates in hair samples and there may be only minor amounts of metabolites. If you are going to consider metabolites in your interpretation then you have to sometimes have significantly better sensitivity for them to be able to detect them and validate their presence.

Our internal quality control samples are prepared using pure materials purchases as methanolic stock solutions. We attempt to obtain different suppliers of each material. If not possible, then we at least try to obtain different lots for the preparation of standards and controls. We verify the purity of the materials and their quantitative accuracy by GC/MS. For our open controls as opposed to blind controls, we prepare a hair extract matrix. We start by obtaining hair from prospectively drug-free individuals. A lot of time they are lab workers. We all know they don't use drugs, especially working tox lab.

Actually, if that were true, then we would never get any positives on our own employment screening. We added hair testing to our pre-employment armamentarium about three years ago and it was amazing, the increase in the number of positives that we got. It has been very helpful for us.

The matrix is washed to remove any contaminants, as it is in the normal analytical procedure. It is then sectioned into short pieces. It is analyzed using GC/MS and MSMS and must fall below the LOD for all the analytes. If this all happens the matrix then is extracted as in the procedure, two hours at 70 degrees with a solvent. That is evaporated and reconstituted with buffer and that is spiked with the analyte for the screening. Once again, we verify the accuracy using GC/MS. Test results are reported and the reporting actually is more reminiscent of our non-DOT testing process than it is federally regulated. With the regulated testing, of course, we don't do any telephone reporting, but we use all the modalities for hair testing results than we do for a non-DOT. We do report to an authorized recipient only. We report using a secured fax or printer and have to have a letter on file for that. In the local area, we use courier delivery. In fact, even up in northern Nevada where we have a significant presence we have the ability to deliver reports quite quickly because there are very frequent flights going back and forth. The manner of reporting these results, if a sample tests negative once, it is negative. Initial test negatives are reviewed by a screening operator and a certifier.

Positives are reviewed also by the operator as far as the confirmation itself, and by the certifier, who also, before reporting a positive, reviews the screening data as well. This data is released to the client or the MRO. Actually, this slide kind of continues. Various entities -- I think it would be fair to say that most of our hair results aren't MROd at this moment. But there is a significant fraction that are. They are reported to MROs, to physicians, to human resource managers, various clinics and third party administrators, government agencies and individual attorneys. They are reported via the usual mechanism, the printed report which can either be mailed, couriered or sent overnight, faxed or autodialed, and we have printers in some of the locations. It is done in a verbal manner as well.

I mentioned the fact that we only report confirmed positives positives. The drug classes, amphetamines, cannabinoids, cocaine and metabolites, opiates and phencyclidines are the so-called NIDA-5. We do identify the specific analyte detected on the report and we report primarily qualitatively based on our cutoffs, and quantitatively if the client demands that.

These are the substances we actually report. At the bottom PCP is cutoff, but with the PCP group, PCP is the analyte reported. For the amphetamines, methamphetamines and amphetamines, for cannabinoids THC and carboxy THC. For cocaine the cocaethylene and benzoylecgonine are also reported. Opiates, as I indicated, morphine, codeine and 6 AM.

Now we get to the subject of interpretation. Hair testing in the employment environment is a fairly new phenomenon and it is not presently subject to the kind of scrutiny and consistent guidelines that have been elaborated for urine testing. As you have heard, the different hair testing labs have differences in the way they approach interpreting issues. Philosophically, it is our intention to try to be conservative in our interpretation of our data and to try to stick as closely to the science as we can. Sometimes that conservatism is expressed on the interpretive level rather than the reporting level, as you will see. What follows is just how our laboratory and its clients deal with several of these thorny problems that you have heard about. You can kind of think about the interpretive problem on several levels. The first involves the analytical result itself, whether the drug is present at all. The second is to place the result in the employment context in which it will be used. The third is to consider the problem cases.

The first one is very similar conceptually to how you would solve the same issues regarding urine. That is based on the consideration of screening and confirmation data, requiring that all quality control requirements be fulfilled, that the chromatography be acceptable in terms of the analyte in the sample having the relative retention time that it should, and its peak shape, its peak separation requiring that the controls be within the acceptable range and that the ion ratios for the mass spectroscopy be within the acceptable range. Then last but not least, to evaluate it in comparison with cutoffs.

Last but not least, there is the issue of whether metabolites are present or absent. One of the things we have observed about metabolites in hair testing, especially if they are more polar in the parent compound, is that they are present in low amounts in relation to that parent compound.

So, while in the case of amphetamines we have adopted this criteria of not reporting meth unless amp is present, actually we are able to do that because amphetamine does occur in hair in reasonable concentrations. It is not as high as the kind of ratios you would typically see in urine, which in our lab recently we looked at that. It was about a four to one ratio of meth to amp. That suggests that somebody ingested methamphetamine and nowadays we seem to be getting a lot of pure amphetamines and 50/50 mixtures of meth and amp.Where somebody has ingested amp, it looks like the ratio is about four to one in urine, and in hair it is about 10 to one. That still allows an adequate amount that it can be detected.

With some of the other situations, cannabinoids, for example, the metabolite carboxy THC is so much more polar than THC itself, that there are some definite situations where it isn't seen in urine, even when we know the person is a user, and I will show you some of that data a little bit later.

One of the things that you can do is to base your interpretation on the purpose for which a test was performed. The considerations, as I alluded to earlier, differ greatly among different types of testing. In pre-employment the goal is to identify drug users with a high degree of fidelity but not absolute fidelity. I am kind of speaking from an employer perspective rather than an analytical perspective.

Many employers are less concerned with the issue, for example, of external exposure because they operate a high cash business and they feel that anyone involved in an intimate relationship with a drug user or exposed to large amounts of drug may be more tempted to steal than someone who is not in that circumstance. I guess in that sense the employers are applying the test results as a discretionary item like they would something like a credit report, rather than an absolute plus or minus.

I talked earlier about post-accident and reasonable suspicion testing. Whatever limited role hair testing may have in that realm is based on being able to establish that someone had a pattern of drug use in the past. It doesn't pertain to an individual incident. It is my view that if you have a situation such as I described earlier where someone returns to work after an injury, for whatever reason doesn't report the injury, doesn't get tested on a timely basis using urine and/or blood, that a hair test is not any kind of an adequate way for an employer to say that an employee was impaired by drugs at the time of an accident or incident. Random testing, while not recommended by our laboratory, is interpreted along the same lines as the pre-employment. If an MRO is involved in the process, then they would begin to be involved at this stage of interpretation, often with the help of the laboratory. Both the lab and the MRO need to be cognizant of an employer's policies and procedures as well at this point.

Of course, as there are with urine, there are acceptable explanations for a drug presence in a sample, but they typically involve an individual having a valid prescription or other legitimate source of the drug. As far as I know, the issue of poppy seed ingestion as an explanation for morphine in hair has never really been the kind of problem that it is in urine. In fact, I am hard pressed of a single case where we had a morphine level, didn't have any 6-acetyl morphine, didn't have any codeine where you could attribute that perhaps to poppy seed ingestion. I think the reason for that just has to do with the polarity of morphine relative to codeine. As I mentioned earlier with some of the other compounds, morphine is a lot more polar than the parent compound, codeine, or heroin, for that matter, 6-acetyl morphine. So, it doesn't gain access to hair in the same proportion. It is not too uncommon for us to find codeine and very little, if any, morphine.

The issue of external contamination does come up with donor explanations. The MRO in many cases, as he does have to do with urine, has to assess the credibility of that source of information in terms of whether somebody just says that they went to a bar at which crack cocaine may have been smoked or marijuana may have been smoked, as opposed to a situation where they live with a drug user. But that is something that can be considered. There are various ways to deal with that in concert with MROs. We have at times -- I should indicate, before we even get into the external contamination issue, kind of where we stand. Based on the data that is available today, I don't think that you can flatly state that somebody is a drug user and was not externally contaminated based upon the finding of a parent drug in hair alone. You can come close to doing that with metabolites but there are still some unanswered questions in that realm. There have been situations where, with employers that have extremely credible donors, we have recommended that if they wish to hire that individual, that they hire them subject to random urine testing over a period of time, such as six months. I am uncomfortable still, I guess, with just flatly stating that somebody is a drug user, even though obviously the majority of people who test positive are drug users.

Here are some of the kinds of parameters that enter in. Again, as I mentioned, metabolites, other than cocaethylene, are generally more polar than the parent compound and therefore occur in much lower concentrations. It is relevant whether metabolites are present. We find for benzoylecgonine, for example, that in the vast majority of our samples it is present at a concentration greater than five percent of the cocaine level. Cocaethylene is present about 30 or 40 percent of the time. So, there are times where it is not helpful. Where it is, I think I would go beyond the statement I just made that if cocaethylene is present, clearly that did go through the individual's body, even though you can think of some scenarios where that wouldn't have had to happen. Quantitation, was it a lot or a little. Without really having data, I have the feeling that if somebody thinks they were passively exposed that they should have a lower rather than a higher level of the drug in the sample. The context, too, that depends on the explanation that the donor is able to provide.

Additional facts that have to be interpreted generally with regard to hair testing include the source of the hair, whether it is head hair or body hair, or the treatments that that may have been subjected to. I would like to take a few minutes to talk about two of the big issues, external contamination and hair color bias. The external contamination issue is based on the fact that several workers have shown that drugs can enter the hair from the outside, either in the form of smoke or the form of aqueous solutions or solvent solutions for that matter, and not just from the blood supply to the hair follicle during hair growth. It is also believed that one of the key mechanisms by which drugs can enter the hair is through the perspiration and skin oil, in which fluid the parent drug predominates.

There is kind of a scientific debate, as you all know, between two philosophies in terms of whether what is known as wash kinetics can be employed to distinguish between use and external exposure. I think in a practical sense wash kinetics probably works. That is probably because most of the samples that test positive is from users. If you turn that equation around, as Dr. Kidwell has done, and you actually run through this system samples which were intentionally externally exposed, you find that the discriminating power of the technique is not complete. My philosophy on that is if it doesn't work virtually 100 percent of the time, what good is it. So, we don't employ that in our laboratory. We do wash our samples, but we don't test or try to interpret the wash solutions. We have done that on a research basis. We certainly find that in a predominant number of cases that the concentration of drug in the hair is vastly greater than that which is in the wash solutions.

I think that to use that kind of a protocol and believe that it works a high enough percentage of the time is not something I am prepared to do right now. In regard to hair color bias -- and I have a couple of slides here, I think -- this one isn't visible anyway but it is available in your handouts. This is a comparison of cannabinoids. We basically took urine positives on people that had been tested with hair and urine for pre-employment. Since they are positive in urine, we know they are users. But about 20 percent of the time they didn't have THC in their hair. That illustrates the kind of interpretive problems you can get into. This is some data from Dr. Cone's group showing binding of cocaine to powdered hair samples, and indicating that male Africoid hair has a much higher level of cocaine binding.

This type of experiment and both in vivo and in vitro experiments have indicated that there is the potential for hair testing to fall disparately upon different populations. However, the real world situation is a little more confusing, as Mitchkowski's paper a few years ago showed, on recent arestees in Florida.

He found, indeed, that African Americans had a higher proportion of positive samples than Caucasians did. But that was true for both urine and for hair. Actually, the factor was somewhat greater for urine.

Recently we have been doing a study in our laboratory correlating the percentage of positives by hair color for various drug groups and comparing that with the overall distribution of hair colors that we see in negative samples. I am going to present this data in more detail at the CAT meeting in November in Las Vegas. The bottom line is that we do see, as you would expect with cocaine, an enrichment of positives in the darker hair colors.

When you look at methamphetamine and cannabinoids, that pattern is reversed. Either that means that blondes do have more fun, or perhaps it means that we need to consider an alternate explanation besides hair color bias in the physical chemical sense. What I allude to there is basically differences in drug preference by ethnic group.

I just have one slide. We do have an external quality assurance program where samples are submitted blindly. We prepare the samples. We spike them from dilute aqueous solution, wash them extensively and shampoo them and try to make them in every respect representative of actual donor samples. Then have them blindly submitted by an external agency to our laboratory. About one a day is submitted and we have about one batch of hair tests a day, so that works out well. I would be happy to entertain questions.

DR. BAUMGARTNER: Dr. Kelly, I assume that you are basically using the non-forensic NIDA method for washing? What is your wash procedure?

DR. KELLY: It is basically a methanol wash, very brief. We have definitely shown that in a spiked sample situation that you can remove 80 or 90 percent of the total. We don't go through the sequential process that you do.

DR. BAUMGARTNER: I think the basic issue is that the more aggressively you wash, you know, the more effective you are in removing any external contaminants. Then, how long do you wash?

DR. KELLY: Five minutes, at ambient temperature. With the marijuana, the cannabinoids, both for screening and confirmation, we do use a multi-step process with phosphate buffers. So, it is a little more aggressive, as you would put it. I don't really want to re-debate the whole wash kinetics scientific controversy today. I guess I could say I don't really find your data convincing. In science something has to be repeatable in order for it to be valid. It may work well in your situation. I am not saying it doesn't. But somehow, the other workers who have tried it have not been able to reproduce that level of success. Even if it works a high proportion of the time -- and again, as I indicated I suspect that it does because I think mostly you are dealing with samples from users in an employment population -- I just wouldn't want to hang my hat on it.

DR. BAUMGARTNER: I think you just used the key word, reproducible. The study the navy did, did not use our method, as you know.

DR. KELLY: I have heard you say that before, sir.

DR. BAUMGARTNER: It is in the literature. They spelled it out in every detail, where it differs. So, they did not evaluate what we are doing. Secondly, there is nothing profound about wash kinetics. It simply tells you how you are doing with the wash, how close you are of not being able to remove any more material, how close you are to the plateau conditions, how effectively you have removed drugs from the accessible and semi-accessible domain. As you know, by dissolving the hair, we are only focusing on the inaccessible domain. If you washed very effectively, I think all the samples would be false negatives. You need to have a more aggressive extraction procedure, preferably one that dissolves the hair sample.

DR. KELLY: I appreciate your suggestion.

DR. BAUMGARTNER: I also suggest that you may want to follow the international consensus on hair testing, where in Europe water washes were suggested.

DR. KELLY: Thank you for your comments.

DR. CAPLAN: You skipped over quickly one slide, which I don't know if you want to make a point on, that wasn't in the handout. It seemed to be a table of data from Mitchkowski with some data on the right-hand side. Was there a point to be made with that?

DR. KELLY: Yes, I guess the point was the same one I made about the fact that the data that I provided, which was equally invisible as well, maybe more invisible, our data looked at individuals who had had hair and urine collected at the same time, and had tested positive for cannabinoids in urine, and then we looked at hair as to whether THC was present, carboxy THC was present. THC was present 100 percent of the time and carboxy THC was present only 80 percent of the time. Those, again, are known users. What we find overall is that about 30 percent of the time THC is present but not carboxy THC. So, the uniqueness of that study that we looked at was that someone had evidently, you know, not only had a history of marijuana use, but they had used recently enough to where the urine tested positive. So, that is a subset of total users. Inasmuch as urine drug testing has been called a liquid intelligence test, maybe they are not representative of those people who test positive in hair and not in urine.

The same take-home lesson can be derived from Mitchkowski's study in which he looked at the hair screening value, the hair confirmation value and MSMS value for both THC and carboxy THC. I don't think I have that thing with me up here. In essence, it turned out that the vast majority had both carboxy THC and THC, but there was about 20 percent or so that had the THC, not carboxy. Then there was about two or three percent that had carboxy but not THC.

We happen to have a reporting criterion in which we don't report positives without THC itself. So, that is kind of a reversal of the usual situation.

These are the two Mitchkowski references, Forensic Science International in both cases, volume 63, page 85 to 98. That is 1993. Then volume 70, page 83 to 91, 1995. That is where those two slides came from. I apologize for not including that. I think you get the same basic take home message from his study that you are not always going to see carboxy THC, even in known users.

MR. VELASCO: When you do your external QC, the samples are submitted from an outside clinic that collects the sample -- not really collects the sample, but has a sample. These samples, are they spiked samples that they submit to you? If so, how do you prevent the drug from being washed out during the wash procedure.

DR. KELLY: Our experience has been that when you prepare a spiked sample and you wash it extensively and you use the kind of hygienic washing procedures that people would typically use on their own hair. You can create a sample that is actually very much like an actual user's sample. That is how we do it. We also submit a certain proportion of user samples, particularly for cannabinoids. For other drugs, that spiking procedure that I outlined seems to produce a sample that is hard to discriminate.

MR. VELASCO: What is your rate of success with your external QC.

DR. KELLY: In terms of getting positives?

MR. VELASCO: Yes.

DR. KELLY: I forgot to mention that 80 percent of them are negatives and 20 percent of them are positives. I think generally speaking if we have a problem it is false negatives. We haven't had any false positives to date.

DR. SELAVKA: Several of my questions can probably be answered easily. Have you published your method anywhere?

DR. KELLY: Not as yet.

DR. SELAVKA: Are you planning to?

DR. KELLY: Yes, we are planning to publish some of these methods. The first one will probably be the ELISA amphetamine method, which we have a manuscript in preparation now. We are also going to discuss the methods in more detail, as I mentioned, at the CAT meeting in November, the first week of November.

DR. SELAVKA: I appreciate that. If you don't mind, I will just ask you a couple of things that weren't clear from the slides.

DR. KELLY: Sure.

DR. SELAVKA: In your cross reactivity data for the THC ELISA you did not list the cross reactivity of the THC itself. What is that?

DR. KELLY: There I was kind of going with the manufacturer's data, so I don't know. I expect it would be, number one, low, and number two, the concentrations of carboxy THC are very, very low. So, I wouldn't expect that to be a useful thing to look for in terms of screening. Most of our levels wind up being in the low picogram per milligram level.

DR. SELAVKA: Your cutoffs weren't listed anywhere. Did you mention them?

DR. KELLY: I did mention them. I just realized last night, hey, I don't have a slide for my cutoffs. I did mention them, basically 300 picograms per milligram for everything except opiates which is 500 and cannabinoids which is 5, THC itself. The cutoff for carboxy THC is 0.1 picogram per milligram.

DR. SELAVKA: In your discussion of the preparation of the positive QC material, I am not sure I understand your answer to Javier's question. In that NIST publication where they basically created positive QC material for the interlaboratory material for hair drug testing results, they kind of demonstrated that just simple spiking onto the hair does not give hair that is going to act the way conventionally created positive hairs do in drug users.

They had to use very rigorous conditions with DSMO soaping and so forth. So, I have a feeling -- this is more a comment than a question -- that your simple methanolic rinse is not doing the job it needs to, to remove external contamination.

I think further, if you don't use a rinse protocol because your feeling is that it is not 100 percent effective, in your opinion, in removing the potential for an external contamination to give rise to positive findings and the eventual extraction in tests, I still think that it gives you very useful information for interpreting positive findings. Just dismissing the process out of hand, I think, is dangerous. I think you should generate the data that would allow you to better interpret those results which may fall into question.

Again, it is a comment, but I think you are mis-serving the clients by not performing rinse protocols on these hairs.

DR. KELLY: Obviously, I don't agree with that. As I said, we have done it on a research basis. We are comfortable with the fact that we are removing the vast majority of external contamination. I don't think the data would permit anyone to state that could be done quantitatively. Maybe we can discuss this later. I would be interested, particularly since you used to be involved in this business yourself, as to what sort of interpretive criteria you think are appropriate.

DR. SELAVKA: You can read my publication on that in the NIST volume. Finally, you did have one statement. You said you interpret hair drug positive findings as having arisen from multiple uses except maybe cocaine. I wonder how you have drawn that conclusion for cocaine for the 3.9 cutoff. I guess for the sake of the board I would commend to you a study for positives in hair. It is a very well thought out demonstration project on his part of reviewing the literature and looking at these critical issues. Thanks very much.

DR. JONES: First, thanks for coming and sharing with us the information on you protocols. In one of your later slides that was entitled, internal QC challenges, you had a statement there, variable matrix effects. I wondered if you could expand on that or if you had any further expansion on that other than your hair color which you made later on.

DR. KELLY: In the initial stages of developing our protocols, we discovered that it was desirable to, particularly for cannabinoid analysis, to have a wash procedure, a several step wash procedure. That had to do with matrix interferences that we saw in the ELISA assay. Basically, those produced false positives, or positives that couldn't be confirmed. We have been able to reduce that significantly by having a series of phosphate methyl washes coupled with the methanol wash and don't seem to see that any longer.

It is a matrix effect. People put a variety of things on their hair. So, there is always the chance for a much less homogeneous sample that you have with urine, that you are going to have these kinds of problems.

MR. SHULTS: First of all, let me say that I appreciate your candor and objective presentation on this area. As you know, this whole issue raises a number of conundrums for medical review officers.

I have two questions for you. If I am the medical review officer and I am trying to evaluate what is going on with a donor, and I have got a urine specimen for amphetamines and I have a positive hair test. Here is the way the facts come out. The urine comes out negative but I have a positive amphetamine hair test.

My usual instinct would be -- my donor is telling me, they use Vixen Hayler. That is what they are using. So, my question is, at 300 picograms per milliliter, are we able to do a pilule(?) isolation to determine whether we are talking about DNL at that level?

DR. KELLY: Remember, we would be talking about 2,100 if it was L, based on the cross reactivities in the screening test.

MR. SHULTS: So, it is achievable?

DR. KELLY: Yes, if that happened, yes, you would have no problem doing that.

MR. SHULTS: The other issue that comes up is not necessarily an MRO issue. If you have a new client that wants to do hair testing, how do you advise them when they ask you this question: I heard there is a racial bias with this testing methodology. Will I be in trouble? Will this cause me any legal problems if I do pre-employment hair testing? What do you tell that client?

DR. KELLY: To date there hasn't been a problem. There haven't been any cases. There is not any evidence that it is a real world phenomenon.

I show them my data on hair color. You can't really conduct a study today in the private sector, at least, in terms of race, per se. You can look at hair color. We are in the process of trying to collect data on a control group which is basically urine test results with associated hair color. We had to do this prospectively once we saw this pattern that I described earlier. That is an important control to be done.

MR. SHULTS: I realize that the courts and the laws don't have a great deal of legal responsibility for interpreting their results. Maybe I would walk in that situation.

DR. KELLY: That is really not what I was advocating. I think we lean toward looking for metabolites as an indication that someone was using.

MR. SHULTS: In the absence of finding metabolites, though, that is not an issue that gives me a great deal of legal concern, given the fact that an employer can be discretionary as long as it isn't along protected class lines.

I do have a little consternation when we are talking about something that potentially raises, whether it is scientific or not, the issue of a Title VII violation. I think, to me, it is not solely a scientific issue. We had this question a couple of months ago. Is nicotine addictive.

What kind of scientific question was that. There is no question that nicotine is addictive. The issue that was raised in Congress is that it was an admission against interest.

Now we are going to change our attitudes about cigarette smoking and the admission was, as a social code, was cigarette smoking okay.

The CEOs of those companies said, no, it is not, despite the fact that every piece of scientific evidence and the fact that everyone who ever smoked a cigarette knows it is addictive.

The question of, is there a racial bias to drug testing is not a scientific question, although we approach it as that.

The question I have -- some of you may have heard of this case -- California versus OJ Simpson. The issue there was not almost an issue of guilt or innocence. It was a question of, would the jury accept the prosecution's attempt to mask the presence of a racial bias with one of the investigating officers. They made a strategic error in trying to ignore that issue. Not to make this extend too far, but the question is, what would Johnny Cochran do with the amount of evidence that we have, making a case that this is a Title VII issue. Again, it is not a question of, is this a real world scientific issue. My concern is, I know what Johnny would do if he gets the right jury composition in a Title VII case. Again, it goes kind of beyond where we are at with this thing. But the issue that will come up -- and it will come up eventually in Congress -- is, is this racially biased or not.

It won't be a scientific question. But my concern is, will that adversely affect other kinds of drug testing. Will the public be able to distinguish this. Don't get me wrong. I actually would like to see hair testing. It is a powerful technology and it has a lot of utilization. I am concerned with how we are approaching this issue and whether we are overlooking the fact that it is more than science that is involved in all of this.

DR. CHILDS: Good morning. That was a very thorough presentation. I had a couple of questions I wanted to ask. The point is, could you give those of us who are often frustrated by doing these screening procedures that do not confirm by GC/MS some sense of the samples that actually screen positive using your technologies, and it may be by drug category, how many of them have identifiable, reportable GC/MS verified drug in them?

DR. KELLY: Off-hand, I would say about 60 to 80 percent. Most of them, about 60 to 70 percent for cannabinoids, a little over 80 percent, I think, for cocaine. You said reportable. That means above the cutoff. We have some that will screen positive that don't confirm, but the compound is there, whether it is a metabolite cross reactivity or whatever. It is reasonably good except where you have disparate groups such as opiates and amphetamines. The confirmation rates, since we are focusing on codeine, morphine and 6-acetyl morphine, are more like in 40 or 50 percent. In our area, hydrocodon is extremely widely used. We don't at present have that as part of the opiate confirmation procedure for hair, although we do in urine. Perhaps that explains the discrepancy. PCP is about one to one. I don't think we have ever had a screen positive PCP that failed to confirm. That amounts to just a couple dozen samples in the past year.

DR. CHILDS: Just another question. How did you establish your cutoffs for confirmation and screening?

DR. KELLY: They were kind of market driven. Then we kind of retrospectively looked at the amphetamine one because, again, we are getting ready to publish that one. We are going to try a similar one, an ROC approach, for the other assays. Unless that analysis shows that we have too low a cutoff, I think we will probably stick with what we have, because it is kind of out there. The clients expect it now.

DR. CHILDS: I am kind of curious what you mean by market driven.

DR. KELLY: I mean, we copied Psychemedics, basically. Let's be blunt about this, improved them a little bit, in a couple of cases. That is about it.

DR. JONES: Since I have the microphone and Dr. Autry has signaled me, I think it is time for a break. Thank you very much, Dr. Kelly, for sharing that information. Fifteen minutes.

(Applause.)

(Brief recess.)

DR. BUSH: Thank you for your attention. I would like to present our next speaker, Dr. Christine Moore.

Dr. Moore, since 1994, is the laboratory -- she is now the laboratory director at U.S. Drug Testing Laboratory in Chicago, Illinois, and she has been working there since 1994. Prior to that she was technical services director at United Chemical Technologies in Bristol, Pennsylvania. She was also research associate at the College of Pharmacy, University of Illinois, Chicago, and she enjoyed a post-doctorate fellowship in the Department of Legal Medicine in Nagoya University in Japan. She was the first ever recipient of a joint Royal Society/Japan Society for the promotion of science fellowship in the field of forensic toxicology. Dr. Moore received her bachelors degree with honors in applied chemistry from Salford University in Manchester, England, a master of science in forensic science from Strathclyde University in Glasgow, Scotland, and received her PhD in forensic toxicology at Glasgow University in Glasgow, Scotland. She is currently the vice president of the Midwest Association for Toxicology and Therapeutic Drug Monitoring. She is a member of many societies, including the American Academy of Forensic Sciences, Southwestern Association of Toxicologists, the American Chemical Society, American Association of Clinical Chemistry, International Association of Forensic Toxicologists, Society of Forensic Toxicologists and the Royal Society of Chemistry. Dr. Moore has published numerous peer reviewed articles in the areas of analytical forensic toxicology, particularly in the analysis of drugs of abuse in urine, meconium and hair. I present to you Dr. Moore.

(Applause.)

Agenda Item: Analytical Procedures for Hair Testing.

DR. MOORE: Thank you. Evidently I have already made a mistake. I was told that I shouldn't be wearing white after Labor Day. I apologize to all the fashion conscious people in the audience for my white jacket.

I would like to thank Donna and the Drug Testing Advisory Board for this opportunity. My discussion today will involve the analysis of hair for drugs of abuse.

It will take the following format: specimen collection and chain of custody, initial and confirmatory procedures, internal and external quality control, and reporting, interpretation and some current areas of concern in hair testing.

Our clients are instructed initially in the site of collection from the head. We rarely receive hair from any other body part. I have written alternative body areas there, but I personally have never told anybody how to collect any pubic hair. To me, that is pretty invasive and one of the issues with hair is its lack of invasion of privacy. So, we will go with head hair at the moment. Sterile scissors and gloves are worn. The collector is shown how to align the hair in aluminum foil and align it along the markings of an envelope so that you know which end is the root end. That foil is then placed inside the envelope. The donor and the collector sign. This is very much like Dr. Kelly explained to you with the diagrams. Then both the donor and the collector sign the outside bag, which is sent to our laboratory.

At the laboratory we continue the chain of custody as required by our own standard operating procedure. We check the seals, the signature. We check that there is enough hair to actually do a test. We record the length of the hair but we only use the three centimeters closest to the root end. The hair is washed in water, methanol and acetol. That is to remove dirt or any other kind of powder or extraneous obvious material. We then powder the hair. This is to homogenize the sample. One of the either advantages of disadvantages, depending on how you look at it, with hair is that it is heterogeneous.

If you want to do some kind of sectional analysis so that you can see a time frame of drug use in the hair -- like look along the hair shaft, then that is great. Some drugs do lend themselves to that type of analysis in the hair. If you want to use the same sample or the same exact specimen for both screening and confirmation, your best chance is to homogenize the hair. So, we do that and we divide the powder into separate aliquots, one for screening and one for confirmation. Then we store any excess powder or hair at room temperature, which is another good advantage to urine testing.

After washing, the hair is hydrolyzed in acid overnight. The acid is decanted and the remaining hair is destroyed in sodium hydroxide. After cooling, the base is neutralized and the supernatant is added to the acid from the overnight incubation. So, we have got them both in the same tube. The whole sample is then extracted using mixed mode, solid base extraction. The mixed mode extraction allows both acidic and basic drugs to be extracted and then eluted into the one tube. This is actually the procedure and it is in your handout.

The combined element is evaporated to dryness and reconstituted for screening analysis. This might seem like an awful lot of work to you for a screen, and it is. It definitely is. But because we only have fluorescence polarization immunoassay capability -- we don't have any of these really sensitive RIAs -- what we have to do is get the sample as clean as we possibly can before we do the screen. Otherwise we wouldn't be able to even approach the type of sensitivity that we need for the screen. That is the reason that we do it on every sample. Because cocaine is the predominant compound that is found in the hair after cocaine use -- that is by far the highest number of positives that we see -- we are looking at other technologies at the moment, because the FPIA is specific for benzoylecgonine. Specifically we are looking at microplate technology. We really like the results we are seeing on that at the moment and hopefully we will be switching fairly soon.

Our confirmatory test procedures are similar to the screening except, of course, for the addition of internal standards. There is no reason to carry out either the basic digestion or the acid dilution for cocaine, amphetamines or PCP. After drying, the cocaine extract is derivatized with MTBSTFA, which gives you the tissue butyldimethylsilol derivative. Amphetamine we use the HFBA derivative and PCP, of course, doesn't require any derivitization. Analysis of these compounds is in standard procedures, electron impact selected ion monitoring GC/MS. We use that the same as Dr. Kelly, plus or minus 30 percent.

The confirmatory test for the opiates is actually a little bit different. We just started working on this in our lab.

I attended a meeting of the European Society of Hair Testing in Strasbourg in France in June, which is actually a very nice little place. I would recommend Strasbourg, actually, to anyone. It is very nice. I discovered that some people add methanol to their hair. They would ultrasonicate that for a couple of hours, pour off the methanol, and use that as their hair extract. That is really good for codeine and monoacetyl morphine. It isn't very good for morphine. They are sacrificing the recovery of morphine. Other people would do the overnight acid hydrolysis which is like a screen. That would give you good recovery of codeine and morphine but you would sacrifice the acetyl morphine. What we have been working on is combining these two procedures. So, we put our hair in the methanol for a few hours, for three hours, and sonicate that, pour off the methanol, add the acid, do the overnight hydrolysis and combine the two. It seems to be working relatively well at the moment. After drying, the opiates are derivatized with BSTFA. Again, we analyze those with regular electron impact GC/MS.

Now, I think that most laboratories could very easily adapt their procedures to analyze for these four groups of drugs we have talked about.

The problem in the United States is THC and carboxy THC. The Europeans don't seem to have any problem with this. I have no idea why. Maybe they have different marijuana or something. They seem to find THC and carboxy with electron impact systems with no problem at all. However, in the USA, the concentrations that we find of carboxy THC, in particular, are very low.

Our extraction procedure is a modification of the isolation of both of these compounds from blood. We just modify that blood procedure. Effectively, we add acetylnitrol to the hair, sonicate that for a couple of hours, pour that off and do solid phase extraction. Are you getting the theme here? I love solid phase extraction. Almost everything in my lab is solid phase extraction. These are the conditions that we use. I won't go through all of those things. It is in your handout. After evaporation, we derivative with MTBSTFA again to give us the tertiary butyl dimethylsiol derivatives and analyze with positive ion, chemical ionization, GC/MS MS.

I think Psychemedics report .05 picograms per milligram -- correct me if I am wrong -- for the detection cutoff for carboxy THC. APL, as we heard, I think was .3. I am sure you will correct me, but it is very low picograms per milligram for carboxy THC. I think both of those laboratories use negative ion CI MSMS. That is obviously more sensitive than our positive ion. We were only able to get to five picograms per milligram, which is a good hundred times less sensitive than Psychemedics. The system that we have is a Saturn 3400 GC and a Saturn MS. When you buy one of these systems, they will sell you options and upgrades, which essentially consist of three chips that they will sell you. One chip is for what they call selected ion storage, which is a similar concept to SIN. One chip is for CI and one chip is for MSMS. We are fortunate enough to have all of those, but some of the people with this instrument wouldn't necessarily have this capability without the upgrades. The ions which we isolated, 376 for deuterated THC, 373 for unlabeled THC, 576 for carboxy deuterated and 573 for unlabeled. Those ions are isolated, fragmented, and the daughter ions that we look for are 352 and 338 for deuterated THC, 349 and 335 for unlabeled, 560 and 444 for deuterated carboxy THC and 557 and 441 for the unlabeled. We used acetone as the reagent gas, in case you are interested. Our internal quality control is very simple. We maintain the chain of custody procedures and we run positive and negative controls at a rate of 10 percent in our screens and confirmations. The control data is recorded, monitored for any kind of significant fluctuation. That information is available to any kind of licensing agency or inspection group. In fact, we had an inspection last week. It was a general lab inspection by CAP, and they asked to look at that data.

For our positive controls, we use negative hair, and it is mine. As you can see, I have short hair. I know I have no chance ever of growing my hair, because when they get short of negative hair, I have to go get my hair cut. It is pretty much all mine and they powder it into a big bottle of powdered hair. We use that to spike positive controls. I wish that we had access to real drug users, but we don't. We have no way of getting drug from actual real drug users for positives.

For screening we use a six-point calibration curve and for confirmation a five-point curve, usually about every three months. With each batch we run a single calibrator at two nanograms per milligram. The lab staff are judged as in accordance with licensing agency suggestions on a regular basis. Externally, we are enrolled in both available proficiency testing programs. Those are the National Institute of Science and Technology and the Society of Hair Testing. So far, the National Institute of Standards and Technology have only included opiates and cocaine in their proficiency samples. Some of those were intact hair and some of those were powdered.

In 1996, the Society of Hair Testing, for the first time, included amphetamines in their proficiencies, and those are all powdered. So far, no group has included THC, carboxy THC or PCP in their proficiency programs. I know we are going to have a presentation this afternoon on proficiency and samples, but I think there are some questions that need to be resolved along these lines. As I mentioned before, hair is very heterogeneous. If you are going to send the same sample exactly to every lab, you pretty much have to homogenize what you are going to send. That is why powdered specimens have been so popular. What they do in Europe is take hair from dead drug users, so they get a lot of it. That is what they tell me. They homogenize that for like hours or days -- that is what I am told -- in some big kind of caldron or something like that, and then they powder it all and send it out to the labs who are enrolled in their program. I think it is a good way to do it and it takes in a lot of variables and heterogeneous nature of the matrix and things like that. You can't send it to the labs that don't use powder as part of the SOP. Obviously, they are proficiency samples. Obviously, they can't treat them in the same way they would a normal sample.

As far as I am aware, Psychemedics and APL, who are the biggest labs that do this in the USA, don't have that as part of their SOP. I think that is a problem. Let's suppose that you did actually use real hair that was spiked instead, so that the labs receive something that looks like a real hair coming in their door as a proficiency sample. If they spiked it, we have the same issues of different labs washing different ways, will wash off different amounts. You are going to get a whole lot of different results. Unless you mandate a washing procedure also, that is a problem. If you use the wash kinetics, they are all going to be negative. Thirdly, the Society for Hair Testing sent out initial of their samples to three reference labs. The problems they had was the three reference labs couldn't agree. They really couldn't grade the performance of everybody else in the program because they didn't have a good result from the three reference labs. If you do that here, who are going to be the reference labs? As far as I know there are only six, maybe seven labs that do this on a routine basis. Presumably they will all be in the program. That is something to think about.

Finally, of course, marijuana, what kind of levels do you spike at. As far as I know, nobody dies of a marijuana overdose. I don't really know what you are going to do with the marijuana in the proficiency program. I am hoping that Dr. Goldberger will answer some of these questions this afternoon. In urine testing, the cutoffs, as we heard yesterday, were assigned for various reasons. One of the implications was that below that level would be passive exposure and above that level would be active use. We all know that that isn't foolproof either, but at least that is a concept that everyone understands.

In June in this meeting in Strasbourg, the European Society decided that they would have no further discussion of meaningful cutoff levels for hair, which is very interesting. The acceptance of cutoffs above which would mean use and below which would mean exposure, to my mind, for hair, is nonsense. I think that I am in a bit of a unique position to say that. The vast majority of our samples that we look at -- and I will get to that in a moment -- come from children in households who have been exposed to drugs, where the parents are users and these children are exposed. I see levels all the time above these proposed cutoff levels, from individuals who have been exposed, not users. I guess you could say, how do you know these children don't use drugs. Well, I don't. I wasn't in the house.

Dr. Smith and I think Dr. Kidwell published some work from Alabama where they looked at hair, skin swabs, saliva and urine from children in this situation. With some outliers, predominantly they found the urine and saliva to be negative and the skin swabs and the hair to be positive, which would suggest that there is exposure and not use by the children. What we do is, we don't have cutoffs. We have limits of detection. This is the level to which we can reliably take our technology on a daily basis.

Our limits of detection are .05 nanograms per milligrams for cocaine, opiates, amphetamines and PCP and .5 and .05 for cannabinoids. For confirmation the levels are .05 nanograms per milligrams for amphetamines, opiates and PCP, .01 for cocaine and .005 which is five picograms per milligrams for THC and carboxy THC. At USDTL we do very little interpretation of drug results. A positive means the subject either used or was exposed to drugs in the last three months. A negative means the drug is either not present or is below our limit of detection. We don't use MROs because of the nature of our client basis. It simply is Child Protective Service who get our results. What we do include in our report is the signature, the amount of hair, the analytical methods, the limit of detection, and quantitative amounts, if that is required. We don't include any caveats for hair color or interpretive statements. The reports are sent by fax and by mail, secure fax, and we keep the positives for a year and we keep the negatives for two weeks to a month.

That is a pretty brief overview of what we actually do, the analytical methods that we use in our lab. I think that hair has some really good advantages. Certainly in terms of the historical aspects of the sample, you can't just stop using drugs a couple of days before the test. Certainly there is the ease of collection, the lack of invasion, the ease of storage, the ease of transport, all those things are really good advantages over urine. Nothing is perfect. Urine is not perfect either. I would say in hair we certainly have, I would say, three areas of concern that we need to look at. These would be things we have heard about all day long, environmental contaminants, racial bias and the adulteration of specimens.

It has been shown by numerous researchers that drug powders and smoke can be incorporated into the hair. It has been suggested that one can distinguish between active use and passive exposure using the wash kinetics that we have heard about. Unfortunately, as Dr. Kelly said, no one has been able to reproduce this. Therefore, you don't have a scientific consensus on whether or not this type of approach actually works. There have been suggestions that we would use cutoff levels. You have just heard my reasons for thinking that is not a good way to good. The presence of metabolites or parent metabolite ratios is a good suggestion. I think it has some merit. Benzoylecgonine, unfortunately, doesn't prove cocaine use. Cocaethylene is probably a very good indicator of use. So is carboxy THC. I don't think you will find that outside the body.

One acetyl morphine, not particularly. You might fine that in heroin street samples. So, what we are looking for is some unique compound synthesized only by the body that would be good markers for actual drug use. It has been shown both in human and animal studies that basic drugs -- not acid drugs, not marijuana -- will incorporate to a higher extent into pigmented hair than non-pigmented hair. Henderson et al showed that cocaine concentrations in hair in Hispanics, Indians and blacks were higher than those in Caucasians who received the same dose of morphine.

Gigy et al -- this was animal work -- showed that basic drugs incorporate more easily into pigmented hair than non-pigmented hair, giving you higher concentrations in the darker hair.

I was asked to give this talk on very short notice. So, I don't have really up to date data. I apologize for that. I am in the process of making this data much larger and much better. But I just want to illustrate a point with the next few slides. As I have said, we carry out predominantly child protective service nature types of assays. We are actually trying to determine exposure and/or use, but predominantly exposure. The children are not tested randomly. There always has to be some other reason that the Child Protective Service will go in and take a hair sample. The protection of the child is actually the intention of the whole program. Most of our clients are from Iowa, or the midwest. There was one county that agreed to give me information on the race of the people tested and the color of their hair as was determined by the collector at the time of the collection. Between January and May of 1996 we did just 37 hair tests for this particular county, just one county. Fifteen were positive and 22 were negative. The age range was from four months to 32 years. Obviously they included a couple of the parents. Of the 15 that were positive, 12 were from black people, two were biracial and one was Hispanic. That didn't turn out very well, did it. You have it in your handout. In fact, no whites tested positive in these particular tests. These are the negative results. You can see that we actually did some tests on white people in there. They just didn't test positively.

Then I asked the chief investigator to tell me what the hair colors were, as determined by the collector. They had a few options. What they chose was blond or light brown, brown/red which I think is my color, brown, black and what they call brown/black which is that they couldn't decide if it was brown or black. That gave the highest number of positives, the brown/black. In the negatives, the highest number of negatives was blonde hair, light brown, and brown hair. That is much better in your handout than on the screen. Until yesterday I thought everyone was in agreement that there was a racial bias in hair testing. I thought that we only disagreed on what to do about it.

I am wrong. I heard two people talk yesterday who seemed to think there was no bias, which to me is just kind of ignorance of the literature. Obviously there is some kind of bias in pigmentation here. I think the problem is how do we deal with it and how differences come and how to eliminate this type of bias and normalize the results. Dupont and Baumgartner suggested that the removal of melanin from the hair would normalize those values. To me, this sounds like a great approach. This sounds like a great idea. If drugs bind to melanin, we spin down that melanin, we normalize the amount of drug in the hair. It sounds like an excellent approach. Unfortunately, I have just read a paper by Joseph, who concluded that the removal of melanin has no effect on the bias. Again, we have a polarized community. We have no consensus what to do about this. What we need is normalization of basic drug concentrations, a similar concept to creatinine in the urine. We need some kind of normalizing factor. I do know a research group working on this. They have some really good ideas. I don't think it will be too long until we have something like this that is acceptable by the community in general. Acid drugs don't seem to produce the same kind of bias, which kind of suggests that drugs incorporate the ion exchange into the hair.

The third area that we haven't heard a whole lot about is adulteration of specimens. Scaup et al recently concluded that cosmetic treatments which included bleaching and permanent waving of hair caused loss of drug that is already in the hair. Now if you take drugs after you bleach the hair, then it seems you have a bigger chance of incorporating more drug into the hair. Confusing? Henderson et al then showed that is true; bleached hair does absorb more drug. But you lose more of it in your sample preparation. A lot of it washes out. Then of course we had Seles who said that the concentration of cocaine in your hair is dependent on the type of shampoo that you use. If all these people are right, are we going to have a cutoff for people who use Head and Shoulders and a cutoff for people who use Pert Plus and a cutoff for black hair and a cutoff for white hair. It is ridiculous.

It is pretty easy to adulterate a urine sample. We all understand that it is not a perfect specimen. At least there are some kinds of controls. You can check pH and you can check specific gravity and you can check creatinine, whatever it is. I don't think you can deny people work and have them lose their job because they bleached their hair. It is a perfectly normal thing to do and most women I know have had some kind of hair treatment.

In summary, I do believe that hair analysis has good advantages over urine in terms of the historical aspects, the collection, et cetera. I think it has good applications -- child protection, compliance, forensic cases. We think it should even be considered for workplace if we get some kind of consensus on these issues, the environmental issue, the racial issue and the effect of various shampoos and things on the concentrations of drugs in hair.

What I have liked that I note over the past couple of days is this idea of complementary testing. If someone has a high morphine positive urine, then let the laboratory recommend that you then could do a hair sample.

I think you have got something to go on with the urine. You can help a lot of MROs. From that standpoint, I think hair can give you a lot of information, but not on its own. It needs to be in conjunction with other types of approved testing so that we can make ourselves a safer workplace. Thanks.

DR. BAUMGARTNER: Hi, Christine. I think if NIDA organized a consensus conference amongst hair testers, we wouldn't have any problems at all with respect to the issues you have raised. First of all, some of the comments that I made to Dr. Kelly apply to you. I am glad to see that you are more effective in your washing. You are following the European consensus which is close to a standard we exceed.

Concerning the racial bias issue hypothesis, I have just opened a paper which I will forward to you that shows that the Joseph study, which is an in vitro study, has absolutely nothing to do with what goes on in the real world. It is not really a contamination study because they used powdered hair. In the Henderson and Harkin data, which is widely cited, there are three hair samples out of 25 that are non-Caucasian and which are very high. What is very definitive is that these three samples do not follow the pulse kinetics. This means that they are very highly contaminated samples from sweat. Since Henderson and Harkin washed only 10 times for less than a minute, they didn't clean out the semi-accessible domain. So, this is an ineffective wash procedure. As you know we use three times 30 minutes, phosphate washes. If we see there is more to be removed from the wash kinetics, we give it an extra two hours after that. Now, one thing that I would like to ask you, why is it that people religiously avoid citing the very definitive studies, the 1,200 human subjects that Mitchkowski published using our methods. There is no racial bias evident. So, this is a very important issue. If there is no racial bias, then you will not see a melanin spin-out which, in fact, Joseph looked for. If there is no bias, then it won't stick.

DR. MOORE: Are you saying that there is no bias at all?

DR. BAUMGARTNER: The studies have focused on cocaine, the Joseph study and the Henderson and Harkin study and we looked at cocaine.

DR. MOORE: If there is no bias, why do you bother spinning down the melanin.

DR. BAUMGARTNER: Yes, we spin it down because it can also interfere with the immunoassay because of a blank reaction.

DR. MOORE: So, you don't do it to get rid of the racial bias.

DR. BAUMGARTNER: We don't accept the racial bias. There is no evidence for it. As we said, you can't have it both ways. It either sticks to melanin or it doesn't stick to melanin. If it sticks to melanin, then you can have a clear color bias and we will spin it out and we would correct for it. You can't have it both ways.

DR. MOORE: With all due respect, I think you kind of changed your angle a little bit there. The reason that Joseph, I believe, did that work was to exactly show that what you were doing did not get rid of racial bias. I think that was the motivation for his work. Therefore, he thought that you were doing it to get rid of the racial bias. Otherwise, he wouldn't have bothered doing it.

DR. BAUMGARTNER: No, we usually spin out melanin because it interferes with radioimmunoassays with a blank reaction. If there were racial biases, then it would also get rid of that. But this is not an agreement because Mitchkowski's data, using our technology with 1,200 subjects, shows there is no bias in cocaine. As I said, a consensus meeting would take care of this problem. Unfortunately, NIDA has not followed the example that NIJ initiated years ago, so that we won't be talking like the tower of Babal.

The adulteration issue, the removal of substances through hair treatment depends, of course, as I said before, on the extraction method. If you don't focus on the inaccessible domain and if you only focus on the accessible and semi-accessible domain, yes, drug users could just go under the shower and wash their hair for two hours or three hours and all these samples would be negative.

DR. MOORE: Maybe not the shower, and if I was to have a hair test and I was a user, I would just go bleach my hair, dye it, perm it, do everything to it before I took the test. I don't think it would be positive.

DR. BAUMGARTNER: If you destroyed the inaccessible domain by bleaching, you may as well cutoff your hair. You would do this to your hair, and it would be gone. It does affect the inaccessible domain to some degree. But if it destroys it completely, you may as well shave it off. As I said, I think a consensus meeting would certainly take care of these problems. Until NIDA follows the suggestion that NIJ practiced many years ago, we wouldn't have the problem.

DR. CHILDS: I have a question, similar to what I asked Dr. Kelly. Could you give us some feel for the samples that actually screen positive using your technology, and then how many of those actually confirm positive. I know you are reporting at the limit of detection. So, you might have a little different proportion than those using cutoffs.

DR. MOORE: I think his lab does considerably more samples than my lab. We are not a big player. I think APL and Psychemedics are the biggest players in this. I knew you were going to ask me that because I heard you ask him. The biggest one that don't confirm is amphetamines. I think that is probably true of any type of urine program as well. Because we use FPIA, I think we are missing some cocaines, I think. That is why we are switching over to some other technologies. I couldn't give you percentages off hand.

DR. CHILDS: I have another question. I was recently scanning the internet for different types of adulterant products. That is a great fascination for me and other people who like to have access to those things. I saw a listing for a company that offered a product that could surely help you beat your drug test for hair testing. I was curious to know if you have had a chance to investigate or have heard any reports. Often our clients call and say, we found this product. Can you investigate. Have you had any experience with those kinds of products?

DR. MOORE: Was it Malibu treatment? Was that the one?

DR. CHILDS: I don't know the name of it. I am just kind of curious, do you know what they are.

DR. MOORE: No, I don't, but my guess is that it is some kind of bleach or dye or something like that.

DR. CHILDS: If I spent $39.95 and I sent that to you, would you do a little research on that?

DR. MOORE: I guess, in my spare time. It would be interesting to know.

DR. CHILDS: I don't know. Have the other laboratories who do hair testing had experience with that? I don't know if that is a fair question to ask the other two.

DR. JONES: There are a couple of products out there that are shampoos that are marketed specifically for that purpose that I am aware of also.

DR. BAUMGARTNER: We have looked at them, as I said just before. The basic effect is that they do a good job on the semi-accessible domain, which we clean out anyhow. As I said, some products influence the inaccessible domain, but in most cases 80 percent of the blood is in the inaccessible domain. If you destroy that, again to repeat myself, if you just touch your hair, it would just fall off. People shave it off, too.

MR. MC CANN: I appreciate your question, particularly in terms of the summary about the use of hair testing in the workplace. I think you are absolutely correct, that you have got certainly a lot more advantages in terms of the child protective environment than in the workplace, although we have a lot of parallels. One of the things that I am concerned about is the use of some of these -- with all due respect to Dr. Caplan, he calls them complementary methods. I am not quite there yet. I think alternative is probably the appropriate word. The danger seems to be the whittling down of the gold standard that has been created, the lack of certification, the lack of MROs. Basically what we are doing is whittling this thing down. Given the use of hair testing in the workplace -- now I realize that is not your specialty -- but as primarily a pre-employment testing methodology, then once we consider the issue of racial bias and adulterance, what we really have is a result that is not --

DR. MOORE: Valid?

MR. MC CANN: Yes, it is no good. The fact that we are denying people employment which is, in fact, probably either racially biased or beauty shop biased, you know, I have some problems with that. Particularly when you think that particularly in the urine testing environment we are seeing laboratories that are going to a 10-panel test, now I didn't get the opportunity to ask Dr. Kelly if the hair testing has expanded into the 10-panel test in the gaming industry.That would be another concern for me, particularly when there is no MRO involved. I can see a lot of this as emerging issues. If you could comment on that -- I don't know if you have tested for any other substances besides the five illegals or not.

DR. MOORE: I think Dr. Kelly could tell you if they do more than five. I don't know.

DR. KELLY: At the moment we are just confining our testing to those five. You know, there is interest in looking for other drugs, methadone for example, but at this moment we are just doing the NIDA 5.

DR. KIDWELL: I wanted to comment briefly on melanin. It depends on where the melanin comes from. If you use sepia melanin that comes from a cuttle fish, you will get different results than if you use synthetic melanin. You will find out that is different from hair melanin. And it depends on how the hair melanin is isolated. If you use the published digestion procedures, you will find that drugs do not bind to that melanin. It suggests that separating that melanin pump out will not eliminate racial bias. During the digestion process it appears that you change the melanin in such a way that the drugs no longer bind to it. If you wish to do studies with hair and hair melanin, I suggest you use a process that dissolves the hair using urea and dithiathriatol. In that case, that melanin will bind drugs substantially and it will vary depending on the race of the individual giving the hair sample.

DR. MOORE: As I said, we are kind of polarized on this issue.

DR. BAUMGARTNER: I guess it is a discussion between me and Dave. Dave, I thought NIDA may have given you my paper that I sent in, because I showed there that our method of isolating melanin leads to drug binding. As a matter of fact, this is a method that has been developed and sort of adapted from a lot of other melanin studies. There are different melanin. If you look at the paper -- I will be happy to give you my copy that I have here, and even I have overhead slides that I can show you that it binds extremely well.

When you combine carotene with melanin, then you will see that the drug prefers carotene. The reason is that it is in about 70 to 30-fold excess. Carotene is far more in excess than the melanin.

DR. MOORE: Did you just say it binds to carotene and not melanin. Is that what you said?

DR. BAUMGARTNER: The standard procedure in this field, which is an extensive investigative field in pharmacology, is to do competitive binding studies. The study of Joseph's was not a competitive binding study. That is the first thing. As I said, I will pass the paper out. If you have a competition of drug binding to carotene versus melanin, you will see that if you bind the drug with isolated melanin, you have substantial binding to melanin, because everything binds to everything. When the drug is competing in the binding to carotene versus melanin, it favors carotene overwhelmingly.

DR. AUTRY: We will take one more question and then we are going to break for lunch. Most of us were aware that there is an ongoing debate.

DR. KIDWELL: I was not trying to make it into a debate again. I think I will defer my comments until a little later this afternoon.

DR. AUTRY: We will have plenty of time for public comment this afternoon. I want to thank you, Christine. Take you off the hot seat now.

(Applause.)

DR. AUTRY: As I mentioned yesterday at the close of the day, we are going to change the agenda a little bit. All the time that was reserved for written questions will, in point of fact, be open for public comment.

I think you have done a remarkably good job of making additional information available to the board and I want to continue that process in as open a way as we can. We are going to break a little bit early for lunch for a couple of reasons. One, it is raining and for those of you who want to get out to eat, it will give you a little extra time to get over to Pentagon City and back. So, we are going to break until 1:15. Second, for those of you who ate here yesterday, you may want to really think about going out. We will meet at 1:15 and hopefully at that point in time Bruce Goldberger will be here. Oh, he is here. Good. We are delighted that you are here, having flown in presumably in the last hour. We will see you at 1:15. Thank you.

(Whereupon, the meeting was recessed, to reconvene at 1:15 p.m., that same day.)

AFTERNOON SESSION (1:22 p.m.)

DR. BUSH: Welcome to our afternoon session. I hope you all had a good lunch and are ready for our next speaker, which is Dr. Bruce Goldberger.

Dr. Goldberger is an assistant professor in the Department of Pathology, Immunology and Laboratory Medicine in the College of Medicine at the University of Florida in Gainesville. Previously, Dr. Goldberger was an assistant toxicologist and laboratory manager for the state of Maryland at the office of the Chief Medical Examiner and a toxicologist with the National Center for Forensic Science in Baltimore, Maryland. Dr. Goldberger received a bachelor of arts degree in zoology from Drew University in Madison, New Jersey, and a master of science and doctor of philosophy degrees in forensic toxicology from the School of Medicine at the University of Maryland in Baltimore. Dr. Goldberger is certified as a toxicological chemist by the National Registry in Clinical Chemistry and a forensic toxicology specialist by the American Board of Forensic Toxicology. At the University of Florida, Dr. Goldberger is the technical and administrative director of the forensic toxicology laboratory. The laboratory is a full service forensic toxicology lab that provides lab support for District Eight (Gainesville) and District Fourteen (Panama City), State of Florida Medical Examiner Offices. The approximate yearly caseload of the laboratory is about 600. Dr. Goldberger's laboratory also provides analytical services for several governmental, academic and private organizations. Dr. Goldberger has published numerous articles related to forensic toxicology and is co-editor of the Handbook of Workplace Drug Testing (AACC Press). Dr. Goldberger's studies in forensic toxicology have included the analysis of alcohol in breath, and the measurement of therapeutic and abused drugs in biological tissues, including alternative matrices such as hair and nails. Dr. Goldberger's most significant contribution to the field of forensic toxicology is the identification and measurement of heroin and its metabolites in hair and other human tissues. Dr. Goldberger.

Agenda Item: Implementation of a PT Program for Hair Testing in Florida.

DR. GOLDBERGER: Thanks, Donna. During the 1993 and 1994 Florida legislative specimens, bills were introduced to permit hair as a specimen for drug testing in drug free workplace programs in the state of Florida. These bills did not make it past their respective committees. During the 1995 session hair again was proposed as a specimen for drug testing. The bill was passed by the House of Representatives and the Senate, but was vetoed by Governor Childs because of the possibility of racial bias and external contamination.

In 1996, Florida's House of Representatives and Senate again considered the issue of hair testing. The House of Representatives and Senate bill contained very detailed language which specified procedures for specimen collection, decontamination, digestion, screening and confirmation. Ultimately the house bill was unanimously approved by the House of Representatives and the Senate. Governor Childs allowed the bill to become law without his signature. Although I have resided in the State of Florida for about three years now, I would like you to know that I have not been involved in any of the legislative processes regarding any of the drug free workplace programs in Florida. I was asked to testify last year at the hearings, but they were scheduled during the AFS meetings. So, I could not attend.

My presentation today is divided into three parts. First, I would like to describe the State of Florida drug free workplace act as it relates to testing of hair for drugs of abuse. Then I would like to describe the implementation of a proficiency testing program for hair analysis in the state of Florida. Finally, I will discuss problems we experienced during the implementation stages of the PT program.

The original bill was published in the laws of Florida as 96-289. It was recently integrated into section 112.0455 of the Florida statutes. To the best of my knowledge, this is the first attempt in the United States at legislating the accreditation processes of laboratories performing hair analysis for the workplace. A complete copy of the drug free workplace act can be found on the internet at the address shown on this slide. As everyone knows, I am internet fiend. I was actually able to find this readily on the internet. Just beware, if you go to download it, it is quite lengthy, probably 40 or 50 pages long. Incidentally, I was recently talking with Dr. Caplan from Maryland, who I am sure everyone knows. He indicated that the state of Maryland recently adopted similar legislation regarding hair analysis. The major differences between Florida and Maryland is that the meat or detail of the rules regarding hair analysis in Florida were placed into the law, and in Maryland they are actually in the rules which are administered by the Department of Health. When they are administered by the Department of Health, they are more easily adapted and modified according to what we know about the pitfalls and so forth.

In Florida, we can't change the law. It is firm as it is now. To the best of my knowledge, I don't think there are any proposed changes to them in the upcoming session. The state of Florida drug free workplace act consists of statutes which were approved by the state legislature and the governor of the State of Florida, and rules prepared by the State of Florida Agency for Health Care Administration. These rules are modeled after the HHS rules, as I would presume most states have done accordingly. The rules that are prepared and modeled by the State of Florida Agency for Health Care Administration describe criteria for specimen collection, chain of custody, protocol, initial and confirmation testing, reporting of results, the requirement for an MRO, quality control and quality assurance procedures and statistical reports.

According to the Florida drug free workplace act, they define a specimen as a tissue, hair or product of the human body capable of revealing the presence of drugs or the metabolites. As you notice, hair is specifically stated in this statute. Under the drug free workplace act, an employer is required to conduct the following types of drug tests to comply with these drug free workplace programs that we have in Florida. They can be pre-employment testing, random testing, reasonable suspicion, routine fitness for duty, post-accident and rehab. You might want to know that in all situations, hair is considered an acceptable specimen by the State of Florida drug free workplace act. I think there is some discussion about that possibly in the future.

This slide lists the cutoff concentrations for initial testing of amphetamines, cannabinoids, cocaine, opiates and PCP in hair according to the State of Florida drug free workplace act. These concentrations should be familiar to all of you. These are the numbers that were put forth by Psychemedics Corporation and used by Psychemedics.

This slide lists the cutoff concentrations for confirmatory testing of amphetamines, cannabinoids, cocaine, opiates and PCP in hair, according to the State of Florida drug free workplace act. Confirmation of amphetamines must include the amphetamine and methamphetamine. Confirmation of cannabinoids must include the acid metabolite of THC. Confirmation for cocaine must include testing for benzoylecgonine and cocaethylene. Finally, confirmation of opiates must include testing for 6-acetyl morphine and codeine. These concentrations should also look familiar to you, as they were the numbers proposed for use by Psychemedics Corporation.

According to the drug free workplace act in Florida, all hair specimens undergoing confirmatory testing must be decontaminated using a wash procedure that has been published in the peer review literature. In addition, an initial 15-minute organic solvent wash followed by multiple 30-minute aqueous washes must be used. A minimum of three of these aqueous washes must be used in this procedure. According to the drug free workplace act, after decontamination, drug in the hair must be released either by digestion using chemical or enzymatic means or by multiple solvent extractions. In addition, all confirmation methods must eliminate the melanin fraction of the hair before analysis.

Finally, according to the drug free workplace act, if a non-digestion method is used, the laboratory must present published data from the peer reviewed literature from a large population study which indicates that the method of extraction does not possess a statistically significant hair color bias.

The wording for my prior slides is directly from the statute.

Let's talk about licensure just for a moment. To obtain licensure in the State of Florida under this act in hair analysis, satisfactory performance is required in three testing events. For hair analysis, each testing event will consist of five samples which includes a combination of drug positive and drug negative hair samples. These are blinded samples, the content is blinded. For renewal of licensure, satisfactory performance is required in three testing events per year.

Now more about the State of Florida program that I have been asked to set up. About six months ago, Mr. Roger Vance of the State of Florida Agency for Health Care Administration contacted me regarding the proficiency testing requirement of the new state statute. He had indicated that he was unable to find a reliable source of appropriate proficiency testing samples. He had contacted Pascal Kintz with the Society of Hair Testing, contacted CAP, ASCLAD and a number of other organizations. None were willing to provide the number and variety of samples that were required by this act. After some discussion and negotiation, it was decided that my laboratory at the University of Florida College of Medicine would provide these samples.

I am not sure this was a good thing or a bad thing, but you will find out in a minute when I am finished. It is a small program with a limited number of labs. So, before I even start to describe the details of the program, you need to understand that. This is totally something that has not really been done before. So, it is our best food forward under the conditions that we have, and they are quite stressful conditions. Mr. Vance also asked if I would consult with him with regard to proper administration of the program and evaluation of the proficiency test data. He is quite proficient with evaluating clinical chemistry and urine drug testing results. When it comes to hair analysis, it is totally out of his realm of understanding. So, I will assist him in the administration of the program at no charge, actually, just help him out as another state agency.

Drug free hair was obtained from laboratory personnel, including myself, with a negative drug use history. All of us have abstained from prescription medications and over the counter medications. These are the same hair samples that we used in our studies that we have published results from already. To provide a variety of different types of hair -- permed hair, colored hair, my gray hair, Caucasian, black hair and so forth -- we have obtained hair from a number of individuals from our laboratory. All of them, as I said, have negative drug use history and can be trusted. All drug free hair were shown to be drug free by GC/MS analysis. Authentic drug positive hair samples were obtained from known drug users from Pascal Kintz, our colleague in France.

The majority of these drug positive samples, to the best of my understanding, were cadaver hairs that Pascal routinely has the ability to collect at the medical examiner's office in Strasbourg. If authentic drug positive hairs were not available, for example with PCP, hair was fortified with drug at realistic drug concentrations. Fortunately, we were able to locate a large number of authentic hairs. So, the fortification process was restricted to only a small number of samples. All proficiency samples will be subjected to validation studies before use in the proficiency testing programs.

All samples were pulverized by Pascal in France using his ball mill. I have a small device in my lab, but it can't handle the gram quantities of hair that are necessary for this program. After the samples were pulverized, they were subjected to solid phase extraction and analysis by gas chromatography mass spectrometry in both Pascal Kintz' lab and my lab with one exception. I don't have the ability to perform THC acid analysis in hair at this moment. So, therefore, except for the acid metabolite of THC, the target concentration of all analites will be based upon analyses performed in two separate laboratories. So far, with the majority of the testing complete, we found that the quantitative results from both labs -- Pascal's lab and my lab -- are well within the 20 percent agreement.

As previously stated, the drug analytes specified in the statutes include those commonly tested by a majority of the hair analysis laboratories in the United States and Europe. The drug class and the specific analites within the drug class are listed here. For amphetamines we test for amphetamine and methamphetamine. For cannabinoids, we have THC acid. For cocaine it involves cocaine, cocaethylene and benzoylecgonine analysis. For opiates it is 6-acetyl morphine, morphine and codeine. For PCP, we look for PCP. Satisfactory performance in the PT program is achieved by maintaining an overall test event score equivalent to passing proficiency scores for other drug testing matrices such as urine. Presently, in the State of Florida, a passing grade is 80 percent. Failure to participate in a proficiency testing event results in a score of zero percent for that specific testing event.

Also, failure to achieve satisfactory performance in two consecutive testing events or two out of three consecutive events is deemed unsuccessful performance. These are the same rules that are used for urine drug testing in the State of Florida. For those who live in Florida -- I guess there is no one here except Bob White -- we are very familiar with these and live with these.

This is where it gets tricky. How are we going to evaluate the laboratory PT data. What I am going to state today is only preliminary and can be changed based upon a laboratory's performance. We expect to send out the initial set of proficiency samples within a month or so. That initial set may be a trial set. We can verify that what Pascal and I have done with these samples is appropriate, by evaluating a laboratory's data. Then we will go forward with three real cycles of samples. The evaluation of the hair proficiency test data initially will be similar to the urine proficiency test data currently underway in Florida. However, as I stated, these rules may and I would expect will be modified, once we get our results. All results will be subjected to qualitative and quantitative evaluation. With a limited number of labs expected to participate -- definitely you can count them on one hand -- the qualitative evaluation will be considered most important.

In the case of a true positive and a true negative test result, obviously no action will be required. In the case of a false negative test result, remediation will occur. In the case of a false positive test result, remediation will occur and loss of laboratory licensure is possible.

Now it gets even trickier, I suppose. The quantitative evaluation criteria is similar to criteria established for the evaluation of quantitative urine data, for good or bad at this moment. Presently it is based upon a comparison of the laboratory data with the peer group mean plus or minus 20 percent or 2SD, whichever is greater. In the case of the hair analysis program, the peer group mean will be calculated based upon the validation data and the participant lab data. During the implementation of the State of Florida proficiency testing program we experienced a large number of programs. We started back in May and it is now September. It has taken quite a bit of time to get this going and work out all the problems. These include but are not limited to: the lack of authentic hair samples. Fortunately, Pascal Kintz was gracious enough to assist with providing the authentic hair samples and laboratory data on those samples.

We were unable to find authentic samples for known PCP users. We had to prepare fortified hair samples. Also, in order to maintain sample homogeneity among the reference labs and the participating labs, it was necessary to pulverize these samples. I realize that a number of labs prefer not to pulverize, but this was our best approach.

As a result, the PT samples cannot be decontaminated according to the procedures that are found in the drug free workplace act. We understand that and the labs will not be penalized for that reason. Further, the number of labs available to serve as reference laboratories limited, as is the number of labs that we expect to participate in this program. Also, the lack of standardization of testing methods among laboratories will certainly impact the test data in a negative manner. As I stated before, I expect the first set of proficiency samples to be sent to the laboratories within a month or so. We hope to have the initial cycle completed by December 1997. These labs, I understand, will also undergo on-site inspection to verify compliance with the drug free workplace act.

I would like to thank the Drug Testing Advisory Board for the opportunity to speak today. We will have more information at the SOF meeting, the methodology used. So, come visit our poster. I think it is scheduled on Wednesday morning at 8:00 o'clock. I also need to acknowledge that the program is supported in part by the State of Florida Agency for Health Care Administration. That completes my talk. I think there is probably some time to entertain some questions.

(Applause.)

MR. VELASCO: The question I have is, in the statement that you made that you are going to eliminate the melanin factor, how are you going to maintain that in laboratories and make sure they are doing that?

DR. GOLDBERGER: I think some people understand my role as a provider. The State of Florida does have an inspector who will go to the laboratories that will participate. It will be their responsibility to verify that the melanin fraction is removed. Otherwise, I would prefer not to comment. My role here, as I stated, is a provider of the samples and also providing some assistance to the agency where they need it.

DR. SELAVKA: Thanks for the presentation, Bruce. It is nice hearing about the development of the program. I guess you are familiar with the NIST's experience with round robin neutral laboratory comparisons and the size of the SDs that came from even the 12 to 15 to even 23 laboratories, depending on the studies. Having a whichever is greater, 20 percent or plus or minus 2 SDs, if you have a respondent cohort of four, it may make the use of that criterion for quantitative assessment worthless.

DR. GOLDBERGER: I agree.

DR. SELAVKA: I wonder what strategy you might employ statistically speaking to change that, if you are allowed to change it, through rule and regulation promulgation.

DR. GOLDBERGER: Roger Vance at the agency, as most bureaucrats are, he wants to apply a statistical means of evaluation to these data. I have told him several times that this may be impossible, knowing that we have maybe four or five participants and reference labs total. I have emphasized to him that we need to look at the qualitative data initially. If, over the course of three or four or five years experience with this, and we have more labs participating and more data, then I think we can apply the quantitative approach.

I put the quantitative approach on the slide because Roger told me to. He is a bureaucrat, but he is a scientist, too. He is a medical technologist by training.

The quantitative approach is in the agency's rules, so they can be changed and there is flexibility. It is not in our law. So, we are very willing to work with the laboratories. I would expect that the initial participant labs will do very well on these rounds of the proficiency test.

DR. SELAVKA: If I can ask you another question, I am presuming -- although you didn't say it and I hope you will tell me it is true -- laboratories which are not compliant with the Florida drug free workplace act may participate by purchasing these proficiency test materials, and perhaps that an invitation could be extended to laboratories that participated in the NIST and Society of Hair Testing round robin for laboratory comparisons?

DR. GOLDBERGER: That is a really great question and I am glad you brought that up. It was a few months ago I got a call from the State of Oklahoma, who also adopted similar legislation for hair analysis. I don't know too much about their rules, but they do have a proficiency testing requirements.So, the fellow from Oklahoma said, well, can you just sell your samples like CAP sells their samples. I said I would consider that. I worked it out on paper and with a calculator, and the cost is just prohibitive. I can be fair here. This program is not cheap. It costs lots of money. For the relatively small number of labs participating, the cost is just outrageous for the lab -- they are paying for licensure, I believe it costs $8,000 for licensure, and also the costs for the state to pay my lab for making these samples. In the process of making these samples, we have about one gram quantity of every sample. So, I believe we are going to send between 50 and 100 milligrams of pulverized hair to each lab, which leaves us with some in storage. I am willing to share that with other laboratories, but we will have to charge for that because it has been an expensive endeavor to prepare these. I don't think we will ever see CAP providing hair analysis samples unless the field of hair analysis really grows dramatically in two years. It is very difficult, but it was a great question.

DR. JONES: One point of clarification. I didn't have your overheads with me. Did I notice you are only using the carboxy metabolite of THC and no THC itself was in your hair samples or spiked samples?

DR. GOLDBERGER: Our hair samples for THC are authentic. They contain THC, THC acid and CDN that Pascal quantitates. I have his numbers and he is providing numbers for all three compounds. These are authentic samples so we would expect to find those other analites.

DR. AUTRY: Other questions and comments? Thank you very much, Dr. Goldberger.

(Applause.)

Thank you, Bruce, we really appreciate that. I must say, I do not envy you your task. Consistent with what we have said yesterday and today, the floor is now open for public comment. We will keep the floor open as long as there are people who want to say things, subject to stopping on time.

I would anticipate that we will probably end up ending a little bit earlier. I would request anybody who does want to comment, to limit their comments to a maximum of five minutes.

I will exercise the chair prerogative that if we seem to be getting into a debate, I will propose that we not get into a debate. Then I have a few closing comments and requests from everybody here.

I do want to thank all of you, even before we go into the public comment period, for your participation. I think you have been great contributors to this process. I also want Donna, next time, to work up bios on the board members, so you can be equally impressed by them, as you have been by the people she has introduced here. I think this is a testament to not only my colleagues expertise, but also a testament to Donna's skill at writing. The floor is open.

Agenda Item: Public Comments

DR. WILLETTE: Somebody asked me, after my presentation, about how do you deal with on-site testing with dilute specimens, let alone adulterated specimens. I want to point out that some of the federal probation offices actually supply their testers with refractometers. If they collected a specimen that is virtually colorless or has low color, they will actually do a specific gravity of them right on the spot before it is tested, and then retain the person until they get a high enough specific gravity. Also, I know of at least two, maybe three companies that are in the process of getting ready to market dipstick tests for creatinine levels. I think this is, in part, being prompted by the need to check for dilution with on-site testing. So, there will be some mechanism available. I haven't seen any data yet, or very limited data, on the effective adulterants in some of these on-site tests. I think that is some critical information that is really needed, to see how well they resist adulteration.

DR. AUTRY: Thank you, Bob. I would also just add a footnote to that, and that is that adulterants is a problem for all of us at this point in time, just to emphasize some of what we heard from the other presenters.

DR. HILDERBRAND: I would just like to make a comment that many years ago we had no authentic THC carboxylic acid standard. We had no deuterated internal standards. We were using glass packed columns and things like that. Out of this came the national laboratory certification program. The real intent of this whole program is to protect against false positives and to guard, I would say, against false accusations. That can even include presumptuous statements against a person. I think that what -- well, in the future as we go through all these things, we need to really address the continuing responsibility to maintain quality in all the test procedures that we look at. It is not as simple an issue as just recommending technology or looking at scientific evidence. I think we heard yesterday that there is much more to it. I would really ask the board to consider quality in any recommendation that is made. Thank you.

DR. AUTRY: Thank you. Other comments, questions, concerns?

MR. DONAHUE: I am a technical director. Our laboratory performs hair analysis for workplace testing. I guess you can call us the third commercial lab that does the work. I believe a board member asked about stripping the agents or adulterants that are on the market. I don't have the data here but we have done some extensive testing. In fact, we find that the stripping agent provides no effect on the results. If someone wants to give me an address or a phone number to call to send some data to, I will be glad to share my results with them. There are about four agents that are on the market that we have looked at.

DR. KIDWELL: We have demonstrated that cocaine, placed on the skin of a drug free individual can be detected for up to three days after the exposure. I am suggesting that some of the better results obtained by sweat patches presented yesterday may be due to prior use or contemporary contamination. The surface contamination may occur from being in an environment where cocaine was used or wearing clothing from previous cocaine use. Now, in sweat testing alcohol is used to clean the skin prior to the patch application. I am suggesting or strongly urging that the alcohol that has been used to clean the skin be tested for drugs to assure that that patch is not measuring prior use rather than current use.

DR. AUTRY: Thank you.

DR. JONES: One comment I would like to make from the board's perspective, as I said yesterday, you have a grid that is a work in progress. We anticipate contacting some of the industry representatives for some additional information that might help us in removing some of the Is or some of the Ps that are in some of the various grid locations.

We will be further refining that table over the coming few weeks and coming back with some additional comments and recommendations to Dr. Autry's office. So, the industrial coordinators, or area coordinators who are present, expect a call. I am putting you on notice for that. On behalf of the board -- I guess I am not speaking for them, they can speak for themselves -- but we want to thank all of you for your participation, your staying with us, your supplying information. We are going to try to take all this information in and go forward from here. Do any other board members have any comments?

DR. AUTRY: Again, let me just add my personal thanks to all of you. As soon as we have the revised grid and explanatory notes available, we will make that available to all the participants.

Skip has promised that he will write them on the way home on the plane, by the way. However, I have talked to his wife since then and I assure you they will not be completed by this evening. We will make that available to you in hard copy. We will also put it up on our web site. The same thing as we did last time, we will have the transcript of this meeting up on the web site.

DR. BUSH: The address is already in the blue folder of the web site. Three weeks from now we hope to have it on the web.

DR. AUTRY: If you have any trouble down loading it or anything like that, give us a buzz and we will see if our technical experts who put it up there can figure out how to get it down.

Again, any last comments?

If not, then this meeting stands adjourned. I thank you. This is an ongoing process, as you have heard. We most likely will not be taking this to December meeting since there is other pressing business that the board has to get to. This will be part of the ongoing dialogue here. Let me see if we have a housekeeping chore coming up.

DR. BUSH: Just to let you know, there are some extra packets out there. For those of you who were asking for them, please feel free to take them. They are out on the table there.

DR. AUTRY: Thank you. The meeting stands adjourned.

(Whereupon, at 2:08 p.m., the meeting was adjourned.)