Journal of Neuro-Ophthalmology, March 1991, Volume 11, Issue 1

Reactive lymphohistiocytosis with recurrence in the optic chiasm.

Histiocytic infiltration of the optic chiasm is rare. We report a patient with a seizure disorder on anticonvulsant therapy, who developed systemic reactive histiocytosis. Treatment with splenectomy, corticosteroids, and anticonvulsant medication change resulted in clinical remission. The patient later developed recurrent lymphohistiocytosis restricted to the optic chiasm. Extensive re-evaluation yielded no evidence for infectious or malignant etiology. Radiation therapy and withdrawal of anticonvulsant therapy resulted in clinical remission. We conclude that both phenytoin and phenobarbital may trigger a lymphohistiocytic process and that infiltration of the visual pathway is possible.

lou mal of Clinical Neuro- ol'htiullmology 11( 1): 25- 30, 1991. © 1991 Raven Press, Ltd., New York Reactive Lymphohistiocytosis with Recurrence in the Optic Chiasm Steven L. Galetta, M. D., Edward A. Stadtmauer, M. D., David G. Hicks, M. D., Eric C. Raps, M. D., Gregory Plock, M. D., and J. Carl Oberholtzer, M. D., ph. D. Histiocytic infiltration of the optic chiasm is rare. We report a patient with a seizure disorder on anticonvul­sant therapy, who developed systemic reactive histiocy­tosis. Treatment with splenectomy, corticosteroids, and anticonvulsant medication change resulted in clinical re­mission. The patient later developed recurrent lympho­histiocytosis restricted to the optic chiasm. Extensive re­evaluation yielded no evidence for infectious or malig­nant etiology. Radiation therapy and withdrawal of anticonvulsant therapy resulted in clinical remission. We conclude that both phenytoin and phenobarbital may trigger a lymphohistiocytic process and that infil­tration of the visual pathway is possible. Key Words: Lymphohistiocytosis- Optic chiasm­Phenytoin- Phenobarbital. From the Departments of Neurology ( S. L. G., E. CR., G. P.), Pathology ( D. G. H., J. CO.), and Hematology-. Oncology ( E. A. S.), University of Pennsylvama School of MedlCme, Phil­adelphia, Pennsylvania, U. S. A. Address correspondence and reprint requests to Dr. Steven L. Galetta at Hospital of the University of Pennsylvama, De­partment of Neurology, 3400 Spruce Street, Philadelphia, PA 19104, U. S. A. 25 Histiocytic disorders are diverse conditions that result from an abnormal proliferation of Langer­hans or non- Langerhans macrophages. Histiocyto­sis X refers to the idiopathic Langerhans cell dis­orders that include eosinophilic granuloma of bone, Hand- Schuller- Christian disease and Let­terer- Siwe disease ( 1,2). Langerhans macrophages are distinguished by the presence of Birbeck gran­ules on electron microscopy and the detection of OKT6 cell surface antigens by immunologic tech­niques ( 2). In contrast, non- Langerhans cell prolif­eration may be neoplastic, paraneoplastic, familial, reactive ( immunologic activation of the mononu­clear phagocytic system), or the consequence of an immunodeficiency disorder ( 1,2). A concurrent lymphocytic infiltration may be present ( 3). In 1979, Risdall described the clinical syndrome of reactive histiocytic hemophagocytosis and pan­cytopenia associated with various viral infections ( 4). Subsequently, the spectrum of organisms be­lieved to stimulate a non- Langerhans histiocytic proliferation broadened to include infection with bacteria, fungi, mycobacteria, rickettsia, and other parasites ( 5). Agents such as beryllium, zirconium, and phenytoin have also been reported to trigger a histiocytic process ( 1,6). The proliferating histiocytes in a reactive process are morphologically mature, cytologically benign, and display prominent erythrophagocytosis. In contrast, a malignant infiltrate usually consists of immature and cytologically malignant cells exhib­iting little or no evidence of hemophagocytosis ( 4). We present the clinical, radiologic, and pathologic features of a patient who developed a reactive non­Langerhans histiocytosis of the optic chiasm. 26 S. L. GALETTA ET AL. CASE SUMMARY A 30- year- old male was well until 1980, when he sustained a skull fracture and cerebral contusion after being struck by a baseball bat. A post­traumatic seizure disorder was treated with phe­nytoin. The patient did well until January, 1988, when he developed night sweats, fevers, abdom­inal pain, and a 13- pound weight loss. He was admitted to an outside hospital with a temperature of 102° F and marked splenomegaly. Laboratory data at admission revealed a leuco­cyte count of 900 cells/ mm3 , with 15% segmented neutrophils, 9% bands, 69% lymphocytes, 1% monocytes, and 4% metamyelocytes. The hemo­globin was 6.7 g/ dl and the platelet count was 22,000/ mm3 . Serum chemistries were remarkable for a lactate dehydrogenase of 509 U/ L, aspartate aminotransferase 136 U/ L, alanine aminotrans­ferase 170 U/ L, a bilirubin 0.5 mg/ dt and a reticu­locyte count of 1%. The direct Coombs, hepatitis surface antigen, and monospot tests were all neg­ative. The sedimentation rate was 2 mrn/ h; LE­prep and rheumatoid factor were negative. Cyto­megalovirus and herpes simplex titers were nega­tive. Epstein- Barr virus, varicella zoster virus, and toxoplasmosis studies were consistent with past or inactive infection. A stool culture was negative for parasites and pathogenic bacteria. Chest radio­graph was normal. Computed tomography ( CT) of the abdomen and pelvis was significant for marked hepatosplenomegaly without lymphadenopathy. A bone marrow biopsy revealed a hypercellular marrow with decreased hematopoietic cells, an in­filtrate of small lymphocytes and histiocytes, ab­normal megakaryocytes, and focal fibrosis. This was felt to be nondiagnostic. A liver biopsy showed a mild peri- portal lymphoid infiltrate and extramedullary hematopoiesis. The patient was transferred to the Hospital of the University of Pennsylvania on February 19, 1988, for further evaluation of his worsening pan­cytopenia. The diagnosis of lymphoma was con­sidered, and the patient underwent an exploratory laparotomy and resection of a massively enlarged spleen. Pathologic examination of the spleen re­vealed extramedullary hematopoiesis, an infiltrate of benign- appearing histiocytes exhibiting marked erythrophagocytosis, and small mature lympho­cytes, consistent with a reactive histiocytosis ( Fig. 1A). Lymphocyte subset analysis performed by both immunohistochemistry and flow cytometry showed a predominance of T cells with no abnor­mal phenotypes and with a normal helper­to- suppressor ratio. Similarly, gene rearrangement studies nf th:: BChalf! of the T cell receptor failed to ? 1 ',,',", 1 I '- I ~~ I show evidence of monoclonal or otherwise abnor­mal cell population. S- 100 staining of the splenic tissue was negative. The abdominal lymph nodes sampled were histologically normal except for the presence of hemophagocytosis wit~ . medull~ sinuses. The patient experienced significant clini­cal and laboratory improvement on prednisone af­ter splenectomy. In addition, his phenytoin was switched to phenobarbital therapy. His platelet count and white count returned to normal, and prednisone was discontinued in March, 1989. In August, 1989, the patient presented to the Neuro- ophthalmology service with decreasing vi­sion of several weeks' duration. There was no fam­ily history of lymphohistiocytosis. His examination was notable for a visual acuity of 20/ 30 in the right eye and 20/ 40 in the left. Visual fields revealed a central scotoma in the right eye and an inferior temporal defect in the left eye that respected the vertical meridian. Brain magnetic resonance imag­ing ( MRI) revealed enlargement and enhancement of the optic chiasm ( Fig. 2), and a two- week em­piric trial of 60 mg of prednisone was adminis­tered. Despite corticosteroids, the patient's visual acuity dropped to 20/ 200 in the right eye and 20/ 80 in the left. He was admitted for intravenous pulse steroid therapy. During his hospital course, his vi­sual acuity improved to 20/ 80 in the right eye and 20/ 50 in the left. The repeat bone marrow, chest, and abdominal CT, complete blood count, and se­rum chemistry panel were unremarkable. A puri­fied protein derivative skin test ( PPO) was nega­tive with a positive skin reaction to mumps. The patient had three lumbar punctures and the results showed WBC ( cells/ mm3 ) 6,4,4, ( greater than 90% lymphocyte predominance); glucose ( mg/ dl) 57, 67, 137; protein ( mg/ dl) 57, 69, 36, respectively. The cerebrospinal fluid ( CSF), cytology, and cul­tures were negative on all specimens, and the elec­trophoresis showed an elevated total protein, but normal CSF- IgG ratio and absent oligocolonal bands. A repeat MRI of the brain was unchanged. The patient was discharged on September 18, 1989, on an oral regimen of prednisone 60 mg. In October, 1989, his visual function was stable and prednisone was tapered to 40 mg a day. The MRI showed marked improvement in the patient's chi­asm lesion with a 50% decline in size ( Fig. 3). In November, 1989, his visual acuity was 20/ 50 0.0., and 20/ 40 0.5., and steroids were tapered to 30 mg a day. In early January, he was noted to have a visual acuity of 20/ 50 in the right eye and 20/ 200 in the left, with a bitemporal hemianopsia. He was readmitted for intravenous pulse corticosteroid therapy. Human immunodeficiency virus, HTLV­1, Lyme, MHA- TP, brucella, and rickettsia titers / LYMPHOHISTIOCYTOSIS IN THE OPTIC CHIASM FIG. 1. Histopathology of Iymphohistiocytic infiltrate. A: Hematoxylin- eosin micro­graph of spleen showing atrophy of the white pulp, extramedullary hematopoiesis and expansion of sinuses by infiltrate of lymphocytes and histiocytes. Inset demonstrates erythrophagocytosis ( arrow). Original magnification x40 and x400, respectively. B: Hematoxylin- eosin micrograph of optic chiasm biopsy showing diffuse infiltration by mononuclear cells. Inset demonstrating infiltrate of mature lymphocytes and benign histiocytes. Original magnification x50 and x400, respectively. 27 were negative. A repeat lumbar puncture showed 2 white blood cells/ mm3 , glucose of 60 mg/ dl, pro­tein of 82 mg/ dl, with negative cultures and cytol­ogy. An MRI of the brain revealed enhancement of " ­, the left optic nerve and chiasm junction. Visual function did not improve with corticosteroid ther­apy, and the patient underwent biopsy of the left optic nerve- chiasm junction. Histology revealed a JClin Neuro- ophthalmol, Vol. 11, No. 1, 1991 28 A S. L. GALETTA ET AL. FIG. 2. T1- weighted MR image with gadolinium demonstrating enlarged and diffusely enhancing optic chiasm in saggital ( A) and coronal ( 8) planes ( arrows). 8 reactive lymphocytic and histiocytic infiltration similar to the original pathology ( Fig. 1B). Electron microscopy showed benign non- Langerhans mac­rophages without viral inclusions. No Birbeck granules were seen. Immunohistologic studies showed a mixed population of mature T lympho-cytes expressing a normal phenotype. A COl1 stain confirmed the presence of benign- appearing macrophages ( Table 1). The patient was treated with radiation therapy 3,600 cGy to the optic chiasm and visual acuity remained stable at 20/ 500.0. and 20/ 200 0.5. Phe- A 8 FIG. 3. T1- weighted MR image with gadolinium showing resolution of optic chiasm enhancement and enlargement following corticosteroid therapy. Saggital ( A) and coronal ( 8). //.-----------......... LYMPHOHISTIOCYTOSIS IN THE OPTIC CHIASM 29 DISCUSSION TABLE 1. Immunohistochemical subset analysis of optic chiasm infiltrate nobarbital and corticosteroid therapy have been discontinued without recurrence of seizures or neurologic dysfunction. Repeat MRl 6 months and 1 year after radiation therapy showed no gadolin­ium enhancement in the optic chiasm region. Fol­lOw- up evaluation 1 year after surgery disclosed no new systemic complaints, and visual function re­mained unchanged. Our patient developed an unusual histiocytic syndrome initially associated with pancytopenia, hepatosplenomegaly, extramedullary hematopoie­sis, erythrophagocytosis, and lymphadenopathy. A complete systemic remission was achieved with corticosteroid therapy and discontinuation of phe­nytoin therapy. However, S months after stopping corticosteroids, and while on phenobarbital, the patient developed progressive visual decline sug­gesting a chiasmal process. A subsequent biopsy revealed mature cytologically benign histiocytes and lymphocytes infiltrating the optic chiasm. Some features of this patient's course suggested the possibility of a malignant hematopoietic pro­cess. The distinction between a reactive and a ma­lignant histiocytosis is based on histologic and cy­tologic features of the infiltrating histiocyte, an ap­proach necessary because of the lack of clonal markers for malignant histiocytes ( 2,4,5). This dis­tinction is complicated by the fact that cases of re­active histiocytosis may have a fulminant presen­tation and may clinically mimic a malignancy ( 4,7). The pathologic material from the current case was interpreted as a reactive histiocytosis on the basis of the mature cytology of the proliferating histio­cytes and the absence of nuclear pleomorphism or atypia. The patient's clinical course also supports a reactive disorder rather than a malignant one. In a series reported by Warnke ( 8), the median survival was 6 months and the mean survival was 14 months for patients with malignant histiocytosis. The natural history of malignant histiocytosis is one of dissemination, multivisceral involvement, and progression to death ( 2). Complete remission may be induced by combination chemotherapy, but in general the prognosis is poor ( 8,9), and it is doubtful that the symptom resolution seen in this patient would occur with malignant histiocytosis. Although an infectious etiology was also consid­ered, a microbiologic evaluation proved unreveal­ing. The biphasic nature of our patient's course is atypical for a virus- associated reactive histiocytic syndrome ( 4,5). In the presence of benign histio­cytes, the possibility of an anticonvulsant- induced lymphohistiocytic syndrome was suggested. The ability of phenytoin to induce a dysregulation in the hematopoietic system and trigger a lymphohis­tiocytic syndrome is well established ( 10). Pheno­barbital, a structurally similar compound, may rarely cause a similar reaction, including a pseudo­lymphoma syndrome ( 11). The distinction be­tween large atypical lymphocytes and true histio­cytes and their respective contributing roles in the pathology of pesudolymphoma and other similar syndromes has only recently been clarified by the routine use of cell surface marker studies ( 2). Phenytoin hypersensitivity is characterized by the triad of fever, lymphadenopathy, and skin rash. However, these signs may be present to a variable degree and be associated with splenomeg­aly, arthralgias, and overt hepatitis. The onset of this syndrome is usually within several weeks of the introduction of the medication, but it can be delayed for months to years. In 1959 Saltzein and Ackerman reported 7 patients with a pseudolym­phoma syndrome secondary to phenytoin and found 75 similar cases in a review of the literature ( 12). Subsequently, one of the seven patients in this report developed a malignant lymphoma ( 13). In 1966, Hymann and Sommers reported 6 pa­tients who developed malignant lymphoma while on anticonvulsant therapy ( 14). Four patients were taking phenobarbital concurrently. During a four­year period, they found that phenytoin was more frequently associated with the development of true lymphoma than with pseudolymphoma. In 1968, Gams et al. reported a patient with a seizure disorder on phenytoin and phenobarbital who developed reactive adenitis 6 months after initiation of anticonvulsant therapy ( 15). Five months later, while being maintained on this ther­apy, he developed a substernal mass lesion with a biopsy revealing malignant lymphoma. Phenytoin was discontinued and his lymphadenopathy re- Rare + cells Scattered + cells 2 + larger oval cells Scattered + cells Negative 1+ larger oval cells 3 + small lymphoid cells 2 + small lymphoid cells 3 + small lymphoid cells 1- 2 + small lymphoid cells Not done T cell markers CD1 ( T6) C02 ( T11) CD4 ( T4) CDS ( T1) CD? ( Leu 9) CD8 ( T8) B cell markers CD19 ( pre- B) CD20 ( B1) Other CD11 ( macrophages) CD1S ( LeuM1) S- 100 ( spleen tissue) ,---.------==_...~.-=--."..".=.. ".'.--" JClin Neuro- ophthDlmol. Vol. 11. No. 1. 1991 30 S. I. GALETTA ET AI. ceded. Twenty months later, on phenobarbital therapy, he developed recurrence of malignant lymphoma and died. Cams et al. divided the pseudolymphoma syn­drome into four categories including hyperplasia, pseudolymphoma, malignant lymphoma, and pseudo- pseudolymphoma. In the latter group, a typical patient will develop lymph node hyperpla­sia with a transient regression when off medica­tions and then have a subsequent malignant re­lapse. As with our patient, all 3 patients grouped under the pseudo- pseudolymphoma syndrome in that study were treated with either phenobarbital or mysoline therapy following discontinuation of phenytoin ( 12,15). In 1970, Anthony charted 41 cases of lympho­adenopathy associated with phenytoin and found that adenopathy may occur within a few hours of dosing up to 10 years from onset of therapy ( 16). He reviewed 85 cases of malignant lymphoma at his institution and found that 4 cases were associ­ated with phenytoin therapy. He concluded that the frequency of lymphoma in patients on anticon­vulsants was 10 times above the frequency ex­pected. The mechanism of the phenytoin pseudolym­phoma syndrome has been examined in several studies. Sorrell examined 63 patients on phenytoin and compared them with 28 lymphoma patients ( 17). He found depressed cellular and humoral im­munity in both groups. Decreased antibody re­sponses to Salmonella typhi, decreased lymphocyte reaction to phytohemagglutinin, and decreased delayed hypersensitivity were found. Rosenthal et al. found increased absolute numbers of T lympho­cytes in patients with phenytoin- induced mycosis fungoides ( 18). In addition, phenytoin use was as­sociated with lymphocyte blastic transformation and an impaired T suppressor lymphocyte re­sponse. In a controlled study, the presence of arene oxide metabolites were implicated as a cause of phenytoin- induced hepatotoxicity ( 19). More re­cently, Knutsen described the features of a pheno­barbital- induced hypersensitivity syndrome and emphasized the structural similarities between phenobarbital and phenytoin ( 20). In susceptible individuals, the ability to metabolize arene oxide metabolites may lead to both lymphocyte and his­tiocyte dysregulation. We postulate that our patient's initial lympho­histiocytic process was induced by phenytoin, and, as in the pseudolymphoma syndrome, remit­ted with discontinuation of phenytoin. Subse­quently, our patient developed a recurrence in the optic chiasm while being maintained on phenobar- JClin Neuro- ophthalmol, Vol. 11, No. 1, 1991 bital therapy alone. Although we cam~ ot ~ l1y e~ ­clude an infectious etiology, our mIcrobIOlogIc evaluation was unrevealing, and the recurrence of a lymphohistiocytic process would ~ e atypical for a virus- associated hemophagocytIc syndrome. Given the structural similarities of phenobarbital and phenytoin, we suggest the us~ of c~ rba­mazepine or valproic acid as alternative anticon­vulsant agents when a patient develo~ s a lympho­histiocytic disorder while on phenytOIn or pheno­barbital therapy. REFERENCES 1. Groopman JE, Golde OW. The histiocytic disorders: a pathophysiologic analysis. Ann Intern Med 1981; 94: 95- 107. 2. Nezelof C, Barbey S. Histiocytosis: nosology and pathiobi­ology. Pediatr Path 1985; 3: 1-- 41. 3. Goldberg J, Nezelof C. Lymphohistiocytosis: a multifacto­rial syndrome of macrophagic activation clinicopathological study of 38 cases. Hematol OncoI1986; 4: 275-- 89. 4. Risdall RJ, McKenna RW, Nesbit ME. Virus- associated he­mophagocytic syndrome. Cancer 1979; 44: 993- 1002. 5. Reiner AP, Spivak JL. Hematophagic histiocytosis. Medicine 1988; 67: 369-- 88. 6. Tsan MF, Mehlman OJ, Green RS, Bell W. Dilantin, agran­ulocytosis and phagocytic marrow histiocytes. Ann Intern Med 1976; 84: 710- 11. 7. Chan, JKC, Ng CS, Law CK, Ng WF, Wong KF. Reactive hemophagocytic syndrome: a study of 7 fatal cases. Pathol­ogy 1987; 19: 43- 50. 8. Warnke RA, Kim H, Dorfman RF. Malignant histiocytosis ( histiocytic medullary reticulosis) I. Clinicopathologic study of 29 cases. Cancer 1975; 35: 215- 30. 9. Zucker JM, Caillaux JM, Vanel 0, Gerard- Marchant R. Ma­lignant histiocytosis in childhood: clinical study and thera­peutic results in 22 cases. Cancer 1980; 45: 2821- 9. 10. Stanley J, Fallon- Pellicci V. Phenytoin hypersensitivity re­action. Arch DermatoI1978; 114: 1350- 3. 11. Black OM, Fivenson DP. Case report of antiepileptic drug hypersensitivity reaction: pseudolymphoma syndrome. Fam Pract Res J1989; 8: 107- 11. 12. Saltzstein SL, Ackerman LV. Lymphadenopathy induced by anticonvulsant drugs and mimicking clinically and pathologically malignant lymphomas. Cancer 1959; 12: 164­82. 13. Harrington WJ, Kissane J, Saltzein SL, et al. Clinicopatho­logic conference. Am JMed 1962; 32: 286- 97. 14. Hyman GA, Sommers Sc. The development of Hodgkin's disease and lymphoma during anticonvulsant therapy. Blood 1966; 28: 416- 27. 15. Gams RA, Neal JA, Conrad FG. Hydantoin- induced pseudo- pseudolymphoma. Ann Intern Med 1968; 69: 557- 68. 16. Anthony JJ. Malignant lymphoma associated with hydan­toin drugs. Arch NeuroI1970; 22: 450-- 4. 17. Sorrell TC, Forbes IJ, Burness FR, Rischbieth RHC. Depres­sion of immunological function in patients treated with phenytoin sodium ( sodium diphenylhydantoin). Cancer 1971; 1: 1233- 5. 18. Rosenthal q, Noguera CA, Coppola A, Kapelner SN. Pseudolymphoma with mycosis fungoides manifestations, hyperresponsiveness to diphenylhydantoin and lympho­cyte disregulation. Cancer 1982; 49: 2305- 14. 19. Spielberg SP, Gordon GB, Black DA, Goldstein DA, Her­long HF. Predisposition to phenytoin hepatotoxicity as­sessed in vitro. N Engl JMed 1981; 305: 722- 7. 20. Knutsen AP, Anderson J, Satayaviboon 5, Slavin RG. Im­munologic aspects of phenobarbital hypersensitivity. JPe­diatr 1984; 105: 558- 63. /:?>-----..,...