Abstracts

Rationale: Rasmussen syndrome (RS) is a chronic progressive unilateral inflammation of the brain. Anti-epileptic drugs (AEDs) are only partially effective. We describe the use of azathioprine (AZA), in the management of RS. Methods: 16 children with a clinical diagnosis of RS (focal epilepsy, progressive motor deficit, and progressive unilateral changes on MRI brain scan or biopsy), received open-label AZA, 1.5mg/kg/day orally, with oral prednisolone 2mg/kg/day for six weeks, then response review and attempt to wean. AZA-response was monitored; specifically seizure control, motor dysfunction, neuropsychological cognitive assessment, annual neuro-imaging, time to epilepsy surgery, and biochemical and haematological surveillance for adverse effects. Outcome was compared to data from 13 historical controls with RS, seen between 1987 and 2005, and not treated with AZA. Results: 16 children were treated with AZA (7 male, 9 female; right hemisphere disease in 9, left in 7). The mean age of symptom onset was 7.1 years (median 7.0, range 3.0-12.6 ); historical controls mean 6.6 years (median 5.8, range 3 - 13.1). All had refractory epilepsy with a mean number of AEDs failed 6, range 2-13 (historical controls mean 6, range 3-9). 3/7 had shown a partial but unsustained response to IV immunoglobulin. The mean duration of follow-up of 13/16 AZA-responders was 4.9 years, median 3.5, range 1.6-15.5. No patient experienced drug side-effects. Seizure frequency was reduced by 50% or more in 13/16 children. In 7/13, subsequent further wean of steroids was possible. The three non-responders had rapidly escalating courses, and promptly progressed to surgery. Of 11 children with serial neuropsychological assessments at yearly intervals on AZA, 10/11 demonstrated further clear cognitive decline, one child stabilised. Development of hemiparesis was observed in 15/17; mean time 3.7 years, median 3.8, range 0.8-5.7 (historical controls mean 1.46 years, median 0.5, range 0.1-4.9). Serial MRI brain studies in AZA-responders showed decreased rate of atrophy and signal change compared to historical controls. 8/13 AZA-responders have come to surgery with time from diagnosis to first surgery at a mean of 6.0 years (median 5 years, range 2-13); historical controls mean 3.2 years (range 0.3 7.7). Conclusions: In this small open label study, AZA was a safe therapy in RS. Seizure frequency decreased. Development of hemiparesis and progression of MRI brain features were delayed, probably contributing to increased time to come to surgery. However, cognitive decline continued. While AZA was initially introduced as a steroid sparing agent, immunocytological studies have shown involvement of both B-cell and T-cell mediated processes in RS. AZA is a cytotoxic immunosuppressant, particularly suppressing T-cells. Further studies are needed to examine the relative benefits of AZA and other immune therapies in RS and to evaluate to what degree they may be modifying the clinical course. It is not known if delaying time to surgery, in the setting of further cognitive decline, will have an impact on long term functional outcome.