Papers

Clinical course of untreated tonic-clonic seizures in childhood: prospective, hospital based study

BMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7078.401 (Published 08 February 1997) Cite this as: BMJ 1997;314:401
  1. C A van Donselaar, neurologista,
  2. O F Brouwer, child neurologistb,
  3. A T Geerts, scientific staffc,
  4. W F M Arts, child neurologistd,
  5. H Stroink, child neurologista,
  6. A C B Peters, child neurologistb
  1. a Department of Neurology, University Hospital Rotterdam Dijkzigt and Sophia, 3015 GJ Rotterdam, Netherlands
  2. b Department of Neurology, University Hospital Leiden, 2333 AA Leiden, Netherlands
  3. c Department of Public Health, Erasmus University, 3000 DR Rotterdam, Netherlands
  4. d Juliana Children's Hospital and Westeinde Hospital, 2566 ER The Hague, Netherlands
  1. Correspondence to: Dr C A van Donselaar St Clara Hospital, Olympiaweg 350, 3078 HT Rotterdam, Netherlands
  • Accepted 29 November 1996

Abstract

Objective: To assess decleration and acceleration in the disease process in the initial phase of epsilepsy in children with new onset tonic-clonic seizures.

Study design: Hospital based follow up study.

Setting: Two university hospitals, a general hospital, and a children's hospital in the Netherlands.

Patients: 204 children aged 1 month to 16 years with idiopathic or remote symptomatic, newly diagnosed, tonic-clonic seizures, of whom 123 were enrolled at time of their first ever seizure; all children were followed until the start of drug treatment (78 children), the occurrence of the fourth untreated seizure (41 children), or the end of the follow up period of two years (85 untreated children).

Main outcome measures: Analyis of disease pattern from first ever seizure. The pattern was categorised as decelerating if the child became free of seizures despite treatment being withheld. In cases with four seizures, the pattern was categorised as decelerating if successive intervals increased or as accelerating if intervals decreased. Patterns in the remaining children were classified as uncertain.

Results: A decelerating pattern was found in 83 of 85 children who became free of seizures without treatment. Three of the 41 children with four or more untreated seizures showed a decelerating pattern and eight an accelerating pattern. In 110 children the disease process could not be classified, mostly because drug treatment was started after the first, second, or third seizure. The proportion of children with a decelerating pattern (42%, 95% confidence interval 35% to 49%) may be a minimum estimate because of the large number of patients with an uncertain disease pattern.

Conclusions: Though untreated epilepsy is commonly considered to be a progressive disorder with decreasing intervals between seizures, a large proportion of children with newly diagnosed, unprovoked tonic-clonic seizures have a decelerating disease process. The fear that tonic-clonic seizures commonly evolve into a progressive disease should not be used as an argument in favour of early drug treatment in children with epilepsy.

Key messages

  • Untreated epilepsy is commonly thought to be a progressive disease

  • Early treatment, preferably after the first seizure, has been advocated

  • In at least 42% of children with newly diagnosed tonic-clonic seizures, the disease has a decelerating pattern, with successively longer intervals between seizures

  • The fear that tonic-clonic seizures commonly evolve into a progressive disorder should not be used as an argument in favour of early treatment of these children

Introduction

Untreated epilepsy may be thought of as a progressive disorder. Such a view has emerged from the observation by Elwes et al that untreated epilepsy shows an accelerating pattern.1 They found that intervals between successive untreated seizures decreased in many patients. In addition, a larger number of seizures before starting treatment seemed to be associated with a poorer control of seizures.2 If these observations are correct, an aggressive approach may be appropriate, as has been advocated by Reynolds et al, who advised starting treatment as soon as possible, preferably after the first seizure, to prevent the development of intractable epilepsy.3 4 5 Early treatment for all patients with newly diagnosed seizures has considerable consequences, however, as all antiepileptic drugs have side effects. A randomised trial comparing early and delayed treatment, the multicentre study of epilepsy and single seizures, is being conducted by D Chadwick.

Intervals between seizures may vary. Patients with accelerating patterns–shortening intervals between seizures–may be more likely to be referred and treated. The retrospective study by Elwes et al 1 was confined to patients for whom the decision to start treatment had already been made. Patients with accelerating seizure patterns may have been overrepresented, and it is doubtful that these findings could be applied to newly diagnosed patients.

The uncertainty about the clinical course of epilepsy and the possible clinical consequences of early treatment of all patients prompted us to study the pattern of seizures in a cohort of children with newly diagnosed, unprovoked tonic-clonic seizures.

Methods

This study is part of the prospective, multicentre Dutch study of epilepsy in childhood, which started in 1988. We enrolled all children aged 1 month to 16 years with newly diagnosed single seizures or epilepsy who had been referred to two university hospitals, a general hospital, and a children's hospital in the Netherlands up to August 1992. The main aims were to study the prognosis of single seizures, the prognosis of newly diagnosed epilepsy, and the consequences of early withdrawal of antiepileptic drugs in children who responded well to treatment. The study was approved by the medical ethical committees of the participating hospitals. Most children were referred directly by their general practitioners; some were seen in casualty departments, and some were referred by paediatricians.

A committee of three child neurologists judged whether the description of the ictal events fulfilled the predefined descriptive diagnostic criteria adapted from van Donselaar et al.6 Children with neonatal, febrile or other acute symptomatic seizures were excluded, as were children who were referred from other hospitals. The panel assessed the diagnosis of epilepsy, classified the seizures according to the revised classification of the International League Against Epilepsy,7 and categorised the epilepsies according to the criteria proposed by Gastaut.8 Etiology was classified as remote symptomatic or idiopathic.9 Children with single seizures were not treated with antiepileptic drugs, but the attending child neurologists were free to start or to delay treatment in children presenting with a status epilepticus or in children who had suffered two seizures or more.

Of the 563 children who were enrolled between 1 August 1988 and 1 August 1992, 261 children had seizures of only the tonic-clonic type, including seizures with partial onset. We excluded 57 of these patients because no accurate data on numbers and dates of all seizures could be obtained. The remaining 204 children formed our study group; table 1 presents their main characteristics. Children enrolled in this study were seen at an early stage of their disease. The mean interval between the first ever seizure and time of recruitment was 4.5 months for all 563 patients, 1.4 months for the study population, and 9.2 months for the 57 children excluded because accurate data were lacking. All children were followed until the occurrence of the fourth untreated seizure, the start of drug treatment, or the end of the follow up period of two years, whichever came first.

Table 1

Main characteristics of 563 children with newly diagnosed seizures and those with tonic-clonic seizures. Figures are number (percentage) of patients unless stated otherwise

View this table:

Multiple seizures within 24 hours were considered as separate events and a status epilepticus as a single event. Disease pattern was categorised as uncertain if the child was given drug treatment after the first, second, or third seizure. We classified the pattern as decelerating if the child became free of seizures without drugs. In cases with four seizures or more the pattern was classified as decelerating if the time intervals from the first ever seizure successively increased and as accelerating if the time intervals successively decreased. The patterns in all remaining children were classified as uncertain.

Results

We enrolled 123 children after their first seizure, 41 after two seizures, 19 after three seizures, and 21 after four or more seizures. Drug treatment was started in 10, 19, 9, and 20 patients, respectively, at time of intake. Table 2 presents the number of seizures before intake and during follow up.

Table 2

Seizures and recurrences in a cohort of 204 children with epilepsy

View this table:

Table 3 shows seizure patterns as from the first ever seizure. We classified the disease pattern as decelerating in 83 patients because they became free of seizures without medication: 60 had only one seizure, 15 had only two, and eight had only three. In two additional untreated children, whose last seizure occurred too close to the end of the follow up period to assess the disease pattern, we classified the disease pattern as uncertain. The pattern could not be assessed in an additional 78 children who were treated after their first seizure (10 children presenting with a status epilepticus) or their second (46 children) or third seizure (22 children). Of the 22 children treated after their third seizure, the intervals had decreased in 14 and increased in eight. Of the 41 patients with four or more untreated seizures, three showed a decelerating pattern, 30 an erratic pattern (classified as uncertain), and eight an accelerating pattern. In summary, the pattern was decelerating in 86 children (42%, 95% confidence interval 35-49%), uncertain in 110 (54%), and accelerating in eight (4%).

Table 3

Timing of seizure patterns in 204 children with epilepsy

View this table:

Discussion

The possibility that untreated epilepsy is a progressive disorder is of major importance in balancing the pros and cons of early treatment with antiepileptic drugs. Our study shows that in a large proportion of children with newly diagnosed, untreated tonic-clonic seizures the intervals between successive seizures become longer. This proportion may be a low estimate because we could not classify the pattern in the substantial number of patients who were given drug treatment after the first or subsequent seizures.

Ideally, the clinical course of epilepsy would be studied in an unselected, untreated group of children enrolled at the onset of the disease and followed for a long time without treatment, but such a study design is ethically not acceptable at this time. Our study reflects clinical practice. We enrolled the patients at an early stage of their disease and were rather reluctant to start treatment with antiepileptic drugs. On the basis of the catchment area of the participating hospitals and the incidence figures from epidemiological studies we estimate that we have enrolled about 75% of the children with new onset epilepsy.

Study design

The design of our study has some limitations. We confined ourselves to children with tonic-clonic seizures (including seizures with partial onset). We expected to get accurate data on the number and dates of these events only in this group of children. We excluded 57 of the 261 eligible patients because accurate data on some seizures were lacking; the mean interval between the first ever seizure and enrolment was much longer in these children. If none of these 57 patients had a decelerating pattern (which is highly unlikely), the proportion of children with a decelerating pattern would decrease to 33%. We thus think that excluding these children did not have a profound impact on our main conclusion.

We included patients enrolled after a single seizure, because some doctors recommend treating these patients as soon as possible.5 Moreover, these children can be followed prospectively from the onset of their “epilepsy.” Had we confined our study to patients with at least two seizures (excluding the 60 patients with a first seizure who did not have a recurrence without treatment and the 10 patients who were treated after their first seizure), the proportion of patients with a decelerating pattern would be much lower (19%) but would still be greater than the proportion with an accelerating pattern.

We defined the occurrence of the fourth untreated tonic-clonic seizure as an end point. By that time most children will be receiving treatment with antiepileptic drugs. Moreover, the timing of seizures in the early phase will be one of the deciding factors in starting or delaying treatment. Determining the proportion of children with an erratic seizure pattern would require delaying the start of treatment much longer so that a larger number of successive time intervals could be evaluated in each patient–an approach that was ethically not acceptable at the time of this study. We believe that our data give a fair approximation of the lower limit of the proportion of children with a decelerating pattern.

Study results

Our results contradict the results of Elwes et al,1 who found an accelerating seizure pattern in “many” (59%) patients with untreated tonic-clonic seizures. The group of patients that we studied was younger than Elwes et al's cohort. Also, that retrospective study was confined to patients in whom the decision to start treatment had already been made. The researchers classified the temporal pattern as accelerating if successive intervals decreased even if the patient had had only three seizures. With similar selection and analysis, 22 of 57 children (39%) in our study had an accelerating pattern. This clearly shows that studies using only patients already allocated to treatment are biased towards exclusion of patients with decelerating seizure patterns.

Length of intervals between seizures may fluctuate. The policy to treat all patients immediately may obscure the fact that the interval between subsequent seizures may be longer, even without drug treatment. Had we treated all our patients at the time of recruitment, we would have missed the favourable outcome of a substantial proportion of children who became free of seizures without drugs. We might even have been convinced that lack of seizures was due to the drugs instead of to the course of their epilepsy. Elwes et al probably missed the favourable prognosis in a substantial proportion of patients by treating them all at the time of enrollment.

We believe our findings are a fair approximation of the clinical course of idiopathic or remote symptomatic, newly diagnosed, tonic-clonic seizures in childhood. The fear that early untreated epilepsy commonly evolves into a progressive disease should not be used as an argument in favour of early treatment of these children.

Acknowledgments

This study was presented in part at the 21st International Epilepsy Congress, Sydney, Australia, 3-8 September 1995.

Funding: Dutch National Epilepsy Fund (Grants A 72 and A 85).

Conflict of interest: None.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
View Abstract