Hair Analysis Panel Discussion:
Section: Appendix C, Robert S. Baratz

Appendix C
Pre-Meeting Comments

Hair Analysis: Exploring the State of the Science

Robert S. Baratz

Hair Analysis

By Robert S. Baratz, MD, PhD, DDS

Introduction

When physicians study a disease or process, they look for ways to evaluate that process in the body. Blood and urine are taken for testing because they are easily obtained and can be readily standardized. Normal values for populations can be set easily with such testing. Blood generally represents what is inside the body, and urine represents what is excreted from the body.

Hair testing has very limited usefulness in medical practice, because it does not represent either the tissues inside the body or what is excreted. Hair analysis is only useful for detecting exotic compounds that are not normally found in the body. Thus, for example, a medicine that someone is taking, might be detected in the hair. Poisons, such as arsenic, also show up in the hair. Elements normally found in the body -- such as copper, chromium, zinc, and even lead, mercury, and uranium-- will show up in the hair, but the levels are quite variable and have little or no practical or clinical significance.

Analysis of hair won't tell you about the source of an element found in the hair. Most minerals obtained by the body come from food or water. Foods are grown all over the country and thus, their constituents, more likely than not, have come from another region, in some cases, from another country. People are more commonly drinking bottled water and juices which also are coming from other regions. Thus, finding something in the hair or body in no way indicates the source of the material. This is especially problematic when dealing with elements that are somewhat ubiquitous in the environment. The most common source of lead, for example, can be from the solder joints of household plumbing. However, lead could also be introduced through any number of foods, and/or beverages. In some cases, even unglazed pottery used for serving food can be a source of lead contamination.

When hair analyses have been done rigorously in quantitative laboratory settings, it has been pointed out that great care must be taken to avoid possible sources of contamination. First, the hair itself must be processed in a uniform fashion to avoid introducing any exogenous material. Metal cutting instruments in sampling hair should be avoided. The hair sample must be standardized as to region of the scalp, length from the scalp and any washing done of the hair during processing. Even so, because hair grows at different rates in different people, there is still a great deal of uncertainty regarding even hair obtained close to the scalp. Many contend that such hair is of more recent vintage and thus more "representative" of the "body composition". There are no data, however, that confirm this idea. Hair seems to grow at an average rate of about 1 cm per month. However, a considerable portion of the hair shaft lies within the skin and thus hair that has been sampled that has already grown out represents hair that may be as many as several months old.

Even if hair analysis was a highly reproducible and valid test, done properly, there are virtually no data for correlation of findings with levels of elemental minerals found in other tissues or organs. Given the element of interest in the Colorado plateau region, it is important to point out that radioactive compounds from tailings are unlikely to go into hair. The agent of most interest is radon which ends up in the body as lead. Because radon exists primarily as a gas, the major organ that is affected is the lung. This reviewer is unaware of any studies that correlate amounts of material found in lungs with amounts found in hair.

As hair is handled in the laboratory, a number of possible contaminants can enter the hair from solutions used in processing. Rigorous care must be taken to check each and every reagent used in the laboratory. Even acid reagents can have significant amounts of trace elements within them when parts per billion are at issue. Water used in the laboratory for washing, hand washing or even wiping down counters may contaminate samples. Use of vaporization techniques such as atomic absorption spectroscopy can release agents into the laboratory air which would then end up contaminating other samples or solutions or both. Laboratory dust must be excluded since it too can act as a source of contamination. Even powder used to cover gloves of laboratory workers can result in significant contamination of the laboratory environment. Analyses are often done at the level of parts per billion and it takes very little contaminating material to change findings dramatically.

Many laboratories that handle hair fail to take into account that exogenous contaminants such as hair shampoos, swimming pools, shower water and the like can all add exogenous agents to hair. These include: selenium, bromine, zinc, copper and even arsenic. Some elements are removed by pre-washing before hair analysis. Acetone, a common washing agent, has been shown to remove sodium, bromine, and calcium. The same solution is known to add copper, iron, manganese, zinc, and mercury. Even the pH of washing solutions can effect the amounts of lead, mercury and cadmium found in hair samples.

Topic I – Analytic Methods

Types of Analytical Methods

The principal means for analysis of hair depends on the object of the analysis(es). Simple small molecules such as trace minerals, can be analyzed either using atomic absorption spectroscopy or mass spectroscopy. Analysis for organic compounds would depend on the specific compound being tested.

Some of these methods are exquisitely sensitive and small quantities of contaminants found in laboratory air from vaporization, dust, or coatings on lab-ware and/or contamination of test solutions can significantly affect results.

For trace minerals, results are commonly in the range of parts per billion, or smaller. Thus, exquisite attention to detail and lab cleanliness must be followed. Reliable analytical methods exist for detection of most trace minerals, however, quantification becomes an issue particularly when multiple overlapping peaks with spectroscopy occur. Moreover, sampling errors, and the nature of the starting material, often inhibit precise quantification. More commonly, qualitative results can be observed reliably.

For reasons to be discussed in later sections numeric quantitative evaluation of trace mineral substances is not clinically, forensically, or for matters of industrial hygiene, useful. This is largely due to the fact that normal ranges have not been established, cannot be established, or are irrelevant. A finding of an exotic substance that is never normally present is significant. Similarly, a change in order of magnitude of a trace substance that is normally present, and may, at high doses, be poisonous, often has clinical and other relevancies.

Hair analysis has been shown to be quantitatively useful for the detection of arsenic, and methyl mercury. Other validated uses of hair analysis are for the finding the presence of drugs of abuse or the presence of certain pharmacological agents.

In this reviewer's experience, commercial laboratories, as opposed to research laboratories, have been observed to have considerable variation in their performance. This variability is the result of inconsistent specimen preparation, source, and handling, inconsistent use of standards, and lack of multiple runs of the same material. Typically, only single samples are run and thus any variability within the laboratory and/or method are often unknown. Where multiple samples have been observed from the same laboratory on the same material, wide variations have been shown to exist. In the case of commercial laboratories, interpretation of results suggests that results are often misleading, inappropriate, and lack sufficient information to make them useful. Two reports by Barrett ( 1985) and Seidel and colleagues (2001) show that, at least in the case of commercial laboratories, reference ranges, results and interpretations vary considerably from laboratory to laboratory. This is not surprising considering the milieu in which this work is done, and the factors described above.

Topic II- Factors Influencing the Interpretation of Analytical Results

Regionally, there can be marked differences in elemental composition of hair even for the same element. For example, in 16 different regions of the scalp, antimony content was shown to vary considerably. Even with a person with a standardized diet and living conditions, the composition of hair at different distances out from the scalp itself can vary. This has been shown, in particular, for copper and zinc.

Additional problems in doing hair analysis show that there are difficulties in trying to measure more than two or three elements at the same time. In atomic absorption spectroscopy, one of the more common methods for hair mineral content analysis, many elements give off multiple peaks which overlap. These absorption peaks obscure each other, negating the ability to do quantitative analyses accurately.

There is often a lack of precision and standardization in the amount of hair taken from any particular subject. Unless a uniform sample was taken from which all analyses were done, the validity of the analysis can be called to question.

Racial differences among subjects have also been found. There is considerable variability in calcium, iron, nickel, chromium, manganese, arsenic and lead levels between Caucasian subjects and blacks.

Similarly, age is a significant factor in metal composition of hair. Paschal and co-workers (1989) found marked differences in concentrations of 28 different metals in hair samples of 199 children compared with 322 adults. Age-dependent increases in calcium, barium, magnesium, zinc and strontium all occur up to about 12-14 years of age. Aluminum is shown to decrease with age. It has been hypothesized that metal composition of hair is related to skeletal and bony growth. Thus, adults undergoing osteoporosis would have differences in their hair composition related to those who did not have this problem. Similarly, anyone with any kind of bone abnormality would have findings that are non-standard.

Most analyses on hair do not correlate positively with concentrations found in organs (Yoshinaga and co-workers 1990). It is intuitively obvious why this is likely so, since tissue concentrations involve both uptake and release which vary over time. Hair is essentially a one-way path out of the body. Likewise, some elements have significant diurnal variation. A good example is chromium (Sheard and co-workers 1980).

Many elements, when analyzed in the presence of other elements, can give false readings. The interaction of chromium with other anions and cations in hair may affect analytical results (Sheard and coworkers 1980). Merely painting the laboratory with particular types of paint, failure to use HEPA filters on the air intake and the presence of dust can easily affect sensitive analytical measurements.

A variety of hair treatments have been shown to alter hair trace element concentrations. (McKenzie, 1978). Other common issues are dyeing and permanent waving, shampooing, hair color, sex, seasonal variations, age, and growth rates. It is generally assumed that hair grows approximately 1 cm per month, however this must be verified in each individual tested. Hair growth is a function of individual factors as well as protein in the diet.

Many commercial laboratories claim to be able to detect and measure more than 20 elements in a single sample of hair, however, this is often accomplished without any specific knowledge of the patient's medical history. What is more troubling, is that there is no definition of a normal range (Hambridge 1982, Rivlin 1983; Manson and Zlotkin 1985, Barrett, 1985, Druyan and others 1998 and Seidel and others 2001;). Quality control in the laboratory is essential towards having useful data. Rigorous attention to detail, methodology, and sampling techniques must be followed. Even when known standards were used, because of the sensitivity of instrumentation, data varied commonly by up to 10%. (Nowak and Kozkowski 1998).

Variations in Sample Collection and Preparation Methods

A number of compounding variables limit interpretation of results from hair analysis. It is well known from the literature that the rate of hair growth varies from person to person, with nutritional and disease states, with the presence of particular drugs, with gender, age, ethnicity/race, with site on the scalp and/or other body parts. While some of the factors may be known in an individual case, others are unknown, or cannot be known. Thus, hair analysis from a particular individual is fraught with a series of uncontrolled variables and unknown data. It should be obvious that these belie making any precise quantitative diagnostic or forensic analysis. This becomes even more of a problem when dealing with trace minerals that are normally found ubiquitously in the environment and characteristically in foods, water, and air. Many trace minerals occur in human hair normally. Thus, finding them there is expected. Making interpretations based upon the quantitative analysis of these is fraught with uncertainty due to the unreliability of the data regarding exposure, timing, hair growth, treatment of the hair, diet, nutrition, and a host of other factors mentioned above.

Similarly, considerable variation exists from laboratory to laboratory in terms of sample preparation, whether a sample is washed, how it is digested, how long it is digested, and how it is handled after digestion.

Even more problematic is the development of "normal ranges" or "reference standards" ("reference ranges"). In most cases, population standards have not been developed. Thus, each laboratory has developed its own "reference range". The major problem with this is that the source of the specimens used to create the "reference range" in a particular lab may be biased. Many commercial laboratories accept samples from a variety of practitioners, patients, and other sources. From reports this reviewer has seen, the precision of medical knowledge and facts regarding the source material is often poorly documented. Careful attention to uniform sample collection techniques is often also a problem.

So-called "normal" reference ranges do not exist for most trace minerals found in hair. The reasons for this are obvious. Considerable variation exists from person to person and the variety of unknown variables enter into the equation. Thus, there are no standardized "reference ranges" for most normal trace minerals. This has to do, in part, with the composition of hair. In essence, hair consists of keratinized or cornified cells packed into tight arrays in the hair shaft. These cells are fundamentally similar to the epidermis however contain proportionally more keratin fibrils and somewhat different materials in the thickened cell membrane that is left when the cells keratinize. Hair, nails, horn, and some portions of the filiform papillae of some animal tongues are a so-called "hard" keratin compared with so-called "soft" keratins found in epidermis and oral and other, mucosae. All keratinized cells contain virtually no aqueous phase after keratinization. It is unclear if minerals are removed from these cells when they mature, or merely remain in the intracellular matrix. Along these lines, even if some minerals were found to be left "inside" such cells, each and every individual trace mineral would have to be studied in pre-keratinized cells and keratinized cells to see how it was handled.

Moreover, epithelium and its derivatives (hair) is an a vascular tissue with little intercellular space or material. What little extracellular material exists is primarily a lipid that forms a barrier to diffusion. The epithelium is neither a gland nor excretory organ but merely forms a protective layer. Thus, substances that would normally be excreted into various body fluids are normally not present in epidermal epithelium. Regionally, there is variability in the thickness of the epidermal epithelium and indeed there is some variability in the consistency of thickness of hair in different regions of the scalp. Hair in other body locations, axillary, pubic, limb, peri-anal, eyebrows, and eyelashes, all vary considerably in their structure, function, and growth rates.

Since hair is principally protein in nature, there is little need for trace minerals in the hair cells themselves. Trace minerals in the body are usually present as co-factors for enzymes. Keratinized cells are generally non-metabollic. After the filaments of keratine aggregate and are coated by other proteinatious material, the cell contents become essentially inert. Nuclear materials, enzymes, carbohydrates, and even lipids are essentially not present in the internal milieu of keratinized cells. Consequently, there would be no need for a regular array of minerals present from a functional point of view.

Some heavy metals may distribute into hair and become complexed with hair proteins. This would be due largely to interaction with free side chains on amino acids and/or forming crosslinks among protein chains as they may be denatured by heavy metals. Some heavy metals are well known as protein denaturants, e.g. mercuric chloride. They may become trapped in hair cells before they become completely keratinized. Whether or not this happens is largely unknown.

Finding trace minerals in hair is neither surprising nor a consistent finding. Because hair shafts consist of essentially of two portions, intra-epithelial and extra-epithelial, the possible absorption of extraneous material is possible in the extra-epithelial portion. The extra-epithelial portion is essentially free in the environment. Thus it is subject to washing, drying, chemical alteration, cosmetics, environmental pollutants present in the water or air, and a host of other chemical and physical insults. Not only may things be adsorbed and absorbed by the hair, but, substances may also be leached from the hair as well. Prolonged immersion and wetting of hair can cause some swelling of the cells of which hair is composed. This can diminish the barriers to diffusion of things both from the outside in, and from the inside out. Moreover, hair is being constantly exposed to scalp oils, and other glandular products excreted into the hair shaft space by sebaceous and other glands present in skin. These provide an additional source of extraneous material to be adsorbed or absorbed onto or into the hair.

A common and highly variable factor in hair is hair growth. Hair growth can occur in several different ways. First, is the fact that hair undergoes a cycle in its normal growth. That is, hair is regularly shed from the scalp and other locations, and replaced by "new hairs". The stages of the growth (catagen, anagen, telogen) each have unknown times associated with them in particular subjects. Further complicating an understanding of growth is the fact that hair in humans is known to grow in a mosaic across the scalp. That is, any particular hair may be in a different state than its neighbors. A long list of drugs, hormones, and other factors can either accelerate or prolong the time a hair stays in a particular part of its growth cycle. Moreover, a number of other factors such as diet, nutrition, age, sex, hair color, and other factors are known to influence growth rates.

Growth can occur both longitudinally and in diameter. Hair in general varies from individual to individual in shape and not all individuals have a circular cross section of their hair. In particular, individuals with highly curled or "kinky" hair have hair that is somewhat flattened to a ribbon-like shape.

In general, hair growth in length is often described at approximately 1 cm per month. However, there is considerable variation in this from individual to individual and results can vary by a factor of 2 either in increase or decrease in rate of growth.

Topic #3: Toxicologic Considerations

As previously mentioned, among the mineral toxic agents studied only arsenic and methyl mercury have been shown to have reliable information on their presence and distribution in hair when viewed in comparison to their distribution in other organs.

To have predictive value, the values obtained from analysis of hair of a particular subject must be capable of yielding data that would be predictive for disease in general. This may prove to be considerably problematic in the case of heavy metals as the agents themselves may affect hair growth directly.

This reviewer is less well versed in arsenic and methyl mercury studies than others on the panel and wishes to defer to their knowledge and experience.

Topic #4: Data Gaps and Research Needs

Many of the data gaps in our knowledge of hair physiology and growth have been discussed earlier.

In one sense each trace mineral must be independently studied with regard to the best source for analytical material. In most cases it is likely that hair will prove to be a problematic source. While hair theoretically gives a longitudinal history of prior events, the speed of that history is largely unknown, and may even change over time. Whether this theory meets practice is also unknown. A better understanding of the physiology of hair growth is obviously an important area of research. This, of course, begs the question that hair may be useful at all for mineral analyses. This later point is yet unproven. A variety of data would suggest that hair is not useful for mineral analyses for most minerals, and that other body sources would be better—e.g. bone, teeth.

Knowledge of the dynamics of incorporation of a variety of environmental toxins, principally organic compounds, into hair would be desirable. Attendant to such a study would be studies of absorption, adsorption and leaching of such compounds.

Studies of the nature of differences in incorporation of materials into hair at different ages, by different sexes, different ethnic groups, and different hair colors would also be useful.

Topic #5: Identifying Scenarios for Which Hair Analysis May Be Appropriate

Hair analysis appears useful only for population studies where much of the individual variability can be eliminated. If a number of factors were known—duration of exposure, rates of incorporation into hair, effects on growth, amounts of leaching, sources of material that were found in hair, etc.—then useful data on exposure could be extracted. Correlating these with clinical findings is more problematic, since such are best done on the individual level, where hair analyses are likely more useful only for population studies.

Particularly for small molecules such as trace minerals, hair is unlikely to prove a reliable source of material for meaningful study.

Organic compounds that can be shown to incorporate into hair may be an area where hair analysis could be appropriate for following exposures to environmental toxins.

Selected References

• Anderson RA et al. "Designing a Biological Monitoring Program to Assess Community Exposure to Chromium: Conclusions of an Expert Panel" Journal of Toxicology and Environmental Heath, 40:555-583, 1993.
• Barrett S. "Commercial Hair Analysis: Science or Scam?"; JAMA; 254(8): 1041-1045; 1985.
• Barrett, S. "Commercial Hair Analysis: A Cardinal Sign of Quackery", www.quackwatch.com, 2000.
• Centers for Disease Control, "Blood and Hair Mercury Levels in Young Children and Women of Childbearing Age—United States, 1999", MMWR, 50 (08); 140-3; March 2, 2001.
• Chittleborough G. "A Chemist's View of the Analysis of Human Hair for Trace Elements", The Science of the Total Environment, 14: 53-75, 1980.
• Clarkson TW. "The Toxicology of Mercury", Crit Rev Clin Lab Sci, 34(4):369-403; 1997.
• Cornelis R. "Neutron activation analysis of hair failure of a mission" J Radioanal Chem; 15:305-316; 1973.
• Deening SB and Wever CW; "Hair Analysis of trace minerals in human subjects as influenced by age, sex and contraceptive drug" Am J Clin Nutr ; 31:1175-1180; 1978.
• Druyan ME et al. "Determination of Reference Ranges for Elements in Human Scalp Hair", Biological Trace Element Research, Vol 62, pp 183-197, 1998.
• Fletcher DJ. "Hair Analysis: Proven and Problematic Applications" Postgraduate Medicine, Vol 72, No 5, pp 79-81,84,87-88, 1982.
• Gaillard Y and Pepin G. "Testing Hair for Pharmaceuticals"; J Chromatogr B Biomed Sci Appl, 733(1-2):231-46; 1999.
• Hambidge KM. "Hair Analyses: worthless for vitamins, limited for minerals", Am J Clin Nutr; 36(5): 943-9; 1982
• Hopps HC. "The Biologic Bases for Using Hair and Nail for Analysis of Trace Elements", The Science of the Toxic Environment, 7: 71-89; 1977.
• Kidwell DA, Lee EH, and DeLauder SF. "Evidence for bias in hair testing and procedures to correct bias"; Forensic Sci Int; 107(1-3):39-61; 2000.
• Klevay LM et al. "Hair Analysis in clinical and experimental medicine" Am J Clin Nutr; 46:233-6; 1987.
• Lamand M., Faviert A., and Pineau A. "La détermination des oligoéléments dans les poils et dans les cheveux: intérêt et limites", Annales de Biologie Clinique, 48, 433-442, 1990.
• Manson P and Zlotkin S. "Hair Analysis- a critical review"; Can Med Assoc J; 133:186-188; 1985.
• McKenzie JM.. "Alteration of the zinc and copper concentration of hair"; Am J Clin Nutr, 31:470-476; 1978.
• Mertz W. "Confirmation: Chromium levels in serum, hair and sweat decline with age"; Nutr Rev; 55(10): 373-5; 1997.
• Nowak B and Chmielnicka J. "Relationship of lead and cadmium to Essential Elements in Hair, Teeth, and Nails of Environmentally Exposed People"; Ecotoxicol Environ Saf; 46(3):265-274; 2000.
• Nowak B and Kozlowski H. "Heavy Metals in Human Hair and Teeth: The Correlation with Metal Concentration in the Environment", Biological Trace Element Research, Vol 62, 1998.
• Paschal DC et al. "Age Dependence of Metals in Hair in a Selected U.S. Population", Environmental Research, 48, 17-28, 1989.
• Ponce RA et al. "Uncertainty analysis methods for comparing predictive models and biomarkers: A Case study of dietary methyl mercury exposure"; Regul Toxicol Pharmacol; 28(2): 96-105; 1998.
• Rivlin RS. "Misuse of Hair Analysis for Nutritional Assessment"; Amer Jour of Med; Vol 75: 489-493; 1983.
• Seidel S. personal communication to John L. Watson, March 6, 2001.
• Sheard EA, Johnson MK, and Carter RJ. "The Determination of Chromium in Hair and other Biological Materials" Hair, Trace Elements and Human Illness, Praeger, NY, 1980.
• Sky-Peck HH. "Distribution of Trace Elements in Human Hair", Clin Physiol Biochem:8:70-80, 1990.
• Teresa M, Vasconcelos SD, and Tavares HMF. "Trace Element Concentration in Blood and Hair of Young Apprentices of a Technical-professional School", The Science of the Total Environment, 205, 189-199, 1997.
• Vir SC and Love AHG. "Zinc and copper nutriture of women taking oral contraceptive agents" Am J Clin Nutr; 34:1479-1483; 1981
• Wennig R. "Potential problems with the interpretation of hair analysis results"; Forensic Sci Int 107(1-3); 5-12; 2000.
• Willhelm M and Idel H. "Hair Analysis in Environmental Medicine", Zeutralblatt for Hygiene and Unweltmedizin; 198:485-501; 1996.
• Yoshinaga J et al. "Lack of Significantly Positive Correlations Between Elemental Concentrations in Hair and in Organs"; The Science of the Total Environment, 99: 125-135; 1990.
• Zlotkin SH, "Hair Analysis: A Useful Tool or a Waste of Money?" International Journal of Dermatology, Vol 24, 161-164; 1985.

Next | Table of Contents

Top of Page