The histology of PR has not attracted much interest due to the fact that since classic PR can be easily diagnosed clinically, a skin biopsy is rarely performed. Indeed the accurate diagnosis of PR is not difficult in the eyes of an experienced clinician, together with a reliable history and adequate follow up. Diagnosis may become difficult in unusual presentations which mimic other dermatological conditions. Cutaneous eruptions related to drug intake or contact allergies may imitate PR clinically. [17-25]

Histopathologic features of PR have largely been thought of as a nonspecific perivascular dermatitis and have even been excluded as a reliable tool for a proper diagnosis by some authors. [21] In contrast, Ackerman et al established a set of criteria that enabled differentiation of PR from other spongiotic diseases except erythema annulare centrifugum. [11] Our study confirms that careful histological examination of PR shows characteristic features which enable separation of the disease from the wastebasket diagnosis of a "non-specific superficial perivascular dermatitis."

Prominent histopathological features of PR include a superficial perivascular dermatitis with multi-focal spongiosis. The parakeratotic mounds and collection of inflammatory cells within the spongiotic foci (simulator of Pautrier's microabscesses) are findings highly suggestive of a diagnosis of PR. These foci of parakeratosis that sometimes house a tad of serum are the histologic equivalent of the pityriasiform scale and they develop consequent to focal spongiosis. Increased keratinocyte proliferation is noted beneath the foci of spongiosis and parakeratosis. Minor features of PR include the presence of scale crusts, acanthosis, and lymphocytes in the epidermis. Inconsistent findings are focal hypogranulosis, dyskeratotic keratinocytes, and intraepidermal red blood cells. The foci of hypogranulosis are observed within spongiotic areas and beneath parakeratotic mounds. In the dermis, a superficial perivascular infiltrate consisting predominantly of lymphocytes is encountered. Variable features in the dermis include red blood cell extravasation and some pigment-laden macrophages.

The results of our study are very similar to early reports of Unna. According to Unna, however, mild thickening of the spinous layer with some mitotic figures, focal hypogranulosis beneath the parakeratotic mounds, and mild edema were less prominent features but they were observed in most of our cases. Unna did not mention lymphocytes within the epidermis and intraepidermal red blood cell extravasation, which were present in 82% and 76% of our cases, respectively.

Our results show differences to the study of Okamoto et al. [26] and Prasad et al. [27] who considered the presence of dyskeratotic cells to be a prominent feature. Vacuolar basal cell alteration, which is observed in skin diseases which predominantly show a lichenoid or interphase inflammatory pattern was not observed in any of our cases. The presence of eosinophils described by Prasad in longstanding lesions of PR was observed in 29 % of our cases but they were very few in number and interstitial in location, compared with cases of allergic contact dermatitis where numerous perivascular and interstitial eosinophils were observed. Delling of the epidermis and hypogranulosis, which were considered to be consistent features in the series by Prasad were seen only in a few of our cases. [27]

Our immunohistochemical results confirm the findings of Sugiura et al. who found a large number of lymphoid cells in the perivascular infiltrate reacting with anti-pan T-cell marker CD3. Sugiura et al. and Neoh et al. suggested, however, that CD4+ cells are predominant over CD8+ cells in a ratio of 2.9. [14,15] Our study showed a mixed infiltrate of helper inducer (CD4 positive) cells and suppressor-cytotoxic cells (CD8 positive) cells without any predominant pattern. Interestingly, the majority of the intraepidermal lymphocytes stained positive with CD8, a finding different from the results of Aiba and Tagami who found no predominant pattern between intraepidermal CD4 and CD8 positive cells. [13] In accordance with Neoh et al. but in contradiction with Hussein and colleagues we found B-cells to be sparse in infiltrates of PR. [15,16]

One interesting immunohistochemical observation was the abundance of CD1a positive cells in the epidermis and dermis of lesions of PR. The relative increase in density within the spongiotic foci, beneath areas of parakeratotic mounds, and the intraepidermal vesicle suggests that they are involved in the pathology of PR. Sugiura et al. as well as Neoh et al. also reported many CD1a+ cells in infiltrates of PR. Sugiura et al. observed that the number of CD1a positive cells decreased significantly in late lesions. [14,15]

The multifocal expression of CK 16 beneath mounds of parakeratosis in a "zebra pattern" and the increased expression of bcl-2 and Ki67 within the spongiotic foci, surrounding simulators of Pautrier's microabscesses and beneath the mounds of parakeratosis are part of the distinctive spongiotic process of PR. Increased expression of CK 16 in the upper and lower portions of the Malphigian layer had been described first in lesions of psoriasis, but it is now considered be a sign of proliferating epidermis. [28] The increased density of CK 16 beneath the parakeratotic mounds probably represents immediate recovery of the epidermis from spongiotic injury.

In sum, our study confirms a century-old observations of Paul Gerson Unna and Hollman that PR has distinct histological features enabling distinction from other forms of spongiotic dermatitides. Multifocal spongiosis, mounds of parakeratosis, simulators of Pautrier's microabscesses, and a superficial perivascular infiltrate with sparse spillage into the interstitium are typical of PR. A skin biopsy can be very helpful in the differentiation from other disorders such as guttate psoriasis, nummular dermatitis, pityriasis lichenoides, secondary syphilis, tinea corporis, and early mycosis fungoides. Immunhistochemical results indicate that the epidermal spongiosis is mediated by CD8 positive lymphocytes and CD1a positive Langerhans cells together with CD68 positive histiocytes, while B lymphocytes and plasma cells are absent from the infiltrate and CD4 positive lymphocytes remain largely in the dermal infiltrate.