Portal de Periódicos da CAPES

Hair Growth Cycle Affects Hair Follicle Destruction by Ruby Laser Pulses

1998; Elsevier BV; Volume: 111; Issue: 1 Linguagem: Inglês

10.1046/j.1523-1747.1998.00227.x

1523-1747

Tai-Yuan David Lin, Woraphong Manuskiatti, Christine Dierickx, William A. Farinelli, Marnie E. Fisher, Thomas J. Flotte, Howard P. Baden, R. R. Anderson,

Skin Protection and Aging

It has been shown that normal mode ruby laser pulses (694 nm) are effective in selectively destroying brown or black pigmented hair follicles in adult Caucasians. This study investigated how the various stages of the hair follicle growth cycle influence follicle destruction by ruby laser treatment, using a model of predictable synchronous hair growth cycles in the infantile and adolescent mice. A range of ruby laser pulse fluences was delivered during different stages of the hair growth cycle, followed by histologic and gross observations of the injury and regrowth of hair. Actively growing and pigmented anagen stage hair follicles were sensitive to hair removal by normal mode ruby laser exposure, whereas catagen and telogen stage hair follicles were resistant to laser irradiation. Selective thermal injury to follicles was observed histologically, and hair regrowth was fluence dependent. In animals exposed during anagen, intermediate fluences induced nonscarring alopecia, whereas high fluences induced scarring alopecia. The findings of this study suggest treatment strategies for optimal laser hair removal. It has been shown that normal mode ruby laser pulses (694 nm) are effective in selectively destroying brown or black pigmented hair follicles in adult Caucasians. This study investigated how the various stages of the hair follicle growth cycle influence follicle destruction by ruby laser treatment, using a model of predictable synchronous hair growth cycles in the infantile and adolescent mice. A range of ruby laser pulse fluences was delivered during different stages of the hair growth cycle, followed by histologic and gross observations of the injury and regrowth of hair. Actively growing and pigmented anagen stage hair follicles were sensitive to hair removal by normal mode ruby laser exposure, whereas catagen and telogen stage hair follicles were resistant to laser irradiation. Selective thermal injury to follicles was observed histologically, and hair regrowth was fluence dependent. In animals exposed during anagen, intermediate fluences induced nonscarring alopecia, whereas high fluences induced scarring alopecia. The findings of this study suggest treatment strategies for optimal laser hair removal. Unwanted excess hair is a major patient concern that can be due to genetics, systemic disease, or drug reactions. The most popular methods of hair removal are temporary (Kvedar et al., 1995Kvedar J.C. Gibson M. Krusinski P.A. Hirsutism: evaluating treatment.J Am Acad Dermatol. 1995; 12: 215-225Abstract Full Text PDF Scopus (28) Google Scholar). In individuals with coarse curly hair, shaving can induce painful "ingrown" hairs and scarring, known as pseudofolliculitis barbae. Electrolysis (galvanic current) or electrothermolysis (RF current) can provide permanent hair removal, but these painful, tedious techniques are dependent on technical skill and require multiple treatments. Complications including folliculitis, scarring, and infection can occur after electrolysis. New methods of permanent hair removal are needed. Normal mode (long-pulsewidth) ruby laser pulses are effective for selectively destroying pigmented hair follicles in adult Caucasians (Grossman et al., 1996Grossman M.C. Dierickx C.C. Farinelli W.A. Flotte T. Anderson R.R. Damage of hair follicles by normal-mode ruby laser pulses.J Am Acad Dermatol. 1996; 35: 889-894Abstract Full Text PDF PubMed Scopus (353) Google Scholar), based on principles of selective photothermolysis (Anderson and Parrish, 1983Anderson R.R. Parrish J.A. Selective photothermolysis: Precise microsurgery by selective absorption of pulsed radiation.Science. 1983; 220: 524-527Crossref PubMed Scopus (2747) Google Scholar). Cutaneous melanin is normally present in the epidermis and follicles, but not in the dermis. Grossman et al., 1996Grossman M.C. Dierickx C.C. Farinelli W.A. Flotte T. Anderson R.R. Damage of hair follicles by normal-mode ruby laser pulses.J Am Acad Dermatol. 1996; 35: 889-894Abstract Full Text PDF PubMed Scopus (353) Google Scholar used ruby laser pulses delivered through a cold sapphire lens to minimize epidermal injury, by thermal conduction from the skin surface. In individuals with fair skin and dark hair, high-fluence ruby laser pulses induced both a growth delay and prolonged hair loss; permanent hair loss was recently reported in this same group of subjects. Melanin within the follicles acts as a target for light absorption, over a broad range of wavelengths (Anderson and Parrish, 1981Anderson R.R. Parrish J.A. The optics of human skin.J Invest Dermatol. 1981; 77: 13-19Abstract Full Text PDF PubMed Scopus (2074) Google Scholar). The spectral region with greatest selective absorption by eumelanin, and excellent optical penetration into the dermis, is near 700 nm. The influence of hair growth cycle on sensitivity of follicles to selective photothermolysis is unknown. This study uses a well-known animal model, the pigmented hairy mouse C57/BL6 (Chase et al., 1951Chase H.B. Rauch H. Smith V.W. Critical stages of hair development and pigmentation in the mouse.Physiol Zool. 1951; 24: 1-8Crossref PubMed Scopus (121) Google Scholar;Paus et al., 1990Paus R. Stenn K.S. Link R.E. Telogen skin contains an inhibitor of hair growth.Br J Dermatol. 1990; 122: 777-784Crossref PubMed Scopus (220) Google Scholar), to study the effects of ruby laser exposure on follicles at various stages of the hair growth cycle. The model was chosen for several reasons. This strain of mouse has nonpigmented skin and darkly pigmented hair follicles, which optimize target selection. In addition, the fur coat of the infantile and adolescent mice is in a predicable synchronous growth cycle for periods of about 24 d. Two full synchronous growth cycles are completed before the fur begins to grow asynchronously. Synchronous hair regrowth enabled us to compare the effects of laser treatment at specific growth stages. Hair follicles undergo profound anatomic and metabolic changes between growth stages that should affect sensitivity to selective photothermolysis. In anagen, the actively growing matrix is heavily pigmented, and the papillae are deep within the dermis and/or fat. During catagen, the lower third of the follicle degenerates by apoptosis (Seiberg et al., 1995Seiberg M. Marthinuss J. Stenn K.S. Changes in expression of apoptosis-associated genes in skin mark early catagen.J Invest Dermatol. 1995; 104: 78-82Crossref PubMed Scopus (91) Google Scholar), including the melanocytes responsible for pigmentation of the hair shaft. In telogen, the nongrowing hair shaft is anchored in a nonpigmented, resting follicle. At the onset of anagen, stem cells from a region near the insertion of the arrector pili (Costsarelis et al., 1990Costsarelis G. Sun T.T. Lavker R.M. Label-retaining cells reside in the bulge area of pilosebaceous unit: implications for follicular stem cells, hair cycle and skin carcinogenesis.Cell. 1990; 61: 1329-1337Abstract Full Text PDF PubMed Scopus (1898) Google Scholar) differentiate to form a new matrix that begins to produce a new hair shaft. At this time, the matrix and papilla are relatively high in the dermis, and the pigmentary system is again stimulated. We therefore hypothesized that early anagen would be the growth stage of greatest sensitivity to laser pulses because of the shallow location of a well-pigmented follicle. In this study, we also examined the effect of laser pulse fluence and number of pulses on destruction of hair follicles. Nominally 2 ms pulses at 694 nm wavelength were delivered by an experimental normal-mode ruby laser (NMRL, Spectrum Medical, Lexington, MA) at up to 0.5 Hz. Glass flats in the laser cavity were used to obtain three different fluences: 1.47 J per cm2 (1.7 ms pulsewidth, exposure spot diameter 1.0 cm), 2.29 J per cm2 (1.8 ms pulsewidth, exposure spot diameter 1.1 cm), and 3.16 J per cm2 (2.2 ms pulsewidth, 1.3 cm exposure spot diameter). Beam diameters were measured by translating a pin hole across the beam, and represent 1/e2 of maximum values. The beam was steered via a series of mirrors into an actively cooled delivery device placed directly and firmly against the skin, described previously (Grossman et al., 1996Grossman M.C. Dierickx C.C. Farinelli W.A. Flotte T. Anderson R.R. Damage of hair follicles by normal-mode ruby laser pulses.J Am Acad Dermatol. 1996; 35: 889-894Abstract Full Text PDF PubMed Scopus (353) Google Scholar). A sapphire lens was used to provide a convergent beam at the skin surface and to increase beam coupling into the skin. The lens was water cooled to 7°C to provide heat conduction from the epidermis before, during, and after each laser pulse. Delivered energy into air was measured with a laser energy meter (Nova, Ophire Optronics, Israel). The C57/BL6 mice were obtained from Charles River Laboratories (Wilmington, MA) and maintained in a conventional animal facility of Massachusetts General Hospital. Laser irradiation was performed in infantile (birth to 21 d) or adolescent (4–6 wk old) mice. The mice were received at the age of 19 d. Their fur was clipped and depilated with NEET cream to reveal actual skin color. They were watched every day to determine the stage of hair growth cycle as described in detail in the Discussion. Briefly, mice are assigned to be in day 4 of anagen when the skin began turning gray, and day 8 when the growing hair began to emerge from the skin surface. Mice were randomly divided into groups of two or more and exposed to laser on a predetermined day throughout the 24 d hair growth cycle according to the schedule in Table 1. The control group consisted of mice that did not receive laser treatment. Before the treatment, each animal's fur was clipped and depilated with NEET to remove excess hair, if necessary. Immediately post-treatment, initial biopsies were obtained from the exposure sites plus unirradiated control. The biopsies were processed in paraffin, vertically sectioned, stained with hematoxylin and eosin, and evaluated by a dermatopathologist. The histologic sections were scored for epidermal and dermal damage (absent, present) and follicular damage (none, mild, moderate, severe). Mild follicular damage was defined as thermal coagulation confined in the matrix and follicular epithelium was relatively intact. Follicular damage was scored as moderate when the follicular epithelium was damaged and the bulb was irregular in shape. Severe follicular damage occurred when individual hair follicles were no longer discernible and associated with dermal and epidermal injury as well. The exposure sites were also subjectively graded for the following immediate responses: erythema (none, mild, moderate, severe), edema (none, mild, moderate, severe), and petechiae (absent, present). Animals were watched every day for signs of discomfort or abnormal behavior.Table ITreatment scheduleNumber of animals usedStage of hair growth (day of cycle)SiteFluence (J per cm2)Number of pulsesNumber of treatments2Anagen (4)11.471122.291133.161143.16214Anagen (6)1-4same as above5Anagen (8)1-4same as above4Anagen (10)1-4same as above6Anagen (12)1-4same as above6Anagen (14)1-4same as above7Anagen (16)1-4same as above4Catagen (18)1-4same as above4Telogen (20)1-4same as above5Telogen (22)1-4same as above4Telogen (24)1-4same as above2Control000 Open table in a new tab Evaluation was again done on day 2 and day 8 post-treatment on all four exposure sites for erythema (none, mild, moderate, severe), edema (none, mild, moderate, severe), and ulceration (absent, present). On day 28 and day 56 post-treatment, the exposure sites were graded for scarring (absent, present) and hair regrowth (none, sparse, moderate, full). Scarring was defined as the lack of skin creases passing through the treatment site. Hair regrowth was subjectively graded as: none (0% regrowth), sparse (0–33%), moderate (33–66%), or full (66–100%). Sites were biopsied if regrowth was less than full. The biopsied specimens were graded for fibrosis (absent, present) characterized by hypercellularity and increased density of collagen matrix. Treated sites for which regrowth was full, for which there was no scarring, and for which the hair had entered telogen, were wax-epilated to induce another growth cycle. Animals were followed for at least 4 mo post-treatment before euthanasia. Anesthesia was achieved with metofane during biopsies, evaluations, and fur clipping. Nonparametric statistical analyses (Glantz, 1992Glantz S.A. Primer of Biostatistics. McGraw-Hill, New York1992Google Scholar) were used. Spearman rank correlation coefficients were calculated for comparisons between histologic evidence (e.g., immediate follicular damage) and clinical observation (e.g., immediate erythema). Friedman's test was carried out to compare regrowth within groups treated during telogen and to compare regrowth within groups treated during anagen, among different fluences tested. The effect of double pulsing of 3.16 J per cm2 was examined by calculating Student-Newman-Keuls test statistic after Friedman's test. Mann–Whitney rank sum test was performed to compare regrowth within each of the fluences tested between exposure to telogen growth stage, and exposure to anagen growth stage. For data representation, parametric statistical analyses were performed for follicular damage (0, no damage; 1, mild; 2, moderate; 3, severe) and hair regrowth (0, no regrowth; 1, sparse; 2, moderate; 3, full) to obtain means and SD. Laser exposure selectively caused damage to the follicular epithelium seen histologically as increased eosinophilia and nuclear elongation without damage to the epidermis and dermis (see Figure 1). Figure 2 shows follicular damage immediately post-treatment for the group treated in different stages of hair growth cycle, using fluence levels of 1.47 J per cm2×one pulse, 2.29 J per cm2×one pulse, 3.16 J per cm2×one pulse, and 3.16 J per cm2×two pulses. For each fluence, anagen follicles showed more damage than catagen or telogen follicles.Figure 2Follicular damage immediately after laser exposure is absent during catagen or telogen, and fluence dependent during anagen. 0, No damage; 1, mild; 2, moderate; 3, severe; A, anagen; C, catagen; T, telogen. The number in parentheses indicates the sample size. Error bars: SD.View Large Image Figure ViewerDownload (PPT) In the groups treated during anagen, there was a heterogeneous but widespread injury to the follicular epithelium that extended along its entire length, increasing with increasing fluence (see Figure 1). In contrast, no follicular damage was observed in catagen or telogen stage follicles at any of the fluences used. Similarly, in the group treated in the anagen stage of hair growth, exposure sites became erythematous and edematous, especially at 3.16 J per cm2. The degree of erythema was again directly related to fluence. In contrast, mice treated in the catagen or telogen stage showed no erythema or edema at any of the fluence levels. Erythema and edema resolved within 3 d after treatment. Petechiae and ulceration were never observed. In addition, erythema immediately post-treatment was found to correlate with follicular damage by histology within groups treated at the same age at different fluence levels. The correlation was statistically significant, with the Spearman rank correlation coefficient of +0.731 (n = 44; p < 0.05). At the highest fluence of 3.16 J per cm2, in addition to follicular damage there was thermal injury to the epidermis and dermis, especially for animals treated on days 8–14 of the anagen stage. Figure 3 illustrates such injury on day 14 of anagen immediately after laser exposure. Table 2 summarizes epidermal and dermal damage for groups treated at different stages of hair growth cycle and with 3.16 J per cm2.Table IIAnalysis o epidermal and dermal injury after laser exposureAnagenCatagenTelogenJ per cmDay 4Day 6Day 8Day 10Day 12Day 14Day 16Day 18Day 20Day 22Day 243.16X10a0, absent.0+ b+, present.+++000003.16X20+++00000a 0, absent.b +, present. Open table in a new tab For each fluence, treatment during anagen caused consistently less hair regrowth compared with laser exposures during catagen or telogen. This difference was striking – full hair regrowth occurred after laser exposures during catagen or telogen, for all fluence levels. Hair regrowth 28 d after laser treatments performed during anagen, was strongly fluence dependent (n = 43;χr = 137.084; p < 0.05 by Friedman's test), with 3.16 J per cm2 causing less hair regrowth than 1.47 J per cm2. Hair regrowth 56 d after laser treatments performed during anagen was also strongly fluence dependent (n = 27, χr = 71.304; p < 0.05 by Friedman's test). At 1.47 J per cm2, at either 28 d (Figure 4) or 56 d (Figure 5) post-treatment, there was moderate to full regrowth, but still significantly less than in the groups treated in catagen or telogen (28 d post-treatment, n for anagen = 43, n for catagen and telogen = 19, ZT = 3.368, p < 0.05; 56 d post-treatment, n for anagen = 27, n for catagen and telogen = 8, ZT = 1.207, p < 0.05 by Mann–Whitney rank sum test). When treated during anagen at 2.29 J per cm2 (Figures 4,5), hair regrowth was sparse to moderate (28 d post-treatment, ZT = 5.932, p < 0.05; 56 d post-treatment, ZT = 3.547, p < 0.05). At 3.16 J per cm2×one pulse (Figures 4,5), regrowth was none to sparse (28 d post-treatment, ZT = 6.526, p < 0.05; 56 d post-treatment, ZT = 4.089, p < 0.05). At 3.16 J per cm2×two pulses (Figures 4,5), regrowth was again none to sparse (28 d post-treatment, ZT = 6.580, p < 0.05; 56 d post-treatment, ZT = 4.068, p < 0.05). Regrowth at 56 d post-treatment was scored relatively higher than regrowth at 28 d.Figure 5Full hair regrowth occurs 56 d after laser exposure given during catagen or telogen, whereas regrowth after exposure given in anagen is fluence dependent. 0, No regrowth; 1, sparse; 2, moderate; 3, full; A, anagen; C, catagen; T, telogen. The number in parentheses indicates the sample size. Error bars: SD.View Large Image Figure ViewerDownload (PPT) Histology of skin biopsies 28 d after treatment at the highest fluence during anagen, showed normal epidermis with dermal fibrosis replacing degenerated hair follicles. In contrast, mice treated during telogen or catagen had normal follicles without fibrosis (Table 3). Fibrosis was not seen in mice treated during anagen at the two lower fluences, even in animals with minimal hair regrowth. Figure 6 shows fibrosis in a treatment site, compared with adjacent unexposed skin, 28 d post-treatment given on anagen day 6 with 3.16 J per cm2×one pulse. Macrophages loaded with melanin granules were scattered among the fibrotic tissue, characterized by dense, linear collagen and hypercellularity. Grossly, mice treated at the highest fluence during anagen showed scarring 28 d after exposure (Table 4). There was generally good correlation between scarring observed grossly, and fibrosis observed histologically. Scarring or fibrosis only occurred after treatments during anagen at 3.16 J per cm2×one or two pulses. No scarring or fibrosis was observed at fluences less than 3.16 J per cm2, even when there was little or no hair regrowth. Note that fibrosis and scarring present 28 d after exposure to the highest fluence during anagen, resolved in many animals by 56 d after exposure (Table 3, 4). There was no discernible difference in skin response or histology, between single and double pulses at 3.16 J per cm2.Table IIIFibrosis by histology after laser exposure is absent during catagen or telogen, and fluence dependent during anagenAnagenCatagenTelogenJ per cm2Day 4Day 6Day 8Day 10Day 12Day 14Day 16Day 18Day 20Day 22Day 2428 d1.47X10a0, absent.00000000002.29X1000000000003.16X1+b+ , present.++++++00003.16X2+++++++000056 d1.47X1000000000002.29X1000000000003.16X1++++++000003.16X2++++++00000a 0, absent.b + , present. Open table in a new tab Table IVScarring by gross observation after laser exposure is absent during catagen or telogen, and fluence dependent during anagenAnagenCatagenTelogenJ per cm2Day 4Day 6Day 8Day 10Day 12Day 14Day 16Day 18Day 20Day 22Day 2428 d1.47X10a0, absent.00000000002.29X1000000000003.16X10+0+++000003.16X2+b+ , present.+++++0000056 d1.47X1000000000002.29X1000000000003.16X10+00++000003.16X20++0++00000a 0, absent.b + , present. Open table in a new tab This study shows that anagen-phase hair follicles are particularly sensitive to damage from ruby laser pulses, presumably because anagen-phase melanin pigmentation is needed to provide the chromophore for selective photothermolysis. The study also demonstrates that alopecia can be induced without fibrosis or scarring in mice, after fluences up to about 3 J per cm2. At higher fluences, scarring alopecia occurs. Despite differences between human and mouse hair follicles, the findings have important clinical implications for hair removal. It should be noted that the hair growth cycle is a continuous process; setting up certain "stages" is entirely arbitrary.Dry, 1926Dry F.W. The coat of the mouse (Mus musculus).J Genet. 1926; 16: 287-340Crossref Scopus (242) Google Scholar divided the cycle into three stages: anagen, the period of activity; catagen, the period of regression; and telogen, the period of quiescence.Chase et al., 1951Chase H.B. Rauch H. Smith V.W. Critical stages of hair development and pigmentation in the mouse.Physiol Zool. 1951; 24: 1-8Crossref PubMed Scopus (121) Google Scholar further divided anagen into six substages (I–VI) in mice. The strain of mice in this study are born without hair, and with nonpigmented skin. As the hair follicle buds enter anagen (days 1–17 of the cycle), melanin is produced and incorporated into the matrix. During the first 3 d (anagen I–II) of anagen stage, no melanin production is apparent and the skin remains pink. At this time, the base of the follicle is characteristically about 250 μm below the skin surface, or at the "resting" level (Chase et al., 1951Chase H.B. Rauch H. Smith V.W. Critical stages of hair development and pigmentation in the mouse.Physiol Zool. 1951; 24: 1-8Crossref PubMed Scopus (121) Google Scholar). In anagen II (days 3–4 of the cycle), the follicle attains its maximum depth, typically ≥500 μm below the skin surface, and usually does not go any deeper (Chase et al., 1951Chase H.B. Rauch H. Smith V.W. Critical stages of hair development and pigmentation in the mouse.Physiol Zool. 1951; 24: 1-8Crossref PubMed Scopus (121) Google Scholar;Chase, 1954Chase H.B. Growth of the hair.Physiol Rev. 1954; 34: 113-126Crossref PubMed Scopus (362) Google Scholar). Melanocytes appear along the papilla cavity, producing and transferring melanin granules to cells in the matrix, and the skin starts to turn gray (Chase, 1954Chase H.B. Growth of the hair.Physiol Rev. 1954; 34: 113-126Crossref PubMed Scopus (362) Google Scholar). By anagen V (day 8 of the cycle), the tip of the hair has broken through the tip of the internal sheath and has grown to about the level of the epidermis (Chase, 1954Chase H.B. Growth of the hair.Physiol Rev. 1954; 34: 113-126Crossref PubMed Scopus (362) Google Scholar). The activity of melanogenesis reaches a peak on days 8–12 (anagen V–VI) (Slominski et al., 1991Slominski A. Paus R. Costantino R. Differential expression and activity of melanogenesis-related proteins during induced hair growth in mice.J Invest Dermatol. 1991; 96: 172-179Abstract Full Text PDF PubMed Google Scholar). Now the skin appears black. Thereafter, the skin begins to turn pink as melanogenesis decreases, whereas the depth of the hair follicle remains at the maximum. In anagen, the hair shaft is fully pigmented for its entire length, which corresponds to our finding of selective laser thermal damage along the entire length of anagen follicles. When the active growth ends, dramatic changes occur in the hair follicle throughout catagen (Straile et al., 1961Straile W.E. Chase H.B. Arsenault C. Growth and differentiation of hair follicles between periods of activity and quiescence.J Exp Zool. 1961; 148: 205-221Crossref PubMed Scopus (100) Google Scholar). Catagen is a relatively brief transitional stage (days 18–19 of the cycle). Melanin production and hair growth cease abruptly (Chase, 1954Chase H.B. Growth of the hair.Physiol Rev. 1954; 34: 113-126Crossref PubMed Scopus (362) Google Scholar). The hair shaft is completely depigmented at least up to the level of the skin surface, hence our finding of no selective laser thermal damage along the entire length of catagen or telogen follicles. Meanwhile, the lower hair follicle undergoes apoptosis and its base moves gradually upward to the resting position. As the hair follicle moves into telogen (days 20–24 of the cycle), hair growth and melanin production remain completely absent (Chase et al., 1951Chase H.B. Rauch H. Smith V.W. Critical stages of hair development and pigmentation in the mouse.Physiol Zool. 1951; 24: 1-8Crossref PubMed Scopus (121) Google Scholar) and stem cells remain largely quiescent. It is currently believed that follicular stem cells exist in the "bulge" area near the origin of the arrector pili muscle, and give rise to follicular matrix cells and melanocytes that migrate to the bulb of a newly growing hair bud (Costsarelis et al., 1990Costsarelis G. Sun T.T. Lavker R.M. Label-retaining cells reside in the bulge area of pilosebaceous unit: implications for follicular stem cells, hair cycle and skin carcinogenesis.Cell. 1990; 61: 1329-1337Abstract Full Text PDF PubMed Scopus (1898) Google Scholar). In the late telogen stage, the dermal papilla is at the level of the bulge region, and the stem cells proliferate rapidly. This activity initiates the anagen phase again, with an upward-mobile hair and a downward-mobile follicle. During anagen, the stem cells of the bulge area remain quiescent. Selective photothermolysis relies on preferential absorption of a pulse of radiant energy delivered to target structures. In theory, the optimal pulse duration for selective photothermolysis is less than or approximately equal to the thermal relaxation time of the target structure; however, the biologically important target structures for inactivation of hair growth have not been fully defined. This study shows that anagen follicles are far more sensitive, which strongly suggests that pigmentation and/or active growth are important determinants of photothermal damage to hair follicles. Thermal relaxation time (tr) is given approximately by the equation tr = d2/gκ where d is a target dimension, κ is thermal diffusivity (≈2×10–3 cm2 per s), and g is a geometric factor (Van Gemert and Welch, 1989Van Gemert M.J.C. Welch A.J. Time constants in thermal laser medicine.Lasers Surg Med. 1989; 9: 405-421Crossref PubMed Scopus (173) Google Scholar). Thermal relaxation in human port-wine stain vessels in vivo shows excellent correlation between thermal transfer theory and biologic endpoints (Dierickx et al., 1995Dierickx C.C. Casparian J.M. Venugopalan V. Farinelli W.A. Anderson R.R. Thermal relaxation of port-wine stain vessels probed in vivo: the need for 1–10-millisecond laser pulse treatment.J Invest Dermatol. 1995; 105: 709-714Crossref PubMed Scopus (236) Google Scholar); presumably the same is true for hair follicles, which are cylindrical dermal structures in the same size range as port-wine stain vessels. The thermal relaxation time in seconds is approximately the square of the target diameter in millimeters. Although the mechanisms of follicular injury are thermal, the relative importance of thermal denaturation versus vaporization and/or mechanical damage from cavitation remain uncertain. Biopsy specimens obtained immediately after laser exposures (Figure 1) showed both thermal coagulation and asymmetric focal ruptures of the follicular epithelium, similar to that previously described in humans (Grossman et al., 1996Grossman M.C. Dierickx C.C. Farinelli W.A. Flotte T. Anderson R.R. Damage of hair follicles by normal-mode ruby laser pulses.J Am Acad Dermatol. 1996; 35: 889-894Abstract Full Text PDF PubMed Scopus (353) Google Scholar). Focal ruptures are most consistent with vaporization (steam formation), and imply that the temperature certainly exceeded 100°C in certain areas of the follicle. The heterogeneous distribution of thermal coagulation may result from local variations in follicular melanin concentration. Various modes of follicle damage have been well documented. It is evident that anagen follicles are more susceptible to most of these methods, including certain drugs, electrolysis, and ionizing irradiation. Drugs may interfere with anagen follicles through two main mechanisms (Tosi et al., 1994Tosi A. Misciali C. Piraccini B.M. Peluso A.M. Bardazzi F. Drug-induced hair loss and hair growth. Incidence, management and avoidance.Drug Saf. 1994; 10: 310-317Crossref PubMed Scopus (106) Google Scholar). First, in anagen effluvium, thallium sulfate (Cavanagh and Gregson, 1978Cavanagh J.B. Gregson M. Some effects of a thallium salt on the proliferation of hair follicle cells.J Pathol. 1978; 125: 179-191Crossref PubMed Scopus (14) Google Scholar), toxic metals (e.g., copper, arsenic, mercury, and cadmium) (Pierard, 1979Pierard G.E. Toxic effects of metals from the environment on hair growth and structure.J Cutan Pathol. 1979; 6: 237-242Crossref PubMed Scopus (20) Google Scholar), and anti-neoplastic agents such as cyclophosphamide (Paus et al., 1994Paus R. Handjiski B. Eichmuller S. Czarnetzki B.M. Chemotherapy-induced alopecia in mice. Induction by cyclophosphamide, inhibition by cyclosporin A, and modulation by dexamethasone.Am J Pathol. 1994; 144: 719-734PubMed Google Scholar), induce an abrupt cessation of mitotic activity in rapidly dividing hair matrix cells. Even topical betamethasone-17-valerate significantly decreases S and G2 + M cell percentages (Schell et al., 1989Schell H. Kiesewetter F. Pausch C. Arai A. Katsuoka K. Hornstein O.P. Influence of topically applied betamethasone-17-valerate on cell cycle kinetics of human anagen hair determined by DNA flow cytometry.Dermatologica. 1989; 178: 12-15Crossref PubMed Scopus (6) Google Scholar). In telogen effluvium, a large number of drugs including anti-coagulants, retinol (vitamin A) and its derivatives, interferons, and anti-hyperlipidemic drugs, precipitate the follicles into premature rest. In addition, the growing anagen follicles are known to be more sensitive to electrolysis (Richards and Meharg, 1995Richards R.N. Meharg G.E. Electrolysis: observations from 13 years and 140,000 hours of experience.J Am Acad Dermatol. 1995; 33: 662-666Abstract Full Text PDF PubMed Scopus (101) Google Scholar) and ionizing irradiation (Ellinger, 1951Ellinger F. Effects of ionizing radiation on growth and replacement of hair.Ann New York Acad Sci. 1951; 53: 682-687Crossref PubMed Scopus (8) Google Scholar;Geary, 1952Geary J.R. Effects of roentgen rays during various phases of the hair cycle of the albino rat.J Anat. 1952; 91: 51-105Crossref Scopus (27) Google Scholar) than the resting telogen hair. On the contrary, some drugs affect only the telogen whereas others affect both the anagen and the telogen follicles. For example, the resting telogen follicles are more sensitive to methylcholanthrene-in-benzene and methylcholanthrene (Chase and Montagna, 1951Chase H.B. Montagna W. Relation of hair proliferation to damage induced in the mouse skin.Proc Soc Exper Biol Med. 1951; 76: 35-37Crossref PubMed Scopus (8) Google Scholar;Wolbach, 1951Wolbach S.B. Hair cycle of the mouse and its importance in the study of sequences of experimental carcinogenesis.Ann New York Acad Sc. 1951; 53: 517-536Crossref PubMed Scopus (23) Google Scholar;Rauch, 1952Rauch H. Effects of topical applications of chemical agents on hair development.Physiol Zool. 1952; 25: 268-272Google Scholar). Etretinate causes an arrest at the onset of anagen and follicular anchorage defect in telogen (Berth-jones and Hutchinson, 1995Berth-jones J. Hutchinson P.E. Novel cycle changes in scalp hair are caused by etretinate therapy.Br J Dermatol. 1995; 132: 367-375Crossref PubMed Scopus (18) Google Scholar). Sensitivity of follicles to selective photothermolysis presents an interesting problem. The most likely target sites for inactivation of hair follicles are the "bulge" (source of follicular stem cells) and/or the papilla (dermal structure supporting the follicle). Although papillae are most superficial during telogen, hairs in telogen did not respond to ruby laser pulses in this study due to lack of pigmentation in the matrix and the portion of the hair shaft imbedded in the dermis. In contrast, during anagen the papillae are deeper and further away from incident laser light – yet anagen follicles were most sensitive and damaged along their entire length. Anagen follicles are more pigmented, suggesting that pigmentation is the primary factor determining sensitivity to selective photothermolysis; however, anatomic or metabolic changes also appear to play a role. In mice, melanogenesis reaches its maximum at days 8–12 of anagen. In contrast, the maximum sensitivity of follicles was at about day 6 of anagen (Figure 4), when the follicles are somewhat less pigmented but may also be more superficial. In humans with brown or black hair (Dierickx et al., 1998Dierickx C.C. Grossman M.C. Farinelli W.A. Anderson R.R. Permanent hair removal by normal-mode ruby laser.Arch Dermatol. 1998; in pressGoogle Scholar), ruby laser pulses delivered through a cold-contact handpiece at 30–60 J per cm2 caused nonscarring temporary and permanent hair loss (Grossman et al., 1996Grossman M.C. Dierickx C.C. Farinelli W.A. Flotte T. Anderson R.R. Damage of hair follicles by normal-mode ruby laser pulses.J Am Acad Dermatol. 1996; 35: 889-894Abstract Full Text PDF PubMed Scopus (353) Google Scholar). In this study, mice treated during anagen developed alopecia without scarring at fluences below 3.16 J per cm2. Scarring alopecia occurred at only 3.16 J per cm2, about one-tenth of the human treatment fluence. Inhibition of mouse hair growth with such low fluences may be explained by the superficial depth and greater pigmentation of mouse compared with human hair follicles; however, scarring alopecia at a fluence less than one-tenth of that causing nonscarring alopecia in humans, is a striking difference. This is best explained by anatomical, rather than biologic differences between mouse and human skin. A relatively unexplored aspect of selective photothermolysis is the influence of distance between adjacent target structures. If the dermis is filled with densely packed hair follicles, it is more difficult to accomplish selective photothermolysis without also damaging the dermis between hair follicles. This appears to be the case in mouse skin during anagen. During anagen, mouse follicles are only about 60 μm apart, and the diameter of a typical mouse anagen follicle is ≈50 μm with a thermal relaxation time, tr, of 2.5 ms. The 2.2 ms pulse duration of the ruby laser used at 3.16 J per cm2, therefore allows time for significant heat conduction during the laser pulse, to tens of micrometers of dermis surrounding each pigmented hair follicle. In mice, this turns out to be a large portion of the dermis. Scaring alopecia in mice may therefore occur as a consequence of the extremely close proximity of adjacent follicles. In contrast, humans do not have fur. Human follicles are much thicker and further apart, such that widespread dermal injury from conducted heat does not occur in this laser pulsewidth domain. It should also be noted that fibrosis seen by histology did not always correlate with scarring seen by gross observation (Table 3, 4), and that both were reversible to some extent between 28 and 56 d after treatment. By definition, scarring is a gross clinical endpoint, whereas fibrosis describes changes in the extracellular matrix of dermis that are associated with different lesions including scars. This study strongly suggests that patients being treated with laser pulses for hair removal, will fare better when their follicles are treated during anagen. Unlike mice, humans do not normally have synchronous hair growth cycles; however,Grossman et al., 1996Grossman M.C. Dierickx C.C. Farinelli W.A. Flotte T. Anderson R.R. Damage of hair follicles by normal-mode ruby laser pulses.J Am Acad Dermatol. 1996; 35: 889-894Abstract Full Text PDF PubMed Scopus (353) Google Scholar found that ruby laser pulses cause both a prolonged growth delay and apparently permanent alopecia in humans (Dierickx et al., 1998Dierickx C.C. Grossman M.C. Farinelli W.A. Anderson R.R. Permanent hair removal by normal-mode ruby laser.Arch Dermatol. 1998; in pressGoogle Scholar). After a single treatment, growth delay was induced for a period of several months in all subjects, at fluences well below that causing significant alopecia in a smaller fraction of the subjects. It is unknown at present whether the laser-induced growth delay is due to a telogen phase, or to recoverable injury during anagen. Regardless of the cause of the laser-induced growth delay, the findings of our animal study strongly suggest that a second treatment given at about the time the hair begins to reappear at the skin surface, may be highly effective. This provides a rational approach for planning laser hair-removal treatments, which fortuitously fits well with the patient's desire for treatment if and when hairs begin to reappear. This study also suggests that hair removal from skin sites with a high fraction of hairs in catagen or telogen, may be somewhat more difficult. In general, these are body sites with shorter hair; however, the anatomic depth and degree of pigmentation of follicles also play important roles in selective photothermolysis, which vary between different body sites. There is little or no detailed information comparing laser hair removal from different body sites. In mice, we found that immediate erythema was an excellent indicator of injury to hair follicles, which also correlated with fluence-dependent hair loss of follicles treated during anagen. Immediate perifollicular erythema and edema also occur in humans treated with ruby laser pulses for hair removal. Our clinical impressions concur with this animal study – it is unlikely that alopecia will be achieved without an inflammatory response. Immediate post-treatment perifollicular erythema therefore provides an endpoint that can be helpful for setting the treatment fluence. In this study, double-pulsing at 3.16 J per cm2 fluence neither increased effectiveness for hair removal, nor increased the severity of fibrosis or scarring. This is consistent with clinical use of the ruby laser delivered through a cold-contact handpiece for hair removal (Dierickx et al., 1998Dierickx C.C. Grossman M.C. Farinelli W.A. Anderson R.R. Permanent hair removal by normal-mode ruby laser.Arch Dermatol. 1998; in pressGoogle Scholar), and implies that overlapping of adjacent pulses delivered at a low repetition rate is neither harmful nor helpful. We thank Dr. Valeria Duque for help with animal preparation, Bart Johnson for his help with biopsy specimens, and Norm Michaud for his help with histologic photographs. We also thank Dr. Gregory Altshuler for helpful discussions and for assistance with modeling of selective thermolysis of mouse hair follicles. This study was supported by grants from the Palomar Medical Technologies.