The Natural History and Prognosis of Epilepsy

Ettore Beghi

Josemir W. Sander

Introduction

For the majority of people with epilepsy, the condition is treatable and the term prognosis generally refers to the probability of attaining seizure freedom after treatment has been started, during treatment, and after drug withdrawal. Little is known about the natural history of the untreated condition as most patients, particularly in industrialized countries, are treated. There are only a few reports, mostly from resource-poor countries, on the prognosis of untreated epilepsy.

Methodologic Issues

As with most other chronic conditions, the prognosis of epilepsy depends on the characteristics of the population concerned, the case definition, the spectrum of severity of the condition, the duration of follow-up, and the choice of prognostic predictors, including treatment. More specifically, the design of an ideal study on the prognosis of epilepsy should include the following:

Well-defined criteria for the inclusion of patients.
Standard (homogeneous) definitions of the prognostic predictors and outcome measures.
Adequate duration of follow-up and proper statistical methods to adjust for those lost to follow-up and limited periods of observation.

Ideally, patients should be enrolled at a similar point in the course of the disease (e.g., the time of diagnosis or the time of the index event) and should be representative of the general population of people with epilepsy. Many patients do not seek medical advice until seizures have recurred; using these people as the population under study will exclude those who have had only one seizure and may well influence the prognosis. Additionally, most studies of the prognosis of epilepsy have been carried out in specialized services, generally including predominantly people with complex, difficult to control epilepsy; thus, selection bias may well affect the prognosis of epilepsy.

Homogeneous (preferably standard) definitions for the most common prognostic indicators should be encouraged. For example, there are inconsistent reports that a family history of epilepsy increases the risk of seizure recurrence; this may be explained by the fact that family history is difficult to ascertain and may be unreliable. Details should also be given of all the putative prognostic predictors considered, to provide a comprehensive overview of the prognosis of the disease and to give the best explanation of the results after controlling for the known prognostic indicators. Prolonged follow-up is required and attempts should be made to obtain information on the outcome in all patients. Outcome measures should be clearly defined and, where possible, reliable indicators should be used. Proper statistical methods (including multivariate analysis models) should be employed to assess the independent role of each prognostic predictor. Most of the differences among studies on the prognosis of epilepsy can be explained by the different methodology, with particular reference to the study design (retrospective or prospective; community based or specialized clinic based), the target population (children and/or adults), the timing of enrollment (interval between first seizure and enrollment), the type of seizure (generalized tonic and/or clonic, partial, and other), the length of follow-up, and the use of antiepileptic drugs.

Overall Prognosis of Epilepsy

The overall prognosis of epilepsy is favorable in the majority of patients. There are several pieces of evidence that support this conclusion. First, reports from several resource-poor countries (where people with epilepsy are largely untreated) give prevalence rates that tend to be broadly similar to those of industrialized countries.³^,⁷^,⁸^,²¹^,²²^,²⁸^,³¹^,⁴⁸^,⁵⁰^,⁵²^,⁵⁹^,⁶⁴^,⁶⁵^,⁶⁹^,⁷¹ Prevalence rates for active epilepsy are usually between 4 and 10 per 1,000 in both settings. In most resource-poor countries the incidence of epilepsy is higher than that in industrialized countries.²⁷^,³⁰^,³⁶^,⁴²^,⁴³^,⁴⁸^,⁵⁰ Incidence rates in developed countries tend to fall within the range of 30 to 60 cases per 100,000, while most reports from resource-poor countries give figures in excess of 70 per 100,000. The increased mortality of epilepsy in resource-poor countries explains only part of the difference between incidence and prevalence. The most likely explanation for the similarity of the prevalence rates, therefore, is spontaneous remission in some people. Additionally, contrary to old reports,⁴⁹ studies done in newly diagnosed patients have consistently shown that 55% to 68% of cases achieve prolonged seizure remission (Table 1).³⁵

Table 1 Population-based Studies on the Prognosis of Epilepsy

Country	Population	% (Duration) remission	Source	Follow-up (yr)	Notes
United States	All ages	55% (10-y) at 20 y	Records linkage	33	Included isolated seizures
United Kingdom	All ages	69% (5-y) at 9 y	General practitioners	9	Included isolated seizures
Sweden	Adults	58% (5-y) at 11 y	Multiple	10	Included isolated seizures
France	All ages	62% (complete) at 5 y	Neurologists	5	Included isolated seizures
Switzerland	All ages	68% (complete) at 10 y	Neurologists	10	Included isolated seizures
Ecuador	All ages	21% (complete) at 4 y	Hospitals	4	Only two or more seizures
Holland	Children	62% (2-y) at 5 y	Hospitals	5	Included isolated seizures
Canada	Children	55% (complete) at 7 y (average)	Child neurologists	>20	Included isolated seizures
Finland	Children	68% (5-y) at 30 y	Multiple	>30	Only two or more seizures
Source: Jallon P. In: Jallon P, Berg AT, Dulac O, et al., eds. Prognosis of Epilepsies. Montrouge: John Libbey; 2003.

Early Prognosis of Unprovoked Seizures and Epilepsy

Prognosis after the First Unprovoked Seizure

The risk of relapse after a first unprovoked seizure has been reported to range from 23% to 71%⁵ and from 14% to 68% when actuarial methods are used. The rates at 2 and 5 years are 21% to 69% and 34% to ≥71%, respectively. Population-based studies²^,²⁹ provide more homogeneous relapse rates at 1 (36% to 37%) and 2 years (43% to 45%). In a systematic review of 16 reports,¹¹ the average overall recurrence risk was 51% (95% confidence interval [CI], 49% to 53%). By 2 years, the recurrence risk was 36% and 47% in prospective and retrospective studies, respectively. After a first unprovoked seizure, the probability of a relapse decreases with time; about
50% of recurrences occur within 6 months of the initial seizure and 76% to 96% within 2 years.¹¹

Risk Factors for Relapse of a First Unprovoked Seizure

The two most consistent predictors of recurrence are a documented etiology of the seizure (as opposed to idiopathic or cryptogenic seizures) and an abnormal (epileptiform and/or slow) electroencephalogram (EEG) pattern.¹¹ The pooled recurrence risk in patients with an idiopathic or cryptogenic first seizure is 32% (95% CI, 28% to 35%), compared with 57% (95% CI, 51% to 63%) for a remote symptomatic seizure (i.e., a seizure with an underlying, nonacute brain problem). The recurrence risk ranges from 27% (95% CI, 21% to 33%) with a normal EEG tracing to 58% (95% CI, 49% to 66%) with an EEG showing epileptiform abnormalities. Epileptiform abnormalities tend to be associated with a higher risk of seizure recurrence than do nonepileptiform abnormalities. The pooled 2-year recurrence risk is lowest for an idiopathic or a cryptogenic first seizure with a normal EEG (24%; 95% CI, 19% to 29%), intermediate for a remote symptomatic seizure (48%; 95% CI, 34% to 62%) with normal EEG or an idiopathic/cryptogenic seizure with an abnormal EEG (48%; 95% CI, 40% to 55%), and highest with a remote symptomatic seizure with an abnormal EEG (65%; 95% CI, 55% to 76%). Seizures occurring during sleep tend to be associated with a higher risk of recurrence in both children and adults.³⁴^,⁵⁷^,⁶⁸ Partial seizures, which are usually associated with a documented brain injury, are also correlated with a higher risk of recurrent seizures, even after controlling for etiology and EEG abnormalities.²^,¹⁵ A positive correlation between seizure relapse and family history of seizures has only been confirmed in patients with idiopathic or cryptogenic first seizures in one study.³³ A history of prior acute symptomatic seizures has occasionally been found to increase the risk of relapse, while evidence is inconclusive or lacking for sex, age, and presentation with status epilepticus.¹¹

Treatment, Risk of Recurrence, and Long-term Prognosis of a First Seizure

There are at least five published randomized studies assessing the effects of treatment of the first unprovoked seizure.¹⁶^,²⁰^,²³^,²⁵^,⁴⁴^,⁴⁶ The results of these studies consistently show that treatment of the first seizure seems to reduce the risk of short-term relapse but is apparently ineffective on the chance of long-term seizure remission. A large multicenter Italian trial²³ (the FIRST study) of 397 children and adults was conducted to assess the effectiveness of treatment of the first seizure on the risk of relapse and the long-term prognosis of epilepsy. Patients seen within 7 days after a first witnessed unprovoked tonic–clonic seizure with or without partial onset were randomized to be treated or to be left untreated until the time of a second seizure. The mean period of observation was 274 days in patients given immediate treatment and 309 days in patients who did not receive treatment. Overall, 36 of 204 treated patients and 75 of 193 untreated patients were referred for seizure relapse. In this trial, the cumulative time-dependent risk of recurrence in treated patients was 17% at 12 months and 25% at 24 months, and in untreated patients was 41% and 51%, respectively. The differences tended to disappear, however, when the end-point was the chance of initiating a 2-year remission. The cumulative probability of long-term remission in the two treatment groups tended to be similar from the second year of follow-up until 15 years after randomization.^41a The results of this study were confirmed by an even larger pragmatic randomized European trial (the MESS study) comparing immediate and deferred treatment for early epilepsy and single seizures.⁴⁴ Patients aged at least 1 month were randomized if both the clinician and the patient were uncertain whether to proceed with treatment. In this trial, 722 patients were randomized to immediate treatment and 721 to deferred treatment. Of these, 404 and 408, respectively, had a single seizure at randomization. Immediate treatment prolonged the time to the first relapse (risk ratio [RR] 1.5; 95% CI, 1.2 to 1.8) and increased the proportion of patients achieving immediate 2-year remission (64% vs. 52%) (p = 0.023). However, at the 2-year follow-up, 32% of those with immediate treatment had had a recurrence compared with 39% of those with deferred treatment. In addition, 92% versus 92% patients achieved 2-year remission at 5 years and 95% versus 96% at 8 years, respectively. The results of these trials tend to confirm several observational reports that the long-term prognosis of the first seizure is substantially unaffected by immediate treatment. However, the comparative effects of the treatment of the first seizure and treatment only on relapse on the chance of long-term remission (without drugs) have not yet been assessed.

Only gold members can continue reading. Log In or Register to continue