The diagnostic label "pityriasis rosea" (PR) was introduced by the French dermatologist Camille Melchior Gibert (1797-1866) in 1860 to describe a distinctive clinical entity characterized by an eruption of "small furfuraceous spots which are very lightly colored, irregular, scarcely exceeding a fingernail in size, numerous and close set, . . . with predilection for the neck, the upper part of the chest, and the upper part of the arms, but which may spread successively from above downwards as far as the thighs, . . . is protracted quite commonly to six weeks or two months." [1] The disease, though, had already been described in earlier literature under the rubric "roseola annulata," by the British dermatologist Robert Willan (1757-1812), who influenced Gibert's approach to dermatology. [2,3] In fact the term furfura, which describes the characteristic scales of individual lesions of PR, was defined by Willan as small exfoliations of the cuticle, which take place after slight inflammation or irritation of the skin. [4] The disease has gone under a variety of names such as pityriasis circinata (Hardy and Horand), pityriais rubra aigu dissemine, arthritide pseudoexanthematique (Bazin), roseola squamosa (Chapard), pityriasis marginee et circinee (Vidal), erytheme papuleux desquamatif (Besnier) but only the name pityriasis rosea persisted. [5]

Comprehensive studies on the histopathology of PR were conducted by Unna and Hollman. [5] They described their findings as a process and argued whether the primary pathology commences from either the epidermis or the dermis. While earlier authors unanimously agreed that the granular layer was intact and that the cornified layer is only focally abnormal, Unna described additional epidermal features such as thickening of the spinous layer and a noticeable quantity of mitotic figures in the basal layer. He believed that the process initiates with intercellular edema of the lower layers of the epidermis with ascent of very sparse leukocytes. He observed that in the higher layers of the epidermis, the spinous cells form round or elliptical subcorneal blisters, the contents of which exit the cornified layer as serum and that the granular layer contiguous to the blister then fades. Unna claimed, however, that the blister, which has an irregular shape, does not translate to a clinically visible lesion. A mound of parakeratosis then appears where the old cornified layer is shed. The dermal changes observed were round cells surrounding dilated blood vessels and increased number of fibroblasts. After the extrusion of the vesicle, there is resolution of the inflammatory process in the dermis and the stratum spinosum is restored back to its normal appearance. [6]

Hollman studied in detail over 200 slide preparations and validated earlier observations by Unna that the basic underlying pathological process in PR is spongiosis. Hollman described changes in early lesions as moderate round cell perivascular infiltrates in the papillary and subpapillary dermis with accompanying dilatation of blood vessels. He said that the process affected the overlying epidermis, which in turn developed moderate intra- and intercellular edema in the stratum spinosum while both the granular layer and the cornified layer remained intact. The older lesions were described by him as having an increase of infiltrates with some migration of leukocytes into the overlying epidermis coinciding with a proliferation of the spinous layer. [7] The development of a subcorneal blister, as described by Unna, was noted by him in only one case. [5]

Recent editions of common textbooks in dermatopathology show minor variations in their description of the histopathology of PR. Table 1 provides a summary of the histopathological findings of PR in the classic Jadassohn textbook where descriptions of Unna and Hollman are reproduced in detail in comparison with descriptions in current dermatopathology textbooks. [8-12] Controversy exists especially in regard to the presence or absence of necrotic keratinocytes and vacuolar alteration in the epidermis. Multinucleated keratinocytes have been mentioned only by a few authors. The composition of the dermal infiltrate is usually described as mainly lymphocytic but according to some authors, eosinophils are encountered in late lesions. While some authors consider the histopathology of lesions of PR to be distinctive, the only differential diagnosis being erythema annulare centrifugum, [11] others mention specifically that the changes are those of a "non-specific subacute or chronic dermatitis." [12]

Immunhistochemical studies of PR are sparse and controversial in regard to the composition of the dermal and epidermal infiltrate. [13-16]

Aiba and Tagami studied skin lesions of 15 patients with PR. They noted focal intercellular edema, epidermotropism of mononuclear cells often associated with the formation of focal intraepidermal collections of mononuclear cells, and a perivascular lymphohistiocytic cell infiltration in the superficial dermis histologically. Immunologic analysis showed that large numbers of lymphocytes in the perivascular infiltrate reacted with anti-pan-T-cell, anti-helper-inducer subset, and anti-HLA-DR antibodies, while the epidermotropic mononuclear cells consisted of helper-inducer cells or suppressor-cytotoxic cells without any predominant pattern. In addition to epidermal Langerhans' cells, some of the dermal infiltrating cells were reactive with OKT6 and there was localized expression of HLA-DR antigen on keratinocytes. Based on their results, the authors suggested that cellular immune reactions are taking place in the lesional epidermis of pityriasis rosea. [13]

Sugiura et al. investigated 15 patients with PR with emphasis on the development of lesions over time. [14] Many CD1a+ cells were seen in the epidermis and dermis of early lesions. In the well-developed lesions, the number of CD1a+ cells greatly increased in the dermis. In the late lesions, CD1a+ cells in the dermis significantly decreased as compared with the well-developed lesions. Early lesions showed a moderate T-cell infiltrate. In the well-developed lesions, the dermal T-cell infiltrate was dense, and the CD4 CD8 ratio was 2.9. The late lesions had a moderate T-cell infiltrate, in which the CD4/CD8 ratio significantly decreased as compared with the well-developed lesions.

Hussein and colleagues investigated PR in comparison with allergic contact dermatitis and atopic dermatitis. [15] Infiltrates in all the three conditions were composed predominantly of CD3(+) T lymphocytes and CD68(+) cells (histiocytes). Some of the CD3(+) cells were granzyme B(+). The counts of CD3(+) and CD68(+) cells were high in allergic contact dermatitis compared with atopic dermatitis and PR. The counts of CD20(+) and granzyme B(+) cells were high in PR compared to allergic contact dermatitis and atopic dermatitis, but not statistically significant.

Only recently, Neoh et al. studied 12 biopsy specimens from 6 patients diagnosed with PR comparing herald patch and a secondary patch. [16] Histopathologically, all specimens showed epidermal changes such as parakeratosis, orthokeratosis, acanthosis, and spongiosis. The dermal infiltrate of lymphocytes stained positively for monoclonal antibodies specific for T cells. The ratio of the CD4+ (helper) vs. CD8+ (cytotoxic) T cells in the dermal infiltrate was increased in most specimens. Increased staining for Langerhans cells was seen within the dermis of lesional skin. There were no marked differences found in histopathology and immunohistochemistry between the herald patch and secondary lesions. There was a lack of natural killer cell and B-cell activities in PR lesions. The authors concluded that their results indicated a predominantly T-cell mediated immunity in the development of PR.

The following study attempts to reassess critically histopathologic criteria of PR and to elucidate further the immunophenotype of the infiltrate in lesions of PR.