Researchers at Great Ormond Street Hospital, London, and Lingfield Centre for Epilepsy, Surrey, UK, studied retrospectively the seizure outcome of 75 children treated over an 8-year period for infantile spasms (IS) with prednisolone or vigabatrin. Prior to the UK Infantile Spasms Study (UKISS) (Lux et al, 2004) vigabatrin was the treatment of choice at this institution; after the study, steroids were preferred. Of 75 children treated, 24 (32%) had cryptogenic and 51 (68%) had symptomatic spasms. Mean age at onset of IS was 6 months. Response to therapy was defined as complete absence of spasms for at least 2 weeks. EEG was not used to categorize response.
IS response rate was 61.1% with steroid therapy and 42.5% with vigabatrin. Response was faster with steroids than vigabatrin, but relapse rates were similar for both. Response to first-line therapy is the main predictor of long-term seizure outcome. Of 51 requiring a second line therapy, 31 received steroids (29 a combination of vigabatrin and steroids) with a response rate of 51.6% (48.2% in the combined group). Of 11 patients receiving vigabatrin as second line therapy, 45.4% responded. Combination therapy was no more effective than vigabatrin alone (p>0.05). Cryptogenic seizures responded significantly better to steroids (100% response rate) and had a better neurodevelopmental outcome (p=0.01), whereas the symptomatic group responded to vigabatrin (48.7%) or steroids (22.2%) only partially (p>0.05). Time to treatment (< or > 28 days of onset) did not influence seizure or developmental outcome. At 12 months follow-up, 78% children with spasms had seizures of other types. 
COMMENT. Since the first report of the effectiveness of ACTH by Sorel et al in 1958, the optimal treatment for infantile spasms is still unclear. ACTH is generally considered superior, and the use of prednisolone in the above study may account for some differences in response. Also, lack of EEG monitoring may have altered the observed response rates. For example, in contrast to the present report, previous studies have shown that the earlier the treatment, younger the patient, and higher the developmental quotient, the better the seizure and EEG outcome. [2, 3]
In agreement with previous studies, the Great Ormond Street Hospital experience shows that cryptogenic in contrast to symptomatic spasms have 100% response to treatment and a good neurodevelopmental outcome. The authors conclude that steroids are the initial treatment of choice for infantile spasms, except for patients with tuberous sclerosis who respond to vigabatrin. That spasms are controlled faster with steroids than vigabatrin is a potential advantage, if the length of exposure to seizures and EEG hypsarrhythmia predicts cognitive outcome. Relapse rates with steroids or vigabatrin are similar. Combination, vigabatrin and steroid, a potential second-line therapy is not more effective than steroid or vigabatrin alone, but a larger controlled study may be indicated.
Add-on ketogenic diet. Short-term use (8 months) of the ketogenic diet 3:1 ratio as add-on treatment was successful in control of infantile spasms and hypsarrhythmia in 14 of 16 patients, without the growth disturbance that occurs with the traditional long-term (2 years) ketogenic diet.