Obtaining a complete history is vital, although often difficult during the first contact. The details surrounding the event are crucial, and past medical history must be reviewed. Provoking factors, such as sleep deprivation, drug use, or hypoglycemia, and specific triggers, such as exercise,⁸¹ hot water immersion,⁹⁴ or a history of stress or sexual abuse, must also be evaluated. In adolescence, it is important to determine whether the patient previously had overlooked minor seizures, namely myoclonus in the morning, or simple partial seizures. Some of this information may not be readily forthcoming until a good relationship with the patient has been established. A careful neurologic examination must be performed to determine the presence of Todd paralysis or a more permanent deficit. Screens to detect cardiovascular and metabolic disorders or drug abuse should be done as circumstances dictate.

Nonepileptic events must be identified (see Chapter 68). It is particularly important to recognize hyperekplexia and paroxysmal cardiogenic disorders (e.g., rhythm disorders, including prolonged QT syndrome, or pulmonary arterial hypertension), because both are treatable syndromes that may produce sudden unexpected death.^58,77,83,117 Slow waves on electroencephalogram (EEG) are a nonspecific abnormality and may be produced by migraine, breath holding, or syncope. Therefore, the EEG is more reliable in determining the type of epilepsy than in confirming the diagnosis of epilepsy itself.

Occasional seizures raise diagnostic issues. Seizures triggered by fever may range from simple febrile seizures to various kinds of epilepsy that range from benign to catastrophic, such as Dravet syndrome (see Chapter 230). Seizures triggered by photic stimulation, such as video games or television, require long-term treatment with AEDs only if the triggering features cannot be avoided.

If the evaluation leads to the conclusion that the seizure was caused by a transient or treatable disorder, AED treatment is not necessary. If the seizure is determined to be of cerebral origin, and no triggering factor is identified, the next step is to determine the risk for and possible consequences of a second seizure. These may differ considerably between individuals. For example, in some children who have benign partial epilepsy with central-midtemporal spikes, even the small risks associated with treatment may not be warranted, because most patients have only a few seizures before permanent remission occurs, and they cause no danger to the brain⁴⁰ (see Chapter 236). On the other hand, for adults working in some occupations, a second seizure could result in harmful consequences. Thus, a specific type of epilepsy should be determined based on seizure type, etiology, family history, and social context, but this may not always be possible when the patient is first seen. An EEG may need to be repeated once or several times, especially during infancy and childhood, and include recordings during sleep.

A number of structural lesions are associated with recurrent seizures, including tumors, cortical dysplasia, and vascular malformations. When these are diagnosed after a single seizure, drug treatment should be started without hesitation.

The recurrence risk for absence and myoclonic seizures is clearly higher than for other types of seizures, and patients referred for treatment usually have already had repeated seizures.

Another step in evaluating the need for treatment is to determine if a seizure was provoked or unprovoked. If a clear provoking factor, such as sleep deprivation or drug abuse, can be identified, the patient should be counseled to avoid the
precipitating stressors. However, a much more common situation is that in which the initial evaluation fails to reveal a specific cause or provoking factor. The risk for subsequent seizures must then be carefully evaluated. Data from clinical studies^{11,16,17,51,54,56} in the last decade have identified specific risk factors obtained from a complete evaluation. These factors can help identify persons who are at higher risk for additional seizures.

In children, the risk for seizure recurrence depends on the type of seizure, EEG characteristics, and cause of the seizure. When a first seizure has a remote etiology, the recurrence risk has been estimated to reach 65%. The presence of EEG abnormalities (44%) significantly increases the risk of recurrence.¹⁰⁶

In the absence of a definable etiology, and with normal interictal EEG findings, the risk of recurrence following a first convulsive seizure ranges from 30% to 60% during 3 years; most seizures recur in the year following the first seizure.^11,56 The presence of EEG epileptiform activity increases the risk of recurrence to 60%⁸⁶ this rate is no different if spikes are generalized or focal (58% vs. 61%). Ambrosetto et al.³ and Hamada et al.⁴⁹ found that a first seizure with rolandic spikes had a recurrence risk of 85% in the first year, but 71% of the patients had rare seizures (six or fewer). Seizures are usually simple partial. The clinical severity of a seizure usually has no prognostic value in children. For example, a single prolonged seizure with Todd paralysis and vomiting is often the only expression of benign occipital epilepsy, whereas so-called minor seizures often are seen in the most serious childhood epilepsies. The use of EEG in determining recurrence risk for convulsive seizures is valid for children and adolescents with generalized tonic–clonic seizures, but not for infants. Indeed, in patients with severe epilepsy beginning in the first year of life, interictal spikes may be lacking during the first 2 or 3 years of the disease.²⁵ On the other hand, 3% of school-age children without any history of seizures or cognitive troubles exhibit spike-waves.¹⁸ These spikes tend to resolve spontaneously, and these children do not develop epilepsy. The incidence of spikes without epilepsy is even higher in children who suffer various types of cognitive troubles, including developmental dysphasia,⁸⁵ mental retardation,⁹ and autism.⁵³ For all these conditions, any nonepileptic paroxysmal event that requires an EEG could therefore be misleading.

In one study of adults and some teenagers,⁵⁶ the recurrence risk of all patients who had no previous history of an insult to the central nervous system (CNS) (idiopathic single seizure) was 10% at 1 year, 24% at 2 years, and 29% at 5 years. This group was evaluated for further risk factors. Family history was important: Among those who had a sibling with seizures, the recurrence rate was 29% at 1 year and 46% at 5 years, compared with 9% at 1 year and 26% at 5 years for those without a sibling having epilepsy. A history of seizures in parents or first-degree relatives in this study was not associated with a risk above that of others with idiopathic single seizures. Another important predictive factor was the EEG. Patients with a pattern of generalized spike-and-wave discharges had a recurrence risk of 15% at 1 year and 58% at 5 years, compared with 9% at 1 year and 26% at 5 years for those with a normal or nonspecific EEG pattern. Another important factor was the history of a previous acute symptomatic seizure; that is, a previous seizure in the context of an illness or a childhood febrile seizure. These persons had a risk of 10% at 1 year and 39% at 5 years. Age, sex, seizure type, abnormal EEG pattern other than
generalized spike-and-wave, and abnormal neurologic examination did not elevate the risk above others in the idiopathic group.⁵⁶ Another group in this study was categorized as remote symptomatic. These were patients with a history of head trauma, stroke, CNS infection, or static encephalopathy from birth, with mental retardation or cerebral palsy. These patients had a risk for a second seizure of 26% at 1 year, 41% at 2 years, and 48% at 5 years—markedly higher than the corresponding rates of 10%, 24%, and 29% in the idiopathic group. Within the remote symptomatic group, increased risk for recurrence was noted in those with Todd postictal paresis: 41% at 1 year and 75% at 5 years. If a previous symptomatic acute seizure occurred, the risk for a second seizure was 60% at 1 year and 80% at 5 years. If status epilepticus (SE) or multiple seizures occurred at onset, the recurrence rate was 37% at 1 year and 56% at 3 years.⁵⁶ Thus, persons with a previous history of CNS insult, a sibling with seizures, prior acute seizure, or an EEG pattern of generalized spike-and-wave discharge have a significantly higher risk of a second seizure than do patients without those risk factors. Most neurologists would treat these patients with AEDs.

A more difficult decision arises in the case of patients with none of the additional risk factors, for whom the probability of a second seizure is less than 10% in the first year and approximately 20% by the end of 2 years after the single seizure.⁵⁶ Is this rate high enough to warrant the risks of treatment? There is no single answer to this question. The decision regarding whether to treat must be based on an evaluation by the patient and physician of the perceived risks and benefits. The risk of treatment with presently available AEDs is generally low. The risk of a second seizure depends on the lifestyle of the patient. In the case of those who need to drive, or for whom a second seizure may pose a significant risk of injury or loss of esteem, treatment may be indicated. There is a great need to emphasize patient choice and involvement in decision making here. The risk for recurrence is greatest in the first 2 years; thus, if treatment is initiated, it probably can be discontinued after the period of highest risk has passed.

Some brain lesions are clearly epileptogenic and occasionally produce severe epilepsy. However, it is often difficult to evaluate whether AED treatment has the ability to prevent epilepsy—particularly severe epilepsy—or the associated motor or cognitive deterioration.⁶³ Risk factors for epilepsy have been disclosed for both adults and children. In children, the risk for epilepsy following head trauma is increased when a focal cerebral lesion is evident, as indicated by focal neurologic signs, depressed skull fracture, or cerebral contusion on computed tomographic (CT) or magnetic resonance imaging (MRI) scan, and when seizures occur in the initial stage for more than 24 hours⁶⁰ (see Chapter 255).

The risk for epilepsy varies in different types of neurocutaneous syndromes. In Sturge-Weber syndrome, the first seizure often consists of SE in the first year of life, related to pial angioma, when the cutaneous angioma involves the first root of the trigeminal nerve.³⁷ Some authors advise prophylactic treatment from the neonatal period to prevent SE.¹¹⁶ The question of prophylactic treatment arises with the detection of tuberous sclerosis in utero. The highest incidence of onset of epilepsy is in the middle of the first year of life. To date, no data are available to make any scientifically based decision.

Sudden unexpected death may occur in various situations (see also Chapter 189). Some data suggest that the risk is significant only for remote symptomatic epilepsy.⁵⁵ In a community-based study of mortality in children with epilepsy, Harvey et al.⁵² found that mortality was not increased in idiopathic epilepsy compared with the general population. However, they found that, in symptomatic epilepsy, the risk of death was 50-fold higher. Gaitatzis et al.⁴⁰ concluded that life expectancy was reduced by up to 10 years. Other studies^69,79 suggest that “subtherapeutic” drug levels, mainly in adolescents with rare seizures and poor compliance, may precipitate cardiac arrhythmia or pulmonary edema. Therefore, the risk for death should not be a factor in the decision to start therapy, although the need for good compliance should be emphasized if treatment is initiated.

The risk for suicide is also higher in patients with epilepsy than in the general population, and some AEDs (i.e., phenobarbital) seem to contribute to this increase.¹³

Falling during a seizure may produce various physical injuries, and patients may drown when alone, either in the bath or when swimming. In adults, the risk depends on the patient’s usual activities and seizure type (e.g., if there is an aura, the patient may be able to anticipate and alleviate danger by shutting off the gas flame or water tap, or stopping the car). A survey of a nonselected series of 1,000 adults with epilepsy disclosed five severe head injuries, 22 fractures, and seven burns.⁴⁸ In a group of 138 children with epilepsy followed over a 10-year period, six patients sustained physical injuries caused by seizures, none of which produced sequelae.¹²² Four of these injuries were the first manifestation of epilepsy.

The incidence of vehicular accidents resulting from epileptic seizures is difficult to estimate. In one population study,⁵⁰ the rate slightly exceeded that of the general population. In another series,⁴² one-third of patients with epilepsy had seizures when driving, which produced accidents in half the patients. The type of seizure influences the risk for accident; a third of accidents involve seizures with an aura, and two-thirds of accidents involve seizures that lack an aura. Driving regulations vary from country to country.⁶⁵ Although a first seizure does not indicate epilepsy, in countries where efficient public
transportation is not available, the risk for a second seizure may be considered sufficient reason to start treatment.

Brain damage and sequelae from very brief seizures are very unusual, and prospective studies³⁴ have failed to demonstrate any brain damage from short seizures. Damage following SE seems to result mainly from the underlying cause of status.⁷⁵ The only exception is prolonged convulsive seizures with fever in infancy, because there is a significant risk of brain damage affecting the mesial temporal structures or the whole hemisphere.^2,101 The overall risk for development of SE in patients with newly diagnosed epilepsy is unknown, but it must be very small.

It is particularly difficult to evaluate the potential impact of drug treatment on the psychosocial consequences of recurrent seizures. The unpredictability of infrequent seizure recurrence can be extremely difficult for patients. Indeed, seizure control has been shown to be inversely related to self-esteem in adolescents.⁵⁷ For younger children, seizures that occur exclusively at home, primarily during sleep, usually have no psychosocial impact.