Interface Dermatitis Timothy McCalmont, MD University of California, San Francisco

At the beginning of the 20th century, a diagnosis of dermatitis was fashioned largely by clinical observation. The clinical pattern observed was matched to the patterns known to the mind of the beholder, and a diagnosis was rendered based upon the closest fit, in the opinion of the observer. Many dermatologists divided inflammatory conditions into "hyperemias," characterized by an "abnormal flux of blood" with lesser inflammatory changes, and "inflammations," in which erythema might also be present but in which inflammation was overt. The infectious nature of many conditions was poorly understood, and thus there was not clear distinction between infectious cutaneous diseases and their strictly dermatitic counterparts. It was clear even to early dermatologists that many seemingly unrelated disorders were sometimes (or perhaps often) indistinguishable, and microscopical examination of skin biopsies began to be utilized as a means to obtain unique findings that would point to a specific diagnosis. In this first conception of dermatopathology, the clinician and pathologist were usually the same individual. Only much later did the present paradigm of an independent pathologist/dermatopathologist evolve. Conceived by Wallace Clark and refined and disseminated by Bernard Ackerman, pattern analysis has moved to the fore as the favored method for the classification and diagnosis of dermatitis, and conventional microscopical assessment has also been utilized as a means to gain insight into the pathogenic mechanisms that underlie these diseases. Now instead of erythemas and inflammations, we ponder whether a dermatitis should be considered spongiotic or psoriasiform or vacuolar or lichenoid in nature, and whether a blistering eruption shows intraepidermal or subepidermal vesiculation, and we further wonder about the exact structural antigen that is associated with the eruption. This discussion will briefly review insights into several diseases that are characterized by changes involving the interface between the epidermis and dermis, a collection of inflammatory diseases known simply as interface dermatitis. Interface dermatitides can be broadly characterized as vacuolar or lichenoid with respect to their microscopical pattern, and as acute or chronic in nature. Vacuolar interface dermatitis is characterized by vacuolar alteration of the lowermost epidermis, generally in concert with a sparse or modest perijunctional infiltrate. Vacuolar alteration represents keratinocytotoxic alteration of the epidermis by a cell-mediated lymphocytic attack. Vacuolar alteration may be acute or chronic in nature (discussed immediately below). Lichenoid dermatitis is characterized by a dense perijunctional infiltrate in association with vacuolar changes. Because of the density of the perijunctional infiltrate, sometimes vacuolar alteration may be less obvious in a lichenoid process. It is important to note that there is overlap between things vacuolar and things lichenoid. The purpose of stratification of a process as vacuolar or lichenoid is merely to enable an algorithmic approach to differential diagnosis to be applied. Generally, the determination as to whether an interface eruption is relatively acute or relatively chronic can be made based upon histopathological parameters, without exact reference to the clinical situation. An acute eruption is typified by little alteration to the superjacent stratum corneum, as well as an unaltered (non-fibrotic) papillary dermis. A chronic eruption is typified by overlying hyperkeratosis and/or parakeratosis, as well as alteration of the epidermal thickness (with either acanthosis or atrophy). Furthermore, the papillary dermis is typically expanded or altered by fibrosis or sclerosis in association with a chronic interface eruption. The categories of greatest value in algorithmic differential diagnosis include acute vacuolar interface dermatitis, such as erythema multiforme, fixed drug reaction, or graft-versus-host disease; chronic vacuolar interface dermatitis, such as discoid lupus erythematosus or dermatomyositis; and (chronic) lichenoid dermatitis, such as lichen planus, lichen nitidus, lichen striatus, or a lichenoid drug reaction. In the paragraphs that follow, 3 eruptions that serve as the basis for today's illustrative presentation of interface dermatitis will be reviewed in greater detail. Erythema Multiforme (EM) In the distant past, erythema multiforme was (poorly) defined as an acute inflammatory disease characterized by reddish macules, papules, or even "tubercles," which occasionally could become vesicular or bullous and which varied in size and shape—i.e. were "multiform" in clinical character. The etiology was commonly thought to be secondary to exposure to a toxin, perhaps derived from food. Clinical patterns such as erythema iris, erythema annulare, and erythema bullosum were often interpreted as variants within the broad and nebulous clinical spectrum accepted as EM. Case 1: 68-year-old female with recurrent eruption 2-3 years, with recent flare of scattered tender to painful, bilateral and asymmetrical discrete, smooth tumid papules and patches (to several cm.) on torso, limbs, and fingers. She is on several medications; CBC normal, ESR-38 (one month ago); 5 mm punch at center of tender 1 1/2 cm disc right inner knee. Slide 1 Click to view with ImageScope Click to view with a Web-Based Viewer By the early 1960s, a clinical pattern characterized by extensive involvement of the mucous membranes had been described by Stevens and Johnson. At that time, there was not a clear distinction between bullous pemphigoid and EM, and the former was often considered a chronic variant of the latter. Descriptions of microscopical findings were vague and included mention of spongiosis and intracellular edema above a lymphocytic infiltrate, with variable mention of changes of "vasculitis." By the early 1970s, the confusing concept of "epidermal" and "dermal" and "mixed" patterns of EM was in play. Some "dermal" examples of EM were probably misclassified examples of urticarial dermatitides, while others were merely a reflection in the zonal variation in microscopical findings within EM lesions (depending upon whether the central necrotic area or the edematous peripheral area of a lesion had been sampled). By the mid to late 1970s, it was generally established that EM represented a type of dermatitis characterized by "hydropic degeneration" or vacuolar change at the epidermal-dermal junctional zone. At present, it is currently accepted that EM is an acute form of vacuolar interface dermatitis, that Stevens-Johnson syndrome represents a clinical variant with pronounced mucosal involvement, and that toxic epidermal necrolysis (TEN) represents EM with extensive, confluent epidermal necrosis. EM minor refers to cases with limited cutaneous disease, typically presenting in papulovesicular fashion and stereotypically presenting with targetoid lesions accentuated on the acra. EM major refers to cases with extensive and/or mucosal involvement, including both Stevens-Johnson syndrome and TEN. Histopathologically, EM presents with the pattern of an acute vacuolar interface dermatitis, with vacuolar alteration at the junction below a normal or near-normal epidermis, and with a perijunctional lymphocytic infiltrate that tends to be restricted to the superficial dermis and tends to be of only modest density. Typically, the infiltrate in EM is strictly lymphocytic in composition. Admixed eosinophils may be encountered, particularly in examples of drug-induced disease. A link between herpes simplex labialis and recurrent EM had been observed clinically, and the inception of antiviral therapy provided further evidence of a relationship when it was observed that acyclovir treatment suppressed the subsequent development of EM lesions in such patients. Clinicians refer to such cases as "herpes-associated EM" or HAEM. Subsequent analysis of paraffin-embedded tissue after PCR expansion confirmed the presence of herpes simplex virus genomic DNA within many EM lesions, and it is presumed that the cytotoxic (vacuolar) pattern represents a specific injury induced by virally-altered keratinocytes. The majority of cases of EM are now thought to be secondary to herpes virus infection, including some cases in which there is no clinical evidence of primary herpetic labialis or genitalis (occult primary disease), and also including some cases with extensive mucosal disease (Stevens-Johnson). As alluded to above, EM can also be induced by medications, and the cause of some cases remains unknown, despite clinical investigation. Fixed drug reaction Fixed drug reaction was originally introduced as a term by Brocq in 1894, after the development of an erythematous pigmented localized eruption after the ingestion of antipyrine. By the 1970s it was well documented that fixed drug reaction represented an interface dermatitis of vacuolar type that could be identified microscopically on the basis of vacuolar interface change, necrosis of individual keratinocytes, and a mixed infiltrate including granulocytes and admixed melanophages. Case 2: 56-year-old male with 2 week pruritic, semi-annular eruption on trunk. Slide 2 Click to view with ImageScope Click to view with a Web-Based Viewer Clinically, the "fixed" nature of a fixed drug reaction stems from the fact that clinical lesions are topographically fixed—lesions recur at exactly the same site when rechallenge with the causative medication occurs. In the first occurrence of a fixed drug reaction, a stage undocumented (and perhaps undocumentable) clinically, the presentation is as erythematous plaques. Just like erythema multiforme, papulovesicular changes can be observed within lesional skin, and sometimes clear vesicular or bullous changes will occur. With repetition, considerable post-inflammatory change occurs, which is unsurprising given the restricted clinical distribution of the inflammatory insult. Thus, in repeat episodes, plaques tend to assume a reddish-brown or even a violaceous hue. In some patients, the distribution of disease will become somewhat greater with each "rechallenge" with the causative agent. Rarely, a generalized fixed drug reaction (the "fixed" equivalent of TEN) can occur. Histopathologically, the primary pattern of a fixed drug reaction is that of an acute vacuolar interface dermatitis, with vacuolar alteration of the junctional zone below a relatively normal-appearing basket-weave stratum corneum. The infiltrate tends to be substantial, relatively deep in distribution, and clearly mixed, with eosinophils and/or neutrophils present near the junctional zone, and with numerous melanophages present in the papillary dermis as well. Much like EM, striking vacuolar alteration can eventuate with confluent necrosis and/or subepithelial vesiculation, yielding a bullous microscopical pattern. Some examples of fixed drug reaction present with ballooning (intracellular edema) of keratinocytes, creating a constellation that includes a viral vesicular dermatitis in the differential diagnosis. Clinical and histopathological findings suggest that drug ingestion elicits cytotoxic destruction of keratinocytes by lymphocytes, and presumably a drug-derived determinant expressed on the keratinocyte surface provokes the process. Recent immunoperoxidase evidence indicates a cytokine-mediated keratinocytotoxic process in which gamma interferon plays a role. Lichen planus (and hypertrophic lichen planus) In general, diseases that present with a lichenoid microscopical pattern represent chronic forms of interface dermatitis. A number of different maladies can be considered in the differential diagnosis when a lichenoid configuration is encountered, including lichen planus (LP), lichen striatus, and even mycosis fungoides. LP, the prototypical lichenoid disorder, represents a relatively common disease with highly variable findings. The diversity comes from a wide spectrum of clinical and histopathological findings. Clinically, LP most common presents as small pruritic polygonal papules that are accentuated on the acra or genitalia. Variant clinical forms include LP that presents with distinctive clinical morphology, such as annular, atrophic, vesicular, or erosive LP, as well as LP that presents with involvement restricted to a specific site, such as ungual LP or follicular LP (lichen planopilaris). Case 3: 82-year-old female with 2 cm red nodule. "Consider SCC, BCC, or trauma" Slide 3 Click to view with ImageScope Click to view with a Web-Based Viewer Histopathologically, LP presents with a dense bandlike lymphocytic or lymphohistiocytic infiltrate that is typically restricted to the papillary dermis. The afflicted epidermis shows a spectrum of chronic alterations, including slight hyperplasia, hypergranulosis, and compact orthohyperkeratosis, typically with little if any parakeratosis present. There is typically clear vacuolar change at the junction; when pronounced, it may appear that the basal epidermis has been lost and that squamous keratinocytes of the stratum spinosum are present abutting the dermis, a configuration known as "squamotization" of the basal layer. Variant microscopical forms include atrophic lichen planus, in which the epidermis is attenuated rather than amplified, and vesiculobullous LP, in which basal vacuolar alteration is present to a degree that yields subepithelial separation. Hypertrophic LP represents an important variation that is distinctive both clinically and histopathologically. The clinical lesions are distinctive in that they present in nodular (rather than papular) fashion and are often restricted to the lower extremities. The clinical lesions of hypertrophic LP may lack associated pruritus. Histopathologically, the lesions of hypertrophic LP show the characteristic vacuolar reaction, squamotization, and lichenoid infiltrate that would be expected in association with conventional LP, but are distinctive for the concurrent presence of irregular epithelial hyperplasia. In a sense, hypertrophic LP is merely conventional LP with concomitant pseudocarcinomatous hyperplasia of the surface epithelium. Because of the pseudocarcinomatous alteration, lesions of hypertrophic LP can be misconstrued as either squamous cell carcinoma or keratoacanthoma. This challenging differential diagnosis is rendered even more difficult by the fact that keratoacanthomas and carcinomas have been documented to arise within lesions of hypertrophic LP. In short, in the evaluation of a potential case of hypertrophic LP, one should also consider the possibility of carcinoma or keratoacanthoma in the differential diagnosis, including the possibility of a keratoacanthoma arising within a lesion of hypertrophic lichen planus. References: Darragh TM; Egbert BM; Berger TG; Yen TS. Identification of herpes simplex virus DNA in lesions of erythema multiforme by the polymerase chain reaction. J Am Acad Dermatol,1991 Jan, 24(1):23-6. Brice SL; Leahy MA; Ong L; Krecji S; Stockert SS; Huff JC; Weston WL. Examination of non-involved skin, previously involved skin, and peripheral blood for herpes simplex virus DNA in patients with recurrent herpes-associated erythema multiforme. J Cut Pathol, 1994 Oct, 21(5):408-12. Aslanzadeh J; Helm KF; Espy MJ; Muller SA; Smith TF. Detection of HSV-specific DNA in biopsy tissue of patients with erythema multiforme by polymerase chain reaction. Br J Dermatol, 1992 Jan, 126(1):19-23. Molin L. Oral acyclovir prevents herpes simplex virus-associated erythema multiforme. Br J of Dermatol, 1987 Jan, 116(1):109-11. Orfanos CE, Schaumburg-Lever G, Lever WF. Dermal and epidermal types of erythema multiforme. Arch Dermatol, 1974, 109:682-88. Korkij W; Soltani K. Fixed drug eruption. A brief review. Archives of Dermatology, 1984 Apr, 120:520-4. Kossard S et al: Hypertrophic lichen planus-like reactions combined with infundibulocystic hyperplasia: pathway to neoplasia. Arch Dermatol. 2004 Oct;140(10):1262-7. Chave TA, Graham-Brown RA: Keratoacanthoma developing in hypertrophic lichen planus; Br J Dermatol. 2003 Mar;148(3):592.