Clinicopathological and genotypic aspects of anticonvulsant-induced pseudolymphoma syndrome

Background Pseudolymphoma syndrome (PLS) is relatively rare but can lead to death if there are extensive skin lesions, severe hepatitis, agranulocytosis and neutropenia. PLS may also give rise to harmful effects if misdiagnosed as malignant lymphoma and patients with PLS are treated unnecessarily with chemotherapy, because it may mimic histologically other lymphomas, including mycosis fungoides (MF). Objectives To examine the clinicopathological and genotypic features of anticonvulsant-induced PLS. Patients and methods We retrospectively reviewed clinical, laboratory and histological findings for eight cases of anticonvulsant-induced PLS, and performed T-cell receptor gene rearrangement using polymerase chain reaction with paraffin-embedded specimens from each case. Results The causative agents were carbamazepine (four cases), phenytoin (two cases), phenobarbital (one case) and valproic acid (one case). A cross-reaction between phenobarbital and phenytoin was observed in one case. The duration from the start of anticonvulsant therapy to skin eruption was 3-24 weeks (mean 7 weeks). The skin lesions were generalized maculopapular eruptions in all cases, including one case accompanied by vesiculopustular lesions. The frequencies of the associated features were as follows: facial oedema (88%), fever (75%), lymphadenopathy (63%), and hepatomegaly (25%). Laboratory findings revealed leukocytosis, atypical lymphocytes, eosinophilia, monocytosis, neutrophilia, lymphocytosis and abnormal liver function. Histopathologically, there was similarity between PLS and MF in that epidermotrophism of atypical lymphocytes (100%) and Pautrier's microabscess-like structures (38%) were observed. However, PLS has some differences from MF that include moderate to marked spongiosis (75%), necrotic keratinocytes (63%), and infiltration of eosinophils (25%) in the epidermis and, in the dermis, papillary dermal oedema (100%), extravasated erythrocytes (100%), lymphocytes within the dermis larger than those within the epidermis (63%), and infiltration of various inflammatory cells including neutrophils (50%). Genotypic analysis demonstrated a rearrangement of the T-cell receptor-gamma gene in one of eight cases studied. There were no deaths and all cases were improved at 2-9 weeks (mean 6 weeks), after the cessation of causative agents, systemic and topical corticosteroid therapy, and symptomatic therapy. There were no significant differences in clinical, laboratory and histological findings between the causative agents. Conclusions PLS may show histopathological findings similar to MF and take a prolonged course even after the cessation of causative agents. Thus, a clear understanding and diagnosis of this disease is considered to have an important effect on treatment and prognosis.