Chapter 22 The paroxysmal dyskinesias
Introduction
The overwhelming majority of individuals with hyperkinetic movement disorders have symptoms that are continuous or continual (e.g., chorea, dystonia, tardive dyskinesia), except for relief with sleep, and with some variation in intensity during periods of stress and relaxation, or other factors such as voluntary movements (e.g., action dystonia, intention myoclonus, intention tremor) or maintaining certain postures (e.g., essential tremor). The symptoms and signs of dopa-responsive dystonia sometimes have a diurnal pattern, being absent or slight in the morning hours and becoming more pronounced as the day proceeds (see Chapter 12). Some dyskinesias (Table 22.1) are characterized as occurring intermittently, such as myoclonus (Fahn et al., 1986; Hallett et al., 1987) and startle syndromes (hyperekplexia) (Andermann and Andermann, 1986) that can be triggered by a variety of stimuli. Restless legs syndrome is best described as a diurnal disorder, manifesting itself primarily (1) in the evenings with abnormal crawling sensations that are relieved by the patient moving about and (2) during the night with periodic movements in sleep (Hening et al., 1986; Walters et al., 1991). Sandifer syndrome is a tilting downwards of the head after eating a meal, occurring in boys, and associated with gastroesophageal reflux (Menkes and Ament, 1988). Movements that occur as a result of akathisia relieve the sensation of inner restlessness. These movements occur intermittently and are usually complex (see Chapter 19). Stereotypies are also complex movements, which appear largely in individuals with learning disability, autism, and schizophrenia (Fahn, 1993). They are not always present but occur frequently and almost continually, but in some patients they appear as intermittent bursts (Tan et al., 1997). The explanation as to why these patients make these movements is not known, but suggestions such as “being in touch with the environment” have been proposed. Perhaps the commonest dyskinesias that occur intermittently are tics, which are suppressible to varying degrees (Koller and Biary, 1989). Although tics and myoclonic jerks, since they commonly occur out of a normal background, could possibly be considered paroxysmal, this term is usually reserved for an entirely different set of hyperkinetic movement disorders, which is the topic of this chapter. The term paroxysmal dyskinesia has been applied to these disorders.
Table 22.1 Classic movement disorders that usually appear in bursts or with specific actions, but are not considered paroxysmal dyskinesias
The movements usually occur so frequently that they are not distinguished with a “paroxysmal” label.
The common neurologic paroxysmal disorders are epilepsy and migraine. Movement disorders that appear “out of the blue” and are transient and recurring are uncommon, and often present to the clinician as confusing diagnostic problems. Not uncommonly, the history that is provided by the patient does not convey the information that the episodes of abnormal movements occur at intermittent intervals, and the clinician might overlook the category of paroxysmal dyskinesias. Therefore, if the examination does not reveal the presence of a movement disorder, the clinician needs to consider the possibility that he or she is dealing with a paroxysmal dyskinesia, and thereby ask the appropriate questions that can lead to the proper diagnosis. To compound the problem, nonfamilial paroxysmal movement disorders are often psychogenic in etiology (Bressman et al., 1988); therefore, the clinician has the problem of determining the etiologic distinction of psychogenic versus organic.
The pathophysiology of paroxysmal dyskinesias is not understood, and “epilepsy of the basal ganglia” has been a serious consideration but difficult to prove. Some paroxysmal dyskinesias are supplementary sensorimotor seizures, including many of the hypnogenic variety (Lüders, 1996). The classification of the paroxysmal dyskinesias is still incomplete and evolving (Demirkiran and Jankovic, 1995), and treatment for many of them is often unsuccessful, but treatment for some can be highly successful. A most welcome burst of research on the genetics of the paroxysmal dyskinesias is shedding new light in the classification of these disorders (Nutt and Gancher, 1997), their diagnoses, and their mechanism of action as channelopathies (Griggs and Nutt, 1995) now that the genes that have so far been discovered for these disorders are those that code for some of the ion channels.
Definitions: transient, paroxysmal, episodic, and periodic
Some pediatric neurologists have utilized the term “transient” for a number of movement disorders in children, but the concept of paroxysmal dyskinesias is becoming more widely recognized among pediatric neurologists (Lotze and Jankovic, 2003). Excluding tic disorders, Fernandez-Alvarez (1998) reviewed the 356 movement disorder cases under the age of 18 years that had been seen in his department’s clinic and reported that 19% of them were classified as transient dyskinesias (Table 22.2). Sometimes these transient movements can be mistaken for seizures (Donat and Wright, 1990). An accurate diagnosis can be reassuring to the family because these movements are not seizures and are almost always benign. The diagnosis depends on the clinical features; diagnostic tests are normal and unnecessary.
Table 22.2 Transient movement disorders in children (clinic of Fernandez-Alvarez, 1998)
Kotagal and colleagues (2002) evaluated their cases of paroxysmal nonepileptic events. These are reviewed in Chapter 25. These were most common in adolescents and the school-age group. In the preschool group, the most common diagnoses were stereotypies, hypnic jerks, parasomnias, and Sandifer syndrome.
The term “periodic” is defined by both Dorland’s and Webster’s as recurring at regular intervals of time. Since the paroxysmal dyskinesias do not recur at regular intervals, the term “periodic” would not be appropriate. Despite this, in neurology this term is used for the condition of familial periodic paralysis (Rowland and Gordon, 2005) and familial periodic ataxia (e.g., Vighetto et al., 1988), even though muscle weakness and ataxia, respectively, in these disorders do not occur at regular intervals. Some authors (e.g., Griggs et al., 1978) had used the term paroxysmal ataxia in preference to periodic ataxia, and others use the term episodic ataxia (Zasorin et al., 1983). But the literature today uses “episodic ataxia.”
According to dictionary definitions, either “paroxysmal” or “episodic” would be an appropriate term for the dyskinesias that are under discussion here. By common usage, and with few exceptions (e.g., Margolin and Marsden, 1982), “paroxysmal” has been chosen in preference over “episodic” for the choreoathetotic and dystonic types and is used in this chapter. The term “episodic” is also applied here to the ataxias in keeping with the current trend in the literature. The term “paroxysmal” has been utilized to indicate that the symptoms occur suddenly out of a background of normal motor behavior. It does not define the frequency, severity, duration, aggravating factors, or type of dyskinesia of the attack. These features vary and are important in the current nosology and classification of the paroxysmal dyskinesias.
Historical aspects
This is a condensed review of historical highlights of the paroxysmal and episodic dyskinesias. For a more complete discussion, see the review by Fahn (1994). A review of paroxysmal dyskinesias in the Japanese literature is available in English by Hishikawa and colleagues (1973).
Earliest descriptions: reported as epilepsy
Although Gowers (1885) is often credited with the first report of movement-induced seizures, it is possible that his cases actually represented paroxysmal dyskinesia. One of his patients was a boy whose attack lasted 15 seconds, but the boy was said to be unconscious during his initial attack. Later, he remained awake during the attacks. Another patient was a girl whose attacks started at the age of 11 years and occurred when she arose suddenly after prolonged sitting. But at least one of her attacks was said to be associated with a terrified expression, flushed facies, and dilated pupils. Subsequent to Gowers, a number of reports of “movement-induced seizures” appeared in the literature. Many of these reports have been published under the designation of reflex epilepsy and tonic seizures induced by movement. But unlike most motor convulsions, there was no alteration in the state of consciousness. Moreover, some of these reports had more than tonic contraction, namely, they included sustained twisting, athetosis, and chorea. These characteristics are today referred to as paroxysmal dystonia and paroxysmal choreoathetosis, rather than convulsive seizures. Even the presence of choreoathetosis did not lead the earliest interpreters of these brief attacks to conclude they were a movement disorder; instead, they were considered to be a form of epilepsy, the cerebral site of these “seizures” being in the basal ganglia or in the subcortical region.
After the report of Gowers, the next report of movement-induced paroxysmal movements appears to be that of Spiller in 1927. Spiller described two patients with brief tonic spasms that were brought on by voluntary movement of the involved limbs and, in one of them, also by passive manipulation. Spiller called this “subcortical epilepsy.” Wilson (1930) described a 5-year-old boy who had brief attacks of unilateral torsion and tonic spasm that lasted up to 3 minutes and were precipitated by fright or excitement. There was no loss of consciousness. The attacks could be preceded by pain. Wilson considered this to be reflex tonic epilepsy and thought it also to be subcortical in origin. In more recent times, the concept that these attacks of tonic, often twisting, contractions without loss of consciousness are uncommon seizure disorders has continued (Whitty et al., 1964; Burger et al., 1972). It would appear that today these movement-induced involuntary movements would be considered paroxysmal choreoathetosis/dystonia rather than convulsive movements of the reflex epilepsy type. Differentiation of the attacks between cortical seizures and paroxysmal dyskinesias is sometimes difficult. Clouding of consciousness, if it occurs, would point to a seizure disorder.
By 1966 and 1967, when papers using the term “paroxysmal choreoathetosis” began regularly to appear (Stevens, 1966; Kertesz, 1967; Mushet and Dreifuss, 1967), particularly those cases induced by movements, there were still occasional papers referring the condition to a seizure disorder. As is discussed below in the section on paroxysmal hypnogenic dyskinesias, nocturnal epilepsy is now considered to be the leading cause.
Reported as a paroxysmal disorder of involuntary movements
In 1940, Mount and Reback (1940) introduced a new concept, that of labeling attacks of tonic spasms plus choreic and athetotic movements as a paroxysmal type of movement disorder. They described a 23-year-old man who had had both large and small “spells” since infancy. Both types were preceded by a sensory aura of tightness in parts of the body or by a feeling of tiredness. The movements involved the arms and legs and were usually a combination of sustained twisted posturing and chorea and athetosis. The small attacks lasted from 5–10 minutes. Longer attacks were considered large and also involved the neck (retrocollis), eyes (upward gaze), face (ipsilateral to the limbs if the limb involvement was unilateral), and speech. These large attacks lasted for as long as 2 hours, and the movements were considered to resemble those seen in Huntington disease. There was never a loss of consciousness or clonic convulsive movements, biting of the tongue, or loss of sphincter control. Drinking alcohol, coffee, tea, or cola would usually bring on an attack. Fatigue, smoking, and concentrating were other precipitating factors. The attacks would clear more rapidly if the patient lay down and would be aborted by sleep. The patient had an average of one large and two small attacks a day. Between attacks, the neurologic examination was normal. Phenytoin and phenobarbital had no effect, and scopolamine was the only drug that was found to reduce the frequency, severity, and duration of the attacks. The family history revealed 27 other members who had similar attacks; the pedigree showed autosomal dominant inheritance with what appears to be complete penetrance. Mount and Reback called this disorder “familial paroxysmal choreoathetosis.”
Mount and Reback’s paper became the seminal paper in the field of paroxysmal dyskinesias. Following its publication, most of the reports in the literature referenced it over the next five decades. However, the next report of a large family with similar attacks of muscle spasms did not refer to it. In 1961 Forssman (1961) described a family with autosomal dominant inheritance in which there were attacks lasting from 4 minutes to 3 hours.
The next large family was described in 1963 by Lance. Like Forssman (1961), Lance also did not relate this to nor refer to Mount and Reback’s report, nor did he mention the report by Forssman. In fact, Lance considered his patients to have a form of epilepsy. Later, Lance (1977) was to write one of the definitive papers in this field, containing a useful classification scheme in which he related his family to those of Mount and Reback (1940), Forssman (1961), and Richards and Barnett (1968).
Although there were reports of patients whose paroxysmal dyskinesias were induced by sudden movement, they were not denoted by any special terminology until 1967, when Kertesz (1967) introduced the label “paroxysmal kinesigenic choreoathetosis” (PKC). This label has developed into a most useful and widely accepted designation, since the kinesigenic feature has proven to be so characteristic. Kinesigenicity has an important place in the classification of the paroxysmal dyskinesias, and Demirkiran and Jankovic (1995) recommended that the term paroxysmal kinesigenic dyskinesia (PKD) be used instead, because the movements can be other than choreoathetotic. That suggestion is followed in this chapter. As is pointed out below, the PKD designation can be applied to some patients who do not have the dyskinesias triggered by sudden movement (or startle).
Kertesz reported 10 new cases of paroxysmal dyskinesia and reviewed the literature. Among the important features of his paper, Kertesz differentiated the kinesigenic variety (induced by sudden movement) from that described by Mount and Reback, by Forssman, and by Lance, which were aggravated not by movement but by alcohol, caffeine, and fatigue. It should be noted that Kertesz differentiated the kinesigenic type from that reported by Mount and Reback (1940) and by Lance (1963), but he failed to mention the paper by Forssman (1961).
Although phenytoin was recognized earlier as a very useful agent for PKD, carbamazepine was later found to be as useful and was introduced as a treatment by Kato and Araki (1969). This drug currently appears to be the one that is most commonly used for this disorder.
After Lance’s paper in 1963, Richards and Barnett (1968) reported the next big family with the same type of paroxysmal dyskinesia as Mount and Reback’s case and thought that Lance’s family (1963) represented a variant, since there were only tonic spasms and no movements in that family. Richards and Barnett emphasized the nonkinesigenic nature of the attacks and felt that the terms rigidity, tremor, dystonia, torsions spasm, athetosis, chorea, and hemiballism could all be used for such movements, often blending into each other. To emphasize the postural and increased tone, they added “dystonic” to the label. They recommended avoiding the term “epilepsy” until the pathophysiology is better known. Richards and Barnett coined the term “paroxysmal dystonic choreoathetosis” (PDC), which was later adopted by Lance in 1977. The terms “paroxysmal nonkinesigenic choreoathetosis” and “paroxysmal dystonia” are sometimes used instead of PDC (Bressman et al., 1988). We have adopted the term paroxysmal nonkinesigenic dyskinesia (PNKD) proposed by Demirkiran and Jankovic (1995).
The original cases that were reported as PNKD were idiopathic and usually familial. It was not long before symptomatic cases began to be reported in which the attacks of movements were reported as a PNKD: perinatal encephalopathy by (Rosen, 1964), encephalitis (Mushet and Dreifuss, 1967), and head injury (Whitty et al., 1964; Robin, 1977). However, earlier reports of symptomatic PNKD had been described as a manifestation of multiple sclerosis, but considered as a form of epilepsy (Matthews, 1958; Joynt and Green, 1962; Verheul and Tyssen, 1990). Many other etiologies have been reported since the cases in the 1950s and 1960s (see below).
One of the most enlightening papers (Lance, 1977) achieved the following:
Instead of Lance’s proposed classification (item 5 in the preceding list) based on duration of the attacks, the classification scheme that is adopted here is the one based on precipitating factors suggested by Demirkiran and Jankovic (1995).
The next historical advances were the recognition that (1) idiopathic PNKD can occur sporadically and not just in families (Bressman et al., 1988) and that (2) sporadic PNKD is often psychogenic in origin (Bressman et al., 1988; Fahn and Williams, 1988).
Paroxysmal hypnogenic dyskinesia (PHD)
Horner and Jackson (1969) described two families in which several members of the family had attacks of involuntary movement that occurred during sleep. These appear to be the first cases of paroxysmal hypnogenic dyskinesia (PHD) to be reported. Family W is of particular interest because some affected members had classic PKD, some had hypnogenic, and others had a combination. Case 3 in this family began with the hypnogenic variety at age 8. By age 11, daytime attacks also occurred, sometimes triggered by sudden movement. Gradually, the hypnogenic episodes disappeared, leaving the patient with kinesigenic dyskinesia that responded to anticonvulsants. Lugaresi and his colleagues (Lugaresi and Cirignotta, 1981; Lugaresi et al., 1986) independently rediscovered and eventually popularized the syndrome of paroxysmal hypnogenic dyskinesias.
In addition to these short-duration attacks, Lugaresi and colleagues (1986) reported long-duration hypnogenic attacks. Such long-duration attacks occur in a minority of individuals with paroxysmal hypnogenic dyskinesia. These longer attacks last from 2 to 50 minutes and do not respond to medication, including anticonvulsants, tricyclics, benzodiazepines, and antipsychotics.
There has long been considerable speculation as to whether the short-duration hypnogenic attacks could be a manifestation of epilepsy, since they respond so well to anticonvulsants. The lack of abnormal electroencephalographic (EEG) findings during the attack has been used to argue against this concept. However, there is accumulating evidence that many paroxysmal hypnogenic dyskinesias are indeed due to seizures. Tinuper and colleagues (1990) described three patients with this disorder who had EEG evidence for frontal lobe seizures as a cause of the attacks. Sellal and colleagues (1991) and Meierkord and colleagues (1992) studied a series of patients with hypnogenic dystonia and concluded that these represent seizure disorders, particularly of frontal lobe epilepsy because repeated nocturnal EEG recordings often reveal epileptic patterns of abnormalities. Seizures arising near the mesial posterior frontal supplementary sensorimotor area may be a particular culprit in inducing paroxysmal hypnogenic dyskinesias in children (Bass et al., 1995). These types of seizures tend to be brief, frequent, and with bilateral tonic posturing, gross proximal limb movements, and preserved consciousness. Dystonic and other dyskinetic features may result from spread of epileptic activity from the mesial frontal region to the basal ganglia because there are close anatomic connections between them. It appears that the short-lasting attacks of paroxysmal hypnogenic dyskinesias are most likely due to seizures, but the question remains whether patients without abnormal EEGs and more prolonged hypnogenic attacks could have something more akin to the paroxysmal dyskinesias. In a family with autosomal dominant nocturnal frontal lobe epilepsy, interictal EEGs were normal, but ictal video-EEG studies showed that the attacks were partial seizures with frontal lobe origin (Scheffer et al., 1995). Fish and Marsden (1994) have reviewed epilepsy masquerading as a movement disorder, and they concluded that most cases of hypnogenic dyskinesias are considered to be due to epilepsy, as did Lüders (1996), with cyclic alternating EEG pattern believed to be a provocative factor (Terzano et al., 1997).
The genetics of hypnogenic dyskinesias/seizures is being explored. A large autosomal dominant Australian family (Oldani et al., 1998) and a Norwegian family (Nakken et al., 1999) have been described with mutations in the nicotinic acetylcholine receptor alpha 4 subunit (CHRNA4) gene, located on chromosome 20q13.2–q13.3. A second acetylcholine receptor subunit, CHRNB2, is also associated with autosomal dominant nocturnal frontal lobe epilepsy (Phillips et al., 2001). Another family with autosomal dominant hypnogenic frontal-lobe epilepsy has been mapped to 15q24 (Phillips et al., 1998).
In addition to epilepsy mimicking hypnogenic paroxysmal dyskinesias, there is a syndrome of infantile convulsions and paroxysmal dyskinesias, referred to as the “infantile convulsions with choreoathetosis” (ICCA) syndrome (Lee et al., 1998). The gene for this disorder has been mapped to the pericentromeric region of chromosome 16 (see the discussion in the section on PKD and also the section “Molecular genetics of paroxysmal dyskinesias”). Single photon emission computed tomography (SPECT) studies revealed alterations in local cerebral perfusion in the sensorimotor cortex, the supplementary motor areas, and the pallidum (Thiriaux et al., 2002).
Transient paroxysmal dyskinesias in infancy
Snyder (1969) introduced a new type of paroxysmal dyskinesia that he called “paroxysmal torticollis in infancy.” He described 12 cases of intermittent head tilting in young infants. The age at onset was between 2 and 8 months of age, except for three cases, in which the first attacks occurred at 14, 17, and 30 months. The attacks would occur about two to three times a month and last from 10 minutes to 14 days, usually 2–3 days. The head would tilt to either side and often rotate slightly to the opposite side. There is no distress unless a parent attempts to straighten the head, upon which the baby cries. In some cases, the head tilting is associated with vomiting, pallor, and agitation for a short period. The infant is normal between attacks, which disappear after months or years, usually around age 2 or 3 years. Subsequently, a number of similar cases have been described (Gourley, 1971; Sanner and Bergstrom, 1979; Bratt and Menelaus, 1992), including familial cases (Lipson and Robertson, 1978). Sanner and Bergstrom (1979) reported a patient whose father had a similar condition in early infancy, suggesting that this disorder is hereditary.
The clinical picture of paroxysmal torticollis in infancy (Video 22.1) that has evolved is that the trunk can also be involved, with lateral curvature concave to the same side as the head tilting, and the ipsilateral leg can be flexed. Onset can be as early as the first months of life and can recur every couple of weeks until they disappear before the age of 2 years. Each attack can last a couple of hours to a couple of weeks. In between attacks, the child is normal. The main differential diagnosis is a posterior fossa tumor and Sandifer syndrome (Menkes and Ament, 1988). 
In 1988, the clinical spectrum expanded with the report by Angelini and colleagues (1988) under the name “transient paroxysmal dystonia in infancy.” They described nine patients who had onset of the paroxysmal dyskinesias between 3 and 5 months of age, except for one patient who had an onset at age 1 month. Three had a history of perinatal brain damage, six did not. The attacks consisted of opisthotonus, increased muscle tone with twisting of the limbs, and, in three, neck and trunk twisting, thereby linking this with “paroxysmal torticollis in infancy.” The attacks lasted several minutes, with a maximum of 2 hours in one patient. They would occur from several times per day (Video 22.2) to once a month. Remission occurred between the ages of 8 and 22 months, with two not yet having reached a remission. 
Dunn (1981) described an infant with head turning and posturing of the right arm lasting from 45 minutes to 18 hours. There were six attacks from age 26 months to age 40 months. The author did not mention the possible diagnosis of paroxysmal torticollis in infancy and made a diagnosis of paroxysmal dystonic choreoathetosis instead. One should consider the possibility that PNKD may occur in infancy and disappear over several months. If so, then the paroxysmal torticollis in infancy of Snyder and the paroxysmal dystonia in infancy of Angelini may represent the lowest age spectrum of PNKD and a benign form of the disorder.
Some patients with benign paroxysmal torticollis of infancy come from kindreds with familial hemiplegic migraine linked to CACNA1A mutation, and after recovering from these episodes as they reach childhood, they might have migraines (Giffin et al., 2002). In fact, a family with individual members of the kindred having one or more of the following: paroxysmal tonic upgaze, benign paroxysmal torticollis of infancy, and episodic ataxia, with familial migraine and a CACNA1A mutation (as in EA-2) has been reported (Roubertie et al., 2008). The clinical picture of benign paroxysmal torticollis of infancy has been summarized to consist of attacks usually lasting less than a week, recurring from every few days to every few months, improving by age 2 years, and ending by age 3; there is very frequently a family history of migraine (Rosman et al., 2009).
Intrauterine cocaine exposure can be associated with multiple transient dyskinesias. Beltran and Coker (1995) described four infants who tested positive for cocaine metabolite at birth with subsequent transient dystonic reactions, beginning at 3 hours to 3 months of age and persisting for several months.
The clinical syndrome of transient paroxysmal dyskinesia of infancy appears distinct from the syndrome referred to as “benign paroxysmal tonic upgaze of childhood” (Ouvrier and Billson, 1988; Deonna et al., 1990; Echenne and Rivier, 1992; Campistol et al., 1993), which is a sustained tonic conjugate upward deviation of the eyes that begins in infancy and eventually disappears in childhood. This appears to be an autosomal dominant disorder (Guerrini et al., 1998). An infant with tonic upgaze was found to have a partial tetrasomy of 15q (Joseph et al., 2005). Ataxia may be present, and there can be clumsiness and delayed walking. The ocular deviations lessen in the morning hours and disappear with sleep. Acetazolamide is not effective; however, Campistol and colleagues (1993) reported levodopa to be effective. Perhaps the tonic upgaze with diurnal fluctuations would be a better term than “paroxysmal.” Not all cases of infantile transient tonic upgaze disturbance are benign, although the mean age of offset is 2.5 years (Hayman et al., 1998). Ouvrier and Billson (2005) published a recent review of this disorder. The family reported by Roubertie and colleagues (2008) suggests that tonic upgaze and transient torticollis and EA-2 may be related. A secondary cause of paroxysmal tonic upgaze has been reported: a 1-year-old girl with a hypomyelinating leukoencephalopathy, who presented in the neonatal period with episodes of sustained paroxysmal tonic upward gaze, roving eye movements, pendular nystagmus, and severe hypotonia, with the later appearance of pyramidal and extrapyramidal signs and no development (Blumkin et al., 2007).
Another paroxysmal ocular disorder, known as paroxysmal ocular downward deviation has been described in normal and brain-damaged infants (Yokochi, 1991; Miller and Packard, 1998). The ocular displacement was accompanied by closure of the upper eyelids, and the episode lasted seconds.
The syndrome of “benign myoclonus of infancy” can be mistaken for infantile spasms, but the benign EEG and clinical course allow for a clear distinction. The movements are sudden myoclonic and shuddering episodes of the head and shoulders (Lombroso and Fejerman, 1977; Fejerman, 1984). The jerks often repeat in a series; consciousness remains intact.
Another benign paroxysmal disorder in infants is spasmus nutans. It consists of a slow (2.4 Hz) tremor of the head, usually horizontal, and often an associated pendular nystagmus (Antony et al., 1980). It tends to disappear within 6 months and must be differentiated from congenital nystagmus (Fernandez-Alvarez, 1998).
Episodic (paroxysmal) ataxias and tremor
Intermittent ataxia can be due to metabolic defects such as Hartnup disease (Baron et al., 1956), pyruvate decarboxylase deficiency (Lonsdale et al., 1969; Blass et al., 1970, 1971), and maple syrup urine disease (Dancis et al., 1967). Fever often triggers the attacks of ataxia. In one case with pyruvate decarboxylase deficiency (Blass et al., 1971), choreoathetosis tended to accompany the chorea. Paroxysmal ataxia and dysarthria have also been reported to occur in multiple sclerosis (Andermann et al., 1959; Espir et al., 1966; DeCastro and Campbell, 1967; Miley and Forster, 1974; Gorard and Gibberd, 1989), which as remarked previously, is a disorder that also can cause paroxysmal choreoathetosis/dystonia. The attacks of paroxysmal ataxia due to multiple sclerosis last seconds, and are thus much shorter than the attacks described below. They also can respond to carbamazepine. Paroxysmal ataxia has also been reported in Behçet disease (Akmandemir et al., 1995).
A second family from the same region in North Carolina also had ocular motility problems. These were abnormal smooth pursuit with normal saccades, dampened opticokinetic nystagmus, inability to suppress the vestibulo-ocular reflex, gaze-evoked nystagmus, and episodic attacks of horizontal diplopia, oscillopsia, ataxia, nausea, vertigo, and tinnitus (Damji et al., 1996). This family’s disorder was considered part of the same neurologic disorder that is referred to as periodic vestibulo-cerebellar ataxia, like Farmer and Mustian’s family. Of special clinical significance is the lack of dysarthria. Genetic studies showed that the ataxia of this family is distinct from the two types of episodic ataxias (EA-1 and EA-2) that were previously characterized genetically (see below).
Hill and Sherman (1968) described another family, in which onset was in childhood in many of the affected members and there was no development of progressive ataxia. White (1969) described another family that experienced childhood onset and a benign course. All the families showed autosomal dominant inheritance.
An important advance was the discovery by Griggs and colleagues (1978) that acetazolamide can effectively prevent attacks. These authors showed this benefit in one kindred with familial paroxysmal ataxia. The following year, Donat and Auger (1979) had similar results in another kindred. Fahn (1983, 1984) reported a woman who had paroxysmal tremor, both intention and resting, associated with ataxia and postural instability during the attack; acetazolamide eliminated the attacks. Factor and colleagues (1991) reported an infant who had three attacks of coarse tremor and an orofacial dyskinesia that resembled that seen with tardive dyskinesia. Each attack lasted several hours before spontaneously clearing. Tetrahydrobiopterin, the cofactor for the enzymes tyrosine hydroxylase and phenylalanine hydroxylase, was reduced. The child responded to levodopa.
Mayeux and Fahn (1982) reported a patient with PNKD in a background of hereditary ataxia. Onset of PNKD was at age 10; onset of ataxia was at age 19. During an attack, which could last from 10 minutes to 4 hours, there was an accompanying increase of ataxia. Initially there was an 8-month response to acetazolamide. After the drug was no longer effective, the patient’s PNKD responded to clonazepam. This patient might be a link between familial PNKD and paroxysmal ataxia.
Several other reports of acetazolamide-responsive familial paroxysmal ataxia have been reported (Aimard et al., 1983; Zasorin et al., 1983; Koller and Bahamon-Dussan, 1987). Although computed tomography (CT) has been normal, magnetic resonance imaging (MRI) studies have revealed selective atrophy in the anterior cerebellar vermis (Vighetto et al., 1988).
Families with a combination of periodic ataxia and persistent, continuous electrical activity in several muscles, reported either as myokymia (Van Dyke et al., 1975; Hanson et al., 1977; Gancher and Nutt, 1986; Brunt and Van Weerden, 1990) or as neuromyotonia (Vaamonde et al., 1991), have been described. Description of the attacks, which are brief and are sometimes preceded by sudden movement, include dyskinetic movements and sustained posturing as well as ataxia, dysarthria, and vertigo. This type of paroxysmal ataxia is now called episodic ataxia 1 (EA-1).
In 1986, Gancher and Nutt (1986) classified the hereditary episodic ataxias into three syndromes. In one group are those cases associated with persistent myokymia or neuromyotonia (now called EA-1). They described the attacks being precipitated by fatigue, excitement, stress, and physical trauma, but the family reported by Vaamonde and colleagues (1991) had attacks triggered by sudden movement, and kinesigenicity is now recognized as a feature. There is no dizziness or vertigo. The attacks last 2 minutes or less. Acetazolamide and anticonvulsants are usually ineffective. The gene for this type of paroxysmal ataxia has been located at chromosome 12p13 (Litt et al., 1994).
The second group (now known as EA-2) is featured by attacks of ataxia (with or without interictal nystagmus and with or without persistent ataxia), responding to acetazolamide or amphetamines. The attacks are precipitated by exercise, fatigue, or stress and occasionally by carbohydrate or alcohol ingestion. In addition to ataxia, the attacks are accompanied by vertigo, headache, nausea, and malaise. The attacks last for several hours or until the patient falls asleep. In recent years, additional families have been reported with these features (Bain et al., 1991; Baloh and Winder, 1991; Hawkes, 1992). The siblings reported by Bain and colleagues (1991) had persistent diplopia due to superior oblique paresis as part of the syndrome. Using [31P] nuclear magnetic resonance spectroscopy, Bain and his colleagues (1992) found the pH levels in the cerebellum to be increased in untreated subjects with acetazolamide-responsive paroxysmal ataxia; the pH dropped to normal with treatment. The gene for this type of paroxysmal ataxia has been mapped to chromosome 19p13 (von Brederlow et al., 1995; Vahedi et al., 1995).
Gancher and Nutt (1986) listed a third group, which is kinesigenic. Typical PKD can occur in some members of the family. The attacks of ataxia last minutes to hours, whereas the PKD lasts seconds. The disorder can resolve with time. Acetazolamide appears to be ineffective, but phenytoin is effective for both the kinesigenic ataxia and the PKD. In their review, Griggs and Nutt (1995) place this third type with associated PKD in the first group of paroxysmal ataxias. Genotyping has now definitively placed this as EA-1 (Nutt and Gancher, 1997).
A case was reported in which a young girl had attacks of ataxia associated with fevers and accompanied by vertical supranuclear ophthalmoplegia (Nightingale and Barton, 1991). The ataxia and eye findings can last days.
Miscellaneous paroxysmal disorders
Keane (1984) reported two patients with post-traumatic periodic, rhythmic movements of the tongue. The attacks occurred about every 20 seconds and each attack lasted 10 seconds. They consisted of three undulations per second. Eventually, the movements diminished. Other cases of episodic lingual dyskinesias were associated with epilepsy (Jabbari and Coker, 1981) and pontine ischemia (Postert et al., 1997).
A few paroxysmal dyskinesias are mentioned here, but are not discussed further. These are Sandifer syndrome (prolonged head tilting in children following eating, due to gastroesophageal reflux) (see the review by Menkes and Ament, 1988); hyperekplexia (excessive startle syndrome with complex movements) (see the review by Andermann and Andermann, 1986 and Chapter 20); stereotypy (Duchowny et al., 1988; Tan et al., 1997), and paroxysmal bursts of myoclonus and tics. Classically, stereotypy, myoclonus, and tics are each recognized as a specific class of movement disorders, and they characteristically present as paroxysmal bursts of their type of movement. As a result, their discussion should be separated from discussion of conditions labeled as “paroxysmal.” Shuddering attacks in children (Holmes and Russman, 1986) are brief bursts of rapid shivering-like movements of the head and arms, occurring up to 100 times per day; they can begin in infancy or in older children, and they resolve over time. The attacks lasts several seconds without impairment of consciousness. The frequency of shuddering movements as seen on EMG or EEG was similar to that of essential tremor (Kanazawa, 2000).
Classification of the paroxysmal dyskinesias
Because the various movement phenomena in any of the paroxysmal dyskinesias can vary from chorea/ballism to the sustained contractions of dystonia in any given patient, Demirkiran and Jankovic (1995) suggested that these disorders be labeled with the more generic names paroxysmal kinesigenic dyskinesia (PKD), paroxysmal nonkinesigenic dyskinesia (PNKD) (whether short-lasting or long-lasting), and paroxysmal exertion-induced dyskinesia (PED) regardless of the duration of the attack. This classification system is used here. But the paroxysmal dyskinesias that last seconds and might not be induced by sudden movement are also included in the PKD category because of their otherwise similar short duration of attacks and response to anticonvulsants.
The highlights of various categories of paroxysmal dyskinesias are presented in outline format in Table 22.3.
| A. Paroxysmal kinesigenic dyskinesia (PKD) |
Paroxysmal kinesigenic dyskinesia (PKD)
Clinical features
Kinast and colleagues (1980) reported as Case 4 a patient with typical brief dyskinesias occurring many times a day, preceded by paresthesias, and responding to anticonvulsants but not induced by sudden movement. Although this was technically not PKD, because sudden movements did not trigger any attacks, the clinical features otherwise resemble the attacks that are seen in PKD. As was mentioned above, this type of paroxysmal dyskinesia is placed in this category of PKD. Since the attacks of PNKD are usually so prolonged and because they do not ordinarily respond to anticonvulsants, this case does not fit into the PNKD category. See the discussion on the classification of these brief attacks as symptomatic PKD below.
Plant (1983) emphasized the focal and unilateral nature of PKD in many patients (Table 22.4).
Table 22.4 Laterality of types of paroxysmal kinesigenic dyskinesia (Plant, 1983)
| Laterality of attacks | No. |
|---|---|
| Unilateral – one side only | 25 |
| Unilateral – either side | 12 |
| Unilateral and bilateral | 11 |
| Bilateral only | 22 |
| Not stated | 3 |
| Total | 73 |
Primary paroxysmal kinesigenic dyskinesia
The etiology of most case reports of PKD has been idiopathic and predominantly hereditary, inheritance being autosomal dominant. For some unexplained reason males are more often affected than females, with a ratio of 3.75 : 1 (75 males and 20 females reported in Fahn, 1994). A large series of 150 cases were reported from a questionnaire in Japan. This gender imbalance was supported by the additional 26 cases reported by Houser and colleagues (1999), consisting of 23 men and 3 women, and another 150 case from Japan (Nagamitsu et al., 1999). Adding all these cases together brings the total to 218 men and 53 women or a ratio of 4.1 : 1. Age at onset shows a wide range, usually starting in childhood between the ages of 6 and 16 years, but can range from 6 months to 40 years (Frucht and Fahn, 1999; Li et al., 2005). Excluding the cases by Nagamitsu and colleagues (1999), the mean age at onset is 12 years; the median age is also 12 years. Familial cases might be more common among the Japanese (Kishimoto, 1957; Fukuyama and Okada, 1967; Kato and Araki, 1969) and Chinese (Jung et al., 1973). The survey reported by Nagamitsu and colleagues (1999) found 53 sporadic cases and 97 familial ones.
There is one report of PKD developing in a patient who had essential tremor (Nair et al., 1991
