PHENYTOIN AND CEREBELLAR ATAXIA: NOT ALL DOWN TO TOXICITY?

Introduction Phenytoin is an effective anticonvulsant for focal epilepsy with the advantage of single daily administration. Its popularity however is in decline, partly because of perception of adverse effects that include gum hypertrophy, hirsutism, osteoporosis and cerebellar ataxia. Whilst phenytoin has been shown to be toxic to cerebellar Purkinje cells in vitro, the frequency, clinical and radiological phenotype and the mechanism of cerebellar degeneration in vivo remains unclear. In this study we aim to identify the frequency and characterise the clinical and radiological phenotype of patients with phenytoin–associated ataxia. Methods Patients were recruited from the epilepsy clinics, Royal Hallamshire Hospital, Sheffield, UK. HLA type and tests for other causes of ataxia were undertaken. The severity of the ataxia was rated using the Scale for the Assessment and Rating of Ataxia (SARA). Patients with clinical evidence of ataxia underwent a 3T MRI including spectroscopy of the cerebellum and had serum phenytoin levels estimated. Results Twenty five patients aged between 32–77 years were recruited. Eleven had focal epilepsy, 6 generalised and 8 unclassified. Age at diagnosis ranged from childhood to 74 years with duration 2–67 years. Fifteen patients (60%) continued taking phenytoin from the time of diagnosis. Mean duration of phenytoin treatment was 21 years and mean dose 350mg. Seventeen patients (68%) complained of poor balance and 13/25 (52%) had clinical evidence of ataxia on examination. All 13 had gait ataxia and the majority also had limb ataxia. SARA scores ranged from 3–17 (median 6). Gait, stance and heel–shin slide subgroups of the SARA scale were mostly affected. Eight out of 17 (47%) patients had the HLA DQ2/DQ8 type, strongly associated with autoimmune ataxias. Volumetric 3T MRI imaging and spectroscopy data were available in 6 of the 13 patients with ataxia. Cerebellar atrophy and abnormal spectroscopy of the vermis was found in three patients. Two of these patients had circulating antigliadin antibodies that signify sensitivity to gluten. Two patients with no ataxia but who had taken phenytoin for more than 20 years had no cerebellar atrophy and normal spectroscopy. Mean phenytoin level was 18 mg/L (5–20 mg/L). One patient had a serum phenytoin level of 65.8 mg/L. Conclusion 68% of patients on long–term phenytoin complained of poor balance and 52% had clinical evidence of ataxia. Cerebellar abnormalities on imaging were seen in 3/6 (50%) patients with ataxia. The primarily vermian involvement seen is similar to that of autoimmune ataxias. In those patients with vermian abnormalities an additional potential cause of ataxia was identified in the form of serological evidence of sensitivity to gluten. It is possible that the mechanism of cerebellar damage in patients on long term phenytoin may involve other mechanisms such as immune mediated damage in genetically (HLA DQ2/DQ8) susceptible individuals and/or sensitivity to gluten.