Acne Keloidalis Is a Form of Primary Scarring Alopecia

STUDY Acne Keloidalis Is a Form of Primary Scarring Alopecia

COL Leonard C. Sperling, MC, USA; Carri Homoky, MD; LTC Laura Pratt, MC, USA; Purnima Sau, MD

Objective: To better define the pathogenesis of acne ke- lar or perifollicular microorganisms. The number and type loidalis (AK). of hairs were also recorded.

Design: Prospective, blinded study of histologic mate- Results: The most common findings in the 19 histo- rial collected from 10 patients with clinically typical AK. logically abnormal specimens were perifollicular, chronic (lymphocytic and plasmacytic) inflammation, most in- Setting: Outpatient dermatology clinic of a military ter- tense at the level of the isthmus and lower infundibu- tiary care medical center. lum; lamellar fibroplasia, most marked at the level of the isthmus; complete disappearance of sebaceous glands, as- Patients: Ten male volunteers 18 years or older with sociated with inflamed or destroyed follicles; thinning of early AK lesions (1- to 4-mm firm papules on the lower the follicular epithelium, most marked at the level of the occipital/nuchal region). isthmus; and total epithelial destruction (superficial and deep), with residual “naked” hair fragments. Even some Data Source: Biopsy specimens from small, early le- “normal” specimens contained true follicular scars, dem- sions and from clinically uninvolved skin, studied his- onstrating that normal-appearing scalp skin had previ- tologically with transverse sectioning. ously been affected by the disease.

Intervention: Three separate 4-mm punch biopsy Conclusions: Acne keloidalis is a primary form of scar- specimens of the scalp (lesional, perilesional, and ring alopecia, and many of the histologic findings closely “normal” scalp) were obtained from each volunteer. resemble those found in certain other forms of cicatri- The specimens were processed using transverse sec- cial alopecia. Extensive subclinical disease may be pres- tioning. ent in patients with AK and can account for some of the permanent hair loss. Overgrowth of microorganisms does Main Outcome Measures: The primary variables for not appear to play an important role in the pathogenesis data analysis were the presence or absence of the follow- of the disease. There is no etiologic relationship being histologic features: premature loss of the inner root tween AK and pseudofolliculitis barbae. Therapies found sheath; eccentric placement of shaft, with thinning of the to be useful in other forms of inflammatory scarring alo- outer root sheath; lamellar fibroplasia surrounding the pecia are useful in the treatment of early AK. follicle; loss of sebaceous glands; evidence of follicular destruction or scarring; inflammation; and intrafollicu- Arch Dermatol. 2000;136:479-484

CNE KELOIDALIS (AK) was and posterior neck. In a minority of cases, first described by Kaposi1 lesions are more numerous on the vertex in 1869 as dermatitis pap- and crown; therefore, we have chosen to illaris capillitii. Kaposi use the term acne keloidalis without the From the Department of concluded that the dis- modifier nuchae. Initially, hairs can be seen Dermatology, Uniformed ease was “a peculiar idiopathic chronic in- exiting the papules. However, with time, Services University, Bethesda, A flammation and connective-tissue new for- the papules tend to coalesce and form firm Md (Dr Sperling); Dermatology mation in the cutis.”2 Bazin originated the hairless keloidlike protuberant plaques that Associates, Kingsport, Tenn term acne keloidalis in 1872,3 and it is still can be painful and cosmetically disfigur- (Dr Homoky); and used today, although some authors4 pre- ing. Abscesses and sinuses exuding pus can Dermatology Department, 5 Portsmouth Naval Medical fer the term folliculitis keloidalis. This dis- be present in advanced cases. Although Center, Portsmouth, Va order typically affects young black men and AK may be asymptomatic, mild symp- (Dr Pratt). Dr Sau is in private begins as small, smooth, firm papules, with toms of burning and itching are often pres- practice in Silver Spring, Md. occasional pustules, on the occipital scalp ent.3,5 Acne keloidalis has been rarely re-

ARCH DERMATOL / VOL 136, APR 2000 WWW.ARCHDERMATOL.COM 479

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 ported to occur in black women and white men,2,6 but is at least 10 times more common in blacks than in whites.3 SUBJECTS AND METHODS The condition includes nearly 0.5% of all dermatologic cases in the black population.7 SUBJECTS Medical treatment for early papular lesions in- This study was approved by the Clinical Investiga- cludes intralesional injections of corticosteroids, topi- tion Committee and the Human Use Committee/ cal steroids, and topical or oral antibiotics (usually Institutional Review Board of the Walter Reed Army tetracycline).5,6,8,9 In some cases, medical treatment is Medical Center, Washington, DC. All subjects en- 10 rolled in the study voluntarily agreed to participate successful in controlling the condition. Once large ke- and gave written informed consent. loidlike plaques have developed, the condition is resis- Ten male volunteers 18 years or older with early tant to all medical treatment, often requiring surgical 11,12 AK lesions (1- to 4-mm firm papules on the lower oc- removal. cipital/nuchal region) were recruited from the derma- The cause of AK remains unclear. Constant irrita- tology clinic of the Walter Reed Army Medical Cen- tion by shirt collars,2,8 low-grade bacterial infection,8 in- ter. Cases were collected over a 2-year period. Patients jury during short haircuts,3,5 curved hair and hair fol- who had been treated for AK with oral medications licles,8,9 and an autoimmune process3 have all been within 6 months or who had large keloidlike nodules suggested as pathogenic mechanisms. Thus far, none of or plaques were excluded. If patients had been using these theories has been proven, and all are based on cir- topical antibiotics or corticosteroids, there was a 3-week cumstantial evidence. The notion that lesions of AK are period during which they were off all medications be- 13 fore scalp biopsy specimens were obtained. caused by ingrowing hairs, analogous to the situation in pseudofolliculitis barbae,14 persists, despite the fact that METHODS there is no evidence to support this hypothesis.3 There Three 4-mm punch biopsy specimens of the scalp were is considerable evidence to the contrary,15 and our find- obtained from each volunteer. Each biopsy site was ings conclusively separate the pathogenic mechanisms anesthetized with 1% lidocaine and 1:100 000 epi- of AK and pseudofolliculitis barbae. nephrine. The first biopsy specimen, called lesional, was Histologic evaluation of AK has been performed in a discrete papule taken from clinically involved skin a few published cases, but the exact sequence of inflam- on the lower occiput/nuchal region. The second bi- matory events in this disorder has not been well de- opsy specimen, called perilesional, was taken from nor- fined. Often the small papular lesions seen early in the mal-appearing scalp skin, within 1 cm of the lesional specimen. The third biopsy specimen, called normal, disease clinically demonstrate histologically advanced was taken from the parietal scalp 4 cm superior to the stages of inflammation, making it difficult to determine apex of the left ear. The specimens were placed in 10% the earliest and perhaps triggering events in the disease formalin and processed using transverse sectioning.16 process.15 Multiple levels from each specimen were examined. Specimen labels were coded in a random fashion so that RESULTS the dermatopathologists were blinded to both the patient and the type of biopsy specimen (lesional, per- Data on specimens showing abnormal pathologic char- ilesional, or normal skin). Any specimen demonstrat- acteristics are presented in the Table. Nine of 10 speci- ing inflammation or scarring was also stained to exclude fungus (periodic acid–Schiff and methenamine sil- mens from lesional skin, 4 of 10 specimens from perile- ver) and bacteria (Brown and Hopps, and Brown and sional skin, and 6 of 10 specimens from clinically normal Brenn). All histologic specimens were reviewed by the skin were histologically abnormal. The most consistent same 2 dermatopathologists (L.C.S. and P.S.). findings among these 19 histologically abnormal specimens were as follows: perifollicular chronic (lympho- DATA ANALYSIS cytic and plasmacytic) inflammation, most intense at the The small number of subjects used in this protocol level of the isthmus and lower infundibulum (n = 18; was based on the expectation of a large difference in Figure 1); complete disappearance of sebaceous glands the histopathologic features of diseased and normal associated with inflamed or destroyed follicles (n = 16); sites and the difficulty in obtaining multiple biopsy thinning of the follicular epithelium, most marked at the specimens from otherwise healthy subjects. level of the isthmus (n = 15; Figure 2); lamellar fibro- Presentation of data is primarily descriptive, pro- viding results of histopathologic analysis through pho- plasia, most marked at the level of the isthmus (n = 14; tomicrographs and a table. The primary variables for focal or total epithelial destruction (superficial and data analysis are the presence or absence of the follow- deep), with residual “naked” hair fragments (n = 11; ing histologic features: premature loss of the inner root Figure 3); dilatation of the follicular canal extending sheath; eccentric placement of shaft, with thinning of down to or below the isthmus (n = 4; Figure 4); prema- the outer root sheath; lamellar fibroplasia surround- ture desquamation of the inner root sheath affecting at ing the follicle; loss of sebaceous glands; evidence of least 1 follicle (n = 3; Figure 4); and acute (neutrophilic) follicular destruction or scarring; inflammation (loca- inflammation surrounding degenerating follicular com- tion and type [ie, acute, chronic, or granulomatous]); ponents (n = 2). and intrafollicular or perifollicular organisms (bacte- One lesional specimen was histologically normal, ria, fungi, and Demodex). The number and type of hairs found in each specimen were also recorded. except for the presence of mild perifollicular epidermal spongiosis and chronic superficial perivascular inflammation. This could account for the palpability

ARCH DERMATOL / VOL 136, APR 2000 WWW.ARCHDERMATOL.COM 480

Complete Thinning Dilatation Premature Chronic Disappearance of the Focal or of the Desquamation No. of Lesion No. Perifollicular of Sebaceous Follicular Lamellar Epithelial Follicular of the Inner Acute No. of Inflamed (Patient No.) Inflammation Glands Epithelium Fibroplasia Destruction Canal Root Sheath Inflammation Follicles Follicles Lesional Scalp Biopsy Specimens

L1 (2) + + + + + − + − 7 4 L3 (3) + + + + + − − − 13 2 L4 (1) + + + + + + − − 8 1 L5 (6) + + + + + + − − 10 3 L6 (8) + + + + − − + − 13 6 L7 (10) + + + + − − + − 10 2 L8 (9) + + + + + − − + 7 2 L9 (4) + + + + − − − − 17 2 L10 (5) + + − − + − − + 26 2 Perilesional Scalp Biopsy Specimens

P1 (6) + − + + − + − − 14 2 P2 (9) + + + + − − − − 7 1 P3 (5) + − − − − − − − 25 2 P4 (2) − + − − + − − − 0 0 “Normal” Scalp Biopsy Specimens

N1 (10) + (Mild) + − − + − − − 13 2 N2 (5) + + + + + − − − 26 4 N3 (6) + − + − − + − − 21 2 N4 (8) + + + + − − − − 17 7 N5 (1) + + + + + − − − 12 5 N6 (4) + + + + + − − − 11 2 Total No. Positive 18/19 16/19 15/19 14/19 11/19 4/19 3/19 2/19 Mean, 14 Mean, 3 (n = 19) Normal specimens, 0/11 0/11 0/11 0/11 0/11 0/11 0/11 0/11 Mean, 16 0/11 pooled data (n = 11)

*Plus sign indicates present; minus sign, absent.

of the lesion that was selected for biopsy. This lesion found in the lesional specimens, but, in general, the find- was felt to be unrelated to the patient’s AK and high- ings were less dramatic. However, even the normal speci- lights one of the pitfalls of selecting very early lesions mens contained true follicular scars (Figure 5), dem- for examination. onstrating that normal-appearing scalp skin had previously Those few follicles that demonstrated premature des- been affected by the disease. quamation of the inner root sheath all showed marked degenerative changes. Therefore, this inner root sheath COMMENT change was not an early finding. Acute inflammation, when present, was found only at the level of the infun- Well-established lesions of AK that are examined in stan- dibulum and was always associated with total destruc- dard vertical sections reveal dense dermal fibrosis with- tion of the epithelium at that level. Small numbers of De- out evidence of keloid formation, as well as chronic in- modex organisms were found in 5 of the 30 specimens flammation and the presence of many plasma cells. Hair studied (normal, n = 1; perilesional, n = 2; and lesional, follicles are disrupted, and microabscesses and/or foreign- n = 2). All the organisms were found within sebaceous body reactions surround hair shafts. There is a conspicu- ducts, and only one of the affected follicles showed any ous absence of sebaceous glands in well-developed le- degree of inflammation. Only rare spores (presumably sions.3,5 Pityrosporum) were identified in specimens showing in- Herzberg and colleagues15 described papular le- flammation or scarring. A few colonies of gram-positive sions of AK using transverse sections in an attempt to cocci were present in the scale/crust on the epidermal sur- characterize the evolution of the disease. They evalu- face of the acutely inflamed follicles. However, bacteria ated 7 discrete papules from the occipital scalp of 4 pa- were rare in the follicular canals and absent in the peri- tients and found that inflammation was most marked at follicular region of all specimens. the level of the lower infundibulum and isthmus of the Despite the fact that the selected biopsy sites ap- follicle. The character of this inflammation changed from peared to be clinically normal, 4 of 10 perilesional and 6 being more acute at the upper levels of the infundibu- of 10 normal specimens showed at least one follicle with lum to chronic and granulomatous at a deeper level marked abnormalities. The changes were similar to those (isthmus). Overall, there was a paucity of sebaceous glands

ARCH DERMATOL / VOL 136, APR 2000 WWW.ARCHDERMATOL.COM 481

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 Figure 1. Lesional biopsy specimen (patient 6) at the level of the isthmus Figure 3. Lesional biopsy specimen (patient 3) at the level of the isthmus showing predominantly lymphocytic inflammation surrounding 2 follicles, showing nearly total destruction of the follicular epithelium at this level. which happen to be sharing a common follicular channel at this level. No Residual hair shafts and follicular debris are surrounded by fibroplasia and inflammation was present at the level of the infundibulum chronic inflammation. The epithelium of these 2 follicles is completely intact (hematoxylin-eosin, original magnification ϫ200). at the level of the infundibulum, but marked chronic perifollicular inflammation is present (hematoxylin-eosin, original magnification ϫ200).

Figure 2. Biopsy specimen from “normal” scalp (patient 1) at the level of the Figure 4. Lesional biopsy specimen (patient 1) at the suprabulbar level (well upper isthmus showing that marked thinning of the follicular epithelium below the isthmus) showing that the inner root sheath has undergone places a hair shaft in close proximity with the dermis. Perifollicular lamellar desquamation and the outer root sheath has begun to cornify. At the level of fibroplasia and epithelial spongiosis are also present. At a slightly lower level, the isthmus, the follicular epithelium of this follicle has totally degenerated, chronic inflammation is also present (hematoxylin-eosin, original suggesting that these changes are late events in the pathogenesis of the magnification ϫ200). disease (hematoxylin-eosin, original magnification ϫ200).

at any stage of the inflammation. Based on their find- subsides. The small papular lesions of AK represent fol- ings, the authors propose the following sequence of in- licles affected by marked inflammation and edema. flammatory events: Acute inflammation begins in the deep However, a spectrum of subclinical disease exists in infundibulum or isthmus or perhaps in the sebaceous which many follicles can be affected by relatively mild gland. This weakens the follicular wall; subsequently, hairs inflammation. are released into the dermis. The naked hairs stimulate We propose the following pathogenesis, based on a foreign-body reaction with acute and chronic granu- our findings in both clinical and subclinical disease. An- lomatous inflammation. Subsequent fibrosis occurs within tigens of the follicular epithelium or within the follicu- the dermis, which then might distort and occlude the fol- lar canal attract perifollicular inflammatory cells at the licular lumen and consequently lead to hair retention level of the isthmus. Potential intrafollicular antigens in- within the deeper follicle. This then leads to further in- clude Demodex, normal skin flora (fungal spores and bac- flammation and scarring. The role of Demodex was con- teria) and their metabolic by-products, cosmetics, se- sidered but not substantiated. bum, and desquamated kertatinocytes. Acne keloidalis We have been able to confirm and expand the find- is a disease of adulthood, suggesting that candidate an- ings of Herzberg and colleagues.15 Furthermore, we can tigens are those that are more prominent after puberty. establish AK as a form of primary inflammatory scarring The sebaceous gland is either an early target of inflam- alopecia. The earliest changes seen, found in visible mation or an “innocent bystander” that is destroyed early lesions as well as nonlesional scalp skin, affect isolated in the inflammatory process. Perifollicular inflamma- follicles or a single follicular unit. Affected follicles tion alters or weakens the follicular wall and is mani- either are destroyed or undergo repair as inflammation fested first by spongiosis and mild lymphocytic exocy-

ARCH DERMATOL / VOL 136, APR 2000 WWW.ARCHDERMATOL.COM 482

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 Figure 5. Biopsy specimen from “normal” scalp at the suprabulbar level showing a true follicular scar (column of connective tissue) that marks the site of a former follicle (hematoxylin-eosin, original magnification ϫ200).

tosis. This weakened epithelial wall may then become increasingly “leaky” to intrafollicular antigens, amplify- ing the inflammatory reaction. A reparative attempt en- sues, in the form of concentric lamellar fibroplasia. The hair shaft eventually migrates through the weakened follicular wall and enters the dermis, inciting intense inflammation and epithelial destruction. A follicular scar is the result. Hair shaft fragments and degenerating epithelial components incite hypertrophic scarring in some indi- viduals, resulting in a keloidal tissue reaction. This chain Figure 6. Crown of the scalp of an African American woman with the of events can be broken at any point (spontaneously or follicular degeneration syndrome (central centrifugal scarring alopecia). with treatment), allowing for at least partial follicular heal- ing. This accounts for viable follicles surrounded by lamellar fibroplasia with loss of sebaceous glands. The suggestion made by Herzberg et al15 that the presence of Demodex might be related to the pathogenesis of the disease was not supported but certainly not entirely disproved by our findings. Similarly, we found little evidence to support the role of pathogenic bacteria in the pathogenesis of AK, although histopathologic analysis is a poor tool for studying small numbers of bacteria and their sometimes potent antigens. However, we have clearly established that ingrown hairs, as found in pseudofolliculitis barbae, play no role in the cause of this condition. The histologic features of pseudofolliculitis barbae have been well described.14 A short distance from the follicular apparatus or within the upper portion of the infundibulum, there is external penetration of the curved hair shaft into the skin. The process is simply that of a foreign-body reaction, with the tip of the hair shaft act- ing as the foreign body. This is totally unlike the situation with AK, and the notion that AK and pseudofolliculitis barbae share a common pathogenesis is erroneous and should be abandoned. Acne keloidalis can be considered to be a bona fide Figure 7. Typical papules of acne keloidalis clustered on the occipital scalp of the same patient as in Figure 6. member of the scarring alopecia family of disorders. Most of the histologic findings seen in early AK are virtually identical to those found in another form of primary scar- tionship between the 2 conditions may exist, however, ring alopecia, namely central centrifugal scarring alope- because both tend to predominate in the African Ameri- cia (CCSA), also known as the follicular degeneration syn- can population, and they commonly coexist in the same drome,17,18 hot-comb alopecia,17,19 and pseudopelade of patient (Figure 6 and Figure 7). the vertex/crown.20,21 This does not imply that AK is the The histologic similarities between AK and CCSA same disease as CCSA since different diseases com- also suggest a similar approach to therapy. In our expe- monly result in similar histopathologic changes. A rela- rience, the combination of potent topical corticoste-

ARCH DERMATOL / VOL 136, APR 2000 WWW.ARCHDERMATOL.COM 483

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 The opinions or assertions contained herein are the private views of the authors and are not to be construed as of- ficial or as reflecting the views of the Department of the Army or the Department of Defense. Corresponding author: COL Leonard C. Sperling, MC, USA, Department of Dermatology, Uniformed Ser- vices University, Bethesda, MD 20814 (e-mail: lsperling @usuhs.mil).

REFERENCES

1. Kaposi M. Ueber die sogennante Framboesia und mehrere andere arten von pa- pillaeren neubildungen der Haut. Arch Dermatol Syphilol. 1869;3:382-423. 2. Adamson H. Dermatitis papillaris capillitii (Kaposi): acne keloid. Br J Dermatol. Figure 8. A male patient with acne keloidalis. Numerous papular lesions are 1914;26:69-83. present, but patches of alopecia without obvious lesions are scattered throughout the area of involvement. Keloidal nodules are absent. 3. Cosman B, Wolff M. Acne keloidalis. Plast Reconstr Surg. 1972;50:25-30. 4. Fox H. “Folliculitis keloidalis” a better term than “dermatitis papillaris capillitii.” Arch Dermatol Syphilol. 1947;55:112-113. roids, such as fluocinonide, along with prolonged use of 5. Dinehart S, Herzerg A, Kerns B, Pollack S. Acne keloidalis: a review. J Dermatol an oral antibiotic, such as tetracycline, is of consider- Surg Oncol. 1989;15:642-647. 18 6. Dinehart SM, Tanner L, Mallory SB, Herzberg AJ. Acne keloidalis in women. Cu- able benefit in the treatment of both CCSA and AK. Both tis. 1989;44:250-252. conditions are characterized by rather extensive sub- 7. Hollander L. Treatment of folliculitis keloidalis chronica nuchea (acne keloid). clinical disease and are chronic and progressive, requir- Arch Dermatol Syphilol. 1952;64:639-640. ing long-term prophylactic treatment. 8. Halder R. Pseudofolliculitis barbae and related disorders. Dermatol Clin. 1988; The keloidal masses seen in some patients with AK 6:407-412. 9. Halder R. Hair and scalp disorders in blacks. Cutis. 1983;32:378-380. represent a peculiar host response to degenerating fol- 10. Kenney JJ. Dermatoses common in blacks. Postgrad Med. 1977;61:122-127. licular components and the resultant inflammation. Once 11. Malherbe W. Sycosis nuchae and its surgical treatment. Plast Reconstr Surg. the integrity of the follicular epithelium is violated, a ni- 1971;47:269-271. dus for superinfection is created. Bacterial overgrowth 12. Kanthak F, Cullen M. Skin graft in the treatment of chronic furunculosis of the incites further tissue inflammation and damage, and the posterior surface of the neck (folliculitis keloidalis). South Med J. 1951;44:1154- 1157. resulting tissue swelling and fibrosis trap both viable and 13. Smith JD, Odom RB. Pseudofolliculitis capitis. Arch Dermatol. 1977;113:328- disintegrated follicles in a hypertrophic scar. Trapped hair 329. shafts and epithelium elicit a foreign-body granuloma- 14. Strauss J, Kligman A. Pseudofolliculitis of the beard. Arch Dermatol. 1956;74: tous reaction and provide a site for ongoing infection. 533-542. These changes, although they represent a dramatic mani- 15. Herzberg A, Dinehart S, Kerns B, Pollack S. Acne keloidalis: transverse microscopy, immunohistochemistry and electron microscopy. Am J Dermatopathol. 1990; festation of the disease, are secondary events and unre- 12:109-121. lated to the events that initiate the disease. In many pa- 16. Frishberg DP, Sperling LC, Guthrie VM. Transverse scalp sections: a proposed tients, the disease behaves like other forms of scarring method for laboratory processing. J Am Acad Dermatol. 1996;35:220-222. alopecia, with patches of hair loss separate from any papu- 17. Sperling LC, Sau P. The follicular degeneration syndrome in black patients: “hot comb alopecia” revisited and revised. Arch Dermatol. 1992;128:68-74. Figure 8 lonodular lesions ( ). In theory, if early disease 18. Sperling LC, Skelton HG III, Smith KJ, Sau P, Friedman K. The follicular degen- could be prevented or effectively treated, the chronic ke- eration syndrome in men. Arch Dermatol. 1994;130:763-769. loidal component of AK would not occur. 19. LoPresti P, Papa CM, Kligman AM. Hot comb alopecia. Arch Dermatol. 1968;98: 234-238. Accepted for publication September 26, 1999. 20. Solomon A. The transversely sectioned scalp biopsy specimen: the technique and an algorithm for its use in the diagnosis of alopecia. Adv Dermatol. 1994;9: This study was funded in part by the Department of 127-157. Clinical Investigation, Walter Reed Army Medical Center, 21. Templeton S, Solomon A. Scarring alopecia: a classification based on micro- Washington, DC. scopic criteria. J Cutan Pathol. 1994;21:97-109.

ARCH DERMATOL / VOL 136, APR 2000 WWW.ARCHDERMATOL.COM 484