Many diseases can cause paroxysmal clinical events. The correct diagnosis of the paroxysmal event is necessary to provide correct treatment. If the event is an epileptic seizure, the seizure type and associated clinical, electroencephalographic (EEG) (see an example in the image below), and neuroimaging findings assist in determining the risk of seizure recurrence and the possible need to begin anticonvulsant therapy. Yet, the correct diagnosis is often missed.
Scheepers et al reported 49 patients with an incorrect diagnosis and 26 patients with an uncertain diagnosis among 214 patients with a diagnosis of epilepsy.[1] In addition, about 30% of patients seen at epilepsy centers for refractory seizures turn out to have been misdiagnosed and do not have seizures.[2]
View Image | An electroencephalogram (EEG) recording of a temporal lobe seizure. The ictal EEG pattern is shown in the rectangular areas. |
This article focuses on 2 related questions, as follows:
Thus, the stepwise approach should be the following:
This article describes the common clinical features of patients with a first seizure, risk factors for seizure recurrence, and a general approach to management.
The definitions of the following terms come from the International League Against Epilepsy (ILAE) Guidelines.[3, 4, 5] In 2017, the ILAE proposed revisions to the terminology, classification, and concepts for seizures and epilepsies, but the ILAE indicates these are guiding principles rather than a presentation of new classification per se.[6, 7]
An epileptic seizure is a clinical event presumed to result from an abnormal and excessive neuronal discharge. The clinical symptoms are paroxysmal and may include impaired consciousness and motor, sensory, autonomic, or psychic events perceived by the subject or an observer.
Epilepsy occurs when 2 or more epileptic seizures occur unprovoked by any immediately identifiable cause. The seizures must occur more than 24 hours apart. In epidemiologic studies, an episode of status epilepticus is considered a single seizure. Febrile seizures and neonatal seizures are excluded from this category.
Idiopathic epilepsy describes epilepsy syndromes with specific age-related onset, specific clinical and electrographic characteristics, and a presumed genetic mechanism.
Epileptic seizures are classified as cryptogenic or symptomatic. A cryptogenic seizure is a seizure of unknown etiology, and it is not associated with a previous central nervous system (CNS) insult known to increase the risk of developing epilepsy. A cryptogenic seizure does not conform to the criteria for the idiopathic or symptomatic categories. Previous studies use the term idiopathic to describe a seizure of unknown etiology. However, the ILAE guidelines discourage use of the term idiopathic to describe a seizure of unknown etiology.
Symptomatic seizure is a seizure caused by a previously known or suspected disorder of the CNS. This type of seizure is associated with a previous CNS insult known to increase the risk of developing epilepsy.
An acute symptomatic seizure is one that occurs following a recent acute disorder such as a metabolic insult, toxic insult, CNS infection, stroke, brain trauma, cerebral hemorrhage, medication toxicity, alcohol withdrawal, or drug withdrawal. An example of an acute symptomatic seizure is a seizure that occurs within 1 week of a stroke or head injury. Studies have reported that 25-30% of first seizures are acute symptomatic seizures.[8]
A remote symptomatic seizure is a seizure that occurs longer than 1 week following a disorder that is known to increase the risk of developing epilepsy. The seizure may occur a long time after the disorder. These disorders may produce static or progressive brain lesions. An example of a remote symptomatic seizure is a seizure that first occurs 6 months following a traumatic brain injury or stroke.
Seizures are also classified as provoked or unprovoked. A provoked seizure is an acute symptomatic seizure. An unprovoked seizure is a cryptogenic or a remote symptomatic seizure.
Compared with an epileptic seizure, a nonepileptic event is a clinical event that can mimic, and be mistaken for, an epileptic seizure. Examples of nonepileptic events that mimic seizures include syncope and psychogenic nonepileptic attacks (PNEAs). Syncope is caused by decreased cerebral perfusion that results mostly from a decrease in the cardiac output, which results in loss of consciousness.
For more information on some of the topics discussed here, see the following:
Causes of epilepsy include the following:
Risk factors for recurrent seizures include the following:
In patients with a first seizure and no known etiology, van Donselaar et al’s pooled analyses showed the following 2-year cumulative risks of seizure recurrence[18] : in patients with epileptiform discharges, 83%; in patients with nonepileptiform abnormalities, 41%; and in patients with normal electroencephalogram (EEG), 12%. The investigators obtained a routine EEG in all cases and a second sleep-deprived EEG if the first EEG did not show epileptiform discharges.
Studies in the 1990s also reported the following risks for seizure recurrence:
It is estimated that 1 in 26 people will develop epilepsy during his or her lifetime.[20]
The incidence of single unprovoked seizures is 23-61 cases per 100,000 persons-years, while the incidence of acute symptomatic seizures is 29-39 cases per 100,000 population per year.[21, 22, 23]
Beghi et al attributed the variability to differences in methodology and definitions.[19] The rates were similar in different geographic areas despite technical differences in the studies.
Racial differences have not been studied, but there appears to be a small to moderate male preponderance of men studies of first adult seizures in most reports.[14, 18, 24, 25] However, in an early study, Annegers et al found a slight overall preponderance of women.[10] Their etiologic categories were neurologic deficit from birth, remote symptomatic, and no known previous etiology. The investigators identified a preponderance of men in the group with neurologic deficit from birth, no sex preponderance in the group with remote symptomatic seizures, and a slight preponderance of women in the group with no known previous etiology.[10] These authors did not determine if these sexual differences were statistically significant.
Among patients who had an initial generalized tonic-clonic seizure, Bora et al found that only 45.5% were men.[15] Patients with partial seizures and structural lesions proven on computed tomography (CT) scan were excluded from this study.
Age does affect the incidence rate of epilepsy, with the highest incidence in the very young and very old groups. The incidence rate in children younger than 1 year is 100-233 per 100,000.[26] The rate decreases in patients aged 20-60 years to 30-40 cases per 100,000, but the rate increases to 100-170 cases per 100,000 in patients older than 65 years.[26]
A first seizure provoked by an acute brain insult is unlikely to recur (3-10%), whereas a first unprovoked seizure has a recurrence risk of 30-50% over the next 2 years. Even among symptomatic seizures, the recurrence rate differs according to the underlying cause. Seizures associated with reversible metabolic or toxic disturbances are associated with a minor risk of subsequent epilepsy (3% or less based on large case series). Seizures provoked by disorders that cause permanent damage to the brain, such as brain abscess, have a higher risk of recurrence (10% or more).[8]
In a 2009 study, individuals with a first acute symptomatic seizure were found to be significantly more likely to die in the first 30 days after the seizure compared with those with a first unprovoked seizure[27] ; however, the risk of 10-year mortality did not differ. In addition, the risk for subsequent unprovoked seizure was 80% lower in the group with first acute symptomatic seizure compared with those with first unprovoked seizure.[27]
Among medically untreated patients in one study, the cumulative 2-year risk of seizure recurrence was 51%.[9] Hauser et al found that variability in the reported risks of seizure recurrence may have been due to the following[11] :
See Etiology for risks factors associated with seizure recurrence.
Counseling patients about driving after a first seizure revolves around 2 issues: the diagnosis and the chance of recurrence.
Patients who have had a single epileptic seizure are at increased risk of having a second seizure. Therefore, these individuals should be informed that they are at increased risk of injury to themselves or others if another seizure occurs. Risk of injury is especially important if patients are driving, operating dangerous machinery, or performing other activities that could put themselves or others at risk. These same concerns also apply to nonepileptic conditions such as syncope that might recur and put the patient or others at risk of injury. Document this discussion in the patient’s medical record.
The patient should be advised to contact the state agency that regulates driving privileges, as driving regulations vary from state to state. The restrictions sometimes apply to any alteration or loss of consciousness from any etiology. This discussion with the patient should also be documented in the medical record.
Patients with a first epileptic seizure and with risk factors such as remote symptomatic etiology or an electroencephalogram (EEG) with epileptiform discharges are at higher risk for a second seizure. Restrictions of hazardous activity should be more emphatic for these patients.
For patient education information, see the Brain and Nervous System Center, as well as Epilepsy.
History remains the key in obtaining a correct diagnosis in patients with first seizure in adulthood. The detailed description of the actual episode in question is particularly important. The description should be obtained separately from the patient and from a caregiver who has witnessed the event.
The patient may be able to report a warning or aura and the feeling after the seizure. The presence of an aura, by definition, makes the diagnosis of a localization-related epilepsy, because auras are “simple partial” seizures with subjective symptoms. However, not every warning symptom is an aura.
Generally speaking, in order to be considered auras, the symptoms should be brief (seconds) and followed, at least some of the time, by more definite seizure. Auras widely vary but tend to be stereotyped in a given patient. Some (eg, déjà-vu, fear, epigastric sensation, lateralized somatosensory or visual phenomena) are very specific and even localizing; others are not (eg, indescribable sensation, whole body sensations, other vague symptoms like dizziness). The patient may not be able to describe the symptoms during the seizure, which speaks to loss of awareness, but says that the “next thing I know is coming to.”
The caregivers or witnesses should then describe what they observe; having the caregivers mimic the types of movements or behaviors they see during the attacks may be helpful. Occasionally, the best witnesses are not present; this may require a telephone call.
The following information should be obtained in the history:
In syncope, several historical features can be helpful. When an accurate description is missing (eg, unwitnessed event), the distinction between syncope and seizures can be difficult, because it is based on history alone; however, several symptoms are helpful in aiding the diagnosis.[27, 29] These include the circumstances of the attacks, because the most common mechanism for syncope (vasovagal response) is typically triggered by known precipitants (eg, pain such as inflicted by medical procedures, emotions, cough, micturition, hot environment, prolonged standing, exercise).
Other historical features that favor syncope include “presyncopal” prodromes (eg, vertigo, dizziness, lightheadedness, chest pain, nausea), as well as age and a history of cardiovascular disease. Historical features that favor seizures include tongue biting, head turning, posturing, urinary incontinence, cyanosis, prodromal déjà-vu, and postictal confusion.[27, 29] A point system using most of these features was designed, with a reported 94% sensitivity and specificity for the diagnosis of seizures.[29]
The neurologic examination should be directed at finding clinical evidence of a focal brain lesion, and a general physical examination should be performed to exclude a non-neurologic cause of the seizure.[30]
Pay attention to the presence of signs traumatic injuries to any part of the body, especially a tongue bite, which is highly specific in epileptic seizures.[31, 32]
Metabolic screening for uremia, hypoglycemia, drug intoxications, and electrolyte disorders should be conducted for patients with first seizure who present to the emergency department.[36] Perform other laboratory investigations as indicated for specific clinical situations.
Neuroimaging should be performed, because discovery of an epileptogenic lesion can have an impact on the diagnosis, prognosis, and treatment of new-onset seizures. Chadwick and Smith concluded that plausible arguments may be made for obtaining routine early computed tomography (CT) scanning and reserving magnetic resonance imaging (MRI) for patients with epilepsy whose seizures are not controlled by antiepileptic drugs.[37]
Magnetic resonance imaging (MRI) improves diagnostic accuracy. Using clinical and electroencephalographic (EEG) data alone, King et al were able to identify 23% of patients as having primary generalized epilepsy, 54% as having partial epilepsy, and 23% as having unclassified seizures.[28] Using clinical, EEG, and MRI data, the investigators were able to determine that 23% of patients had primary generalized epilepsy, 58% had partial epilepsy, and 19% had unclassified seizures.[28]
Computed tomography (CT) scanning may miss surgically remedial brain lesions that would otherwise be detected by MRI. King et al found that CT scanning detected only 12 of the 28 brain lesions that were detected by MRI; 7 of the missed lesions were brain tumors.[28]
Neuroimaging is unlikely to detect brain lesions in patients with clinical and EEG features of idiopathic generalized epilepsy or benign rolandic epilepsy. King et al found that MRI did not detect any brain lesions in 49 patients with clinical and EEG features of idiopathic generalized epilepsy or in 11 patients with benign rolandic epilepsy.[28]
Electroencephalography (EEG) should be performed within 24 hours of the seizure, because this study is significantly more sensitive when obtained during that period (see the following images).[25] If the routine EEG findings are normal, a sleep-deprived EEG should be performed. Standard EEG detects epileptiform discharges in 29% of patients; however, standard EEG combined with sleep-deprived EEG shows epileptiform discharges in 48% of patients.[18]
View Image | An electroencephalogram (EEG) recording of a temporal lobe seizure. The ictal EEG pattern is shown in the rectangular areas. |
View Image | An electroencephalogram (EEG) recording from a patient with primary generalized epilepsy. A burst of bilateral spike and wave discharge is shown in th.... |
View Image | An electroencephalogram (EEG) recording of a seizure from a subdural array in a patient evaluated for epilepsy surgery. The subdural electrodes record.... |
EEG significantly improves diagnostic accuracy in patients with a first seizure. Using clinical data alone, King et al were able to determine that 8% of patients had primary generalized epilepsy, 39% had partial epilepsy, and 53% had unclassified seizures.[25] When using clinical and EEG data together, the investigators were able to determine that 23% of patients had primary generalized epilepsy, 53% had partial epilepsy, and 23% had unclassified seizures.[25]
Schreiner and Pohlman-Eden studied the value of an EEG taken within 48 hours of the first seizure in an adult.[38] They found that 38% of patients without seizure recurrence had normal EEGs, while only 10.2% of patients with seizure recurrence had normal EEGs. Focal epileptiform activities were found significantly more frequently (26.5% vs 13%) in patients with seizure recurrence than in patients without seizure recurrence.
Unfortunately, although EEG can be helpful, it is often harmful, because normal EEGs are frequently overread as epileptiform, leading to the misdiagnosis of seizures.[33, 39] The tendency to overread normal EEGs is common and has numerous causes.[40] The most common reason for misdiagnosis is that the history is not suggestive of seizures, but the entire diagnosis is essentially based on the EEG.
Whenever cardiovascular causes are considered as the cause of a patient's spells but cannot be proven with conventional investigations, the use of the insertable loop recorder should be considered. Simpson et al described a case in which the placement of an insertable loop recorder, an important new tool in the diagnostic evaluation of patients with syncope, led to an unexpected diagnosis of a seizure.[41]
Many patients who have a single seizure do not require anticonvulsant therapy. The physician and patient or family should decide jointly whether to institute anticonvulsant therapy after a single seizure. This decision is based on a discussion of the risk of seizure recurrence, the effectiveness of anticonvulsant treatment, and the adverse medical and socioeconomic effects of anticonvulsant treatment.
Many patients who have a seizure recover spontaneously and fully with normal consciousness after a short time interval. Patients with incomplete recovery or a prolonged postictal state may require inpatient hospitalization.[30]
Inpatient management may be necessary if the clinical course is complicated by other medical problems requiring inpatient management. A short hospitalization may also be necessary for patients who are at risk of recurrent seizures and have no adequate supervision at home. Patients admitted from an emergency department had a 16.8% risk of an early recurrent seizure during their brief hospitalization.[36] This risk of early recurrent seizures was higher than reported in other studies.[10, 9, 11]
In 2015,the American Academy of Neurology (AAN) and the American Epilepsy Society (AES) released a new guideline on the prognosis and treatment of first unprovoked seizures.[42, 43]
According to the guideline, immediate antiepileptic drug (AED) therapy, as compared with delay of treatment pending a second seizure, is likely to reduce recurrence risk within the first 2 years but may not improve quality of life. Clinicians’ recommendations whether to initiate immediate AED treatment after a first seizure should be based on individualized assessments that weigh the risk of recurrence against the AEs of AED therapy and that consider educated patient preferences. Patients should be advised that risk of AED adverse events (AEs) may range from 7-31% and that these AEs are likely predominantly mild and reversible.[42, 43]
Immediate anticonvulsant treatment reduces the likelihood of a second seizure by half.[9] According to a report by Chandra, valproate treatment reduced seizure recurrence rates from 63% to 4.3%.[44] However, immediate anticonvulsant therapy does not affect the long-term prognosis for achieving 1- or 2-year seizure-free remission and exposes many patients who would never have a recurrent seizure to anticonvulsant side effects.[24]
The general consensus is that anticonvulsant treatment is needed after 2 seizures. The decision to provide anticonvulsant treatment after 1 seizure should be individualized. Two situations that are often encountered in clinical practice and should be distinguished are a first seizure and new-onset epilepsy with more than 1 unprovoked seizure. Berg and Shinnar emphasized the need to distinguish between these 2 entities in clinical studies.[12]
Seizure recurrence risk is substantially higher after 2 or more unprovoked seizures than after just 1.[11]
The chosen antiepileptic drug should have high efficacy, long-term safety, good tolerability, and low interaction potential, and the agent should allow a good quality of life, especially because half of all patients never have another seizure without treatment.[8] These agents prevent seizure recurrence and terminate clinical and electrical seizure activity.[45]
The accepted principle is that one should begin with monotherapy. All newer (second-generation) agents are acceptable choices and are likely just as effective as older agents. An American Academy of Neurology (AAN) evidence-based guideline recommended lamotrigine, oxcarbazepine, topiramate, and gabapentin as appropriate for initial monotherapy; however, this guideline did not include newer antiepileptic drugs, such as levetiracetam and pregabalin.
Selected Anticonvulsant Trial Results
In an early report, phenytoin, carbamazepine, valproate, and phenobarbital were equally effective in treating newly diagnosed epilepsy; however, phenobarbital had more adverse effects.[46] Other authors agree that barbiturates should be avoided because of neurotoxic and cognitive side effects.[8]
SANAD study results
The Standard and New Antiepileptic Drugs (SANAD) study reported that lamotrigine was significantly better in terms of time to treatment failure than the current standard treatment, carbamazepine, and the newer drugs gabapentin and topiramate for treatment of partial seizures.[47] (The study compared carbamazepine, gabapentin, lamotrigine, oxcarbazepine, and topiramate.) For time to 12-month remission from seizures, carbamazepine was not significantly advantageous compared with lamotrigine.[47] Lamotrigine also has the lowest incidence of treatment failure and has better outcome than all drugs except oxcarbazepine.
The same SANAD study compared valproate, lamotrigine, or topiramate for generalized and unclassifiable epilepsy seizures, and found that valproate is the drug of choice and is better tolerated than topiramate.[48]
Epilepsy monotherapy trial results
Another study compared 8 antiepileptic drugs used in 20 randomized trials and reported that in patients with partial seizures, the results favored carbamazepine, oxcarbazepine, and lamotrigine.[49] For generalized tonic-clonic seizures, the results favored valproate and phenytoin.[49]
NEAD study results
Regarding sodium valproate, the Neurodevelopmental Effects of Antiepileptic Drugs (NEAD) Study Group recommended that sodium valproate not be used as a first-line drug in the treatment of epilepsy for women of child-bearing age because of its side effects.[50]
Anticonvulsant Safety and Tolerability
Although phenytoin, carbamazepine, valproate, and phenobarbital were initially reported as equally effective in treating newly diagnosed epilepsy, phenobarbital had more adverse effects.[46] In general, barbiturates should be avoided because of neurotoxic and cognitive side effects.[8]
Overall, the newer antiepileptic drugs appear safer and better tolerated; however, they have not been used for as long or in as many patients as the older drugs. None of the common side effects of the older drugs (eg, gum hyperplasia, neuropathy) have been identified; however, ruling out potential new problems with long-term use of the newer antiepileptic drugs is difficult. Head-to-head studies have demonstrated favorable side effect profiles for gabapentin when compared with carbamazepine, and for oxcarbazepine and lamotrigine when compared with both phenytoin and carbamazepine.
All newer antiepileptic drugs, although "officially" approved as adjunctive therapy, are acceptable options for monotherapy (off-label use).
Privitera et al found topiramate 100 mg/d is as effective as therapeutic doses of carbamazepine and sodium valproate and has the fewest discontinuations due to adverse events.[51]
Clinical Context: Lamotrigine is a triazine derivative that inhibits the release of glutamate and voltage-sensitive sodium channels, leading to stabilization of the neuronal membrane. Lamotrigine is indicated for both adjunctive treatment and monotherapy for epilepsy.
Clinical Context: The pharmacological activity of oxcarbazepine is primarily by the 10-monohydroxy metabolite (MHD) of oxcarbazepine. It may block voltage-sensitive sodium channels, inhibit repetitive neuronal firing, and impair synaptic impulse propagation. This drug's anticonvulsant effect may also occur by affecting potassium conductance and high-voltage activated calcium channels.
Clinical Context: Carbamazepine has anticonvulsant actions that may involve depressing activity in the nucleus ventralis anterior of the thalamus, resulting in a reduction of polysynaptic responses and blocking posttetanic potentiation. It reduces sustained high-frequency repetitive neural firing. It is indicated for partial seizures, generalized tonic-clonic seizures, and mixed seizures.
Clinical Context: Topiramate is a sulfamate-substituted monosaccharide with a broad spectrum of antiepileptic activity that may have a state-dependent sodium channel blocking action, potentiation of the inhibitory activity of the neurotransmitter gamma-aminobutyrate (GABA), and, possibly, blocking of glutamate activity. Topiramate is indicated for both adjunctive treatment and monotherapy for epilepsy.
Clinical Context: Levetiracetam may inhibit presynaptic calcium channels by binding to a synaptic vesicle glycoprotein, SV2A. Levetiracetam is rapidly and almost completely absorbed after oral intake, with peak plasma concentrations approximately 1 hour after oral administration. It is predominantly excreted unchanged through the kidneys, with only about 27% metabolized.
Clinical Context: Valproic acid is chemically unrelated to other drugs used to treat seizure disorders. Although its mechanism of action is not established, the activity of valproic acid may be related to increased brain levels of GABA or enhanced GABA action. This agent may also potentiate postsynaptic GABA responses, affect potassium channels, or have a direct membrane-stabilizing effect.
Clinical Context: Phenytoin may act in the motor cortex, where it may inhibit the spread of seizure activity. The activity of brain stem centers responsible for the tonic phase of grand mal seizures may also be inhibited. The dose administered should be individualized.
Anticonvulsant agents, including lamotrigine, levetiracetam, oxcarbazepine, topiramate, and valproic acid, are commonly used for the treatment of seizures. Initial treatment includes monotherapy. Newer agents are acceptable choices and are likely just as effective as older agents.
An electroencephalogram (EEG) recording of a seizure from a subdural array in a patient evaluated for epilepsy surgery. The subdural electrodes record from the left anterior temporal (LAT), left middle temporal (LMT), and left posterior temporal (LPT) regions. The EEG seizure pattern is seen best in bipolar EEG channels LAT 3-4 and LMT 3-4 (rectangular areas).
An electroencephalogram (EEG) recording of a seizure from a subdural array in a patient evaluated for epilepsy surgery. The subdural electrodes record from the left anterior temporal (LAT), left middle temporal (LMT), and left posterior temporal (LPT) regions. The EEG seizure pattern is seen best in bipolar EEG channels LAT 3-4 and LMT 3-4 (rectangular areas).