Cutaneous immunotorescence study of erythema multiforme: Correlation with light microscopic patterns and etiologic agents

Cutaneous immunotorescence study of erythema multiforme: Correlation with light microscopic patterns and etiologic agents

I I I III I I Cutaneous immunofluorescence study of erythema multiforme: Correlation with light microscopic patterns and etiologic agents Marian ...

3MB Sizes 0 Downloads 30 Views






Cutaneous immunofluorescence study of erythema multiforme: Correlation with light microscopic patterns and etiologic agents Marian C. Finan, M.D., and Arnold L. Schroeter, M.D.

Rochester, MN Direct immunofluorescence microscopy was positive in 88% of forty-one skin biopsy specimens from thirty-four patients with the clinical diagnosis of erythema multiforme. The most common finding, present in 67% of positive specimens, was the cytoid body, or fluorescent keratinocyte, which stained most often with IgM (homogeneously) or with C3 (speckled). Other findings included basement membrane zone (BMZ) fluorescence, primarily with fibrinogen and C3, and vascular fluorescence, most commonly C3 in a granular pattern. Correlation of direct immunofluorescent and light microscopic findings revealed that (1) the fluorescent keratinocyte was prevalent only in epidermal and mixed patterns, correlating with the eosinophilic necrotic keratinocyte by light microscopy, and (2) vascular fluorescence was most prominent in dermal forms. Herpes simplex-associated erythema multiforme showed exclusively a mixed histologic pattern, whereas the drug-related form was primarily epidermal. (J AM ACADDERr,tATOL 10:497-506, 1984.)

Although erythema multiforme is a relatively common disorder, accounting for up to 1% of outpatient visits, ~ and has been clinically recognized for more than 100 years since its initial description by Hebra in 1866, 2 clues to its pathogenesis have been slow to unfold. Over the past several years, evidence has continued to accrue suggesting that erythema multiforme is an immunologically mediated disease process that may be precipitated by a wide variety of antigenic stimulants. In 1967, Shelleya implicated such a hypersensitivity mechanism by the production of classic lesions of erythema multiforme through skin testing with

From the Department of Dermatology, Mayo Clinic and Mayo Foundation. Accepted for publication Aug. 10, 1983. Reprint requests to: Dr. Marian C. Finan, e/o Section of Publications, Mayo Clinic, Rochester, MN 55905.

herpesvirus. Although initial direct immunofluorescence studies of cutaneous lesions were negative, 4-6 subsequent studies demonstrated positive immunofluorescence of the superficial cutaneous vasculature from affected skin, involving 33% to 94% of specimens. 7-9 Circulating immune complexes, as detected by the 125I-Clq binding assay, 7'1~ the Raji cell technic, 11 and the cryoprecipitin I~ and monoclonal rheumatoid factor 8 assays, have been demonstrated. Depressed levels of serum complement lz and also decreased complement in blister fluid taken from bullous lesions of erythema multiforme 13 have been observed. Lozada et al, 14 in assessing the immunologic status of thirteen patients with oral erythema multiforme, noted depressed responses to delayed hypersensitivity skin testing to common antigens and lowered in vitro lymphocyte transformation to antigens and phytohemagglutinins. In summary,



Finan and Schroeter

the body of evidence implicating immunologic mechanisms in erythema multiforme is becoming ever more compelling. In the present retrospective study, direct immunofluorescence microscopic technics were employed to demonstrate the presence of immunoglobulins, complement (C3), and fibrinogen in forty-one cutaneous biopsy specimens from thirtyfour patients with the clinical diagnosis oferythema multiforme. The aims of this study were fourfold: (1) to confirm and add to previously reported direct immunofluorescence findings in erythema multiforme; (2) to determine whether a meaningful correlation exists between light microscopic patterns and patterns of direct immunofluorescence; (3) to relate findings from direct immunofluorescence with etiologic agents; and (4) to determine whether the above information, taken collectively, can supply further clues to the clinicopathologic relationship between antigenic exposure and the cutaneous expression of the lesion of erythema multiforme. METHODS AND PATIENTS The results of cutaneous immunofluorescence microscopy performed at the Mayo Clinic on all patients with a clinical diagnosis of erythema multiforme who were seen between April, 1979, and May, 1982, were reviewed. A diagnosis of erythema multiforme was based on the clinical picture of an acute, self-limited cutaneous or mucous membrane eruption with characteristic discrete or confluent target lesions. Cases were excluded that were either clinically atypical (i.e., appearance or evolution of lesions more typical of urticaria or other process, with no target lesions) or that showed histologic features distinctive of another specific disease (such as bullous pemphigoid or dermatitis herpetiformis). In addition, any biopsy specimen in which evidence of epidermal regeneration was present, rendering interpretation difficult or inaccurate, was excluded. Of the thirty-four patients who qualified for the study, nineteen (56%) were male and fifteen (44%) were female, and they ranged in age from 4 to 80 years (mean 42 +__20 years). Forty-one cutaneous irnmunofiuorescence specimens were obtained. Biopsy specimens were taken from early lesions of involved skin in most cases and submitted for both direct immunofluorescence and routine histologic study. The exact morphology of biopsy specimens was not specified in all cases. When indicated, however, the specimens

Journal of the American Academy of Dermatology were obtained from the rim of an iris lesion. Cutaneous specimens were obtained by punch biopsy, for which 2% lidocaine hydrochloride was used for anesthesia. In thirty-seven of forty-one specimens submitted for direct immunofluorescence study, the tissue was divided a n d half of the specimen was sent for routine light microscopic study. Four specimens sent for light microscopy were obtained from involved sites anatomically distant from the site chosen for direct immunofluorescence. These four cases were excluded from a correlative analysis of light microscopic and direct immunofluorescence results. Specimens sent for direct immunofluorescence m i croscopy were immediately frozen in liquid nitrogen and stored at - 7 0 ~ C until processed. Unfixed frozen cryostat sections of 3 to 4/zm in width were stained f o r immunoreactants by means of fluorescein-labeled g o a t antihuman conjugates of IgG, IgA, IgM, C3, a n d fibrinogen by a previously described technic. '5 The dilutions and concentrations of conjugate antisera u s e d are summarized in Table I. Biopsy specimens were examined for fluorescence of the dermal blood vessels, basement membrane zone (BMZ), and epidermis. T h e fluorescence was graded on a scale ranging from (weak) to 3 + (strong). Only readings of 1+ or more were regarded as positive. All positive results were photographed. Tissue submitted for light microscopy was fixed in formalin and subsequently stained with hematoxylin and eosin. Additional stains, namely alcian blue, periodic acid-Schiff, and Smith elastic Giemsa, were also used. Each specimen was studied histologically a n d assigned to one of three categories--"epidermal," "mixed," or "dermal" erythema muItiforme'6-18-according to the following light microscopic criteria: Epidermal pattern. Epidermal necrosis was a pronounced feature, with generalized involvement of the entire epithelial surface. Hydropic degeneration of the basal layer, with or without subepidermal vesiculation, was variably present. Dermal changes were minimal i n all cases, and little or no lymphocytic infiltrate was present about the superficial blood vessels (Fig. 1). Dermal pattern, Epidermal necrosis, either focal o r generalized, was notably absent. A significant lymphocytic perivascular infiltrate was notable within the dermis; eosinophils and neutrophils were variably present. Papillary dermal edema was generally present; it was occasionally pronounced and resulted in subepidermal vesiculation (Fig. 2). Mixed pattern. Both dermal and epidermal changes were prominent. Within the epidermis, a few (<10) necrotic keratinocytes were present, arranged singly o r

Volume 10 Number 3 March, 1984

Immunofluorescence study of erythema multiforme


~ -(


Z' i '!

Fig. 1. Epidermal pattern of erythema multiforme, demonstrating pronounced epidermal necrosis, hemorrhagic subepidermal bulla formation, and minimal dermal lymphocytic infiltrate. (Hematoxylin-eosin stain; x64.) in clusters, displaying a strongly eosinophilic cytoplasm and pyknotic or absent cellular nuclei. Within the dermis, a perivascular lymphocytic infiltrate was present with papillary edema, sometimes to the extent of subepidermal vesiculation. Lymphocytes were frequently seen along the dermoepidermal junction, often in association with hydropic degeneration of the basal layer of cells (Fig. 3). Classification of the histologic diagnosis into each of the three categories cited above was reproducible, in each case, when reviewed by independent observers. Determination of etiologic agents believed to be involved in the pathogenesis of erythema multi forme was made on the basis of a close temporal relationship between antigenic exposure and outbreak of erythema multiforme--the antigenic exposure preceded or was concurrent with the outbreak. Recurrence of erythema multiforme was in some cases associated with recurrent antigenic exposure. RESULTS Positive direct immunofluorescence findings were present in thirty-six of forty-one biopsy specimens (88%), corresponding to twenty-nine of the thirty-four patients (85%). Five specimens were negative. A summary of positive direct im-

Table I. Summary of conjugate antisera dilutions and concentrations used in the direct immunofluorescence technic

Antisera L Dllution ~ Concentration(~g/ml) IgG IgM IgA C3 Fibrinogen

1: 5 1:4 1: 5 1: 40 1: 140

80 I00 80 40

munofluorescence findings is presented in Table II. The most common positive finding, present in twenty-four or 67% of positive specimens, was the cytoid body, or fluorescent necrotic keratinocyte, which was present either just above the dermoepidermal junction or scattered within the epidermis, in some cases as high as the stratum granulosum. In no specimens were cytoid bodies found below the BMZ. These rounded bodies stained uniformly with IgM, IgA, and fibrinogen conjugates; with the C3 conjugate, a peppered or speckled quality was present (Figs. 4 and 5). The immunoreactants most commonly found were IgM (nineteen cases)


Journal of the American Academy of Dermatology

Finan and Schroeter

Fig. 2. Dermal pattern of erythema multiforme, demonstrating papillary edema and lymphocytic infiltrate in the papillary and reticular dermis, with perivascular accentuation. Note complete absence of epidermal necrosis. (Hematoxylin-eosin stain; x 100.) Table II. Direct immunofluorescence results in thirty-six positive biopsy specimens (of forty-one studies) from lesions of erythema multiforme

T Positive staining

Cytoid bodies Basement membrane zone Blood vessels


Total No.

% of positive specimens

24 23 18*

67 64 50


13 12 16

(Ni" p~





6 18 6

19 2 5

5 0 0

* In sixteen, papillary vessels alone were positive; in one, vessels in both papillary and reticular dermis were positive; in one other, vessels only in reticular dermis were positive.

and C3 (thirteen), followed by fibrinogen (six) and IgA (five). Fluorescence, in various patterns at the BMZ, with one or more conjugates, was the next most frequent finding--in twenty-three (64%) positive specimens. Fibrinogen was found in eighteen specimens and was described as a thick, shaggy pattern in the majority of cases. Granular deposition of C3 was noted along the BMZ in twelve specimens. IgM was the only immunoglobulin found along the BMZ, and it was present in only two cases. Vascular fluorescence was found in eighteen (50%) of the positive specimens, most com-

monly C3 in a granular pattern (Fig. 6). IgM and fibrinogen were also present, although less frequently. The combination of cytoid bodies plus BMZ immunofluorescence was the most common pattern of cutaneous immunofluorescence found overall, being present in eleven specimens. Other patterns of fluorescence, combining cytoids, BMZ, or superficial dermal blood vessels, were less often found (Table III).

Correlation with dermatopathologie patterns. Of forty-one cutaneous biopsy specimens submitted for light microscopy and classified according to the histologic criteria cited in the "Methods and

Volume 10 Number 3 March, 1984

Immunofluorescence study of elythema multiforme 501

Fig. 3. Mixed pattern of erythema multiforme, demonstrating combined dermal and epidermal changes, with both lymphocytic infiltrates of the dermis and dermoepidermaI junction and focal hyaline necrosis of keratinocytes. (Hematoxylin-eosin stain; x 100.) Patients" section, most (twenty-five, or 62%) were of the " m i x e d " pattern. Ten specimens (24%) were classified as "dermal" erythema multiforme, and six (14%) were of the "epiderm a l " variety. A correlation of direct immunofluorescence and light microscopic findings could be made in thirty-four specimens in which tissue sent for both technics was obtained simultaneously from the same lesional skin. Table IV summarizes the immunofluorescence findings according to corresponding light microscopic patterns (epidermal, mixed, or dermal). Of the six specimens showing the epidermal pattern of erythema multiforme, all demonstrated cytoid bodies or fluorescent keratinocytes. Less commonly, in three specimens (50%), fibrinogen was present along tbe BMZ; in only one specimen (17%) was vascul~ fluorescence found (flbrinogen). In mixed patterns of erythema multiforme, cytoid bodies were found in the majority of specimens (fifteen biopsies of the twenty-three in this category, or 65%), although less frequently than in the epidermal form, as noted above. BMZ

fluorescence, consisting of fibrinogen alone or in combination with C3, was found in fourteen biopsies, or 61% of specimens. Vascular fluorescence with granular C3, alone or in combination with fibfinogen, was found less c o m m o n l y - - n i n e specimens, or 39%. Direct immunofluorescence patterns found in eight patients with dermal forms of erythema multiforme by light microscopy revealed that vascular fluorescence was significantly more common in this group than in the two others (present in five or 63% of specimens). In addition, IgM, C3, and fibrinogen were present, singly or in various combinations, within the papillary dermal vasculature. BMZ fluorescence was found in only two specimens (25%) (consisting of C3 in both), and cytoid bodies were notably rare (one specimen, 13%). No dermoepidermal junctional fibrinogen was found in the dermal group, in contrast to its prevalence in the epidermal (three specimens, 50%) and mixed (twelve specimens, 52%) groups. Negative immunofluorescence studies were significantly more prevalent in the dermal group than in the two others, accounting for three (38%) of


Journal of the American Academy of Dermatology

Finan and Schroeter

Table HI. Patterns of cutaneous immunofluorescence in thirty-six positive biopsy specimens from lesions of erythema multifomae with one or more immunoreactants Positive fluorescence

Cytoid bodies + B M Z

Cytoid bodies + BMZ + blood vessels Blood vessels only Cytoid bodies only Cytoid bodies + blood vessels BMZ z- blood vessels BMZ only Total


% of total positive





5 4 4 4 3 36

14 11 11 11 8 100

biopsy specimens, compared with none of six specimens in epidermal erythema multiforme and two (9%) of specimens in the mixed form. Correlation with etiologic agents. The etiologic agents involved in the pathogenesis of erythema multiforme in the thirty-four patients were reviewed retrospectively and correlated with the results of direct immunofluorescence studies. In the majority of cases (eighteen patients, 53%), the etiologic agent was not identified or could not be specifically implicated in pathogenesis among multiple potential offending antigens. Herpes simplex virus (HSV)-related erythema multiforme accounted for eight cases (herpes labiaIis, most commonly, herpes genitalis, less often), or 24% of the total. Drugs accounted for six cases, or 18% of the total, and the specific agents implicated included ampicillin, phenothiazine, sulfonamide, oxacillin, and tetracycline. In two cases, infections other than HSV were causally linked to erythema multiforme, consisting of a bacterial dental infection in one and a viral upper respiratory tract infection in the other (Table V). Of the twenty-one biopsies submitted for light microscopy in which a causative factor was unknown or unidentified, the majority, fourteen specimens (67%) were of the mixed pattern, while seven (33%) were derma/ by light microscopy. All HSV-related erythema multiforme displayed mixed patterns histologically (ten specimens, 100%), and the two patients with infections other than HSV had the mixed pattern as well. Of all

eight drug-related cases of erythema multiforme, six (75%) were of the epidermal pattern by light microscopy and one (12.5%) each was of the derma/ or mixed pattern. Interestingly, all biopsy specimens that displayed an epidermal pattern by light microscopy were related to drugs etiologically (Table V); in no instance was epidermal erythema multiforme believed to have been caused by infection or unknown agents. DISCUSSION

Our findings are distinct from those of previously published immunofluorescence microscopy studies of erythema multiforme in several respects. First, cytoid bodies or fluorescent keratinocytes were the most frequent finding, and this has not been reported previously. Prior studies of direct immunofluorescence findings in erythema multiforme have emphasized the presence of delicate, granular deposits of C3 alone or in combination with IgM globulin within the microvasculature of the papillary dermis, r-'l with no mention of epidermal or cytoid fluorescence. Second, BMZ fluorescence was relatively prevalent--a finding only occasionally recorded 8"1~ in prior studies. However, it should be noted that many of the immunofluorescence findings in erythema multiforme may be nonspecific, for the deposition of immune reactants in blood vessels or along the BMZ has been seen as a nonspecific reaction pattern to injury. 2~ Third, direct immunofluorescence findings, when correlated with light microscopic results, revealed deposition of immune reactants corresponding closely to the level of injury seen by light microscopy. Fourth, in the correlation of presumed etiologic agents with light microscopic patterns, erythema multiforme related to HSV was exclusively mixed, whereas that related to drugs was primarily epidermal. The prevalence of cytoid bodies or fluorescent keratinocytes by direct immunofluorescence is noteworthy. Cytoid bodies (also known as "Civatte bodies," "colloid bodies," "single cell necrobiosis," "individual cell keratinization," "sunburn cells," "dyskeratotic cells") 21 have been found in many benign dermatoses, especially those associated with presumed immunologic attack on the epidermis--lichen planus, fixed drug

Volume 10 Number 3 March, 1984

Immtmofluorescence study of erythema multiforme 503

Fig. 4. Epidermal cytoid bodies in erythema multiforme. Homogeneous staining of cytoid bodies with anti-IgM. (x 625.) Fig. 5. Epidermal cytoid bodies in erythema multiforme. Speckled or granular staining of cytoid bodies with anti-C3. (x625.) Fig. 6. Vascular fluorescence with granular C3 in the papillary dermis in erythema multiforme. (x450.) eruptions, the skin lesions of graft-versus-host reactions, and others. 2'-' Fluorescent cytoid bodies seen on immunoftuorescence microscopy are probably identical to the eosinophilic bodies seen by light microscopy2:3,'~ and generally are located at the dermoepidermal junction and in the upper part of the dermis. In the current study, the cytoid body or fluorescent keratinocyte was found above the dermoepidermal junction, and it appeared to have been derived from individually necrotic epidermal cells. This is in contrast to the findings in lichen planus, in which necrotic epidermal keratinocytes are characteristically found at the dermoepidermal junction or in the papillary dermis. These necrotic epidermal keratinocytes may be apoptotic cells as

well, and they appear to be specific for those forms of erythema multiforme in which necrosis of keratinocytes is seen to variable degrees--i.e., the epidermal and mixed forms. Correlation of light microscopic with immunofluorescence findings revealed deposition of immune reactants corresponding closely to the level of light microscopic changes. In dermal patterns, vascular fluorescence was the most common finding; in epidermal patterns, cytoid bodies or fluorescent keratinocytes were present in all specimens; whereas in mixed patterns, approximately equal fluorescence of cytoid bodies and BMZ was noted, with vascular fluorescence somewhat less, suggesting an intermediate position between the dermal


Journal of the American Academy of Dermatology

Finan and Schroeter

Table IV. Summary of cutaneous direct immunofluorescence results in relation to light

microscopic pattern Component (No. positive)

Light microscopic _ pattern

Epidermal Mixed Dermal Other* Total

No. of patients

No. of biopsies

4 18 8 4 34

6 23 8 4 4l

Cytoid bodies



positive (%)

6 21 5 4 36

(100) (91) (63) (100)

[ Fibrin-

Blood vessels














5 14 0 2 21

3 9 1 1 14

3 3 0 1 7

2 2 0 1 5

0 1 0 0 1

0 9 2 1 12

3 12 0 2 17

0 0 3 0 3

0 8 4 2 14

1 2 2 1 6

*In these specimens, correlation between light microscopy and direct immunofluorescence was not valid because of anatomic and temporal variations in biopsy specimens obtained.

Table V. Results o f direct immunofluorescence microscopic studies in erythema multiforme in

relation to presumed etiologic agent Area of positive staining


No. of


No. of biopsies



8 (24) 2 (6) 6 (18)

10 2 8

l0 2 8

4 2 3

8 1 4

8 1 7


18 (53)







34 (100)






Etiologic agent

HSV Other infection Drug

(% of total)

Light microscopic pattern, No. (%)


Mixed: 10 (100) Mixed: 2 (100) Epidermal: 6 (75) Mixed: 1 (12,5) Dermal: 1 (12.5) Mixed: 24 (67) Dermal: 7 (33)

HSV: Herpes simplex virus.

and epidermal forms. These results are consistent with the view originally proposed by MacVicar et al in 196325 that dermal versus epidermal predominance m a y reflect the depth of vascular insult, the dermal being deeper and the epidermal more superficial. It must be emphasized, however, that the study of erythema multiforme is complicated by the evolving nature of the individual lesions, and the histologic findings are dependent on both the stage of evolution of the lesion and the morphologic zone of the target lesion from which the biopsy specimen has been taken. Nonetheless, the light microscopic and direct immunofluorescence findings, albeit transient and evolving, can be meaningfully correlated, as in the current study, when taken from the same area of the same lesion simultaneously.

Prominent epidermal injury, as noted in the mixed forms of erythema multiforme, was largely herpes-related in the current series. Theoretically, epidermal necrosis in mixed patterns, as seen associated with herpes simplex infections, could result from one of two mechanisms: (1) a primary phenomenon, with specific attack of complement and mediators of the inflammation on both the epidermis and the underlying dermis, brought about by herpes-containing immune complexes being deposited in both dermal vessels and the epidermis; or (2) deposition of herpesvirus antibody immune complexes in the dermal vasculature, with activation of complement, leading to severe dermal inflammation and edema. In the latter case, epidermal necrosis would occur as a secondary phenomenon. Orfanos et al ~6 suggested

Volume 10 Number 3 March, 1984

Immunofluorescence study of erythema multiforme 505

that, in erythema multiforme, epidermal changes develop independently of dermal changes; this theory is in accord with a mechanism of primary epidermal attack. On the other hand, Bedi and Pinkus 2~ noted that in up to 60% of bullous and iris lesions, the epidermis is completely necrotic and is indistinguishable from lesions of adult toxic epidermal necrolysis, apparently being secondary to underlying dermal edema and inflammation. In the current series, 100% of the cases demonstrat'ing epidermal erythema multiforme distinguished themselves in being drug-related. Interestingly, prior immunofluorescence studies suggested an epidermotoxic nature of certain drugs. Stein et a127 demonstrated basal cell immuno fluorescence with immunoglobulins and complement in two cases of drug-induced toxic epidermal necrolysis; they postulated that the drug binds to the intercellular epidermal protein and that the basal cells serve as target sites in druginduced toxic epidermal necrolysis. Similarly, Shelley et al 2s demonstrated direct immunofluorescence findings confined to the basal cell layer and two to three adjacent suprabasilar layers in a patient with a fixed bullous eruption due to phenolphthalein. In graft-versus-host disease, which is histologically analogous to erythema multiforme, Merritt et a129 found an IgM antibody in the serum which was cytotoxic for epidermal cells. De Dobbeleer and Achten3~ noted that the dyskeratotic cells in fixed drug eruption are a result of severe epidermal injury, with many features in common with the dyskeratotic ceils found in epidermal erythema multiforme, toxic epidermal necrolysis, and a morbilliform eruption due to penicillin. They noted that whereas these disorders display varying intensity of epidermal injury, their common denominator is the necrotic keratinocyte. Therefore, drug-induced epidermal erythema multiforme may lie within a spectrum of disorders (including fixed drug eruption, lichenoid drug reaction, and others) in which the epidermis is the primary shock organ for immune reaction and toxicity. It appears distinct from the dermal and mixed forms by its notable lack of significant changes in the upper dermis by light microscopy and the striking epidermal fluorescence in all cases, with relative sparing of the underlying tissue. We propose that it be once

again separated from the other forms of erythema multiforme, as Lyell'al had initially done in his description of drug-induced toxic epidermal necrolysis. REFERENCES

1. Hellgren L, Hersle K: Erythema multiforme: Statistical evaluation of clinical and laboratory data in 224 patients and matched healthy controls. Acta Allergol (Kbh) 21:45-51, 1965. 2. Hebra F: Diseases of the skin. (English translation edited by CH Fagge.) London, 1866, New Sydenham Society, vol. 1, p. 285. 3. Shelley WB: Herpes simplex virus as a cause of erythema multiforme. JAMA 201:153-156, 1967. 4. Jordon RE, Beutner EH, Witebsky E, et al: Basement zone antibodies in bullous pemphigoid. J A M A 200: 751-756, I967. 5. Kay DM, Tuffanelli DL: Immunofluorescent techniques in clinical diagnosis of cutaneous disease. Ann Intern Med 71:753-762, 1969. 6. Burnham TK, Fine G, Neblett TR: Immunofluorescent "band" test for lupus erythematosus. Arch Dermatol 102:42-50, 1970. 7. Imamura S, Yanase K, Taniguchi S, et al: Erythema multiforme: Demonstration of immune complexes in the sera and skin lesions. Br J Dermato1102:161-166, 1980. 8. Bushkell LL, Mackel SE, Jordon RE: Erythema multiforme: Direct immunofluorescenee studies and detection of circulating immune complexes. J Invest Dermatol 74:372-374, 1980. 9. Kazmierowski JA, Wuepper KD: Erythema multiforme: Immune complex vasculitis of the superficial cutaneous microvasculature..I Invest Dermatol 71:366-369, 1978. 10. HuffJC, Weston WL, Carr RI: Mixed cryoglobulinemia, t2~I C1q binding and skin immunofluorescence in erythema multiforme. I Invest Demaatol 74:375-377, 1980. 11. Wuepper KD, Watson PA, Kazmierowski JA: Immune complexes in erythema multiforme and the StevensJohnson syndrome. J Invest Dermato174:368-371,1980. 12. Laurell A-B, Lundh B, Malmquist J, Siboo R: Complement components in three pathological sera: Relation to clinical states. Clin Exp Immunol 1:13-23, 1966. 13. Safai B, Good RA, Day NK: Erythema multiforme: Report of two cases and speculation on immune mechanisms involved in the pathogenesis. Clin Immunol Immunopathol 7:379-385, 1977. 14. Lozada F, Spitler L, Silverman S Jr: Results of immunologic testing in patients with erythema multifomae. J Dent Res 59:567-572, 1980. 15. Beutner EH, Chorzelski TP, Jordon RE: Autosensitization in pemphigus and bullous pemphigoid. Springfield, IL, 1970, Charles C Thomas, Publisher, pp. 133-172. 16. Orfanos CE, Schaumburg-Lever G, Lever WF: Dermal and epidermal types of erythema multiforme. Arch Dermatol 109:682-688, 1974. 17. Ackerman AB, Penneys NS, Clark WH: Erythema multiforme exudativum: Distinctive pathological process. Br J Dermatol 84:554-566, 1971. 18. Lever WF, Schaumburg-Lever G: Erythema multiforme,


Journal of the American Academy of Dermatology

Finan and Schroeter

in I-tistopathology of the skin, ed. 5. Philadelphia, 1975,



21. 22.

23. 24.


J. B. Lippincott Co., pp. 122-124. Telner P, Leznoff A, Frank H: A relationship between erythema multiforme and histoplasmosis. J Invest Dermatol 45:135-138, 1965. Miller RAW, Griffitlas WAD: Experimentally induced complement and immunoglobulin deposition along the basement membrane zone (BMZ) and in dermal blood vessels. Br J Dermatol 106:275-279, 1982. Person JR, Rogers RS III: Ashy dermatosis: An apoptotic disease? Arch Dermatol 117:701-704, 1981. Weedon D, Searle J, Kerr JFR: Apoptosis: Its nature and implications for dermatopathology. Am J Dermatopathol 1:133-144, 1979. Ueki H: Hyaline bodies in subepidermal papillae: Immunohistochemical studies in several dermatoses. Arch Dermatol 100:610-617, 1969. Hashimoto K: Apoptosis in lichen planus and several other dermatoses" Intra-epidermal cell death with filamentous degeneration. Acta Derm Venereol (Stockh) 56:187-210, 1976, MacVicar DN, Graham JH, Burgoon CF Jr: Dermatitis

26. 27.


29. 30.


herpetiforrnis, erythema multiforme and bullous pemphigoid: A comparative histopathological and histochemical study. J Invest Demaatol 41:289-299, 1963. Bedi TR, Pinkus H: Histopathological spectrum of erythema multiforme. Br J Dermatol 95:243-250, 1976. Stein KM, Schlappner OLA, Heaton CL, Decherd JWDemonstration of basal cell immunofluorescence in drug-induced toxic epidermal necrolysis. Br J Dermatol 86:246-252, 1972. Shelley WB, Schlappner OLA, Heiss HB: Demonstration of intercellular immunofluorescence and epidermal hysteresis in bullous fixed drug eruption due to phenolphthalein. Br J Dermatol 86:118-125,. 1972. Merritt CB, Mann DL, Rogentine GN: Cytotoxic antibody for epithelial cells in human graft versus host disease. Nature 232:638-639, 1971. De Dobbeleer G, Aehten G: Fixed drug eruption: Ultrastructural study of dyskeratotic cells. Br J I3ermatol 96:239-244, 1977. Lyell A: Toxic epidermal necrolysis: Eruption resembling scalding of skin. Br J Dermatol 68:355-36 I, 1956.

BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the JOURNALOFTHE AMERICANACADEMYOFDERMATOLOGYa r e availableto subscribers (only)for the 1984 issues from the Publisher at a cost of $53.50 ($65.00 international)for volume 10 (January-June)and volume 11 (July-December). Shipping charges are included. Each bound volumecontains a subjectand author index and all advertisingis removed. Copies are shipped within30 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. ContactThe C. V. Mosby Company, CirculationDepartment, l 1830 Westline Industrial Drive, St. Louis, Missouri 63146, USA; phone (80(3)325-4177, ext. 351. Subscriptions must be in force to qllalify. Bound vohlmes are not available in place of a regular journal subscription.