HEMATOLOGICAL ONCOLOGY, VOL. 10,331-337 (1992)

CYTOPHAGIC HISTIOCYTIC PANNICULITIS: A DIAGNOSTIC DILEMMA PHAIK-LENG CHEAH, LAI-MENG LQOI, PUAY-ENG TAN, JOHN BOSCO* AND PONNUDURAI KUPERAN

Departments of Pathology and *Internal Medicine. Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia

SUMMARY Cytophagic histiocytic panniculitis (CHP) is a recently recognized entity that frequently poses a perplexing diagnostic problem. Although the classical case presents with a relapsing fever, subcutaneous nodules, pancytopenia and liver dysfunction, most patients have in addition a multitude of other manifestationswhich confuse the clinical picture. Notwithstanding the variable clinical course, the disease frequently terminates in fatal hemorrhage. Diagnosis is based on histological features. A lobular panniculitis with an infiltrate of cytologically benign cytophagocytichistiocytes in skin nodules is the sine qua non of CHP. Hence, a deep skin biopsy which includes subcutaneous fat is mandatory to establish the diagnosis. Published information regarding this newly described entity remains scarce and we report two cases of CHP, one occurring in a 30-year-old Kadazan man and another in a 17-year-old Chinese woman seen at the University Hospital, Kuala Lumpur. The latter case presented with exudative ascites, an unusual feature, possibly due to intra-abdominal panniculitis. In addition, we record the development of cirrhosis in the same patient. KEY WORDS

Cytophagic histiocytic panniculitis Mesentericpanniculitis Cirrhosis INTRODUCTION

Cytophagic histiocytic panniculitis (CHP) was first described by Winkelmann and Bowie in 1980.' This newly recognized entity is characterized by fever, subcutaneous nodules, pancytopenia, liver dysfunction and a progressively downhill clinical course frequently terminating in fatal hemorrhage. Diagnosis is often difficult in view of the wide spectrum of clinical manifestations. Nevertheless, CHP can be confirmed by identifying an infiltrate of cytologically benign, cytophagocytic histiocytes associated with panniculitis in biopsies of skin lesions. Published information on CHP is however scarce and CH P remains a poorly-understood condition. We report our experience with two cases of CH P encountered at the University Hospital, Kuala Lumpur and highlight the unusual manner of clinical presentation and subsequent development of liver cirrhosis in one case. CASE REPORTS case 1

A 30-year-old Kadazan man presented at a hospital in Sabah (East Malaysia) with a 2-month history of intermittent high grade fever and recurrent, non-tender subcutaneous nodules over the trunk and extremities. During his stay in hospital, he also developed pleural effusion, arthralgia Addressee for correspondence: Dr P-L Cheah, Department of Pathology, Faculty of Medicine, University of Malaya, 59100 Kuala Lumpur, Malaysia.

0278-0232/92/06033 1-07 $08.50 0 1992 by John Wiley & Sons, Ltd.

Received 21 July 1992 Revised 10 November 1992

332

P.-L. CHEAH E r AL.

involving the knee and ankle and oedema. Investigations showed a raised erythrocyte sedimentation rate (ESR) of 52 mm/h (Westergren). Blood counts showed anemia (hemoglobin level of 11.2 g/100 ml) and leukopenia (1700/pl), while platelet count was 150 x 103/pl. Bone marrow aspiration failed to reveal an underlying cause for the hematological findings. Biopsy of a skin nodule showed non-specificinflammatory changes. Hypoalbuminemia (24 g/L) and raised alanine aminotransferase (ALT) (174 IU/L) and alkaline phosphatase (ALP) (1622 IU/L) were revealed on the liver profile. Anti-nuclear factor, rheumatoid factor and LE cells were not detected. His sputum grew beta-hemolytic streptococci (sensitive to penicillin) but repeated cultures for Mycobacterium tuberculosis were negative. Despite treatment with a course of penicillin and a trial of anti-tuberculous chemotherapy his fever persisted. Three months after initial presentation, the patient was referred to us as a diagnostic problem. On admission, oral ulcers were noted. The liver was palpable 5 cm below the costal margin. The spleen and lymph nodes were not enlarged clinically. Apart from leukopenia (total white count of 26OO/pl; 65 per cent neutrophils, 22 per cent lymphocytes, 7 per cent monocytes, 5 per cent metamyelocytes and 1 per cent atypical lymphocytes) the hemoglobin level and platelet count were normal. Liver function tests showed a normal total bilirubin but low total plasma protein (56 g/L) and albumin fraction (18 g/L). ALP (839 IU/L), aspartate aminotransferase (AST) (41 1 IU/L) and ALT (297 IU/L) were elevated. Prothrombin ratio was increased and partial thromboplastin time was prolonged. The bone marrow aspirate and trephine biopsy were normal except for the presence of occasional benign-looking histiocytes exhibiting hemophagocytosis. Hypocalcemia (plasma calcium = 1.89 mmol/L) was noted. Other laboratory investigations including urine microscopy and measurements of 24-h urine protein, blood urea, sodium, potassium and chloride were normal. Bacteriological cultures and Widal-Weil Felix test were similarly negative. Meanwhile, the patient's liver function continued to deteriorate and his blood counts dropped progressively. Biopsy of the liver was forestalled in view of the prolonged prothrombin and partial thromboplastin times. Instead, abdominal ultrasonography was performed which revealed hepatomegaly. A repeat skin biopsy was inconclusive but at the third attempt a diagnosis of CHP was reached. He was commenced on intravenous vincristine (1 mg) and etoposide (100 mg) and oral prednisolone (100 mg). Vincristine was subsequently omitted on finding the patient allergic to the drug. The patient's general condition, blood counts and liver function improved rapidly. By the second cycle, his fever and most of his skin lesions had subsided. The blood counts and liver function tests also returned to normal. The patient was transferred back to the referring hospital in Sabah. Since then he has been lost to follow-up. Case 2 A 17-year-old Chinese girl first presented at another hospital with exudative, bacteriologicallynegative ascites and low-grade fever. A full blood count, urinalysis, blood urea and serum electrolytes, liver function tests, and an ultrasonographic examination of the abdomen did not reveal any abnormality. No definitive diagnosis was reached. The ascites improved following treatment with diuretics and antibiotics but the low-grade fever persisted. In the course of the next 9 months she developed oral ulcers, tender subcutaneous nodules and jaundice. The oral ulcers and skin lesions subsided temporarily with high dose prednisolone but jaundice and fever persisted. No history of blood or blood product infusion was elicited. The patient was then referred to us. On admission, she was icteric, pale, oedematous and had a temperature of 37.6"C. The liver was enlarged but no other abnormality was detected on examination of the other systems. Investigations showed anemia (hemoglobin level = 6.7 g/ 100 ml), but normal leukocyte and platelet counts. The ESR (4 mm/h; Westergren) was normal. Total plasma protein (54g/L) and plasma albumin (20g/L) were low. Total plasma bilirubin

CYTOPHAGIC HISTIOCYTIC PANNICULITIS

333

(441 pmol/L) was elevated with a predominant conjugated fraction (329 pmol/L). Liver enzymes were elevated (ALP = 274 IU/L, AST = 247 IU/L, ALT = 21 3 IU/L). Prothrombin ratio and partial thromboplastin time were normal. Anti-mitochondria1 antibody, anti-smooth muscle antibody, anti-nuclear factor, VDRL, hepatitis B surface antigen, anti-hepatitis A antigen and anti-hepatitis C antibody were negative on serological testing. Ultrasonographic examination revealed an enlarged liver and a minimally enlarged spleen. A liver biopsy was performed. The patient was managed symptomatically. She felt generally better and was discharged. During the next few months, non-tender plaques and nodules formed over the trunk and limbs. In addition she complained of severe pain in her hip. Blood counts revealed pancytopenia (hemoglobin level = 8.9 g/100 ml; white cell count = 2100/p1,73 per cent neutrophils, 17 per cent lymphocytes, 7 per cent monocytes, 3 per cent atypical lymphocytes; platelet count = 123 x 103/p1). Liver profile showed no improvement. Plasma calcium level (1 -90mmol/L) was low. Bacteriological cultures of the blood, urine and stools and sputum culture for Mycobacterium tuberculosis were negative. Abdominal computerized tomographic examination revealed hepatosplenomegaly and enlargement of some mesenteric and iliac lymph nodes. Bone marrow aspirate and trephine biopsy showed increased numbers of hemophagocytic, benign-looking histiocytes. The other hemopoietic components were normal in amount and maturation. The patient's condition continued to worsen rapidly. Blood counts dropped precipitously (hemoglobin level of 5.7 g/100 ml, total leukocytes of 200/p1 and platelets of 10 x 103/pl). Liver function deteriorated further. In view of the declining liver function, a laparotomy was performed whence a wedge biopsy of the liver and a mesenteric lymph node were sampled for histopathological examination. The second attempt at biopsy of the skin lesion clinched the diagnosis of CHP. The patient was commenced on etoposide 110 mg, adriamycin (doxorubicin) 60 mg, and vincristine 2 mg intravenously and prednisolone 90 mg orally for two cycles. Etoposide was then substituted with cyclophosphamide 400 mg intravenously, the other drugs remaining the same. This regime was administered for five cycles. Adriamycin and vincristine were then withdrawn and etoposide was re-instituted for another six cycles. A slow but progressive improvement in the patient's condition was observed. She was transferred back to the referring hospital and remained well. Cytotoxic therapy was stopped after 14 months of treatment. At the time of chemotherapeutic termination, a needle biopsy of the liver was repeated, i.e. 20 months after the first attempt. Repeat blood counts then revealed only mild anemia (hemoglobin level of 9.8 g/L). Plasma bilirubin and protein levels had returned to normal but enzymes continued to be elevated. The patient remains well at the time of writing, 41 months since first presentation and 10 months since discontinuing active therapy. She has also been relapse-free. PATHOLOGY Histological sections of biopsies from both patients performed at the University Hospital, Kuala Lumpur were reviewed. Recuts from the paraffin blocks were stained for Mac 387, UCHLl and L26 using the peroxidase-antiperoxidase method. Mac 387 (Dako) is a monoclonal antibody which stains for a human cytoplasmic antigen expressed in monocytes, tissue histiocytes and granulocytes* while UCHLl (Dako-CD45RO) and L26 (Dako-CD20) are monoclonal T and B-lymphocyte markers respectively. Initial biopsies of the skin lesions of both cases which did not include subcutaneous tissue were non-diagnostic. Case 1 The diagnostic lesional skin biopsy showed an essentially normal epidermis. Cytologically benign lymphocytes and plasma cells were seen around the dermal vessels and appendages. The most

334

P.-L. CHEAH ETAL.

Figure 1. Necrosis of subcutaneous adipose tissue with infiltrateof histiocytes and lymphocytes. (Inset) Examples of cytophagocytic histiocytes seen in other fields, (a) histiocytes showing phagocytosis of red blood cells and a lymphocyte and (b) another stuffed with erythrocytes.(Hematoxylin-eosin, x 315) prominent features were seen in the subcutaneous fat. In addition to necrosis of the adipose tissue, an infiltrate of histiocytes, some of which had engulfed erythrocytes, lymphocytes and nuclear debris was evident (Figure 1). Neutrophils were also observed in the areas of fat necrosis. The histiocytic cells, whether cytophagocytic or non-cytophagocytic, expressed strong immunoreactivityfor Mac 387. The lymphocytes in the dermis expressed UCHLl predominantly. Case 2 Histological and immunohistochemical examination of the diagnostic skin biopsy revealed features identical to those of case 1. The liver was biopsied at 12, 16 and 32 months after initial presentation. Histologically, an increasingly severe picture was noted in these serial biopsies. At 12 months, periportal steatosis, mild cholestasis,bile ductular proliferation with a mixed benign portal inflammatory infiltrate and hemosiderin loading of Kupffer cells were seen. Scattered apoptotic hepatocytes and focal necrosis of fat-laden hepatocytes were also observed. At 16 months, a marked panacinar macrovesicular fatty change had developed (Figure 2 ) . Otherwise, the other histological features were generally unchanged. The amount of steatosis and inflammation were decreased at 32 months. The liver architecture was however disrupted by thick fibrous bands surrounding nodules of hepatocytes (Figure 3) indicative of cirrhosis. No cytophagocytichistiocytewas identified at any stage. Mac 387 positivity confirmed the histiocytic nature of some of the infiltrative cells in the portal tracts in all three biopsies. In addition, some of the mononuclear portal inflammatory cells also exhibited UCHLl positivity. Curiously, many of the fat-filled hepatocytes were rimmed by Mac 387 positive histiocytic cells.

CYTOPHAGIC HISTIOCYTIC PANNICULITIS

335

Figure 2. Severe hepatic macrovesicular steatosis. A mononuclear inflammatory infiltrate and bile ducturlar proliferation are evident in the portal tract. (Hematoxylin-eosin,x 180)

A prominent lymphocytic depletion with loss of germinal centres was seen in the mesentenc lymph node. The sinuses were expanded and contained cytologically benign non-cytophagocytic histiocytes. Only occasional sinus cells stained positively with Mac 387. DISCUSSION These two cases illustrate the diverse clinical spectrum of CHP. The multitude of clinical manifestations exhibited by these two cases in general parallel those described previously.1s3-4However we wish to draw attention to the unusual manner of presentation in case 2. Although the 'patient eventually manifested the full-blown picture of CHP, presentation with ascites has hitherto not been described. The exudative nature of the ascites in this patient reflects an inflammatory process in the peritoneal cavity. Clinical and laboratory evaluations excluded a bacteriological etiology. Even though the exact cause could not be established, nonetheless mesenteric panniculitis should be considered in this situation. In the face of a confusing conglomeration of symptoms, signs and investigatory abnormalities, histological examination of the skin lesions remains the mainstay of diagnosis. The presence of typical, cytologicallybenign, 'bean-bag' histiocytes' stuffed with erythrocytes, white cells, platelets and nuclear fragments together with panniculitis clinched the diagnosis of CHP in our two cases. The characteristic histiocytes seen in conjunction with panniculitis also aid in differentiating CHP from Weber-Christian disease and malignant histiocytosis, two conditions with which it is most likely to be confused clinically. Cytophagocytic histiocytes do not occur in Weber-Christian

336

P.-L. CHEAH ETAL.

Figure 3. Thick fibrous band separating nodules of liver parenchyma indicating the development of cirrhosis. (Hematoxylin-eosin, x 200).

disease. The benign appearance of histiocytic cells in CHP is in contrast to the malignant appearance in malignant histiocyt~sis.~ The necessity of a deep biopsy to include subcutaneous fat cannot be overemphasised. Although severe impairment of hepatic function has been noted clinically, the histological changes in the liver are only briefly mentioned in most reports.'343c9Included among the ante-mortem changes observed in CHP are fatty change, non-specific periportal inflammation, hemosiderin in Kupffer cells and mild cholestasis. The presence of cytophagocytic histiocytes in the liver has only been reported at autopsy examination.6Focal necrosis of steatotic hepatocytes and bile ductular proliferation, constant features in all biopsies of case 2 have not been previously described. Progression to cirrhosis has similarly not been documented before. However in a review of cases of fatal panniculitis, Aronson et af., noted that bile ductular proliferation, hepatocyte necrosis and cirrhosis were present in some autopsied cases.' It is possible that among them were cases of what we know today as CHP. Although Mac 387 positive histiocytes were identified in the liver biopsies, these did not exhibit the classical feature of cytophagocytosis. It would appear that cytophagocytic histiocytes are not essential to the alterations in the liver. The exact role of noncytophagocytic histiocytes in the liver is still unclear. Whether these partake in the breakdown of the lipid metabolic pathway in the liver or are attracted there as a result of tissue injury is speculative. Aronson et al., proposed that the striking amount of fatty change in the liver is secondary to a panniculitis elsewhere, whereupon excess fatty acids and triglycerides are mobilized to the liver

CYTOPHAGIC HISTIOCYTIC PANNICULITIS

337

from breakdown of fat.7 Although the underlying etiology of the prominent fatty change is not clear, it is nevertheless evident that the liver damage is progressive and cirrhosis can supervene. The significance of the cytologically benign UCHLl positive T-lymphocytes in the skin lesions and liver is debatable. Whether they predispose to the development of T-cell lymphoma is unclear although cutaneous T-cell lymphoma has been reported to occur concurrently with CHP.6 Both patients responded well to cytotoxic therapy. The first case was lost to follow-up but the second case is alive and well, 41 months after initial presentation and 10 months after terminating treatment. She has also enjoyed a relapse-free interval of 21 months. It is difficult to predict her eventual outcome at this moment. Although usually fatal, prolonged survivals have been recorded in CHP.8-'oIn a review from the Mayo Clinic, Rochester, White and Winkelmann documented a survival of 15 years in one patient."

REFERENCES 1. Winkelmann, R. K., Bowie, E. J. W. Hemorrhagic diathesis associated with benign histiocytic cytophagic panniculitis and sytemic histiocytosis. Arch. Intern. Med. 1980,140, 146&1463. 2. Flavell, D. J., Jones, D. B., Wright, D. H. Identification of tissue histiocytes on paraffin scctions by a new monoclonal antibody. J. Histochem. Cytochem. 1987,35,1217-1226. 3. Crotty, C. P., Winkelmann, R.K. Cytophagic histiocytic panniculitis with fever, cytopenia, liver failure and terminal hemorrhagic diathesis. J. Am. Acad. Dermatol. 1981,75,734-738. 4. Willis, S. M., Fitzpatrick, J. E. Cytophagic histiocytic panniculitis. Systemic histiocytosis presenting as chronic, nonhealing, ulcerative skin lesions. Arch. Dermatol. 1985,121,91&913. 5. Pinol-aguade, J., Ferrando, J., Tomas, J. M., Mieras, C., Peyri, J. Necropsy and ultrastructural findings in histiocytic medullary reticulosis. Br. J. Dermatol. 1976; 95,3544. 6. Aronson, I. K., West, D. P., Variakojis, D., Ronan, S. G., Iossifides, I., Zeitz, H. J. Panniculitis associated with cutaneous T-cell lymphoma and cytophagocytic histiocytosis. Br. J. Dermatol. 1985,112,87-96. 7 . Aronson, I. K., West, D. P., Variakojis, D., Malkinson, F. D., Wilson, H. D., Zeitz, H. J. Fatal panniculitis. J . Am. Acad. Dermatol. 1985,12,535-551. 8. Barron, D. R.,Davis, B. R., Pomeranz, J. R., Hines, J. D., Park, C. H. Cytophagic histiocytic panniculitis. A variant of malignant histiocytosis. Cancer 1985,55,2538-2542. 9. Alegre, V. A., Fortea, J. M., Camps, C., Aliaga, A. Cytophagic histiocytic panniculitis. Case report with resolution after treatment. J. Am. Acad. Dermatol. 1989,20,875-878. 10. White, J. W., Winkelmann, R.K. Cytophagic histiocytic panniculitis is not always fatal. J . Cutan. Pathol. 1989,16,137-144.

Cytophagic histiocytic panniculitis: a diagnostic dilemma.

Cytophagic histiocytic panniculitis (CHP) is a recently recognized entity that frequently poses a perplexing diagnostic problem. Although the classica...
2MB Sizes 0 Downloads 0 Views