[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 50.16.107.222. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
Table 1. 
Clinical Staging of Acute Graft-vs-Host Disease
Clinical Staging of Acute Graft-vs-Host Disease
Table 2. 
Overall Clinical Grading of Acute Graft-vs-Host Disease
Overall Clinical Grading of Acute Graft-vs-Host Disease
Table 3. 
Clinical Parameters According to Biopsy Status*
Clinical Parameters According to Biopsy Status*
Table 4. 
Clinical Parameters According to Biopsy Result*
Clinical Parameters According to Biopsy Result*
Table 5. 
Outcome of Bone Marrow Transplantation According to Skin Biopsy Status*
Outcome of Bone Marrow Transplantation According to Skin Biopsy Status*
Table 6. 
Outcome of Bone Marrow Transplantation According to Skin Biopsy Result*
Outcome of Bone Marrow Transplantation According to Skin Biopsy Result*
1.
Deeg  HJStorb  R Graft-versus-host disease: pathophysiological and clinical aspects. Ann Rev Med. 1984;3511- 24Article
2.
Glucksberg  HStorb  RFefer  A  et al.  Clinical manifestations of graft-versus-host disease in human recipients of marrow from HLA-matched sibling donors. Transplantation. 1984;18295- 304Article
3.
Beatty  PGClift  RAMickelson  EM  et al.  Marrow transplantation from related donors other than HLA-identical siblings. N Engl J Med. 1985;313765- 771Article
4.
Vogelsang  GBHess  ADSantos  GW Acute graft-versus-host disease: clinical characteristics in the cyclosporin era. Medicine (Baltimore). 1989;67163- 174
5.
Deeg  HJCottler-Fox  MBurakoff  SJedDeeg  HJedFerrara  JedAtkinson  Ked Clinical spectrum and pathophysiology of acute graft-versus-host disease. Graft-vs.-Host Disease: Immunology, Pathophysiology, and Treatment New York, NY Marcel Dekker1990;311- 336
6.
Saurat  JH Cutaneous manifestations of graft-versus-host disease. Int J Dermatol. 1981;4249- 256Article
7.
Thomas  EDStorb  RClift  RA  et al.  Bone marrow transplantation. N Engl J Med. 1995;292832- 843Article
8.
Przepiorka  DWeisdorf  DMartin  P  et al.  Consensus conference on acute GVHD grading. Bone Marrow Transplant. 1995;15825- 828
9.
Storb  RPrentice  RLSullivan  KM  et al.  Predictive factors in chronic graft-versus-host disease in patients with aplastic anemia treated by marrow transplantation from HLA-identical siblings. Ann Intern Med. 1983;98461- 466Article
10.
Storb  RPrentice  RLBuckner  CD  et al.  Graft-versus-host disease and survival in patients with aplastic anemia treated by marrow grafts from HLA-identical siblings: beneficial effect of a protective environment. N Engl J Med. 1983;308302- 307Article
11.
Woodruff  JMEltringham  JRCasey  HW Early secondary disease in the rhesus monkey, I: a comparative histopathologic study. Lab Invest. 1969;20499- 511
12.
Woodruff  JMButcher  WIHellerstein  LJ Early secondary disease in rhesus monkey, II: electron microscopy of changes in mucous membrane and the external epithelia as demonstrated in the tongue and lip. Lab Invest. 1972;2785- 98
13.
Slavin  RESantos  GW The graft-versus-host reaction in man after bone marrow transplantation: pathology, pathogenesis, clinical features and implications. Clin Immunol Immunopathol. 1973;1472- 498Article
14.
Woodruff  JMHansen  JAGood  RASantos  GWSlavin  RE The pathology of the graft-versus-host reaction (GvHR) in adults receiving bone marrow transplants. Transplant Proc. 1976;8675- 684
15.
Sale  GELerner  KGBarker  EAShulman  HMThomas  ED The skin biopsy in the diagnosis of acute graft-versus-host disease in man. Am J Pathol. 1977;89621- 633
16.
Hymes  SRFarmer  ERLewis  PGTutschka  PJSantos  GW Cutaneous graft-versus-host reaction: prognostic features seen by light microscopy. J Am Acad Dermatol. 1985;12468- 474Article
17.
Elliott  CJSloane  JPSanderson  KVVincent  MShepherd  VPowles  R The histological diagnosis of cutaneous graft versus host disease: relationship of skin changes to marrow purging and other clinical variables. Histopathology. 1987;11145- 155Article
18.
Kohler  SHendrickson  MRChao  NJSmoller  BR Value of skin biopsies in assessing prognosis and progression of acute graft-versus-host disease. Am J Surg Pathol. 1997;21988- 996Article
19.
Drijkoningen  MDe Wolf-Peeters  CTricot  GDegreef  HDesmet  V Drug-induced skin reactions and cutaneous graft-versus-host reaction: a comparative immunohistochemical study. Blut. 1988;5669- 73Article
20.
Sviland  LOearson  ADJEastham  EJHamilton  PJProctor  SJMalcom  AJ Histological features of skin and rectal biopsy specimens after autologous and allogeneic bone marrow transplantation. J Clin Pathol. 1988;41148- 154Article
21.
Darmstadt  GLDonnenberg  ADVogelsang  GBFarmer  ERHorn  TD Clinical, laboratory, and histopathologic indicators of the development of progressive acute graft-versus-host disease. J Invest Dermatol. 1992;99397- 402Article
22.
Barksdale  SKOberlender  SABarnhill  RL "Rush" skin biopsy specimens in a tertiary medical center: diagnostic yield and clinical utility. J Am Acad Dermatol. 1998;38548- 554Article
Study
June 2000

Clinical Significance of Skin Biopsies in the Diagnosis and Management of Graft-vs-Host Disease in Early Postallogeneic Bone Marrow Transplantation

Author Affiliations

From the Divisions of Dermatology (Drs Zhou and Rivers) and Hematology (Dr Barnett), Department of Medicine, Vancouver General Hospital and University of British Columbia, Vancouver.

Arch Dermatol. 2000;136(6):717-721. doi:10.1001/archderm.136.6.717
Abstract

Objective  To determine the value of skin biopsies in the management of suspected graft-vs-host disease (GVHD) within 30 days of allogeneic bone marrow transplantation (BMT).

Design  Retrospective study based on review of a BMT database.

Setting  Leukemia/BMT ward of a tertiary care, university teaching hospital.

Patients  One hundred and eighty-seven consecutive patients who received allogeneic BMT between January 1, 1994, and June 30, 1997, at Vancouver General Hospital, Vancouver, British Columbia.

Main Outcome Measures  (1) Skin biopsy frequency for patients with rashes suggestive of acute GVHD; (2) clinical significance of skin biopsy in the management of patients with suspected acute GVHD after BMT; (3) relationship between severity of clinical GVHD and the likelihood to receive GVHD therapy; and (4) relationship between biopsy status or biopsy result and outcome of BMT (acute and chronic GVHD, transplant-related mortality, and overall and event-free survival).

Results  During the early post-BMT period (<30 days after BMT), 88 patients had rashes suggestive of acute GVHD; of these, 51 (58%) underwent skin biopsy to confirm the diagnosis. Skin biopsies were performed more often for higher clinical stages of cutaneous GVHD. There was no significant difference between the patients with positive biopsy findings and those with negative findings, either in the clinical severity of acute GVHD or in likelihood to receive treatment for GVHD. Most (85%) of the patients who underwent biopsies and received GVHD therapy had treatment initiated before skin biopsies were performed or before the results were available. The higher the clinical grade of overall acute GVHD, the more likely it was that the patients were treated for GVHD (P<.001). The outcome of BMT was not influenced by the skin biopsy status or biopsy result.

Conclusions  The biopsy findings correlated poorly with the clinical severity of skin rash suggestive of acute GVHD soon after BMT. The decision to treat suspected acute GVHD depended not on skin biopsy findings but rather on clinical severity of acute GVHD. In this regard, skin biopsy has a limited role in the management of patients early after allogeneic BMT.

GRAFT-VS-HOST disease (GVHD) after allogeneic bone marrow transplantation (BMT) is an immunological process in which activated donor lymphocytes mount an attack on recipient tissues. This process involves secretion of cytokines and recruitment of additional inflammatory cells. The severity of acute GVHD is one of the most important prognostic factors that predict outcome of BMT.1

The likelihood of developing acute GVHD (ie,within 100 days of BMT) depends on a number of factors, including the degree of histoincompatibility, number of T lymphocytes in the graft, patient's age, and the prophylactic immunosuppressive regimen used.25 Clinically relevant acute GVHD occurs in 35% to 50% of patients given HLA-identical marrow grafts.1 The skin is often the initial organ to manifest acute GVHD.1,6 Clinically, acute cutaneous GVHD is characterized by a maculopapular rash that involves the palms and soles and, as the disease progresses, the upper trunk, neck, cheeks, and ears.1,2,6

In addition to the skin, the liver and gastrointestinal system are often involved in acute GVHD.1 As summarized in Table 1 and Table 2, acute GVHD can be graded according to its clinical severity.1,2,7,8 Outcome analyses have established the prognostic value of this staging system in that the higher the overall clinical grade, the higher the transplant-related mortality.1,9,10

Skin biopsies are frequently used to confirm the clinical diagnosis of acute GVHD. Histologically, acute GVHD of the skin is characterized by varying degrees of damage to the epidermal keratinocytes.1118 In grade 1, vacuolization of the basal keratinocytes is present; in grade 2, both basal keratinocyte vacuolization and dyskeratotic keratinocytes are present; in grade 3, focal clefting of the basal layer is formed; and in grade 4, the epidermis is totally separated from the underlying dermis.

The interpretation of skin biopsy results after BMT can be difficult, however, because similar histologic findings may result from the high-dose chemotherapy/radiotherapy of the preparatory regimen before transplantation or other drug treatment afterwards.17,1921 The situation is further complicated by the fact that frequently there is a lack of correlation between clinical features and histologic findings.16,17 For example, the skin in a clinically affected area may seem unremarkable histologically, or clincally uninvolved skin can display basal keratinocyte vacuolization and necrosis consistent with well-established GVHD.17 Therefore, the diagnostic value of skin biopsies has been questioned.1521

Several investigators have studied the prognostic value of skin biopsy findings after BMT.18,21 In 1 report, dyskeratotic keratinocytes, number of exocytosed lymphocytes, and presence of follicular involvement did not correlate with outcome.18 In another, more recent study, no significance was found in any of the skin histologic parameters analyzed, including necrotic (dyskeratotic) keratinocytes, basal vacuolization, satellitosis, and appendageal involvement.21 These studies suggest that skin biopsies are of little prognostic value.

Despite these controversies, skin biopsies are still performed regularly on patients after BMT, sometimes as "rush" specimens.22 To our knowledge, there have been few studies to define the role, if any, of skin biopsies in the management of such patients. A retrospective review of data from a leukemia/BMT ward between 1994 and 1997 was performed to determine the frequency and clinical utility of skin biopsies within 30 days of BMT in the management of patients with suspected GVHD.

PATIENTS AND METHODS
DATABASE

A BMT database is maintained by the Leukemia/BMT Program of British Columbia at Vancouver General Hospital. For each patient undergoing BMT, the following information is recorded with regard to acute GVHD of the skin, liver, and gut: onset, clinical stage, date and result of biopsies, and therapeutic interventions. In addition, clinical outcome variables are recorded, including (1) peak cutaneous and overall clinical severity of acute GVHD; (2) development of chronic GVHD (limited vs extensive); (3) mortality due to GVHD; and (4) mortality due to other causes such as infection, graft failure, regimen-related toxic effects, and relapse of malignancy. A widely used clinical staging/grading system8 is used for the construction of the database and is summarized in Table 1 and Table 2.

POLICY FOR DIAGNOSIS AND TREATMENT OF ACUTE GVHD

The diagnosis of acute GVHD (ie, within 100 days of BMT) was based on clinical symptoms and signs as well as the results of tests (ie, liver function, endoscopy, and skin, gut, or liver biopsies). There was no specific guideline to indicate when a skin (or other) biopsy should be carried out, although histologic confirmation of the diagnosis was obtained routinely. Primary treatment of established acute GVHD was with corticosteroids and started within 12 hours of the diagnosis. Once initiated, the duration of treatment was for a minimum of 6 weeks and did not depend on the results of diagnostic tests. Similarly, cyclosporine was generally continued for 3 to 4 months (related donor) or 6 months (unrelated donor) after BMT and then, if the GVHD was quiescent, tapered over a few weeks. The exception to this was for patients deemed to be at high risk for relapse, for whom the cyclosporine therapy was discontinued earlier.

STUDY DESIGN

Patients who developed skin rashes before day 30 after BMT were placed into 1 of 2 groups: those undergoing biopsy and those not. The group undergoing biopsy was divided into 2 subgroups, patients with positive skin biopsy findings and those with negative findings. A "positive" biopsy finding was recorded when the pathology report stated "diagnostic for GVHD" if no histologic grade, or a grade of 2 or higher, was given. The patients were also divided into treated and nontreated groups. Treatment was defined as initiation of anti-GVHD therapy at or after the first clinical diagnosis of acute GVHD.

The following parameters were then analyzed for each patient category: (1) clinical severity of GVHD (skin and overall) at the time of first diagnosis; (2) the skin biopsy frequency; (3) biopsy result; (4) initiation of anti-GVHD therapy as a result of the diagnosis of GVHD; and (5) the BMT outcome (acute and chronic GVHD, transplant-related mortality, and overall event-free survival).

STATISTICAL ANALYSIS

We used χ2 and t tests where appropriate to calculate the statistical significance of the GVHD findings. Survival analysis using the Kaplan-Meier method was performed to estimate overall and event-free survival (with death or relapse as events) and the actuarial probabilities of transplant-related mortality and GVHD.

RESULTS
PATIENT POPULATION PROFILE

A total of 187 patients underwent allogeneic BMT between January 1, 1994, and June 30, 1997, at Vancouver General Hospital. Relevant information was available on all patients. In the early period after BMT (ie, within 30 days), 88 patients developed a rash suggestive of acute GVHD. The analysis centers on these patients. There were 54 men and 34 women, with a median age of 40 years (range, 20-55 years). The indications for BMT were as follows: chronic myeloid leukemia (n=21), acute myelogenous leukemia (n=21), myelodysplastic syndrome (n=13), multiple myeloma (n=10), acute lymphoblastic leukemia (n=9), non-Hogkin lymphoma (n=9), and other (n=5). The source of marrow was as follows: unrelated donor (n=48), HLA-matched sibling (n=39), and mismatched sibling (n=1). Prophylaxis for GVHD was with cyclosporine and methotrexate (n=69) or T-cell depletion of the allograft and cyclosporine (n=19).

BIOPSY FREQUENCY

Eighty-eight patients developed skin rashes suggestive of acute GVHD within 30 days of BMT, and 51 (58%) of them underwent skin biopsy to confirm the diagnosis. The average clinical cutaneous stages and overall grades of acute GVHD are given in Table 3. The mean±SEM skin stage (1.71±0.18) and overall grade (2.02±0.12) of the group undergoing biopsy did not differ significantly from that of the group not undergoing biopsy (1.41±0.11 and 1.84±0.12, respectively; P>.05, t test). All patients with skin stage 4 (n=2) and 10 of 13 patients with stage 3 GVHD underwent biopsy, while the biopsy rates for stages 1 and 2 were lower (44% and 67%, respectively). The group undergoing biopsy was more likely to receive GVHD treatment than the group without biopsy (80% vs 57%; P<.02, χ2 analysis [Table 3]).

CLINICAL CORRELATION OF SKIN BIOPSY RESULTS

The results of skin biopsy were positive in 27 patients and negative in 24. As outlined in Table 4, the 2 subgroups did not differ in their average clinical severity of acute GVHD, either in skin-specific stages or in overall grades. The percentage of positive biopsy findings for stages 3 and 4 (50%) was no higher than the lower clinical stages of cutaneous GVHD (57% and 52%, respectively). More importantly, the patients with positive or negative biopsy findings did not differ in their likelihood to receive GVHD treatment. Specifically, 85% of the patients with a positive biopsy finding received anti-GVHD therapy compared with 75% of the patients with a negative finding (P>.05, χ2 analysis).

TEMPORAL RELATIONSHIP BETWEEN SKIN BIOPSY AND INITIATION OF GVHD THERAPY

Of the 51 patients who underwent skin biopsy, 41 received treatment for GVHD. However, in most of the treated cases, therapy began before the biopsy results were available (29 patients) or even before biopsies were performed (7 patients). The biopsy results were available for only 5 patients at the time of therapy initiation. For 3 of these 5 patients, the therapy was initiated when there was clinical deterioration of the GVHD (1 with new development of diarrhea and 2 with worsening of skin GVHD from stage 1 or 2 to stage 3 or 4). Only 2 patients (3%) received treatment for GVHD after (presumably because of) positive skin biopsy findings in the absence of other independent indications for GVHD therapy. Both patients had histologic grade 2, clinical stage 2 skin GVHD and overall grade I GVHD when the therapy was initiated.

RELATIONSHIP BETWEEN CLINICAL SEVERITY OF GVHD AND THERAPEUTIC INTERVENTION

Sixty-two of the 88 patients with skin rashes before day 30 following BMT received treatment for GVHD around the time of diagnosis. Patients in the treated group, compared with the nontreated group, had significantly higher skin GVHD stages (1.67±0.10 vs 1.35±0.12; P<.05) and overall GVHD grades (2.19±0.10 vs 1.35±0.10; P<.001). Indeed, all patients with overall III (n=16) and IV (n=3) GVHD grades were treated. Most of the overall grade II patients (79%; n=42) also received GVHD therapy. In contrast, only 37% of the patients with mild clinical overall GVHD (grade I, n=27) received GVHD therapy.

OUTCOME OF BMT ACCORDING TO BIOPSY STATUS AND RESULT

The peak clinical severity of acute GVHD before day 100 post-BMT was analyzed according to whether a skin biopsy was performed and, if so, according to the biopsy findings (Table 5 and Table 6). In addition, actuarial probabilities of acute (at 100 days) and chronic (at 1 year) GVHD, transplant-related mortality (at 3 years), and event-free and overall survival (at 3 years) for the patient groups are given. There was no statistically significant difference in any of the outcomes analyzed.

COMMENT

In this retrospective study we set out to determine the frequency and clinical utility of skin biopsy within 30 days of allogeneic BMT in the management of patients with a suspicious rash. We found that more than half of patients suspected to have skin GVHD underwent skin biopsy to confirm the diagnosis, and that the biopsy rates were related to clinical severity of the rash, with moderate to severe rashes more likely to be tested. The biopsy findings, however, did not correlate with the clinical severity of skin and overall GVHD, nor did they have any impact on the likelihood for the patients to receive GVHD therapy. Furthermore, the outcome of BMT was not influenced by whether a biopsy was done or by the result.

Most patients undergoing biopsy who were treated for GVHD had therapy initiated before the biopsy was performed or before the results were available. Indeed, skin biopsy findings were the reason for initiation of GVHD therapy in only 3% of all treated patients.

In our patient population, the decision to treat GVHD was dependent on clinical severity of the GVHD rather than on the skin biopsy findings. Similar results were recently reported by another group of investigators who demonstrated that "rush" skin biopsies did not have any clinically significant relevance in the management of patients with GVHD after BMT.22 Specifically, of 36 biopsy specimens processed on an urgent basis, only 2 resulted in an immediate change in management.

In summary, skin biopsy findings within 30 days of allogeneic BMT do not correlate with clinical severity of GVHD, and skin biopsy seems to play only a minor role in directing a physician to initiate therapy for acute GVHD. Therefore, we suggest that the routine practice of subjecting patients to such biopsies could be abandoned without compromising their care.

Back to top
Article Information

Accepted for publication December 22, 1999.

Presented in part at Resident's Forum, American Academy of Dermatology annual meeting, New Orleans, La, March 21, 1999.

We would like to thank Daphne Brockington, CCHRA[C] and Janet Nitta, CCHRA[C], for assistance in searching the BMT Database and performing the bone marrow transplantation outcome analysis.

Reprints: Jason K. Rivers, MD, FRCPC, 835 West 10th Ave, Vancouver, British Columbia, V5Z 4E8, Canada (e-mail: jrivers@interchange.ubc.ca).

References
1.
Deeg  HJStorb  R Graft-versus-host disease: pathophysiological and clinical aspects. Ann Rev Med. 1984;3511- 24Article
2.
Glucksberg  HStorb  RFefer  A  et al.  Clinical manifestations of graft-versus-host disease in human recipients of marrow from HLA-matched sibling donors. Transplantation. 1984;18295- 304Article
3.
Beatty  PGClift  RAMickelson  EM  et al.  Marrow transplantation from related donors other than HLA-identical siblings. N Engl J Med. 1985;313765- 771Article
4.
Vogelsang  GBHess  ADSantos  GW Acute graft-versus-host disease: clinical characteristics in the cyclosporin era. Medicine (Baltimore). 1989;67163- 174
5.
Deeg  HJCottler-Fox  MBurakoff  SJedDeeg  HJedFerrara  JedAtkinson  Ked Clinical spectrum and pathophysiology of acute graft-versus-host disease. Graft-vs.-Host Disease: Immunology, Pathophysiology, and Treatment New York, NY Marcel Dekker1990;311- 336
6.
Saurat  JH Cutaneous manifestations of graft-versus-host disease. Int J Dermatol. 1981;4249- 256Article
7.
Thomas  EDStorb  RClift  RA  et al.  Bone marrow transplantation. N Engl J Med. 1995;292832- 843Article
8.
Przepiorka  DWeisdorf  DMartin  P  et al.  Consensus conference on acute GVHD grading. Bone Marrow Transplant. 1995;15825- 828
9.
Storb  RPrentice  RLSullivan  KM  et al.  Predictive factors in chronic graft-versus-host disease in patients with aplastic anemia treated by marrow transplantation from HLA-identical siblings. Ann Intern Med. 1983;98461- 466Article
10.
Storb  RPrentice  RLBuckner  CD  et al.  Graft-versus-host disease and survival in patients with aplastic anemia treated by marrow grafts from HLA-identical siblings: beneficial effect of a protective environment. N Engl J Med. 1983;308302- 307Article
11.
Woodruff  JMEltringham  JRCasey  HW Early secondary disease in the rhesus monkey, I: a comparative histopathologic study. Lab Invest. 1969;20499- 511
12.
Woodruff  JMButcher  WIHellerstein  LJ Early secondary disease in rhesus monkey, II: electron microscopy of changes in mucous membrane and the external epithelia as demonstrated in the tongue and lip. Lab Invest. 1972;2785- 98
13.
Slavin  RESantos  GW The graft-versus-host reaction in man after bone marrow transplantation: pathology, pathogenesis, clinical features and implications. Clin Immunol Immunopathol. 1973;1472- 498Article
14.
Woodruff  JMHansen  JAGood  RASantos  GWSlavin  RE The pathology of the graft-versus-host reaction (GvHR) in adults receiving bone marrow transplants. Transplant Proc. 1976;8675- 684
15.
Sale  GELerner  KGBarker  EAShulman  HMThomas  ED The skin biopsy in the diagnosis of acute graft-versus-host disease in man. Am J Pathol. 1977;89621- 633
16.
Hymes  SRFarmer  ERLewis  PGTutschka  PJSantos  GW Cutaneous graft-versus-host reaction: prognostic features seen by light microscopy. J Am Acad Dermatol. 1985;12468- 474Article
17.
Elliott  CJSloane  JPSanderson  KVVincent  MShepherd  VPowles  R The histological diagnosis of cutaneous graft versus host disease: relationship of skin changes to marrow purging and other clinical variables. Histopathology. 1987;11145- 155Article
18.
Kohler  SHendrickson  MRChao  NJSmoller  BR Value of skin biopsies in assessing prognosis and progression of acute graft-versus-host disease. Am J Surg Pathol. 1997;21988- 996Article
19.
Drijkoningen  MDe Wolf-Peeters  CTricot  GDegreef  HDesmet  V Drug-induced skin reactions and cutaneous graft-versus-host reaction: a comparative immunohistochemical study. Blut. 1988;5669- 73Article
20.
Sviland  LOearson  ADJEastham  EJHamilton  PJProctor  SJMalcom  AJ Histological features of skin and rectal biopsy specimens after autologous and allogeneic bone marrow transplantation. J Clin Pathol. 1988;41148- 154Article
21.
Darmstadt  GLDonnenberg  ADVogelsang  GBFarmer  ERHorn  TD Clinical, laboratory, and histopathologic indicators of the development of progressive acute graft-versus-host disease. J Invest Dermatol. 1992;99397- 402Article
22.
Barksdale  SKOberlender  SABarnhill  RL "Rush" skin biopsy specimens in a tertiary medical center: diagnostic yield and clinical utility. J Am Acad Dermatol. 1998;38548- 554Article
×