Few population-based studies of infantile spasms have been done, and most reports have not included comparison groups. In spite of these limitations, this review attempts to summarize what is known about the distribution, etiology, and natural history of infantile spasms in populations; discusses the limitations of current data; and includes suggestions for further population-based research. Most estimates of the incidence of infantile spasms are between 0.25 and 0.42 per 1000 live births per year. Among children less than 10 years of age, the annual prevalence of infantile spasms is 0.14 to 0.19 per 1000. The peak age at onset of spasms is 4 to 6 months, and there appears to be a slight excess of male cases. The etiology of infantile spasms is unknown for 40% to 50% of affected children. Selected syndromes (eg, Aicardi syndrome, Down syndrome, etc) and inherited disorders (eg, tuberous sclerosis) may account for as many as one third of cases. Other factors, such as intrauterine infection, "birth injury," and head trauma have not been systematically evaluated, and thus, their contribution to the etiology of infantile spasms is uncertain. Areas for future etiologic research include controlled studies of immunologic factors and in utero and postnatal infections and further exploration of the interrelationship between infantile spasms and Lennox-Gastaut syndrome.