Neurodevelopmental status in infantile spasms and West syndrome: the challenge of assessment

Abstract

It has long been recognized that infantile spasms (epileptic spasms with onset before age 2y) are associated with neurodevelopmental impairment. However, the causal relationships between the clinical epileptic spasms, electroencephalogram (EEG) patterns (particularly hypsarrhythmia), and underling aetiologies—whether acquired, genetic, metabolic, structural, or combinations of these—have been matters for some debate. Many traditional descriptions of West syndrome have included developmental delay or regression as part of a definitional triad, alongside infantile spasms and hypsarrhythmia (an interictal EEG pattern characterized predominantly by asynchronous, high-amplitude, slow waves and multifocal epileptiform discharges). Developmental delay or regression at the time of presentation was not included in the definition of West syndrome proposed in theWest Delphi consensus statement. That proposal collated and iterated views from 42 respondents in 16 countries, and it suggested that West syndrome should be defined by the combination of two presenting features: epileptic spasms that occur in clusters and the presence of hypsarrhythmia in the EEG. The main reasons for suggesting these changes from the traditional triadic definition were, first, that it was uncommon for infantile spasms to occur singly and so such a feature would suggest a different seizure type; and second, that it is often difficult to assess neurodevelopmental status reliably at the time of presentation of infantile spasms. The study by Sumanasena et al. has made a significant contribution to the literature by systematically assessing developmental status, at the time of clinical presentation and before randomized treatment allocation as part of a clinical trial, in a cohort of 95 infants who had epileptic spasms and hypsarrhythmia. More than 90% of the infants were found to have developmental delay in several domains at the time of presentation; only one of the infants had scores within the average range for all the developmental domains. The authors discuss the fact that some earlier studies reported lower frequencies of developmental delay at the time of presentation. This difference is possibly due to these earlier studies including infants whose EEG was consistent with infantile spasms but not categorized as hypsarrhythmia. It is possibly due also to a less systematic and sensitive approach to neurodevelopmental and sensory assessments. With respect to EEGs, however, causal inferences are complicated by the observation that categorization as hypsarrhythmia has relatively poor interrater reliability. This is a factor that currently limits the usefulness of the interictal EEG alone as a biomarker. Clinicians charged with making treatment decisions continue to be exercised by the question of how much weight to put on factors such as the resolution of the epileptic spasms, changes in the EEG, and overall neurodevelopmental status. So, what happens in the longer term? A meta-analysis of studies reporting later neurodevelopmental outcomes found significant statistical heterogeneity, but, using a random effects model, it suggested a pooled estimate of 23.6% (95% confidence interval 19.3–28.6) for the proportion with good neurodevelopmental outcome. It also reported the consistent and important finding that longer lead-times to treatment from the onset of epileptic spasms are associated with poorer long-term neurodevelopmental outcomes. This is in the context of studies reporting that lead-times to diagnosis and effective treatment are often long. The ASSIST Study in California, for example, reported that the median time to evaluation by an effective provider was 3.5 weeks, and that fewer than one-third of infants saw an effective provider within one week of initial parental concerns. It seems imperative that, when concerns are raised about possible infantile spasms, we are able to respond with timely, systematic, and reliable assessments.