Current issues of ACP Journal Club are published in Annals of Internal Medicine


Younger age, first seizure during sleep or on awakening, tongue biting, and epileptic or abnormal results on electroencephalograms predicted seizure recurrence in adults with an idiopathic first seizure

ACP J Club. 1991 Sept-Oct;115:58. doi:10.7326/ACPJC-1991-115-2-058

Source Citation

van Donselaar CA, Geerts AT, Schimsheimer RJ. Idiopathic first seizure in adult life: who should be treated? BMJ. 1991 Mar 16;302:620-3.



To determine the recurrence rate of seizure in adult patients with an untreated idiopathic first seizure.


Inception cohort assembled at the time of idiopathic first seizure and followed for 1 to 2 years.


1 university hospital and 3 general hospitals in the Netherlands.


All patients aged ≥ 15 years with an idiopathic first seizure, as determined by 3 neurologists according to prespecified diagnostic criteria, were included. Patients were excluded who had a seizure other than febrile convulsions in the past, who had major abnormalities on computed tomography, who were treated immediately with antiepileptic drugs, who presented with status epilepticus or a second seizure within 24 hours, or whose seizure lasted > 30 minutes. 226 patients were referred and the study cohort for the assessment of risk for recurrence comprised 151 patients.

Assessment of prognostic factors

Data pertaining to 9 potential predictors of subsequent seizure were collected (age, sex, family history, interval between first seizure and first visit, time of occurrence, provocative circumstances, standard electroencephalogram, combined results of standard and sleep-deprivation electroencephalogram, and tongue biting).

Main outcome measure

Occurrence of a second seizure.

Main results

The overall recurrence rate was 40% (95% CI 32% to 48%) at 2 years. Most recurrences (71%) occurred within 6 months. Higher recurrence rates at 2 years were associated with younger age (15 to 24 years, 50% [CI 36% to 64%] vs 25 to 44 years, 39% [CI 25% to 43%] vs 45 to 85 years, 29% [CI 16% to 42%]), occurrence of the first seizure during sleep or on awakening (72% [CI 54% to 90%] vs 32% for seizure occurring during the day [CI 23% to 41%]), and tongue biting (present, 55% [CI 43% to 67%] vs absent, 27% [CI 17% to 37%]). The cumulative risk for recurrence at 2 years was 81% (CI 66% to 97%) in patients with epileptic discharges on a standard or partial sleep-deprivation electroencephalogram, 39% (CI 27% to 51%) in patients with other electroencephalographic abnormalities, and 12% (CI 3% to 21%) in patients with normal electroencephalograms. Family history, provocative circumstance, sex, and interval between first seizure and first visit to the hospital did not correlate with the risk for recurrence.


Younger age, occurrence of first seizure during sleep or on awakening, tongue biting, and epileptic or abnormal results on standard and sleep-deprivation electroencephalograms were predictive of a second seizure in adult patients with an untreated idiopathic first seizure.

Source of funding: TNO Research Committee on Epilepsy.

Address for article reprint: Dr. C.A. van Donselaar, Department of Neurology, University Hospital, Rotterdam-Dijkzigt, 3015 GD Rotterdam, the Netherlands.


The decision to treat or follow a patient who presents with a first unprovoked seizure is a frequent problem in the daily practice of emergency medicine and neurology, and should be based on the risk for recurrence. In the hospital-based inception cohort study of van Donselaar and colleagues, the patients are described in detail, but possible referral biases are not addressed. Nearly complete follow-up was obtained, but the definition of recurrence is not elaborated and possible recurrences of partial seizures may have been overlooked. In addition, no mention is made of blinding of the follow-up observers to details of the initial presentation, allowing possible diagnostic suspicion bias.

The 40% 2-year seizure recurrence rate is about twice that recently reported (1) (in which 80% of patients were treated with anticonvulsants at presentation), but is similar to the 36% 2-year recurrence rate observed in a mostly untreated group of children (2). Generalized spike-wave discharges on electroencephalogram were found to markedly increase the recurrence risk in all of the above studies; various studies have found other prognostic factors to be predictive of seizure recurrence, but without uniform agreement.

The study by van Donselaar and colleagues is among the best of those assessing the outcome of untreated first unprovoked seizures. Additional studies will need to add a careful, blinded search in follow-up for recurrent partial seizures, especially partial complex seizures, that may have been overlooked by studies to date and whose discovery may increase significantly the assessment of seizure recurrence risk.

Preston C. Calvert, MD
The Neurology CenterFalls Church, Virginia, USA


1. Hauser WA, Rich SS, Annegers JF, Anderson VE. Seizure recurrence after a 1st unprovoked seizure: an extended follow-up. Neurology. 1990;40:1163-70.

2. Shinnar S, Berg AT, Moshe SL, et al. Risk of seizure recurrence following a first unprovoked seizure in childhood: a prospective study. Pediatrics. 1990;85:1076-85.