Presentation and course
| || |
• First seizure usually manifests as either generalized tonic-clonic or focal onset with secondary generalization.
| || |
• A careful history is required to elicit possible prior events that may not have been recognized as seizures.
| || |
• Neonatal seizures, febrile seizures, and status epilepticus require separate parameters for workup and carry different prognoses from other first seizures.
In patients who present with a single seizure, including multiple events in one day, the range of seizure manifestations is fairly limited. By far, the most common is a generalized tonic-clonic seizure (39). Seizures of focal onset with secondary generalization are also common (39). Less common are complex partial seizures and atypical prolonged absence seizures. Although patients with these seizure types can present after a single episode, they rarely do so. More commonly, on careful history, there is a several-month history of events often not identified as seizures by the patient or family, with ultimate presentation to medical attention with a convulsive event. Typical absence seizures, myoclonic seizures, and atonic seizures almost never present with a single episode and are, thus, not part of the spectrum of a single seizure presentation (22). Febrile seizures are a common seizure type in children, but are considered a form of acute symptomatic or provoked seizure and, therefore, are not part of this discussion. Neonatal seizures, defined as those occurring in the first month of life, are more difficult to classify as unprovoked, require a more extensive evaluation, and are managed differently. They are, therefore, also excluded from the category of first unprovoked seizures.
First seizures can be prolonged and fit the diagnostic criteria for status epilepticus (lasting 30 minutes or more). The diagnostic evaluation of these patients is more extensive and, therefore, they are excluded from most parameters addressing diagnostic evaluation of a first unprovoked seizure, which is fairly limited (22; 36; 29). It should be noted, however, that recurrence frequency and long-term prognosis is similar in such patients, and it is unclear if management is significantly different than in those presenting with a brief initial seizure (18; 37).
Prognosis and complications
The long-term prognosis of both children and adults who present with a first unprovoked seizure is favorable. Fewer than half will have a recurrence. Of those that do, most have a very mild epilepsy with only occasional seizures; the vast majority of these are in 2-year remission within 5 years of diagnosis (08; 35; 48; 03; 32). Initial seizure frequency is a strong predictor of future intractability, and a single seizure is the extreme low end of that spectrum. In addition, the epilepsy syndromes that tend to be intractable, such as Lennox-Gastaut in children and mesial temporal lobe epilepsy in adolescents and adults, rarely present with a first unprovoked seizure, but usually first present to medical attention with a clear history of multiple events over the prior months.
In more than half of these patients, although the first unprovoked seizure is a frightening event, it is an isolated incident and no further episodes will occur. Recurrent seizures occur in 40% to 50% (19; 39; 05; 32; 38). When recurrences do occur, they occur early, with 50% occurring in the first 6 months and 80% in the first 2 years. Late recurrences are uncommon (41; 18).
First unprovoked seizures can be classified etiologically as remote symptomatic (involving a structural or metabolic abnormality), or as of unknown cause (previously cryptogenic/idiopathic) (10; 11; 04; 47). Determination of etiology may be dependent on workup and does affect prognosis, though it often won’t alter immediate management.
Risk factors for recurrence include a remote symptomatic etiology (including known brain lesion or insult or a new significant brain imaging abnormality), an abnormal EEG, and the occurrence of the first seizure in sleep (28). Age is not an important risk factor for recurrence, though age is clearly relevant when assessing the potential consequences of a recurrent seizure and when making treatment decisions. The number of seizures in 24 hours, family history of seizures, duration of seizure, and prior history of febrile seizures also have not been proven to increase recurrence risk (05; 41; 26).
Remote symptomatic etiology has consistently been shown to increase risk for subsequent seizures, with one meta-analysis suggesting a recurrence rate of 57% for patients with this etiology compared with 32% for those without (05). Other analyses have generally confirmed this finding (40; 46; 32; 27).
Similarly, abnormal findings on EEG indicate an increased risk of recurrence (19; 20; 05; 22; 27; 38). Distinction between types of abnormalities is of interest to the epileptologist, but in the initial assessment, the important point is that any abnormality, particularly an epileptiform one, confers additional risk (05; 27; 38).
In studies of first seizures that occur out of sleep, even after controlling for the initial presentation of syndromes that typically include such events (eg, benign epilepsy of childhood with centrotemporal spikes), recurrence is increased (24; 38). However, in these cases, recurrences also tend to occur in sleep, reducing the likelihood of patient injury (38).
A focal seizure is associated with a higher recurrence risk, but is also more common both in those with a remote symptomatic etiology and those with an abnormal EEG. Once we adjust for etiology and EEG, there are no clear differences on multivariate analysis in recurrence risk for those who present with a focal seizure compared to those who present with a generalized seizure. Therefore, a first focal seizure does not need to be approached differently, except for the need for imaging studies (22; 23; 29; 18).
Treatment with chronic antiepileptic drugs following a first unprovoked seizure reduces recurrence risk by approximately half, but has no effect on long-term prognosis. This very significant statement will be discussed at greater length in the Management section below.
Although the majority of patients with a first unprovoked seizure will never have another seizure, the first seizure is the initial presentation of a seizure disorder or epilepsy in a substantial minority. The question arises: at what point should we classify someone as having epilepsy? The traditional definition has been two or more unprovoked seizures more than 24 hours apart (10; 11; 04). In 2014, an ILAE report proposed a new and controversial definition that would include patients with a single unprovoked seizure and a high (> 60%) risk of seizure recurrence within 10 years (14). From a practical standpoint, this in no way alters the management of a first seizure. The difficulty continues to lie in appropriately assessing this statistical risk. The ILAE report concedes that “No formula can be applied for additive risks since data are lacking on how such risks combine; such risks will have to be decided by individualized considerations.” Management choices, including the use of antiseizure medications, must therefore be individualized as well.
Following a second seizure, the risk of further seizures in both children and adults is approximately 70% (21; 41); in other words, after a second seizure, there is no group with a low recurrence risk of additional seizures, and both the diagnosis of epilepsy and the recommendation for treatment is clear. With the exception of children with some of the age-dependent self-limited epilepsy syndromes, such as benign epilepsy of childhood with centrotemporal spikes, treatment is generally indicated following a second seizure. These patients now meet the diagnostic criteria for epilepsy.
Independent risk factors for subsequent seizures include remote symptomatic etiology and the occurrence of the second seizure within 6 months of the first seizure. Treatment, again, does reduce recurrence risk by approximately 50%, similar to the effect following a first seizure; once again, there is little evidence that it alters long-term prognosis.
The population with a single seizure has a much lower frequency of comorbidities than patients with newly diagnosed epilepsy. To address perhaps the most common concern, there is no convincing evidence that brief seizures cause brain damage (16; 06; 23). Patients and their families should be reassured on this point.
Other comorbidities are also less frequent in this population. For example, in a cohort of children who presented with a first unprovoked seizure and were assessed 10 years later, those with two or more seizures had more school difficulties and cognitive and behavioral problems than those with a single seizure. The group with an isolated seizure did not appear any different than their siblings (45).
Another major concern today is the risk of sudden unexplained death in epilepsy (SUDEP). The first seizure population is not high risk for SUDEP. There have been a few cases of SUDEP reported in studies of first seizures. All have occurred in patients with established epilepsy following at least three seizures (32; 43; 44). Thus, concern about SUDEP, although influencing treatment decisions in patients with newly diagnosed epilepsy, should not affect whether or not to treat following the first seizure.
All of the following cases are fictitious.
Clinical vignette 1. Marie D. was a 5-year-old girl with no significant medical history. She was in her usual state of good health until this evening, when her parents awoke to noises coming from her room. Her mother found her thrashing in bed and thought she was having a nightmare, but when she uncovered the child she saw jerking movements of her whole body and foam coming from her mouth. Marie did not respond to her mother calling her name or touching her for 2 minutes, when the child stopped moving and then “woke up,” but was confused. Marie’s father called an ambulance while her mother was watching her.
By the time the family arrived in the emergency room, Marie was awake and crying, but easily consoled with a coloring book. However, her mother described to the emergency room physician that Marie was sleepy during the ambulance ride and did not seem “herself” for several minutes after the event. At this point, Marie’s parents were more upset than she was. After a careful examination in the emergency department, the physician explained to the parents that immediate workup beyond basic blood work was unnecessary. The physician referred the child for an outpatient EEG, which revealed spikes over the centrotemporal regions of the head that increased during sleep. After the syndrome of childhood epilepsy with centrotemporal spikes was explained to the parents, they were reassured. Medication was not started. When Marie had a second and similar seizure 8 months later, her parents did not call the ambulance and instead took her to her pediatrician in the morning, who reassured the family that she was well. Marie had no other medical issues and went on to have a total of four brief seizures at night over the next 3 years before “outgrowing” her epilepsy syndrome, as expected.
This is an example of how recognition of the pattern of onset of a syndrome can prevent unnecessary testing and allow reassurance and a reliable prognosis for the family.
Clinical vignette 2. Joe, a 17-year-old man, presented to the emergency department with a tonic-clonic seizure. He had been out late at a party the night before and was sleep deprived. No prior history of seizure was obtained, and he was referred to clinic as a first unprovoked seizure. Outpatient EEG and MRI were normal. On evaluation in the office, he denied a history of prior seizures. However, on careful questioning, he gave an excellent description of what were best described as déjà vu phenomena, with sensations that he had been experiencing the same situation over and over, that had been occurring for the past 6 months. When asked if his friends could tell he was having these symptoms, he replied that they told him that he looked “spacey.” Although Joe was referred as a first unprovoked seizure, he had in fact had epilepsy for the past 6 months and finally presented to medical attention when, in the context of sleep deprivation, he had a convulsive event. Over time, his syndrome turned out to be temporal lobe epilepsy, based on subsequent EEGs.
Distinguishing a case like Joe’s from a true first unprovoked seizure is very important; a careful history is crucial.
Clinical vignette 3. Ms. Smith, a 25-year-old woman living in Manhattan, had a generalized tonic-clonic seizure lasting 1 minute. Neurologic and physical evaluation in the emergency department was unremarkable. Subsequent EEG and MRI were normal. At her follow-up visit in the office, she wanted to know whether she needed to be on antiepileptic drugs. She did not drive or operate heavy machinery. She was in graduate school and in a relationship, planned to get married in the near future, and was considering having children in a few years. A careful history elicited no prior suspicious events.
In this young woman, the risk-benefit ratio clearly favored no treatment unless a second event occurred. She was counseled to take seizure precautions, but no treatment was started. She did not have a second event in the next few years, and the burdens of taking medication and repeated testing were thus avoided.
Clinical vignette 4. Mrs. Jones, a 70-year-old woman with osteoporosis, hypertension, and heart disease who lived in an assisted living residence, had a generalized tonic-clonic seizure. On arrival to the emergency department, she was somewhat confused. EEG was normal. MRI showed diffuse atrophy as well as a small, old infarct in the right anterior cerebral artery distribution. Neurologic examination when she returned to baseline showed mild dementia. No clear history of prior events was elicited. In this patient, the recurrence risk was high and, more importantly, the potential morbidity of another seizure in a patient with heart disease and osteoporosis was significant. This justified treatment after only one seizure, although, given all the other medications she was taking, a drug with minimal enzyme induction and drug-drug interactions was preferred, if effective.