The Risk of Lower Mental Outcome in Infantile Spasms Increases after Three Weeks of Hypsarrhythmia Duration


Address correspondence and reprint requests to Dr. Zvonka Rener-Primec at Department of Child Neurology, University Children's Hospital, Vrazov trg 1, 1525 Ljubljana, Slovenia. E-mail:


Summary:  To assess the correlation between hypsarrhythmia duration and mental outcome in infantile spasms (IS) the medical records of 48 infants with IS were reviewed retrospectively and psychological assessments undertaken at follow-up at the age of 3 to 13 years. We found 18 (38%) cryptogenic IS cases with typical hypsarrhythmia and 30 symptomatic with modified hypsarrhythmia—further classified into 15 cases as multifocal, 10 as pseudoperiodic and 5 as unilateral hypsarrhythmia. A short treatment lag (one to two weeks) occurred in 25, three to four weeks in 10 cases. Spasms ceased within one month after treatment in 23 infants. At follow-up 15 children had normal mental outcome (borderline included). A correlation between hypsarrhythmia duration longer than three weeks and lower mental outcome was found using the logistic regression model. The duration of hypsarrhythmia represents a sensitive prognostic parameter in IS; the risk of mental retardation increases after three weeks of hypsarrhythmia.

Infantile spasms (IS), i.e., West syndrome, as an epileptic encephalopathy, represents one of the major causes of acquired mental retardation in early childhood (Roger and Dulac, 1994). Prognosis largely depends on etiology, being most unfavorable in prenatal cases with evolution to resistant epilepsy and severe mental retardation. However, infants with cryptogenic/idiopathic etiology can have favorable outcome with lifelong remission of seizures and normal or nearly normal mental development (Lombroso, 1983; Dulac et al., 1993; Riikonen, 1996).

Among prognostic factors, the treatment lag and duration of spasms have been recognized as being significant (Lombroso, 1983; Eisermann et al., 2003; Goh et al., 2005). In most studies a treatment lag of up to one month correlates with a better prognosis. Prolonged duration of spasms despite treatment increases the risk of mental retardation (Eisermann et al., 2003; Goh et al., 2005). In some infants hypsarrhythmia continues despite the cessation of spasms. Therefore we explored the impact of hypsarrhythmia duration on mental outcome in 48 children with IS; the epidemiologic part of the study was published in 2002 (Rener Primec et al., 2002).


The medical records of 48 infants diagnosed with IS were studied retrospectively. IS were defined as a syndromic condition (Roger and Dulac, 1994; Lux and Osborne, 2004). The methodology applied has been described previously (Rener Primec et al., 2002), when epidemiologic features of 47 children (permanent residency in Slovenia) were presented. One additional child (immigrant) was included here. The essentials for this study were standard 12-channel EEG recordings taken in wakefulness and sleep at the time of diagnosis of IS and subsequent EEG recordings obtained every 7 to 10 days during the first 2 months or until remission (whichever came first), then monthly for the first 6 months after diagnosis of IS. If the spasms reoccurred an infant was admitted for investigations on the clinical-need basis. All EEG recordings were interpreted by pediatric electroencephalographer; the EEG reports were reviewed and hypsarrhythmia classified as typical or atypical (Lux and Osborne, 2004). Video-EEG was not available; therefore, when clinical spasms occurred during EEG recording clinical data were included in the technician's report (form and duration of consecutive spasms, asymmetries, etc.). Written detailed information based on instructed clinical observation of the infant's behavior during all daily activities was obtained from nurses and/or parents as a part of routine diagnostic workup of each patient. This approach enabled better recognition of spasms or partial seizures, followed by spasms and assured that infants with relapses were promptly referred to the hospital.

Data on the age at onset of spasms, perinatal and family history were collected from medical records. The psychomotor development prior to the IS was obtained from his/her pediatrician (children in Slovenia are entitled to free community pediatric care [Rener Primec et al., 2002]). Initial treatments were extracted from medical records. The choice of initial treatment changed from steroids to vigabatrin in 1992; however, in symptomatic cases parents often refused steroids due to possible side effects in their already handicapped child.

The duration of hypsarrhythmia (in weeks) was estimated from the infant's age at the onset of spasms, the first EEG and from consecutive EEGs. If an infant had partial seizures before the IS, then the onset of spasms was considered as the time of hypsarrhythmia onset. The remission of IS was assessed when an infant was spasm-free and hypsarrhythmia resolved over a one-month period.

Neuroradiological reports with original films were re-reviewed by a pediatric neuroradiologist (CT in all children, after 1990 brain MRI was additionally performed in 23 children, in whom etiology was unknown). All infants were classified into symptomatic or cryptogenic cases. At follow-up, the idiopathic subgroup was identified, according to published criteria (Vigevano et al., 1993; Lux and Osborne, 2004).

During the follow-up all children underwent psychological examination. The minimum follow-up time was 3 years after the last onset of IS. A psychomotor assessment was undertaken with respect to the child's age and mental level by the same clinical psychologist using standardized psychometric tests—the Brunet Lezine developmental scale, Vineland maturity scale, and Wechsler intelligence scale for children (WISC). The results were expressed as the developmental or intelligence quotient (DQ, IQ). Only three children (severely mentally retarded with tetraplegia) were not tested—information on these three was obtained from their pediatricians. An outcome was considered good in cases with normal mental development (including borderline) and poor in cases with mental retardation (IQ or DQ ≤ 70). The study was approved by the Medical Ethics Committee of Slovenia.

Data analysis

Data were analyzed using standard statistical methods and the logistic regression model with splines was used to analyze the relationship between duration of hypsarrhythmia and mental outcome (Harrell, 2001).


Patients' characteristics

The 48 infants (23 male) fulfilled the diagnostic criteria of IS. The age at onset of spasms ranged from two to 10.5 months (mean 6.1, SD 2.17); with adjusted age (in preterm infants) 1.7 to 10 months (mean 5.7, SD 2.05). Symptomatic etiology was found in 30 infants (62.5%): seven tuberous sclerosis, three vascular insult, three cerebral malformation, two Down syndrome, and 14 children had perinatal hypoxic-ischemic encephalopathy. One child suffered asphyxia during cardiac surgery.

EEG data

All 30 symptomatic cases had modified (atypical) hypsarrhythmia, further classified into hypsarrhythmia with multifocal spikes in 15 (31%), pseudoperiodic patterns in 10 (21%) and unilateral hypsarrhythmia in 5 cases (10%). All 18 cryptogenic IS cases (38%) had typical hypsarrhythmia.

Treatment lag, duration of spasms, and hypsarrhythmia

A short treatment lag of one week was found in 11 infants (23%), two weeks in 14 (29%) and three to four weeks in 10 infants (20%), i.e., treatment started within one month after the onset of IS in 35 infants (73%). A longer treatment lag of five to nine weeks occurred in eight children; even longer, three months or more in five children (spasms had not been recognized as an epileptic condition).

For the initial treatment (after pyridoxine iv.) hydrocortisone or a synthetic adrenocorticotropic hormone were used in 20, vigabatrin in 7, and conventional AED in 21 (44%) infants.

The duration of spasms and hypsarrhythmia are presented in Table 1.

Table 1. Duration of spasms and hypsarrhythmia in 48 infants with IS
Time intervalDuration of spasms N (%) of infantsHypsarrhythmia duration* N (%) of infants
  1. *Hypsarrhythmia duration in two children (4%) was impossible to assess due to incomplete data.

1 week4 (8)3 (6)
2 weeks 8 (17) 5 (10)
3 weeks 5 (10) 6 (12)
4 weeks 6 (12)4 (8)
5–8 weeks12 (25)17 (36)
>8 weeks13 (27)11 (23)

Remission of spasms was achieved in 27 (56%) infants, but spasms reoccurred in 7 of these before the first year of life. In the second year of life seizures remitted in 9, while even later in another 3 cases. The remaining 16 out of 48 children had refractory epilepsy; four died. At follow-up 28 children were seizure-free; 18 of whom were medication-free.

Mental outcome

At follow-up, normal mental outcome (including borderline) was found in 15 out of 48 children (Table 2). Infants with typical hypsarrhythmia had much better outcome: 11 out of 18 children with typical and only 4 out of 30 with atypical hypsarrhythmia had normal mental development.

Table 2. Etiology and mental outcome at follow-up in 48 children after IS
Etiology of ISMental outcomeN (%)
  1. χ2= 11.953, p < 0.0005 (Fisher's exact test p < 0.001).

  2. *Four children died during the study period: three were severely and one moderately retarded.

Symptomatic (%) 4 (13)26 (87) 30 (100)
Cryptogenic and idiopathic (%)11 (61) 7 (39) 18 (100)
All* (%)15 (31)33 (69)48* (100)

The level of mental retardation in 33 (69%) children was: mild in 6 (IQ 51 > 70), moderate in 9 and severe/profound in 14 cases (29%, IQ ≤ 35). Among four children who died, three were severely and one moderately retarded.

In the further statistical analysis mental outcome was considered a binary variable (0 = normal, 1 = not normal) and its correlation with the hypsarrhythmia-duration (in weeks) was analyzed using the logistic regression model. Since the relationship between the log odds for nonnormal outcome and the hypsarrhythmia-duration was clearly nonlinear, restricted cubic splines were used to assess the correlation (Harrell, 2001).

The analysis showed (see Fig. 1) that the curve of log odds for nonnormal outcome begins to increase steeply after three weeks of hypsarrhythmia and their values level off at about 6 weeks. This means that hypsarrhythmia of longer then three weeks' duration correlates with a higher risk of mental retardation. Inclusion of other variables in this model was not possible given the sample size, as the method used already required estimation of five coefficients.

Figure 1.

Dependence of the log odds for nonnormal outcome on hypsarrhythmia duration. The dotted lines represent the 95% confidence intervals. The p-value for the model is 0.004.


The aim of the study was to assess the correlation between the duration of hypsarrhythmia and the mental outcome in children with IS. It has been recognized for decades that prolonged duration of spasms is a negative prognostic factor with evolution to resistant epilepsy and mental retardation, often reflecting the severe underlying brain pathology. In most studies the treatment lag up to one month is accepted as the limit with regard to a favorable mental outcome and the significantly better outcome in the early treatment groups was found for both cryptogenic and symptomatic IS (Lombroso, 1983; Koo et al., 1993; Kivity et al., 2004). It is of interest that in the early treated cryptogenic cases the remission of spasms occurred within the three days in the majority of infants (Kivity et al., 2004), while the opposite is reported for the late treated, where also the proportion of the lower mental outcome is significantly higher. Whether this reflects the consequence of prolonged hypsarrhythmia or the underlying disorder itself (biochemical or neurotransmitter) remains an open question. It is of note that in the same study (Kivity et al., 2004) some of the late treated infants (and none from early treated) later developed Lennox–Gastaut syndrome—another form of age related epileptic encephalopathy. This supports a hypothesis that early steroid treatment may prevent such evolution (Dulac, 2001).

In our study a short treatment lag, i.e., initiation of therapy within one month of the spasms onset was found in 73% of infants, in half of them (25 cases) even within 2 weeks, which is a higher percentage then usually reported (Lombroso, 1983; Koo et al., 1993; Kivity et al., 2004). It was therefore possible to analyze the correlation between the shorter intervals of hypsarrhythmia duration (in weeks) and the mental outcome. Using the logistic regression model we found that after three weeks of hypsarrhythmia the risk of mental retardation increases. The short interval of three weeks can be explained by higher responder rate of early treated cases, while those with longer duration of hypsarrhythmia belong either to nonresponders (severe brain pathology) or to the long treatment lag. Due to small number of cases with specific etiology in our study separate analysis was not feasible.

In the symptomatic IS cases the underlying pathology may itself cause mental retardation, so it is difficult to assess a deleterious effect of hypsarrhythmia alone on the brain development. However, two studies recently addressed the mental outcome in children with specific etiology of IS—tuberous sclerosis complex and Down syndrome—and they showed, that the risk of mental retardation increases significantly with a prolonged duration of spasms and with prolonged time from treatment onset until remission of spasms (Eisermann et al., 2003; Goh et al., 2005).

The resolution of hypsarrhythmia is considered the essential marker of primary electroclinical response, together with cessation of spasms (Lux and Osborne, 2004), while its persistence has been shown to correlate with cognitive deterioration (Gaily et al., 1999; Dulac, 2001). Prolonged hypsarrhythmia may either reflect an underlying pathology (drug unresponsive cases), or may result from a prolonged duration of IS before the initial treatment (the long treatment lag). In our study, 27% of cases received medical attention with a delay of more than one month, mainly because spasms were not recognized as an epileptic condition. Among these, some children with cryptogenic IS were moderately mentally retarded, despite the fact, that remission of spasms and hypsarrhythmia was achieved.

The importance of early treatment has been stressed by numerous studies (Lombroso, 1983; Koo et al., 1993; Eisermann et al., 2003; Goh et al., 2005). Hypsarrhythmia reflects a severe underlying disorder, which, although poorly understood due to its intrinsic complexity, has a negative influence on the developing brain. Prolonged duration of such continuous epileptic activity may affect the cortical areas in the process of rapid brain development and contributes to mental retardation (Dulac, 2001; Rho, 2004). According to our study more than three weeks of hypsarrhythmia may already represent a risk of mental retardation.

In conclusion, the duration of hypsarrhythmia should be considered a sensitive prognostic parameter in IS with regard to mental development.


Acknowledgments:  We are grateful to our clinical psychologist Vali Glavic Tretnjak for the assessment of children.

This work was supported by the Ministry of Science and Technology of the Republic of Slovenia, Grant No. J3-6163-312/2001.