The misdiagnosis of epilepsy is common in infancy and childhood,1,2,3,4 This book is devoted to epilepsy treatment, but if therapy is given for epilepsy that does not exist, then the result may be deleterious to the patient. This chapter outlines the conditions in neonates, infants, and toddlers that may beguile the physician (and even the surgeon) into misdiagnosis.
Past the neonatal period, the definition of an epileptic seizure is fairly straightforward, with the presumption that were an EEG recorded in a seizure, paroxysmal ictal discharges would be evident. Epilepsy is defined as repeated unprovoked epileptic seizures. However, in neonates and very young infants the epileptic “discharge” may appear as a suppression or attenuation of the EEG5 and the epileptic discharge must be inferred by the observer.
For this chapter, the disorders or phenomena are arranged according to median age of first appearance to the parent or the doctor. With notable exceptions,6,7,8,9 median age of onset (MA) is seldom specified in the literature. Episodic events that are liable to be misdiagnosed as epilepsy with a MA of up to age 2 years will be reviewed.
The MA of 1 day does not take into account the onset of symptoms in fetal life (aside from hyperekplexia, “convulsive” fetal movements are best known in pyridoxine dependency but are also seen in cerebral malformations, and other rarer causes of epilepsy). Those with dominantly inherited hyperekplexia due to a mutation in a glycine receptor subunit, especially GLRA1, are stiff infants or neonates, and while they may be at risk of apnea and sudden death they are not usually misdiagnosed as having epilepsy. However, in “sporadic” cases10,11,12,13,14,15,16,17 and in those with mutations in the glycine transporter gene18 prominent nonepileptic neonatal seizures are an important feature.
Vigevano et al10 and Bernardina et al11 gave the first detailed clinical descriptions of the nonepileptic seizures in neonatal hyperekplexia. Aside from massive myoclonic jerks, symptoms were marked by diffuse muscle stiffness, extension of all limbs, and intense cyanosis with bradycardia. The attacks were resistant to intravenous diazepam and oral phenobarbitone, nitrazepam and clonazepam.10 Spontaneous and induced generalized attacks with “sudden massive stiffness and shaking of the limbs lasting one to several minutes, and the mimicked generalized clonic or myoclonic fits” might induce severe bradycardia necessitating cardiac massage.11 A published polygraphic recording11 of one of these attacks that was tonic and vibratory revealed high-voltage repetitive muscle potentials in the upper and lower limb muscles, on most of the EEG channels and on the ECG channel. This polygraphic appearance consisting of bursts of EMG “spikes” is pathognomonic of neonatal hyperekplexia12 (Figs. 16–1 and 16–2).
Figure 16–1.
Neonatal hyperekplexia. A still from the video of an episode of profound syncope is superimposed on the simultaneous EEG and ECG (recorded with Oxford Medical Systems), with 8 seconds of the trace displayed. The infant is stiff and apneic. The EEG is isoelectric but shows a slow “artefact” (duration less than 1 sec) indicated by the white arrows, clinically associated with a (nonepileptic) anoxic spasm. The ECG shows gross bradycardia with atrioventricular dissociation (junctional rhythm). Superimposed on all but two of the EEG channels and on the ECG channel are very high-voltage compound muscle action potentials (CMAP). For most of the visible recording, the CMAPs are rhythmic at about 7 Hz, but at the time of the arrowed spasm the CMAP rate is around 22 Hz (see Fig. 16–2 for further detail). The two EEG channels that do not show these repetitive CMAPs are from the sagittal midline and over bone rather than over muscle. The clinical history was that he developed severe convulsions at age 40 hours, up to 6 per day, most often precipitated by bathing.18,19 He seemed “normal” and not stiff in between episodes of twitching, but after immersion in warm water he had rapid quivering of his limbs, an interrupted cry with fast grunting, and then silence with intense stiffening in a semiflexed posture leading to deep cyanosis. Nose tap was positive but the response gradually waned, such that in due course he was able to play club rugby football. In due course, he was found to have a pathogenic mutation in the GlyT2 gene SLC6A5 (16—patient 6) (www.stars.org.uk. or www.stars-us.org).
Figure 16–2.
This figure shows an extract from the EEG of the same infant with neonatal hyperekplexia due to a mutation in the gene for the presynaptic glycine transporter GlyT2 that was introduced in Figure 16–1 (16—patient 6). The middle trace is virtually isoelectric in keeping with the profound syncope. The first and third channels show superimposed the high-voltage CMAPs mainly at about 8 per second that are typical of neonatal hyperekplexia. In addition, at the solid black arrow, there is a slow deflection of duration about 500 ms with irregular CMAPs superimposed at about 24 Hz. This is the EEG/EMG signature of a nonepileptic anoxic spasm (www.stars.org.uk. or www.stars-us.org).
Pascotto and Coppola13 noted that attacks occurred when the infant was asleep, lasting 20 seconds to several minutes. Polygraphy recorded rhythmic 12 Hz myoclonus from both deltoid muscles. Bernardina et al11 previously emphasized that while the frequent massive and violent jerks that occurred spontaneously or in response to unexpected stimuli disappeared during sleep, the generalized attacks characterized by massive generalized stiffness with violent shaking of the limbs, apnea, cyanosis, and tachycardia followed by bradycardia, could occur spontaneously while asleep.
Misdiagnosis of epilepsy is common from the neonatal period until adult life. A typical example was the reported case of a girl with “fits” from birth, who was diagnosed with myoclonic epilepsy by neurologists when she was 16 years old.20
A simple diagnostic maneuver for hyperekplexia is the nose-tap test. Kurczynski21 first proposed the nose-tap response as a characteristic feature, noting: “A prominent startle response consisting of eye blinking and a flexor spasm of the trunk could be elicited easily and repetitively by tapping the tip of the nose but no other part of the nose, forehead, or face” and “This response seemed to show no habituation. Sudden unexpected loud noises could also elicit a similar startle response.”
It has been known since the first description of hyperekplexia that extremely strong brain stem reflexes are a constant feature.22 It is thus important to emphasize that the response to nose tap (percussing the tip of the nose) in hyperekplexia is not simply a startle (with trunk flexion) but involves prominent head retraction.22,23,24 In this regard, Shahar’s description23 is instructive: “nose tapping was followed by extension of the head accompanied by a brief episode of generalized myoclonic jerking, which recurred with repeated taps.” These two aspects—startle and head retraction reflex—have been confused or conflated in the literature. For instance, Nigro and Lim24 referred to the “hyperekplexic startle response” to nose tapping, and published photographs of an infant showing “startle with sudden adduction of the arms and extension of the head following nose tap.”
Formerly known as familial rectal pain syndrome,25,26 this rare disorder was once thought to be a form of reflex epilepsy27 and responds to some extent to carbamazepine26,27,28 because of the properties of the mutated sodium channel SCN9A.29
Episodes are provoked by maneuvers such as perineal stimulation and consist of flushing—usually with the harlequin phenomenon, such as the right side of the face red with tonic posturing. The “distant” look in the neonate’s eyes is likely a result of extreme pain. Tachycardia is followed by extreme bradycardia or asystole and often profound syncope.26 Severe cyanotic breath-holding28 is not an uncommon manifestation.
The main importance of benign neonatal sleep myoclonus (BNSM)9 is for it to be recognized by observation. BNSM begins in otherwise completely normal neonates but is very easily mistaken for epilepsy,9,30,31,32,33,34 including status epilepticus.33 Rhythmic myoclonus noted only in sleep may affect one or more limbs, and facial muscles are occasionally involved.31,35,36 Although jerking may continue for 90 minutes,37 polygraphy (and on video) reveals a series of very short bursts of 4–5 per second jerking.31 Perhaps the best way to ensure a correct diagnosis is for all neonatal caretakers to view symptoms on video recordings of BNSM.31 It has been reported that rocking the crib of the infant will reproduce BNSM,38 but otherwise it is helpful to show the mother a video of another baby with BNSM.31,39 Antiepileptic medications may worsen BNSM;30 oversedated infants have been admitted to neonatal intensive care units for ventilator support.40
It is now clear that BNSM is overrepresented after maternal opioid drug use in pregnancy.41
Jitteriness is said to occur in 50% of normal-term newborns, but occurs in abnormal neonates as well. Pediatricians and neonatologists do not have problems recognizing this tremor-like phenomenon in the waking state, and it should therefore not prove a diagnostic problem, even when it persists.42
When neonatal seizures have no EEG correlates, they are assumed to be nonepileptic.43 Behaviors such as pedaling come into this category.
It is most important to note that some neonatal behaviors, notably a combination of multifocal myoclonus with no EEG change, and peculiar eye movements and grimacing may indicate the beginnings of pyridoxine dependent (or pyridoxal phosphate responsive) epilepsy.44
Whether or not the opsoclonus-like eye movements and the apneas that may indicate the start of GLUT1 deficiency are epileptic in mechanism is not clear, but it is also important to think of this treatable disorder in that context45,46,47 and indeed in any unexplained paroxysmal movement disorder in later infancy and childhood.47
The peculiar eye movements (such as monocular nystagmus) that herald alternating hemiplegia of childhood (AHC) commonly begin in the newborn period,8 but this condition is discussed later under MA 3 months.
Vigevano and Lispi48 described 13 infants aged between 1 and 3 months with nonepileptic seizures characterized by tonic stiffening with apnea and cyanosis in response to tactile stimulation and almost exclusively when the infants were held in a vertical position. The description is reminiscent of the “awake apnea of Spitzer”49 in which gastroesophageal reflux is the mechanism, but the Italian patients are not reported to have had such reflux. Spontaneous remission within 2 months was the rule.48
