49 Paroxysmal choreic, athetotic, or dystonic attacks

We have grouped attacks showing choreic, athetotic, and dystonic movements together in this section. Sometimes it is difficult to determine whether the movement disturbance in a particular type of paroxysm is choreic or dystonic. This can be due to the brief duration of the bout, the presence of both elements, or inadequate descriptions in the literature. Thus, it has been believed that the movements in paroxysmal kinesigenic dyskinesia tend to be choreic, while those in paroxysmal dystonic dyskinesia and paroxysmal exercise-induced dyskinesia are dystonic. However, this may not necessarily be the case and a range of dyskinesias can occur in these disorders (Demirkirin and Jankovic 1995, Bhatia 1997, Houser et al. 1999). Similarly, there are a variety of choreic, myoclonic, and dystonic actions in the disorder we have chosen to call ‘paroxysmal hypnic dystonia’ but sustained posturing seems to be the most prominent.

The borderland between dystonic and tonic attacks is inevitably somewhat hazy and perhaps somewhat artificial. We have chosen to include paroxysmal hypnic dystonia here and other forms of tonic seizures in Chapter 50, although there is evidence to suggest that both are epileptic and may involve similar structures. Similarly, the dystonic posturing that is sometimes seen as a part of a complex partial seizure is discussed here, although it could also have become part of Chapter 50. The tonic attacks of multiple sclerosis sometimes have a dystonic element and are mentioned in both Chapters 49 and 50. Rarely patients with this disease have clear choreic movements and this is covered here.

Another issue that should be mentioned is that of familial spontaneous or exercise-induced cramps. At times these have been the only manifestation of paroxysmal exercise-induced dyskinesia, while on other occasions they may have been an expression of neuromyotonia. They are discussed in Chapter 45, although the pedigree of Kurlan et al. (1987), in which some members clearly had exercise-induced dyskinesia, is included here.

Most patterns of choreic or dystonic movements can occur in idiopathic or symptomatic disorders. Precipitation of involuntary movements by sudden activity at one time seemed almost diagnostic of primary paroxysmal kinesigenic dyskinesia. It is now clear, however, that this feature can be symptomatic of a wide diversity of disorders, including head trauma, cerebral infarction, hypocalcaemia, hyperthyroidism, and progressive supranuclear palsy.

It has been claimed that prolonged spasms, with stiffening and abnormal posturing of limbs, which occurs in inbred immature Syrian hamsters may represent a model of paroxysmal dystonia (Löscher et al. 1989). Mechanisms underlying this behaviour are uncertain. It has been postulated that the GABAergic system might be involved and that delay in maturation of this might cause the disturbance (Löscher et al. 1989, Richter and Löscher 1993, Löscher et al. 1995). The movements occur both spontaneously and in response to environmental stimuli. The maximal severity occurs between 20 and 40 days and thereafter spontaneously resolves, except for a tendency to recur in pregnant or nursing females. The relationship of this animal model to human disorders, if any, is unknown. It has been suggested that it may be similar to the transient dystonic posturing that is seen in otherwise normal infants (Löscher et al. 1995) and which we have included in Chapter 36.

However, it has become clear that the paroxysmal movement disorders share common features with and therefore may have similar pathophysiology to other episodic neurological disorders, such as the episodic ataxias which are now known to be disorders of ion channels or channelopathies (Bhatia et al. 2000).

Gowers (1881) has generally been credited with the first description of movement-induced dyskinesia, although more recently it has been suggested that an Italian physiologist, Angelo Mosso, may have mentioned the disorder in 1884 (Martinelli and Gabellini 1991). Neither of Gowers’ two cases is completely typical of paroxysmal kinesigenic dyskinesia and they may be examples of symptomatic or secondary dyskinesias induced by movement. The first had ‘irregular fixation of the limbs without loss of consciousness’ precipitated by movements. These followed an illness in which there was ‘generalized powerlessness’ with subsequent episodes of generalized rigidity and loss of consciousness produced by passive spinal movements. The second case had ‘attacks of a very unusual character’ brought on by movement after ‘sitting still’. They lasted about a minute and were associated with scratching the face, tremor, dilated pupils, and post-ictal confusion. Subsequently attacks occurred without being provoked by movement.

Pitha (1938) provided the first definite description of paroxysmal kinesigenic dyskinesia in a medical student who suffered attacks from the age of 6 years. They were provoked by active or passive limb movements after a period of rest, or by startle. Pitha noted that movement of one limb resulted in tonic spasm on the same side lasting 10–30 seconds. Before 17 years of age attacks had been bilateral, but subsequently they were unilateral. An unusual feature was that occasional attacks were induced by watching or even reading about running races.

Shortly before Pitha's (1938) description the issue had been clouded by reports of brief focal tonic or clonic spasms induced by various stimuli, including movement, in patients who had other underlying neurological diseases (Spiller 1927, Wilson 1930, Strauss 1940). Further confusion occurred after Mount and Reback (1940) described a familial disorder in which prolonged attacks of chorea and athetosis were precipitated by certain foodstuffs, fatigue, and stress, but not by movement. The following year, under the title of ‘periodic dystonia’, Smith and Heersema (1941) reported three cases of typical paroxysmal kinesigenic dyskinesia, observing that it was familial in two. They noted improvement with barbiturates. Michaux and Granier (1945) were the first to report marked improvement of paroxysmal kinesigenic dyskinesia with phenytoin. Their patient's dystonic and clonic left-sided attacks were provoked by sudden movements of the left leg following inactivity. They noted startle or surprise produced attacks and felt this was secondary to sudden movement.

Pryles et al. (1952) noted an association between epilepsy and familial paroxysmal kinesigenic dyskinesia, although they confused this latter condition with the disorder described by Mount and Reback (1940).

In spite of these reports, paroxysmal movement disorders were not widely recognized and as late as 1957 Kishimoto described four pedigrees of paroxysmal kinesigenic dyskinesia and thought it was the first account of the disorder.

Subsequent reports continued to confuse familial paroxysmal kinesigenic dyskinesia with the disorder described by Mount and Reback (1940) and with symptomatic varieties of paroxysmal dyskinesia. In 1963 Lance described eight patients under the title of ‘sporadic and familial varieties of tonic seizures’. These comprised several disorders including paroxysmal kinesigenic dyskinesia and a family with the condition described by Mount and Reback (1940).

Kertesz (1967) introduced the term ‘kinesigenic’ to describe dyskinesias induced by movement. He clearly separated them from the non-kinesigenic attacks of Mount and Reback (1940) and Lance (1963). Kertesz described 10 patients with sporadic or familial paroxysmal kinesigenic dyskinesia. There is now an extensive literature on this disorder and well over 100 patients with the idiopathic form have been reported (Fahn 1994, Demirkirin and Jankovic 1995, Houser et al. 1999). In addition, kinesigenic dyskinesia can occur secondary to a variety of different diseases, including multiple sclerosis, head injury, stroke, cerebral palsy, hypoparathyroidism, and progressive supranuclear palsy [see Fahn (1994) for a list of published cases]. These symptomatic cases are discussed later in this chapter under the various disorders that cause them.

Paroxysmal kinesigenic dyskinesia is characterized by attacks of chorea, athetosis, or dystonia, which are provoked by sudden movement or startle, last less than 5 minutes, usually only seconds, and are unassociated with impairment of consciousness. Between attacks examination is usually normal. The essential features are that they are induced in this way and are brief. Many last only seconds and the majority are less than half a minute. Although there is a considerable variation in the type of involuntary movement that occurs (see later under ‘Clinical features’), the picture is otherwise remarkably stereotyped.

A variety of other names have been used by different authors to refer to this condition, including paroxysmal choreoathetosis (Stevens 1966, Tassinari and Fine 1969, Jung et al. 1973), familial paroxysmal chorea-athetosis or choreoathetosis (Pryles et al. 1952, Williams and Stevens 1963, Hudgins and Corbin 1966, Wagner et al. 1966), paroxysmal kinesigenic choreoathetosis (Kertesz 1967), periodic dystonia (Smith and Heersema 1941), conditionally responsive extrapyramidal syndrome (Kishimoto 1958), seizures induced by movement (Lishman et al. 1962, Whitty et al. 1964, Burger et al. 1972), movement epilepsy (De Bolt 1967), reflex epilepsy (Pitha 1938, Michaux and Granier 1945, Whity et al. 1964, Perez-Borja 1967), hereditary kinesthetic reflex epilepsy (Fukuama and Okada 1968), and the reflexogenic extrapyramidal syndrome (Kishimoto and Okada 1965). Not only do some papers confuse paroxysmal kinesigenic dyskinesia with the syndrome of Mount and Reback (1940) and with symptomatic (secondary) choreic athetotic or dystonic attacks, but a number also include reflex and non-reflex epilepsy. Sometimes there may be confusion between the kinesigenic variant with the form which is precipitated by prolonged exercise, but close enquiry usually reveals differences, such as induction of involuntary movements only after vigorous activity (Tabaee-Zadeh et al. 1972). Furthermore, exercise-induced paroxysmal dystonia (Lance 1977, Plant et al. 1984) differs in that dyskinesia persists for between 5 and 30 minutes.

There have only been two brief reports of autopsy findings in this disorder. Stevens (1966) mentions slight asymmetry of the substantia nigra pars compacta, with that on the left ‘being dorsal’ and that on the right being ‘medially located’. Kertesz (1967) records the ‘presence of some melanin pigment in macrophages in the area of the locus ceruleus’ in a 13-year-old boy and assumed this ‘suggested slight loss of neurons from this nucleus’. It seems unlikely that either of these minor abnormalities is significant.

There have been a few reports of radiological abnormalities suggesting pathology. Slight dilatation of lateral ventricles on pneumoencephalography was reported in a 35-year-old man (Kertesz 1967). CT brain scan in a 22-year-old sporadic case, who had experienced attacks for 8 years, was reported as showing ‘severe enlargement of the 4th ventricle and moderate enlargement of the circummesencephalic cisterns’ suggestive of ‘brain stem and possibly central cerebellar atrophy’ (Watson and Scott 1979). A 39-year-old sporadic case, with left-sided attacks since the age of 5 and hyper-reflexia in the left leg, had an ill-defined, non-enhancing mass effect in the right fronto-temporal region on CT scan, resulting in asymmetry and possible effacement of the right lateral ventricle (Gilroy 1982). Arteriography confirmed a mild avascular mass effect without midline displacement, and on pneumoencephalography the right lateral ventricle did not fill. Brain biopsy was normal and a follow-up CT scan 15 months later showed no change.

In spite of these abnormalities, the great majority of radiological studies have been normal (see later under ‘Investigations’). It seems likely that the above abnormalities may have been incidental findings or examples of symptomatic paroxysmal movement disorders.

There have been no studies of cerebral biochemistry in this disorder. Serum dopamine-beta-hydroxylase was reported slightly decreased in one patient but pyruvate, lactate, acetoacetate, beta-hydroxybutyrate, homovanillic acid, and 5-hydroxyindolacetic acid were normal in cerebrospinal fluid (CSF) (Busard et al. 1984). A proton magnetic resonance (MR) spectroscopy study looking at the basal ganglia, thalami, and supplementary motor areas in five patients with idiopathic paroxysmal kinesigenic dyskinesia has been reported (Kim et al. 1997). In this report significantly decreased peak ratios of Cho/Cr (choline/creatine) and Ml/Cr (myoinositol/creatine) were found in the basal ganglia contralateral to the side of attacks in three of five patients but not in matched controls. The authors suggested a dysfunction of the cholinergic system in the basal ganglia in paroxysmal kinesigenic dyskinesia, but the significance of these findings is not entirely clear.

Neurophysiological changes associated with the involuntary movements are largely unknown. The Contingent Negative Variation (CNV) is a slow cerebral potential measured over the scalp following a warning stimulus preparing the subject that a second stimulus will follow, which requires performance of a rapid movement. The amplitude of this potential has been reported to be increased in paroxysmal kinesigenic dyskinesia and to return to normal following successful treatment with phenytoin or valproic acid (Franssen et al. 1983, Busard et al. 1984). Contingent negative variation amplitude is increased in normals by anxiety. This electrophysiological abnormality in paroxysmal kinesigenic dyskinesia may thus just reflect an anxiety related change occurring prior to the involuntary movements, which are sometimes enhanced by stress and anticipation before the motion (Lishman et al. 1962, Kertesz 1967, Kato et al. 1969, Burger et al. 1972, Kinast et al. 1980).

Lee et al. (1999) measured the changes of forearm flexor H reflexes produced by conditioning radial nerve stimulation at delays of –2, 0, 2, 4, 7.5, 10, 25, and 75 ms in 10 patients with paroxysmal kinesigenic dyskinesia and six with generalized seizure disorder. The results were compared with 12 normal volunteers. Follow-up studies were done in eight paroxysmal kinesigenic dyskinesia patients after they responded to the anticonvulsant treatment. At each delay, patients with seizure disorders showed comparable results to controls. Patients with paroxysmal kinesigenic dyskinesia, however, showed paradoxical facilitation at a delay of 0 ms, enhanced facilitation between 2 and 7.5 ms delays, and attenuated inhibition at a delay of 75 ms. There were no significant differences in the amount of reciprocal inhibition according to the severity of clinical symptoms. Follow-up studies showed no significant changes in reciprocal inhibition compared to the baseline data. The authors concluded that in paroxysmal kinesigenic dyskinesia, paradoxical facilitation and enhanced first relative facilitation period may be caused by defective spinal interneurons. Also, in addition to the defective reciprocal inhibition, abnormalities of supraspinal inputs seemed to be involved in the genesis of paroxysmal kinesigenic dyskinesia (Lee et al. 1999).

Movement-related potential (Bereitschaftspotential) recorded in two patients with paroxysmal kinesigenic dyskinesia was said to be abnormal, with changes [reduced amplitude of early negativity (NS1) and relatively steep late negativity with delayed onset] similar to those seen in other extrapyramidal disorders (Houser et al. 1999).

Several pathophysiological mechanisms have been proposed to underlie paroxysmal kinesigenic dyskinesia. Pitha (1938) considered the possibility of hysteria, but rejected it. Nonetheless, a psychiatric basis is often suspected (Waller 1977), but there is no evidence supporting this.

Many early descriptions considered the disorder to be epileptic. Gowers’ (1881) first report of movement-induced ‘seizures’, included in his book on epilepsy, have sometimes been misinterpreted as referring to paroxysmal kinesigenic dyskinesia.

Confusion between paroxysmal kinesigenic dyskinesia and symptomatic movement-induced epilepsy fostered this notion. Demonstration that symptomatic attacks could be relieved by excision of a cortical scar was a powerful support for an epileptic basis (Falconer et al. 1963). Features in favour of paroxysmal kinesigenic dyskinesia being epileptic include 1) the paroxysmal nature of attacks, 2) the sensory aura, 3) association with other varieties of epilepsy in some patients and relatives, 4) paroxysmal encephalographic abnormalities in some patients, and 6) abolition of attacks with anticonvulsants. Arguments against an epileptic basis include 1) preservation of consciousness and lack of tongue biting or incontinence during generalized attacks, 2) failure of progression from dyskinesia to grand mal seizure during attacks, 3) the infrequent occurrence of definite epilepsy in patients and relatives, and 4) lack of paroxysmal electroencephalographic (EEG) abnormalities during attacks (these features are reviewed below under ‘Clinical features’ and ‘Investigations’).

Smith and Heersema (1941) were the first to suggest that a basal ganglia abnormality might underlie the disorder. This possibility has subsequently been examined by many authors (Lishman 1962, Whity et al. 1964, Lance 1977, Goodenough et al. 1978). Factors in favour of this interpretation include 1) the choreic and dystonic nature of the movements, 2) occasional persistence of chorea between attacks, 3) reported improvement with levodopa (Loong and Ong 1973), 4) reported exacerbation of symptoms by the anticholinergic drug diphenhydramine (Perez-Borja et al. 1967), 5) the possibility that response to phenytoin could be via its dopamine-blocking action (Chadwick et al. 1976, Elliott et al. 1977), and 6) abnormalities of contingent negative variation (CNV) and Bereitschaftspotential (Houser et al. 1999) as mentioned earlier. Indirect support for basal ganglia involvement is derived from those features, mentioned above, that are against an epileptic basis. Results of electrical stimulation can be used to support both possibilities. Stimulation of the supplementary motor cortex can produce a sensory aura and tonic contralateral posturing, similar to that seen in some patients (Penfield and Jasper 1954). Stimulation of the oral part of the ventral lateral thalamic nuclei produces slow torsion movements of the head and contralateral upper limb, similar to some attacks (Hassler et al. 1960).

Two studies have investigated patients with idiopathic paroxysmal kinesigenic dyskinesia using transcranial magnetic stimulation (Mir et al. 2005, Kang et al. 2006). In one of these a proportion of patients were studied on and off treatment. Mir et al. (2005) found reduced inhibitory mechanisms including decreased short intracortical inhibition, a reduced early phase of transcallosal inhibition, and a reduced first phase of spinal reciprocal inhibition. The cortical silent period, the startle response, and the second and third phases of reciprocal inhibition were normal. The abnormalities in transcallosal inhibition were normalized by carbamazepine treatment, while the other parameters did not change with medication. The authors concluded that the abnormalities in cortical and spinal inhibitory circuits may be useful to differentiate paroxysmal kinesigenic dyskinesia from primary dystonia and also from epilepsy which show a different pattern of electrophysiological responses. Kang et al. (2006) reported normal measures for thresholds, intracortical facilitation, and silent period. In contrast to Mir's study, they found normal short intracortical inhibition.

Various authors have speculated on basal ganglia abnormalities which might cause paroxysmal movements. Impaired maturation of extrapyramidal systems has been frequently mentioned (Kertesz 1967, Loong and Ong 1973, Chadwick 1976, Homan et al. 1980, Busard 1984). Faulty cortical control of the neostriatum and thalamus (Busard et al. 1984) and pathological oscillations in ‘feedback’ circuits in the basal ganglia (Kertesz 1967) are but two postulated mechanisms. None is supported by experimental evidence.

The two main theories regarding causation of paroxysmal kinesigenic dyskinesia are thus reflex movement-induced epilepsy and disturbed basal ganglia function. These hypotheses are not mutually exclusive and, although normal EEG and retained consciousness during generalized seizures excludes widespread cortical involvement, subcortical seizures remain possible. Paroxysmal disturbance of neurotransmitter or receptor function in the basal ganglia is an intriguing possibility but at present without experimental support.

A unifying hypothesis comes from the observation that the paroxysmal dyskinesias share many similarities to other episodic neurological disorders, many of which are known to be due to genetic disorders of ion channels or channelopathies (Bhatia et al. 2000) and thus may have the same pathophysiology. In this context it is interesting to observe the similarities between paroxysmal kinesigenic dyskinesia and episodic ataxia type 1 (EA1), which is known to be caused by mutations of the potassium channel KCNA1 gene (Browne et al. 1994). Both conditions begin early in life, and both tend to abate in adulthood. Like paroxysmal kinesigenic dyskinesia, episodes of ataxia in patients with EA1 can be provoked by kinesigenic stimuli and tend to be brief and frequent (Bryne et al. 1991). Although EA1 typically responds to acetazolamide, like paroxysmal kinesigenic dyskinesia anticonvulsants may reduce EA1 attacks in some patients and also help the interictal myokymia seen in this disorder (Griggs et al. 1978, Brunt et al. 1990).

There is also overlap with other periodic disorders. Clinically periodic paralysis, and other episodic disorders including migraine and epilepsy syndromes share the common feature of episodic attacks on a normal interictal background and many have similar precipitating factors as well as treatment, as do the episodic ataxias, and the paroxysmal dyskinesias (Bhatia et al. 2000). Carbamazepine, for example, prevents epileptic seizures and is also very effective in patients with paroxysmal kinesigenic dyskinesia. Acetazolamide is helpful not only for patients with periodic paralysis and myotonia, but also, as mentioned above, episodic ataxias (Griggs et al. 1978), and some paroxysmal dyskinesias. There are also reports of families with multiple episodic disorders, for example paroxysmal dyskinesia in a family with episodic ataxia, and association of episodic problems like migraine and epilepsy in families with paroxysmal dyskinesias (Munchau et al. 2000). This suggests that the idiopathic paroxysmal dyskinesias may also be channelopathies.

In addition to numerous single case reports, there are two studies reviewing clinical features in larger cohorts: the study by Houser et al. (1999) on 26 patients with paroxysmal kinesigenic dyskinesia and the more recent study by Bruno et al. (2004) who reviewed features of 121 affected individuals with a presumptive diagnosis of idiopathic paroxysmal kinesigenic dyskinesia that were referred for genetic studies. On this basis the latter authors also proposed new diagnostic criteria.

Paroxysmal kinesigenic dyskinesia has occurred world-wide and although increased incidence among Orientals has been suggested (Kinast et al. 1980, Fahn 1994), there is no established racial predilection. The incidence is unknown but by inference may be in the order of 1:100,000,000 (Lishman et al. 1962). Approximately 70% of reported cases have been familial and 30% sporadic (Lance 1977). The majority of familial cases show autosomal dominant inheritance (Kishimoto 1958, Hudgins and Corbin 1962, Whitty et al. 1964, Stevens 1966, Fukuyama and Okada 1968, Kato and Araki 1969, Houser et al. 1999). Involvement of siblings without apparent abnormality with parents is compatible with autosomal dominant inheritance with incomplete penetrance or autosomal recessive inheritance (Smith and Heersema 1941, Pryles et al. 1952, Whity et al. 1964, Kertesz 1967, Jung et al. 1973). Although autosomal dominant inheritance is most common, other factors must be involved in the expression of the trait as males are affected four times as commonly as females in both familial and sporadic varieties (Lance 1977). This is reflected in the report of 26 idiopathic paroxysmal kinesigenic dyskinesia cases by Houser et al. (1999) of whom 23 were male. In the majority of cases symptoms commence in childhood or adolescence. Reported familial cases have commenced between 5 and 15 years of age. There is more variation in sporadic cases and onset has ranged from 6 months (Perez-Borja et al. 1967) and 1 year (Williams and Stevens 1963) to at least 33 years (Hudgins and Corbin 1966). This may reflect contamination of the sporadic group by symptomatic cases. The mean age of onset in the series of 26 idiopathic cases reported by Houser et al. (1999) was 13.6 years.

The frequency of attacks is extremely variable and can range from one every few months or less to more than 100 per day (Kato and Araki 1969, Jung et al. 1973). Reports suggest 95% of patients experience at least daily attacks at some stage (Goodenough et al. 1978); 57% of patients of the 26 cases reported by Houser et al. (1999) had between 1 and 40 attacks a day.

Paroxysms are usually precipitated by voluntary movement. Most commonly, this is preceded by a period of inactivity, such as rising swiftly from a chair or starting to walk after standing still (Lance 1977, Goodenough et al. 1978). Occasionally changing the direction or speed of movement may induce attacks (Hishikawa et al. 1973). Anticipation of performing an act or anxiety may increase the chances of a paroxysm when the act is initiated (Goodenough et al. 1978). In some (Pitha 1938, Plant 1983) but not all (Kertesz 1967) patients, attacks can be provoked by passive limb movements. Precipitating movements usually involve the legs or legs plus arms. It is uncertain how often arm movement alone induces attacks (Hishikawa et al. 1973). Swimming has occasionally been provocative (Kertesz 1976). In virtually all cases, however, movement has been the precipitant (Houser et al. 1999). An exception was Pitha's (1938) patient who had attacks induced by watching or reading about athletic events. Startle occasionally produces attacks (Houser et al. 1999), but some have regarded these as being due to an accompanying small jump or muscle stiffness (Stevens 1966, Kertesz 1967). Hyperventilation infrequently induces a bout (Williams and Stevens 1963, Stevens 1966), but in most patients this manoeuvre is ineffective (Kertesz et al. 1967). Both improvement (Gilroy 1982) and aggravation (Homan et al. 1980) of paroxysms by alcohol have been recorded. Occasionally attacks seem to be spontaneous. The majority of these are probably symptomatic of other neurological diseases. Occasionally, however, they seem to be due to paroxysmal kinesigenic dyskinesia [case 4 in Kinast et al. (1980)].

Many patients learn manoeuvres to prevent or suppress attacks. Thus, avoiding sitting or standing still by continually shifting weight (Hishikawa et al. 1973), making small ‘warming up’ movements before a major action, and starting off slowly may prevent attacks (Goodenough et al. 1978). If a paroxysm is just starting it may be suppressed by stopping, tensing muscles, or other manoeuvres, such as flexing or holding the affected limb (Lishman et al. 1962, Whity et al. 1964, Kertesz 1967). Attacks occurring during sleep must be extremely rare and some authors have claimed that they never occur (Hishikawa et al. 1973). Smith and Heersema (1941), however, reported that paroxysms occasionally occurred at night and a patient of Stevens (1966) had attacks if turning briskly during sleep.

Some authors have claimed nearly all patients experience an aura like numbness or ‘pins and needles’ in the affected limb or the epigastric region prior to the dyskinesia (Goodenough et al. 1978) and 69% of patients reported by Houser et al. (1999) and 63% in the series by Bruno et al. (2004) had an aura. In a review of the literature Plant (1983) found an aura mentioned in approximately half the cases. Thus, focal or unilateral attacks may be preceded by numbness, tingling, warmth, tightness, ‘crawling’, or similar sensations. They are usually felt briefly in a limb, often the leg, and then rapidly spread. Focal and generalized attacks may be preceded by a non-localized aura (Gilroy 1982, Plant 1983). These consist of a strange awareness of an impending attack and other indescribable feelings. Focal auras do not precede generalized attacks (Plant 1983). The aura lasts only a second or so and is immediately followed by the abnormal movements. These usually occur within seconds of the precipitating event, but rarely a delay of several minutes has been noted (Homan et al. 1980). Jung et al. (1973) reported an unusual case with a prodrome lasting hours or days, consisting of a peculiar oral sensation, which progressed into an aura of ‘mild stiffening or twitching of the tongue’.

Mild attacks may consist of an aura associated with a brief interruption of normal movement (Lishman et al. 1962, Kertesz 1967, Goodenough et al. 1983, Plant 1983). Minor movement of digits or dystonic posturing of an extremity may occur. Larger attacks may involve a limb, be unilateral, or generalized (Fig. 49.1). The character of movements shows considerable variation between individual attacks, and more particularly between patients. Occasionally it is tonic spasm with little movement. More commonly it is dystonic or athetotic. Although almost any direction of movement occurs, a commonly described position is extension at the hip and knee, plantar flexion and inversion of the foot, adduction at the shoulder, flexion at the elbow and wrist, pronation of the forearm, and either flexion or extension of fingers. This posture often alternates with its opposite, resulting in sinuous dystonic movements (Kertesz 1967, Plant 1983). At times movements may be choreic and patients with generalized attacks have frequently been described as looking as though they had advanced Huntington's disease. The character of movements has been described as choreoathetotic or athetotic in approximately 65% of reports and tonic in 20% (Goodenough et al. 1978). Similarly, Houser et al. (1999) reported that 57% of their cases had primarily dystonic movements, while choreodystonic movements were noted in about 30%. A few patients experience ballismus (Hishikawa et al. 1973, Plant 1983) or clonic jerking (Plant 1983). The trunk may flex or rotate and there may be torticollis, retrocollis, or inability to turn the head (Kertesz 1967, Hisikawa et al. 1973, Jung et al. 1973, Franssen et al. 1983). Facial movements are very common, involving the ipsilateral side in unilateral attacks. Forced deviation of eyes towards the affected side has rarely been reported (Jung et al. 1973). There may be blepharospasm, retraction of the eyelids, baring of the teeth, and grimacing. Protrusion or lateral deviation of the jaw and trismus occur (Hishikawa et al. 1973, Franssen et al. 1983, Plant 1983). The tongue may be moved in the mouth or protruded. A sensation of lingual enlargement has been reported (Kertesz 1967). Dysarthria is common and some patients are incomprehensible (Kertesz 1967, Hishikawa 1973, Garello et al. 1983, Houser et al. 1999). Forced laughter has been reported in one case (Whitty et al. 1964).

Plant (1983) noted only 16% of patients had both unilateral and bilateral attacks. In 30% these were only bilateral and in 50% they were only unilateral, although in some the side varied from attack to attack (Table 49.1). Unilateral or focal attacks can sometimes be induced by moving a limb on that side. Thus, tapping, twisting, shaking, or hopping on one foot may induce a paroxysm in that leg or over the whole of that side (Pitha 1938, Smith and Heersema 1941). Occasionally movements of the leg may induce an attack restricted to the ipsilateral arm (Plant 1983). In some patients attacks can be induced independently on either side by such manoeuvres, but in others generalized paroxysms result (Goodenough et al. 1983). Occasionally involvement in spontaneous attacks is ‘crossed’ such that an arm and leg are involved on opposite sides

(Kertesz 1967), but this has not been recorded in attacks induced by unilateral manoeuvres. Many authors have noted a refractory period following an attack, during which further paroxysms cannot be induced. This varies from about 3 to 15 minutes (Lishman et al. 1962, Stevens 1966, Watson and Scott 1979, Franssen et al. 1983, Goodenough et al. 1983, Plant 1983). Following a unilateral seizure only that side is refractory and a further bout may be induced on the opposite side immediately (Plant 1983).

Approximately 80% of attacks are less than a minute in duration and all are less than 5 minutes (Goodenough et al. 1968, Lance 1977). Many paroxysms last only seconds and two thirds of the 26 cases of Houser et al. (1999) had attacks lasting between 30 and 60 seconds. In only three of their patients did the attacks last more than 2 minutes.

Attacks are painless, consciousness is preserved, incontinence or tongue biting do not occur, and there is no postictal confusion (Hishikawa et al. 1973). One patient reported by Houser et al. (1999) reported unilateral headaches and occasional numbness of the same arm and leg for about 2 hours after an attack.

Formal neurological examination during an attack is virtually impossible. Only occasional reports mention neurological findings immediately after a paroxysm. Lishman et al. (1962) found generalized hyper-reflexia and extensor plantar responses which persisted for several minutes. Others, however, (Burger et al. 1972) found no abnormality.

Several authors have noted a tendency for severity to increase in adolescence and subside in adult life (Kishimoto 1957, Goodenough et al. 1968, Hishikawa et al. 1973, Jung et al. 1973, Houser et al. 1999). Occasional patients may experience years of freedom between attacks (Gilroy 1982) and occasionally the paroxysms cease (Stevens 1966, Kertesz 1967, Hishikawa et al. 1973, Homan et al. 1980, Busard et al. 1984).

In the great majority of cases neurological examination between attacks is normal. Kertesz (1967) noted one patient was hyperactive and a few authors have reported mild, persistent chorea (Hudgins and Corbin 1966, Stevens 1966, Perez-Borja et al. 1967, Bird et al. 1978, Homan et al. 1980). In one family paroxysmal kinesigenic dyskinesia occurred in some patients and persistent chorea in others. An association with benign hereditary chorea has been suggested (Perez-Borja 1967, Bird et al. 1978, Homan et al. 1980). Occasional patients have been reported with hemiatrophy, hemiparesis, and hyper-reflexia on the side of the dyskinesia (Hishikawa et al. 1973, Kinast et al. 1980, Gilroy 1982). It seems likely that involuntary movements in such cases are secondary to focal intracranial disease and are not examples of idiopathic paroxysmal kinesigenic dyskinesia. Paroxysmal kinesigenic dyskinesia has been noted to develop in a patient with essential tremor (Nair et al. 1991).

Hishikawa et al. (1973) noted almost 15% of reported cases had at least one generalized convulsion. Many, however, may have been febrile seizures and only 6% had convulsions over the age of 5 years. Nonetheless, in some pedigrees association of paroxysmal kinesigenic dyskinesia with epilepsy seems marked and Homan et al. (1980) reported a family in which three siblings had both conditions. Tonic clonic (Witty et al. 1964, Hudgins and Corbin 1966, Jung et al. 1973), absence (Pryles et al. 1952), and possible temporal lobe attacks (Hudkins and Corbin 1966) have all been reported. A family history of epilepsy has been recorded in almost 30% of familial cases (Goodenough et al. 1978). An association between infantile convulsions and paroxysmal dyskinesias, which from descriptions resemble paroxysmal kinesigenic dyskinesia, has been reported in some families (Szepetowski et al. 1997, Lee et al. 1998). This was given the eponym ICCA (infantile convulsions choreoathetosis) syndrome by the authors. These families have been linked to a locus in the pericentromic region of the short arm of chromosome 16 (see later under ‘Genetics’).

The prognosis is overall good. Attack frequency decreases with age and patients have a normal life expectancy. Although the attacks are not in themselves harmful, they can, however, result in injury and endanger life (Smith and Heersema 1941). Most patients find them embarrassing and they can cause social disadvantage. Suicide has been reported (Kertesz 1967). During pregnancy, 50% of affected women may note improvement (Bruno et al. 2004).

Routine haematology, plasma biochemistry, and CSF analysis are normal. Negative results of more elaborate biochemical testing are mentioned above under ‘Biochemistry’. Plain skull radiology and radioisotopic brain scanning are normal (Kertesz 1967, Kinast et al. 1980). CT and MRI brain scans are normal in the majority of cases (Goodenough et al. 1978, Kinast et al. 1980, Suber and Riley 1980, Bortolotti and Schoenhuber 1983, Franssen et al. 1983, Busard et al. 1984). Likewise arteriography and pneumoencephalography have usually been unremarkable (Lishman et al. 1962, Witty et al. 1964, Kertesz 1967). Exceptions are mentioned above under ‘Anatomical pathology’. Functional imaging studies using single photon emission computed tomography (SPECT) have shown altered perfusion of the basal ganglia contralaterally to the affected side or bilaterally supporting the basal ganglia theory (Ko et al. 2001, Joo et al. 2005).

In the majority of patients EEGs recorded between attacks are normal. In a review of published familial cases, however, Goodenough et al. (1978) noted 24% of EEGs were abnormal but non-epileptiform, and a further 10% were suggestive of epilepsy. These figures are similar to a review carried out by Hishikawa et al. (1973) which included sporadic cases. They found generalized or localized spikes and sharp waves in approximately 10% of patients, slow wave abnormalities in a further 30%, while the remainder had normal recordings.

EEGs performed during attacks usually show no change in the baseline activity (Hishikawa et al. 1973, Goodenough et al. 1978). Occasionally interictal abnormalities have disappeared or diminished (Witty et al. 1964, Kertesz 1967). One patient showed rhythmic 12–15 Hz activity over the vertex and right hemisphere, which commenced at onset and disappeared before the end of the attack, but this was not clearly differentiated from artefact (Perez-Borja et al. 1967). Another had generalized rhythmic 5 Hz discharges (Hirata et al. 1991). Invasive long-term monitoring has been reported to show ictal discharge in the supplementary motor cortex and ipsilateral caudate in a young girl, without significant spread to other neocortical areas (Lombroso 1995). The significance of this observation to primary paroxysmal kinesigenic dyskinesia is uncertain.

Visual, somatosensory, and brainstem auditory evoked potentials have been normal, as have electrophysiologically recorded blink responses (Bortolotti and Schoenhuper 1983, Franssen et al. 1983).

Abnormalities in contingent negative variation, Bereitschaftspotential, and transcranial magnetic stimulation responses have been mentioned above under ‘Neurophysiology’.

Szepetowski and colleagues (1998) described four French families with infantile convulsions and paroxysmal dyskinesias which by description resembled paroxysmal kinesigenic dyskinesia and called this the ‘ICCA syndrome’ (see earlier under ‘Clinical features’). This was linked to the pericentromeric region of chromosome 16. Linkage to the same locus was further confirmed in a Chinese family with a similar syndrome (Lee et al. 1998). The clinical characteristics of some of the paroxysmal dyskinetic episodes in families with the ICCA syndrome were very similar to those described for paroxysmal kinesigenic dyskinesia given their brevity, frequency, and onset by kinesigenic stimuli in some episodes. It was therefore not surprising that in subsequent reports of eight Japanese families (Tomita et al. 1999) and a three-generation African-American kindred (Bennet et al. 2000) with typical paroxysmal kinesigenic dyskinesia attacks, the disease locus for paroxysmal kinesigenic dyskinesia was also mapped by linkage analysis to the pericentromeric region of chromosome 16. The paroxysmal kinesigenic dyskinesia region in the Japanese families spans 12.4 cM and overlaps by 6.0 cM the ICCA region. As there was an increased prevalence of afebrile infantile convulsions in the Japanese families with paroxysmal kinesigenic dyskinesia, it has been postulated that one gene may be responsible for both paroxysmal kinesigenic dyskinesia and ICCA. However, the paroxysmal kinesigenic dyskinesia interval identified in the African-American family in which individuals have paroxysmal kinesigenic dyskinesia alone (and no infantile seizures) overlaps by 3.4 cM with the ICCA region and by 9.8 cM with the paroxysmal kinesigenic dyskinesia region identified in Japanese families. Thus, at the moment it is unclear whether there are two genes or a single gene in this interval which could give rise to both ICCA and paroxysmal kinesigenic dyskinesia in these families.

A family with paroxysmal kinesigenic dyskinesia from India has been linked to a second locus on the long arm of chromosome 16 distinct from the locus of the Japanese families with paroxysmal kinesigenic dyskinesia (Valente et al. 2000). This suggests that there may be a family of genes causing paroxysmal disorders on the pericentromeric region of chromosome 16 (Valente et al. 2000).

The gene/genes causing these disorders are yet unknown but there are a group of ion channel genes which lie in this pericentromeric region of chromosome 16, within the ICCA and paroxysmal kinesigenic dyskinesia intervals, and represent excellent candidates for these paroxysmal disorders. However, a detailed analysis of 157 genes in this area revealed no mutations that were shared by unrelated families (Kikuchi et al. 2007).Two non-synonymous substitions affecting the SCNN1G and ITGAL gene, which were segregated with disease in two families, have been proposed to play a role (Kikuchi et al. 2007). This needs further investigation.

Also see Table 49.2 for an overview.

A variety of drugs have been used in treatment, but there are no controlled trials. Anticonvulsants are the most useful. Carbamazepine and phenytoin are the drugs of choice (Kato and Araki 1969,  Kinast et al. 1980, Gilroy 1982, Carrilho et al. 1994). Eighty six percent of patients in the study by Bruno et al. (2004) responded well to either carbamazepine or phenytoin. As with phenytoin, small doses of carbamazepine (100–200 mg daily) are sufficient (Hishikawa et al. 1973, Houser et al. 1999). Acetazolamide is also a useful alternative or adjunct to carbamazepine in the treatment of paroxysmal kinesigenic dyskinesia, especially when due to demyelinating lesions.

Phenytoin also abolishes or substantially reduces attacks (Kertesz 1967, Goodenough et al. 1978, Kinast et al. 1980, Gilroy 1982, Plant 1983). Occasionally, however, it is not helpful (Smith and Heersema 1941). Frequently the dose required to suppress attacks is smaller than that needed for epilepsy. It has been suggested that adults are more responsive than children and and lower dosage is sufficient. Plasma concentrations as low as 2–4 µg/ml may be effective in the former (Homan et al. 1980). Houser et al. suggested doses of 5 mg/kg/day for phenytoin for adults. Most studies have reported using between 100 and 300 mg of phenytoin daily. As mentioned above under ‘Pathophysiological mechanisms’, the anticholinergic or dopamine-blocking effect of phenytoin, rather than its anticonvulsant action, could underlie the response and possibly explain the unusual sensitivity to this drug.

A few authors have also reported that phenobarbitone and primidone improved attacks (Smith and Heersema 1941, Lishman et al. 1962, Hishikawa et al. 1973, Goodenough 1978, Gilroy 1982, Garello et al. 1983, Busard et al. 1984). Exacerbation with primidone has been described (Kertesz 1967) but definite confirmation is lacking. Sodium valproate has been used less frequently and has been claimed to be both effective (Suber and Riley 1980, Busard 1984) and non-effective (Garello et al. 1983).

Generally benzodiazepines seem less useful, although they have been found beneficial in patients with HIV-associated forms. Diazepam has been reported as being partially effective (Gilroy 1982) and non-effective (Kinast 1980, Garello et al. 1983). Chlordiazepoxide may occasionally be useful (Perez-Borja et al. 1967).

In contrast to the widely reported effectiveness of phenytoin and carbamazepine, the effect of drugs that primarily modulate dopaminergic and cholinergic function is less certain. Loong and Ong (1973) reported that l-dopa was useful in suppressing attacks in a single non-familial case, whereas Garello et al. (1983) and Busard et al. (1984) found no improvement. Amphetamines, which release dopamine and noradrenaline stores, have been noted both to be unhelpful (Smith and Heersema 1941) and to improve attacks (Pryles 1952). Haloperidol has been reported ineffective (Garello et al. 1983), but tiapride, which also weakly blocks dopamine receptors, improved attacks in a single patient (Busard et al. 1984).

Scopolamine was reported to be effective by Pitha (1938) and Lishman et al. (1962). Other workers, however, have not been able to confirm this effect. Smith and Heersema (1941) found hyoscine to be unhelpful. Flunarizine was said to virtually abolish attacks in a child with what was suggested to be paroxysmal dystonic dyskinesia of the type originally described by Mount and Reback (1940), but the description suggests it was really primary paroxysmal kinesigenic dyskinesia (Lou 1989).

Because of its relatively benign nature and the excellent response to anticonvulsants, surgery has little, if any, part to play in the management of this disorder. Hirota et al. (1971) reported that stereotactic ventrolateral thalamotomy was effective in abolishing movements in three cases who did not respond satisfactorily to phenytoin.

By using manoeuvres mentioned above under ‘Clinical features’, it may be possible to avoid precipitating attacks or to suppress them before a full-blown paroxysm has developed. Occasionally attacks are infrequent. In such patients drug therapy may be avoided. If attacks are uncontrolled, dangerous situations such as swimming and being at heights may need to be restricted. It is usually not necessary to be as rigorous with restrictions as in uncontrolled epilepsy.

In 1940 Mount and Reback described a patient with choreic and athetotic movements precipitated by fatigue, concentration, and several foodstuffs, including coffee and alcohol. Attacks were prolonged, lasting between 5 minutes and 2 hours. Twenty-six relatives spanning four generations had been similarly affected. In 1961 Forssman reported a similar large pedigree with cases spanning six generations. Provoking factors were similar to those reported by Mount and Reback and attacks lasted up to 4 hours. The movements, however, were somewhat different, with painful dystonia predominating.

Lance (1963) reported a family in which six individuals were affected over three generations. Similar prolonged dystonic spasms were induced by alcohol, stress, fatigue, and sometimes relaxing. Observations on this family have subsequently been extended (Lance 1977). Richards and Barnett (1968) recorded a family in which four generations had prolonged choreic and athetotic movements similar to those described by Mount and Reback (1940). Richards and Barnett (1968) were the first to recognize that similar pedigrees had been described, although they appeared unaware of Forssman's (1961) account. They felt the character of movements placed their family with that of Mount and Reback (1940), but distinguished it from that of Lance (1963). Like Kertesz (1967), they separated these disorders from those in which dyskinesia was provoked by voluntary movement (kinesigenic). Subsequently, similar families with dystonic, athetotic, and choreic movements have been reported (Tibbles and Barnes 1980, Walker 1981, Przuntek and Monninger 1983, Jacome and Risko 1984). Sporadic cases have been recognized (Kinast et al. 1980, Dunn 1981, Fahn and Bressman 1983, Bressman et al. 1988, Demirkiran and Jankovic 1995) and similar attacks have been described secondary to a variety of neurological disorders. These symptomatic paroxysmal dystonias are not discussed further here but are described later in this chapter under the different diseases that cause them. They include multiple sclerosis, cerebral palsy, head injury, transient ischaemic attack, stroke, encephalitis, hypoglycaemia, diabetes, thyrotoxicosis, hypocalcaemia, basal ganglia calcification, AIDS, and psychiatric disturbance [see Fahn (1994) for a list of published cases].

Primary paroxysmal non-kinesigenic dyskinesia is characterized by acute episodes of dyskinesia which are principally dystonic in nature. The familial variety shows 1) episodic attacks of dystonia, athetosis, and chorea, which last between 5 minutes and several hours, and 2) that these tend to be precipitated by certain foodstuffs, lack of sleep, fatigue, and emotional stress. They are not provoked by sudden movements or sustained exercise. In some families attacks are mainly dystonic (Foresman 1961, Lance 1977), while in others they are mainly choreic and athetotic (Mount and Reback 1941, Richards and Barnett 1968). This has led some authors to classify these two groups separately. The existence of families and cases displaying all of these movements, as well as common factors such as precipitation, duration of attacks, and response to therapy, justifies putting these groups together.

The disorder has been referred to by a number of different names including ‘hereditary intermittent tetanus’ (Forssman 1961), ‘familial tonic seizures’ (Lance 1963), ‘familial paroxysmal choreoathetosis’ (Mount and Reback 1941), ‘familial paroxysmal dystonic choreoathetosis’ (Lance 1977), and ‘paroxysmal non-kinesigenic dyskinesia’ (Goodenough et al. 1978). Kurlan et al. (1987) described a similar family with dystonic and choreic attacks which were precipitated by prolonged exercise and cold. Although they compared these cases with the disorder described by Mount and Reback (1940) and in their original description they did not mention an exercise-induced element (Kurlan and Shoulson 1983), we have chosen to include them under ‘paroxysmal exercise-induced dyskinesia’ because of the nature of the triggering factors.

There is no known anatomical pathology associated with this disorder. No macroscopic abnormality was noted in the brain of one adult and no relevant macroscopic or microscopic neuropathology was found in a child who died a ‘crib death’ (Lance 1977). Pneumoencephalography or CT brain scan have been reported to show slight enlargement of lateral and third ventricles in one patient and minor irregularity of the roof of one lateral ventricle in another (Lance 1977). SPECT imaging revealed hyperperfusion of the right caudate and thalamus (del Carmen Garcia et al. 2000). 18F-dopa and 11C-raclopride positron emission tomography (PET) demonstrated reduced density of presynaptic dopa decarboxylase activity in the striatum and increased density of postsynaptic dopamine D2 receptors and this was interpreted as chronic upregulation of postsynaptic dopa receptors (Lombroso and Fischman 1999). 18-fluor-desoxyglucose and 11C-dihydrotetrabenazine (DTBZ) PET did not show any metabolic abnormalities or abnormal binding (Bohnen et al. 1999). It has been suggested that dopaminergic abnormalities, if present, may be due to altered regulation of dopamine release or to postsynaptic mechanisms, rather than to an altered density of nigrostriatal innervation.

Two separate groups identified a gene locus for paroxysmal non-kinesigenic dyskinesia on chromosome 2q (Fink et al. 1996, Fouad et al. 1996). Fink and co-workers performed a genome-wide search in a large American kindred of Polish descent with 28 affected members and mapped the disorder to chromosome 2q33-q35 (Fink et al. 1996). In a five-generation Italian family with 20 affected members, Fouad and co-workers also showed tight linkage between the condition and microsatellite markers on distal 2q (2q31-q36) (Fouad et al., 1996). The smallest region of overlap of the candidate intervals identified by the two groups then placed the locus in a 6-cM interval. In a six-generation British family, Jarman and co-workers (1997) confirmed linkage to distal chromosome 2q and narrowed the candidate region to a 4-cM interval. Linkage to the same genetic location, designated FPD1 (familial paroxysmal dyskinesia type 1) and later also the DYT 8 locus (see Table 35.8), was further confirmed by Hofele et al. (1996) in a German family originally described by Przuntek and Monninger (1983) as classical of the Mount and Reback type of paroxysmal dyskinesia. Also linked to the same locus are among others two other typical families – one North American and one of German descent (Raskind et al. 1998) – as well as a Japanese family (Matsuo et al. 1999).

The causative gene was subsequently identified by Rainier and co-workers (2004) who detected that a missense mutation in the myofibrillogenesis regulator (MR-1) gene with substitution of valine for alanine at amino positions 7 and 9 resulted in alteration of the amino-terminal alpha helix in the two unrelated kindreds. These findings have since been confirmed by others (Hempelmann et al. 2006, Stefanova et al. 2006). Lee et al. (2004) could show by functional studies that there are two MR isoforms, MR-1L and MR-1S, of which the former is exclusively expressed in the cell membrane of the brain, while the latter is ubiquitously expressed and shows diffuse cytoplasmic and nuclear localization. Functionally, they may be involved in a stress-related pathway. The homologue, the hydroxyacylglutathione hydrolase (HAGH) gene, plays a role in a pathway to detoxify methylglyoxal, a compound present in coffee and alcoholic beverages and produced as a by-product of oxidative stress, thus suggesting a possible mechanism whereby alcohol, coffee, and stress precipitate attacks in paroxysmal non-kinesigenic dyskinesia (see Fig. 49.2 for metabolic pathway).

Richards and Barnett (1968) considered the involuntary movements might have a psychological basis. They found, however, that intravenous sodium amytal did not abolish movements until the patient was anaesthetized and that they returned before the patient regained consciousness, thus excluding hysteria.

Mount and Reback (1940) postulated that as alcohol produced turns, they might be due to the ‘release of lower motor centres from cortical control’. Similarly Lance (1963) suggested a ‘temporary release of basal ganglia and reticular postural mechanisms from cortical control’. He postulated the disorder might be akin to migraine and caused by cerebral vasospasm or ‘spreading depression’.

Byne et al. (1991) reported one pedigree in which some members have also had myokymia, and attention has been drawn to the fact that periodic ataxia with myokymia is associated with abnormality of potassium-channel KCNA₁ (Fink et al. 1996).

The discovery that at least in some families there is tight linkage between the disorder and microsatellite markers on chromosome 2q 33-35, which is close to a cluster of sodium channel genes, has led to speculation that primary paroxysmal dystonic dyskinesia may be a so-called channelopathy (Fink et al. 1996, Fouad et al. 1996).

Based on the more recent discovery of the gene locus on 2q Fink et al. (1997) suggested a model of paroxysmal non-kinesigenic dyskinesia pathophysiology from analyzing neurophysiologic effects of alcohol and caffeine (which provoke attacks of paroxysmal non-kinesigenic dyskinesia), the variably beneficial effects of levodopa-carbidopa, and the occurrence of dystonia and paroxysmal dyskinesia in biopterin synthesis disorders. They proposed that nigrostriatal neurons in paroxysmal non-kinesigenic dyskinesia patients have either marginally deficient dopamine synthesis or excessive alcohol- and caffeine-induced dopamine release; and that following alcohol- and caffeine-induced dopamine release patients experience a period of dopamine deficiency. The same group used PET with [¹¹C]dihydrotetrabenazine (DTBZ) to study striatal dopaminergic innervation in paroxysmal non-kinesigenic dyskinesia (Bohnen et al. 1999). The results did not reveal abnormal DTBZ binding potential in the paroxysmal non-kinesigenic dyskinesia striatum. This suggested that dopaminergic abnormalities, if present, may be due to altered regulation of dopamine release or to postsynaptic mechanisms, rather than to an altered density of nigrostriatal innervation (Bohnen et al. 1999). Taking this forward, Jarman et al. (2000) in one subject with familial paroxysmal non-kinesigenic dyskinesia found CSF monoamine metabolites were increased during an attack compared with baseline. Magnetic resonance spectroscopy of brain and basal ganglia performed both during and between attacks was normal. PET using the D2 receptor ligand ¹¹C-raclopride to measure D2 receptor binding also showed no abnormalities. Further light was shed on the pathophysiology when mutations in the MR1 gene were discovered. The enzyme plays a role in detoxification (see previously under ‘Genetics’ and Fig. 49.2).

The incidence of paroxysmal non-kinesigenic dyskinesia is unknown, but judging by the paucity of reports it seems rare. Cases have been recorded in the USA, Canada, Sweden, Italy, Australia, the UK, and Japan. Familial cases have been inherited as an autosomal dominant trait and the linkage to chromosome 2q 33-35 with mutations in the gene encoding the myofibrillogenesis regulator-1 has been mentioned above. Mount and Reback (1940) thought their cases were inherited recessively, but the family tree clearly shows inheritance to be dominant. They pointed out that no cases were inherited through phenotypically normal people. Forssman (1961), however, found the disorder was passed on through a symptomatically unaffected individual, although he did not interview this person. Nonetheless, this and the considerable range of severity among cases, suggests variable penetrance. Fouad et al. (1996) found one individual who carried the disease haptotype and was phenotypically normal. She was, however, only 8 years of age. The seven sporadic cases reported by Fahn and Bressman (1983) and Bressman et al. (1988) differed from other patients in several respects. Attacks were not triggered by foodstuffs, with the exception alcohol in one man. In two cases occasional bouts of dystonia were brought on by movement, although the majority of attacks in these same patients were not triggered by activity. Lack of sleep was not a precipitant and, interestingly, episodes of involuntary movement tended to occur on wakening in one male. Persistent dystonic posturing was present in two patients and in another dystonia could linger for weeks after an attack. Two further patients had tremor and a final one had familial scoliosis. In most of the nine sporadic and idiopathic cases reported by Demirkiran and Jankovic (1995), the movement was unilateral. These authors pointed out that a similar disorder could occur secondary to known neurological conditions and they had 17 such examples in their series. The relationship of idiopathic sporadic to familial cases remains to be clarified, but it seems likely that some of the former may really be examples of occult symptomatic or secondary dystonia.

In the familial cases mentioned above, there have been 84 affected individuals, 56 of whom have been male, giving a male:female ratio of approximately 2:1. This male predominance is not as great as in paroxysmal kinesigenic dyskinesia and it has not been present in all kindreds (Bryne et al. 1991, Fink et al. 1996). It is of doubtful significance. There does not seem to be any gender predominance among sporadic cases (Fahn and Bressman 1988, Demirkiran and Jankovic 1995). Familial paroxysmal non-kinesigenic dyskinesia frequently becomes symptomatic at an earlier age than paroxysmal kinesigenic dyskinesia. Thus some authors report onset around age 5–8 years of age (Mount and Reback 1940, Forssman 1961, Richards and Barnett 1968, Lance 1977, Tibbles et al. 1980) and in several patients symptoms have become apparent in the first year of life (Byrne et al. 1991). Occasionally onset may be delayed until late adolescence or early adult life (Forssman 1961, Lance 1977). Byrne et al. (1991) noted commencement at 40 years of age in a single patient, although the presence of myokymia made this family atypical. The age of onset is more variable in sporadic cases and has ranged from childhood to late middle life (Fahn and Bressman 1983, Bressman et al. 1988), and it is even more variable in secondary cases (range 2–80 years) with a peak in the 20s when caused by trauma and a mean age of 60 years when due to vascular events (Blakeley and Jankovic 2002). In one patient described by Dunn (1981) there was torticollis and posturing of an upper limb which started at 6 months and cleared at 40 months. It is possible, however, that this represents a variant of benign paroxysmal infantile torticollis or benign idiopathic dystonia of infancy (see Chapters 36 and 50 and Table 1 in the ‘Introduction’ to Section 11).

In familial cases attacks may occur spontaneously but tend to be precipitated by alcohol, coffee, tea, cocoa, chocolate, and cola drinks (Mount and Reback 1940, Forssman 1961, Richards and Barnett 1968, Lance 1977, Tibbles and Barnes 1980, Jarman et al. 1995), and in gene-proven cases this was the case in almost 100% of patients (Bruno et al. 2007). It is thus one of the proposed criteria for making the diagnosis. Forssman (1961) demonstrated onset of muscle spasms approximately half an hour after oral or intravenous administration of alcohol. One of his cases, who was an alcoholic, was reported to lie ‘in the ditch paralyzed by both drink and the attacks’. Caffeine is probably the precipitant in some foodstuffs and one of Forssman's (1961) patients had attacks precipitated by caffeine-containing medicine. Lack of sleep, fatigue, prolonged concentration, irritation, and anxiety have also been noted to trigger turns.

Forssman (1961) reported that the emotion of being a pallbearer caused an attack, with the result that the coffin slid down. Extremes of temperature have also been recorded to trigger attacks in some patients (Mount and Reback 1940, Forssman 1968). Both Mount and Reback (1940) and Forssman (1961) noted attacks did not occur during sleep. Unlike paroxysmal kinesigenic dyskinesia, paroxysms are not triggered by movement or hyperventilation (Mount and Reback 1940, Forssman 1968). There was marked diurnal variation in some of Forssman's (1961) patients and attacks did not occur in the morning. Unlike the inherited disorder, however, attacks in the sporadic variety usually occur without precipitation, although fatigue, menses, heat, and emotional stress may trigger them in some patients (Fahn and Bressman 1983, Demirkiran and Jankovic 1995). In a few of these non-familial subjects movement has brought on occasional bouts (Bressman et al. 1988).

In the inherited disorder a brief aura immediately preceding an attack is common and may consist of parasthesiae in the limb (Lance 1977), a feeling of weakness on one side (Tibbles and Barnes 1980), a ‘tugging’ in the muscles (Forssman 1961), stiffness throughout the body (Richards and Barnett 1968), or sensations about the trunk, throat, or face (Mount and Reback 1940, Lance 1977). Occasional patients have been reported to look ‘dazed’ at the start of an attack (Lance 1977). Forty one percent of the genetically proven cases reviewed by Bruno et al. (2007) had an aura such as sensation of tightness in a limb or involuntary movements of the mouth or anxiety.

Motor symptoms usually commence after the aura, but occasionally they can be suppressed by undertaking physical activity and ‘fighting it off’ (Lance 1963, Richards and Barnett 1968). In some patients muscle spasm seems to unpredictably commence anywhere (Forssman 1961), but in others it follows a stereotyped sequence (Lance 1963). Involuntary movements and posturing may come on gradually over 10 minutes or more, or they may be so abrupt as to cause the patient to fall (Forssman 1961, Lance 1977). Paroxysms may range from being mild and focal to severe and generalized.

A minor attack may consist of slight stiffness or mild cramps, whereas a major one can appear similar to advanced Huntington's chorea (Richards and Barnett 1968). In a 4-year-old child reported by Tibbles and Barnes (1980) they consisted of severe generalized weakness, although other family members showed typical involuntary movements. Dystonia and athetosis tend to predominate, and in some pedigrees these have been the only movements (Forssman 1961). Eighty percent of genetically proven cases were found to have a combination of dystonia and chorea; 12% had dystonia only. (Bruno et al. 2007) These tonic spasms may cause severe pain. Other patients show a variety of choreic and athetotic movements (Mount and Reback 1940, Richards and Barnett 1968, Jarman et al. 2000). Occasionally wild ‘flinging’ of limbs occurs (Mount and Reback 1940, Lance 1977). Clonic jerking (Forssman 1961, Richards and Barnett 1968) and rotatory shoulder movements (Lance 1977) have been reported. In some patients movements may start as one type and change to another. They can commence in one place and spread. Sometimes an attack will appear to wane, only to recur at another site (Forssman 1961).

In a typical dystonic episode the upper limb may be abducted or adducted at the shoulder, with elbow and wrist flexion and forearm pronation. The fingers may be flexed or extended. The foot plantar-flexes and inverts and there may be flexion or extension at the hip and knee (Fig. 49.3) (Mount and Reback 1940, Lance 1977). A variety of postures and movements occur, however, so that none can be regarded as completely typical.

Cephalic involvement is common with ‘twisting of the face’, blepharospasm, ptosis (Forssman 1961), or facial weakness (Fig. 49.3). One of Lance's (1977) patients was only able to ‘blink and move his eyes’. Normal eye movements have been reported in some patients (Tibbles and Barnes 1980) but occasionally have been restricted, and forced ocular deviation has been reported (Mount and Reback 1940). Blurring of vision and diplopia occurred occasionally in Mount and Reback's (1940) patient.

Although some patients speak and swallow normally during attacks (Lance 1977, Tibbles and Barnes 1980), in others these functions are impaired. Speech can be difficult, slurred, or impossible and replaced by grunts and groans (Mount and Reback 1940, Forssman 1961, Lance 1977, Tibbles and Barnes 1980). Coughing, swallowing, and tongue movements can be impaired and there may be drooling (Forssman 1961, Richards and Barnett 1968, Lance 1977). Two of Mount and Reback's (1940) cases bit their tongues during attacks and 45% of the genetically proven cases reported by Bruno et al. (2007) developed dysarthria or anarthria, however, with full awareness.

Involuntary neck and trunk movement can occur (Mount and Reback 1940, Forssman 1961). Some patients are able to stand and walk during attacks, but many cannot do this (Mount and Reback 1940, Richards and Barnett 1968, Tibbles and Barnes 1980).

Although painful muscle spasms and distress may make patients reluctant to communicate, loss of consciousness and incontinence do not occur. In fact, patients are able to micturate normally during attacks (Lance 1977).

In most patients paroxysms subside gradually, but occasionally they cease abruptly, over about a minute (Lance 1977). Usually patients recover without sequelae, but postictal headache and hunger were reported in one patient (Lance 1977).

The types of involuntary movements in spasmodically occurring cases seem to be similar, including forced jaw opening, tongue protrusion, truncal extension, and limb posturing. In the majority, however, these are asymmetrical or unilateral (Bressman et al. 1989).

Sleep seems to have a definite relationship to the familial disorder. Some patients have noted an attack is aborted by sleeping, even if just for a few minutes (Mount and Reback 1940, Forssman 1961, Lance 1963, Tibbles and Barnes 1980, Bryne et al. 1991, Fink et al. 1997, Jarman et al. 2000). Forssman (1961) reported waking had to be spontaneous in order for sleep to be effective. In addition, as mentioned above, paroxysms that are precipitated by lack of sleep (Forssman 1961) do not occur during sleep and may be absent ‘in the morning after a good night's sleep’ (Forssman 1961). As noted earlier, Bressman et al. (1988) discovered the opposite effect in a single patient whose bouts occurred after sleeping. In the inherited disorder the duration of attacks shows considerable variation, but they are always longer than paroxysmal kinesigenic dyskinesia. Although they are usually 15 minutes to 1 hour, they have ranged from 5 minutes to 11 hours (Forssman 1961). In most patients attacks occur no more than once daily and they are often separated by weeks, months, or occasionally years. Sporadic cases show more variability and although most spasms continue for 1–2 hours they have been reported to range from seconds to weeks. Similarly they have occurred from 30 times daily to several times a year (Bressman et al. 1988). Formal neurological examination during attacks is difficult. Mount and Reback (1940) had the impression of hypotonia in a patient whose movements were largely choreic. Conversely, Forssman (1961) noted rigidity in a patient with tonic spasm. Tendon reflexes have been reported normal (Richards and Barnett 1968) or difficult to elicit (Mount and Reback 1940). Plantar responses are flexor (Mount and Reback 1940, Forssman 1961, Richards and Barnett 1968).

In most cases the disorder persists but in some severity decreases with age (Forssman 1961, Tibbles and Barnes 1980, Bryne et al. 1991). In occasional patients, the disorder remits (Lance 1977). Although distressing and sometimes painful, the condition is not usually dangerous as in most cases onset of attacks is gradual. A ‘crib death’ has occurred at 21 months (Lance 1977), but it is uncertain if a dystonic paroxysm was involved.

Interictal abnormalities have been described in a few patients with the familiar disorder. Lance (1977) noted intellectual impairment in one patient, but this seemed to have been present from birth and there was no indication of progression. It is likely to be conincidental and normal intellect has been reported by others (Mount and Reback 1940, Forssman 1961, Richards and Barnett 1968). Fine tremor was present in one of Forssman's (1961) family. One patient had pes cavus and an ‘atypical Kayser-Fleisher’ ring (Mount and Reback 1940). Copper studies were not performed but seem unlikely to have been relevant in view of the otherwise typical paroxysmal non-kinesigenic dyskinesia and the dominant family history.

In the family described by Lance (1977) one patient had tonic-clonic epilepsy and another had prolonged absence attacks. A relative without paroxysmal dyskinesia possibly had adult onset absence attacks. In none did EEGs show epileptiform activity. The nature of these absence attacks is uncertain as two other patients in that pedigree are said to have looked ‘dazed’ or ‘glassy’ eyed at the onset of attacks of dyskinesia. Other reports do not mention epilepsy in patients or relatives, and it is possible this association is fortuitous. Thus, unlike paroxysmal kinesigenic dyskinesia, increased incidence of epilepsy in paroxysmal non-kinesigenic dyskinesia is not established. An affected father and son in the pedigree reported by Bryne et al. (1991) had post-contraction and spontaneous myokymia of limb muscles.

As noted above, interictal abnormalities are not uncommon in the sporadic variety, and mild dystonia, tremor, and scoliosis have been recorded (Bressman et al. 1988).

Routine haematology, biochemistry, and cerebrospinal fluid indicies have been normal (Forssman 1961, Richards and Barnett 1968). Interictal EEGs have usually been unremarkable, although a few have had a non-specific excess of slow activity (Mount and Reback 1940, Forssman 1961, Richards and Barnett 1968, Lance  1977, Tibbles and Barnes 1980, Bressman et al. 1988). Two patients have shown paroxysmal slow activity on hyperventilation and, although the authors suggested this may have been epileptiform (Mount and Reback 1940, Forssman 1961), such findings would nowadays be regarded as non-specific and probably within normal limits. One case in a large pedigree reported by Fouad et al. (1996) was said to have bilateral spike and wave complexes at 4 Hz with a dominance in the central brain region on interictal EEG. In a patient of Jacome and Risko (1984) photic stimulation induced epileptiform discharges in the cerebral hemisphere contralateral to the side of paroxysmal hemidystonia. Several patients, however, have had EEG during paroxysms and these have shown no change from baseline recordings (Richards and Barnett 1968, Tibbles and Barnes 1980). Overall there is thus little convincing evidence of EEG abnormality associated with this disorder.

Radiology, including pneumoencephalography, angiography (Lance 1977), and brain scan, have shown only minor, probably incidental, abnormalities, as outlined above under ‘Anatomical pathology’.

Electromyography and nerve conduction studies (Forssman 1961, Bryne et al. 1991) have been normal. Laboratory investigations, other than the genetic studies mentioned above, have thus shown no abnormality that seems likely to be relevant to this disorder.

Avoiding precipitants mentioned above under ‘Clinical features’ may reduce frequency of attacks, and suppressive manoeuvres as bouts commence may occasionally be effective. Once a paroxysm has started in the familial variety, sleep may terminate it. Most cases, however, require medication.

Avoidance of dangerous situations, as mentioned above under ‘Paroxysmal kinesigenic dyskinesia’, is usually not so crucial because of the gradual onset.

Unlike paroxysmal kinesigenic dyskinesia, major anticonvulsants are usually not helpful in inherited paroxysmal non-kinesigenic dyskinesia. Thus, phenytoin and carbamazepine have been ineffective and in some cases seem to have induced attacks (Mount and Reback 1940), although they appear to have helped occasional early-onset sporadic cases (Bressman et al. 1988). Valproic acid (Fouad et al. 1996), primidone (Lance 1963) and phenobarbitone (Forssman 1961), have occasionally improved paroxysms.

Benzodiazepines have been the most effective drugs. In the study by Bruno et al. (2007) of 49 genetically proven MR1 mutations carriers, 97% of those who had tried benzodiazepines had a beneficial response. Clonazepam, 6 mg daily, may prevent attacks (Lance 1977, Fahn and Bressman 1983, Bressman et al. 1988, Fink et al. 1996) and clorazepate has been said to completely suppress them (Fouad et al. 1996). Diazepam, up to 50 mg daily (Lance 1977), and chlordiazepoxide (Walker 1981) have also been reported to benefit patients. Acetazolamide has been helpful in sporadic paroxysmal non-kinesigenic dyskinesia (Fahn and Bressman 1983, Bressman et al. 1988). Adrenocorticotrophic hormone has been claimed to be effective in some sporadic cases (Fahn and Bressman 1982). Haloperidol, gabapentin, and acetazolamide, as well as l-dopa may also be mildly beneficial. Botulinum toxin can improve symptoms secondary to stroke (Blakeley and Jankovic 2002).

In 1977 Lance described a family with paroxysmal dyskinesia, which was precipitated only by sustained exertion. He separated this condition from paroxysmal kinesigenic dyskinesia and paroxysmal dystonic dyskinesia, which have been discussed above. Subsequently, a similar family was reported by Plant et al. (1984). Kurlan and Shoulson (1983) studied a pedigree with what seemed to be paroxysmal non-kinesigenic dyskinesia, but in a subsequent publication it was clear that spasms were precipitated by prolonged activity (Kurlan et al. 1987) and thus these cases are included here. Munchau et al. (2000) described an autosomal dominant family with five members who had dystonic attacks induced by exercise. Three of the affected members in the family also had migraine without aura. Wali (1992) has published details of a sporadic case of hemidystonia brought on by prolonged exercise. Four patients noted by Demirkiran and Jankovic (1995) were sporadic. Bhatia et al. (1997) also reported eight sporadic cases of paroxysmal exercise-induced dystonia and reviewed the then existing literature. They found there were in all 20 cases, putting together their cases and those in the literature; five were from two reported families (Lance 1977, Plant et al. 1984) and 15 sporadic cases. Thus paroxysmal exercise-induced dystonia was relatively rare.

Paroxysmal exercise-induced dyskinesia is characterized by attacks of involuntary movement which are precipitated by sustained exercise. Bhatia et al. (1997) found that most cases had episodes of dystonia following prolonged walking or running, and some cases have followed exposure to cold (Wali 1984). The duration of the episodes in most of the cases was between 5 and 30 minutes. They have thus been intermediate between the brief episodes of paroxysmal kinesigenic dyskinesia and the prolonged ones of paroxysmal non-kinesigenic dyskinesia. Occasional spasms have been as brief as 5 seconds and the longest have been 48 hours (Kurlan et al. 1987).

The literature on paroxysmal kinesigenic dyskinesia contains examples of patients with attacks precipitated by sustained exercise, including walking and running [cases 5 and 7 in Hishikawa et al. (1973); family A, case 1 and sporadic case 1 in Jung et al. (1973); cases 2, 4, 5, and 6 in Lishman et al. (1962)]. Lishman et al.'s (1962) case 2 had attacks during races. In their case 4 and Jung et al.'s (1973) sporadic case 1, paroxysms were induced by ‘vigorous activity’. In all of these, however, episodes were brief, typical of paroxysmal kinesigenic dyskinesia, and most had other attacks triggered by sudden movement after inactivity. Paroxysms during sustained activity may have resulted from sudden change in pace or direction, which is known to provoke paroxysmal kinesigenic dyskinesia, and thus these cases seem quite separate from paroxysmal exercise-induced dyskinesia. A family with an apparently recessive disorder characterized by rolandic epilepsy, episodes of exercise-induced dystonia, and writer's cramp (RE-PED-WC syndrome) affecting three members of the same generation has been linked to chromosome 16p12-11.2 (Guerrini et al. 1999). This family has some overlap with the paroxysmal kinesigenic dyskinesia/ICCA families (see earlier) but appears to be distinct from typical familial paroxysmal exercise-induced dyskinesia as those are autosomal dominant in inheritance (Lance 1977, Plant et al. 1984, Munchau et al. 2000) and interictally the patients in the latter are normal.

In a pedigree described by Auberger et al. (1996) 18 members from four generations had a disorder that they termed autosomal dominant choreoathetosis with spasticity. Attacks could be precipitated by physical exercise, although they did not give details. Emotional stress, lack of sleep, and alcohol might also act as triggers. Because of the presence of spastic paraplegia in some cases, we have chosen to discuss this separately below. There is however suggestion that mutations in the same gene, namely the one encoding the GLUT1 transporter (see later), may underlie both disorders.

In Lance's (1977) family there were three affected patients in three generations, and in Plant et al.'s (1984) family there were two affected patients in two generations. In Kurlan et al.'s (1987) pedigree there were five individuals in five generations with the disorder. The inheritance in these families was compatible with an autosomal dominant condition. Six out of 10 patients were female. The onset of symptoms varied between 3 and 23 years of age. In the family with five affected members described by Munchau et al. (2000), the mean age of onset in affected members was 12 (range 9–15 years). Male to female ratio was 3:1. The relationship of sporadic (Wali 1992, Demirkiran and Jankovic 1995, Bhatia et al. 1997) to inherited cases is uncertain.

The frequency of attacks has been between one a day to monthly. Initial sensory symptoms may have been due to the motor abnormality and there was no definite sensory aura. Immediately ceasing activity at the first sign of an attack may abort it (Lance 1977). Paroxysms usually commence after sustained walking, but the distance necessary to provoke symptoms has ranged from only 200 m to a mile or so. This can show considerable variation in the same patient at different times and sometimes long distances can be covered without difficulty (Lance 1977). Paroxysms have also been caused by other types of exercise, including climbing stairs, playing tennis, lifting children, writing, hammering, and doing housework. In the family described by Munchau et al. (2000) attacks in affected members were predominantly dystonic and lasted between 15 and 30 minutes. They were consistently precipitated by walking but could also occur after other exercise. Generalization did not occur. In some members even chewing gum could precipitate attacks sometimes localized to the jaw (Munchau et al. 2000). Emotional stress may facilitate production of attacks by exercise (Plant 1984, Demirikan and Jankovic 1995). Such episodes are usually restricted to the legs, although occasionally they may involve the trunk, arms, or face, or be generalized. Cold and heat also triggered attacks in the patients of Kurlan et al. (1987) and cold in the sporadic cases of Wali (1992) and Demirikan and Jankovic (1995). Menstruation and alcohol have been mentioned as possible precipitants in sporadic cases (Demirikan and Jankovic 1995).

Initial symptoms have been ‘heaviness’ in the legs, ‘cramping of the feet’, or a sensation that the ‘knees are being drawn towards the abdomen’. Earliest movements may consist of dystonic posturing with plantar flexion and inversion of the feet. It may be possible to continue to walk on the lateral borders of the feet, but eventually the patient is forced to stop or may fall. Other movements include extension of the large toes with hip and knee flexion (Plant et al. 1984). More severe attacks can result in twisting and writhing motion of the legs (Fig. 49.4). Such paroxysms normally affect both sides at once. In their review Bhatia et al. (1997) pointed out that exercise-induced dystonia most commonly affected the feet (15 of the 20 cases); but in two of their cases after starting in the feet the dystonia could become generalized, and in another it could affect the lower trunk. A hemidystonic distribution was the next most common presentation (four cases).

Although one of Lance's (1977) patients could induce attacks by polishing floors, spasms did not seem to involve the arms. In contrast, one of Plant et al.'s (1984) patients could produce dystonia of an upper limb by exercising it. Attacks provoked in this way might involve the ipsilateral lower limb. Prolonged writing was sufficient to cause an attack. Adduction at the shoulder, flexion of the wrist, and repeated flexion and extension of the fingers occurred. Attacks produced by upper limb activity only lasted 4 or 5 minutes, whereas those provoked by leg exercise persisted for 5–30 minutes. In the sporadic case of Wali (1992), the paroxysms could be brought on by vigorous exercise of the right, but not the left, arm, and involved only the right side. In Kurlan et al.'s (1987) family dystonic spams predominated but there was also ‘mild choreoathetosis’ and movements involved all the limbs, trunk, neck, and face. Attacks were painful and symmetrical. There are no postictal symptoms, apart from mild discomfort in the limbs.

Plant et al. (1984) noted attacks could be precipitated by 30–60 seconds of passive limb movement, a few seconds of vibration, an electric shock to the fingers, or upper limb ischaemia. Dystonia produced in these ways was largely confined to the stimulated limb (Fig. 49.4), although with passive movements there could be ipsilateral spread. Vibration on the forehead produced blepharospasm and on the sternum precipitated truncal dystonia. Thus, although dystonia provoked during normal daily activities is usually precipitated by sustained exercise and limited to the legs, there may be a wider underlying spectrum of abnormality.

Neurological examination between attacks has usually not revealed definite abnormality, although one patient developed fixed dystonia (Kurlan et al. 1987) and another had particularly well-developed calf muscles (Lance 1977).

The disorder has generally seemed non-progressive, although in occasional patients attacks became more troublesome with increasing age (Plant et al. 1984). In Kurlan et al.'s (1987) family eight other members spread throughout four generations had been troubled by painful but not disabling cramps in limbs after prolonged activity, and the relatives affected by paroxysmal dystonia had suffered similar cramps before onset of their frank spasms.

Abnormalities of anatomical pathology, biochemistry, and neurophysiology remain poorly understood. EEG recordings are typically normal. Other neurophysiological studies including somatosensory evoked potentials by stimulation of the median nerve (MN-SEPs), somatosensory evoked potentials by posterior tibial nerve stimulation (PTN-SEPs), brainstem auditory evoked potentials (BAEPs), visual evoked potentials (VEPs), motor evoked potentials (MEPs), and electromyography (EMG) are suggestive of hyperexcitability at the muscular and brain membrane levels (Margari et al. 2000). That attacks can be produced by several sensory stimuli, including vibration, may suggest that proprioceptive inputs activate an abnormal central mechanism. Between attacks, cortical excitability and inhibitory neuronal mechanisms (response threshold and amplitudes, duration of the silent period ipsilaterally and contralaterally, corticocortical inhibition and facilitation) have been found normal (Meyer et al. 2001). This is in contrast to focal task-related dystonia where abnormal motor cortex inhibition is also present during isometric muscle contraction (Rona et al. 1998). SPECT studies during motor attacks demonstrated reduced perfusion of the frontal cortex and basal ganglia and increased perfusion of the cerebellar, a pattern compatible with other forms of idiopathic and symptomatic forms of dystonia (Kluge et al. 1998).

No abnormalities have been reported in routine haematology, biochemistry, skull radiology, brain scans, or interictal and ictal EEG. A twofold increase of homovanillic acid and 5-hydroxyindoleacetic acid was measured in cerebrospinal fluid after motor attacks compared to baseline (Bhatia 1999), supporting the hypothesis of dopamine involvement in the pathophysiology. In gene-proven cases affected members had CSF glucose levels at or below the lower limit of the normal range (Weber et al. 2008).

Recently, mutations in the gene encoding the glucose transporter type 1 (GLUT1), located on chromosome 1p35-p31.3 have been identified in a three generation family with paroxysmal exercise-induced dyskinesia (Weber et al. 2008). Suls et al. (2008) could confirm these findings. The authors confirmed the findings by studying two independent families in which two additional mutations were detected. The gene encodes the transporter of glucose into erythrocytes and across the blood-brain barrier. This explains why glucose levels in the CSF have been found at or below the lower limit of the normal range. However, this gene SLC2A1, shows phenotypic variability in that it has only been associated with paroxysmal exercise-included dyskinesia, but mutations in this gene can also cause the so-called GLUT-1 deficiency syndrome (Leen et al. 2010). Thus the clinical presentation of GLUT1 disordes ranges from every severe (labelled as “classic GLUT1 deficiency syndrome”) to relatively mild (pure paroxysmal exercise-induced dyskinesia) (Fig. 49.5). The severe syndrome of “GLUT1 deficiency”, as recognized by paediatricians, is characterized by delayed development with microcephaly, drug-resistant

seizures, ataxia, spasticity, hypoglycorrhachia, and decreased erythrocyte glucose uptake (Seidner et al. 1998, Brockmann et al. 2001). Atypical forms of GLUT1 deficiency syndrome have a milder phenotype or late-onset, without seizures or presence of intermittent ataxia or dyskinesias triggerd by exercise or coffee (de Saint-Martin et al. 2007).

As for the condition of paraxysmol exercise-induced dyskinesia, this is likely to be genetically heterogeneous and SLC2A1 gene mutations may explain only some of the families with paroxysmal exercise-induced dyskinesia or sporadic cases (Schneider et al. 2010).

As mentioned above, a family with an apparently recessive disorder characterized by RE-PED-WC syndrome affecting three members of the same generation has been linked to chromosome 16p12-11.2 (Guerrini et al. 1999). This is in the same region as the defect in the families with paroxysmal kinesigenic dyskinesia and the ICCA syndrome (see previous section on ‘Paroxysmal kinesigenic dyskinesia’). The exact standing of the family described by Guerrini et al. (1999) is unclear as the clinical features of the attacks had some overlap with paroxysmal kinesigenic dyskinesia. Also this family is recessive in inheritance while most of the other families with paroxysmal exercise-induced dyskinesia described in the literature are autosomal dominant (Lance 1977, Plant et al. 1984, Munchau et al. 2000, Weber et al. 2008).

The characteristics of the different paroxysmal dyskinesias are shown in Table 49.2.

Avoiding sustained exercise is important in preventing attacks. Based on the molecular findings in GLUT1-related cases ketogenic diet has been recommended (Weber et al. 2008). Benzodiazepines may be helpful and clonazepam and oxazepam have produced improvement (Plant et al. 1984, Kurlan et al. 1987), although Lance (1977) and Wali (1992) noted no effect. Phenytoin, carbamazepine, phenobarbitone, and l-dopa have not generally been useful, although one sporadic case of Demirikan and Jankovic (1995) was said to be improved by l-dopa. Another of their cases was helped by carbamazepine. L-tryptophan was said to improve two of Kurlan et al.'s (1987) subjects. Bhatia et al. (1997) found that anticonvulsants were generally not found useful, but one reported case showed some benefit on acetazolamide. Another patient worsened by acetazolamide (Guimaraes et al. 2000). One case responded to pallidotomy (Bhatia et al. 1998).

Auburger et al. (1996) described a family in which a dominantly inherited disorder was manifest over four generations by episodes of involuntary movement, dystonic posturing of arms, legs and toes, imbalance, dysarthria, diplopia, and parasthesiae affecting the perioral region and lower limbs. These were sometimes accompanied or followed by headache. The attacks lasted approximately 20 minutes and occurred between twice daily and twice yearly. Age of onset ranged from 2–15 years. Bouts were triggered by exercise, emotional stress, sleep deprivation, and alcohol. Interictal examination was usually normal, but five out of 18 affected subjects had a spastic paraplegia. Although cerebellar signs were not noted in attacks, the authors seemed to consider the disorder was related to the episodic ataxias. They referred to it as paroxysmal choreoathetosis/spasticity. There was genetic linkage to chromosome 1p, and there is suggestion that on a genetic basis this condition may be a form of glucose deficiency syndrome which has been discussed in detail previously.

In 1967 Horner and Jackson reported a large pedigree with paroxysmal choreoathetotic dystonia affecting four generations. Nine out of 12 affected individuals only had attacks occurring during sleep. Niedermeyer and Walker (1971) subsequently described five cases of largely tonic generalized nocturnal paroxysms, who had negative EEGs using scalp electrodes. Stereotactic studies in four cases revealed epileptic foci in the cingulate gyrus. They ascribed the disorder to mesio-frontal epilepsy.

Paroxysmal hypnic dystonia, however, did not become a recognized entity until 1981 when Lugaresi and Cirignotta published a report of five patients with sleep-induced paroxysms of largely dystonic involuntary movements. They named the condition ‘hypnogenic paroxysmal dystonia’ but subsequently suggested it should be called ‘nocturnal paroxysmal dystonia’ (Lugaresi and Cirignotta 1982). The exact relationship between the sporadic disorder described by Lugaresi and Cirignotta (1981) and the cases of Horner and Jackson (1969) and Niedermyer and Walker (1971) remains uncertain. The same can be said of two cases described by Morley (1970), a father who had attacks of sleep-induced dyskinesia and a son with paroxysmal kinesigenic dyskinesia, and a large pedigree reported by Scheffer et al. (1994) with similar mainly nocturnal attacks due to ‘frontal lobe epilepsy’ (see later).

Paroxysmal hypnic dystonia or nocturnal paroxysmal dystonia is a syndrome characterized by dystonic, choreic, or ballistic movements occurring during non-rapid eye movement (non-REM) sleep, in particular during sleep stages 2 and 3. This is probably not just a single disorder and a subclassification has been proposed depending on the duration of attacks. Thus, Lugaresi et al. (1986) divided cases into those with short lasting (less than a minute) and long lasting (2 minutes to hours) attacks. It has also been suggested that there may be a third category with paroxysms of intermediate duration (2–5 minutes) (Lugaresi (1990). Justification for such categorization is that patients with short attacks present a relatively stereotyped picture, may have associated epilepsy, and their dyskinetic attacks respond to anticonvulsants. Attacks of long duration do not appear to be associated with epilepsy and do not respond to anticonvulsant drugs. Attacks of intermediate duration sometimes also occur during wakefulness and may be induced by sudden arousal from sleep or by physical exercise (Lugaresi 1990). It has been proposed that these short, long, and intermediate duration attacks might be analagous to paroxysmal kinesigenic dyskinesia, paroxysmal non-kinesigenic dystonia, and paroxysmal exercise-induced dystonia respectively (Lugaresi et al. 1986, Lugaresi 1990). The vast majority of cases seem to fit into the short duration category.

There has been no autopsy reports of this disorder, but radiology, including CT and MRI scans, has been almost uniformly unremarkable (Lugaresi et al. 1986, Tinuper 1990). We have found a possible suprasellar epidermoid cyst in a single case (Fig. 49.6). A temporal cavernous angioma was present in another patient and removal resulted in improvement of the attacks (Sellal et al. 1991). Two small posterior putaminal lesions were seen on MR scan in a post-traumatic case in which there was mild hemiparesis and paroxysmal nocturnal hemidystonia (Biary et al. 1994). This has been included below in the secondary (symptomatic) paroxysmal dyskinesias. SPECT demonstrated hyperperfusion of the anterior cingulate gyrus (Schindler et al. 2001).

There is no known biochemical abnormality associated with paroxysmal hypnic dystonia and catecholamine studies of urine and CSF have been normal in familial cases (Lee et al. 1985).

In almost all cases scalp recordings of wake and sleep EEGs have been normal both ictally and interictally (Lee et al. 1985, Lugaresi et al. 1986, Lehkuniec et al. 1988). As mentioned above, Niedermyer and Walker (1971) noted nocturnal generalized tonic seizures in five patients and found that in four there was stereotactic evidence of epileptic activity in the cingulate gyrus, although scalp EEGs

had been unremarkable. In 1972 Tharp reported three patients with frequent nocturnal seizures, which included vocalization and semi-purposeful motor activity. They were found to have repetitive high amplitude frontal sharp and slow waves on surface EEG, which were most marked during sleep. One patient had a lateral frontal focus on electrocorticography, leading the author to conclude that the seizures originated from this area. Rajna et al. (1983) studied a patient with frequent nocturnal attacks of arousal, vocalization, ‘changes of posture, and smaller complex orientational movement series’. Scalp EEGs were unremarkable, but the patient was found to have an epileptiform discharge in the right cingulate cortex using depth electrodes. Surgical removal of the presumed focus, however, did not produce improvement.

Clinical descriptions in the cases of Niedermyer and Walker (1971), Tharp (1972), and Rajna (1983) suggest that there may be a relationship with the disorder described by Lugaresi and Cirignotta (1981), but lack of detail prevents adequate evaluation.

Many surface EEG studies have shown that the nocturnal paroxysms commence during non-REM sleep and are usually preceded by a few seconds of electrophysiological evidence of arousal. In addition, K-complexes or vertex waves frequently seem to trigger this sequence (Rajna et al. 1983, Lee et al. 1985, Lugaresi et al. 1986). Tinuper et al. (1990) studied two patients in whom typical episodes of paroxysmal hypnic dystonia sometimes merged into generalized convulsive seizures. Anteriorly placed K complexes or sharp-waves accompanied the dystonic attacks. In the episodes that progressed to convulsions the EEG developed clear-cut generalized epileptiform spiking. Epileptiform discharges, either localized or generalized, were said to be present in four out of eight patients reported by Besset and Billard (1985), and Sellal et al. (1991) noted interictal abnormalities in 18 and ictal paroxysmal activity in nine out of 23 patients. In a single case Oguni et al. (1992) found right frontal interictal epileptiform discharges in a patient with attacks which involved only the contralateral arm. Ictal recordings, however, did not generally show this. Nonetheless, in most reported cases EEG evidence of seizure activity has been absent or inconspicuous.

Some authors have considered paroxysmal hypnic dystonia to be a conversion reaction (Kovacevic-Ristonovic et al. 1988) and Berger et al. (1987) reported cure with psychotherapy. The patient's attacks, however, were not typical of paroxysmal hypnic dystonia.

Lugaresi et al. (1986) postulated that paroxysmal dystonia characterized by short duration attacks is an epileptic phenomenon. Seizures in mesio-orbital cortex produce complex motor behaviour with vocalization, bilateral arm and leg activity, and movements of the head and trunk. Genital manipulation sometimes occurs. Such seizures are often nocturnal (Geier et al. 1977, Wada and Purves 1984, Williamson et al. 1985, Waterman et al. 1987, Wada et al. 1988) and not detected in scalp electrodes (Wada and Purves 1984, Williamson et al. 1985).

On the basis of the cases sited above, in which short duration attacks of paroxysmal hypnic dystonia progressed to convulsive seizures, Tinuper et al. (1990) proposed that these episodes are due to mesio-orbital epileptic seizures. They have further suggested that these may be triggered by K-complexes. They have pointed out that at times during non-REM sleep brief dystonic posturing may occur every 20-40 seconds in subjects without arousal and that this is similar to the frequency of K-complexes. In addition, the same authors (Montagna et al. 1990) suggested that paroxysmal arousals, which consist of brief awakening associated with a start or cry, eye opening, a frightened or ‘questioning’ expression, dystonic posturing, and tremor, may be a minor version of paroxysmal hypnic dyskinesia (Montagna 1992). The former attacks last only seconds, are sometimes preceded by K-complexes, and are not directly recalled by the patient, although there may be an awareness of having slept badly or day-time somnolence. It has also been postulated that episodic nocturnal wanderings with screaming, unintelligible speech, walking, leaping, head banging, kicking, and the like, for which there is subsequent amnesia, may be related. These may be at one end of a spectrum of epileptic disorders which has paroxysmal arousals at the other extreme and paroxysmal hypnic dyskinesia in between (Montagna 1992). Whatever the merits of this argument, further discussion in this section is limited to paroxysmal hypnic dyskinesia. Paroxysmal arousals are discussed in Chapter 47.

Other authors have also supported the notion that paroxysmal hypnic dyskinesia is epileptic in origin, possibly involving mesio-frontal regions (Hirsch et al. 1994). Among 23 subjects with this disorder reported by Sellal et al. (1991), nine had suffered previous tonic-clonic seizures and four had experienced partial seizures. In six of these polysomnography was said to show typical paroxysmal hypnic dyskinesia, which then progressed to generalized convulsive seizures. Meierkord et al. (1992) compared nine patients who had paroxysmal hypnic dyskinesia with eight who suffered from day-time frontal lobe seizures and another eight with nocturnal motor attacks of known epileptic origin. They all underwent video-EEG monitoring. There were no clinical features that allowed distinction between the groups and no single phenomenon that was unique to paroxysmal hypnic dyskinesia. Overall, the data obtained by such monitoring, the frequency of day-time epilepsy in subjects with this disorder, and the response to anticonvulsants support an epileptic basis for the dyskinetic bouts of short duration.

The pathophysiological mechanisms underlying the less frequent longer lived paroxysmal hypnic dystonic attacks are uncertain, and it seems unlikely that these are associated with epilepsy.

The incidence of paroxysmal hypnic dystonia is quite uncertain, but judging by the substantial number of reports that occurred in the decade after Lugaresi and Cirignotta's description in 1981 it may not be all that uncommon. The majority of published cases have been sporadic, although a dominantly inherited disorder with a somewhat similar picture has been described (Horner and Jackson 1967, Lee et al. 1985). Scheffer et al. (1994) reported six families with what they termed ‘frontal lobe epilepsy’, which was inherited as an autosomal dominant disorder. An eponym ADNFLE (autosomal dominant frontal lobe epilepsy) was given to describe this disorder (Scheffer et al. 1995). They state it was often misdiagnosed as night terrors, nightmares, hysteria, or paroxysmal nocturnal dystonia. The predominant seizure pattern was clustered nocturnal motor seizures of 20 attacks nightly, lasting 30-40 seconds each. The clinical description was very similar to that of paroxysmal hypnic dyskinesia. In one family day-time complex partial seizures predominated and in 64% of patients there were secondary generalized seizures. In the patients with short duration attacks reported by Lugaresi et al. (1986) age of onset varied from 3-47 years (mean 21.8) and duration of illness at time of observation ranged from 1-34 years. There does not appear to be any sex preference (Lugaresi et al. 1986). In the family of Scheffer et al. (1994), mentioned above, age of onset varied from 6 months to 55 years.

Some attacks are associated with an expression of fear, grimacing, or feelings of panic (Stoudemiere et al. 1987, Tinuper et al. 1990). Feelings of choking, apnoea, and changes in heart rate with both tachycardia and bradycardia have been reported (Lugaresi et al. 1986, Lehkuniec et al. 1988 and Maccario and Lustman 1990).

The whole episode may last between a few seconds and a minute or so. After a variable period the patient goes back to sleep, commonly to be woken again by similar episodes on a number of occasions during the night. Although attacks vary considerably between patients they are usually quite stereotyped in individual cases. They occur on most nights and many patients are troubled by them every time they sleep. They can vary from 1–20 or more paroxysms during a single night.

Typical attacks may not be difficult to recognize but there are a number of patients in whom the diagnosis is less obvious. Occasional patients may have some episodes during the day (Lugaresi et al. 1986, Stoudemiere et al. 1987, Lehkuniec et al. 1988).

In the short duration attacks the patient suddenly arouses from sleep with dyskinetic movements. Although dystonia tends to predominate and there may be briefly sustained dystonic posturing, chorea and ballistic elements may occur (Lugaresi et al. 1986). There may be head turning, truncal arching, flexion or extension of limbs, fanning of fingers, forced wrist flexion, or inversion and plantar flexion of the feet (Fig. 49.7). Some patients show wild thrashing of the trunk and limbs or pedaling movements with the legs (Lugaresi et al. 1986, Tinuper et al. 1990).

The onset of movements may occur while the patient is asleep but consciousness usually occurs, although recollection may be hazy. There may be spontaneous vocalization, including screaming, but some patients are unable to talk in spite of being fully conscious (Lee et al. 1985, Lehkuniec et al. 1988).

Attacks may commence focally or be limited to one limb (Tinuper et al. 1990) and precipitation by tactile stimulation to one foot was reported in a case by Lehkuniec et al. (1988). Occasionally movements may be quite inconspicuous and bouts are dominated by feelings of terror, panic, tightness in the throat, dyspnoea, and tachycardia (Stoudemiere et al. 1987).

Tuniper et al. (1990) noted that in addition to the major attacks, which have been described above, patients may show very frequent smaller episodes and these seem to be fragments of the main attacks. Thus, there may be ‘minimal’ episodes consisting of slight movement lasting less than 5 seconds, or ‘minor’ bouts with a little more exaggerated dystonic posturing lasting 5-10 seconds. These are only different in degree from ‘major’ attacks and may be associated with arousal lasting only a few seconds or they may occur during sleep. They can occur periodically every 20-40 seconds during some portions of non-REM sleep. These episodes, paroxysmal arousals, and episodic nocturnal wanderings have been discussed above under ‘Pathophysiological mechanisms’.

Some authors have documented excessive tiredness in patients suffering from paroxysmal hypnic dystonia (Lee et al. 1985) and it has even been claimed that sleepiness may be the only symptom of this disorder. Maccario and Lustman (1990) noted paroxysmal episodes of stereotyped muscular activity, vocalization, apnoea, and tachycardia, associated with increased wakefulness and decreased periods of REM sleep, in patients presenting with excessive day-time somnolence but without a history of definite nocturnal awakening. The exact relationship of this to typical paroxysmal hypnic dystonia is uncertain.

There is an increased incidence of epilepsy in patients with short duration attacks of paroxysmal hypnic dystonia and Lugaresi et al. (1986) noted this in two thirds of patients. It consisted of sporadic day-time partial sensory-motor seizures and both day- and night-time grand mal seizures. Some of these had complex partial seizure symptomatology, including viceral sensations. Day-time complex partial and secondary generalized seizures have also been frequent in occasional pedigrees in which a clinical picture resembling paroxysmal hypnic dyskinesia has occurred (see earlier).

Patients with long duration attacks of paroxysmal hypnic dystonia show similar dystonic, choreic, or ballistic movements which may fragment sleep and reduce both REM and non-REM periods. Lugaresi et al. (1986) reported such attacks preceding typical Huntington's disease by 20 years.

Patients with intermediate duration attacks have been reported to have bouts occurring when awake as often as during sleep. Movements may be short and jerky and alternate between limbs and the trunk, producing a puppet-like appearance. Paroxysms have been triggered by physical exercise or sudden arousals from sleep (Lugaresi 1990, Montagna et al. 1992[b]). These long and intermediate duration attacks appear to be quite different from the short duration paroxysms described above.

In patients with a history suggestive of paroxysmal hypnic dystonia the most relevant investigation is nocturnal video and EEG monitoring. Although surface EEGs are likely to be negative, this is not always the case and in some attacks mesio-frontal discharges may be recorded (Scheffer et al. 1995). Recording with depth electrodes would rarely be justified. CT or MRI scanning is likely to be negative.

The paroxysmal dyskinesias have been suspected to be disorders of ion channels and ADNFLE, which has been found to be caused by a mutation in the gene encoding the neuronal acetylcholine receptor, a ligand gated ion channel (see later), is thus the first paroxysmal dyskinetic disorder to be proven so. Phillips et al. (1995) mapped a ADNFLE locus on chromosome 20q13.2 in an Australian family reported earlier by Scheffer et al. (1995). The obvious candidate was the one encoding the alpha 4 subunit of the neuronal acetylcholine receptor (CHRNA4). Two different mutations – a missense mutation and a 3-bp insertion – were then identified in the CHRNA4 gene in the Australian family and in a Norwegian family respectively (Steinlein et al. 1995, 1997). However, another family with ADNFLE was not linked to CHRNA4 on chromosome 20q but to a novel locus on chromosome 15q24 close to a CHRNA3/CNRNA5/CHRNB4 nicotinic acetylcholine receptor gene cluster (Philips et al. 1998). Also, in seven other families with ADNFLE and in seven sporadic cases, linkage to the ADNFLE loci on chromosome 20q13.2 and 15q24 was excluded, thereby suggesting the existence of at least a third ADNFLE locus and supporting the fact that ADNFLE is a genetically heterogeneous disease (Philips et al. 1998). How these mutations in the CHRNA4 gene cause epileptogenesis is not clearly understood, although impaired calcium entry into cells may be a possible mechanism (Kuryatov et al. 1996, Weiland et al. 1996).

Short duration attacks usually respond well to anticonvulsants. Carbamazepine has been effective in the majority of cases and sometimes quite small doses will be effective (Lee et al. 1985, Lugaresi et al. 1986, Stoudemiere 1987, Tinuper 1990, Scheffer et al. 1995). As little as 1.5 mg/kg/day has occasionally been used. Other anticonvulsants including phenobarbitone, primidone, phenytoin, and benzodiazepine, have usually been ineffective, although some patients have responded to the latter (Lehkuniec et al. 1988, Nair 1990). Neuroleptic drugs have not been helpful (Lugaresi et al. 1986). Four out of seven of Maccario and Lustman's (1990) patients with excessive day-time somnolence were said to be improved on anticonvulsants, including valproic acid, carbamazepine, and phenobarbitone.

Therapy for intermediate and long duration attacks has not been established.

A list of secondary causes is shown in Table 49.3.

Chorea or dystonia due to vascular causes are usually long lasting or permanent (see Chapters 24 and 43). There are a few reports of brief episodes of chorea, athetosis, or dystonia associated with probable transient ischaemic attacks, infarction, or haemorrhage.

Antin et al. (1967) described a woman with a haemorrhage in the region of the right thalamus and left-sided hemiparesis. Choreic and ballistic movements developed in the paretic limbs during pneumoencephalography and persisted for 8 hours. Follow-up pneumoencephalography over 3 months later showed the previous right thalamic mass had been replaced by a cystic cavity communicating with the third ventricle. It is uncertain whether the involuntary movements resulted from further haemorrhage or from shift of the swollen brain during pneumoencephalography. Although these movements were transient, they were not episodic in the sense that they were recurrent.

It has not been possible to prove involuntary movements have resulted from transient cerebral ischaemia, but the association of cerebrovascular disease in the absence of any other apparent cause makes this aetiology likely. Ganshirt et al. (1978) had a patient with recurrent episodes of hemiballismus lasting minutes, which they attributed to vertebrobasilar insufficiency as there were supporting clinical and angiographic features. The involuntary movements stopped with anticoagulation. Calzetti et al. (1980) reported a man with right carotid stenosis, proximal left subclavian occlusion, and left-sided subclavian steal syndrome. Right-sided hemiballismus developed while performing moderate activity with his arms. There were recurrent attacks lasting a few minutes, followed by freedom for the same duration. Marked improvement occurred after 24 hours and they disappeared on the 4th day after onset.

Margolin and Marsden (1982) reported four patients with attacks of hemiathetosis in one arm or hemiballismus lasting between 1 minute and 1 hour. The patients were all over 50 years of age and three had established vascular disease, while the fourth had hypertension. In two patients arteriography revealed atheroma in the contralateral carotid artery and endarterectomy was performed. One patient was hyperglycaemic and this has been associated with readily reversible involuntary movements (see Chapter 24). Probable episodes of choreoathetosis in one arm were noted in a 53-year-old patient with contralateral carotid stenosis and did not recur following carotid endarterectomy. The coexistence of seemingly inactive systemic lupus erythematodes, however, makes assessment of this case difficult (Stark 1985).

Baguis et al. (1995) reported seven patients with what were thought to be carotid territory transient ischaemic attacks in whom there were brief bouts of involuntary coarse irregular wavering or trembling movements of one upper limb, with or without involvement of the ipsilateral leg. They were termed ‘shaking’ movements, but from the descriptions given there is nothing to clearly differentiate them from chorea or ballism, and this possibility was not addressed by the authors. All but one patient had other attacks suggestive of carotid territory transient ischaemia and each of them had major occlusive atheroma in the contralateral carotid.

Sunohara et al. (1984) investigated a patient with a hemiparesis who had no involuntary movements at rest, but developed dystonic writhing of the weak limbs on attempted movement of any body part. CSF 5-hydroxyindoleacetic acid was low and homovanillic acid was high. Administration of 5-hydroxytryptophan and clonazepam abolished these movements. There was an infarct in the contralateral posterolateral ventral thalamus. A similar case has been reported in whom dystonic episodes lasted less than 1 minute and were triggered by moving the left hand. They followed infarction of the right posterior, lateral, and ventral thalamus and to a lesser extent the medial globus pallidus and internal capsule. These movements involved the left face and limbs and were abolished by phenytoin. In this respect they resembled idiopathic kinesigenic dyskinesia (Camac et al. 1990). They differed from those of Sunohara et al. (1984) in that they were not precipitated by every attempted movement and could thus more genuinely be considered episodic.

Sustained chorea due to thyrotoxicosis is well documented (see Chapter 24). Fischbeck and Layzer (1979) described an epileptic patient with myasthenia gravis and thyrotoxicosis who had continuous mild chorea affecting mainly the left hand and foot. In addition, 1- to 2-minute attacks occurred 1-4 times daily, consisting of tonic extension of the trunk and legs and flexion of arms with intermittent shaking of hands. The right upper limb was affected more than the left. These episodes were described as ‘choreoathetosis’. EEG showed diffuse slowing. Attacks ceased following treatment of the thyrotoxicosis leaving moderate residual chorea. It was postulated that residual striatal damage due to earlier documented status epilepticus predisposed to development of chorea, but the mechanism underlying such paroxysmal attacks remains uncertain. Paroxysmal attacks of asymmetrical dystonic and choreic movements, lasting minutes and induced by voluntary activity, have been reported to result from over-medicating with thyroid extract (Drake 1987).

Hypocalcaemia may produce a variety of psychiatric and neurological features as outlined in Chapter 24. Movement disorders are relatively common and are often intermittent. Tetany is the most frequent and, along with epileptic seizures, is discussed in Chapter 50.

Most chorea, athetosis, or dystonia occurring in hypocalcaemic patients is sustained (see Chapter 24), but there are a few records of paroxysmal attacks, some of which are exercise induced. Simpson (1952) reported generalized chorea, which spontaneously subsided over a week.

Tabaee-Zadeh et al. (1972) described a 30 year old with an 11-year history of attacks lasting 30 seconds, occurring several times weekly, which were preceded by vigorous activity. Torticollis, dystonic posturing of the arm, and semiflexion of the leg were accompanied by intermittent choreic and athetotic movements. Basal ganglia calcification was present but attacks were abolished by correction of hypocalcaemia. The sustained tonic spasms in this case are similar to those of tetany and the extent to which central and peripheral mechanisms were involved is uncertain (see Chapter 50).

There is, however, no doubt that hypocalcaemia may be associated with episodic but otherwise typical chorea or dystonia. This is usually generalized but may be asymmetrical and can last between seconds and several hours. In some patients bouts of involuntary movement are precipitated by initiation of movement and resemble idiopathic kinesigenic dyskinesia (Barabas and Tucker 1988, Micheli et al. 1989). Identical paroxysms can occur in pseudohypoparathyroidism (Siejka et al. 1988). These attacks disappear with correction of hypocalcaemia (McKinney 1962, Soffer et al. 1977, Siejka et al. 1988) and are not necessarily associated with calcification on plain skull radiology.

The intermittent occurrence of these attacks, abolition by correcting plasma calcium levels, and absence of overt calcification of basal ganglia suggest functional, rather than structural, mechanisms. Nonetheless, microscopic and brain scan evidence of calcification may occur with normal plain skull radiology and the interaction of structural and biochemical mechanisms cannot be excluded in some cases (see Chapter 50).

As mentioned in Chapter 24, hypoglycaemia is frequently accompanied by restlessness, but there is a paucity of convincing accounts of chorea. Transient abnormal posturing of limbs due to hypoglycaemia during diabetic treatment has been recorded (Marsden 1983). Newman and Kinkel (1984) noted repeated attacks of choreoathetosis of all limbs, facial grimacing, and opisthotonus in a patient during episodes of hypoglycaemia. The movements were terminated by intravenous glucose. It was postulated that there may have been some underlying basal ganglia abnormality, which was uncovered by the hypoglycaemia. Similar involuntary movements have been noted by others (Parajua et al. 1986) and may even occur with low normal blood glucose levels in diabetes (Haan et al. 1989). Winer et al. (1990) reported recurrent episodes of generalized dystonic posturing associated with confusion, impairment of speech, sweating, and facial flushing in a woman with hypoglycaemic episodes due to an insulinoma. Interestingly, the movements were increased after reversal of hypoglycaemia, raising the possibility that rate of change, rather than absolute values, of blood glucose might underlie such movements.

Sudden onset athetotic, choreic, and ballistic movements have been reported in a few patients with non-ketotic hyperglycaemia (see Chapter 24). These have sometimes been the first indication of diabetes. Movements have ceased within 18-24 hours of correcting glucose and electrolyte abnormalities and it is thus uncertain if they would have resolved spontaneously (Rector et al. 1982, Totoritis et al. 1982). Occasionally, such bouts may be triggered by movement or excitement and resemble paroxysmal kinesigenic dyskinesia. One reported case had an atrophic lesion in the parieto-temporal area contralateral to the affected arm, probably as the result of a previous infarct (Clark et al. 1995), raising the possibility of focal epileptic seizures. These are a recognized but uncommon entity in non-ketotic hyperglycaemia (see Chapter 50).

Paroxysmal episodes of dystonia lasting hours rarely complicate varicella infection and may cause oculogyric crises and oromandibular spasms, including tongue protrusion (Gallomp and Fahn 1987). Sustained chorea and dystonia seem more frequent but are still extremely uncommon (see Chapter 24).

Careful analysis of videotapes of complex partial seizures suggests that in about 15% of such attacks there is dystonic posturing of one upper limb, which occasionally spreads to involve the leg. In most cases this is associated with choreoathetotic movements or coarse 2–3 Hz tremor of the same limb (Kotagal et al. 1989[a]). There is usually automatism in the opposite limbs. Kotagal et al. (1989[a]) found that the dystonic limb was always contralateral to the ictal discharge, which was most commonly in the basal temporal lobe, with minimal involvement of the convexity. These features led them to postulate that such attacks might involve the basal ganglia and they differentiated them from the tonic seizures caused by ictal discharge in the supplementary motor area (Kotagal et al. 1989[b]) (see Chapter 50). Newton et al. (1992) confirmed that the affected temporal lobe was always contralateral to the dystonicly postured upper limb and found that the direction of head rotation was of no value in localizing the epileptic temporal lobe. They discovered using SPECT scans that there was hyperperfusion of the basal ganglia in the involved hemisphere in all cases. Others, however, have claimed similar dystonic and choreoathetotic posturing can occur with epileptic activity arising elsewhere, such as the supplementary motor area, cingulate gyrus, and lateral premotor regions (Bennett et al. 1989, Bokstein et al. 1995) (see Chapter 50).

In most cases the impairment of conscious level which accompanies complex partial seizures would prevent their classification as episodic movement disorders as defined in this book (see ‘Introduction’ to Section 11). It is possible, however, that brief loss of consciousness might go unnoticed. Although we have chosen to classify the movements associated with complex partial seizures or those arising in the temporal lobe as dystonic and those originating elsewhere in the brain as tonic, this is based largely on the descriptions in the literature and these are often imprecise. For example, Neville and Boyd (1995) described two children with disordered gait which was thought to result from epilepsy. In one of them the right arm was held above the head in a ‘dystonic’ position with the elbow flexed and the hand partly open. This lasted for 3 days. EEGs showed left fronto-temporal abnormalities, including sharp waves. Consciousness was preserved. It is uncertain here exactly what structures were involved and it is probable lateral premotor or supplementary motor areas could have been affected. Thus distinction between dystonic and tonic seizures is somewhat artificial and there is considerable overlap and variation in the types of movements seen in these disorders. The section on ‘Focal epilepsy’ in Chapter 50 should be read in conjunction with the present description.

Choreic and dystonic movements occasionally occur in multiple sclerosis (Tranchant et al. 1995) (see Chapter 24). These are caused by a relapse and as such usually persist for several weeks or months. They would thus not be considered episodic movement disorders, as defined here. Occasionally, however, movements may disappear within a week or two [case 3 in Sarkari (1968)] and it is thus possible they could be included.

Tonic spasms of cerebral and spinal aetiology are considered under ‘Tonic attacks’ in Chapter 50 and recurrent episodes of dysarthria and ataxia are dealt with under ‘Intermittent ataxias’ in Chapter 51. It should be noted, however, that many such brief tonic bouts are associated with dystonic posturing (Berger et al. 1987) and could equally well be included here.

Dystonia is an early feature of progressive supranuclear palsy (see Chapter 7) and has been reported in up to about a quarter of patients (Rafal and Friedman 1987). In these cases, however, the involuntary movements are usually persistent. In rare cases the dystonia may be quite episodic and precipitated by attempts at voluntary activity, leading to a picture superficially resembling idiopathic paroxysmal kinesigenic dyskinesia (Adam and Orina 1986).

Paroxysmal dystonic dyskinesia has been reported in association with inherited cerebellar ataxia (Fig. 49.8). Mayeaux and Fahn (1980) described a family with dominantly inherited progressive cerebellar ataxia, in which two siblings also had paroxysmal dystonia and choreoathetotic episodes. These involved the limbs, trunk, and oro-mandibular structures lasting between half an hour and 4 hours. Three siblings with a recessively inherited childhood-onset disorder showing nystagmus, dysarthria, hypertonia, hyper-reflexia, extensor plantar responses, incoordination, and ataxia were described by Graff-Radford (1986). Two of these also had late-onset dystonic spasms affecting the face, bulbar muscles, and limbs lasting 30 seconds to several minutes. As mentioned above under ‘Paroxysmal dystonic dyskinesia with spasticity’, Auberger et al. (1996) described a pedigree with episodes of dystonia and ‘imbalance’ but without observed ataxia. Some family members had spastic paraplegia. An argument could be made for including this condition here, but we have chosen to put it with the primary causes of paroxysmal dyskinesia because the paraplegia was found in only a small proportion of affected individuals.

Phenytoin and carbamazepine improved the briefer bouts in Graff-Radford's (1986) patients whereas acetazolamide was ineffective. The converse was the situation with Mayeaux and Fahn's (1980) cases.

Alternating hemiplegia of childhood usually becomes apparent shortly after birth or in the first 18 months of life. The disorder is usually sporadic, although an autosomal dominant form has been described (Silver and Andermann 1993). There is persistent hypotonia and paroxysmal dystonic posturing, which appear early. This is then followed by episodes of hemiplegia which can occur with or follow the bouts of dystonia. The hemiplegia usually persists for hours or days but can last only minutes. It frequently alternates sides but can be bilateral. As the disorder progresses the dystonic attacks become less marked, but in later childhood or adolescence persistent chorea or dystonia becomes apparent. A variable degree of mental retardation is usual. A family history of migraine has been commonly recorded. Therapy with flunarizine, a calcium channel blocker, commonly improves the disorder by reducing the severity of the attacks. Less often the frequency is decreased (Verret and Steel 1971, Hosking et al. 1978, Krägeloh and Aicardi 1980, Casaer and Azon 1984, Aicardi 1987, Casaer 1987). Abnormalities in MRs have raised the possibility of an underlying disorder of mitochondrial function.

Early onset hypotonia with bouts of dystonia involving the eyes, neck, trunk, and limbs has been suggested to be a variant of alternating hemiplegia of childhood. Attacks have lasted minutes to hours. The subsequent development of episodes of hemiplegia in the teens and response to flunarizine support this (Andermann et al. 1994).

Neurological deficit dating from birth has been present in several patients with paroxysmal movement disorders (Alajouanine and Gastaut 1955, Lance 1963, Rosen 1964, Kertesz 1967). Underdevelopment of limbs on one side and pyramidal signs have been described (Lance 1963, Rosen 1964). The hemiparetic patient reported by Rosen (1964) had typical paroxysmal kinesigenic dyskinesia with bilateral attacks of involuntary movement. In Kertesz's (1967) case the attacks are not described, but other family members had typical kinesigenic dyskinesia without neurological deficit. In Lance's (1963) patient minor involuntary movements in the paretic arm were present between spasms, which were accompanied by diffuse pain. These occurred spontaneously or were precipitated by surprise, rather than on initiation of movement. This case does not fit the criteria of paroxysmal kinesigenic dyskinesia and it is likely attacks were of a different type. A 15 year old described by Hamano et al. (1995) had a large left porencephalic cyst with changes in the adjacent lateral putamen, contralateral arm paresis, and homonymous hemianopia. There were bouts of paroxysmal kinesigenic choreoathetosis. Blood flow studies showed hypoperfusion of the lenticular nucleus and region of the cyst.

Robin (1977) reported typical unilateral attacks of paroxysmal kinesigenic dyskinesia developing in a 33-year-old man 8 months after head injury, which resulted in 20 minutes of unconsciousness and 18 hours of post-traumatic amnesia but without focal neurological deficit. Onset of paroxysmal kinesigenic dyskinesia at this age is rare and it seems likely it was related to the head injury.

Subsequent reports have appeared of kinesigenic and non-kinesigenic paroxysmal choreic and dystonic movements developing weeks or months after head injury. These have occurred in hemiparetic (Drake et al. 1986) and normal (Richardson et al. 1987) limbs and been associated with normal (Drake et al. 1986) and abnormal (Richardson et al. 1987) brain scans and EEGs. When present, however, the scan abnormalities have not directly involved the basal ganglia and the EEGs have not shown paroxysmal epileptiform activity during the attacks. In most cases movements have been well controlled by anticonvulsants, including phenytoin, carbamazepine, and phenobarbitone. Perlmutter and Raichle (1984) noted paroxysms of right facial and limb dystonic posturing with weaker involvement of the left foot, which lasted seconds and recurred every few minutes. These spasms commenced only 20 minutes after minor head and neck trauma and were actually lessened by voluntary movements. They were associated with increased blood flow and decreased oxygen metabolism in the left basal ganglia on PET scanning. Although these responded to phenytoin and trihexyphenidyl, they seem different to the other cases and rather more like the tonic spasms associated with disorders such as multiple sclerosis (see Chapter 50).

Similarly, episodes of exclusively nocturnal hemidystonia in a 4-year-old boy, commencing some months after head injury and arising from non-REM sleep, seem unlike most other cases. They were associated with a mild hemiparesis in the involved limbs and small posterior putaminal lesions on MR scan. The attacks did not respond to anticonvulsants but were suppressed by acetazolamide (Biary et al. 1994).

Most examples of drug-induced chorea and dystonia are discussed in Chapters 24 and 43. Paroxysmal kinesigenic dyskinesia has been reported after methylphenidate administration, but the relationship is uncertain as it continued long after drug withdrawal and responded to carbamazepine (Gay and Ryan 1994).

Choreic and athetotic movements have been reported in alcoholics (see Chapter 24) and sometimes first appear following alcohol withdrawal. Movements may be most prominent in the face and upper limbs. They can disappear within hours or days, although persistence for up to a year has been reported. The mechanism is uncertain (Mullin et al. 1970, Fornazzari and Carlen 1982). Although these bouts can be brief, they are not usually episodic in the sense that they are recurrent.