SUMMARYObjective: There is scant evidence to guide the management of infantile spasms after successful response to initial therapies. There is significant risk of relapse, largely because effective pharmacologic treatments cannot be continued long term because of concern for significant adverse events. Zonisamide (ZNS) and topiramate (TPM) are commonly used to prevent relapse, and the purpose of this study was to specifically evaluate the efficacy of ZNS and TPM as agents for secondary prevention of infantile spasms. Methods: Patients with video-electroencephalography (EEG) confirmed resolution of infantile spasms were retrospectively identified. Relevant clinical data were systematically collected, including lead time from onset of spasms to successful treatment response, etiology of infantile spasms, number of treatment failures prior to response, timing of relapse, and detailed exposure data for ZNS and TPM. Results: We identified 106 patients with response to hormonal therapy (n = 58), vigabatrin (n = 25), or surgery (n = 23). To prevent relapse of infantile spasms, 37 patients received ZNS, 34 received TPM, 3 received both ZNS and TPM, and 38 patients received neither ZNS nor TPM. There were 44 relapses, occurring a median of 6.9 (3.2-10.8) months after initial response. Time to relapse was not affected by treatment with ZNS or TPM. Relapse was less likely among patients who were older (hazard ratio 0.97 [per month], p = 0.036) and those who responded to surgical resection (hazard ratio = 0.28, p = 0.017). Of note, we identified a relatively refractory cohort with multiple treatment failures and long lead time to initial response. Significance: In this refractory cohort, neither ZNS nor TPM was successful in preventing relapse of infantile spasms, despite relatively high dosages. At this time, aside from surgical resection in eligible candidates, there is no known treatment that is efficacious in the prevention of relapse of infantile spasms.