Infantile spasms: diagnosis and assessment of treatment response by video-EEG

Eija Gaily, PO BOX 280, Hospital forChildren and Adolescents, FIN-00029 HYKS, Finland.E-mail eija.gaily@ hus.fiThe purpose of this study was to investigate early electroclinicalmanifestations and evaluate treatment responses by video-EEGin infants with newly diagnosed spasms. Spasms were recordedin 44 infants (27 males, 17 females) before adequate treatment.Mean ages at onset of spasms and at first video-EEG were 5.3months (range 0.9 to 9 months) and 5.9 months (range 2.4 to11.5 months) respectively. Thirteen infants had cryptogenic and31 had symptomatic aetiology. First treatment was vigabatrinin 36 infants. All infants were followed until 12 months of ageor death. Treatment response in the first months of therapywas assessed by repeated video-EEG studies in 23 infants. Onthe first video-EEG, 34 infants had typical symmetric motorspasms, three infants showed asymmetric or asynchronousbehaviour, and seven infants had only subtle spasms.Interictal EEG showed hypsarrhythmia in 27 infants andmultifocal spikes with normal or nearly normal background in17 infants. Subtle spasms, asymmetric or asynchronousspasms, and asymmetric ictal or interictal EEG abnormalitieswere associated with symptomatic aetiology and poor cognitiveand seizure outcome at 12 months. Serial video-EEGrecordings showed a transition from motor to subtle spasmsduring the first 2 weeks of vigabatrin therapy in four infantsand only subtle spasms in two therapy-resistant infants at 12months. Cessation of spasms usually preceded disappearance ofhypsarrhythmia or multifocal spikes, but persistence ofmultifocal spikes over several weeks was always associatedwith existing spasms. Transition of hypsarrhythmia intomultifocal spikes was observed during vigabatrin therapy evenin infants with intractable spasms. Initial diagnosis of infantilespasms requires video-EEG studies especially in infants withsymptomatic aetiology who may show only subtle spasms.Video-EEG is the only reliable method for assessing treatmentresponse as spasms and interictal EEG abnormalities aremodified by treatment and may become subtle.The most common ictal behaviour in infantile spasms is sym-metric flexion, extension, or a mixture of both (Kellaway etal. 1979, Fusco and Vigevano 1993). Asymmetric as well assubtle spasms have also been described (Kellaway et al. 1979,Donat et al. 1991, Fusco and Vigevano 1993, Gaily et al. 1995,Haga et al. 1995). Dulac and coworkers (1993) and Fuscoand Vigevano (1993) separated ‘independent spasms’, wherehypsarrhythmia reappears between spasms, and ‘clusteringspasms’ where no hypsarrhythmia is seen between spasms.The ictal EEG phenomena associated with infantile spasmsinclude fast (or ‘spindle-like’) activity, high voltage spikes,spike-waves, slow waves, and attenuation (Jeavons andBower 1964, Kellaway et al. 1979, King et al. 1985, Fusco andVigevano 1993, Haga et al. 1995).Two previous studies (Fusco and Vigevano 1993, Haga etal. 1995) have investigated ictal behavioural and EEG phe-nomena by video-EEG in infants with newly diagnosed infan-tile spasms and hypsarrhythmia before effective treatment.According to their results, the motor pattern of symmetricalspasms does not appear to have any prognostic significance,but asymmetrical spasms and partial seizures are stronglyassociated with symptomatic aetiology and poor prognosis.The possible prognostic implications of subtle spasms havenot been studied before. Fusco and Vigevano (1993) foundthat independent spasms predict a better outcome thanspasm clusters; this was disputed by Haga and colleagues(1995). No data are available on how therapy modifies semi-ology of behavioural spasms in the first months after treat-ment onset.Hypsarrhythmia is observed in the initial EEG of 66 to 99%of children with infantile spasms in at least one state of alert-ness (most commonly stage 2–3 sleep; Jeavons and Bower1964, Watanabe et al. 1993). Hypsarrhythmia may be preced-ed and followed by focal and multifocal spikes (Watanabe etal. 1973, Kotagal 1995). The prognostic significance of non-hypsarrhythmic EEG on presentation has only been studiedby Jeavons and Bower (1964) who found it a favourable sign.Different symmetrical variants of hypsarrhythmia as describedby Hrachovy and coworkers (1984) do not appear to correlatewith prognosis, but more ‘severe’ hypsarrhythmia, as deter-mined by a detailed scoring system, has some associationwith poor outcome according to Kramer and colleagues(1997). Asymmetrical hypsarrhythmia is associated with symp-tomatic aetiology (Drury et al. 1995) and poor outcome(Dulac et al. 1993).Before the era of adrenocorticotropic hormone (ACTH) orvigabatrin therapy, hypsarrhythmia persisted until 12 monthsof age in all and until 24 months in one-third of those infantswho had hypsarrhythmia at onset of spasms (Jeavons andBower 1961). Lombroso (1983) reported no change in interic-tal EEG 10 months after onset of ACTH therapy in 27 of 128(21%) children, of whom at least 67% had initial hypsarrhyth-mia. Few data are available on how interictal EEG is modifiedin the first months after therapy onset. Jeavons and Bower(1964) reported cessation of spasms by clinical observation in70% and normalization of EEG in 41% of 66 children whoreceived ACTH or steroid therapy, but observed no distincttemporal correlation between spasm and EEG response.Dulac and colleagues (1993) defined normalization of interic-tal EEG within 15 days after treatment onset as one criterion ofa favourable outcome in cryptogenic infantile spasms. The aims of the present study were: (1) to describe the initial