SEARCH

SEARCH BY CITATION

Keywords:

  • Acute symptomatic seizure;
  • Epidemiology;
  • Epilepsy;
  • Seizure recurrence;
  • Mortality

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Purpose: To compare mortality and subsequent unprovoked seizure risk in a population-based study of acute symptomatic seizure and first unprovoked seizure due to static brain lesions.

Methods: We ascertained all first episodes of acute symptomatic seizure and unprovoked seizure due to central nervous system (CNS) infection, stroke, and traumatic brain injury (TBI). Subjects were residents of Rochester, Minnesota, identified through the Rochester Epidemiology Project’s records-linkage system between 1/1/55 and 12/31/84. Information was collected on age, gender, seizure type, etiology, status epilepticus (SE), 30-day and 10-year mortality, and subsequent episodes of unprovoked seizure.

Results: Two hundred sixty-two individuals experienced a first acute symptomatic seizure and 148 individuals experienced a first unprovoked seizure, all due to static brain lesions. Individuals with a first acute symptomatic seizure were 8.9 times more likely to die within 30 days compared to those with a first unprovoked seizure [95% confidence intervals (CI) = 3.5–22.5] after adjustment for age, gender, and SE. Among 30-day survivors, the risk of 10-year mortality did not differ. Over the 10-year period, individuals with a first acute symptomatic seizure were 80% less likely to experience a subsequent unprovoked seizure compared with individuals with a first unprovoked seizure [adjusted rate ratio (RR) = 0.2, 95% CI = 0.2–0.4].

Discussion: The prognosis of first acute symptomatic seizures differs from that of first unprovoked seizure when the etiology is stroke, TBI, and CNS infection. Acute symptomatic seizures have a higher early mortality and a lower risk for subsequent unprovoked seizure. These differences argue against the inclusion of acute symptomatic seizures as epilepsy.

For many years, epilepsy was operationally defined as recurrent unprovoked seizures (Hauser & Kurland,1975; Hauser et al., 1993; ILAE, 1993). In the landmark article by Hauser and Kurland (1975), epilepsy was conceptually defined as “a condition in which seizures … tend to recur chronically because of a persistent and possibly progressive structural or physiologic abnormality of brain tissue.” Recurrent epileptic attacks (i.e., epilepsy) were operationally defined as “at least two convulsive episodes that occurred at different times, without an identifiable causative metabolic or acute structural abnormality.” Later empirical evidence for the validity of the 1975 operational definition came from an evaluation of the risk of recurrence after two unprovoked seizures (Hauser et al., 1998).

In 2005, a new International League Against Epilepsy (ILAE) definition of epilepsy described epilepsy as “a disorder of the brain characterized by an enduring predisposition to generate epileptic seizures and by the neurobiologic, cognitive, psychological, and social consequences of this condition” (Fisher et al., 2005). In a section devoted to a history of at least one seizure, this paper clearly states that “at least one seizure is required to establish the presence of epilepsy” and “the definition does not include a requirement that the seizure be ‘unprovoked.’” Therefore, the new definition includes single seizures, even if acute symptomatic, under the rubric of epilepsy.

If, in fact, the circumstances of a seizure do not need to be taken into account when identifying epilepsy, then one might predict that the outcomes following a seizure in the setting of an acute insult (i.e., provoked acute seizure) would be the same as those following a seizure that occurs in the presence of the same type of insult but after the acute phase of the initial insult has subsided (i.e., unprovoked remote symptomatic). These are conditions that the ILAE definition of epilepsy might consider evidence of an enduring predisposition to seize. To test this assertion, we undertook an analysis of the Rochester Epidemiology Project data to assess whether the prognosis of a first acute symptomatic seizure differs from that of a first unprovoked seizure in terms of short- and long-term mortality and the occurrence of subsequent unprovoked seizures. We reasoned that if the prognosis differed for acute symptomatic seizures and unprovoked seizures, then the two were likely different, and if prognosis was similar, then they may be the same. To optimize the comparability of the groups, we considered only those first seizures that are associated with specific static brain lesions, common to both acute symptomatic and unprovoked seizures.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

We compared the short- and long-term mortality as well as the risk for subsequent unprovoked seizure among the Rochester, Minnesota, residents who experienced a first acute symptomatic seizure or a first unprovoked seizure between 1/1/55 and 12/31/84. Seizures were ascertained through the records-linkage system of the Rochester Epidemiologic Project, which links all medical records from medical facilities in Southeastern Minnesota and includes inpatient, outpatient, emergency room, doctor’s private office, and home visits. To identify first seizures, we reviewed the medical records of all residents with a diagnosis of seizure, convulsion, epilepsy, or conditions known to be related to seizures.

Acute symptomatic seizure

Acute symptomatic seizure was defined as a seizure in close temporal association with a transient central nervous system (CNS) insult or transient systemic disturbance. These seizures were presumed to be an acute manifestation of the insult. The definition of “close” varied with the insult (Annegers et al., 1995). Individuals with unprovoked seizures before their acute symptomatic seizure or with prior acute symptomatic seizures were excluded in order to compare the prognosis of first acute symptomatic seizures and first unprovoked seizures.

First unprovoked seizure

We defined first unprovoked seizure as a seizure occurring at least one week after the etiologies included in this report. Individuals with acute symptomatic seizure before their first unprovoked seizure were excluded.

Etiologies included

Individuals with first seizures were included only if their seizure occurred in association with stroke, traumatic brain injury (TBI), and CNS infection. For stroke and TBI, close was defined as within 7 days of the event or associated with extension or progression of the stroke or recovery from the TBI (Annegers et al., 1995). For CNS infection, close was defined as during the active infection, confirmed by laboratory tests or clinical symptoms (Annegers et al., 1995).

Mortality

We examined deaths during the first 30 days after either an acute symptomatic seizure or after the first unprovoked seizure (short-term mortality) as well as death in the subsequent 10 years among subjects surviving the first 30 days (long-term morality). The latter was done to eliminate the influence of the early events related to the underlying etiology of the seizures under the assumption that deaths occurring soon after the first seizure and those occurring later would be related to different factors.

Subsequent unprovoked seizure

We defined unprovoked seizure as a seizure without an identified proximate precipitant (Hauser et al., 1993). For the group with acute symptomatic seizure, recurrence was defined as the occurrence of a first unprovoked seizure. For the group with first unprovoked seizure, recurrence was defined as the occurrence of subsequent epilepsy.

Factors studied

Information was collected on age at the first seizure, gender, seizure type, etiology, and whether or not the first seizure was an episode of status epilepticus (SE).

Age at acute symptomatic seizure

Age at the incident seizure was grouped as: <1, 1–64, and ≥65 years for analysis of short-term mortality and occurrence of subsequent unprovoked seizure. Because there were no deaths over the long-term among infants, age was categorized as <65 years and ≥65 years for the analysis of long-term mortality.

Etiology of acute symptomatic seizure

Etiology was categorized as CNS infection, TBI, cerebrovascular disease.

Seizure type

Seizure type was categorized by one of us (WAH) based on criteria of the International League Against Epilepsy (ILAE, 1993). Seizures were categorized as: generalized not otherwise specified (NOS), secondarily generalized, focal seizures without secondary generalization, and other (including primary generalized seizures) and unknown. Generalized seizures NOS included cases with generalized tonic, clonic, and tonic–clonic seizures where focal onset could not be ruled out. Focal seizures included cases with simple or complex partial seizures, which did not generalize. The classification of seizure type was based upon the description written by the attending medical staff, neurologists, nurses, and trained medical personnel such as electroencephalography (EEG) technicians and the ambulance attendants. Where the clinical characteristics of the seizure could not be determined, seizures were characterized as unknown seizure type.

Status epilepticus (SE)

We defined SE as a single clinical seizure, lasting more than 30 min or a series of seizures lasting more than 30 min without intervening recovery of consciousness (Hauser, 1982).

Statistical analysis

We used Student’s t-test to compare continuous variables and the chi-square statistic to compare categorical variables. Kaplan-Meier survival analysis (Breslow & Day, 1987) was used to calculate the cumulative incidence of short-term and long-term mortality, comparing residents with a first acute symptomatic seizure to residents with a first unprovoked seizure, excluding those with CNS tumor. Surviving patients were followed until the end of the study (February 1996), date of last medical contact in Southeastern Minnesota or, if they migrated from the area, the date of last Southeastern Minnesota residency.

Kaplan-Meier survival analysis was also used to calculate the 10-year cumulative incidence of subsequent unprovoked seizure, comparing residents with a first acute symptomatic seizure to residents with a first unprovoked seizure, excluding those with CNS tumor. Subjects were followed until they experienced a subsequent unprovoked seizure, and censored if they died before they experienced a recurrence or if they migrated out of Southeastern Minnesota.

The log-rank statistic was used to compare the risk of death or of subsequent unprovoked seizure for subjects with and without a first acute symptomatic seizure.

Cox proportional hazards regression (Breslow & Day, 1987) was used to examine the effect of a first acute symptomatic seizure compared to a first unprovoked seizure on the risk short-term mortality, long-term mortality, and subsequent unprovoked seizure. Separate models examined the risk of each of these endpoints controlling for age, gender, SE, and stratifying by etiology of the first seizure.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Among residents of Rochester, Minnesota, 262 individuals experienced a first acute symptomatic seizure caused by a neurological insult and 148 individuals experienced a first unprovoked seizure associated with the same insults during 1/1/55 through 12/31/84. Etiology of acute symptomatic seizure was stroke in 34.7%, TBI in 34.7%, and CNS infection in 30.6% (Table 1). Etiology of the first unprovoked seizure was stroke in 68.2%, TBI in 25.0%, and CNS infection in 6.8%.

Table 1.   Descriptive information on acute symptomatic seizure or first unprovoked seizure from 1955–1984 in Rochester, Minnesota by common underlying etiology
VariableStrokeTBICNS infection
Acute (N = 91)Unprovoked (N = 101)Acute (N = 91)Unprovoked (N = 37)Acute (N = 80)Unprovoked (N = 10)
  1. Data are frequency (percent).

  2. aExcludes neonatal seizure type.

  3. bExcludes deaths within first 30 days.

Gender
 Male45 (49.4)45 (44.6)54 (59.3)23 (62.6)55 (68.8)7 (70.0)
 Female46 (50.6)56 (55.5)37 (40.7)14 (37.8)25 (31.2)3 (30.0)
Age (years)
 <10 (0.0)1 (1.0)5 (5.5)1 (2.7)17 (21.2)0 (0.0)
 1–192 (2.2)2 (2.0)44 (48.4)13 (35.1)43 (53.8)6 (60.0)
 20–6426 (28.6)31 (30.7)28 (30.8)18 (48.7)14 (17.5)4 (40.0)
 ≥6563 (69.2)67 (66.3)14 (15.4)5 (13.5)6 (7.5)0 (0.0)
Seizure typea
 Generalized NOS15 (16.5)31 (30.7)49 (53.8)15 (40.5)25 (31.2)1 (10.0)
 Secondarily generalized12 (13.2)0 (0.0)2 (2.2)0 (0.0)10 (12.5)0 (0.0)
 Focal only61 (67.0)66 (65.3)37 (40.7)20 (54.1)42 (52.5)7 (70.0)
 Other/Unknown3 (3.3)4 (4.0)3 (3.3)2 (5.4)3 (3.8)2 (20.0)
Status epilepticus
 Yes33 (36.3)23 (22.8)10 (11.0)6 (16.2)47 (58.8)9 (90.0)
 No58 (63.7)78 (77.2)81 (89.0)31 (83.8)33 (41.2)1 (10.0)
Deaths
 30 days38 (41.8)5 (5.0)10 (11.0)0 (0.0)8 (10.0)0 (0.0)
 10 yearsb36 (67.9)62 (64.6)11 (13.6)9 (24.3)7 (9.7)0 (0.0)

Individuals age 65 years and older accounted for 31.7% (N = 83) of the group with first acute symptomatic seizure and 48.7% (N = 72) of the group with first unprovoked seizure (p = 0.001). Most cases were male (55.9%). The majority of acute symptomatic seizures were either generalized NOS (34.0%) or focal (53.4%), compared with 31.8% and 62.8% in the group with first unprovoked seizure (p = 0.31).

Short-term mortality

During the first 30 days, there were 56 deaths in the group with first acute symptomatic seizure compared to five deaths in the group with first unprovoked seizure. A first acute symptomatic seizure was associated with a significantly higher risk of 30-day mortality (21.4%, 95% CI = 16.9–26.9%) compared to a first unprovoked seizure (3.4%, 95% CI = 1.4–7.9%, p < 0.001) (Fig. 1). Individuals with a first acute symptomatic seizure were 6.9 times more likely to die within 30 days compared to those with a first unprovoked seizure (95% CI = 2.8–17.3). This association remained even after adjustment for age, gender, and SE (rate ratio [RR] = 8.9, 95% CI = 3.5–22.5) (Table 2). Stratified by etiology of the first seizure, individuals with acute symptomatic seizure were 11.6 times more likely to die within 30 days compared to those with a first unprovoked seizure (95% CI = 4.6–29.1). This association remained even after adjusting for age, gender, and SE (RR = 11.4, 95% CI = 4.5–29.0).

image

Figure 1.  Cumulative risk of death in the first 30 days after first acute symptomatic seizure and first unprovoked seizure.

Download figure to PowerPoint

Table 2.   Prognosis of acute symptomatic seizures compared to first unprovoked seizures, restricting to seizures in association with stroke, CNS infection, and traumatic brain injury
Variable N (% dead)Crude RR (95%CI)Adjusted RR (95%CI)
30-day mortality
 First seizure
  Acute symptomatic seizure262 (21.4)6.9 (2.8–17.3)8.9 (3.5–22.5)
  First unprovoked seizure148 (3.4)1.0 (Referent)1.0 (Referent)
 Age (years)
  <124 (4.2)0.2 (0.02–1.2)0.1 (0.01–0.7)
  1–64231 (10.4)0.4 (0.3–0.7)0.3 (0.2–0.6)
  ≥65155 (23.2)1.0 (Referent)1.0 (Referent)
 Gender
  Female181 (14.9)1.0 (0.6–1.7)1.0 (0.6–1.7)
  Male229 (14.9)1.0 (Referent)1.0 (Referent)
 Status epilepticus
  Yes106 (18.9)1.4 (0.8–2.4)1.0 (0.6–1.7)
  No304 (13.5)1.0 (Referent)1.0 (Referent)
10-year mortality in 30-day survivors
 First seizure
  Acute symptomatic seizure206 (26.2)0.5 (0.4–0.7)0.7 (0.5–1.0)
  First unprovoked seizure143 (49.7)1.0 (Referent)1.0 (Referent)
 Age (years)
  <65230 (12.6)0.09 (0.06–0.1)0.09 (0.06–0.1)
  ≥65119 (80.7)1.0 (Referent)1.0 (Referent)
 Gender
  Female154 (39.6)1.3 (0.95–1.9)0.8 (0.6–1.1)
  Male195 (32.8)1.0 (Referent)1.0 (Referent)
 Status epilepticus
  Yes86 (45.4)1.6 (1.1–2.3)1.4 (0.9–2.0)
  No263 (32.7)1.0 (Referent)1.0 (Referent)
  N (% subsequent US)Crude RR (95%CI)Adjusted RR (95%CI)
10-year risk of subsequent unprovoked seizure (US)
 First seizure
  Acute symptomatic seizure262 (13.0)0.2 (0.1–0.3)0.2 (0.16–0.4)
  First unprovoked seizure148 (48.7)1.0 (Referent)1.0 (Referent)
 Age (years)
  <124 (16.7)0.2 (0.1–0.7)0.5 (0.2–1.6)
  1–64231 (19.0)0.3 (0.2–0.5)0.5 (0.3–0.7)
  ≥65155 (37.4)1.0 (Referent)1.0 (Referent)
 Gender
  Female181 (28.7)1.3 (0.9–1.8)0.9 (0.6–1.3)
  Male229 (23.6)1.0 (Referent)1.0 (Referent)
 Status epilepticus
  Yes106 (34.9)1.8 (1.2–2.7)1.9 (1.2–2.8)
  No304 (22.7)1.0 (Referent)1.0 (Referent)

We evaluated short-term mortality in groups defined by etiology. Among individuals with stroke, 30-day mortality was 41.9% (95% CI = 32.5–52.8%) for first acute symptomatic seizure compared to 5.0% (95% CI = 2.1–11.5%, p < 0.001) for first unprovoked seizure. For individuals with TBI and CNS infection, the risk of 30-day mortality was 11.0% (95% CI = 6.1–19.6%) for acute symptomatic seizures caused by TBI and 9.9% (95% CI = 5.1–18.8%) for those with acute symptomatic seizures caused by CNS infection. There were no deaths observed within 30 days after the first unprovoked seizure among those with TBI and CNS infection.

Long-term mortality

Over the next 10 years among 30-day survivors, 71 deaths occurred in the first unprovoked seizure group and 54 deaths occurred in the first acute symptomatic seizure group. Individuals with a first unprovoked seizure had a significantly higher risk of 10-year mortality (54.9%, 95% CI = 46.2–64.0%) compared to those with a first acute symptomatic seizure (33.4%, 95% CI = 26.3–41.7%, p ≤ 0.001) (Fig. 2). Individuals with a first acute symptomatic seizure had a significantly lower risk of 10-year mortality compared to those with a first unprovoked seizure (RR = 0.5, 95% CI = 0.4–0.7). This association was no longer significant after adjusting for age, gender, and SE (RR = 0.7, 95% CI = 0.5–1.0) (Table 2). Stratified by etiology of the first seizure, acute symptomatic seizure was associated with the same risk of 10-year mortality as a first unprovoked seizure (RR = 1.0, 95% CI = 0.7–1.5). This association remained unchanged after adjusting for age, gender, and SE.

image

Figure 2.  Cumulative risk of 10-year mortality in 30-day survivors of first acute symptomatic seizure and first unprovoked seizure.

Download figure to PowerPoint

We evaluated 10-year mortality in groups defined by etiology. Among individuals with stroke, the risk of 10-year mortality was 73.8% (95% CI = 60.4–85.6%) for first acute symptomatic seizure and 71.6% (61.0–81.4%, p = 0.68) for first unprovoked seizure. The risk of 10-year mortality was also similar for a first acute symptomatic seizure (21.7%, 95% CI = 11.9–37.5%) and a first unprovoked seizure (28.1%, 95% CI = 15.6–47.4%, p = 0.40) among individuals with TBI. However, among those with a first seizure in association with CNS infection, only those with a first acute symptomatic seizure experienced deaths in the 10-year period (13.6%, 95% CI = 6.6–27.1%).

Subsequent unprovoked seizure

Over 10 years of follow-up, subsequent unprovoked seizures occurred in 34 individuals with first acute symptomatic seizure and 72 with first unprovoked seizure. The first unprovoked seizure group had a significantly higher risk of subsequent unprovoked seizure (64.8%, 95% CI = 55.1–74.4%) compared to the first acute symptomatic seizure group (18.7%, 95% CI = 13.7–25.4%, p < 0.001) (Fig. 3). A first acute symptomatic seizure was protective against having a subsequent unprovoked seizure (RR = 0.2, 95% CI = 0.1–0.3). This association remained even after adjusting for age, gender, and SE (RR = 0.2, 95% CI = 0.2–0.4) (Table 2). Similar results were observed after stratification by etiology of the first seizure.

image

Figure 3.  Cumulative risk of subsequent unprovoked seizure after first acute symptomatic seizure and first unprovoked seizure.

Download figure to PowerPoint

The risk for subsequent unprovoked seizure was examined for each seizure etiology. Among those with stroke, the risk of subsequent unprovoked seizure was 33.0% (95% CI = 20.7–49.9%) among those with a first acute symptomatic seizure and 71.5% (95% CI = 59.7–81.9%, p = 0.001) among individuals with a first unprovoked seizure. Among those with TBI, the risk of subsequent unprovoked seizure was 13.4% (95% CI = 7.0–24.8%) for first acute symptomatic seizure and 46.6% (95% CI = 30.4–66.3%, p < 0.001) for first unprovoked seizure. The risk of subsequent unprovoked seizure in the first acute symptomatic seizure group was 16.6% (95% CI = 9.5–28.0%) and 63.5% (95% CI = 21.2–98.6%, p = 0.010) in the first unprovoked seizure group among those with CNS infection (Table 2).

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

We found a significantly different prognosis for first acute symptomatic seizure compared to first unprovoked seizure when the etiology was stroke, TBI, and CNS infection. Compared with first unprovoked seizures, individuals with a first acute symptomatic seizure were significantly more likely to die in the first 30 days after the seizure and less likely to experience a subsequent unprovoked seizure over the next 10 years.

In the framework of discussions about concepts that should be applied to classification in epilepsy, we have taken the approach of Berg and Blackstone (2006) to determine if a first acute symptomatic seizure is or is not epilepsy. As Berg and Blackstone have discussed, it is necessary to directly compare two or more entities to one another in order to determine if they are the same or different. Such head-to-head comparisons have been suggested as an approach to classifying epilepsy syndromes, but they are central to any endeavor to define and classify an entity.

Our data suggest that for static brain lesions, etiologies that are shared by acute symptomatic seizure and unprovoked seizure, these seizures are in fact different from one another. Furthermore, a first acute symptomatic seizure does not seem to be associated with “enduring predisposition to generate epileptic seizures” (Fisher et al., 2005), despite the presence of common brain insults. The risk for subsequent unprovoked seizure was 80% less in the group with first acute symptomatic seizure compared to that with first unprovoked seizure.

There are no published population-based studies on the “neurobiologic, cognitive, psychological, and social consequences” of acute symptomatic seizures; therefore, it is not currently possible to evaluate whether they meet this criteria of the new epilepsy definition.

We studied first acute symptomatic seizures and first unprovoked seizures that shared the same etiologies. This design is a strength of our study, because we were able to examine prognosis in the setting of the same type of underlying brain insult. It is also a weakness, because we are unable to draw inferences about first acute symptomatic seizures due to other causes, since this other group of acute symptomatic seizures has no parallel comparison group in the Rochester data. We were also unable to assess whether the severity of the underlying brain insult differed for the acute symptomatic seizures group compared with the unprovoked seizure group as these data were not collected. Nonetheless, the etiologies analyzed in this study are the most likely to result in a static brain lesion that could be associated with an enduring predisposition to seize.

Our epidemiologic data do not provide evidence that first acute symptomatic seizures should be included in the diagnosis of epilepsy when these seizures are caused by static brain lesions. Future studies should compare the prognosis of first acute symptomatic seizure and first unprovoked seizure caused by metabolic derangements. For example, prognosis of individuals with acute symptomatic seizures caused by hyponatremia compared to first unprovoked seizure in association with sodium levels in the normal range. Such comparisons are likely possible in settings different from ours and could further our understanding of differences and similarities between first acute symptomatic seizure and first unprovoked seizure. Such understanding can help provide a rationale for proposed definitions of epilepsy.

Acknowledgments

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

The authors thank Dr. Anne Berg for her thoughtful comments. In addition, we confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Disclosures: None of the authors have any financial interest in relation to this study or its results.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References