Infantile Spasms

by Michael Wong, MD, PhD
Dr. Wong is an Assistant Professor of Neurology at Washington University School of Medicine and an epileptologist in the Pediatric Epilepsy Center at St. Louis Children’s Hospital.


In 1841, West originally reported infantile spasms as an unusual and severe form of convulsions witnessed in his own son. Despite the medical advances over the past century and a half, infantile spasms remain one of the most complex and challenging conditions in pediatric neurology. This article will review the major features of infantile spasms, including clinical features, etiology, treatment, and prognosis.

Clinical Features

Even the terminology surrounding infantile spasms is complex, as the term infantile spasms has been used to refer to both a specific seizure type and a complete epilepsy syndrome. The seizure type of infantile spasms (also referred to as epileptic spasms or simply spasms) is characterized by brief, but often repetitive, muscle contractions usually involving the head, trunk, and extremities. The epilepsy syndrome of infantile spasms (also known as West syndrome) consists of the classic triad of the epileptic spasms, hypsarrhythmia on EEG, and mental retardation.

Spasms, the seizure type, may have variable features, but have been categorized primarily into three subtypes (flexor, extensor, and mixed flexor-extensor) based on postural manifestations and patterns of muscle involvement during the seizure. Flexor spasms involve flexion of the neck, trunk, and extremities, resulting in jack-knifing at the waist and a self-hugging motion of the arms. Extensor spasms consist of extension of the neck, trunk, and extremities. Mixed flexor-extensor spasms involve combinations of the above. While often confused with myoclonic or tonic seizures, spasms represent a distinct seizure type.

Individual spasms typically last for less than 1 second to up to 5 seconds. In many patients, spasms exhibit characteristic temporal patterns. Fifty to eighty percent of epileptic spasms occur in clusters of 2 to more than 100 seizures. Patients may have dozens of clusters and several hundred spasms per day, but individual variability in seizure frequency is often large. Although spasms rarely occur during sleep, clusters of spasms are frequently activated after awakening from sleep. Spasms are occasionally triggered by loud noises with associated arousal from drowsiness and sleep, but are not sensitive to photic stimulation.

Approximately one-third to one-half of patients with epileptic spasms also have other seizure types preceding or accompanying the onset of the spasms. Associated seizure types include partial, myoclonic, tonic, and tonic-clonic seizures. Spasms usually cease spontaneously by age 5 but are often replaced by other seizure types. Mental retardation and cerebral palsy occur in about 75% and 50% respectively of children with infantile spasms. The common combination of epileptic spasms, mental retardation, and hypsarrhythmia on EEG (see below) constitutes the classic infantile epilepsy syndrome of infantile spasms or West syndrome.

The EEG characteristics of patients with infantile spasms have been well-described. The classic interictal (baseline) EEG pattern of patients with epileptic spasms is hypsarrhythmia. Hypsarrhythmia represents a completely chaotic and disorganized background pattern consisting of high amplitude slow waves and spikes that are asynchronous, non-rhythmic, and variable in duration and topography. The spikes usually alternate randomly between focal, multifocal, and generalized discharges at different moments within a brief record. Hypsarrhythmia represents a very dynamic pattern that may change dramatically on a time scale of minutes, hours, and weeks or longer. On a larger time scale, hypsarrhythmia usually develops during early infancy and disappears by early childhood. The ictal (seizure) EEG correlates of epileptic spasms have been studied in detail using EEG-video monitoring. Kellaway et al. first described eleven different types of ictal EEG patterns consisting of various combinations of generalized sharp or slow wave discharges, generalized voltage attenuation (electrodecremental discharges), and fast activity. Overall, electrodecremental discharges represented the most common ictal feature, occurring in over 70% of recorded spasms, but could also occur in the absence of an obvious clinical seizure.


From an etiological standpoint, infantile spasms can be divided into three main categories: symptomatic, cryptogenic, and idiopathic. Because of advances in neurodiagnostic testing, symptomatic infantile spasms now represents the largest category. Patients with symptomatic infantile spasms have an identified underlying neurological disorder that directly causes the spasms. The number of neurological diseases that can result in spasms is enormous, but some of the major categories include intrauterine insults and infections, hypoxic-ischemic encephalopathy, malformations of cortical development, metabolic disorders, other genetic or chromosomal defects, meningitis, tumors, and neurocutaneous syndromes.

In both cryptogenic and idiopathic infantile spasms, no specific cause is known. However, due to the coexistence of other neurological symptoms, especially severe developmental delay, cryptogenic cases are presumed to have an unidentified underlying neurological disorder. Idiopathic cases of infantile spasms are extremely rare and, by definition, consist of patients with normal development and a good prognosis without any coexisting neurological conditions.


Due to the poor prognosis of infantile spasms, treatment is usually initiated quickly and aggressively after diagnosis, often at the risk of serious side effects, with the hope of changing the natural history of the disease. For approximately fifty years hormonal therapy with ACTH or prednisone has been the staple of treatment for infantile spasms. However, studies detailing the therapeutic properties of these compounds have been fraught with uncertainty. Many controversies still exist concerning the relative efficacy, optimal dose and timing, and predictive factors for good responsiveness to ACTH and corticosteroids. No placebo-controlled trials of ACTH or steroids have been performed. In most open-label or retrospective studies, ACTH or prednisone induces a reduction or complete cessation of spasms, as well as an improvement in the EEG, in approximately 50-75% of patients. This effect is usually realized within a couple weeks. While some studies report similar efficacy of ACTH and prednisone, others indicate that ACTH is more effective. Some patients who do not initially respond to ACTH may respond to prednisone and vice-versa. A large variety of doses of ACTH have been used, but there is no evidence that larger doses (150 units / meter2/day) are more effective than lower doses (20-30 units/day). Longer treatment periods usually do not improve remission rates. While relapses occur in about one-third to one-half of patients, a second course of ACTH is often effective.

A variety of findings have been reported with regard to predicting responsiveness to ACTH or steroids. Some studies, but not others, demonstrate that shorter time lag between diagnosis and treatment improves initial remission rate. Age of onset of the spasms has occasionally been correlated with treatment efficacy, with later onset (>8 months) having a better seizure control. Whether etiology of the infantile spasms influences responsiveness to treatment is also controversial, as some studies report equal efficacy in symptomatic and cryptogenic groups, but others find a better response in the cryptogenic group. Finally, the effect of hormonal therapy on long-term neurodevelopmental outcome is unclear. While some studies report an association between initial responsiveness to ACTH and improved long-term intellectual development, others found no significant difference in prognosis between initial responders and non-responders to hormonal therapy.

Hormonal therapy with ACTH or corticosteroids may have significant, potentially fatal, side effects. While weight gain and cushingoid features are among the most common side effects, ACTH or corticosteroids may also produce hypertension, metabolic abnormalities, severe irritability, osteoporosis, sepsis, and congestive heart failure. Given the serious morbidity of hormonal therapy, a number of other therapies have received attention for infantile spasms.

Among conventional antiepileptic drugs, valproate and nitrazepam have been shown to be effective as first-line therapy in placebo-controlled clinical trials of spasms. Newer antiepileptic drugs, such as felbamate, lamotrigine, topirimate, and zonisamide have been tried with some success as adjunctive therapy in infantile spasms. Few studies have directly compared these other agents with hormonal therapy, making conclusions about relative efficacy difficult. The most exciting recent development in the treatment of infantile spasms had been the emergence of vigabatrin as a potential first-line therapy comparable to ACTH, as supported by multiple studies. Vigabatrin may be especially effective for spasms in patients with tuberous sclerosis, with some series reporting complete control in about 95% of patients. Unfortunately the initial enthusiasm for vigabatrin has been tempered by recent reports of visual field constriction due to the drug, which will likely prevent vigabatrin from becoming an approved treatment in the United States and limit its utility in other countries. Thus, other new anti-epileptic drugs may be investigated more thoroughly as potential first-line therapy for infantile spasms.

In addition to medication, there are some potential surgical options for infantile spasms, although they may only be applicable to a small percentage of patients. Although in most patients the precise source of the spasms in the brain cannot be localized, there is a small minority of patients who have secondarily generalized spasms from focal cortical lesions, which can be surgically removed. For patients without a surgically-resectable lesion, corpus callostomy has been reported to dramatically improve spasms, as well as other seizure types, such as drop attacks.


Many studies have examined the long-term prognosis of patients with infantile spasms. Although there is substantial variability in specifics reported from different studies, by all accounts the majority of patients with infantile spasms suffer a poor outcome with respect to chronic epilepsy, mental retardation, and other neurodevelopmental disabilities.

Epileptic spasms usually resolve with or without treatment in the majority of patients, generally by mid-childhood. However, other seizure types arise in 50-70% of patients. Similarly, on long-term follow-up, chronic intractable epilepsy is present in approximately 50% of patients with a history of infantile spasms. A close relationship between infantile spasms and Lennox-Gastaut syndrome has been consistently observed. Twenty to fifty percent of patients with infantile spasms evolve into Lennox-Gastaut syndrome and conversely, a similar percentage of patients with Lennox-Gastaut syndrome have a history of infantile spasms.

Mental retardation occurs in 70-90% of patients with infantile spasms, usually involving severe to profound retardation. Other neurological deficits, such as cerebral palsy, may be seen in about 30-50% of patients. By far, the most important factor in predicting neurological prognosis, including developmental outcome and long-term epilepsy, is etiology. In some studies, patients with cryptogenic infantile spasms have only a 30-50% chance of mental retardation compared to 80-95% for patients with a symptomatic etiology. Factors that have been associated with a good prognosis include normal neurological exam and development at onset, absence of other seizure types at onset, older age of onset, short duration of spasms, and early effective treatment of spasms (reported with ACTH).


Children with infantile spasms represent one of the greatest challenges in pediatric neurology. Accurate diagnosis depends on a thorough understanding of the clinical and electrographic features of spasms. Diagnostic evaluation should employ a rational approach for identifying potential etiologies of symptomatic infantile spasms. Decisions about treatment should take into account the poor prognosis of infantile spasms, conflicting data about the impact of treatment on long-term outcome, and the potentially serious side effects of treatment. As more is learned about the causes and pathophysiology of infantile spasms, hopefully more effective and safer treatments (perhaps including neuroprotective agents) will be developed. Given the profound neuro-developmental sequellae of IS, the ultimate goal should be the primary prevention of IS.

Selected References

  1. Chugani HT. Infantile spasms. Curr Opin Neurol 1995;8:139-144.
  2. Commission on Pediatric Epilepsy of the International League Against Epilepsy. Workshop on infantile spasms. Epilepsia 1992;33:195.
  3. Dulac O, Chugani HT, Dalla Bernardina B, eds. Infantile Spasms and West Syndrome. London: WB Saunders; 1994.
  4. Snead OC. Treatment of infantile spasms. Pediatr Neurol 1990;6:147-150.
  5. Wong M, Trevathan E. Infantile spasms. Pediatr Neurol 2001;24:89-98.