Introduction
It is generally accepted that chemical testing of biological fluids is the most objective means of diagnosis of drug use. The presence of a drug analyte in a biological specimen can be used to document exposure. The standard in drug testing is the immunoassay screen, followed by gas chromatographic-mass spec-trometric confirmation conducted on a urine sample. In recent years, remarkable advances in sensitive analytical techniques have enabled the analysis of drugs in unconventional biological specimens, such as saliva, sweat, meconium and hair. The advantages of these samples over traditional media, like urine and blood, are obvious: collection is almost noninvasive, is relatively easy to perform, and in forensic situations may be achieved under close supervision of law enforcement officers to prevent adulteration or substitution. Moreover, the window of drug detection is dramatically extended to weeks, months or even years. The aim of this article is to document the usefulness of these alternatives in forensic situations.
Hair testing for drugs of abuse in humans was first demonstrated in 1979 in the United States, and was rapidly followed by some German results. Since then, more then 300 papers have been published, most of them being devoted to analytical procedures. After an initial period during which drugs were analyzed using immunological tests, the standard is now to use gas chromatography coupled with mass spectrometry. During the early stages of hair testing, opiates and cocaine were the predominant analytes, followed by amphetamine derivatives. Only recently have can-nabis, benzodiazepines and, very recently, doping agents been evaluated.
Hair analysis was initially developed to document forensic cases; however, today, numerous applications have been described in clinical, occupational and sporting situations.
Incorporation of Drugs in Hair
This is one of the major points of disagreement among scientists actively involved with hair analysis. The time at which a drug appears in hair after administration is highly variable. According to several authors, this delay can be some hours or even days. Based on these differences, a complex model has been proposed to account for drug incorporation. Both sweat and sebum have been suggested to complement blood in this process but the exact mechanism is still under discussion. It has been demonstrated that, for the same administered dose, black hair incorporates more drug than blond hair, clearly indicating the influence of melanin. For some authors, these findings suggest a racial element in hair composition. Cosmetic treatments, like bleaching or waving, affect the drug content, producing a 50-80% reduction in the original concentration.
In almost all cases, the major compound detected in hair is the parent drug, much more so than its metabolites. For example, cocaine is present at concentrations 5-10 times greater than benzoylecgonine, and 10-30 times greater than ecgonine methylester, although, in blood, both metabolites are found at higher concentrations than cocaine. The very short half-life of heroin and 6-acetylmorphine makes their detection quite impossible in blood, but these two compounds are found in larger amounts than morphine in hair. This is also the case for sweat, and thus confirms the implication of sweat in the incorporation of drug in hair.
Environmental contamination has also been proposed as a potential risk of incorporation, leading to false positives. Drugs that are smoked, like cannabis,crack or heroin, are of concern, and it is therefore necessary to include a decontamination step to eliminate false-positive findings. Various procedures have been described in the literature: these involve the use of organic solvents, aqueous solutions or a sequence of solvent and buffer. To minimize the influence of external contamination, several authors have proposed various appraches including an analysis of the wash solution, or to carry a kinetic of the wash ratios. Others have proposed the identification of specific or unique metabolites, such as norcocaine or cocaethylene, a compound that is formed when concomitant cocaine and ethanol are used. The detection of specific markers is not easy, as their concentrations in hair are generally low. Therefore, some authors have proposed the use of drug ratios, like morphine-codeine, to document heroin abuse, or benzoylec-goninecocaine greater than 0.05, to document cocaine abuse. Finally, positive cutoffs have been published to insure international uniformity (Table 1).
After collection, the hair specimen is best stored at ambient temperature. Once incorporated in hair, drugs are very stable. Cocaine and benzoylecgonine have been detected in hair from Peruvian mummies, clearly demonstrating 500 years of stability.
Hair Collection and Analysis
Strands of hair (about 60 to 80) are cut as close as possible to the skin, in the posterior vertex region, dried and stored in tubes at room temperature. The root-to-end direction must be indicated. In the case of very short hair, pubic hair can also be collected.
Typical hair preparation involves the following steps:
• decontamination of the strand in organic solvent or buffer;
• pulverization of 100 mg in a ball-mill;
• hydrolysis of a 50 mg sample in acid or alkaline buffer;
• purification by solid-phase or liquid-liquid extraction;
• derivatization;
• analysis with gas chromatography coupled with mass spectrometry (GC-MS).
Table 1 Proposed positive cutoffs
| Analytes | Cutoff (ng mg 1) |
| Heroin | 0.5 for 6-acetylmorphine |
| Cocaine | 0.5 for cocaine |
| Amphetamine, MDMA | 0.5 for both drugs |
| Cannabis | Not decided |
Care is necessary to prevent the conversion of cocaine to ecgonine, or 6-acetylmorphine to morphine, in alkaline solution. Differences in efficiency between enzymatic and acid hydrolysis are not statistically significant.
A critical element in the acceptance of hair analysis for detection of drugs of abuse is laboratory performance. Laboratories must be able to demonstrate that they can accurately determine what drugs are present in unknown hair samples and at what concentrations. Several international interlaboratory comparisons of qualitative and quantitative determinations of drugs have been organized in the United States (National Institute of Standards and Technologies, Gaithersburg, MD) and Europe (Society of Hair Testing, Strasbourg, France). Interlaboratory comparisons of hair analysis have been published for opiates, cocaine, cannabis and amphetamines. In most cases, GC-MS was used for the analyses. However, no one extraction method could be identified as being superior to others.
In 1999, the following compounds were reported to be detectable in hair:drugs of abuse (opiates, cocaine, cannabis, amphetamines, methadone, phencyclidine, narcotics); pharmaceuticals (barbiturates, antidepressants, benzodiazepines, neuroleptics, etc.); nicotine; doping agents (p-adrenergic drugs, anabolic steroids and corticosteroids); pesticides.
Measured concentrations are expressed in picograms or nanograms per microgram.
Detection of Drugs of Abuse in Hair
Opiates
Three methods of screening for opiates, cocaine, cannabinoids and amphetamine, including its derivatives, dominate in the literature; these are briefly described in Table 2. Liquid-liquid extraction after HCl hydrolysis and solid-phase extraction after enzymatic hydrolysis with p-glucuronidase/sulfatase lead to similar results, both with the disadvantage that heroin and 6-O-acetylmorphine (MAM) might be hydrolyzed to morphine. The methanol method is undoubtedly the simplest, with high sensitivity for heroin and cocaine but poor sensitivity for their metabolites, morphine and benzoylecgonine; and high sensitivity for A9-tetrahydrocannabinol (THC) but no sensitivity for THC-COOH. In 1995, it was confirmed by systematic extraction studies that methanol and water had the best extraction capability for opiates, but using hydrophobic solvents like dioxane and acetonitrile, a low extraction rate was found. With toluene, almost no extraction occurred. The range of positive results using these procedures is listed in Table 3. Pubic hair showed higher drug levels than scalp hair. This can be due to the slightly lower growth rate of pubic hair than scalp hair but, additionally, pubic and scalp hair have totally different telogen:anagen ratios and concentrations cannot be directly compared. Regarding individual growth rate and the problem of telogen:anagen ratios, dose-concentration relation studies should only be performed with hair samples grown from the shaved skin before drug administration and under control of the growth speed of the hair.
Table 2 Screening procedures for the detection of illegal drugs in hair
| Kauert (1996) | Kintz (1995) | Moeller (1993) | |
| Analytes | Heroin, 6-MAM, | Heroin, 6-MAM, | Heroin, 6-MAM, dihydrocodeine, |
| dihydrocodeine, codeine, | dihydrocodeine, codeine, | codeine, methadone, THC, | |
| methadone, THC, cocaine, | methadone, THC, cocaine, | cocaine, amphetamine, | |
| amphetamine, MDMA, | amphetamine, MDMA, | MDMA, MDEA, MDA | |
| MDEA, MDA | MDEA, MDA | ||
| Decontamination step | Ultrasonic 5 min each | 5 ml Cl2CH2 | 20 ml H20(2x) |
| 5 ml H20 | (2 x 5 min) | 20ml acetone | |
| 5 ml acetone | |||
| 5 ml petrolether | |||
| Homogenization | 100 mg hair cut into small sections in a 30 ml vial | Ball-mill | Ball-mill |
| Extraction | 4 ml methanol ultrasonic | 50mg powdered hair, 1 ml 0.1 | 20-30 mg powdered hair, 2 ml |
| 5 h at 50°C | N HCl, 16hat56°C | acetate buffer + p-glucuronidase/ arylsulfatase, 90min at 40°C | |
| Clean-up | None | (NH4)2HP04; extraction 10 ml CHCl3/2-propanol/n-hepta-ne (50:17:33); organic phase purified with 0.2 N HCl; HCl phase to pH 8.4; re-extraction with CHCl3 |
NaHC03; SPE (C18), elution with 2 ml acetone/CH2Cl2 (3:1) |
| Derivatization | Propionic acid anhydride | 40 nl BSTFA/1% TMCS; 20min at 70 °C | 1000 nl PFPA/75 nl PF-n-propanol; 30min at 60°C; N2 at 60°C; 50 nl ethylacetate |
In hair specimens of 20 subjects receiving intravenous heroin hydrochloride, no correlation between the doses administered and the concentrations of total opiates in hair was observed. However, when considering a single analyte, it was noted that the correlation coefficient seemed to be linked to its plasma half-life. A weak correlation coefficient corresponds to a drug with a short plasma half-life, and the correlation coefficient increases when plasma half-life increases, from heroin, 6-acetylmorphine to morphine.
The so-called poppy seed problem could by solved by examining hair for morphine after poppy seed ingestion, as morphine is not detected in hair after consumption of seeds, or at least only in traces.
Cocaine
The fact that the parent drug is found in higher concentrations in the hair of drug users has been well known since 1991. Typical concentration ranges are listed in Table 4.
Contrary to the case in heroin abuse, cocaine consumption can be detected by measurable metabolites which cannot be caused by cocaine contamination, like norcocaine or cocaethylene. The determination of the pyrolysis product of cocaine, the anhydroecgo-nine methyl ester (AEME), can be helpful in distinguishing cocaine and crack users.
The literature and the scientific meetings concerning cocaine are dominated by discussion as to whether decontamination procedures can remove external contamination completely, and whether a racial element exists. This is important when hair analysis is used as ‘stand-alone’ evidence for workplace testing.
Table 3 Published ranges of 6-O-acetylmorphine concentration in the hair of heroin users
| Authors | Concentration (ngmg 1) |
| Kauert (1996) | 0.03-79.8 |
| Kintz (1995) | 0-84.3 |
| Moeller (1993) | 2.0-74 |
| Pepin (1997) | 0.3-131.2 |
Table 4 Published ranges of cocaine concentration in the hair of drug users
| Authors | Concentration (ng mg 1) |
| Kauert (1996) | 0.04-129.7 |
| Kintz (1995) | 0.4-78.4 |
| Moeller (1993) | 0.3-127.0 |
| Pepin (1997) | 0.89-242.0 |
An important study with controlled doses of cocaine-ii5 was published in 1996. The deuterium-labeled cocaine was administered intravenously and/ or intranasally in doses of 0.6-4.2 mg kg-1 under controlled conditions. A single dose could be detected for 2-6 months; the minimum detectable dose appeared to be between 22 and 35 mg; but within the range of doses used in the study, the hair test did not provide an accurate record of the amount, time or duration of drug use.
Cocaine, benzoylecgonine and ecgonine methyl-ester have also been found in the mummified bodies of ancient Peruvian coca-leaf chewers. In contrast to today’s cocaine users, the cocaine:benzoylecgonine ratio was less than 1.
Cannabis
In 1996, the first results on levels of cannabis in hair measured by using GC-MS were reported, simultaneously with the determination in the same run of THC and its major metabolite THC-COOH. The measured concentrations were low, particularly in comparison with other drugs. Some authors suggested the use of negative chemical ionization to target the drugs, or the application of tandem mass spectrometry. More recently, a simpler method was proposed, based on the simultaneous identification of cannabinol, cannabidiol and THC. This procedure appears to be a screening method that is rapid and economical and does not require derivatization prior to analysis. As THC, cannabinol and cannabidiol are present in smoke, to avoid potential external contamination the endogenous metabolite THC-COOH should be secondarily tested to confirm drug use.
As shown in Table 5, the concentrations measured are very low, particularly for THC-COOH, which is seldom identified. To date, there is no consensus on cutoff values for cannabis. An international debate must be held to discuss the differences noted between
Table 5 Reported concentrations for cannabis in hair
| Authors | Compound | Number of positives | Concentration (ng/mg) |
| Cairns (1995) | THC-C00H | + 3000 | (0.0007) |
| Jurado (1995) | THC | 298 | 0.06-7.63 (0.97) |
| THC-C00H | 298 | 0.06-3.87 (0.50) | |
| Kauert (1996) | THC | 104 | 0.009-16.70 (1.501) |
| Kintz (1995) | THC | 89 | 0.10-3.39(0.64) |
| Cannabidiol | 306 | 0.03-3.00 (0.51) | |
| Cannabinol | 268 | 0.01-1.07(0.16) | |
| THC-C00H | 267 | 0.05-0.39(0.10) | |
| Moeller (1993) | THC | 10 | 0.4-6.2 (2.0) |
| THC-C00H | 2 | 1.7-5.0(3.3) | |
| Wilkins (1995) | THC | 8 | 0.03-1.1 |
American laboratories, which reported THC-COOH in the low picogram per milligram range, and some European laboratories, which gave concentrations in the low nanogram per milligram range, as is obvious from the measured concentrations shown.
Amphetamine derivatives
Almost all of the literature dealing with amphetamines in hair has been written by Japanese researchers. In most cases, amphetamine and methamphtetamine have been the target drugs. More recently, particular attention has been focused on methylenedioxy-amphetamine (MDA) derivatives, like methylenedi-oxymethamphetamine (MDMA). Most techniques published used acid or alkaline hydrolysis, or a combination of hydrochloric acid and methanol, followed by a purification step (liquid-liquid extraction or solid-phase extraction) and derivatization.
When comparing four different procedures for amphetamine, MDA and MDMA (methanol sonica-tion, acid hydrolysis, alkaline hydrolysis and enzymatic hydrolysis) it was demonstrated that best recovery rates were observed after alkaline hydrolysis; however, it was not possible to determine which method performed best, based on recovery rate, precision and practicability. Lower concentrations were observed after methanol sonication, together with dirty chromatograms.
It must be pointed out that, since the first identification of MDMA in human hair in 1992, this compound, particularly in Europe, is one of the most frequently identified and must therefore be included in all screening procedures.
Typical findings for amphetamine derivatives are given in Table 6.
Although there are still controversies surrounding the interpretation of results, particularly concerning external contamination, cosmetic treatments, ethnic bias or drug incorporation, pure analytical work in hair analysis has more or less reached a plateau, having solved almost all the analytical problems. Conferences on hair analysis in Genoa, Strasbourg, Tampa and Abu Dhabi, between 1992 and 1996, indicate the increasing role of this method for the investigation of drug abuse.
Table 6 Analytical parameters and results for a general screening procedure for amphetamine derivatives
| Compound | Ions monitored (m/z) | Linearity (r) | Precision | Concentration |
| (at2ngmg-1,%) | (ngmg-1) | |||
| Amphetamine | 91, 118, 240 | 0.998 | 6.9 | 2.3-20.6 (n = 5) |
| Methamphetamine | 169, 210, 254 | 0.995 | 8.4 | |
| MDA | 135, 240, 375 | 0.994 | 9.1 | 0.4-8.0 (n = 13) |
| MDMA | 210, 254, 389 | 0.996 | 10.2 | 0.3-42.7 (n = 14) |
| MDEA | 240, 268, 403 | 0.997 | 13.0 | 0.6-69.3 (n = 6) |
| MBDB | 176, 268, 403 | 0.994 | 8.7 | 1.41-3.09 (n = 2) |
| BDB | 135, 176, 389 | 0.996 | 9.4 | 0.21 (n =1) |
Applications
In the case of segmental analysis, to evaluate the pattern of drug abuse for example, proximal and distal portions of hair must be identified. Given the variation in hair growth rates, generally 1.0-1.3 cm per month, results from a multisectional analysis should not be used to determine a precise period of drug exposure. The further away from the hair root, the more cautious the interpretation of quantitative findings of the individual hair sections must be.
Table 7 lists the major characteristics of both urine and hair analyses.The major practical advantage of hair testing compared with urine testing for drugs is its larger surveillance window: weeks to months in hair, depending on the length of the hair shaft, versus 2-4 days in urine for most xenobiotics, except can-nabis. In fact, for practical purposes, the two tests complement each other. Urinalysis provides short-term information on an individual’s drug use, whereas long-term histories are accessible through hair analysis. While analysis of urine specimens cannot distinguish between chronic use or single exposure, hair analysis makes this distinction. Its greatest use, however, may be in identifying false negatives, as neither abstaining from a drug for a few days nor trying to ‘beat the test’ by diluting urine will alter the concentration in hair. Urine does not indicate the frequency of drug intake in subjects, who might deliberately abstain for several days before screening.
It is always possible to obtain a fresh, identical hair sample if there is any claim of a specimen mix-up or breach in the chain of custody. This makes hair analysis essentially fail-safe, in contrast to urinalysis, as an identical urine specimen cannot be obtained at a later date. Clearly, hair analysis can thus function as a ‘safety net’ for urinalysis.
Numerous forensic applications have been described in the literature where hair analysis was used to document the case: differentiation between a drug dealer and a drug consumer, chronic poisoning, crime under the influence of a drug, child sedation, child abuse, doubtful death, child custody, abuse ofdrugs in jail, body identification, survey of drug addicts, chemical submission, obtaining a driving license and doping control. It appears that the value of hair analysis for the identification of drug users is steadily gaining recognition. This can be seen from its growing use in preemployment screening, in forensic sciences and in clinical applications. Hair analysis may be a useful adjunct to conventional drug testing in toxicology. Methods for evading urinalysis do not affect hair analysis. Specimens can be more easily obtained with less embarrassment, and hair can provide a more accurate history of drug use. Costs are too high for routine use but the generated data are extremely helpful in documenting positive cases.
Table 7 Comparison between urine and hair for testing drugs of abuse
| Parameter | Urine | Hair |
| Drugs | All | All |
| Major compound | Metabolites | Parent drug |
| Detection period | 2-4 days, except cannabis | Weeks, months |
| Type of measure | Incremental | Cumulative |
| Screening | Yes | Yes |
| Invasiveness | High | Low |
| Storage | -20 °C | Ambient temperature |
| Risk of false negative | High | Low |
| Risk of false positive | Low | Undetermined |
| Risk of adulteration | High | Low |
| Control material | Yes | Yes |
