Hair analysis for the detection of drugs of abuse for workplace drug-testing programs has been in existence for nearly 20 years. Much of the literature on hair analysis from infancy to current practices addressed the complicated interpretive issues surrounding hair analysis for drugs of abuse. Hair analysis has provided the fields of forensic and clinical toxicology with an acceptable alternative to analysis of blood and/or urine when its limitations are considered for interpretive purposes.
Although the analytical technologies of hair analysis have improved considerably over time, many challenges for interpretation have yet to be fully investigated and resolved. As interpretive and analytical considerations are not the primary focus of this chapter, a short discussion of each will be presented below.
Potential for Environmental Contamination Influencing Tests Results
Drug analytes may be found on and in hair as a result of environmental contamination. A drug may gain entry to an individual’s body through association with a drug-contaminated environment by breathing dust or vapors, swallowing saliva contaminated with dust or vapors, and eating food or drinking liquids that have been in contact with the drug. If an unsuspecting individual has been contaminated through the environment, and the laboratory procedures cannot distinguish this contamination from intentional drug use (i.e., the purpose of drug testing), the drug test must be considered susceptible to “false positive’ results.. All matrices used in drug testing are susceptible to contamination through passive ingestion.
The concern in hair testing is not so much passive ingestion of drug, but contamination of the exterior surface of the hair by drug in the environment adsorption) and subsequent absorption of the drug into the hair matrix. Forensic laboratories implemented practices intended to eliminate the potential for false-positive results emanating from environmental contamination. First, most forensic laboratories include a decontamination step in their analytical procedures for hair analysis designed to remove or minimize the contribution of environmental contamination to the drug concentrations. In addition, some laboratories analyze the wash fractions obtained during their decontamination procedure for the drug analyte. This information has been used to assist with interpretation. One approach has been to develop a mathematical model incorporating the drug analyte concentration found in the wash fractions to determine when the wash has removed the majority of the contaminating drug and deduct the amount of drug found in the last wash from the amount of drug found in the hair. The resulting value is then compared to a cutoff concentration that is used to establish whether the specimen is positive or negative for a given analyte. Tsanaclis and Wicks recently reported on the results from the analysis of 216 hair samples and the corresponding methanolic wash fractions. The samples were obtained from a population where external contamination could be expected (drug-related police investigation cases) and used the analyses to determine a wash-to-hair (W-H) ratio. The WISH ratios were determined for 13 analytes (acetylmorphine, acetylcodeine, codeine, heroin, morphine, dihydrocodeine, cocaine, benzoylecgonine, amphetamine, MICA, MDMA, THC, THCGOOH) found in these hair samples. The W/H ratios were categorized into 4 groups: (1) Group null – in which no drug levels were detected in the wash residue (drug use likely and no environmental contamination); (2) Group <0.1 – for W/H <0.1, implying the presence of the analyte in the wash residue less than 10% Possible drug use; environmental contamination cannot be excluded, (3) Group <0.5 – for W/H <0.5, implying the presence of the analyte in the wash residue <50% (possible drug use; but environmental contamination cannot be excluded; and (4) Group ≥0.5 – implies much of the drug is from external contamination. Results for the last group were not reported directly but could be calculated from the reported data. From the data, it would appear that the frequency at which definite external contamination existed ranged from 0% for MDA to 25% for heroin. At the other extreme, samples with no drug in the wash, Group null, implying drug use and no contamination, ranged from 24% of the samples containing 6-acetylmorphine to 92% of the samples containing MDA. For drugs that are smoked and more likely to have environmental contamination, evaluation of the parent metabolite indicated that drug use was probable with the following frequencies: 40% cannabis (n = 61), 29% amphetamine (n = 52), 34% cocaine (n = 95).
The authors opined that “the likelihood of external contamination confounding the interpretation of hair tests can be reduced to manageable proportions” since only 216 of 30,000 (c3$^o~ actually tested positive for the investigated analytes. Closer scrutiny of the data shows that of the 891 positive test results, 97 had a W/H >0.5, demonstrating that approximately 11% of all positive tests were consistent with a high probability of contamination and were inconclusive concerning drug use.
Several additional conclusions may be drawn from this study. Given the high frequency of undetermined drug use for this highly probable drug use population (e.g., drug use could not be confirmed with proposed W/H ratio 2% to 24% of the time), wash evaluations for interpretation need to be investigated and the W/EI criteria and washing techniques further refined. Secondly, to provide evidence of drug use, a laboratory’s analytical methodologies should identify compounds that can only be present as a product of metabolism. Finally, contamination and passive exposure studies investigating the potential for environmental contamination should be used to provide data that can be used to establish conservative cutoff concentrations that will determine whether an individual is positive or negative for a drug.
Variability of Results
The impact from hair analysis continues to be weakened by the variability of the results found within individual laboratories and between groups of hair-testing laboratories. This problem has been observed in clinical studies where dose-response relationships could not be demonstrated because of intra-individual variability. Similarly, investigators reporting on proficiency testing results for most drug classes have demonstrated up to an average 50% variability rate among laboratories testing the same hair samples. In addition to the variability due to analytical procedures, it appears that variability can be the result of the sample collection process, the coloration of the hair, hair growth patterns, and other variables that affect drug incorporation into hair. To date, the attempts to control all of these sources of variability have achieved only partial success.
A major contributor to the variability is the nature of the hair matrix. It is a non-homogeneous solid matrix with variations in water content based on the amount of moisture in the air. Even under the best of conditions, drug analytes are not expected to be evenly distributed throughout each hair. An even distribution might be expected if the donor drug concentrations in the blood were at steady state for the entire period represented by the hair strands and if all of the hair strands were growing at the same rate. However, as previously discussed, we know that different populations of hair exist based on the growth patterns.
In addition to the contribution of the physical characteristics of hair to variability, there are variables induced by the conditions place on the analytical methods by the hair matrix. Because hair grows slowly laboratories are required to use small amounts of hair and have sensitive methods to detect the drugs within the matrix. Since hair is a solid matrix, the methods must be designed to extract the drug from the hair and the notorious difficulty in obtaining consistent recovery from a solid matrix obviously accounts for some of the variability. Also, variability is induced by the washing procedures conducted by the laboratories to remove possible drug contamination, which contributes to the differences within and between laboratories can viability be reduced. One possible step that might reduce variability would be standardization of analytical techniques. Currently, each lab has its unique procedures for sampling the hair, preparing the hair, decontaminating the hair, isolating the drug from the hair matrix, and analyzing the hair. For urine testing, although laboratories are allowed to develop their own procedures as they do in hair testing; overall the procedures are much more standardized with established methods. Given the analytical issues stemming from the complexity of the hair matrix standardization of analytical techniques is one of the few options available for improving the variability of hair testing results.
Non-standardized Techniques between Laboratories
There are a myriad of factors that make analysis for drugs of abuse in hair a most difficult analytical challenge. Laboratories must develop methodologies that efficiently free the drug from its solid matrix. The lack of clear evidence indicating which of the many processes currently in use best standardize and prevent the variability among laboratories remains a challenge. The sensitivity required detecting drugs of abuse in hair and the cost of equipment that meets the sensitivity requirements has hindered industry transition to the appropriate instrumentation. A review of the literature, especially reports in which multiple laboratories analyze the same sample, demonstrates that non-standardized analytical processes promulgate high variability.
Forensic laboratories have minimal agreement on each step of the hair sample preparation, including weighing and physical processing decontamination, digestion extraction, and derivatization procedures. Some laboratories weigh their aliquots to the nearest milligram, whereas others weigh to the nearest fraction of a milligram. The efficiency of removal of drug analytes from pulverized or powdered hair is different than its removal from small or large hair strand segments. Moreover, the extent of decontamination obtained through a brief 5- to 10-min methanolic or organic solvent wash would not be expected to equal that of a multiple-step/ multiple-hour wash procedure. Several authors have demonstrated that compounds that are stable in harsh conditions can withstand alkaline digestion procedures and yield better drug recoveries in comparison to enzymatic or acid digestions. Likewise, analytical technologies vary greatly among laboratories. Although there are some screening immunoassays that have been developed for use with hair some laboratories use immunoassays that were manufactured for processing blood and urine and have modified these immunoassays to work with hair. These modifications may well lead to variation in sensitivity and specificity. Different methods for sample introduction ionization techniques and mass spectral technologies that are not equivalent have further contributed to variability of hair testing results.
Over the years of urine testing for workplace drug testing, procedures have become more standardized and refined by forensic laboratories through mandated requirements and/or the need to maintain competitiveness. For the same reasons, hair-testing laboratories will most likely experience some standardization before hair testing in workplace drug-testing programs progresses.
Need for Consistent Proficiency Testing Material and Laboratory Certification
Proficiency testing material and laboratory certification programs help to regulate and maintain the status and accountability of any drug testing program. Both allow individual and systematic assessment of a laboratory’s ability to perform testing and to ensure forensic defensibility. Laboratory guidelines from professional organizations and federal programs support implementation and maintenance of these quality assurance measures when they are available. Several organizations have continued efforts to evaluate the proficiency of hair-testing laboratories, although participation is voluntary and not all laboratories choose to take part. Currently, there are no established laboratory certification programs for hair drug testing.
Proficiency testing (PT) programs use both authentic drug-user hair and manufactured proficiency samples that are sent to laboratories to analyze and report results. Results axe then compared to reference values or mean values of the participants. The results of PT programs are not routinely reported outside of the program.
The SoHT has conducted an international Pl program for several years. The program has not adhered to a consistent schedule for sending samples to the participating laboratories. In recent years, SoHT has generally sent 5 authentic drug user samples once a year. Between 15 and 20 laboratories participate worldwide, with most of the laboratories located in the European Union. Qualitative and quantitative results from the past 5 cycles (2001-2006) are available at the SoHT website. The Istitut Superiore di Sanita (Rome, Italy) in cooperation with Institut Municipal d’Investigacion Medica (Barcelona, Spain also established an external proficiency programs HAIRVEQ, in 2002.
Since 2000, the National Laboratory Certification Program NLCP, under the direction of SAMHSA of HHS, has conducted a pilot PI program for US laboratories. Five to 13 laboratories have participated in this pilot PT program during its existence. The NLCP program in the United States uses both authentic drug-user hair and manufactured hair PT materials, but the majority of the samples are manufactured by spiking hair strands with drugs of abuse. Figures 6-3 and 64 present the correlation with expected concentrations over a year period at the beginning of the program. For all drugs, laboratories performed well qualitatively, but quantitatively the percentage coefficients of variation (%CV) were >20% and accuracy was low as well. Figures 6-3 and 6-4 compare the average CV for hair and urine analysis, both as pilot programs and urine as a current maintenance cycle. Similar to urine as a pilot program in 1987, hair demonstrated a very large variation as measured by %CV.
Since this time, changes in the focus of the NLCP pilot PT program for hair to improve repeatability within and between laboratories has led to better performance in quantitative analysis of participating laboratories, but continued efforts are still warranted to achieve performance such as that maintained by the existing urine NLGP program. Figure 6-5 depicts %CV data for participating laboratories for amphetamine and THCA data during multiple pilot PT cycles during a 1-year period (2006 – 2007).
Kintz and Cirimele, on behalf of the Society of Hair Testing, first presented an inter-laboratory comparison of quantitative determination of amphetamines in hair samples Four different preparation procedures were used to test amphetamine, MDA, and MDMA: a direct methanol extraction, an acidic digestion (HCl 0.1 N), an alkaline digestion (NaOH 1 N), and an enzymatic (β-glucuronidase/arylsulfatase) hydrolysis. Of these preparations, best recoveries were observed after alkaline hydrolysis. In the study, the same hair sample was powdered and sent to 16 laboratories in the United States (4), Germany (6), France (3), Spain (1), Japan (1), and Korea (1). Amphetamine tested positive 13 trifles with concentrations raining from 3,300pg/mg to 17,500 g/mg. Only 2 laboratories identified and incorrectly reported methamphetamine, using GC/MS, at low concentra1ions (800 and 1,800 pg/mg). MDA and MDMA both tested positive in 14 cases, with concentrations ranging from 1,800 to l9500 and 8,900 to 100,000pg/mg for MDA and MDMA, respectively.
These scattered results clearly indicated that new exercises were needed to guarantee quality in hair testing, a major aim of the SoHT. In 2003, Jurado and Sachs indicated that qualitative results for the SoHT program were good, but continued to demonstrate false-positive and -negative results. They continued to scrutinize high variation in the quantitative results and noted that laboratories using an extraction procedure of >12 h were able to produce results with <20% CV.
Original reports of the HAIRVEQ program indicated that 82% of the participating European laboratories reported incorrect qualitative results, which the authors attributed to use of inappropriate methods for hair testing. More lately, Ventura et al. reported the performance of 34 laboratories participating in the analysis of 12 PT specimens over a 2-year period. All samples were authentic drug-user hair samples, some were pulverized, and some were submitted to the laboratories for replicate analysis. A portion of the hair samples was also submitted to laboratories participating in the SoHT program. Errors in qualitative results approached 50% for both 2004 and 2005. Quantitative results were comparable for both the HAIRVEQ and SoHT laboratories, with superior viability demonstrated by the former. The discrepancy of the results remained high and comparable to previous years, leading the authors to conclude that guidelines to focus laboratory method justification and data evaluation were needed to perk up performance.
Information obtained through PT programs is fundamental to the development of forensic laboratories. Once the industry, governments, and the public are made aware of laboratory performance, there will certainly improve methods and techniques. Without improvement the utility of hair testing in the United States and possibly throughout the world will be extremely limited and its relevance will be continually questioned.
Potential Bias of Color or Ethnic Differences in Drug Incorporation into Hair
Melanin content in hair has been and continues to be an unresolved interpretive issue. Some researchers believe drugs vary in their affinity for different types of melanin, while others believe if these contributions exist they are minimal and insignificant. In vitro studies by Joseph et al. He established that more drug content was detected in black or dark brown hair when compared to blond hair. These investigators projected that melanin was the primary binding site for the drug. Similarly animal studies with coexisting dark and light hair on individual rats showed that the drug content was highest in black hair, followed by brown then white. Likewise, Reid et al. established that pigmented hair from graying individuals had 1.3 to 6.0 times higher concentrations of cocaine analytes than non-pigmented hair from the same individual. In contrast, Kelly et al. showed in a statistical study an unimportant hair color bias for amphetamines and THCA in authentic human hair samples. In this study, although there was a higher incidence of cocaine positives in dark-haired individuals, the same pattern was seen in complementary urine specimens. Nonetheless, the extent to which drug incorporation into hair is influenced by melanin content or the mechanism by which the drug-melanin relationship exists/remains uncertain.