cri

tal, 2

pilet coars.copein oa, pondacks

The wrong diagnosis of epilepsy is unfortunately common. Ofo are smisdiaacrossic attacat reh syncce settmisdiae.g., men givh expl

PNEAs constitute by far the most common (>90%) conditionmisdiagnosed as epilepsy, at least at referral epilepsy centers[9,10]. They are probably common in the general population, withan estimated prevalence of 2 to 33 per 100,000 [11]. The majorityof patients with PNEAs are young women, but no group is sparedand PNEAs also affect men and elderly patients relatively often

pic has been variable and continues to be confusing. Various terms

received AEDs for some time before the correct diagnosis ismade [14].

s A very high frequency of seizures (multiple daily episodes) thatis completely unaffected by AEDs.

s Specic triggers that are unusual for epilepsy (e.g., stress, gettingupset, pain, certain movements, sounds), especially if they arealleged to consistently trigger a seizure.

Epilepsy & Behavior 15 (2009) 1521

Contents lists availab

&

.e lE-mail address: sbenbadi@health.usf.edunear certainty, the delay in diagnosis remains long at about 7 to10 years [7,8]. This suggests that neurologists may not have a highenough index of suspicion to question the diagnosis of seizureswhen drugs fail. This article reviews the main conditions that canmimic and be misdiagnosed as epilepsy.

2. Psychogenic nonepileptic attacks

suspicion that seizures may be psychogenic rather than epileptic.

2.2. History

The most helpful historical features are the following:s Resistance to AEDs: This is usually the reason for referral to the

epilepsy center, and most (about 80%) patients with PNEAs haveto make a diagnosis of epilepsy (and its main mimic PNEA) with

tic delay and its consequences [46]. Amazingly, despite the ability PNEAs are initially suspected in the clinic on the basis of the his-

tory and examination. A number of red ags are useful in raising apatients diagnosed with epilepsy wh20% to 30% are found to have beencentage is astonishingly consistentcontinents. Psychogenic nonepileptthe most common condition foundand epilepsy monitoring units, thougmon in a general neurology practiconditions can also occasionally beis true of other chronic conditions (a wrong diagnosis of epilepsy has beated without being questioned, whic1525-5050/$ - see front matter 2009 Elsevier Inc. Adoi:10.1016/j.yebeh.2009.02.024een at epilepsy centers,gnosed [13]. This per-centers, countries, andks (PNEAs) are by farferral epilepsy centersope may be more com-ing. Other paroxysmalgnosed as epilepsy. Asultiple sclerosis), whenen, it is easily perpetu-ains the usual diagnos-

have been used, including pseudoseizures, nonepileptic seizures,and psychogenic seizures. Strictly speaking, terms such as pseudo-seizures and nonepileptic seizures include both psychogenic andnonpsychogenic (i.e., organic) epilepsy mimics. On the other hand,a term such as psychogenic nonepileptic seizures should be preferredbecause it adds the important connotation of a psychological ori-gin. Lastly, the word seizures is confusing to patients, and for thosereasons, psychogenic nonepileptic attacks will be used here.

2.1. Suspecting the diagnosis1. Introduction [12,13]. In addition to being common, PNEAs represent a challengein both diagnosis and in management. The terminology on the to-Review

The differential diagnosis of epilepsy: A

S. BenbadisComprehensive Epilepsy Program, University of South Florida and Tampa General Hospi

a r t i c l e i n f o

Article history:Received 18 February 2009Accepted 19 February 2009Available online 21 February 2009

Keywords:SeizuresNonepileptic seizuresDifferential diagnosisParoxysmal neurologic symptoms

a b s t r a c t

The wrong diagnosis of eattacks are by far the mosaverage delay of 710 yenon-epileptic attacks. Synis probably more commonlepsy include hypoglycemiTIAs, migraines, and TGA. Cing spells, and shudder attmisdiagnosis of epilepsy.

Epilepsy

journal homepage: wwwll rights reserved.tical review

Tampa General Circle, Tampa, FL 33606

psy is common. At referral epilepsy centers, psychogenic non-epilepticmmon condition found to have been misdiagnosed as epilepsy, with anThere are many red ags that can raise the suspicion of psychogenicis the second most common condition misdiagnosed as epilepsy, and itutpatient populations. Other conditions more rarely misdiagnosed as epi-anic attacks, paroxysmal movement disorders, paroxysmal sleep disorders,itions specic to children include nonepileptic staring spells, breath-hold-. At all ages, the over-interpretation of EEGs plays an important part in the

2009 Elsevier Inc. All rights reserved.le at ScienceDirect

Behavior

sevier .com/locate /yebeh

toms. The difference between these two conditions is that inmalingering, the reason for doing so is tangible and rationally

Behs The circumstances in which attacks occur: PNEAs tend to occurin the presence of an audience, and occurrence in the physiciansofce or waiting room is particularly suggestive of PNEAs [15].Similarly, PNEAs tend not to occur in sleep, although they mayseem to and be reported as doing so [16].

s The presence of fashionable diagnoses, such as bromyalgiaand chronic pain [15] (and probably others such as chronicfatigue syndrome and Lyme disease). Similarly, a orid reviewof systems, especially written lengthy lists of symptoms ordiagnoses suggesting somatization [17], should raise thesuspicion.

s The psychosocial history, including associated psychiatricdiagnoses.

2.3. Detailed description of the spells

This often includes characteristics that are inconsistent withepileptic seizures. In particular, some characteristics of the motor(convulsive) phenomena are associated with PNEAs, but this is bet-ter assessed with video/EEGmonitoring (see below). However, wit-nesses accounts are rarely detailed enough to describe theseaccurately, and in fact, even seizures witnessed by physicians areoften wrongly diagnosed.

2.4. Examination

The examiner should pay particular attention to mental statusevaluation, general demeanor, affect, level of concern, overdrama-tization, and histrionic features. Specically, the examination mayuncover histrionic behaviors such as give-way weakness and tightroping. Performing the examination can in itself act as an inductionin suggestible patients, making an attack more likely to occur dur-ing the history taking or examination.

In contrast to the above, certain symptoms when present arguein favor of epileptic seizures. These include signicant postictalconfusion, incontinence, occurrence out of sleep, and, most impor-tant, signicant injury, although injuries may be reported by pa-tients with PNEAs. In particular, tongue biting is highly specicto generalized tonicclonic seizures [18] and, thus, is a very helpfulsign when present.

2.5. Conrming the diagnosis

2.5.1. EEGBecause of its low sensitivity, routine EEG is not very helpful in

making a diagnosis of PNEA. However, the presence of repeatednormal EEGs, especially in light of frequent attacks and resistanceto medications, certainly can be viewed as a mild red ag [19].Ambulatory EEGs can contribute to the diagnosis by recordingthe habitual episode and documenting the absence of EEG changes.However, because the lack of EEG changes needs to be interpretedin light of the clinical attack and of the difculties in conveying thisdiagnosis (see Section 2.6), it should always be conrmed by video/EEG monitoring. (The use of prolactin is often discussed in the con-text of suspecting PNEAs versus seizures [20], but in the age of vi-deo/EEG monitoring, it has little value unless resources are verylimited.)

2.5.2. Video/EEG monitoringThe gold standard for diagnosis [1,9,10,21], video/EEG moni-

toring is indicated in all patients who continue to have frequentseizures despite medications. In the hands of experienced epilep-

16 S. Benbadis / Epilepsy &tologists, the combined electroclinical analysis of both the clinicalsemiology of the ictus and the ictal EEG ndings allows a denitivediagnosis in nearly all cases. If an attack is recorded, the diagnosisis usually not difcult, and it is very rare that the simple questionof PNEA versus epilepsy cannot be answered.

The principle of video/EEG monitoring is to record an episodeand demonstrate that: (1) no change occurs in the EEG duringthe clinical event, and (2) the clinical attack is inconsistent withseizure types that can be unaccompanied by EEG changes. IctalEEG has limitations because it may be negative in some partial sei-zures [22]. Ictal EEG may also be uninterpretable if movementsgenerate excessive artifact [23]. Analysis of the ictal semiology(i.e., video) is as important as (and sometimes more importantthan) the ictal EEG, as it often shows behaviors that are obviouslynonorganic and incompatible with epileptic seizures or other or-ganic conditions, although this can be difcult to put into words.

Certain characteristics of the motor phenomena are stronglyassociated with PNEAs. These include very gradual onset or termi-nation; pseudosleep; discontinuous (stop and go), irregular, orasynchronous (out of phase) activity; side-to-side head move-ments; pelvic thrusting; opisthotonic posturing; stuttering; weep-ing; preserved awareness during bilateral motor activity; andpersistent eye closure [1,10,2426]. It should be pointed out thatalthough some of these behaviors are highly specic for PNEAs,none is in itself pathognomonic and establishes the diagnosis inisolation. False positives do exist, and for example, preservedawareness during bilateral motor activity can be seen in somefrontal lobe seizures [27]. Provocative techniques, activation pro-cedures, or inductions can be very useful for the diagnosis of PNE-As, particularly when the diagnosis remains uncertain and nospontaneous attacks occur during monitoring [21,2831]. Further,in difcult situations where the combination of semiology and EEGdoes not allow the conclusion that an episode is psychogenic in ori-gin (e.g., uninterpretable EEG due to movement-related artifacts orsymptoms consistent with a simple partial seizure), the very pres-ence of suggestibility (i.e., suggestion triggers the episode in ques-tion) is the strongest argument to support a psychogenicmechanism [32].

2.5.2.1. Short-term outpatient video/EEG monitoring with activa-tion. An extension of the use of inductions is that, when patientsare strongly suspected to have PNEAs on clinical grounds, theycan undergo outpatient video/EEG monitoring with activation. Thiscan be very cost effective while retaining the same specicity and areasonably high sensitivity, thus obviating the need for prolongedvideo/EEG monitoring [33,34].

2.6. Psychopathology

PNEAs are by denition a psychiatric disorder. According to theDiagnostic and Statistical Manual of Mental Disorders classication,physical symptoms caused by psychological causes can fall intothree categories: (1) somatoform disorders, (2) factitious disorders,and (3) malingering. Somatoform disorders involve the uncon-scious production of physical symptoms because of psychologicalfactors, which means that symptoms are not under voluntary con-trol; that is, the patient is not intentionally trying to deceive.Somatoform disorders are subdivided into several disordersdepending on the characteristics of the physical symptoms andtheir time course, and the two most relevant to PNEAs are conver-sion disorder and somatization disorder. In contrast to the uncon-scious (unintentional) production of symptoms of somatoformdisorders, factitious disorder and malingering imply that the pa-tient is purposely deceiving the physician, that is, faking the symp-

avior 15 (2009) 1521understandable, whereas in factitious disorder, the motivation is

Beha pathological need for the sick role. It is generally accepted thatmost patients with PNEAs fall into the somatoform category(unconscious production of symptoms), rather than the intentionalfaking type (malingering and factitious).

The literature often implies that PNEAs represent a unique dis-order, but in reality PNEAs are but one type of somatoform disorder(or malingering or factitious disorder). Only the expression of thepsychopathology is different. Fundamentally, the underlying psy-chopathology, the prognosis, and the management are no differentfor PNEAs than they are for other psychogenic symptoms. What-ever the manifestations, psychogenic symptoms represent a chal-lenge both in diagnosis and in management. Every medicalspecialty deals with symptoms that can be psychogenic [35].Among such symptoms, PNEAs are unique in one principal charac-teristic. With video/EEG monitoring, they can be diagnosed withnear certainty. This is in sharp contrast to other psychogenic symp-toms, which are almost always a diagnosis of exclusion.

2.7. Management

The role of the neurologist or epileptologist is to determinewhether organic disease is present. Once the attacks have beenshown to be psychogenic, the exact psychiatric diagnosis and itstreatment should be best handled by the mental health profession-als (psychiatrist, psychologist, counselor). However, implementa-tion of this in practice is plagued by many difculties [35]. Therole of the neurologist should not end when the diagnosis of PNEAsis made. In fact, arguably the most important step in initiatingtreatment is the delivery of the diagnosis to patients and families.Patients reactions can include disbelief, denial, and anger. Anotherobstacle is that psychiatrists tend to be skeptical about the diagno-sis of psychogenic symptoms, even for PNEAs where video/EEGmonitoring allows a near-certain diagnosis [35,36].

3. Syncope

Syncope is common, and although it is a distant second to PNEAsin terms of conditions misdiagnosed as epilepsy at referral epilepsycenters, itmay bemore common in general neurology practices. Therst reason that syncope ismisdiagnosedas seizures is theerroneousbelief that seizures can cause a accid motionless episode of loss ofconsciousness (LOC) for seconds to minutes. In reality, no seizuretype does this. Generalized tonicclonic seizures have obvious mo-tor manifestations, myoclonic seizures are very short jerks with nodetectable LOC, atonic seizures may cause abrupt falls but no pro-longed LOC, and complex partial seizures or absence seizures causealteration of awareness but not limpmotionless LOC. In general, epi-sodes of LOC with eyes closed for several seconds to minutes areeither psychogenic or syncopal, but not epileptic (nor transientischemic attacks!). The second reason for themisdiagnosis is the fre-quency with which syncopal events are convulsive. Although theconventional teaching is that syncopal episodes are limpmotionlessevents, they in fact frequently involve brief body jerks. In a study ofpatients with an implantable debrillator in whom syncope wasdeliberately induced, 45% of episodes included tonic or clonicmotoractivity [37]. In another study of patients diagnosed with epilepsywho underwent tilt-table testing, 63% of induced episodes of syn-copewere convulsive [38]. Lastly andmost impressively, when syn-cope was induced in healthy volunteers (using hyperventilation,squatting, andValsalvamaneuver), 38 [90%]of42episodeshadsomeclonic-like jerking activity [39]. Motor symptoms associated withsyncope are clonic- or myoclonic-like, tend to last only a few sec-onds, and terminate once the patient is horizontal, in sharp contrast

S. Benbadis / Epilepsy &to the typical generalized tonicclonic seizure duration of 30 to 90seconds. EEGs are very sensitive to decreased cerebral ow, and bythe time LOC occurs, EEG changes are present. When syncope (con-vulsive or not) is recorded on video/EEG monitoring, the EEG pro-ceeds through a very stereotyped pattern of changes (deltaslowing and suppression) [40,41]. When an accurate description ismissing (e.g., unwitnessed event), the distinction between syncopeand seizures can at times be difcult, as it is based on history alone.However, several symptoms are helpful in pointing one way or theother [42,43]. Among these are the circumstances of the attacks, asthe most common mechanism for syncope (vasovagal response) istypically triggered by clear precipitants (e.g., pain such as inictedby medical procedures, emotions, cough, micturition, hot environ-ment, prolonged standing, exercise). Other historical features thatfavor syncope include presyncopal prodromes (malaise, sweating,dizziness, lightheadedness, nausea, chest pain, palpitations), as wellasageandahistoryof cardiovasculardisease.Historical features thatfavor seizures include tongue biting, head turning, posturing, uri-nary incontinence, cyanosis, prodromaldj-vu, andpostictal confu-sion [42,43]. A point system using most of these features has beendesigned and reportedly has 94% sensitivity and specicity for thediagnosis of seizures [43].Management of syncope depends entirelyon its cause. Themajority of syncopal episodes are benign vasovagalepisodes, but theconcerningetiologies are cardiac related.Evenwithextensive evaluations, a large proportion of syncopal episodes re-main unexplained. Many patients with unexplained syncope (orpresyncope) probably have psychogenic pseudo-syncope, andwhenred ags are present (identical to those for PNEAs), video/EEGmon-itoring shouldbeperformedas it caneasily beused tomake thediag-nosis [40].

4. Other organic conditions

4.1. Hypoglycemia

Hypoglycemia rarely causes complete LOC. When it does, itresembles syncope and is also preceded by orid prodromes ofhunger, weakness, tremulousness, malaise, and abnormal behav-iors. Hypoglycemia typically occurs in reasonably obvious settings(e.g., diabetic patients, insulin or oral antihyperglycemics, fasting).

4.2. Panic attacks

Panic attacks are paroxysmal manifestations of anxiety or panicdisorder and may be mistaken for seizures [44]. The border be-tween panic attacks and PNEAs may be imprecise and some over-lap may exist [45]. Conversely, fear is a relatively common(psychic) aura in patients with mesiotemporal epilepsy. The iden-tication of fear as an epileptic aura is easy when it evolves into aclear seizure, but can be difcult in the absence of other seizuretypes [46]. Panic attacks include intense autonomic, especially car-diovascular and respiratory, symptoms. Abrupt and intense fear isaccompanied by at least four of the following symptoms: palpita-tions, diaphoresis, tremulousness or shaking, shortness of breathor sensation of choking, chest discomfort, nausea or abdominal dis-comfort, dizziness or lightheadedness, derealization or depersonal-ization, fear of losing control, fear of dying, paresthesias, and chillsor hot ashes. The symptoms typically peak within 10 minutes.Panic disorder often coexists with other manifestations of anxietysuch as agoraphobia and social phobia and also with depressivedisorders. Thus, the diagnosis is not usually difcult and rarely re-quires video/EEG recordings.

4.3. Paroxysmal movement disorders

4.3.1. Acute dystonic reactions

avior 15 (2009) 1521 17Acute dystonic reactions are caused by dopamine receptorblockers such as antipsychotics (including atypical ones) and antie-metics, although other dugs can be involved (e.g., carbamazepine,

sies of the LennoxGastaut type.

Behlithium, trazodone, illicit drugs). They typically occur within 1 to4 days of beginning the medication and are characterized by twist-ing movements affecting the cranial, pharyngeal, and cervical mus-cles. The oculogyric crisis is a dramatic subtype characterized byacute conjugate eye deviation, usually in an upward direction.The typical attack lasts 1 to 2 hours, during which the abnormalmovement occurs repetitively for seconds to minutes. These dys-tonic reactions respond very well and rapidly to anticholinergics(trihexyphenidyl, benztropine, diphenhydramine) and levodopa[47].

4.3.2. Hemifacial spasmHemifacial spasm (HFS) may supercially resemble a simple

partial seizure with motor symptoms of the face, or more simplya facial clonic seizure, but clear differences make the differentia-tion easy. HFS is a chronic progressive (rather than paroxysmal)disorder. While facial motor seizures typically involve the perioralarea (because of a large representation on the motor homunculus),the unilateral facial twitching of HFS typically affects the periorbi-tal muscles rst and then spreads to other (ipsilateral) facial mus-cles over a period of months to years. Over time or withexacerbations, the clonic movements can result in a sustained to-nic contraction causing forceful (unilateral) eyelid closure (blepha-rospasm) [48].

4.3.3. Nonepileptic myoclonusNonepileptic myoclonus is dened as myoclonus that is not of

cortical origin, that is, not visible on EEG. Hiccups and hypnic jerksare examples of normal nonepileptic myoclonus, but abnormalnonepileptic myoclonus can be seen in metabolic or toxic enceph-alopathies and neurodegenerative diseases.

4.4. Sleep disorders

4.4.1. ParasomniasParasomnias are the most likely sleep disorders to present a

diagnostic challenge because they are, by denition, short-livedparoxysmal behaviors that occur out of sleep. In particular, thenon-REM parasomnias (night terrors, sleepwalking, and confu-sional arousals) can supercially resemble seizures, as they includecomplex behaviors and some degree of unresponsiveness andamnesia for the event. The non-REM parasomnias are most com-mon between ages 4 and 12, and night terrors are particularlycommon. They are often familial and may be worsened by stress,sleep deprivation, and intercurrent illnesses. Similarly, rhythmicmovement disorder is a parasomnia typically seen at transitionor stage 1 sleep, and can also resemble partial seizures. One com-mon example is head banging (jactatio capitis). Among REM sleepparasomnias, nightmares rarely present a diagnostic challenge,but REM behavior disorder may with violent and injurious behav-iors during REM sleep. The diagnosis of REM behavior disorder isusually easy as it affects older men and the description of actingout a dream is quite typical.

Several historical features can help in differentiating parasom-nias from seizures [49], but occasionally video/EEG monitoringmay be necessary, provided that the episodes are frequent enough.Video/EEG monitoring will usually conrm the absence of an EEGseizure and usually shows that the behavior arises from a specicstage of sleep [50]. Occasionally, in the absence of ictal EEGchanges, the differentiation between seizure and parasomnia canbe difcult.

4.4.2. Cataplexy

18 S. Benbadis / Epilepsy &Cataplexy is part of the narcolepsy tetrad and consists of anabrupt loss of tone. As such, it could theoretically be mistakenfor atonic seizures or drop attacks, but there are several distin-4.4.3. Hypnic jerksHypnic jerks or sleep starts are benign myoclonic jerks that

everyone experiences on occasion. Although they resemble thejerks of myoclonic seizures, their occurrence only on falling asleepstamps them as benign nonepileptic phenomena. They occur at allages and can lead to evaluations for seizures, especially when thejerks are unusually violent. They are easily identied on video/EEG monitoring by the fact that they occur in wake to stage 1 tran-sition and have no EEG correlate associated with the jerks [53].

4.5. Transient ischemic attacks

Transient ischemic attacks (TIAs) rarely present a diagnosticchallenge, because symptoms of TIAs are typically negative, andsymptoms of seizures are typically positive. In addition, focalsymptoms in TIA are strokelike, that is, maximal at onset, whereasfocal seizure symptoms tend to march or evolve over seconds.Although rarely seizures can manifest with ictal negative symp-toms (e.g., aphasia, negative myoclonus), they usually also includepositive symptoms, which make the diagnosis easier. Similarly,limb-shaking TIAs occur but are very rare [54]. The confusion be-tween TIAs and seizures may be more likely when the seizure isunwitnessed and the patient appears with a focal decit (e.g., Toddparalysis or aphasia), especially because both will improve overtime (minutes). Usually, age and associated symptoms help differ-entiate the two. Contrary to a common misconception, TIAs excep-tionally (if ever) cause a LOC.

4.6. Migraines

Complicated migraines and migraine auras can cause positivefocal symptoms in all ve senses and, as such, may mimic focal(simple partial) seizures or epileptic auras. In addition, both mi-graine and seizure focal symptoms march. The key differentiat-ing factor, as usual, is the time course: migraine symptoms tendto evolve in minutes, whereas seizure symptoms evolve in seconds.Usually, associated symptoms (migrainous headache or more obvi-ous seizure symptoms) make the diagnosis easy. Basilar migrainecan also cause LOC.

4.7. Transient global amnesia

Transient global amnesia consists of dramatic episodes of anter-ograde amnesia. Patients are alert and otherwise cognitively intactbut cannot form new memories, and they ask repetitive questionsabout their environment. This lasts several hours and then re-solves. The cause is not known, but transient global amnesia isnot thought to represent a TIA or a seizure and usually does not re-cur [55].

5. Conditions and issues specic to young children:Misdiagnosis of epilepsy in children

The differential diagnosis of seizures is broader in children thanin adults [56,57], with many nonepileptic but nonpsychogenic con-guishing features: The most characteristic feature of cataplexy isthat it is typically triggered by emotions, most commonly laughter[51,52]. Severe daytime sleepiness characteristic of narcolepsy isalmost always present. Lastly, atonic seizures usually occur in acompletely different context of symptomatic generalized epilep-

avior 15 (2009) 1521ditions to be considered. Physiological and organic events predom-inate in infants and young children, and psychiatric disordersbecome more common in later childhood and adolescence. About

sonnel, but are neither epileptic nor psychogenic. Video/EEGrecordings will usually clarify the situation, but because many of

Beh50% have psychological disorders (90% being PNEAs), but otherpsychiatric diagnoses are also found (e.g., episodic dyscontrol withrage attacks, behavioral outbursts, panic/anxiety disorder, and fac-titious disorder by proxy). PNEAs occur in older children and ado-lescents and have the same characteristics as in adults. However,the gender difference of female predominance is not seen untiladolescence [52], so that PNEAs are as common in preadolescentboys as in girls. The other 50% have nonpsychogenic conditions,the most common of which is nonepileptic inattention with staringspells [58]. Other diagnoses include stereotyped mannerisms, hyp-nic jerks, parasomnias, tics, gastroesophageal reux with posturingor laryngospasm, arousals, shuddering attacks, and apneas[56,57,59,60].

Nonepileptic staring spells represent a common challenge. Chil-dren are occasionally inattentive, and the families report brief epi-sodes of staring and unresponsiveness with no motormanifestations. Several features can help distinguish absence sei-zures from benign nonepileptic staring spells in otherwise normalchildren [58]. Three features suggest nonepileptic events: (1) theevents do not interrupt play; (2) the events were rst noticed bya professional such as a schoolteacher, speech therapist, occupa-tional therapist, or physician (rather than by a parent); and (3)the child, while staring, is responsive to touch or interruptible byother external stimuli. Other features that have been found to sug-gest nonepileptic or behavioral rather than epileptic staring in-clude lower age and lower frequency [61]. By contrast, factorsthat suggest an epileptic etiology include twitches of the extremi-ties, urinary incontinence, and upward eye movement. Such be-nign nonepileptic staring spells are particularly likely to benoticed and reported by overvigilant parents in a child who hasor has had clear seizures.

Tics can supercially resemble simple partial seizures with mo-tor symptoms, but several features distinguish them [62]. Tics arenot episodic and tend to occur throughout the day, although theycan uctuate. They are sporadic rather than repetitive, stereotyped(the same movement repeats itself without evolving, and the samemuscle group is involved), and disappear in sleep. They are pre-ceded by an urge to move that is temporarily suppressible and fol-lowed by a sense of relief. Tics are particularly common betweenages 5 and 10.

Shuddering attacks are benign paroxysmal spells of older in-fants and young children, and can mimic several seizure types,including tonic, absence, and myoclonic. These episodes are usu-ally benign, have no association with increased morbidity or mor-tality, and tend to remit spontaneously. The condition is seen inolder infants and young children. Parents describe the paroxysmalepisodes as a sudden exion of the neck and trunk and adductionof the arms. A shiverlike movement of the trunk (like a chill) oc-curs, and the body may stiffen. Consciousness does not seem to bealtered. The episode usually lasts 5 to 15 seconds. Unlike epilepticseizures, shuddering attacks do not occur during sleep [63,64].

Breath-holding spells (cyanotic infantile syncope) affect chil-dren aged 6 months to 5 years. Typically, a clear trigger is present,with the child being upset and crying. At the end of expiration, thechild is unable to relax and inhale and becomes apneic and cya-notic. The child may appear angry and upset about this uncomfort-able feeling, loses consciousness, may have urinary incontinence,and becomes stiff or even opisthotonic. The EEG during the eventtypically shows high-amplitude slowing followed by suppression,as is seen in syncope of any cause. When the child relaxes andbreathes again, consciousness is gradually regained. These cyanoticbreath-holding spells could be easily confused with epilepticevents, but they are not primarily epileptic phenomena. They can

S. Benbadis / Epilepsy &evolve into epileptic seizures and even status epilepticus, but theinitial event is not epileptic. Anemia should be ruled out andmay require treatment. Cyanotic breath-holding spells are to bethe mild nonspecic symptoms mimic simple partial seizures orauras rather than more severe seizures, the mere presence of a nor-mal ictal EEG does not in itself exclude seizures. However, the vi-deo, that is, the characteristics of the movements, usually does,as they are nonclonic, nontonic, and not myoclonic [69]. At timesthe distinction can be difcult, and when in doubt it is preferableto be conservative rather than label the episodes as seizures.

6.1. The role of misread (overread) EEGs in the misdiagnosis of epilepsy

Many patients (about a third) who have been misdiagnosed ashaving epilepsy have had previous EEGs interpreted as epilepti-form that contributed to the misdiagnosis of epilepsy [3,70,71].In fact, sometimes patients are diagnosed with epilepsy and trea-ted based solely on an EEG, despite that fact that they have nosymptoms or that their symptoms are not at all suggestive of sei-zures. This has led many experts to point out that EEG can in factbe bad for you. There are many well-described normal variantsthat can be misread as epileptiform, but in reality the vast majorityof overread patterns are simple uctuations of sharply contouredbackground rhythms or fragmented alpha activity [7073]. Thedistinguished from pallid infantile syncope, which is associatedwith brief cardiac asystole and overlap with seizures [65].

Gastroesophageal reux associated with laryngospasm in in-fants may cause events that look like seizures, with limb posturing,eye deviation, and even opisthotonus (Sandifer syndrome). Theseevents tend to occur in sleep and in the postprandial period, andrecurrent vomiting is typically associated with them [66]. Thediagnosis can be made with an esophageal pH probe, and treatingthe reux usually resolves the problem.

Benign myoclonus of infancy [67] must be differentiated frominfantile spasms. This requires video/EEG monitoring as the eventsappear clinically similar. This benign phenomenon resolves spon-taneously within a year without neurologic sequelae.

Mannerisms are common in young children, in particular thosewith a mental handicap. Mannerisms can look odd and unnaturaland occasionally mimic motor seizures. Similarly, self-stimulatingbehaviors, including masturbation, can be erroneously interpretedas seizures.

Spasmus nutans is a benign triad of head nodding, head tilt, andpendular nystagmus, which typically occurs between 4 and12 months of age [68].

6. Benign nonspecic symptoms misinterpreted as seizures

This phenomenon has no name and is not written about be-cause it does not t under psychogenic seizures or other organicconditions described above. It is best described as overvigilanceand is commonly seen at epilepsy centers. It basically consists ofthe overinterpretation of benign or nonspecic symptoms as sei-zures. Unexplained symptoms are common in everyday life and in-clude transient dizziness, limb numbness, head sensations, andvarious mild and brief involuntary movements. The misinterpreta-tion of these symptoms as seizures is more likely to occur in anx-ious patients (or caregivers) with hypochondriacal tendencies. It isalso more common in patients who also have or have had seizuresor who have other organic conditions. Another setting is the inten-sive care unit, where many patients who are very ill can have non-specic abnormal movements such as shivers, twitches, andtremors that are of concern to caregivers or intensive care unit per-

avior 15 (2009) 1521 19reasons for the overinterpretation of EEGs are complex, and havebeen discussed elsewhere [71,72], but the fact that the diagnosisof seizures should be clinical cannot be overemphasized. In chil-

Behdren, benign centrotemporal (rolandic) spikes on EEG are a com-mon red herring because they occur frequently in asymptomaticchildren. Most errors in diagnosis are made because the EEG isoverread as abnormal and is interpreted outside of the clinical con-text [71,72].

References

[1] Benbadis SR. Differential diagnosis of epilepsy. Continuum Lifelong LearnNeurol 2007;13:4870.

[2] Chadwick D, Smith D. The misdiagnosis of epilepsy. Br Med J 2002;324:4956.[3] Smith D, Defalla BA, Chadwick DW. The misdiagnosis of epilepsy and the

management of refractory epilepsy in a specialist clinic. Q J Med1999;92:1523.

[4] LaFrance Jr WC, Benbadis SR. Avoiding the costs of unrecognized psychologicalnonepileptic seizures. Neurology 2006;66:16201.

[5] Martin RC, Gilliam FG, Kilgore M, et al. Improved health care resourceutilization following videoEEG-conrmed diagnosis of nonepilepticpsychogenic seizures. Seizure 1998;7:38590.

[6] Nowack WJ. Epilepsy: a costly misdiagnosis. Clin Electroencephalogr1997;28:2258.

[7] Carton S, Thompson PJ, Duncan JS. Non-epileptic seizures: patientsunderstanding and reaction to the diagnosis and impact on outcome. Seizure2003;12:28794.

[8] Reuber M, Fernandez G, Bauer J, et al. Diagnostic delay in psychogenicnonepileptic seizures. Neurology 2002;58:4935.

[9] Benbadis SR, Heriaud L, ONeill E, et al. Outcome of prolonged EEGvideomonitoring at a typical referral epilepsy center. Epilepsia 2004;45:11503.

[10] Benbadis SR. Psychogenic non-epileptic seizures. In: Wyllie E, editor. Thetreatment of epilepsy: principles and practice. Philadelphia: LippincottWilliams & Wilkins; 2006. p. 62330.

[11] Benbadis SR, Hauser WA. An estimate of the prevalence of psychogenicnonepileptic seizures. Seizure 2000;9:2801.

[12] Behrouz R, Heriaud L, Benbadis SR. Late-onset psychogenic nonepilepticseizures. Epilepsy Behav 2006;8:64950.

[13] Duncan R, Oto M, Martin E, Pelosi A. Late onset psychogenic nonepilepticattacks. Neurology 2006;66:16447.

[14] Benbadis SR. How many patients with pseudoseizures receive antiepilepticdrugs prior to diagnosis? Eur Neurol 1999;41:1145.

[15] Benbadis SR. A spell in the epilepsy clinic and a history of chronic pain orbromyalgia independently predict a diagnosis of psychogenic seizures.Epilepsy Behav 2005;6:2645.

[16] Benbadis SR, Lancman ME, King LM, Swanson SJ. Preictal pseudosleep: a newnding in psychogenic seizures. Neurology 1996;47:637.

[17] Benbadis SR. Hypergraphia and the diagnosis of psychogenic attacks.Neurology 2006;67:904.

[18] Benbadis SR, Wolgamuth BR, Goren H, et al. Value of tongue biting in thediagnosis of seizures. Arch Intern Med 1995;155:23469.

[19] Davis BJ. Predicting nonepileptic seizures utilizing seizure frequency, EEG, andresponse to medication. Eur Neurol 2004;51:1536.

[20] Chen DK, So YT, Fisher RS. For the Therapeutics and Technology AssessmentSubcommittee of the American Academy of Neurology. Use of serum prolactinin diagnosing epileptic seizures: report of the Therapeutics and TechnologyAssessment Subcommittee of the American Academy of Neurology. Neurology2005;65:66875.

[21] Benbadis SR. Clinical features and role of EEG-video monitoring. In: SchachterS, LaFrance WC, editors. Gates and Rowans nonepileptic seizures. 3rd ed.London/New York: Cambridge University Press; in press.

[22] Devinsky O, Sato S, Kufta CV, et al. Electroencephalographic studies of simplepartial seizures with subdural electrode recordings. Neurology1989;39:52733.

[23] Benbadis SR. The EEG in nonepileptic seizures. J Clin Neurophysiol2006;23:34052.

[24] Bergen D, Ristanovic R. Weeping as a common element of pseudoseizures.Arch Neurol 1993;50:105960.

[25] Chung SS, Gerber P, Kirlin KA. Ictal eye closure is a reliable indicator forpsychogenic nonepileptic seizures. Neurology 2006;66:17301.

[26] Vossler DG, Haltiner AM, Schepp SK, et al. Ictal stuttering: a sign suggestive ofpsychogenic nonepileptic seizures. Neurology 2004;63:5169.

[27] Kanner AM, Morris HH, Luders H, et al. Supplementary motor seizuresmimicking pseudoseizures: some clinical differences. Neurology1990;40:14047.

[28] Benbadis SR. Provocative techniques should be used for the diagnosis ofpsychogenic nonepileptic seizures. Arch Neurol 2001;58:20635.

[29] Benbadis SR, Johnson K, Anthony K, et al. Induction of psychogenicnonepileptic seizures without placebo. Neurology 2000;55:19045.

[30] Schachter SC, Brown F, Rowan AJ. Provocative testing for nonepileptic seizures:attitudes and practices in the United States among American Epilepsy Societymembers. J Epilepsy 1996;9:24952.

[31] Cohen LM, Howard 3rd GF, Bongar B. Provocation of pseudoseizures bypsychiatric interview during EEG and video monitoring. Int J Psychiatry Med

20 S. Benbadis / Epilepsy &1992;22:13140.[32] Hinson VK, Haren WB. Psychogenic movement disorders. Lancet Neurol

2006;5:695700.[33] Benbadis SR, Siegrist K, Tatum WO, et al. Short-term outpatient EEGvideowith induction in the diagnosis of psychogenic seizures. Neurology2004;63:172830.

[34] McGonigal A, Oto M, Russell AJC, Greene J, Duncan R. Outpatient videoEEGrecording in the diagnosis of non-epileptic seizures: a randomised controlledtrial of simple suggestion techniques. J Neurol Neurosurg Psychiatry2002;72:54951.

[35] Benbadis SR. The problem of psychogenic symptoms: is the psychiatriccommunity in denial? Epilepsy Behav 2005;6:914.

[36] Harden CL, Burgut FT, Kanner AM. The diagnostic signicance of video-EEGmonitoring ndings on pseudoseizure patients differs between neurologistsand psychiatrists. Epilepsia 2003;44:4536.

[37] Aminoff MJ, Scheinman MM, Grifn JC, Herre JM. Electrocerebralaccompaniments of syncope associated with malignant ventriculararrhythmias. Ann Intern Med 1988;108:7916.

[38] Zaidi A, Clough P, Cooper P, et al. Misdiagnosis of epilepsy: many seizure-likeattacks have a cardiovascular cause. J Am Coll Cardiol 2000;36:1814.

[39] Lempert T, Bauer M, Schmidt D. Syncope: a videometric analysis of 56 episodesof transient cerebral hypoxia. Ann Neurol 1994;36:2337.

[40] Benbadis SR, Chichkova R. Psychogenic pseudosyncope: an underestimatedand provable diagnosis. Epilepsy Behav 2006;9:10610.

[41] Sheldon RS, Koshman ML, Murphy WF. Electroencephalographic ndingsduring presyncope and syncope induced by tilt table testing. Can J Cardiol1998;14:8116.

[42] Hoefnagels WA, Padberg GW, Overweg J, et al. Transient loss of consciousness:the value of the history for distinguishing seizure from syncope. J Neurol1991;238:3943.

[43] Sheldon R, Rose S, Ritchie D, et al. Historical criteria that distinguish syncopefrom seizures. J Am Coll Cardiol 2002;40:1428.

[44] Merritt TC. Recognition and acute management of patients with panic attacksin the emergency department. Emerg Med Clin North Am 2000;18:289300.

[45] Vein AM, Djukova GM, Vorobieva OV. Is panic attack a mask of psychogenicseizures? A comparative analysis of phenomenology of psychogenic seizuresand panic attacks. Funct Neurol 1994;9:1539.

[46] Biraben A, Taussig D, Thomas P, et al. Fear as the main feature of epilepticseizures. J Neurol Neurosurg Psychiatry 2001;70:18691.

[47] Dressler D, Benecke R. Diagnosis and management of acute movementdisorders. J Neurol 2005;252:1299306.

[48] Colosimo C, Bologna M, Lamberti S, et al. A comparative study of primary andsecondary hemifacial spasm. Arch Neurol 2006;63:4414.

[49] Derry CP, Davey M, Johns M, et al. Distinguishing sleep disorders fromseizures: diagnosing bumps in the night. Arch Neurol 2006;63:7059.

[50] Iranzo A, Santamaria J, Rye DB, et al. Characteristics of idiopathic REM sleepbehavior disorder and that associated with MSA and PD. Neurology2005;65:24752.

[51] Guilleminault C, Gelb M. Clinical aspects and features of cataplexy. Adv Neurol1995;67:6577.

[52] Krahn LE, Boeve BF, Olson EJ, et al. A standardized test for cataplexy. Sleep Med2000;1:12530.

[53] Montagna P, Liguori R, Zucconi M, et al. Physiological hypnic myoclonus.Electroencephalogr Clin Neurophysiol 1988;70:1726.

[54] Han SW, Kim SH, Kim JK, et al. Hemodynamic changes in limb shaking TIAassociated with anterior cerebral artery stenosis. Neurology 2004;63:151921.

[55] Quinette P, Guillery-Girard B, Dayan J, et al. What does transient globalamnesia really mean? Review of the literature and thorough study of 142cases. Brain 2006;129(Pt. 7):164058.

[56] Kotagal P, Costa M, Wyllie E, Wolgamuth B. Paroxysmal nonepileptic events inchildren and adolescents. Pediatrics 2002;110:e46.

[57] Wyllie E, Benbadis SR, Kotagal P. Psychogenic seizures and other paroxysmalnonepileptic events in children. Epilepsy Behav 2002;3:4650.

[58] Rosenow F, Wyllie E, Kotagal P, et al. Staring spells in children: descriptivefeatures distinguishing epileptic and nonepileptic events. J Pediatr1998;133:6603.

[59] Bye AM, Kok DJ, Ferenschild FT, Vles JS. Paroxysmal non-epileptic events inchildren: a retrospective study over a period of 10 years. J Paediatr ChildHealth 2000;36:2448.

[60] Desai P, Talwar D. Nonepileptic events in normal and neurologicallyhandicapped children: a videoEEG. Pediatr Neurol 1992;8:1279.

[61] Carmant L, Kramer U, Holmes GL, et al. Differential diagnosis of staring spellsin children: a video-EEG study. Pediatr Neurol 1996;14:199202.

[62] Rickards H. Tics and ts: the current status of Gilles de la Tourette syndromeand its relationship with epilepsy. Seizure 1995;4:25966.

[63] Holmes GL, Russman BS. Shuddering attacks: evaluation usingelectroencephalographic frequency modulation radiotelemetry andvideotape monitoring. Am J Dis Child 1986;140:723.

[64] Kanazawa O. Shuddering attacks: report of four children. Pediatr Neurol2000;23:4214.

[65] Horrocks IA, Nechay A, Stephenson JB, Zuberi SM. Anoxic-epileptic seizures:observational study of epileptic seizures induced by syncopes. Arch Dis Child2005;90:12837.

[66] Frankel EA, Shalaby TM, Orenstein SR. Sandifer syndrome posturing: relationto abdominal wall contractions, gastroesophageal reux, and fundoplication.

avior 15 (2009) 1521Dig Dis Sci 2006;51:63540.[67] Maydell BV, Berenson F, Rothner AD, et al. Benign myoclonus of early infancy:

an imitator of Wests syndrome. J Child Neurol 2001;16:10912.

[68] Gottlob I, Wizov SS, Reinecke RD. Spasmus nutans: a long-term follow-up.Invest Ophthalmol Vis Sci 1995;36:276871.

[69] Chen S, Melo M, Benbadis SR. The differential diagnosis of seizures in the ICU:a videoEEG study. Presented at the 62nd Annual Meeting of the AmericanEpilepsy Society, Seattle, WA, USA; 2008.

[70] Benbadis SR, Tatum WO. Overinterpretation of EEGs and misdiagnosis ofepilepsy. J Clin Neurophysiol 2003;20:424.

[71] Benbadis SR, Lin K. Errors in EEG interpretation and misdiagnosis of epilepsy:which EEG patterns are overread? Eur Neurol 2008;59:26771.

[72] Benbadis SR. Errors in EEGs and the misdiagnosis of epilepsy: importance,causes, consequences, andproposed remedies. EpilepsyBehav2007;11:25762.

[73] Krauss GL, Abdallah A, Lesser R, et al. Clinical and EEG features of patients withEEG wicket rhythms misdiagnosed with epilepsy. Neurology2005;64:187983.

S. Benbadis / Epilepsy & Behavior 15 (2009) 1521 21

The differential diagnosis of epilepsy: A critical reviewIntroductionPsychogenic nonepileptic attacksSuspecting the diagnosisHistoryDetailed description of the spellsExaminationConfirming the diagnosisEEGVideo/EEG monitoringShort-term outpatient video/EEG monitoring with activation

PsychopathologyManagement

SyncopeOther organic conditionsHypoglycemiaPanic attacksParoxysmal movement disordersAcute dystonic reactionsHemifacial spasmNonepileptic myoclonus

Sleep disordersParasomniasCataplexyHypnic jerks

Transient ischemic attacksMigrainesTransient global amnesia

Conditions and issues specific to young children: Misdiagnosis of epilepsy in childrenBenign nonspecific symptoms misinterpreted as seizuresThe role of misread (overread) EEGs in the misdiagnosis of epilepsy

References