33 Other specific causes of symptomatic generalized myoclonus

In this chapter we will consider, in broad groups, other causes of generalized symptomatic myoclonus not dealt with elsewhere. In particular, we will concentrate on post-anoxic myoclonus, Creutzfeld–Jakob disease, viral and post-infectious myoclonus, myoclonus in dementia, toxic and drug-induced myoclonus, and metabolic myoclonus.

In discussing the effects of cerebral hypoxia Courville (1939) noted that ‘muscular twitching of the extremities… at times assumed the proportions of true convulsive movements’. Aigner and Mulder (1960), in a review of 94 patients with myoclonus, noted that two had developed symptoms following an episode of hypoxia at the time of anaesthesia. Swanson et al. (1962), in a review of 67 patients with myoclonus, included two cases that followed cardiac arrest. Hassler (1968) used the term anoxic myoclonus for those patients who suffered cerebral damage in infancy, presumably at the time of birth.

The definitive description of post-anoxic myoclonus was that of Lance and Adams (1963), who described four patients who recovered from hypoxic coma, but who were left with a permanent disabling neurological syndrome dominated by action or intention myoclonus. Their patients had suffered respiratory obstruction or cardiac arrest. They pointed out that generalized spontaneous myoclonus is a feature of the early stage of recovery from an episode of cerebral hypoxia, but that later on myoclonus is precipitated particularly by voluntary movement, more so if co-ordination is required. Passive movement and, in some cases, touch, tendon tap, pinprick, and sound were also effective stimuli. These authors used the term action myoclonus to describe muscle jerks occurring on attempted voluntary movement, and noted that the myoclonus was aggravated more as the task required increasing precision and as the target of aim was approached. They drew attention to the original use of the term action myoclonia by Wohlfart and Hook (1951), who reported 10 patients with myoclonus on movement due to a variety of conditions including Unverricht-Lundborg disease, the Ramsay-Hunt syndrome, and Wilson's disease. Lance and Adams (1963) also commented on the superficial similarity between action myoclonus and intention tremor. They described ‘an arrhythmic fine or coarse jerking of a muscle or a group of muscles in disorderly fashion, excited mainly by muscular activity, particularly when a conscious attempt at precision was required, worsened by emotional arousal, suppressed by barbiturates, and superimposed on a mild cerebellar ataxia’.

In 1971 [a and b] Lhermitte et al. described a detailed pharmacological analysis of the response to a variety of treatments in single cases of post-anoxic myoclonus. Their major observation was that administration of 5-hydroxytryptophan (the precursor of serotonin) almost completely relieved the myoclonus.

A syndrome similar to that following cerebral anoxia has been described after head injury (Hallett et al. 1979) and heat stroke (Marsden et al. 1982).

A period of cerebral hypoxia can produce widespread cerebral pathology of various types. It has proved difficult to disentangle the lesions seen specifically in those who develop the syndrome of post-anoxic myoclonus from those not associated with myoclonus. Only a few pathological studies of this aftermath of cerebral hypoxia have been published (Castaigne et al. 1964 and 1968, Wolf, 1977, Chadwick et al. 1978, Masson et al. 1979, De Lean et al. 1986, Hauw et al. 1986).

Hauw et al. (1986) summarized the pathological findings in four cases of post-anoxic myoclonus, two of which had been described elsewhere (Castaigne et al. 1964 and 1968). The four patients had suffered a period of cerebral hypoxia, which was due to the complications of anaesthesia or shock. The time from onset to death varied from 10 to 70 days. The pathological findings varied from case to case (Table 33.1). The authors acknowledged that it was difficult to distinguish changes due to agonal effects in patients subjected to resuscitation and intensive care from the more long-lasting ischaemic injuries that might be relevant to the development of myoclonus. In this regard, the findings in their cases 3 and 4, who survived longest, were thought to be more significant. Even then, there was no consistent pattern. In case 3, who developed severe action myoclonus after a period of coma for 3 days following cardiopulmonary arrest, the main pathological finding was bilateral necrosis of the mamillary bodies. There were milder changes in the medial nuclei of the thalamus. However, in general, the cerebral cortex, basal ganglia, brainstem, and cerebellum appeared intact. The findings in this case resembled those of Wernicke-Korsakoff encephalopathy.

In case 4 who had a period of prolonged hypotension after cardiac ischaemia and who developed stretch-induced myoclonus, the pathological changes were more widespread. In particular, the cerebellar cortex showed massive Purkinje cell loss and there were marked changes in cerebellar nuclei, particularly the dentate nucleus. Milder changes were found in the medial nuclei of the thalamus and the griseum centrale mesencephali, along with a few ischaemic changes in the inferior olivary nuclei and cerebral cortex.

Hauw et al. (1986) concluded that the only coherent pathology that they could discern from their cases and those previously published was involvement of the thalamus (particularly the medial nuclei), the brainstem raphe system (which is the source of major serotonergic pathways), and the cerebellum.

De Lean et al. (1986) described in detail the pathological findings in the case of a 72-year-old woman who took a barbiturate overdose. On recovering consciousness 36 hours later she had violent spontaneous, action-, noise-, and tactile-induced myoclonus. She remained bedridden for nearly 3 years before dying as a result of choking on food. Pathological examination showed no laminar type damage in the cerebral cortex. The hippocampus and medial temporal lobes showed age-related changes but nothing else. The basal ganglia including the caudate, putamen, pallidum, and subthalamic nucleus were normal, as was the thalamus. There was gliosis in the midbrain and olivary nuclei, but it was difficult to assess neuronal loss. The cerebellar cortex and deep nuclei were normal. The authors acknowledged that it was difficult to separate the effects of ‘age’ from those of anoxic pathology. They reviewed their findings in light of other reports in the literature, including that of Hauw et al. (1986). They concluded that the findings fell into two categories. First, there were cases showing few significant pathological changes. Second, there were others demonstrating multiple areas of anoxic damage with no specific pattern. Indeed, they felt that the pathology seen in the second group was similar to that found in patients with anoxic encephalopathy without myoclonus.

In conclusion, such pathology as exists in humans has been insufficient to identify any changes that appear specific for the development of post-hypoxic myoclonus.

The development of an animal model of post-hypoxic myoclonus has not completely clarified the situation. Following a period of 10 minutes cardiac arrest in rats, seizures and spontaneous myoclonus appear, but they resolve over the first few days, while auditory-induced myoclonus deteriorates and reaches a peak 2 weeks later, before gradually declining (Truong et al. 1994). Global cerebral anoxia in experimental animals causes damage in a number of brain regions, including the cortex, hippocampus, basal ganglia, thalamus, subthalamus, substantia nigra pars reticulata, and cerebellar Purkinje cells (Ichord et al. 1997). The interneurons in the thalamic reticular nucleus are perhaps particularly vulnerable (Freund et al. 1984, Smith et al. 1984, Ross and Graham 1993, Romergryko et al. 2000).

Chadwick et al. (1978) described pathological findings in two cases of post-anoxic myoclonus, with limited biochemical study of the brain. The pathology seen was widespread, and similar to that evident in cases with post-hypoxic encephalopathy without features of myoclonus. Concentrations of serotonin (5HT) and 5-hydroxyindoleacetic acid in cerebral cortex and striatum were normal. They were unable to demonstrate selective damage of the serotonin-containing nerve cells in the dorsal raphe nuclei of the brainstem. In the rat model of post-anoxic myoclonus, significant changes in cortical serotonin and cortical and mesencephalic 5-hydroxyindoleacetic acid appear to have significant correlation with the presence of myoclonus (Matsumoto et al. 1995[a]). 5-HT receptor agonists, particularly those acting on 5-HT2 and 5-HT3 receptors, may diminish myoclonus (Jaw et al. 1994, Matsumoto et al. 1995[b]), although receptor antagonists acting at 5-HT1 and 5-HT2 receptors have been reported to have a similar effect (Pappert et al. 1999, Goetz et al. 2000).

In addition, abnormalities in monoaminergic, GABAergic, and glutamatergic systems have been postulated. Using a variety of ligands microinjected into the ventricles of normal rats Matsumoto et al. (2000) found that only GABAA antagonists produced myoclonus, and microinjections of GABAA antagonists into the thalamic reticular nucleus or caudate resulted in myoclonus. GABA uptake inhibitors attenuate post-hypoxic myoclonus in rats (Jaw et al. 1996[a]). Similar results have been reported with glutamate receptor antagonists (Jaw et al. 1996[b]). The significance of these observations, if any, to the human disorder remains uncertain.

Lance and Adams (1963) in their original description of the syndrome noted that the myoclonic jerks were associated with a brief electromyographic (EMG) discharge in affected muscles, usually followed by a silent period or pause in the continuing EMG activity. The pauses were felt to be as significant as the myoclonic discharges themselves, for they were associated with postural lapses, including falling when attempting to walk. Thus, their patients had not only positive myoclonus but also negative myoclonus (asterixis).

The myoclonic jerks were associated with spikes or runs of spikes often increasing in amplitude as ‘an augmenting series’ in three of the four patients that they described. These electroencephalographic (EEG) changes were maximal over the sensorimotor cortex contralateral to myoclonic jerking. The latent interval from the initial positive deflection of the cortical spike to the myoclonic jerk was some 12 ms for biceps, 16 ms for wrist extensors, and 32 ms for quadriceps. They calculated that the spinal cord efferent conduction velocity was about 30–40 m/s.

Lance and Adams (1963) proposed that the myoclonus was caused by a synchronous and repetitive discharge of thalamocortical pathways, because of the similarity of the augmenting series of myoclonic jerks to the augmenting response which could be obtained by stimulation of the thalamus. However, Lhermitte et al. (1971[b]) undertook stereotactic recording from the ventrolateral thalamus, basal ganglia and internal capsule in a patient with post-anoxic myoclonus, and found that activity in the thalamus followed the cortical spike. Subsequent stimulation and then destruction of that thalamic nucleus had no effect on the myoclonus. However, myoclonic jerks could be evoked by electrical stimulation of the motor cortex or internal capsule.

Hallett et al. (1979) studied the physiology of action and stimulus-sensitive myoclonus in patients with post-anoxic myoclonus in detail [see Hallett (2000) for a review]. They described one form of myoclonus as having characteristics suggesting origin in the somatosensory cerebral cortex – cortical action myoclonus and cortical reflex myoclonus. The evidence presented by Hallett et al. (1979) (summarized in Table 33.2) indicated that this form of post-anoxic myoclonus arose from focal discharge in the sensorimotor cortex propagated down fast corticomotor neuron pathways to the muscles involved in the local jerk (Fig. 33.1). A single discharge in part of one sensorimotor cortex would generate a contralateral focal myoclonic jerk. Multiple sensorimotor cortical discharges would generate multifocal jerks.

Spontaneous jerks were due to spontaneous discharges in sensorimotor cortex. Action-induced jerks were the result of voluntary movement triggering a sensorimotor cortical discharge. Stimulus-sensitive jerks were caused by a variety of peripheral somaesthetic stimuli (touch, pinprick, passive stretch) triggering a cortical discharge.

Peripheral stimuli were found to generate giant somatosensory evoked potentials (SEPs), recorded over the contralateral sensorimotor cortex in some patients (Rothwell et al. 1984, Obeso et al. 1986[a]). The latency of the giant SEP (upper limb, 18–25 ms; lower limb, 30–35 ms) is approximately half that of the latency of the reflex muscle jerk produced by the stimulus (upper limb, 36–50 ms; lower limb, 60–70 ms).

Details of the giant SEPs reveal that the arrival of the sensory volley in the cortex, reflected by the N1 component (which is analogous in normal subjects to the N20 response), is usually of normal size. The abnormal response begins with the first positive component (the P1, corresponding in normals to the P₂₅) followed by the large N2 (corresponding to the normal N₃₃). The peak of the P1 SEP response precedes the EMG burst responsible for the focal myoclonic jerk with a latency identical to that produced by direct-electrical or magnetic stimulation of the sensorimotor cortex. In other words, the cortical myoclonus originating in sensorimotor cortex is transmitted to the involved muscles via fast conducting corticomotoneuron pathways.

Werhahn et al. (1997) in a study of 14 patients with chronic post-hypoxic myoclonus found that all had multifocal action-myoclonus, 13 showing a time-locked cortical correlate preceding the limb jerks on back-averaging of EEG. Two also had pathologically enlarged somatosensory evoked potentials and a C-reflex.

Cortical myoclonus is not the only type of jerking that occurs in patients with post-anoxic myoclonus. Hallett et al. (1977) demonstrated that some patients with the syndrome may exhibit brainstem reticular myoclonus (Fig. 33.2) (see Chapter 31). Brown et al. (1991[a]) showed that there are patients with post-anoxic myoclonus who exhibit an exaggerated startle response (see Chapter 31). Many patients after cerebral anoxia may exhibit combinations of cortical and brainstem reticular myoclonus (Chadwick et al. 1977; Brown et al. 1991[b], Oguro et al. 1997, Werhahn et al. 1997) or cortical myoclonus and exaggerated startle responses (Brown et al. 1991[a], Werhahn et al. 1997).

Another type of post-anoxic myoclonus was described by Thompson et al. (1989) in seven cases. These patients had generalized jerks synchronous in comparable muscle groups, provoked by movement, although sometimes they were confined to the limb voluntarily activated. The jerks were preceded by generalized runs of cortical spike-waves, which were of maximum amplitude over the vertex and the cortex contralateral to the activated limb. The cortical spike-waves were time-locked to the following jerks. Muscles activated in the generalized jerks fired with a rostro-caudal pattern, suggesting a cortical rather than brainstem origin. SEPs were of normal size. The bilateral myoclonic jerks, and the time-locked generalized spike-waves preceding them, indicated that both cerebral hemispheres were synchronously activated, presumably by a subcortical discharge. Thompson et al. (1989) called this subcortical-cortical myoclonus (Fig. 33.3). Whether this type of post-anoxic myoclonus is driven by thalamocortical discharge, as suggested by Lance and Adams (1963), is not known.

Thus there are a variety of pathophysiological types of myoclonus in patients with post-anoxic encephalopathy (see Table 33.3). In addition, myoclonus may occur in the acute stage, shortly after an episode of hypoxia while the individual is still comatose. It is characterized by generalized and often massive jerks (Young et al. 1990, Van Cott et al. 1996). These can be associated with seizures. The myoclonic jerks consist of spontaneous or stimulus-sensitive jerks involving the face, limbs, and axial muscles. The EEG shows generalized spike and polyspike activity at the time of the jerks and can be silent between bouts (burst suppression). The physiology of this type of myoclonus is uncertain, but it may arise from a brainstem generator (Hallett 2000).

Fahn (1979) reviewed a total of 29 papers in the literature describing 59 cases of post-anoxic myoclonus. In subsequent reviews (Fahn 1986, Frucht and Fahn 2000) he added further cases, bringing the total up to 122. The causes of cerebral hypoxia in these 122 cases are shown in Table 33.4. Respiratory arrest, particularly related to asthma, was the commonest cause, accounting for 36%. Anaesthesia and surgical accidents were the next most frequent and made up 32%. The duration of coma after cerebral hypoxia ranged from less than a day to 120 days (Table 33.5). The median was 4.5 days and the vast majority lasted less than 2 weeks. The age of the patients varied from infants (rarely) through childhood and the rest of life.

During the period of initial coma 31 of 43 patients had generalized seizures (their presence or absence was not stated in the remaining cases) and 34 of 47 patients had widespread spontaneous myoclonus. Generalized myoclonus during the coma may be very severe and disrupt mechanical ventilation. Asynchronous distal limb myoclonus also can be seen (Snyder et al. 1980[a]). As mentioned above, the myoclonus during the period of coma may be stimulus-sensitive (Niedermeyer et al. 1977, Wolf 1977). Myoclonic status in this situation has been associated with a very poor prognosis and of 40 such patients with this disorder reported by Wijdicks et al. (1994) all died. Occasional patients, however,

have had a similar picture, but survived (Arnoldus and Lammers et al. 1995).

As the patients recover consciousness (Snyder et al. 1980[b]) generalized spontaneous myoclonus tends to lessen, but action-induced and stimulus- induced myoclonus becomes more and more prominent. Some patients may have myoclonus only on movement (Thompson et al. 1989). Without treatment, such myoclonus usually persists, and only rarely does spontaneous remission occur after 5 or 6 months (Chee and Poh 1983). The longest duration of post-anoxic myoclonus described has been 44 years (Goldberg and Dorman 1976).

During the chronic stage of the illness, after recovery from coma, the syndrome is dominated by myoclonus of variable severity. At its worst, patients are bedridden. They may exhibit spontaneous myoclonus, but it is much worse on any attempted voluntary movement. As a result their speech may be unintelligible, they may have considerable difficulty swallowing, their hands are rendered useless for feeding, dressing, or any skilled task, and they cannot stand or walk. A characteristic feature when they attempt to stand is irregular or semi-rhythmic bouncing of the legs beneath them, as a result of postural lapses. In addition to the action-induced myoclonus, a variety of stimuli may provoke myoclonic jerks, particularly if the stimulus carries an element of surprise. These include visual threat, a sudden noise, touch, or pinprick to the extremities or mantle area, and passive muscle stretch. The myoclonus may be localized to a limb moved or stimulated (cortical action or reflex myoclonus), or may be generalized (either reticular reflex myoclonus or subcortical-cortical myoclonus). A single large myoclonic jerk may throw the seated patient off balance, and subsequent voluntary movements initiated to compensate for this may themselves provoke action myoclonus, appearing as a crescendo of myoclonic jerks. Sometimes this is so rhythmic that it gives the impression of a jerk oscillation or shuddering attack, developing in response to a stimulus or movement, then fading away over a few seconds. This phenomenon has been called oscillatory myoclonus (Obeso et al. 1983[a]).

In less severely affected cases, myoclonus may only be evident on attempted movement or in response to peripheral stimuli. In our experience there is lessening of the severity of the myoclonus in the first 6 months after the acute episode. Thereafter progress is slower. Werhahn et al. (1997) reported 14 patients who were first seen a mean of 2.5 years after the hypoxic event and followed up for 3.7 years. Late improvement in the myoclonus and the level of disability occurred in all but one patient (Fig. 33.4). Five were eventually able to walk unaided and three could discontinue antimyoclonic medication.

In general, patients with post-anoxic myoclonus recover reasonably from the other consequences of the period of cerebral hypoxia. Confusion and subsequent cognitive impairment may resolve completely, although a proportion is left with an amnesic syndrome of variable degree. Of the 14 patients reported by Werhahn et al. (1997) cognitive defects were found in seven and were generally mild. The verbal IQ on the patients tested was a mean of 87.2 (range 78.3–94.5). One patient scored at a severely demented level on the Mini-Mental State Examination. Initial incontinence also usually resolves and pyramidal signs often disappear. Cerebellar ataxia may persist, but it is exceedingly difficult to distinguish true cerebellar ataxia from the effects of the action-induced myoclonus. Only when the latter responds to drug treatment does it become evident that there is an underlying cerebellar dysarthria, incoordination, and intention tremor of the arms, and ataxia of gait. In those less affected, there may be no residual cerebellar ataxia. In the series described by Fahn (1979 and 1986, Frucht and Fahn 2000) dysarthria was present in 29 of the 32 cases in which this symptom was mentioned, and residual ataxia was evident in 30 of 39 patients. Speech was regarded as being normal, or only slightly abnormal, in 10 out of 13 of Werhahn et al. (1997) patients at presentation and five improved during follow-up. Eleven had no neurological signs when last seen, apart from myoclonus, seven were able to walk unaided, seven were independent, and two needed only a little help, mainly when toileting. The other four were largely dependent on the assistance from others.

Seizures may persist in the chronic syndrome. Amongst 56 patients reveiwed by Fahn (1979 and 1986, Frucht and Fahn 2000) in which seizures were specifically mentioned, only 24 were seizure-free.

EEG findings have been described in many reports (Lance and Adams 1963, Chadwick et al. 1977, Fahn, 1979, Thompson et al. 1989). In some cases the EEG is normal. Other patients may exhibit spike and spike-slow-wave complexes, with background slowing. Such spike or spike-wave complexes may be triggered by movement, are most prominent at the vertex, and may occur in runs (Thompson et al. 1989, Werhahn et al. 1997). The spike complexes may or may not appear time-locked to the myoclonic EMG bursts in the routine record. However, back-averaging reveals a cortical correlate preceding the jerk, over the contralateral sensorimotor cortex in many cases, or sometimes as a generalized phenomenon. A slightly different pattern seen in some cases consists of cortical fast activity (Thompson et al. 1989) (see ‘Pathophysiology’).

The movement-induced myoclonus may have a semi-rhythmic quality, making it difficult to distinguish from cerebellar intention tremor. The duration of EMG bursts in positive myoclonus is around 15–40 ms, which may occur in salvos at 16–20 Hz (Thompson et al. 1989). The duration of silent periods varies from 100 to 340 ms (Lance and Adams 1963) and they are associated with slow waves in the EEG.

Not every patient has enlarged SEPs (Chadwick et al. 1977). However, the P1—N2 component may be enlarged, particularly in those with focal cortical action or reflex myoclonus (Rothwell et al. 1984, Obeso et al. 1986[b], Kakigi and Shibasaki 1987).

Brain imaging (CT or MRI scan) may be normal or may reveal evidence of atrophy or focal ischaemic change. No consistent focal abnormality has been described. Of the 12 patients of Werhahn et al. (1997) who had imaging, four showed mild cerebral cortical or cerebellar atrophy, four had probable small infarcts, which were cerebellar in one case, and four showed no abnormality. Positron emission tomography (PET) imaging showed a significant bilateral increase in glucose metabolism in the ventrolateral thalamus and pontine tegmentum in patients relative to controls (Frucht et al. 2004).

Guilleminault et al. (1973) described a decrease in the CSF concentration of 5-hydroxyindoleacetic acid in one patient, a finding confirmed by Van Woert and Sethy (1975), Van Woert et al. (1977), and Chadwick et al. (1977 and 1986). Some patients, however, have normal levels (Frucht and Fahn 2000). Chadwick et al. (1977) also measured CSF tryptophan levels and found them to be normal. The number of patients who have undergone extensive CSF analysis is small and while they have suggested damage to the raphe system, it remains uncertain whether this is really connected with myoclonus.

Lance and Adams (1963) found little benefit from anticonvulsants such as phenobarbitone or primidone. In general, the response to phenytoin has also been disappointing (Growdon et al. 1976), as has that to carbamazepine (Boudouresques et al. 1971). These conclusions were supported by a literature survey of Frucht and Fahn (2000). However, such routine anticonvulsant treatment may be successful in controlling generalized tonic-clonic seizures.

More successful has been the use of sodium valproate (Carroll and Walsh 1978, Bruni et al. 1979, Fahn 1979, Rollinson and Gilligan 1979). Frucht and Fahn (2000) noted that it had been helpful in controlling myoclonus in 45% of cases reported in the literature. Benzodiazepines, in particular clonazepam, have also been found to be effective (Boudouresques et al. 1971, Chadwick et al. 1975 and 1977, Goldberg and Dorman 1976, Fahn et al. 1979). In Frucht and Fahn's (2000) survey 51% of patients who had been tried on clonazepam were said to have shown significant improvement in myoclonus. Other benzodiazepines, however, may not be so useful and Frucht and Fahn (2000) noted that only two out of 14 patients who had been given diazepam were said to have benefited and nitrazepam appeared ineffective. Obeso et al. (1989) pointed out that the best control of myoclonus can be obtained by the combination of clonazepam and sodium valproate. They also advocate the addition of primidone in recalcitrant cases, and piracetam.

Piracetam was first noted to have an effect in post-anoxic myoclonus by Terwinghe et al. (1978) and Cremieux and Serratrice (1979). Subsequent experience (Obeso et al. 1989) in open-labelled studies confirmed its value in a variety of causes of myoclonus, including that following cerebral hypoxia. Brown et al. (1993) have shown in a double-blind placebo-controlled withdrawal trial that it is an effective anti-myoclonic agent. The best control of myoclonus may be obtained with a combination of piracetam, sodium valproate, and clonazepam. Although some authors favour the addition of primidone in recalcitrant cases (Obeso et al. 1989), the overall evidence for benefit is not convincing (Frucht and Fahn 2000).

Lhermitte et al. (1971a and 1972) first demonstrated the effectiveness of oral 5-hydroxytryptophan, the immediate precursor of serotonin, in post-anoxic myoclonus. This observation was confirmed by Guilleminault et al. (1973). The treatment was developed by Van Woert and colleagues (1975 and 1977) and Chadwick et al. (1975 and 1977). De Lean et al. (1976), Growdon et al. (1976), Magnussen et al. (1978), and Thal et al. (1980) also reported efficacy of 5-hydroxytryptophan in post-anoxic myoclonus [see Chadwick et al. (1986) and Van Woert et al. (1986) for reviews].

Optimum benefit was found by the use of the levo-rotatory form of 5-hydroxytryptophan (1000–2000 mg/day) combined with a peripheral dopa-decarboxylase inhibitor such as carbidopa (100–200 mg/day) (Lhermitte et al. 1975, Van Woert et al. 1986). Not every case has responded (Chadwick et al. 1977). L-tryptophan combined with a monoamine oxidase inhibitor to increase brain serotonin action also can reduce myoclonus (Chadwick et al. 1975 and 1977, Van Woert et al. 1977), but in general has been found to be less effective than 5-hydroxytryptophan combined with carbidopa.

The most common side-effects of 5-hydroxytryptophan, despite co-administration with carbidopa, have been anorexia, nausea, vomiting, and diarrhoea (Van Woert et al. 1986). Although these side-effects tend to lessen with long-term treatment (Van Woert et al. 1977), they have often continued to be troublesome. Other side-effects have included mental changes including euphoria, hypomania, restlessness, rapid speech, insomnia, and agitation.

In general, the success of treatment with a combination of sodium valproate, clonazepam, and piracetam (Obeso et al. 1989) has led to less interest in the effects of 5-hydroxytryptophan combined with carbidopa, which is now rarely employed. However, the latter method of treatment has been of considerable theoretical interest because it suggests that the syndrome of post-anoxic myoclonus is due to deficient serotonin neurotransmitter action within the central nervous system. This concept is reinforced by the finding of low concentrations of the metabolite of serotonin, 5-hydroxyindoleacetic acid, in CSF in many such patients.

A variety of other drugs have been employed in an attempt to control post-anoxic myoclonus (Table 33.6). Of interest is the observation that serotonin-uptake inhibitors (paroxetine and fluoxetine) can be of benefit in some cases, as can monoamine oxidase inhibitors. Occasionally, the serotonin antagonist methysergide can produce improvement. Levodopa was found to be of benefit by some observers, but its effect was not as great as that of 5-hydroxytryptophan. Dopamine antagonists, whether directly acting, such as phenothiazines or haloperidol, or indirectly acting, such as tetrabenazine, generally have been ineffective, as have cholinergic and anticholinergic drugs.

Unilateral thalamotomy had no effect on the myoclonus in the case described by Lhermitte et al. (1971[b]). A case reported by Fahn (1979) had bilateral thalamotomies without benefit. Chronic cerebellar stimulation (Cooper et al. 1976) in a single patient was reported to produce some decrease in the myoclonus.

Creutzfeldt (1920) described a 23-year-old female whose progressive illness began at age 16 years with difficulty in walking. By age 20 she had mental changes and thereafter she progressively demented, developing also ataxia, seizures, and paralysis. At autopsy there was cerebral atrophy with profound neuronal loss in the cortex, thalamus, dentate nucleus, and spinal cord. Reactive gliosis and status spongiosus were not mentioned. The following year, Jakob (1921[a] and [b]) described three further cases, two of whom exhibited involuntary movements towards the end of their illness. Heidenhain (1929) reported three further cases with predominant cortical blindness or visual agnosia and myoclonus; he also noted the characteristic microscopic spongiform change in the cerebral cortex. In 1936 Gerstmann, Straussler, and Scheinker described a family with an autosomal dominant condition causing cerebellar ataxia and dementia, associated with amyloid plaques and apparent olivopontocerebellar atrophy.

In 1968, Gibbs et al. established that Creutzfeldt–Jakob disease (CJD) could be transmitted to chimpanzees by intracerebral inoculation of brain tissue from patients. In 1981[a], Masters et al. found that Gerstmann-Straussler-Scheinker disease also was transmissible. Thus CJD was established, along with Kuru and Gerstmann-Straussler-Scheinker syndrome (GSS), as a transmissible encephalopathy, related to scrapie of sheep and goats, mink encephalopathy, chronic wasting disease of mule deer and elk, and bovine spongiform encephalopathy. All share common features (Table 33.7). These transmissible diseases are due to infection with unusual agents, christened prions (proteinaceous infectious particles containing no nucleic acid) (Prusiner 1982, Hsiao and Prusiner 1990). In 1986 fatal familial insomnia (FFI) was described by Lugaressi et al. and this was subsequently also shown to be caused by prions. The prion diseases of humans and animals along with their mechanisms of pathogenesis are shown in Table 33.8.

In the late 1980s there was an increase in the incidence of bovine spongiform encephalopathy in Britain which peaked in 1992 (Fig. 33.5). It has been estimated that almost one million cattle were infected with prions but most did not develop the

disease because they were slaughtered. It has been postulated that a change in the rendering process by which offal was converted into a high-protein nutritional supplement and fed to cattle resulted in them becoming infected (Prusiner 1998). In 1996 the National CJD Surveillance Unit in Britain reported the appearance of an apparently new variant (vCJD) (Will et al. 1996[a]) and the following year the features of this were more fully delineated (Zeidler et al. 1997[a] and [b]). There is now convincing evidence to link the emergence of vCJD to the outbreak of spongiform encephalopathy in cattle (Stewart and Ironside 1998) and similar outbreaks of vCJD have not occurred elsewhere in Europe (Lundberg 1998, Pals et al. 1999).

Prions are transmissible particles that are devoid of nucleic acid and are composed entirely of a modification of a normal cellular protein (PrP^C). PrP^C can be converted into the infectious prion form (PrP^Sc) through a post-translational process whereby a portion of its α–helical and coil structure is refolded into a β-sheet. This structural transition is associated with marked changes in the physico-chemical properties of the protein. PrP^Sc then acts as a template upon which more PrP^C is refolded into PrP^Sc molecules. PrP^Sc thus triggers a type of cascade process in which it is self-replicated. PrP^Sc is resistant to degradation and its accumulation results in formation of amyloid and the neurological symptoms.

The human prion diseases occur in inherited, acquired, and sporadic forms (see Table 33.9). The majority of patients have sporadic forms (about 85–90%). Approximately 10–15% are inherited and associated with coding mutations in the PRNP (PrP gene). These form the basis of inherited Creutzfeld–Jakob disease (iCJD), GSS, and FFI. The PRNP gene is located on chromosome 20. It contains two exons and encodes a glycoprotein, the non-pathogenic cellular human prion protein, PrPC. In CJD, there is an unstable region of five variant tandem octapeptide coding repeats between codons 51 and 91. Overall, more than 20 mutations of the prion protein gene have been recognized in human prion diseases. In addition there are recognized polymorphisms which may differ in geographic areas, for example Val/Met129. There is no known genetic abnormality associated with sporadic Creutzfeld–Jakob disease (sCJD) (Prusiner 1998).

How CJD is acquired remains uncertain. Theories include somatic mutation of the prion protein gene, spontaneous conversion of PrP^C into PrP^Sc, or infection with prions from animals or humans. CJD has been transmitted from man to man iatrogenically via corneal implants, EEG depth electrodes (Davanipour et al. 1985), dura mater grafts (Thadani et al. 1988, Nisbet et al. 1989), and from growth hormone and gonadotropin derived from contaminated human pituitary glands (Brown et al. 1985, Cochius et al. 1992). GSS also has been transmitted from contaminated human pituitary gonadotropin. Over 90 patients are known to have been infected with CJD from pituitary extracts (Prusiner et al. 1998) and over 70 from dural grafts (Lang et al. 1998), these being the commonest means of iatrogenic disease transmission. Overall, approximately 200 patients worldwide have died or suffer from variant Creutzfeldt–Jakob disease. Other potential modes of transmission have been suggested, including surgical sutures and ocular tonometers (Alter 2000). There is also concern about the possibility of spread via blood transfusion. PrP^Sc is known to accumulate in lymphoid tissue in vCJD, raising the possibility that circulating lymphocytes might also be capable of transmitting the disease (Stewart and Ironside 1998). Such lymphoid accumulation does not seem to be a feature of sCJD. The amino acid sequence of PrP^Sc corresponds to that encoded by the PrP gene of the last mammalian host in which is replicated. It thus seems significant that all patients described with vCJD have been homozygous for methionine at codon 129, which is the same as the situation found in cattle. This contrasts significantly with normal control individuals who are approximately 37% homozygous for methionine and with sCJD with the figure of 79% (Stewart and Ironside 1998, Allroggen et al. 2000).

In CJD the brain may grossly appear normal or show cortical atrophy and ventricular dilatation. The cerebellum may or may not be involved. Microscopically, there is widespread spongiform change, astrocytosis, and gliosis, with varying degrees of neuronal loss (Fig. 33.6). Spongiform change sometimes is absent in atypical cases which transmit to experimental animals (Manuelidis 1985). All regions of the cerebral cortex are involved, but the degree and distribution vary from case to case. In typical cases the neocortex, striatum, thalamus, brainstem grey matter, and cerebellum are all involved. The hippocampus may be relatively spared. The spongiform change, which is distinct from hypoxic/ischaemic damage, is due to intraneuronal cytoplasmic vacuoles, leaving the nucleus intact. The gliosis is prominent in those areas with severe neuronal loss.

In GSS the neuropathology is similar to that of CJD, but there is also widespread amyloid plaque formation throughout the brain (Masters et al. 1981[b]). Amyloid can also be seen in CJD, although it is generally less marked in sCJD. It may be particularly florid in vCJD where numerous large fibrillary amyloid plaques composed of PrP^Sc may be found in both cerebral and cerebellar cortex, usually surrounded by spongiform change (Will et al. 1996[a], Ironside 1997, Ironside and Bell 1997). The occipital cortex is especially involved and spongiform change is very prominent in the caudate, while severe cell loss and gliosis tend to be marked in the posterior thalamic nuclei. Marked amyloid plaque formation has also been a feature of CJD associated with dura mater grafts (Shimizu et al. 1999). Immunohistochemistry with antibodies against prion protein may be especially valuable (Fig. 33.7) and show deposition of PrP^Sc not appreciated using other techniques (Kretzschmar et al. 1996).

GABAergic neurons which stain positively for parvalbumin appear to be particularly affected in CJD with the development of abnormal dendritic trees followed by selective and severe loss. These changes occur at an early stage in the disease (Ferrer et al. 1993, Guentchev et al. 1998). As these are important inhibitory neurons, their loss may play a part in the development of myoclonus.

The neuropathology in FFI particularly shows neuronal loss and gliosis in the thalamus without amyloid deposits. Spongiosis may occur in cerebral cortex (Delisle et al. 1999).

The incidence of CJD is usually in the order of 1 per million people per year. Approximately 10% of these may be inherited CJD. The sexes tend to be affected approximately equally (Lundberg 1998, Pals et al. 1999). Generally speaking, no obvious racial or geographical predominance has been noted, although occasional clusters of sCJD have been reported, in addition to the appearance of vCJD in Britain. However, molecularly, there is variation and point and insertion mutations in the PRNP gene which vary in frequency between countries. The incidences of GSS and FFI are much less. In a study of 72 cases of pathologically confirmed spongiform encephalopathy in Sweden all were felt to be CJD, apart from one example of GSS (Lundberg 1998).

CJD usually presents in the sixth decade, but cases in the third and ninth decades have been described. In a large series from France, the average age of onset was 60 years and the mean duration of illness was 7.6 months (Brown et al. 1987). This is similar to other studies. About two-thirds of cases will be dead within a year. However, a minority may have a clinical course lasting 2 or more years (Brown et al. 1984). Pals et al. (1999) reviewed 800 patients from Belgium. The mean age at death was 63 years and the median disease duration was 9 months. Lundberg (1998) in a Swedish study reported that 62% died within 6 months, 13% within 6–12 months, 12% within 1–2 years, and 13% survived over 2 years. However, among the inherited forms this may vary according to the genetic deficit, for example familial CJD due to the E200K mutation has a relatively short mean disease duration (6 months) compared to CJD patients with a 24-base pair expansion (5–9 extra repeats) (Goldfarb 2007).

Cases of vCJD have had a much younger age of onset, with a mean of 29 years. Most patients survive for longer after symptom onset, the median having been about 14 months (Zeidler et al. 1997[a]).

GSS differs from CJD in that many cases are inherited as an autosomal dominant trait, onset is earlier in the third or fourth decades, and the course of the illness lasts several years (Hsiao and Prusiner 1990).

sCJD usually presents with subtle symptoms of tiredness, anxiety and fear, depression, insomnia, confusion, and odd sensory and visual complaints. Occasionally the disorder commences with hallucinations (Heinz et al. 1995). Soon dementia becomes evident, which is rapidly progressive, sometimes to stupor within 4 months or so. Memory, visuospatial functions, praxis, and language are destroyed. Finally, the unfortunate patient is mute, rigid, decerebrate or decorticate, and bedridden. Severe involvement of the occipital cortex leads to cortical blindness (Heidenhain variant). Loss of anterior horn cells may cause muscle weakness and atrophy (amyotrophic variant) (Worrall et al. 2000).

Myoclonus is the most characteristic finding after dementia and occurs at some time in the course of the illness in around 80–100% of cases. Myoclonus develops earlier in patients with methionine/methionine (MM) or methionine/valine (MV) alleles at codon 129 of the prion protein gene, PRNP, and with Type I PrPSc (see later and see Maltête et al. (2006) for review). Myoclonus gets worse as the disease progresses but tends to disappear in the end stage. Myoclonus may be confined to one part of the body or may become multifocal, occurs spontaneously, and is sometimes rhythmic. Such myoclonus often can be triggered by touch, pinprick, or muscle stretch. Alternatively, there may be generalized jerks and exaggerated startle to sudden noise or visual stimuli. Conjugate involuntary jerky eye movements may also occur. Pyramidal, cerebellar, and extrapyramidal signs and seizures occur but are less common. To elaborate, dystonia may occur and is listed in the revised diagnostic criteria for possible vCJD (Criteria II D, World Health Organization, 2001), but has also been observed in many cases of sporadic or familial CJD, either as an isolated feature or as part of a complex movement disorder (Maltete et al. 2006). Chorea may occur in late stages of vCJD. Tremor is often an early feature of sporadic CJD. Parkinsonism, or more precisely akinetic mutism, is a characteristic feature of the terminal stages of sporadic CJD or vCJD. Parkinsonian presentations, progressive supranuclear palsy-like and corticobasal degeneration-like, have also been described.

Attempts have been made to link the genetic and biochemical features to the clinical appearance in sCJD. The translation product of the PRNP gene is polymorphic at residues 129 (methionine or valine) and 219 (glutamic acid or lysine). The 129 codon particularly seems to affect phenotypic expression. In addition, the chemical properties of the PrP^Sc appears to be important. Two types of protease-resistant forms have been identified that differ in size and glycosylation. Approximately 50% of Caucasians are heterozygous at codon 129 whereas this applies to only 15% of patients with sCJD, the large majority of whom are homozygous for either allele. Parchi et al. (1996) identified four groups of patients with sCJD showing distinct clinicopathological features. The typical myoclonic variant and the Heidenhain variant, which was found in 11 of 19 subjects, had methionine homozygosity and Type I protease-resistant PrP^Sc. These patients showed a rapid clinical course with early dementia, myoclonus, and periodic sharp waves on EEG. Pathologically they showed mild to moderate involvement of the cerebral cortex with the brunt of the change occurring in the occipital lobe, striatum, thalamus, and cerebral cortex, while the brainstem, hippocampus, and hypothalamus were virtually spared. They argued that there was little difference between these two disorders and that the apparent lack of visual disturbance in the typical myoclonic variant, in spite of marked pathology in the occipital lobes, might result from the early dementia.

The patients who were also homozygous for methionine but had Type II protease-resistant PrP^Sc had symptoms of longer duration, without sustained myoclonus or periodic sharp waves on EEG. They had more severe histological changes. Methionine/valine heterozygotes and valine homozygotes, both with Type II PrP^Sc, had a short history, with most pathology affecting the basal ganglia, brainstem, and cerebellum. These heterozygotes also showed Kuru-like plaques and both cognitive and cerebellar signs. The valine homozygotes presented with cerebellar involvement and dementia occurred later. In addition, valine homozygosity in association with Type I PrP^Sc has been described presenting as a syndrome of cerebral cortical impairment progressing to dementia, without ataxia, myoclonus, or periodic complexes on EEG. Cortical and striatal lesions predominate and brainstem and cerebellum are relatively spared (Worral et al. 1999). Overall, these conclusions generally seem valid (Tranchant et al. 1999), but overlap in variations can be expected (Samman et al. 1999).

While psychiatric symptoms occur early in the clinical course of sCJD in only about one third of patients, they are almost invariable in vCJD (Zeidler et al. 1997[b]). The majority of the latter are likely to be diagnosed as suffering from depression, although occasionally delusions and hallucinations point to a psychosis. With time, however, neurological symptoms develop. Early somatic sensory symptoms tend to be prominent with dysaesthesiae, paresthesiae, and feelings of coldness in the lower limbs, often with associated pain. Incoordination, ataxia, myoclonus, and pyramidal features subsequently develop, along with primitive reflexes. Upgaze paresis, an uncommon feature in sCJD, has been present in about one third of patients (Zeidler et al. 1997[a], Stewart and Ironside 1998).

GSS presents in middle life, around the age of 40 years (range 19–66), as a cerebellar ataxia, with incoordination of limbs and gait, dysarthria, and dysphagia. Dementia occurs late in the illness. Seizures, myoclonus, pyramidal and extrapyramidal features sensory signs, and amyotrophy may occur. Progression occurs more slowly than in CJD, death occurring on average 5 years after the onset (range 1–11).

Several different point mutations of the prion protein gene have been reported to underlie GSS. Often these involve codon 102 (CCG→CTG) leading to substitution of leucine for proline. This has been particularly associated with an ataxic form of the disease, while a different mutation at codon 117 (GCA→GTG), resulting in the substitution of valine for alanine, has been linked to a predominantly dementing illness, although there is significant variability (Mastrianni et al. 1995). By and large, myoclonus is less prominent than in CJD.

FFI is associated with a mutation at codon 178 of the PRNP gene co-segregating with the methionine polymorphism at codon 129 of the mutated allele. Symptoms usually start between 20 and 60 years of age and patients survive between about 6 months and 2 years (Manetto et al. 1992, Almer et al. 1999). The disorder is characterized by disruption of circadian rhythms, insomnia, and dysautonomia, along with a variety of motor disturbances including dysarthria, ataxia, corticospinal tract involvement, and myoclonus (Delisle et al. 1999). There is considerable phenotypic variation between the inherited prion disorders and some families show features of CJD, GSS, and FFI (Zerr et al. 1998[a], Harder et al. 1999).

Routine CSF analysis is usually normal, although protein content may be raised. Specific analysis for 14-3-3 protein is usually positive. Several studies have suggested a sensitivity of 90–95% (Hsich et al. 1996, Zerr 1998[b], Beaudry et al. 1999). False positives, however, are not uncommon and other disorders causing impairment of cognition and myoclonus, including hypoxic encephalopathy, Hashimoto thyroiditis associated encephalopathy, metabolic encephalopathy, and paraneoplastic neurological disorders may have elevated levels (Saiz et al. 1998 and 1999, Hernandez 2000). The presence of this protein, however, has been said to discriminate better between CJD and other rapidly progressive dementias than either MRI or EEG (Poser et al. 1999). The 14-3-3 protein has also been reported in iCJD. It is probably less frequently positive in vCJD (Will et al. 1996[b], Zeidler 1997[a]). In addition, neuron-specific enolase may be present and may have similar sensitivity for CJD but is not so specific (Beaudrey et al. 1999, Brandel 1999). The presence of an astroglial protein, S100, has also been proposed as a useful marker and is perhaps even more sensitive for CJD than 14-3-3 protein, although it is less specific (Otto et al. 1997, Beaudrey et al. 1999). In addition, it has been claimed that high concentrations of S100 protein in the serum may be useful in diagnosis (Otto et al. 1998).

In CJD the EEG is abnormal at some stage of the illness, usually with generalized slowing and pseudoperiodic sharp waves. There is initial slowing of background rhythms, followed by the development of frontal intermittent rhythmic delta (FIRDA). The evolution of periodic discharges, which mature into sharp waves or spike-waves repeated every 0.5–1.5 seconds or so, generally occurs at a late stage in the disease (Fig. 33.8) (Burger et al. 1972, Bortone et al. 1994, Hansen et al. 1998). The periodic discharges have a variable relation to the myoclonic EMG bursts: EEG discharges may or may not be associated with EMG bursts, sometimes the two are associated, but on occasions the EMG burst may precede or follow the EEG discharge. Back-averaging may reveal a contralateral sharp wave preceding the myoclonic EMG burst by 50–85 ms (Shibasaki et al. 1981). The interval between the cortical correlate and the myoclonus is too long for conduction in fast corticomotoneuron pathways. Both the myoclonus and the periodic synchronous EEG discharge may be due to discharge in deep brain structures, perhaps the rostral brainstem (Nelson and Leffman 1963). Rayport (1963) described scalp, pial, and intracerebral recordings in two cases of CJD with periodic (0.5–20 times per second) surface positive-negative waves or triphasic waves, but no definite intracerebral source was localized. It should be emphasized that periodic EEG discharges are not specific to CJD.

Flash electroretinograms may show a significant drop in the amplitude in the B1 wave and the ratio of the B/A waves. Flash visual evoked potentials are of normal latency, but amplitude may be increased, particularly if there is myoclonus. In the Heidenhain variant giant visual evoked potentials can be recorded during the early stages of the disease, even in the absence of myoclonus, but subsequently become diminished. Latency of pattern visual evoked potentials may initially be increased (de Seze et al. 1998, Finsterer et al. 1999). Somatosensory evoked potentials have generally been of normal size, but occasionally have been reported to be enlarged and associated with enhanced C-reflexes and negative myoclonus (Matsunaga et al. 2000).

CT scan may reveal brain atrophy, which can occur quite early in the course of the disease (Hayashi et al. 1992), although it is not usually seen until the later stages. MRI is more sensitive, but, none-the-less, standard studies are often unrewarding early and do not assist diagnosis. In patients with significant basal ganglia pathology, however, hyperintensity may be apparent on T2-weighted images and this can involve the caudate, putamen, and globus pallidus (Fig. 33.9). Such symmetric bilateral, hyper-intense signal changes on T2-weighted and proton density-weighted MRIs have been observed in 67–100% of patients with sporadic CJD. On the other hand, in vCJD hyper-intensities in the posterior (pulvinar) and dorsomedial thalamic nuclei, and to a lesser extent in the caudate head on fluid-attenuated inversion recovery (FLAIR) MRI sequences are striking features. These changes are also referred to as ‘pulvinar’ and ‘hockey-stick’ signs and are incorporated in the revised diagnostic criteria for possible vCJD (Criteria II D, World Health Organization, 2001) (Onofrj et al. 1993, Nagaoka et al. 1999, Kropp 2000, Maltete et al. 2006). The signal has also been reported on T1-weighted images (De Priester et al. 1999), but this is generally not the case (Yoon et al. 1995). It has been postulated that the imaging changes may represent spongiform change and gliosis and they can also be seen at times in other sites, including cerebral cortex (Yoon et al. 1995, Nagaoka et al. 1999) (Fig. 33.10) The change in the pulvinar has been correlated with gliosis on histological examination (Ziedler et al. 2000).

EMG studies may reveal signs of denervation with normal motor and sensory conduction in patients with amyotrophy (Worral et al. 2000). Although nerve conduction studies have generally been normal, occasional patients have been reported with marked slowing and a demyelinating peripheral neuropathy (Neufeld et al. 1992, Antoine et al. 1996). As mentioned above, in vCJD PrP^Sc accumulates in the tonsils and this may be able to be demonstrated on biopsy (Hill et al. 1997).

In cases with a genetic basis, DNA analysis may be used to show the underlying mutation responsible for the disease.

Unfortunately there is no effective treatment for CJD, GSS, or FFI. Genetic counselling can now be approached in familial cases. Research has addressed the following possibilities (Vana et al. 2007):

To achieve this, adeno-associated virus systems for antiprion components, including antibodies and siRNAs, and vaccination have been explored (Ludewigs et al. 2007).

Myoclonus may occur in a variety of acute encephalitic illnesses (Table 33.10), but it is a fairly uncommon phenomenon and there are relatively few reports in the literature. Rarely, myoclonus is a major component of an encephalitic illness.

Swanson et al. (1962) described three such cases. One was a 19-year-old woman who developed, subacutely, severe headache, agitation, hallucinations, and fever, had three seizures, and then lapsed into stupor with a stiff neck a few days after an upper respiratory infection. She exhibited repetitive myoclonus of the eyelids, both arms, and the left leg, subsequently spreading to involve the tongue, jaw, trunk, and right leg. The myoclonic jerks were said to be forceful, generally synchronous, at 60–100 per minute. They were not helped by paraldehyde. The CSF contained 85 lymphocytes and a protein of 160 mg/L. The EEG showed diffuse slow wave activity but no seizure discharges. Serum titres were negative for psittacosis, mumps, influenza, rickettsia, typhus, St Louis encephalitis, Eastern and Western equine encephalitis, Rocky Mountain spotted fever, and lymphocytic choriomeningitis. The patient died 8 days after admission to hospital. At autopsy the brain contained some polymorphonuclear cells around superficial cortical vessels and slight perivascular cuffing around a few brainstem vessels.

This case illustrates the general picture of those few patients with a parainfectious encephalopathy who develop marked myoclonus. Often viral studies are negative, the cause of the encephalitis is not established, and it is unknown whether it is due to direct viral invasion of the brain or a post-infectious demyelinating encephalopathy. However, MRI scan may show white matter lesions suggestive of active demyelinating encephalomyelitis (Takahashi et al. 1992).

Lance (1968) described a syndrome of action myoclonus appearing 2 weeks after an apparent encephalitic illness, which resolved

completely over a period of 7 weeks. Baringer et al. (1968) described eight patients with acute oscillopsia and action myoclonus of the trunk muscles as a sequel to respiratory or gastrointestinal infections, who subsequently recovered over 6–8 weeks. These may, however, be more appropriately considered as examples of the opsoclonus-myoclonus syndrome.

Bhatia et al. (1992) reported two patients with ‘isolated’ post-infectious myoclonus. The first was a 23-year-old woman who had an influenza-like illness followed by a generalized myoclonic syndrome and an EEG which showed polyspikes-spikes at 5–6 per second. The second case had segmental myoclonus involving the right upper limb following uncomplicated chickenpox. Back-averaging of EEG did not show any time-locked cerebral events. These patients had no other clinical features and the authors proposed that the term ‘isolated post-infectious myoclonus’ be used if there was 1) sudden onset of generalized multifocal or segmental myoclonus, 2) a history of recent preceding infectious illness, 3) no features of encephalitis or the opsoclonus-myoclonus syndrome, 4) a non-progressive course without seizures, ataxia, or dementia, and 5) recovery in a short but variable period of time. They identified nine other cases reported in the literature who appeared to be examples of this problem (Campbell and Garland 1956, Aigner and Mulder 1960, Bradshaw 1969, Silverskiold 1969).

There are, however, a few specific encephalitic illnesses in which myoclonus can be a prominent feature. For example, rhythmic myoclonus of the face, neck, limbs, or trunk was described in encephalitis lethargica (Walshe 1920, Reimold 1925, Krebs 1952). The myoclonus in encephalitis lethargica could be localized to a muscle or a group of muscles or could occur synchronously in many muscles in different parts of the body. It could be rhythmic, up to 50–70 per minute. Hall (1924) describes both focal and generalized myoclonus in the acute phase of the illness (the myoclonic type of encephalitis lethargica was particularly prominent in the epidemic of 1919–1920), which often recovered, and chronic persistent residual myoclonus. The latter often was focal (see Chapter 29) but could be generalized. Hall (1924) was of the opinion that the ‘electric chorea’ described by Dubini (1846) might well have been a minor epidemic of the myoclonic type of encephalitis lethargica.

Coxsackie (Hirayama et al. 1998) and enteroviruses are examples of viruses that can result in an encephalopathy with myoclonus. Entervirus 71, which causes hand-foot-and-mouth disease in young children, produces vesicles on the hands and feet, ulcers in the mouth, fever, and vomiting. When the initial illness resolves, aseptic meningitis or an encephalomyelitis may develop. Central nervous system involvement has been reported in 35% of cases (Wang et al. 1999), with the predominant neurological features being myoclonus (68%), vomiting (53%), and ataxia (35%). MRI scan may show rhombencephalitis with involvement of the midbrain, pons, and medulla. Cranial nerve involvement and long tract signs occur and while the disorder can be mild and reversible, persistent neurological abnormality, including myoclonus, may occur and a mortality of about 15% has been reported (Huang et al. 1999). The majority of such cases have been reported from China.

Myoclonus has also been reported in association with nipah virus encephalitis, a paramyxovirus which was contracted from pigs in Malaysia (Chua et al. 1999, Goh 2000). Involvement of the brainstem adjacent upper cervical spinal cord has been a prominent feature and mortality of over 30% has been reported. Another new agent that has been reported to cause myoclonus is tonate virus, which is associated with the Veuezuelan equine encephalitis complex (subtype IIIB). (Hommel et al. 2000, Talarmin et al. 2001).

Myoclonus occurs in a small number of patients with HIV infection and AIDS. While it has most often been focal (Bartolomei et al. 1999) or segmental (Nath et al. 1987) and possibly resulted from secondary infection (De Mattos et al. 1993), it has occasionally been generalized and thought to result from the human immunodeficiency virus itself. In this situation it has occurred in the setting of AIDS dementia complex (Maher et al. 1997). Sudden noise may trigger myoclonus and it can resemble a startle response, suggesting subcortical involvement. Other viral infections involving the brainstem, such as herpes simplex, have also been recorded as causing widespread myoclonus (Urushitani et al. 1993).

There are a number of other important viral or para-infectious myoclonic syndromes which deserve separate consideration. These include subacute sclerosing panencephalitis, Creutzfeldt–Jakob disease, myoclonic encephalopathy of infancy (the opsoclonus-myoclonus syndrome), and brainstem encephalitis in adults.

SSPE is a chronic persistent infection of the central nervous system caused by an altered form of the measles virus. Historically, in 1933, and again in 1934, Dawson described two children from Tennessee with a progressive encephalitis characterized by progressive mental deterioration and a peculiar type of myoclonus. At autopsy, intranuclear Cowdry type A inclusions were found in many cortical neurons. In 1945, Van Bogaert reported three cases of what he called subacute sclerosing leukoencephalitis, because there was marked inflammation of the white matter in the cerebral hemispheres.

Oshiro et al. (1975) and Weil et al. (1975) also have reported chronic rubella encephalitis resembling SSPE, with ataxia and myoclonus. Chronic rubella encephalitis, though usually occurring after congenital rubella infection, may follow aquired infection in the adult (Townsend et al. 1975).

SSPE is a progressive encephalopathy in children and adolescents, with marked mental deterioration, myoclonic jerks, spasticity, and blindness due to persistent measles infection of the brain.

Macroscopically, there may be cerebral atrophy. Microscopically, the major features are widespread gliosis (and demyelination) in white matter, with neuronal degeneration in the cerebral cortex, thalamus, and brainstem, and to a lesser extent in basal ganglia and cerebellum. There is perivascular cuffing of blood vessels with lymphocytes and plasma cells. Cowdry type A intranuclear inclusions are found in neurons and oligodendroglia (Fig. 33.11). Electron microscopy of such inclusions shows that they contain viral particles and tubules of the measles–distemper group of myxoviruses. Chen et al. (1969) and Horta-Barbosa et al. (1969) isolated measles virus from the brains of patients, and the illness has been transmitted by brain tissue from patients to animals (Lehrich et al. 1970, Thein et al. 1972). The disease is caused by an aberrant measles infection of the brain in which the virus is defective in its production of the matrix (M) protein (Hall and Choppin 1981). The exact pathogenesis is not understood, but the defective virus probably is not fully recognized and inactivated by antibody. In some cases defective immunity may play a part and it has been described in association with HIV infection in children (Koppel et al. 1996).

Subacute sclerosis panencephalitis is now relatively rare in developed countries and the incidence has decreased dramatically since the introduction of the measles vaccine, with only four to five new cases being registered annually in the United States (Dyken et al. 1989, Asher 1991). In Canada the incidence of SSPE was estimated to be 0.06 per million children per year based on data of the Canadian Paediatric Surveillance Program for the years 1997–2000 (Campbell et al. 2005). It is much more common in other areas of the world such as Africa, the Middle East, India, and parts of South America (Moodley and Moosa 1990, Yalaz et al. 1992, Yaqub 1996, Nunes et al. 1999).

The illness usually begins between the ages of 5 and 15 years (Modlin et al. 1979), commonly at about 7–8 years. SSPE very occasionally can present in early adult life (Cape et al. 1973, Prashanth et al. 2006). In a review of reported cases Singer et al. (1997) found the mean age of onset in adult cases to be 25.4 years (range 20–35 years). Onset as late as the sixth decade has been reported (Tanaka et al. 1987). Males tend to be affected 2–3 times more common than females (Modlin et al. 1979). Typically, affected individuals have been exposed to measles virus before the age of 2–3 years (Dick 1973, Dyken 1985, Okuno et al. 1988). It has been postulated that immaturity of the immune system may contribute to failure to eliminate the virus (Johnson 1982, David et al. 1990). Patients with adult onset of subacute sclerosing panencephalitis seem to have had their primary measles infection at either an unusually early or unusually late age (Singer et al. 1997).

From their review of over 200 cases, Farrell and Swanson (1975) identify three stages to the illness:

Intellectual function continues to deteriorate and the child becomes dependent, wheelchair-bound, incontinent, and dysarthric. Sometimes chorea, dystonia, and tremors occur. With time there is increasing spasticity and loss of facial expression and spontaneous movement. The myoclonus begins to subside and visual impairment intrudes.

Some authors, however, have noted a different pattern of onset. In a Brazilian series, myoclonus or tonic-clonic seizures were said to have been the initial symptom in 46%, while only 14% started with a behavioural disturbance (Nunes et al. 1999). Occasional adult-onset disease has also been reported as starting with myoclonus (Vela et al. 1997). Singer et al. (1997) in a review of adult-onset cases found that 73% complained of disturbance of vision early in the course of the disease and in 45% presentation was exclusively ophthalmological. Involvement of retina and/or choroid, optic nerve, pre- and retrochiasmatic pathways have all been reported. This prominence of visual system involvement has not been recognized in childhood- and adolescent-onset cases, although there are isolated reports of similar abnormalities (Green and Wirtschafter 1973, Gravina et al. 1978, Higashi et al. 1992, Park et al. 1997, Nguyen et al. 1999). Adult-onset cases not starting with visual disturbance have generally commenced with behavioural/mental change (Singer et al. 1997). While myoclonus in adults is common, it was absent in three of the 13 cases reviewed by these authors and the motor problems consisted of an akinetic rigid syndrome and pyramidal tract dysfunction.

The majority of affected individuals die within a year of onset, but death may be as early as 6 weeks or as late as 10 years (Farrell and Swanson 1975). Fulminant disease, with lapsing into coma within weeks and death within 4 months, has been said to occur more commonly in females and initial visual symptoms, even in children, have been a prominent presenting complaint (Takayama 1994, Gokcil 1998).

The course is usually progressive but stuttering, with 5% dying under 3 months and 20% living for 4 or more years. The mean survival has been put at 1½ years (Modlin et al. 1979, Dyken 1989, Asher 1991). Spontaneous remission uncommonly occurs and has been reported in between 5 and 8% of cases (Landau and Luse 1958, Cobb and Morgan-Hughes 1968, Resnick et al. 1968, Risk et al. 1978, Asher 1991). Remission may be more likely to occur in adult patients (Singer et al. 1997).

Antibody titres to measles in blood and CSF are very high. The CSF contains greatly increased IgG, with a marked oligoclonal band directed to measles antigens. Cell content is normal and total protein may be normal or slightly raised. More detailed analysis showed that some inflammatory markers, for example matrix metalloproteinase-9 (MMP-9), may be increased, whereas levels of other inflammatory markers, such as metalloproteinase-1 (TIMP-1), are normal (Ichiyama et al. 2008).

The EEG shows typical changes (Cobb and Hill 1950). Initially there is slowing of background rhythms, then periodic complexes appear, synchronous with the myoclonic jerks. They consist of large, widely distributed multiphasic waves every 5–15 seconds or so. They have been reported in about 75% of childhood-onset (Nunes et al. 1999) and adult-onset (Singer et al. 1997) cases. There is no invariant relationship between the EEG complexes and the myoclonic EMG bursts, which usually are long (up to 200 ms). The EMG burst may precede the EEG complex (Fig. 33.13); indeed the associated eye movement may occur up to a second before it (Lombroso 1968). During sleep the EMG silences and the periodic movements cease, but the EEG complexes go on unaltered (Petre-Quadens et al. 1968). The origin of the periodic EEG complexes is not clear. Some consider them to arise in the cerebral cortex (Bogacz et al. 1959), while others have suggested that they are driven by some thalamic or midbrain generator (Cobb 1966, Yagi et al. 1992).

Brain imaging is not very sensitive in subacute sclerosing panencephalitis. CT scan has been reported to show atrophy in 17% and be normal in 70% of cases (Anlar et al. 1988), but the exact proportion showing abnormality depends on the stage of the disease. Even MRI does not show a good correlation with the clinical status and only 40% of patients with advanced disease have been noted to have marked abnormalities, while 30% have been reported as having normal, or almost normal, findings (Brismar et al. 1996). Conversely, patients with clinically mild disease may have quite marked MRI changes. The earliest pathological change seems to be focal white matter high signal on T2-weighted images, which gradually increases and is perhaps most marked in the frontal and posterior parts of the cerebral hemispheres. Changes also occur in basal ganglia, particularly the putamina. Subsequent atrophy of white matter, including the corpus callosum and cortical grey matter, occurs (Kunika et al. 1995, Brismar et al. 1996). Lesions in the visual system, including retina and geniculate bodies, may be seen (Higashi et al. 1992, Takayama et al. 1994). Most marked changes in the frontal lobes were also found using serial diffusion-weighted MRI in one case (Kanemura and Aihara 2008) and there was correlation with the clinical course in that regional apparent diffusion coefficient of the frontal lobe decreased significantly with clinical progression and increased to normal range with clinical improvement. Magnetic resonance spectroscopy (MRS) showed reduction in N-acetyl aspartate (NAA) and increase in myoinositol (Aydin et al. 2006).

A patchy reduction in blood flow in the cerebral cortex, basal ganglia, and cerebellum has been noted using SPECT, with a similar decrease in [18^F] fluorodeoxyglucose on PET (Kunika et al. 1995, Singer et al. 1997). Accumulation of 6-¹⁸F fluoro-l-dopa on PET scan may also be decreased in the striatum (Singer et al. 1997).

There is also some correlation between imaging and EEG. Praveen-kumar et al. (2007) studied 58 patients with SSPE, 37 of which had typical EEG and 21 atypical EEG. CT and MRI scans were normal in 16 patients while others had white matter (15), cerebral oedema (8), cerebral atrophy (8), basal ganglia (2), and thalamic (2) changes. The authors found an independent association of frontally dominant slowing of background activity and typical periodic complexes with diffuse cerebral oedema on imaging, while white matter changes correlated with slowing of background activity only.

Sadly, there is no cure for this devastating illness. Prevention by vaccination against measles is the best strategy. Unfortunately, however, this does not guarantee immunity, and with the falling incidence the proportion of vaccine-related cases has increased. In the United States it has been put at 37% (Dyken et al. 1989).

The usual treatment is isoprinosine with or without alpha-interferon, the former being given orally and the latter administered intraventricularly or intrathecally (Dyken et al. 1982, Yalaz et al. 1992, Nunes et al. 1995, Anlar et al. 1997, Fayad et al. 1997, Singer et al. 1997, Gokcil et al. 1998). No control studies have been undertaken and the results of this therapy remain uncertain. Although improvement has been reported in about half of patients after alpha-interferon, subsequent deterioration is usual and it is doubtful whether isoprinosine significantly influences the prognosis (Anlar et al. 1997). Beta-interferon has been used instead of alpha-interferon but is of uncertain benefit (Anlar et al. 1998).

If myoclonus or seizures present a management problem, they may be able to be improved by valproate or benzodiazepines, but response is often only partial (Namer et al. 1986, Malherbe et al. 1991, Sharp et al. 1991). Myoclonus has also been reported to respond to trihexyphenidyl (Nunes et al. 1995).

Opsoclonus, first described by Orzechowsky (1927) in encephalitis, refers to chaotic, violent, usually conjugate bursts of random oscillations of the eyes in various directions (Smith and Walsh 1960). Others defined it as a dyskinesia consisting of involuntary, arrhythmic, chaotic, multidirectional saccades without intersaccadic intervals (Wong et al. 2001).

In 1962, Kinsbourne described six infants (aged 6–18 months) who developed, subacutely over a week, irregular multifocal myoclonus (dancing feet) and opsoclonus (dancing eyes). There was a dramatic response to adrenocorticotrophic hormone (ACTH) treatment.

In 1968, Solomon and Chutorian described the association of this syndrome with neuroblastoma.

Kinsbourne's first case (1962) had a negative frontal cortical biopsy, and his case 5 died at age 33 months; autopsy showed no more than perivascular lymphocytic cuffing in the brainstem. Other reports of the brain at autopsy have been largely unremarkable (Bray et al. 1969, Lemerle et al. 1969). There are occasional reports of abnormality in the cerebellum with Purkinje and granular cell loss plus gliosis affecting both the vermis and hemispheres (Ziter et al. 1979, Tuchman et al. 1989, Clerico et al. 1993).

The presence of opsoclonus suggests brainstem dysfunction, and the association with neuroblastoma has led to speculation on an immunological basis for the syndrome. Based upon the response to steroids, Kinsbourne (1962) suggests that the illness might have an autoimmune basis. Stephenson et al. (1976) suggested that the neuroblastoma and the cerebellum were ‘joint targets of an immunological attack’. Changes in CSF and response to medications known to cause immunosuppression have been regarded as supporting this concept, but direct evidence has been scanty [(see Pranzatelli (1996) for a review]. Anti-Ri antibodies, which have been associated with adult opsoclonus-myoclonus, have not been reported and anti-Hu antibodies, which have also been documented in some adult cases, have only occasionally been found in children (Fisher et al. 1992). Anti-neurofilament antibodies have been reported in sera (Noetzel et al. 1987). IgM and IgG antibodies from children with both neuroblastoma-associated and post-viral opsoclonus-myoclonus have been demonstrated to bind to cytoplasm of cerebellar Purkinje cells and some axons in white matter. This binding includes a subunit of neurofilament. Control sera have not shown similar reactivity (Connolly et al. 1997).

In about half of the cases the disorder is thought to be post-infectious and follow a viral illness (Kinsbourne 1986, Pranzatelli 1996). There may be a prodromal febrile illness, with upper respiratory tract symptoms. The onset typically is acute, developing over 24–48 hours, but may take up to a week to develop. Subsequently, the course varies from complete recovery within 3 months to persistence of symptoms for 1–3 years.

Based upon more than 100 cases described since their original report, Lott and Kinsbourne (1986) described the following picture: initial symptoms often consist of irritability, vomiting, and behavioural change (Table 33.11). Vomiting and head tilt may raise the suspicion of a posterior fossa tumour. Gait ataxia is obvious and, if the child is old enough, there may be an ataxic tremor of the arms (as when threading beads on a stick) and legs. Titubation of the head may be evident. This syndrome thus may be diagnosed as acute cerebellar ataxia of childhood (Weiss and Carter 1959, Weiss and Guberman 1978). The myoclonus usually becomes obvious with widespread, irregular, asymmetrical, multifocal, spontaneous, shock-like jerks. It may be increased by movement or startle; it disappears in deep sleep. Rarely, the movements appear choreic (Lott and Kinsbourne 1986).

The opsoclonic bursts of conjugate eye movements occur irregularly and on any direction of gaze. It is present during fixation, smooth pursuit, and convergence, and persists during sleep or eyelid closure. Each burst consists of up to 10–15 movements or rotations per second and may thus frequently cause visual blur and oscillopsia (an illusion of movement of the seen world) (Wong 2007). The direction of eye dancing may be horizontal, vertical, or rotatory, or any combination. Opsoclonus differs from nystagmus in that the phase that takes the eye off the target is always a saccade, not a slow eye movement (Wong 2007). Cogan (1954) distinguished opsoclonus from ocular dysmetria (overshoot with corrections as the eyes gain the point of fixation) and ocular flutter (spontaneous periodic horizontal bursts of eye movement consisting of back-to-back saccades that are confined to the horizontal plane, whereas opsoclonus is multidirectional). Dyken and Kolar (1968) described both opsoclonus and ocular flutter in infantile myoclonic encephalopathy. Zee and Robinson (1979) suggested that opsoclonus is the result of spontaneous activity of saccadic burst cells released from inhibition by pontine pause cells.

Gradual recovery, assisted by treatment, occurs in the majority of cases over 3 months to a year. Relapses may occur, often during intercurrent febrile illnesses or after steroid withdrawal (Lott and Kinsbourne 1986, Rodrigues-Barrionuevo et al. 1998). Recovery may be prolonged for several years and often is incomplete. More than 50% of children remain permanently handicapped, with varying degrees of mental retardation, educational difficulty or attention deficit, dysarthria, incoordination, and ataxia (Hammer et al. 1995, Papero et al. 1995, Rodrigues-Barrionuevo et al. 1998). Sixteen of 26 children with myoclonic encephalopathy were found to have significant neurological sequellae despite marked initial improvement (Lott and Kinsbourne 1986).

Confusion with acute cerebellar ataxia of childhood and posterior fossa tumour was mentioned above. Opsoclonus may be mistaken for the ocular bobbing seen in palatal myoclonus (see Chapter 29). The chaotic irregularity of the bursts of opsoclonus distinguishes this disorder from the jerk rotatory nystagmus of Pelizaeus-Merzbacher disease and from the characteristic eye and head movements of spasmus nutans. Other differential diagnoses of opsoclonus or ocular flutter to consider are listed in Table 33.12 (Wong 2007).

The EEG may show single or multiple spikes, but there is no correlation with the myoclonic EMG bursts (Pampiglione and Maria 1972). Routine CSF examination is normal or may show an increase in lymphocytes and elevation of IgG and IgM (Dyken and Kolar 1968, Bellur 1975, Rivner 1982). Oligoclonal bands have also been reported (Kostulas et al. 1987, Kalmanchey and Veeres 1988). Homovanillic acid and 5-hydroxy indoleacetic acid concentrations in CSF have been reported to be about 30–40% lower than in control subjects (Pranzatelli et al. 1995). CT or MRI brain scan is usually normal, even during the acute phase (Pranzatelli 1996).

A variety of viruses have been reported in association with the disorder, including Epstein–Barr, coxsackie B3, mumps, and St Louis encephalitis-virus (Estrin 1977, Kuban et al. 1983, Delreux et al. 1989, Sheth et al. 1995).

Neuroblastoma is found in about half of the reported cases (Table 33.11). Altman and Baehner (1976) described 28 such patients. The tumour is usually found within 3 months of the onset of the neurological syndrome. The majority lie in the thorax, usually in the mediastinum; the rest are in the abdomen. Urinary catecholamine levels may be raised but often are normal.

Survival in those with neuroblastoma and the myoclonic encephalopathy syndrome is more favourable (90%), compared with a 2-year survival of 30–40% in those with neuroblastoma alone (Koop and Johnson 1971, Altmann and Baehner 1976). Why this is so is not entirely clear. Neuroblastoma, however, has the highest incidence of spontaneous regression of any childhood solid tumour and the paraneoplastic opsoclonus-myoclonus syndrome may mean that an immune response has been mounted against the cancer. Perhaps the opsoclonus-myoclonus leads to discovery of the tumour at an earlier stage than might have otherwise been the case, as they are usually small, non-metatstatic, and asymptomatic, apart from neurological disorder (Pranzatelli 1996). In contrast, paraneoplastic opsoclonus in adults is associated with (small cell) lung, breast, and ovarian cancers (Wong et al. 2007, Musunuru and Kesari 2008).

The clinical features of the opsoclonus-myoclonus syndrome do not differ substantially in those with a neuroblastoma and those without (Table 33.12). The neurological illness responds as well to steroids in those with a tumour (around 60%) as in those without (Pinsard et al. 1981, Pranzatelli 1996).

Resection of the neuroblastoma has a variable effect on the neurological syndrome. Some improve dramatically (Moe and Nellhaus 1970, Senelick et al. 1973) but may subsequently relapse. Others deteriorate (Pranzatelli 1996) and the encephalopathy may even begin 15 or 19 months after a neuroblastoma has been removed (Delalieux et al. 1975, Rupprecht and Mortier 1975). In a follow-up study of ten children with neuroblastoma and opsoclonus myoclonus diagnosed at age 10–24 months and an interval between the first signs and starting treatment of 2–12 weeks, it was found that using different treatments consisting of different immunosuppressants, remission was achieved within 5 months in seven and relapses were present in seven. At a mean follow-up of 6–7 years, only one child had mild ataxia (Klein et al. 2007).

Even if the disorder is not overtly parainfectious, long-term follow up of infants with myoclonic encephalopathy does, however, often not lead to discovery of a neuroblastoma (Marshall et al. 1978).

Bray et al. (1969) put forward three possible reasons for the association:

At present, the third possibility seems most plausible. The anti-Hu antigen (ANNA-1) is found in approximately 75% of childhood neuroblastomas, although anti-Hu antibodies are present in blood in only about 4% (Dalmau et al. 1995). This antibody has, however, been associated with both opsoclonus and myoclonus plus neuroblastoma in a patient with Turner's syndrome (Fisher et al. 1992), which is a genetic condition that seems particularly associated with this tumour (Blatt et al. 1997).

Other antibodies associated with paraneoplastic opsoclonus in childen or adults (also see later and Table 33.13) supporting the notion of a humoral immune mechanism include anti-Ri (ANNA-2) (Anderson et al. 1988), anti-Yo (PCA-1) (Peterson et al. 1992), anti-Ma1 (Rosenfeld et al. 2001), anti-Ma2 (Wong et al. 2001), antiamphiphysin (Saiz et al. 1999), anti-CRMP-5/anti-CV2 (Yu et al. 2001), anti-Zic2 (Bataller et al. 2003), antineurofilaments (Noetzel et al. 1987), and new antineuroleukin antibodies in two girls with poststreptococcal opsoclonus-myoclonus syndrome [see Wong (2007) for review].

As with the parainfectious condition, the neurological disorder in those with neuroblastoma is likely to relapse even after apparently successful treatment. Recovery is gradual and the majority of this occurs over several months. Persistent neurological deficit is also common and has been reported in about 70% of cases (Koh et al. 1994, Hammer et al. 1995, Russo et al. 1997).

The child should be investigated for evidence of infection and neuroblastoma. Treatment should be undertaken for these disorders as indicated. Overall, neuroblastoma removal seems to have little influence on the neurological syndrome (Koh et al. 1994, Hammer et al. 1995, Klein 2007). The response to other forms of therapy directed towards the opsoclonus-myoclonus seems about the same, regardless of the underlying aetiology. In spite of the demonstration of low 5-hydroxyindoleacetic acid levels in CSF, patients do not improve with oral 5-hydroxy-L-tryptophan (Pranzatelli et al. 1994). If the movement disorder is causing disability, which is usually the case, administration of steroids is the treatment of first choice. There is an impression that ACTH is superior to other forms of steroids (Pranzatelli 1996), but there is no controlled trial and prednisone or prednisolone have also commonly been used. Initial response to ACTH has said to be in the order of 80–90% of cases, but relapse on withdrawal, or even while therapy is continuing, is not uncommon, which often leads to prolonged treatment (Hammer et al. 1995). In fact, a 40-week programme of ACTH has been suggested (see Wong et al. 2007). Steroid treatment has been reported to further decrease CSF homovanillic acid levels, without an effect on 5-hydroxyindoleacetic acid, in spite of symptomatic improvement (Pranzatelli et al. 1998).

Intravenous immunoglobulin may also result in benefit (Pranzatelli 1996, Russo et al. 1997) and sometimes may need to be used in association with steroids (Veneselli et al. 1998). Plasmapheresis, which has been useful in adult cases of opsoclonus-myoclonus, is technically difficult in small children. The effect of other immunosuppressive treatments has not been systematically investigated, although it has been claimed that patients with neuroblastoma who have received chemotherapy may be more likely to make a complete recovery than others (Russo et al. 1997).

For symptomatic therapy of nystagmus and oscillopsia, propranolol, nitrazepam, baclofen, clonazepam, and thiamine have been mentioned in the literature [see Wong et al. (2007) for revew]. Myoclonus can be treated with antiepileptic drugs.

A syndrome reminiscent of the opsoclonus-myoclonus encephalopathy of infants is occasionally encountered in adults. While the disorder is most commonly paraneoplastic (also see earlier), some cases are idiopathic and a variety of other less common causes have been reported, including intracranial tumours, hydrocephalus, thalamic haemorrhage, multiple sclerosis, metabolic disturbance, and toxic encephalopathies (Dropcho and Payne 1986, Caviness et al. 1995, Pranzatelli 1996).

Cogan (1954) described such a 61-year-old man who died. At autopsy there was encephalitis with perivascular lymphocytic infiltration in the hypothalamus, midbrain, and pons. Ross and Zeman (1967) described another such patient with an occult lung carcinoma, who at autopsy had mild diffuse Purkinje cell loss and some involvement of the olive. Before death, this 63-year-old man exhibited opsoclonus-myoclonus of the face, shoulders, and feet. This syndrome of paraneoplastic brainstem encephalitis with opsoclonus and myoclonus has subsequently been described in association with a number of different cancers. Tumours of the lung, especially small cell carcinoma, and breast have been particularly involved, but others include ovary, uterus, fallopian tube, bladder, thyroid, stomach, colon, larynx, Hodgkins disease, and non-Hodgkins lymphoma (Ellenberger et al. 1968 and 1970, Digre 1986, Kay et al. 1993, Caviness et al. 1995, Khoris and Pages 1999). As mentioned above, the associated antibodies include Anti-Hu (Hersh et al. 1994, Lucchinetti et al. 1998), anti-Ri (ANNA-2) (Anderson et al. 1988, Dropcho et al. 1993), anti-Yo (PCA-1) (Peterson et al. 1992), anti-Ma1 (Rosenfeld et al. 2001), anti-Ma2 (Wong et al. 2001), antiamphiphysin (Saiz et al. 1999), anti-CRMP-5/anti-CV2 (Yu et al. 2001), anti-Nova-1 (Buckanovich and Darnell 1997), anti-Zic2 (Bataller et al. 2003), and antineurofilaments (Noetzel et al. 1987). Sometimes there have been features of a brainstem encephalopathy with opsoclonus but without myoclonus (Caviness et al. 1995). While the vast majority of these tumours have been peripheral to the nervous system, intracranial lymphoma may rarely cause this syndrome (Tsuzaka et al. 1993).