Skin Lesions of Sweet Syndrome and Its Dorsal Hand Variant Contain Vasculitis: Title and subTitle BreakAn Oxymoron or an Epiphenomenon?

IN THIS ISSUE of the ARCHIVES, 2 articles address intriguing features pertaining to neutrophilic dermatosis: (1) a clinicopathologic study that demonstrated the presence of vasculitis in 6 of 28 lesional skin biopsy specimens from 21 patients with Sweet syndrome¹ and (2) a retrospective study of the clinical and histopathologic characteristics found in 7 women with neutrophilic dermatosis of the dorsal hands (NDDH).² How are we to integrate this information into our preexisting concepts of these conditions? My interpretation of these findings has been formulated by asking (and answering) several questions.

What is Sweet syndrome? In 1964, Sweet³ published a report in which he described "a distinctive and fairly severe illness" that he had encountered in 8 women during the 15 years from 1949 to 1964.³ Sweet initially commented that the condition was known in his department as the Gomm-Button disease in "eponymous honour of the first 2 patients," and he subsequently recommended that the name of the condition remain descriptive.^{3 - 4} However, the eponym Sweet syndrome has become well established.

In his original report, Sweet defined the following 4 cardinal features of the dermatosis: "fever, neutrophil polymorphonuclear leukocytosis of the blood, raised painful plaques on the limbs, face and neck and histologically a dense dermal inflammation with mature neutrophil polymorphs."³ He also stated that "the finer superficial part of the dermis . . . is hardly affected at all nor, structurally, are the blood vessels at any depth."³ As a result, one of the major diagnositic criteria for Sweet syndrome is the "histopathologic evidence of a dense neutrophilic infiltrate without evidence of leukocytoclastic vasculitis."^{4 - 5}

This definition of Sweet syndrome has evolved since the original description of the dermatosis.^{6 - 8} Depending on the associated clinical setting, the dermatosis may be classified as classic or idiopathic,^{9 - 13} malignancy associated,^{14 - 29} or drug induced.³⁰ Although cytokines such as granulocyte colony-stimulating factor are suspected to have an etiologic role in the development of Sweet syndrome, the definitive pathogenesis remains to be determined.^{31 - 32} Systemic corticosteroid therapy is still the therapeutic gold standard for this condition.^{33 - 34} However, potassium iodide and colchicine are effective first-line therapies, especially for individuals in whom corticosteroid therapy is contraindicated. Additional studies may be warranted to further confirm the efficacy of other less frequently reported agents such as indomethacin, clofazimine, dapsone, and cyclosporine.³⁵

What is NDDH? Pustular vasculitis refers to a group of conditions that clinically are characterized by "pustular lesions that occur on purpuric bases"³⁶ and that microscopically "range from fully developed leukocytoclastic vasculitis to a neutrophilic vascular reaction [which is the preferred designation instead of the previously used Sweet-like vasculitis] with vessel changes such as those seen in Sweet syndrome."³⁶ In 1995, Strutton et al³⁷ introduced the term pustular vasculitis of the dorsal hands for "a peculiar cutaneous eruption limited to the dorsa of the hands and fingers." Subsequently, Galaria et al³⁸ proposed the term neutrophilic dermatosis of the dorsal hands for this condition.

Presently, NDDH has been reported in 19 patients: 3 men^{38 - 39} and 16 women.^{2 ,37 - 38 ,40} Skin lesions, morphologically similar to those of Sweet syndrome, are predominantly restricted to the dorsal hands; however, 8 (42%) of the 19 individuals also had lesions at other sites: oral mucosa (4 patients: lip [n = 3]^{2 ,37 ,40} and gingiva [n = 1]² ), arm (3 patients),² leg (2 patients),² back (1 patient),² and face (1 patient).² Biopsy findings range from a neutrophilic infiltrate without vasculitis^{38 ,40} to leukocytoclastic vasculitis of variable severity.^{2 ,37 ,39 - 40} Similar to those of Sweet syndrome, lesions of NDDH rapidly resolve after initiation of therapy with systemic corticosteroids or dapsone or both.^{2 ,37 - 40} Based on the efficacious response of patients with Sweet syndrome who have been treated with either potassium iodide or colchicine, it is reasonable to postulate that these agents would also be likely to provide prompt improvement of the skin lesions in patients with NDDH.

In summary, as the spectrum of clinical and pathological manifestations of NDDH continues to expand, this condition increasingly appears to represent a variant of Sweet syndrome that is localized primarily to the dorsal hands.^{2 ,37 - 39} It is not unexpected that one of DiCaudo and Connolly's² patients (patient 2) had a concurrent, biopsy-confirmed lesion of Sweet syndrome on her back. Indeed, if the pathologic changes of NDDH are those of a secondary vasculitis instead of a primary vasculitis, several of the other patients with NDDH would also fulfill the diagnositic criteria for Sweet syndrome.

What is vasculitis? Malone et al¹ have used the concept of vasculitis defined by Fauci.⁴¹ Specifically, vasculitis implies leukocytoclastic vasculitis, "usually involving postcapillary venules; infiltration of polymorphonuclear leukocytes with leukocytoclasis (presence of nuclear debris), fibrinoid necrosis, and extravasation of erythrocytes."⁴¹

The microscopic presence of vasculitis on hematoxylin-eosin–stained sections of lesional skin biopsy specimens does not always implicate its pathogenesis. However, the finding of immunoglobulin, complement, or both within the affected vessel walls on immunofluorescent staining suggests an immune-complex–mediated primary vasculitis.^{42 - 43} In contrast, the finding of pathologic changes of vasculitis on light microscopy in the absence of both immunoglobulin and complement on immunofluorescent microscopy could be consistent with a secondary vasculitis or a vasculitis that has occurred as an epiphenomenon, such as the vasculitis that is occasionally observed to involve the superficial vessels adjacent to a follicle that has recently ruptured.

What is an oxymoron? An oxymoron is "a combination of contradictory or incongruous words."⁴⁴ Hence, vasculitis in Sweet syndrome would be an oxymoron if vasculitis refers to a primary immune complex–mediated leukocytoclastic vasculitis. However, what if our current interpretation of the histopathologic diagnostic criteria for Sweet syndrome were clarified by defining the pathologic changes of a lesional skin biopsy specimen to be a dense neutrophilic infiltrate without evidence of a primary immune complex–mediated leukocytoclastic vasculitis? Then, the observation of a secondary leukocytoclastic vasculitis (demonstrated by both fibrinoid change and intramural inflammation of even a single vessel) in a skin biopsy specimen from a lesion that was otherwise suspected to be Sweet syndrome would not exclude the possibility of that dermatosis.

What is an epiphenomenon? An epiphenomenon is "an accidental or accessory event or process occurring in the course of a disease but not necessarily related to that disease."⁴⁵ I am pleased that Malone et al¹ incorporated my classication of vasculitis-associated epiphenomenon in their report. Specifically, this refers to a "de novo epiphenomenon" (in which the involved vessels are located in an area showing only perivascular inflammation) and an "innocent bystander epiphenomenon" (in which the involved vessels are located in an area of diffuse or bandlike neutrophilic inflammation) (P.R.C., written communication, January 31, 2001).

What are the observations of vasculitis in Sweet syndrome and NDDH? First, Malone et al¹ appropriately acknowledge the following earlier observations of vasculitis in Sweet syndrome skin lesions that were reported by Jordaan⁴⁶ and von den Driesch⁴⁷ : features of vasculitis, such as intramural inflammatory cells (in 21 [39%] of 54 biopsy specimens) and fibrinoid necrosis (in 10 [19%] of 54 biopsy specimens),⁴⁶ and "vessels with vasculitis-like changes in about 30% of cases."⁴⁷

Second, Malone et al¹ found vasculitis involving at least 1 blood vessel in 6 of the 28 biopsy specimens that had been taken from 21 patients with Sweet syndrome. Two specimens each were evaluated from 7 patients; of these, specimens from 1 patient (both of which did not show vasculitis) were obtained 12 days apart, whereas the specimens from the other patients were obtained concurrently. Only 1 specimen was evaluated microscopically in 5 of the patients with Sweet syndrome whose lesions showed vasculitis, and 2 specimens were obtained concurrently from separate lesions of 60 days' duration in the sixth patient; only 1 of the 2 specimens demonstrated vasculitis. Malone et al¹ hypothesized that this "varying evidence of vasculitis may have resulted from the clinician's or pathologist's sampling error." I postulate that the discordant findings from these lesions of similar duration, location, and degree of inflammation were both valid, and that the presence of secondary vasculitis in only 1 of the specimens may have occurred as the result of a local etiologic or focal factor that was not simultaneously present in both lesions.

Third, the immunofluorescence studies showed no immunoglobulins or complement within the affected vessel walls when Malone et al¹ evaluated the 6 biopsy specimens that demonstrated vasculitis. A possible explanation for these findings is that the specimens had been taken from older lesions (ie, 2 of the Sweet syndrome lesions demonstrating vasculitis were of >8 weeks' duration) in which the immunoglobulin and complement were no longer present. However, the absence of immune complex deposition in the vessel walls after immunofluorescent staining of specimens from more acute Sweet syndrome lesions that showed vasculitis (ie, 1 lesion of less than 1 week's duration and 2 additional lesions of 1-2 weeks' duration) makes this hypothesis less likely. As postulated by Malone et al,¹ the absence of both immunoglobulin and complement in the affected vessel walls and " . . . the varying pattern of dermal inflammation suggests . . . secondary vasculitis"; ie, a vasculitis-associated epiphenomenon.

Fourth, Malone et al¹ found that the inflammatory infiltrate ranged from sparse to moderate to extensive in the 6 specimens that demonstrated vasculitis. An extensive neutrophilic infiltrate was present in 3 of the specimens. The other patients had either a sparse (2 specimens) or moderate (1 specimen) infiltrate.

Fifth, in the biopsy specimens from patients with Sweet syndrome that showed vasculitis, the pattern of inflammation was reported but not defined. It is reasonable to assume that diffuse represents inflammation of uniform density throughout both the papillary and reticular dermis, bandlike represents diffuse inflammation restricted to the papillary dermis and perhaps extending into the upper reticular dermis, and perivascular represents inflammation in which the infiltrate is present only around blood vessels. Diffuse (3 specimens) or bandlike (2 specimens) inflammation, either alone (3 patients) or in combination with perivascular inflammation (2 patients), was present in 5 patients. Perivascular inflammation was present in 3 patients; it was the only pattern of inflammation in 1 of these patients, whereas the other 2 patients also had either diffuse or bandlike inflammation. Hence, the pattern of vascular inflammation in the patients with Sweet syndrome and secondary vasculitis corresponds to an innocent bystander pathogenesis of vasculitis-associated epiphenomenon (3 patients), a de novo mechanism of vasculitis-associated epiphenomenon (1 patient), or possibly both (2 patients).

Sixth, Malone et al¹ state that "the presence of vasculitis was significantly associated with lesions that had been present a longer duration before their biopsy (P = .02)"; however, they do not specify the variables of duration used for their calculation. The relationship between lesion duration and the presence or absence of vasculitis in the biopsy specimens and patients with Sweet syndrome is summarized in Table 1. Only 2 (33%) of the 6 patients with Sweet syndrome whose biopsy specimens demonstrated vasculitis had lesions that had been present for longer than 8 weeks. In comparison, vasculitis was absent in the specimens from 2 (13%) of the 15 patients with Sweet syndrome (corresponding to 5 [23%] of the 22 specimens) whose lesions were of longer than 8 weeks' duration. Seven lesions from a total of 4 patients with Sweet syndrome had been present for longer than 8 weeks before biopsy; 1 patient had a biopsy of a single lesion that showed vasculitis , and the other 3 patients underwent biopsy of 2 lesions: in these 3 patients, vasculitis was absent in 5 specimens and present in 1 specimen. Hence, only 2 (50%) of the 4 patients with Sweet syndrome (corresponding to only 2 [29%] of the 7 specimens) with lesions of longer than 8 weeks' duration showed vasculitis. Also, biopsy of an early lesion did not preclude the possibility of discovering vasculitis, since the lesions from 3 (50%) of the 6 patients with Sweet syndrome whose biopsy results showed vasculitis were of no longer than 2 weeks' duration.

Seventh, in NDDH, Strutton et al³⁷ claim that "histologically these lesions resembled Sweet syndrome except for the presence of a definite vasculitis of small dermal vessels." Therefore, if these pathologic changes are truly those of a primary vasculitis, then NDDH should be considered a form of cutaneous vasculitis. However, the changes of vasculitis—at least in some of the patients with NDDH—appears to be secondary instead of primary. In the subsequently described patients with this condition, the degree of vascular changes within the lesions were of variable severity; indeed, leukocytoclastic vasculitis was absent in 3 individuals.³⁸

Eighth, direct immunofluorescence studies of lesional skin were performed in 2 patients with NDDH.^{2 ,40} The results of these studies in the patient described by Hall et al⁴⁰ and in patient 2 of DiCaudo and Connolly² were negative for immunoglobulins or complement. Hence, it is reasonable to postulate that the vasculitis observed in these patients with NDDH (and perhaps also in other patients with this condition) was a secondary leukocytoclastic vasculitis instead of a primary leukocytoclastic vasculitis.

What does it all mean? The observation by Malone et al¹ of secondary vasculitis in skin lesions of Sweet syndrome is not an oxymoron, but an epiphenomenon in which the affected vessel is most frequently an innocent bystander within moderate to extensive, diffuse or bandlike dermal inflammation. Also, the presence of pathologic changes consistent with a secondary leukocytoclastic vasculitis on hematoxylin-eosin–stained slides of a lesional skin biopsy specimen does not absolutely exclude the diagnosis of Sweet syndrome. Furthermore, it is reasonable to consider NDDH to be a localized variant of Sweet syndrome in which the lesions are predominantly restricted to the dorsal hands, especially if the pathologic changes are interpreted to represent a secondary leukocytoclastic vasculitis similar to that observed in some lesions of Sweet syndrome.