Article Text

Download PDFPDF

The management of infantile spasms
  1. Isabelle Desguerre1,
  2. Rima Nabbout1,2,3,
  3. Olivier Dulac1,2,3
  1. 1
    APHP, Paediatric Neurology, Hôpital Necker, Paris, France
  2. 2
    INSERM, U663, Paris, France
  3. 3
    Université Paris Descartes, Paris, France
  1. Olivier Dulac, APHP, Paediatric Neurology, Hôpital Necker, 149 Rue de Sèvres, 75015 Paris, France; o.dulac{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Infantile spasms (IS) is the most frequent epilepsy syndrome in infancy. Although the consensus triad of spasms in clusters with hypsarrhythmia adversely affecting psychomotor development1 seems quite straightforward, the great variability of each component is a challenge not only to the clinician involved in management of the condition but also to the methodologist designing trials for treatment issues.

Indeed, the spasms may be symmetrical, asymmetrical or mixed, and florid and easy to identify or reduced to eye movements with crying that require video polygraphy including muscular recording for identification as ictal events.2 For weeks the diagnosis may be misdiagnosed as gastrointestinal disturbance until an EEG is performed. Ictal EEG may consist of a diffuse high amplitude slow wave, electrodecrement or low amplitude fast activity, or even be combined with focal discharge. The distinction between spasms and myoclonic or brief tonic seizures may be difficult, particularly in very young infants. The interictal tracing, instead of showing a continuous and asynchronous high amplitude delta theta slow wave with spike activity indicating hypsarrhythmia, may demonstrate asymmetry, periodic flattening reminiscent of the suppression burst pattern, high amplitude fast or slow activity with few spikes, or even single or multifocal spike focus.3 Psychomotor delay may have preceded the onset of spasms or appear to be absent even after several weeks of seizures, and may range from speech delay to autistic behaviour.

Such diversity corresponds to the long list of possible causes that range from focal or multifocal, pre-, peri- or postnatal brain damage, to apparent lack of any structural lesion. Nevertheless, some correspondence has been established, for instance the high amplitude fast activity associated with lissencephaly4 or the “split-brain” pattern of Aicardi syndrome. However, although the clinical and EEG pattern is determined in part by the aetiology, the correspondence is far from being a one to one relationship because the extent and location of lesions vary (ie, post-ischaemic lesions or tuberous sclerosis).

Although a small proportion of patients show no radiological evidence of brain lesions and have an excellent long term outcome, suggesting that IS may in some cases be a purely functional and transient event considered idiopathic,5 an apparent lack of structural lesions is also the rule in genetically determined encephalopathies, such as symptomatic Down syndrome, 1p36 deletion, or ARX or STK9 mutations.

Long term outcomes range from total recovery to severe encephalopathy, with most patients experiencing an intermediate condition. Total recovery can occur despite the presence of brain lesions provided they are not destructive. A very small number of patients are able to attend normal schools, including some who have idiopathic IS and some who have either focal dysplasia or tuberous sclerosis. Distinguishing the mechanisms of persisting encephalopathy is also challenging since the underlying pathology and the epileptic encephalopathy (ie, seizures and the interictal epileptic activity) are likely to depend on aetiology. ARX mutations determine autistic behaviour, even in the absence of overt epilepsy.6 On the other hand, there is growing evidence that the duration of epileptic encephalopathy contributes to the development of autism,7 particularly in tuberous sclerosis8 and Down syndrome.9 The persistence of intractable epilepsy also seems to depend on both the aetiology and the duration of the epileptic encephalopathy. It is rarely possible to control the epilepsy of lissencephalic patients with previous IS.10 However, early therapeutic control of IS seems to have improved the long term outcome of epilepsy in both tuberous sclerosis and Down syndrome.9

There seem to be no restrictions to early treatment of IS. The old advice of avoiding hormonal therapy in symptomatic cases has long been abandoned. The challenging issues are what medication to use and how long it should be used for. It is agreed that two treatments have shown efficacy: hormonal therapy (ie, steroids and ACTH) and vigabatrin. The real question is which treatment is first line or should they be combined from the onset. The efficacy of the combination is currently being tested in a large randomised prospective trial to which the reader is invited to contribute ( Both hormonal11 and vigabatrin12 treatments have been shown by double blind prospective studies to be effective. It is noteworthy that only one high quality study has been performed on each treatment. On the other hand, two other studies compared both treatments. The first, focusing on a specific aetiology (tuberous sclerosis) was not blinded because of the recognisable side effects of steroid treatment, and showed vigabatrin to be clearly superior.13 The second, the UKISS study, was blinded and included all causes except tuberous sclerosis, given the wide consensus in favour of vigabatrin to treat the latter.14 15 One major finding was that although at 2 weeks hormonal therapy was significantly superior, this benefit was lost at 14 months’ follow-up because of a higher relapse rate. Hormonal therapy was claimed to benefit cognition,15 but this was not a robust finding.2

As regards the side effects of both treatments, the data are also weak. Metabolic deficiency is not a contra-indication for hormonal therapy provided biochemistry is monitored.16 The major concern with hormonal therapy is fatal infections, particularly in a population that is at high risk due to age and the underlying condition. The majority of authors, including those most in favour of hormonal therapy as a first line treatment, have reported death due to immune suppression.17 This was the main reason why hormonal therapy was not recommended for symptomatic cases in the 1970s.18 However, the true incidence of death is difficult to determine because the underlying disease itself may cause death, including sudden and unexplained death (personal unpublished observation). Perhaps this is the reason the UKISS study did not emphasise this issue although five patients died, including at least one from infection during steroid treatment.15

Retinal toxicity with constriction of the visual field is the major concern with vigabatrin treatment, which has resulted in the delayed registration of the compound in many countries outside Europe, including the USA and Japan. Investigating visual fields remains difficult in children before the age of 8 years. In addition, psychomotor delay in children with previous IS hinders visual field assessment. A single retrospective study showed that cumulative dose is a major risk factor, and is linked to treatment duration.19 This side effect should therefore not influence the choice of first line treatment because the risk only becomes significant after treatment has lasted for more than 6 months. An attempt to reduce the duration of treatment to 6 months was successful in Down syndrome IS,20 but patients with focal dysplasia or tuberous sclerosis, and thus epileptogenic cortical malformations, seem to require vigabatrin treatment for several years.21

Patients with tuberous sclerosis clearly benefit from first line treatment with vigabatrin. For other patients, in the absence of the clear superiority of one treatment over the other, choice is based a comparison of major side effects, with a low but serious risk during the first weeks of hormonal therapy compared to a delayed and non-serious risk for patients who require prolonged treatment with vigabatrin. For this reason, the authors prefer to start treatment with a 1–2-week vigabatrin trial whatever the aetiology. If treatment with vigabatrin fails, hormonal therapy is clearly indicated. Indeed, in the UKISS study three-quarters of patients responded to the alternative drug after failure of the first compound.15 At this point, the unresolved question is whether hormones should be administered as monotherapy or if the combination with vigabatrin is likely to increase the success rate. However, the ICISS study will probably give some guidance since it is examining whether the combination is superior to hormonal therapy alone.

Surgery is an option for patients who do not respond to either medication or whose epilepsy changes to focal seizures following drug treatment; surgery is not a first line treatment because focal lesions, including dysplasia, may determine transient drug-responsive IS. Although asymmetry of spasms or interictal EEG may indicate a cortical brain lesion, MRI in combination with a PET scan will determine whether surgery is likely to be beneficial.22 There is no point delaying surgery once both vigabatrin and hormonal therapy have failed.

For other pharmacoresistant patients, third generation compounds may be considered since data suggest they might have some effect, except for oxcarbazepine which, like carbamazepine, may worsen the condition.23 Based on personal experience, the authors have a preference for topiramate. Pyridoxine and pyridoxal phosphate are weak alternatives. The efficacy of a ketogenic diet also needs to be determined.24 Total callosotomy may help older patients with severe drop attacks.25

The difficulties accompanying therapeutic decisions require that parents are kept fully informed regarding mechanisms of seizures and behavioural problems. They need to be aware of the toxicity of both main treatments and understand that major reductions in spasm frequency do not obviate the need for more aggressive therapy. They need to know that although IS has a very poor reputation, particularly on the internet chat pages that many parents visit within the first days of a diagnosis, a sizeable proportion of patients will recover, that treatment does help many of the remaining children, and that the prognosis is not “carved in stone” based on aetiology.



  • Competing interests: None.