The Cross-Section Trichometer: A New Device for Measuring Quantity, , and Hair Growth à BERNARD COHEN,MD

BACKGROUND Office physicians are unable to measure hair quantity, hair loss, and hair growth in a simple and meaningful manner. One solution is to measure the cross-sectional area of a bundle of hair that is growing within a premeasured cross-section of scalp. OBJECTIVE The objective was to design a mechanical device that precisely measures the cross-sec- tional area of a bundle of hair and design a device that can precisely delineate an area of scalp. It was assumed that density and diameter changes are evidenced by changes in the bundle cross-sectional area and that growth and loss are the result of density and diameter changes. These assumptions were confirmed using various sized bundles of known diameter non-hair filaments. MATERIALS AND METHODS Bundles of hair and surgical silk fibers were tested using a mechanical device that compressed the bundle and measured its cross-sectional area. Balding patients were cat- egorized according to their observed severity of the loss. Bundles of their uncut hair from 4-cm2 scalp sites were measured and the values were compared to the patient’s category of hair loss severity. RESULTS In patients with balding, there was a direct correlation between the bundle’s cross-sectional area and the observed severity of the loss. The cross-sectional area was expressed as square millimeters of hair per square centimeter of  100 (mm2/cm2  100) and named the trichometric index (TI). Using surgical silk fibers, there was a direct correlation between the bundle’s cross-sectional area and the number of filaments, the diameter of the filaments, and the dry weight of the filament bundle. Using aggregates of cut human hair, there was a direct correlation between the cross-sectional area and the dry weight of the bundle. CONCLUSION This prototype device shows promise as a diagnostic instrument for measuring changes in hair quantity (mass), hair diameter, and hair density, as evidenced by preliminary studies using silk sutures, cut human hair, and patients with various degrees of balding. Formal clinical studies are need- ed. Although the device itself showed a high degree of precision, the accuracy and reproducibility of the measurements can be compromised if the sampling method is not carefully performed using magni- fication. The device is intended for use on uncut hair that is more than 1 inch in length. Dr. Cohen holds patents on the method and device described in this report and will receive royalties on the sales.

air loss affects 75% of men and 10% of hair growth result when there is a fluctuation in one Hwomen, but office physicians are unable to and/or the other. An ideal hair-measuring technology measure its parameters in a simple and meaningful should reflect the simultaneous influence of both way. Precise instrumentation and methodologies density and diameter. The present devices and have been limited to research centers and industry methods have advantages and limitations laboratories where clinical studies and drug evalua- (see Table 1). tions are performed. The office physician needs a rapid, easy, and precise method for measuring a In discussions of hair density and hair diameter, it is patient’s clinical status. helpful to draw the distinction between the termsF shedding and thinning, since both result in hair loss. Hair quantity is determined by measuring the hair’s In states of shedding, with diameters of normal density (n/cm2) and diameter (mm). Hair loss and size fall out. It is normal to shed about 50 to 100

ÃDepartment of Dermatology and Cutaneous Surgery, University of Miami, Miller School of Medicine, Coral Gables, Florida

& 2008 by the American Society for Dermatologic Surgery, Inc.  Published by Wiley Periodicals, Inc.  ISSN: 1076-0512  Dermatol Surg 2008;34:900–911  DOI: 10.1111/j.1524-4725.2008.34175.x

900 TABLE 1 Comparison of Trichometer Device and Method with Existing Hair-Measuring Methods

Reflects influence of Measures hair density and diameter mass within a Category Description Advantages Disadvantages Scale = 0 to 5 delineated area

Hamilton-Norwood Comparison Black & white Quick, simple and Incomplete hair Density = 1 Not measured. Scale1,2 chart drawings depict 12 widely accepted. loss not depicted by Diameter = 0 stages of balding. shades of gray. Hair Loss Gross visual Observe visual ratio of hair Quick, no chart Imprecise. Has Density = 3 Accuracy = 2 Severity Scale3 analysis to skin, choose between 6 required. not gained Diameter = 3 Precision = 1 categories. acceptance. Hair count4,5 Magnified Hairs 430 microns are Suitable for FDA Hair must be cut. Density = 5 Accuracy = 3 image counted using scalp studies or casual of- Office video cams Diameter = 2 to 3 Precision = 5 analysis images of trimmed hair. fice evaluations. are not precise. Clinical Gross image Popular, simple, informal. Quick, easy, Quality varies with Density = 2 Not measured. Photography analysis Cameras and light sources inexpensive. technique and hard- Diameter = 2. vary. ware. Global Gross image Technically formalized High quality control. Density = 3 to 4 Not measured. Photography4,5 analysis hardware. Photos images suitable Compares B&A of Diameter = 3 to 4 evaluated by panel. for FDA studies. same patient only. Dry hair weight5 Laboratory Hair is cut and Industry gold Technically difficult. Density = 5 Accuracy = 5 quantitative weighed after a fixed standard. Suitable Requires strict hu- Diameter = 5 Precision = 5 analysis period of growth. for FDA studies. midity control. Trichoscan Digital image Hair is cut & dyed. Com- Most precise and Hair must be cut Density = 4 to 5 Accuracy = 4 software6,7 analysis puter analyses scalp accurate of and dyed. Not Diameter = 4 to 5 Precision = 5 images of trimmed hair. office methods. widely popular. Scalp Biopsy8 Surgical A 4 mm punch biopsy is Identifies vellus Path interpretation is Density = 2 Sample too small sectioned transversely and intermediate, and difficult to learn. Diameter = 4 to 5 for adequate mea- examined. terminal hairs Post op care. surement. Growth rate9 Direct Hair growth in mm is Accurately quantifies Clinical relevance Density = 0 Not measured. measurement of measured over a fixed anagen activity. is of questionable Diameter = 0. hair length period of time (30 days). value. 47JL 08901 2008 34:7:JULY Trichometer Mechanical Measure X-sectional area No cut hair. Quick and Sampling can be Density = 5 Sampling & Device device & method measurement of hair bundle from deli- simple. Device is pre- imprecise. Hair must Diameter = 5 Accuracy = 4

of intact hair neated area of scalp. cise. be 1 inch. Precision = 4 COHEN

FDA, Food and Drug Administration. THE CROSS-SECTION TRICHOMETER

hairs per day, but in pathologic states of effluvium and , shedding can be quite profound. Underlying skin becomes more and more visible as the shedding progresses. The hair density analysis or hair count will accurately reflect this disorder.

Hair thinning is a disorder characterized by the gradual miniaturization in the length and diameter of individual scalp hairs. Underlying skin becomes more and more visible as the hairs become smaller and smaller. Thinning affects an estimated 75% of men, and although it occurs in 10% of healthy women, it might indicate an endocrine abnormality in a small group of those affected. Unlike shedding, thinning is not diffuse in its distribution over the Figure 1. In hair loss due to balding or thinning, the hairs entire scalp, but almost always appears in a pattern have a wide range of diameters. When a hair count is performed, any hair with a diameter larger than 30 mmis that spares the posterior and sides of the lower scalp, counted as one hair. creating a familiar horse-shaped fringe that persists in spite of the most advanced cases. Thinning will eventuate in lowered density as the affected hairs hair length varies with style, it is not considered in vanish. the calculation of hair quantity.

Thinning, in its earliest stages, cannot be visualized Objectives and is difficult to diagnose and quantify. Simple density counts comparing the permanent occi- The broad objective was to develop a technology for pital fringe to an area of balding are of limited measuring the quantity of hair in a defined area of value because the balding area has a mixed popula- scalp. If the density and diameter of hair determines tion of normal-sized and miniaturized hairs (see its quantity, and the cross-sectional area of an Figure 1). aggregate of hair reflects the range of densities and diameters within that aggregate, then the cross- The anatomic fluctuations of density and diameter sectional area of the aggregate reflects the quantity result in hair loss and growth, but it should be noted of hair within the aggregate. Based on this theory, a that the changes in hair diameter are considerably technology was designed to measure the cross- more influential than changes in hair density. Coarse sectional area of all the hairs in an aggregate of hair hair has a diameter of about 80 mm, average hair from a premeasured area of scalp. This value could about 70 mm, and fine hair about 60 mm. Surpris- be used to quantify the hair that is present and then, ingly, an 80-mm hair has almost twice the mass of a by comparison, quantify the amount of hair that has 60-mm hair because it has approximately double the been lost or gained. cross-sectional area (3.14 Â r2 = cross-sectional area). If two individuals have the same number of hairs, all To measure the collective cross-sections of hair in the the same length, the one with coarse hair has almost premeasured area, the ideal method/device must first twice as much hair mass as the one with fine hair. capture the hairs and compress the loose aggregate A 10% change in hair density will result in a 10% into a rectangular bundle, before the measurement is change in hair quantity. A 10% change in diameter performed. The device should always compress the will result in a 20% change in hair quantity. Because bundle with the same exact load, regardless of the

902 DERMATOLOGIC SURGERY COHEN

2 TABLE 2. Correlation between Silk Strand Diame- the hair density (n/cm ) and/or diameter (mm) ter and the Bundle Cross-section (see Tables 2–4).

Bundle Cross- 3-0 4-0 5-0 6-0 section (mm2) The device is a self-contained mechanical unit with a rectangular anodized aluminum body that is held in Strands per bundle the right hand like a hypodermic syringe (see Figure 80 3.81 60 20 3.63 2). Extending from one end of the body is a hook- 40 40 3.07 shaped arm, and from the other end of the body, a 20 60 2.70 spring-loaded shaft with a retainer cap. The hook 80 2.36 60 20 2.14 and the cap are at opposite ends of one contiguous 40 40 2.03 shaft. An electronic sensing unit with LED display is 20 60 1.65 externally mounted on the side of the housing and 80 1.37 attached internally to the shaft. When the cap at the 60 20 1.26 40 40 1.09 end of the shaft is pressed with the thumb, the hook 20 60 0.82 arm extends out of the body at the opposite end, and 80 0.66 its travel distance is displayed on the LED screen in

Eighty-strand bundles of mixed-size suture material were measured. hundredths of a millimeter. When the thumb is re- leased, the hook retracts back toward the body. hair sample size. The compressive force should com- pletely compact the bundle, but not with a force so To measure the cross-sectional area of a hair sample, excessive to damage the captured hair. The load the arm is extended and hooked around a bundle of should be incrementally applied with mechanical hair that has been gathered from a 2 Â 2-cm area of precision rather than casual hand-applied force. A scalp (see Figures 3–5). When the thumb is lifted, the precise scientific instrument should be included to bundle of hair is captured in a 1 Â 4-mm chamber measure the cross-sectional area of the compressed created on the ledge of the metal housing through and fully compacted bundle. It is preferable that no which the hook passes. The number and diameter of hair is cut and no physician supervision or oversight hairs in the captured bundle determine the height of be required. The device should be of small size, sen- the hair within the chamber. When initially captured, sibly priced, widely available, easy to use, and capable the bundle is slightly compressed and not yet com- of generating results in a short period of time. pacted.

The mid portion of the long shaft has a threaded The Resultant Device and the Method portion that passes through a large threaded knob at for Testing the base of the body. When the knob is turned A device/method that fulfilled the above objectives clockwise on the shaft, it compresses a heavy internal was created. It was tested using silk filaments and spring. As a result, the spring delivers a precise load human hair to determine if changes in the cross- to the upper and lower surface of the captured sectional area were correlated to changes in rectangular bundle and compacts it.

TABLE 3. Correlation between Silk Strand Density and the Bundle Cross-section

No. of Strands 20 40 60 80 100 120 140 160

Cross-Section (mm2) .36 .75 1.17 1.58 1.93 2.36 2.75 3.14 Cross-section divided by No. of strands .018 .019 .020 .020 .019 .020 .020 .020

Twenty-strand bundles of 5-0 suture material were measured.

34:7:JULY 2008 903 THE CROSS-SECTION TRICHOMETER

TABLE 4. Correlation between Silk Strand Weight and Bundle Cross-section

Cross- Grams Number of section Divided by Strands Grams (mm2) Cross-section

150 0.2754 2.74 0.101 148 0.2728 2.72 0.100 146 0.2705 2.71 0.100 144 0.2672 2.65 0.101 142 0.2647 2.62 0.101 140 0.2590 2.52 0.103 138 0.2565 2.48 0.103

The device is engineered to deliver no more or less Figure 3. A close-up view of the capture hook at the end of the shaft, in its extended position. than the same predetermined load each time it is engaged, regardless of the bundle size. When the between 75 to 100 1 (fine to coarse hair). This internal spring is compressed exactly 1 cm, the height value has been arbitrarily named the trichometric of the compacted bundle in its rectangular capture index (TI). chamber is displayed as millimeters on the LED screen. For an 800-hair sample (about average It should be noted that extensive testing was per- 2 for 4 cm of scalp), the normal range of values falls formed using various sized chambers, compression between 3.00 mm (for fine hair about 60 mm) to loads, and sample sizes before the optimal mechanics 4.00 mm (for coarse hair about 80 mm). The value of the device were determined. A load was chosen displayed on the screen is expressed as square that would compress the loose bundle to a point of 2 millimeters of bundle cross-section per 4 cm of scalp complete compactionFthe point beyond which no surface. When divided by 4 and multiplied by 100, further compaction would occur. When this load was the normal range of values conveniently falls determined, sample hair bundles were microscopi- cally examined to determine if fracture or visible distortion of the hair had occurred. On the basis of

Figure 2. The trichometer device is shown in its neutral position. When the shaft is pressed using the thumb, a J- shaped hook extends from the body. The travel distance of the shaft within the body is displayed as millimeters on the Figure 4. The skin is marked with a four- or eight-legged LED screen. marking template moistened with waterproof ink.

904 DERMATOLOGIC SURGERY COHEN

its cross-sectional area. A fourth test sought to establish a correlation between the observed hair loss severity and the TI (mm2 hair per cm2 scalp  100).

Diameter measurements were performed using strands of 3-0, 4-0, 5-0, and 6-0 nonsterile braided surgical silk (Havel’s Inc., Cincinnati, OH). Each material was hand-measured using an electronic micrometer (L.S. Starrett Co., Athol, MA), and strand diameters were found to vary within 15 to 20 mm along their length. The average diameters were 3-0 = 150 mm, 4-0 = 100 mm, 5-0 = 75 mm, and 6-0 = 50 mm. Thirteen mixed-size bundles, composed Figure 5. The device shown applied to an isolated bundle of of 80 strands each, were prepared. The bundles were hair from a 2 Â 2-cm area of scalp in the mid frontal area. incrementally reduced in size by mixing large caliber strands with smaller caliber strands. Each bundle this, an optimal and safe standard compression load was then measured using the device. When the re- was chosen. A working prototype was then designed sults were plotted, the cross-sectional area of each to deliver that same optimal load regardless of the bundle was reduced in a sequence that mirrored the height (or the amount) of hair in the chamber. A self- incremental reduction of suture caliber (see Table 2). braking mechanism was added to prevent the oper- ator from overtightening the threaded knob and Density measurements were performed using the exceeding the optimal load. The spring material same braided silk suture material. Bundles of 5-0 and its design were further refined to insure that silk, the diameter of which is approximately 75 mm the compression spring was in the center of its path (equivalent to the diameter of average-sized hair), from fully opened to fully compressed, i.e., in the were prepared. Eight bundles, containing 20 fila- mid range of its spring constant curve. This was done ments each, were prepared. First the device was used to minimize the imprecision that metal springs to measure the cross-sectional area of one bundle. typically display at both ends of their compression The device was then opened and a second bundle curves. was added. A second measurement was made, the device was then reopened, and a third bundle was added, etc. Bundles were incrementally placed in the device until they totaled 160 fibers. The results are Results of Testing posted in the chart below (see Table 3). The cross- The device was intended to indirectly measure den- sectional area of the bundle was increased in a se- sity and diameter by directly measuring the cross- quence that mirrored the incremental increase of sectional area of all the hairs in a premeasured area filaments added. The ratio of the number of fila- of scalp skin. Four tests were performed. The first ments to the cross-sectional diameter remained con- was designed to determine if the device could detect stant. and measure small changes in surgical silk diameter (mm). A second test was designed to determine if the Weight measurements were performed using a device could detect and measure small changes in the 150-strand bundle of 5-0 surgical filaments to de- bundle densities of hair and surgical silk (n/cm2). termine if the weight and the cross-sectional area of A third test was designed to determine if there was a a bundle were directly correlated. The bundle was correlation between the weight of the bundle and weighed on an electronic analytic balance, and its

34:7:JULY 2008 905 THE CROSS-SECTION TRICHOMETER

TABLE 5. Correlation between Hair Weight and Bundle Cross-section

Cross- Grams Number of section Divided by Hairs Grams (mm2) Cross-section

X 0.5536 2.08 0.265 X minus 2 0.5528 2.04 0.271 X minus 4 0.5519 1.98 0.278 X minus 8 0.5511 1.95 0.284 X minus 10 0.5506 1.94 0.283 X minus 12 0.5494 1.92 0.286 X minus 14 0.5480 1.89 0.289 X minus 16 0.5467 1.87 0.292 X minus 20 0.5434 1.82 0.298

Hairs were cut from a bundle of approximately 600 hairs; the bundle was reweighed, and its cross-section remeasured

cross-sectional area was measured using the new device. The measurements were performed in a lab- oratory with no humidity control. Then two fila- Figure 6. The Hair Loss Severity Scale uses the visible ratio ments were cut from the bundle, the bundle was of hair to skin to quantify the hair within a localized area of reweighed, and its cross-section was remeasured. scalp.3 (A) minimal, (B) mild, (C) moderate, and (D) severe. This was repeated six times. The cross-sectional area of the bundle was reduced in a sequence that mir- and hair, the cross-sectional areas of the bundles rored the incremental reduction of bundle weight. were reduced in a sequence that mirrored the incre- The ratio between the weight and the cross-section mental reduction of bundle weight. remained constant (see Table 4). Twelve male patients, age 23 to 67 years with Weight measurements were then performed using a balding, were examined without magnification. The bundle of cut human hair to determine if the weight observed ratio of hair to skin in the vertex area of and the cross-sectional area of a bundle were directly each patient was estimated by an observer familiar proportional (see Table 5). The bundle contained an with the Hair Loss Severity Scale (HLSS) technique. aggregate of hairs (approximately 400) collected Each patient was placed into one of four categories: from three different women, all of whom had un- minimal = much more hair than skin, mild = more dergone or permanents in the previous hair than skin, moderate = more skin than hair, and 3 months. The measurements were performed in a severe = much more skin than hair (see Figure 6). The laboratory with no humidity control. Hairs were hair within the 4  4-cm center of the vertex area incrementally cut from the bundle, the bundle was was isolated and measured with the device and ex- reweighed, and its cross-section was remeasured. pressed as TIs (mm2 hair per cm2 skin  100). The TI This was repeated eight times. The cross-sectional and hair loss severity of each patient were charted to area of the bundle was reduced in a sequence that determine if there was a correlation (see Table 6). mirrored the incremental reduction of bundle weight. The ratio between the weight and the cross- Collection of the Hair Sample section progressively increased over a period of 30 minutes. This was thought to be the result of It was clear from the onset that the sampling method ambient moisture absorption during the time re- would be quite influential in determining the sys- quired to collect the data. With both silk filaments tem’s total precision. A 2  2-cm square of hair-

906 DERMATOLOGIC SURGERY COHEN

magnifying loupes, and a sharp wooden toothpick TABLE 6. Correlation between Observed Hair Loss Severity and the TI (mm2 Hair per cm2 Scalp  100) were used. The easiest method of demarcating the in 12 Patients with Vertex Balding 2  2-cm area was marking dots with a four-legged or eight-legged template moistened on a pad of wa- Patient Severity Score TI terproof ink. Stencils with four and eight holes were 1 Minimal 47 also used to place the dots on the scalp skin surface. 2 Minimal 44 3 Minimal 44 4 Mild 36 When wet hair was combed away from the 2  2-cm 5 Mild 33 square of scalp, it remained immobilized and firmly 6 Mild 33 in place. When the sampling was performed on dry 7 Moderate 25 8 Moderate 28 hair, the peripheral hair required immobilization 9 Moderate 26 with hair clips or gummed tape. Several methods of 10 Severe 22 demarcating the area, without placing any ink marks 11 Severe 20 on the skin, have since been developed and will be 12 Severe 21 the subject of a future report. bearing scalp skin was chosen because it could be The second sampling issue to be addressed was re- easily handled without magnification and repre- turning to precisely the same site for a subsequent sented a somewhat generous sample (see Figures 4 measurement. Typically, when industry standard hair and 5). On average, this area contained about 800 counts and hair weight measurements are performed, hairs in patients with no loss. The sampling precision dots are tattooed on the scalp to identify the previous was important since fewer than 3 hairs would, for test site. Tattoos were performed on several patients, example, change the cross-sectional value of the although our goal was to design a method that bundle from 3.30 mm2 to 3.31 mm2. The collection would not require a permanent tattoo. Initially we method was not formally tested for reproducibility used a quick and easy method of simply extending a because there were simply too many different combs, nonstretchable 4-mm tape from the junction of the tapes, clips, magnifying devices, marking pens, tem- upper lip and nose columella, on to and over the tip plates, stencils, bundling contraptions, etc., to be of the nose, then up the forehead and along the evaluated. A standardized method for isolating the midline of the scalp along which the hair was cleanly hair sample from the 4-cm2 area (and immobilizing parted. An ink dot was simultaneously placed on the the adjacent hair) has not been formalized at the tape and the scalp. The tape was removed and the time of this publication. It should be pointed out distance between the two points on the tape was that the 4–cm2 site may be the shape of a triangle, measured with a ruler attached to a tabletop. How- rectangle, or hexagon and is not limited to a ever, vertical movement of nose tip caused a 5- to 7- 2  2-cm square. mm variation in the anterior to posterior dot placement. Several other issues needed to be addressed. Should the hair be wet or dry? It was easier to gather a The movement from side to side was considered less sample with precise margins on the 2  2-cm square critical when returning to the exact same site because if the hair was wet. Wet hair measurements were the gradation of hair loss in a balding individual is performed in a salon setting where each client almost much greater along the sagittal axis than the coronal always presented to the operator immediately fol- axis. Several methods and devices that enable a re- lowing a . Almost all the samples gathered turn to precisely the same area, without skin mark- in the physician’s office were performed on dry hair. ing, have since been developed and will be the In both situations, a fine-toothed comb, 2.5  subject of a future report.

34:7:JULY 2008 907 THE CROSS-SECTION TRICHOMETER

It is important to note that the sampled hair must be direct mechanical measurement of the hair and skin a minimum of 1 inch (2.5 cm) in length at time of cross-sections instead of imprecise visual determi- testing. If not, the distance between the scalp surface, nation. The HLSS method was used to determine the the hook/anvil will be too small; and the hair might correlation between hair loss severity and bundle be painfully tugged when the device is fully engaged. cross-sectional area (see Table 6 and Figure 6). Furthermore, if one anticipates newly emerging hairs, be aware that the new hairs might not be of The notion of using an instrument to measure the sufficient length for capture. If the hair is too short, cross-sectional ratio of hair to skin, for the purpose the testing should be postponed until it has grown to of measuring the quantity of hair, has been previ- adequate length. If a clinical trial is being performed, ously described. To the best of our knowledge, it was one should consider the time frame for anticipated first described in a 1936 patent by Nessler,12 who emergence of new hairs and their rate of growth. designed a rudimentary device with a rectangular Dates for retesting should be appropriately planned. slot for capturing hair and then hand-compressed the hair with a blunt, guillotine-like anvil. The height of At first we assumed that the quantity of hair in the hair in the slot was measured using ruler-like normal individuals was evenly distributed over the markings engraved on the side of the brass hand-held entire scalp. We also assumed that in women with device. The Nessler device appears to have never telogen effluvium, diffuse hair loss, or sheddingF gained popularity and no references could be found the loss was evenly distributed over the entire surface in the medical literature. of the scalp. During the pilot studies, it was clear that neither of these assumptions were correct. In 2001, Arnold13 formally presented a method for A significant number of patients with no complaints measuring the quantity of hair in a premeasured area of hair loss had values higher on the top of the head of scalp. Although the work was not published, Ar- than in the occipital region. The same unexpected nold deserves full credit for introducing the concept distribution pattern was seen in women with com- of measuring hair quantity using hair/skin cross- plaints of excessive shedding. Further search of the sectional ratio to the hair science community. Ar- literature revealed that the hair density changes nold’s work served as the inspiration for the method/ dramatically with age and that density is in fact device described in this report. unequally distributed over the scalp, often highest on the top of the head.10,11 These observations are sig- Arnold isolated the hair from a premeasured area of nificant if one compares the occipital and midscalp scalp, but he chose to measure the hair bundle using values when attempting to distinguish between a thread wrapped snugly around the bundle’s pe- diffuse and pattern loss in women with complaints of riphery. An ink mark was made on the circumfe- hair loss. rentially applied loop of thread at the point where the strand crossed over itself. The thread was then removed and stretched out, and the distance between Conclusions and Discussion the two marks was measured. Arnold had measured The method/device described in this report is a me- the circumference of the bundle and called this value chanical refinement of the author’s previous pub- the hair mass index. Neidel and Bretschneider14 have lished HLSS.3 When using the HLSS, the observer is described and published the details of Arnold’s hair asked to determine the ratio of grossly visible hair to mass measurement technique. grossly visible skin. A series of photographs give examples of the categories to be chosen. The Nessler and Arnold did not standardize the load trichometer device/method described in this report applied to the bundle or control its application with likewise compares the ratio of hair to skin, but uses a mechanical apparatus. Because the bundle of hair

908 DERMATOLOGIC SURGERY COHEN

is soft and quite compressible, the variability of their TABLE 7. Possible Applications of the Trich- load and their method of application introduced ometer Technology significant imprecision. Nessler’s ruler and Arnold’s 1. Quantify hair mass in a localized area of measuring thread technique were significantly im- thinning and/or shedding precise as well. The methods of Nessler, Arnold, 2. Quantify the efficacy of proven and unproven Hamilton-Norwood, Ludwig, and Cohen (HLSS) are hair growth products and devices 3. Quantify the medical treatment response in all imprecise and not suitable for scientific studies, thyroid disease, iron deficiency, etc. and although hair weight measurement, global pho- 4. Quantify the results of hair transplantation tography, and hair counts are precise, and suitable surgery for scientific studies, they too have the following 5. Quantify the donor hair available for hair transplantation surgery15 limitations as mentioned in Table 1. 6. Detect the reversal and recovery from post partum effluvium Global photography requires special equipment and 7. Detect balding before it is visible 8. Determine a new drug’s potential for causing hairstyle conformity. It is designed to compare the hair loss relative difference between the before and after ap- 9. Determine the incidence of hair loss caused by pearance of a single patient. It does not generate a FDA-approved drugs single quantitative value for a localized area of the 10. Provide a simple 1–100 scoring system for hair loss scalp. Hair counts on the other hand do generate a 11. Provide a new measuring modality for hair single quantitative value but the value does not re- science research flect the wide variation of diameters seen in condi- 12. Enable the office physician to track and mea- tions of thinning, i.e., androgenetic alopecia. Hair sure a patient’s clinical status 13. Improve and simplify communication between weight measurement, the gold standard, is simply hair professionals too difficult and time-consuming to perform as an office procedure, and hair weight, hair counts, and FDA, Food and Drug Administration. Trichoscan all require that hair be cut. The tricho- meter technology overcomes many of these limita- same area for retesting without using a tattoo com- tions and generates a value that simultaneously promises the measurements as well. Both of these reflects the influence of density and diameter alone. issues will be the subject of a subsequent report.

Preliminary study results, using both silk fibers and The general availability of a simple hair-measuring hair, were the same. The incremental changes in the technology introduces a number of possibilities (see filament number, filament diameter, and bundle Table 7). Any clinical condition characterized by weight were reflected as equal and proportionate in- shedding and/or thinning could be informally quan- cremental changes in the bundle cross-sectional area. tified and tracked. A patient’s hair growth response It was concluded that the device could be used as a to minoxidil, finasteride, and iron supplement could reliable substitute for every instrument and method be easily measured. The efficacy of popular modal- that is presently used to measure the parameters of ities like low-intensity laser, biotin, and saw pal- hair loss and growth, including the dry hair weight metto could be informally determined by practicing measurementFthe industry gold standard. physicians. Unsubstantiated anecdotes could be challenged, and hair growth scams revealed. Although the device itself showed a high degree of precision, it should be emphasized that the accuracy Hundreds of common drugs, prescription and over and reproducibility of the measurements can be the counter, are known to cause hair loss. These compromised if the sampling method is not carefully include retinoids, anticoagulants, cholesterol-lower- performed using magnification. Returning to the ing agents, anticonvulsants, antidepressants, gastric

34:7:JULY 2008 909 THE CROSS-SECTION TRICHOMETER

acidity suppressants, cardiac arrhythmia and anti- 5. Price VH, Menefee E, Strauss PC. Changes in hair weight and hypertensive agents, anti-inflammatory agents, hor- hair count in men with androgenetic alopecia, after application of 5% and 2% topical minoxidil, placebo, or no treatment. mones, and weight reduction drugs . . . plus the J Am Acad Dermatol 1999;11:41.

entire category of antineoplastic agents. Patients 6. Van Neste D, Dumrotier M, De Coster W. Phototrichogram taking these drugs could be evaluated to determine analysis: technical aspects and problems in relation with automated quantitative evaluation of hair growth by computer- the incidence and magnitude of their hair loss. New assisted image analysis. In: Van Neste D, Lachapelle JM, Antoine drugs could be screened, before FDA approval, to JL, editors. Trends in and alopecia research. Dordrecht: Kluwer (Amsterdam); 1989. p. 155–65. determine if they have the potential side effect of causing hair loss. 7. Hoffman R. TrichoScan: combining epiluminescence microscopy with digital image analysis for the measurement of hair growth. Eur J Dermatol 2001;11:362–8.

Marritt observed that a man must lose 50% of his 8. Headington J. Transverse microscopic anatomy of the human hair mass before the loss can be seen with the naked scalp: a basis for a morphometric approach to disorders of the . Arch Dermatol 1984;120:449–56. eye.16 This was confirmed in Table 6. The TI’s of 9. Barth JH, Rushton DH. Measurement of hair growth. In: Serud J, men with minimal hair loss were about 50% lower Jemec G, editors. Non-invasive methods and the skin. Ann Arbor: than the TI’s of men with no loss at all. (Normal CRP Press; 1995. p. 543–8. range for TI is 75 to 100 plus.) Logically, the 10. Olsen EA, Canfield D. Age-related changes in scalp hair density. trichometer might be used to identify men in very Tokyo: European Hair Research Society; 2001. early stages of balding, when diameter reduction 11. Van Neste D. Female patients complaining about hair loss: documentation of defective scalp hair dynamics with contrast- silently precedes visible loss. By measuring and enhanced phototrichogram. Skin Res Technol 2006;5:83–8.

comparing the frontal and occipital regions of men 12. Nessler C. Means for ascertaining the hair production of a with normal-appearing hair, a loss as small as 5 or subject. US Patent 1,962,518, United States Patent Office, June 10% could be detected y perhaps 10 or 15 years 12, 1934. before balding was actually visible. The speed of 13. Arnold J. Hair mass index, 4th Annual Congress European Society of Hair Restoration Surgery. Barcelona, 2001. progression and response to treatment could be 14. Neidel FG, Bretschneider P. Measuring hair mass. In: Unger W, easily monitored. Shapiro R, editors. Hair Transplantation. New York: Marcel Dekker; 2004. p. 876.

15. Parsley W. Donor site measurement. In: Haber R, Stough D, References editors. Hair Transplantation. Philadelphia: Elsevier; 2006. p. 101. 1. Hamilton JB. Patterned loss of hair in men; types and incidence. Ann NY Acad Sci 1951;53:708–28. 16. Marritt E. The death of the density debate. Dermatol Surg 1999;5:654–60. 2. Norwood OT. Male pattern baldness: classification and incidence. South Med J 1975;68:1359–65.

3. Cohen BH. Hair loss profile, index, and severity scale. In: Haber R, Stough D, editors. Hair Transplantation. Philadelphia: Elsevier; Address correspondence and reprint requests to: Bernard 2006. p. 12. Cohen, MD, 4425 Ponce de Leon Boulevard, Suite 230, 4. Canfield D. Photographic documentation of hair growth in Coral Gables, FL 33145, or e-mail: thehairlosscenter@ androgenetic alopecia. Dermatol Clin 1996;14:713–21. mac.com

COMMENTARY

The current standards for measuring hair loss and hair growth are laden with problems and inefficiencies. The ongoing search for a simple and accurate method of measuring hair quantity has finally found respite in a novel device recently christened the cross-section trichometer. This device, developed by Dr. Bernie Cohen, is featured in this issue of Dermatologic Surgery. As Dr. Cohen has quoted many times, ‘‘Medicine is a language of numbers. Simple numbers are used to make a diagnosis like hypertension, diabetes, fever,

910 DERMATOLOGIC SURGERY COHEN

and obesity. In fact, it’s the manipulation of these numbers that determines the manner by which we treat these disorders. In simple terms, if it can’t be measured, it can’t be managed.’’ The challenge when evaluating all hair growth drugs is obtaining consistent methodology for measure- ment. In the past, global photography has been utilized to document the overall change in the appearance of hair from a baseline setting. Using serial global photographs, subjects were classified into distinct categories, i.e., (1) greatly decreased in appearance, (2) slight decrease in appearance, (3) no change in appearance, (4) slight increase in appearance, (5) moderate increase in appearance, and (6) great increase in appearance. Unfortunately, the results are influenced by the F-stop settings on the camera.

Additionally, the lighting, film quality, grooming practices, and length of hair must be kept identical to the original baseline photography or an erroneous change of appearance results. Keeping all of the above factors constant is daunting and often not feasible. Global photography is a very crude and often in- accurate method of assessing hair quantity in terms of both loss and growth. Global photography cannot properly reflect changes in hair counts. To assess changes in hair counts, microphotographic techniques are utilized. These techniques involve computer overlay in comparison to baseline photographs. The addition of or loss of hairs is determined by a numerical value generated by the computer when comparing photographs. This seems to be a more precise method, but fails to account for the changes in hair shaft diameter, which often show the visible results. Since the positive effect of minoxidil and 5a-reductase inhibitors (finasteride and dutasteride) use is in part due to changes in hair shaft; the hair counts may be minimally affected. In other words, a positive effect may be observed in a study patient by global photography with no change over baseline in actual hair count. It is obvious from the above that our current methods are inadequate in providing a precise change in hair mass. Dr. Cohen’s device offers a solution. When it is applied to a bundle of hair, any change in density and diameter will be evident and measured numerically.

The article presented herein by Dr. Cohen is well written and deserves the attention of those involved in hair research, clinical evaluations of hair disorders, and practicing hair transplant surgeons. It is not unreasonable to project that at some point in the future residents in dermatology will utilize a device such as the cross-section trichometer to routinely evaluate the success or failure of hair loss treatments in their clinical patients. Practitioners look forward to this device being available to use on their patients. It will be important for clinicians to produce the same, reproducible accurate results as presented in this article. Dr. Cohen is to be applauded for both his success in bringing this to the field and his contributions in advancing hair research.

DOW STOUGH,MD Hot Springs, AR

34:7:JULY 2008 911