CASE

REPORTS

Skin ulcerations as manifestation of vasculopathy in neurofibromatosis Bo Ljunggren, MD, PhD,a Rolf Holst, MD,a Mecislovas Simanaitis, MD, PhD,b and Alf Rausing, MD, PhDc ¨ and Barseba ¨ ck Sweden Malmo Vascular involvement in neurofibromatosis sometimes occurs and usually affects larger blood vessels of internal organs. We describe an unusual case of a now 41-year-old man with extensive skin involvement with ulcerations and histopathologic features of fibromuscular dysplasia, which has not been previously reported in the skin. ( J Am Acad Dermatol 2004;51:656-9.)

N

eurofibromatosis (NF) is a multisystem genetic disease involving many organs, such as the skin, skeleton, nerves, and the vascular system. Skin manifestations often include cafe´ au lait spots, neurofibromas, and axillary freckles.1 Larger blood vessels are sometimes affected, causing bleeding from aneurysms or hypertension from stenosis of renal arteries.2 Vascular skin involvement has not been reported previously and we, therefore, want to describe a patient with NF who developed extensive ulcerations on his trunk and left arm. Histopathology revealed blood vessel abnormalities consistent with findings reported in internal vasculopathy in NF.

CASE REPORT The patient, a 41-year-old man, first visited our department in 1983 because of tenderness, ulceration, and cyanotic erythema over the left scapular region (Fig 1). He described having had similar lesions earlier, sometimes with ulceration, and healing with atrophic scars. He was noted to have 5 cafe´ au lait spots larger than 1.5 cm, a chest deformity, and kyphoscoliosis. An eye examination revealed a number of pigmented lesions, probably representing Lisch nodules. He had no family history of NF. Routine laboratory tests produced normal findings, and several skin biopsy specimens were From the Departments of Dermatologya and Pathology,b Lund University, Malmo¨ University Hospital; and Mikroklin AB,c Barseba¨ck. Funding sources: None. Conflicts of interest: None identified. Reprint requests: Bo Ljunggren, MD, PhD, Department of Dermatology, Lund University, Malmo¨ University Hospital, S-20502 Malmo¨, Sweden. E-mail: [email protected]. 0190-9622/$30.00 ª 2004 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2004.04.004

656

inconclusive, although nervelike structures were described, consistent with NF. Because of progression of the painful ulcers over his left shoulder and spreading to the back of his arm, the patient was referred for plastic surgery in 1988, although the diagnosis was unclear. In 1989 and 1990, and twice in 1994, he was treated with excision and split skin grafts of the affected areas. However, no stable healing was achieved and the ulcerations soon recurred. In 1990, he had a malignant melanoma (Clark III, Breslow 5.5 mm) over the right scapular region excised with uncomplicated postoperative healing. He also had a gastrointestinal stromal tumor, which is seen with increased frequency in NF, excised in 1991. The ulcerated area on the left side of the shoulder increased steadily, and measured 40 3 9 cm in January 2000 (Fig 2). He was admitted as an inpatient to the departments of dermatology and angiology, and extensive investigations were performed. The hematologic tests, including a bone marrow evaluation, produced normal findings, as did the coagulation status. Vasculitis serology, collagenosis tests, and cardiolipin antibodies were all negative. Tests for HIV, tuberculosis, and Borrelia, and multiple cultures for micro-organisms, were all negative. Angiography showed normal vessels in the left arm. On the right side, however, there was stenosis of the proximal part of arteria brachialis and occlusion of arteria ulnaris. Magnetic resonance angiography showed no root or medulla affection corresponding to the ulcerated areas. Because a diagnosis of pyoderma gangrenosum was suggested, various treatment schedules with high-dose prednisolone (1 mg/kg/d), cyclosporine, mycophenolate mofetil, and Dapsone were tried in the spring of 2000, without success.

J AM ACAD DERMATOL VOLUME 51, NUMBER 4

Fig 1. Early skin lesions of left scapular area in 1983.

Fig 2. Extensive skin ulceration of left arm and shoulder with maximal area involved, as seen in 1999.

In mid-2000, a new, painful ulcer with bluish borders appeared over the left sternal area as did a small cyanotic area with incipient necrosis over the right wrist. In October 2000, thalidomide (100 mg/d) was started in the patient and gradual improvement was noted during the next few months. In March 2001, excision of the medial portion of the ulcer overlying scapula was performed with good postoperative healing. In July of the same year, the remaining ulcers over the left arm and sternum were treated with split skin grafts. The condition was stable over the summer, but from September 2001 the lateral parts of the ulcer area started to break up again, and the pain worsened. Between 2002 and 2003, the condition was relatively stable with remaining ulceration over the left elbow, whereas the shoulder and upper aspect of the arm stayed almost healed. During the whole disease course areas of marginal cyanosis of the borders of the ulcers and areas of vasoconstriction, sometimes with livedo, well demarcated, in otherwise normal skin, preferentially outside of the ulcerated areas and on the right arm, could be seen (Fig 3).

HISTOPATHOLOGY Tissues from the skin lesions have been examined on many occasions. The first biopsy specimen was

Case reports 657

Fig 3. Area of vasoconstriction and livedo of opposite, nonulcerated right arm.

seen in 1983. It was a punch biopsy specimen from a reddish-cyanotic but not ulcerated lesion on the lateral aspect of the left scapular area. The clinical suggestion was cutaneous arteritis. In the reticular dermis and in the subcutis there were many small nerve twigs surrounded by sheaths of spindle cells, and a small subcutaneous artery was also affected. In 1985 a further punch biopsy specimen was examined, now from an ulcerous lesion. The clinical suggestion still was polyarteritis. Histologically, the ulcer appeared as an epidermal necrosis and the dermis was unaffected, but had neural and vascular lesions as in 1983. Another biopsy specimen in 1985, again from a nonulcerated lesion, showed the same picture as the one from 1983, and now was considered as consistent with NF. From 1989 to 2001, many skin resections were performed from ulcerous tissue and surrounding skin, and many further punch biopsy specimens were seen. All specimens demonstrated changes of varying degree. The superficial part of the ulcers had an unspecific picture of a loose granulation tissue with extensive inflammatory infiltrates. There was a variable, sometimes compact fibrosis around the ulcers, often with deep subcutaneous extensions. In the dermis surrounding the ulcers and in the adjacent subcutis there were many nerves surrounded by spindle cell sheaths of varying thickness. They were loose and edematous and quite paucicellular. The spindle cells were very slender and had light-staining nuclei. Small arteries and veins in the neighborhood of the nerves often were included in the spindle cell sheaths. However, there were also many subcutaneous and reticular dermal arteries without accompanying nerve surrounded by similar loose spindle cell sheaths (Fig 4). They often had a thickened and hypercellular muscular wall and sometimes a split lumen and a disorderly appearance reminiscent of a fibromuscular dysplasia (Fig 5). Thrombosis or complete occlusion was never seen. A thin-walled vein with a widely patent, semilunar lumen was often included within the sheath (Fig 6). The cells of

658 Case reports

Fig 4. Subcutaneous tissue adjacent to ulcer. Two small arteries surrounded by spindle cell sheaths. Adjoining the upper one there is small nerve encircled by a similar spindle cell structure. (Original magnification: 335.)

the cellular wall were positive in staining for smooth muscle actin and negative for S-100. The spindle cell sheath contained occasional cells positive for S-100 and a few were positive for smooth muscle actin. Generally, the vascular lesions were most pronounced close to the ulcer base where the tissue was also edematous, but the changes also were found in subcutaneous tissue far from the tissues with reactive lesions (Fig 6). The patient had an operation for a malignant melanoma on his right shoulder in 1990. This specimen did not show any spindle cell lesions. A further 3 nevi were excised (from the abdomen, thigh, and lower leg). They also were situated on normal skin. In 1991, the patient had an ileal tumor. A 3-cm intramural tumor was removed together with 10 cm of gut and appendix. A diagnosis of leiomyoma was rendered. The tumor was positive for CD117 and negative for S-100. The diagnosis today would be gastrointestinal stromal tumor. Many sections of the intestinal wall and subserosal fat tissue were examined. There were no neurofibromas or vascular or neural lesions like those of the skin.

DISCUSSION Vascular abnormalities in NF were first described by Ruebi2 in 1945. He thought that such lesions could be found in all cases if searched for, a supposition that has not been confirmed in later studies. Salyer and Salyer3 found vascular lesions in 7 of 18 cases in an autopsy series. They did not report such lesions in the skin. They thought that the arterial lesions were caused by the proliferation of Schwann cells within the wall. Later studies have disproved this theory. Lehrnbecher et al4 found the cells of the vascular wall to be positive for smooth muscle actin and negative for S-100, as in our case, and the proliferation, therefore, appears to be myogenic. Most reports mention renal artery affection but lesions have been

J AM ACAD DERMATOL OCTOBER 2004

Fig 5. Subcutaneous tissue close to base of ulcer. Artery with disordered wall structure with muscular hypercellularity, split lumen, and irregular demarcation. Note edema of spindle cell sheath and surrounding tissue. (Original magnification: 335.)

Fig 6. Subcutaneous tissue, not in close contiguity with ulcer base. Thick artery with hypercellularity but also some fibrosis of muscular wall and spindle cell sheath. Thin-walled vein included within sheath. Adjoining small nerve with spindle cell sheath of its own. (Original magnification: 335.)

described in arteries of all calibers from aorta to arterioles.5 The majority of the cases of the report by Salyer and Salyer3 were asymptomatic, but many cases with various types of clinical disease supposed to be a result of the vasculopathy have been described. However, to our knowledge, skin vascular lesions, whether symptomatic or not, have never been described before. The skin lesions in our case were clinically intriguing. During the long disease course during more than 15 years, several differential diagnoses were considered. Based on the clinical appearance, and early biopsy specimens, pyoderma gangrenosum was suggested as a possible diagnosis, and appropriate treatment schemes were applied. No healing response was obtained, however, either with high doses of oral steroids or with immunosuppressant drugs. With further biopsy specimens from the ulcerations, however, the peculiar character of the histopathologic changes was revealed. The lesions

J AM ACAD DERMATOL VOLUME 51, NUMBER 4

of our case are indeed peculiar. They consist of loose spindle cell lesions encircling nerves and vessels. Those around the nerves sometimes have a considerable resemblance to pictures that may be seen in some cases of neurotropic malignant melanoma,6 a disease that our patient obviously does not have. They are not like ordinary plexiform neurofibromas at all, clinically or histologically. It may be argued that the perivascular sheaths are edematous and expanded adventitial tissue, but the structures surrounding the nerves are identical. Further, there is a disorderly proliferation of smooth muscle and abnormalities of elastic fibers in the wall. The picture may well be described as fibromuscular dysplasia. The changes are not simple reactive fibrosis around the ulcer base, as they are found also in large operation specimens far from reactive changes in the ulcer base. The relation to the clinical symptoms is unclear. Angiographically, the patient had large artery narrowing, although it was not topographically corresponding to the enormous ulcer and, therefore, it appears reasonable to assume that this is because of the small vessel lesions. Small nerve dysfunction may also be involved. Are the lesions really a manifestation of NF or do they represent another disease? The muscular hypercellularity of the arterial wall is identical to the

Case reports 659

literature reports of NF vasculopathy and so are the arteriographic findings, and we think that this is really yet another, although rare, manifestation of this protean disorder. Symptomatic neurovascular skin lesions in NF must be very uncommon as they have not been reported before. Interestingly, our patient seems to have such lesions exclusively in the skin, and only in limited regions of the skin, to judge from the normal picture in the other tissues examined during his long medical history. We are grateful to Bernhard Zelger, MD, PhD, for fruitful discussions. REFERENCES 1. Gutmann DH, Aylsworth A, Carey JC, Korf B, Marks J, Pyeritz RE, et al. The diagnostic evaluation and multidisciplinary management of neurofibromatosis 1 and 2. JAMA 1997;278:51-7. 2. Ruebi F. Neurofibromatosis et lesions vasculaires. Schweiz Med Wochenschr 1945;75:463-7. 3. Salyer WR, Salyer DC. The vascular lesions of neurofibromatosis. Angiology 1974;25:510-9. 4. Lehrnbecher T, Gassel AM, Rauh V, Kirchner T, Huppertz H-I. Neurofibromatosis presenting as a severe systemic vasculopathy. Eur J Pediatr 1994;153:107-9. 5. Singh S, Riaz M, Wilmshurst AD, Small JO. Radial artery aneurysm in a case of neurofibromatosis. Br J Plast Surg 1998; 51:564-5. 6. Weedon D. Skin pathology. 2nd ed. London: Churchill-Livingstone; 2002. p. 827.