|Home | About | Journals | Submit | Contact Us | Français|
Eosinophilic meningitis is an uncommon condition often associated with helminthic infections. We present a patient with idiopathic eosinophilic meningoencephalomyelitis occurring 9 months after an episode of Well syndrome, an idiopathic eosinophilic cellulitis. The disorder mimics Gordon phenomenon, an experimentally induced encephalitis in animals associated with eosinophil-derived neurotoxicity.
A 55-year-old woman developed a rash on her legs. Skin biopsy revealed an eosinophilic cellulitis consistent with Well syndrome (figure, A and B). A peripheral eosinophilia of 1,134 cells/μL was present. The cellulitis resolved with a short course of oral steroids, and the eosinophilia diminished.
Nine months later, she developed vertigo and ataxia, followed by progressive dysarthria, double vision, headache, and tinnitus. A peripheral eosinophilia of 1,900 cells/μL was present. An initial brain MRI was unremarkable. Two weeks later, the MRI showed mild leptomeningeal enhancement and 2 globoid temporal white matter lesions. The CSF revealed a lymphocytic pleocytosis with 140 white blood cells and protein of 104 mg/dL. Three weeks later, the eosinophilic fraction increased (351 cells; 81% lymphocytes, 8% eosinophils, 8% monocytes). Cytology was negative for malignancy. A FiP1L1 platelet-derived growth factor receptor test was negative, making eosinophilic leukemia unlikely.1 After a short course of empiric IV methylprednisolone, the patient was transferred to our hospital.
On admission, 5 weeks after onset, her examination revealed cerebellar dysarthria, extremely slow saccades in all directions, profound dysmetria, and gait ataxia. Despite an empiric course of albendazole, these findings gradually worsened, and she developed downbeat nystagmus and mild left hemiparesis. Serial CSF examination revealed 76, 78, and 61 white blood cells, with an increasing eosinophilic fraction (4%, 24%, 33%). Oligoclonal bands were present. An infectious evaluation was negative for strongyloides, toxocara, Bartonella, enterovirus, Whipple, coccidiomycosis, cysticercosis, Epstein-Barr virus, cytomegalovirus, varicella zoster virus, JC virus, parvovirus, hepatitis B and C, HIV, Lyme, syphilis, toxoplasma, mycoplasma, rickettsia, and histoplasmosis. Rheumatologic serologies and paraneoplastic testing were unremarkable. Whole body CT scan and breast ultrasound revealed no abnormalities. MRI 8 weeks after presentation revealed mild cerebellar vermal enhancement, new enhancement of the right temporal lesion, and 2 nonenhancing cervical spine lesions. Biopsy of the right temporal lobe revealed a brisk leptomeningeal, perivascular, and intraparenchymal mixed infiltrate with florid gliosis (figure, C and D). High-dose IV methylprednisolone was initiated 1 week after completing albendazole, and the patient's saccades improved rapidly. She was discharged on oral prednisone. The eosinophilic cellulitis recurred after tapering the prednisone several months later, but resolved after reinitiating it. Eighteen months after presentation, she continued to improve neurologically, with mild downbeat nystagmus, dysmetria, and ataxia. The white matter lesions improved and enhancement resolved on MRI. Peripheral eosinophilia has not recurred.
Our patient presented with eosinophilic meningoencephalomyelitis 9 months after a bout of Well syndrome, an eosinophilic cellulitis. Well syndrome may occur as reaction to medications or malignancy, or be idiopathic, as in our patient.2 One of the unusual features of our patient was the prominent cerebellar findings in the absence of an obvious cerebellar inflammatory infiltrate. This observation is consistent with experimental evidence that neurologic damage associated with eosinophilia may result from neurotoxic factors such as eosinophil-derived-neurotoxin (EDN) and the eosinophil cation protein (ECP).
The association between eosinophils and neurotoxicity was discovered serendipitously after Gordon3 injected lymph node suspensions from patients with Hodgkin disease into rabbits’ thecal sacs. Some of the rabbits developed a neurologic syndrome characterized by diffuse spongiform demyelination, most prominent in the cerebellum, brainstem, and spinal cord.4 This “Gordon phenomenon” was only observed when an eosinophilic infiltrate was found in the lymph nodes. Subsequently, Durack et al.5 purified the specific toxin, EDN. Injection of purified EDN in rabbits produces similar clinical and histopathologic findings found in the Gordon phenomenon including ataxia. Pathologically, spongiform lesions in the white matter and loss of cerebellar Purkinje cells have been noted. Our patient had prominent cerebellar findings without an obvious inflammatory disease burden in this location, suggesting that some of her deficits may have resulted from the remote effect of eosinophilic neurotoxicity.
Recently, ECP was demonstrated in the eosinophils of an inflammatory infiltrate from a brain biopsy of a patient with idiopathic hypereosinophilic syndrome. Histologic examination revealed perivascular and parenchymal inflammatory infiltrates without tissue or neuronal damage.6 The lack of correlation between the eosinophilic infiltrates and tissue damage seen in eosinophilic encephalitis along with the role of the EDN and ECP in experimental and clinical cases suggests that nervous system injury associated with eosinophilia may result from remote injury from toxins secreted by such cells.
Disclosure: Dr. Turkeltaub receives salary support and educational funds from the American Academy of Neurology Foundation. Dr. Guzman reports no disclosures. Dr. Lee has a pending patent related to antibody treatment of Alzheimer and related diseases and has received support from the NIH [T32 AG00255 (postdoctoral fellowship)]. Dr. Galetta serves on the editorial boards of Neurology® and the Journal of Neuro-Ophthalmology; served as a consultant for Medtronic, Inc.; serves on a speakers’ bureau for Biogen Idec; and has received speaker honoraria from Teva Pharmaceutical Industries Ltd. and Biogen Idec.
Received July 15, 2009. Accepted in final form August 19, 2009.
Address correspondence and reprint requests to Dr. Peter E. Turkeltaub, Department of Neurology, University of Pennsylvania Health System, 3400 Spruce Street, 3 West Gates Building, Philadelphia, PA 19104; firstname.lastname@example.org