50 Tonic attacks

Episodic movement disorders are defined in the introductory Section 11. Tonic attacks are episodic movement disorders characterized by continuous spasm of a muscle or group of muscles resulting in rigidity of the affected part. Movement, if it occurs, is limited to the start of spasm and once the attack is established the affected part remains stationary. To some extent such movements overlap with dystonic posturing, but a sinuous athetotic component, which is the distal counterpart of dystonia, is not present. The distinction between tonic and dystonic attacks is perhaps somewhat artificial. Chapters 49 and 50 should be considered together. Tonic attacks causing episodic movement disturbances are listed in Table 1 of the Introduction to Section 11.

Generalized tonic seizures may be epileptic or due to transient brainstem compression. As there is loss of consciousness these do not present as episodic movement disturbances and are not considered further. Generalized clonic seizures can sometimes be associated with preservation of awareness (Manford et al. 1996) but are best considered as myoclonic and are not discussed here. Focal epileptic attacks, however, may produce localized or generalized abnormalities of movement, without loss of consciousness. These can be classified according to the brain region involved in the attack. Thus, frontal, temporal, parietal, and occipital lobe seizures have all been recorded as having tonic motor components.

Several patterns of frontal lobe seizures have been defined (Salanova et al. 1995[a], Manford et al. 1996). These include the following:

Dystonic or tonic movements due to lesions restricted to the temporal lobe are probably uncommon and may be associated with widespread radiation of epileptiform activity. This may particularly apply if posturing is the first localizing manifestation of the seizure (Manford et al. 1996). Such attacks are included in the discussion on complex partial seizures in Chapter 49 (also see ‘Tonic limb and axial seizures’ below).

When epilepsy is due to a parietal lobe lesion somatosensory symptoms are common. Focal clonic activity has been reported in 82% of cases, tonic posturing of the extremities in 21%, and version in 15% (Salanova et al. 1995[b]). Complex partial seizures also occur but are less frequent (Williamson et al. 1992[a], Cascino et al. 1993). The relevant aspects of these attacks are included in this chapter.

Seizure disorders involving the occipital lobe tend to produce visual symptoms. Complex partial seizures occur in almost half of the cases. Asymmetrical and focal tonic and clonic motor activity may be present in a small number with or without version (Williamson et al. 1992[b], Iijima et al. 1994, Yokoyama et al. 1994). These are included in this chapter.

Although further discussions could be based on the brain areas in which seizures are thought to originate, this is often uncertain, and in the case of ‘paroxysmal hypnic dyskinesia’ the evidence for epilepsy may be circumstantial. Thus, we have chosen to explore the topic under the clinical categories of ‘tonic limb and axial seizures’, ‘versive seizures’, and ‘reflex epilepsy’.

In the early 20th century the concept developed that motor seizures arising from cerebral cortex produced clonic jerking and those from deeper structures resulted in tonic spasm. Tonic attacks of varying aetiology were used to support this notion. Sterling (1924) reported painful tonic spasms during the acute phase of encephalitis lethargica. These involved the face, neck, or limbs, and particularly the hands or feet. He postulated they were due to ‘extrapyramidal epilepsy’. In 1925 Wimmer recorded a patient who presented with unilateral tonic spasms lasting seconds, unaccompanied by loss of consciousness. Subsequently athetotic movements developed and eventually the appearance was that of generalized torsion spasm. Wimmer (1925) regarded the tonic attacks as ‘striatal epilepsy’.

Spiller (1927) used the term ‘subcortical epilepsy’ to describe two cases with tonic spasms, which were unilateral in one and affected both arms and a leg in the other. They were precipitated by movement and accompanied by severe pain. Although the diagnoses were not clear, the presence of tonic spasm was taken as sufficient to suggest epilepsy arising from deep cerebral structures.

In a critical appraisal of the topic, Wilson (1930) cautioned against assuming that tonic attacks were necessarily epileptic and that they arose from subcortical structures. He described a child with paroxysms precipitated by startle, excitement, a knock, or stimulating the skin of the left foot, during which the left lower limb extended and inverted, the ipsilateral arm abducted and flexed at the elbow, the trunk arched with concavity to that side, the head inclined to the left, and there was slight left upper facial spasm. Consciousness was preserved but speech was suspended. After some years left leg weakness appeared. Contrary to the popular view, Wilson (1930) considered these attacks were due to epilepsy arising from cerebral cortex. His description closely fits that of attacks arising in the region of the supplementary motor cortex and the adjacent superior margin of the cerebral hemisphere, as defined by later workers.

Seizures arising in the region of the pre and post central gyri produce clonic jerking of the contralateral body part, corresponding to the motor homunculus. Although seizures restricted to the post central gyrus are usually sensory, motor attacks can occur and electrical stimulation of this cortex produces contralateral movement at approximately 25% of sites (Penfield and Rasmussen 1950). Seizures arising in the sensorimotor cortex near the Sylvian fissure involve the contralateral face, frequently commencing at the angle of the mouth. Involvement of the adjacent superior cortex produces movement of the contralateral hand, particularly the thumb. Discharge of the sensorimotor cortex near the vertex, or on the medial aspect of the hemisphere, affects the contralateral leg. Localized seizure discharge produces focal epilepsy, but spread to the adjacent cortex results in a classical Jacksonian march. Thus, an attack may involve the whole of a limb, face, and arm, both limbs on the same side, or the whole of one half of the body. Hemiclonic activity is often associated with ipsilateral head turning (Manford et al. 1996). Spread to the opposite hemisphere may follow, but the seizure then commences synchronously over the whole of the contralateral side, rather than gradually spreading (Wilson 1930). In a group of patients with such Jacksonian motor seizures studied by Manford et al. (1996) only lesions in the perirolandic cortex were associated with this type of attack.

Although movements due to epilepsy arising in pre and post central gyri are usually clonic, it has generally been considered that they may be preceded by tonic stiffening, which usually lasts seconds (Rasmussen 1974). Occasionally in aborted seizures such tonic spasm may constitute the entire attack. Aphasia or dysphasia occur only if the dominant hemisphere is affected. There is evidence, however, to suggest that tonic limb posturing, often associated with version, may be the first manifestation of seizures originating in the lateral premotor cortex. Progression to other motor manifestations is common, particularly simple automatisms. By contrast, the appearance of tonic limb posturing occurring later in the attack has little localizing value and occurs as often in temporal lobe as in frontal lesions (Manford et al. 1996). Rarely idiopathic generalized epilepsy may present solely with hemiconvulsive attacks, which are mainly tonic (Kiley et al. 2000). Three adolescent patients with tonic attacks of one side of the body were found to have three cycles per second generalized spike and wave discharges on ictal and/or interictal EEG and all responded to sodium valproate treatment (Kiley et al. 2000).

Seizures arising in the region of the supplementary motor cortex, which is on the medial surface of the cerebral hemisphere rostral to the sensorimotor cortex, and in the adjacent superior border of the cerebral hemisphere often have tonic manifestations. Unilateral discharge may produce complex bilateral movements, resembling the assumption of a posture. Commonly the contralateral arm is externally rotated and abducted at the shoulder and flexed at the elbow, with rotation of the face towards that limb. The opposite arm may pronate and extend by the side, while the legs are rigidity postured in extension or semiflexion (Fig. 50.1). Apart from the positive motor phenomena mentioned above, negative phenomena, i.e. brief lapses of postural tone (negative myoclonus), can also be provoked by discharges of the supplementary sensorimotor area (Melliti et al. 2000). Slowing or arrest of speech may occur in dominant or non-dominant hemisphere involvement. Consciousness is usually preserved, but if it is lost there may be sustained loud vocalization at the onset (Erickson and Woolsey 1951, Penfield and Welch 1951, Kennedy 1959, Rasmussen 1974, King and Smith 1996). The features in children are similar to those in adults (Bass et al. 1995). Occasionally there is laughing or crying (Salanova et al. 1995[a]). Falling sometimes occurs, but incontinence and tongue biting are not features. In a few cases there may be clonic movements in the contralateral or ipsilateral limbs (Kennedy 1959). There may also be accompanying visceral, genital, or somatic sensations, which can spread and usually ascend from below (Kennedy 1959). The features in children are similar to those in adults (Bass et al. 1995). Occasionally attacks may be precipitated by startle (Gastaut 1954), anxiety (Kennedy 1959), and touching (Kennedy 1959) or moving (Falconer et al. 1963) affected limbs.

Partial attacks may occur with recurrent episodes of stiffness in one limb (Kennedy 1959). Unilateral tonic arm movements are more common than bilateral ones and whole body movements are even less frequent (Salanova et al. 1995[a]). While the classical picture is readily diagnosed, such fragments are easily misinterpreted. In addition, spread of epileptiform activity to adjacent brain areas may complicate the situation. Manford et al. (1996) found that in practice it was difficult to differentiate supplementary motor area from lateral premotor seizures and this has been noted by others (Quesney et al. 1990). Using ictal SPECT with 99mTc HMPO Ebner et al. (1996) found focal hyperperfusion in the expected brain region in only 40% of cases of focal tonic seizures felt to be typical of supplementary motor area seizures. Laich et al. (1997) also found hyperperfusion of the supplementary motor using SPECT in their patients concordant with either lateralizing clinical signs, scalp EEG activity, or ictal onset of seizures by intracranial electrodes.

Rarely seizures arising in supplementary motor or sensorimotor corticies are preceded by inability to move the affected contralateral part. This may be accompanied by tingling in the same somatic distribution. In isolation, these somatic inhibitory seizures may also be difficult to diagnose (Rasmussen 1974).

Seizures originating in the supplementary motor area often commence during sleep, although the patient rapidly awakens and may be conscious during the attack. There is stereotyped posturing, which may variably involve extension of the legs but almost always includes abduction of the arms. Opisthotonus is unusual. A prolonged monotonous ictal cry and thrashing movements are common. The entire bout lasts less than 40 seconds and is usually shorter than in pseudoseizures, in which thrashing movements are more prominent (Kanner et al. 1990). Multiple recurrences during one night are frequent (King and Smith 1996). There is marked similarity between these attacks and those of paroxysmal hypnic dystonia, leading some authors to postulate they involve activation of the same mechanism (see Chapter 49).

In a versive seizure the head and eyes tonicly turn to one side. Such attacks arising in the anterior third of the frontal lobe or in associated subcortical regions have been said to be associated with loss of consciousness at commencement of the attack (Walker 1966). If rotation is unaccompanied by or precedes loss of consciousness it has been suggested the epileptic focus is more posteriorly situated, commonly in the intermediate frontal area (Rasmussen 1974). Quesney et al. (1990) found that retention of consciousness during head turning probably occurs in about one third of cases in which version arises due to frontal lobe seizures and they postulated such attacks resulted from stimulation of the supplementary motor area or the intermediate frontal cortex. They felt that unconscious head rotation was more likely to be due to activation of anterolaterodorsal parts of the frontal lobe. Such attacks have usually been regarded as adversive or contraversive, in which rotation is away from the side of epileptic discharge. Solanova et al. (1995) reported that in 40% of patients with frontal lobe seizures there was head turning, and in all the movement was contraversive. Using ictal SPECT scanning in children with frontal seizures Harvey et al. (1993) found that if head turning occurred it was always contraversive to the site of the presumed seizure. Some authors, however, have found occasional patients who have ipsiversive seizures due to pure frontal lobe lesions, but this seems to represent under 10% of cases (Quesney et al. 1990).

Abversive or ipsiversive seizures, in which rotation is towards the abnormal cortex, were at one stage considered to arise from temporal or parietal corticies (Rasmussen 1974) and to be infrequent. The majority of versive epileptic attacks that result from parietal lobe lesions, however, are away from the side of the focus (Cascino et al. 1993, Sveinbjornsdottir and Duncan 1993, Salanova et al. 1995[b]). In a few patients seizures are ipsiversive (Williamson et al. 1992[a]). Head turning is also recorded in occipital lobe epilepsy, but deviation of the eyes alone may be more frequent (Williamson et al. 1992[b]). Such ocular movement is usually tonic (Munari et al. 1984). Williamson et al. (1992[b]) found over half of a group of patients with an occipital lobe focus had deviation of the head or eyes or both contralateral to the affected side, and in 12% it was ipsilateral.

Thus, most studies suggest that contraversion is substantially more frequent than ipsiversion, even when seizures arise at widely different sites throughout the cerebral hemisphere. It has not been shown, however, that such movement does not depend ultimately on activation of the same common motor pathway, even though the origins of the epileptic attacks can vary widely. Not all authors agree with such conclusions and the direction of movement has sometimes been found to have little relationship to the laterality of seizure onset (Robillard et al. 1983, Ochs et al. 1984). Differences in the definition of what constitutes a versive seizure may underlie this conflict. Some workers have distinguished between forced deviation of the head or eyes, in which the movement is said to be sustained and unnatural, and non-forced movements, the latter being said to resemble volitional turning. Kernan et al. (1993) reported that 89% of forced movements were contralateral, whereas non-forced movements were not of lateralizing significance. Versive seizures are commonly followed by a generalized epileptic attack and as such they are not likely to be confused with paroxysmal movement disorders. Diagnostic difficulty may arise if they occur in isolation.

Stimuli, such as bright or flashing lights, patterns, reading, sounds, music, or startle, may induce reflex seizures, but such paroxysms are usually generalized and associated with loss of consciousness. These attacks need not be considered in the diagnosis of episodic movement disturbances. Such stimuli, however, occasionally produce tonic or clonic attacks without loss of consciousness. These are usually localized or unilateral. More commonly, such attacks are precipitated by local somatic stimulation, such as touch, tap, or movement, and the response is frequently restricted to or centres on the part stimulated. These tonic and clonic seizures unaccompanied by loss of consciousness are considered below. Reflexly induced myoclonic attacks are discussed in Section 7 on ‘Myoclonus’.

Gowers (1881) discussed a variety of reflexly induced epilepsies and provided the first description of seizures precipitated by movement, although these have sometimes mistakenly been regarded as examples of paroxysmal kinesigenic dyskinesia (see Chapter 49). In many reported cases the site of epileptic discharge is uncertain. There is an impression that in cases in which localization is established, involvement of the supplementary motor cortex and the superior border of the hemisphere is unduly common.

Excitement or anxiety may occasionally cause attacks. In the child with probable supplementary motor area seizures reported by Wilson (1930), excitement was a precipitant. Kennedy (1959) described tonic postural movements arising from this site, which were sometimes triggered by anxiety. Startle was also provocative in Wilson's (1930) patient and in two of Gastaut's (1954), who had tonic lower limb supplementary motor area seizures. Sudden noise precipitated tonic spasm of an arm, plus rotation of the face and eyes to that side, in a boy described by Allen (1945).

Startle-induced tonic seizures are sometimes encountered in patients with structural brain disorders, including perinatal hypoxia, Down's syndrome, Sturge-Weber syndrome, and porencephaly (Saenz-Lope et al. 1984). There is an abrupt startle response followed by a tonic spasm lasting 10–15 seconds. Unilateral attacks tend to be associated with ipsilateral hemiparesis, a contralateral cerebral hemisphere lesion, focal EEG abnormalities, and better preservation of intelligence than in patients with generalized seizures. The latter usually have severe intellectual impairment, widespread cerebral damage, and generalized EEG change (Saenz-Lope et al. 1984), and are unlikely to present as a movement disorder as awareness is impaired.

Epilepsy precipitated by smell or taste is extremely rare and adequate documentation is lacking (Merlis et al. 1974). Scollo-Lavizzari and Hess (1967) reported a boy with parasthesiae in the tongue, grimacing, jaw opening and closing, plus tonic flexion spasm of the upper limbs precipitated by eating or drinking. Strauss (1940) described a patient with right sided hemiparesis in whom clonic ipsilateral jerking was brought on by a variety of right sided stimuli. Light and sound seemed more effective applied to the ipsilateral eye and ear. Radiology showed an irregular calcification just behind the motor area, with ventricular dilatation and cortical atrophy on the left.

Touch or percussion may occasionally trigger focal motor attacks. In Strauss's (1940) patient, right sided clonic seizures could also be precipitated by widespread ipsilateral stroking or pricking of skin, stimulation of mucous membrane, or percussion. Allen (1945) described several patients in whom focal clonic movements or tonic spasm were produced by cutaneous stimulation. Stroking one foot produced tonic attacks in the patients of Wilson (1930) and Kennedy (1959). Forster et al. (1949) reported a child with a capillary hemangioma involving pre and post central gyri in whom clonic attacks in the contralateral face and arm were precipitated by tapping that shoulder. Painful spasm of the arm, followed by jerking and stiffness of the ipsilateral face, could be elicited by touching, stretching, or sleeping on the upper limb in a patient with contralateral mild cerebral atrophy (Scollo-Lavizzari and Hess 1967). EEG suggested an epileptic focus in the post central region of the atrophic hemisphere. Tactile stimulation of the hands and feet produced runs of spike and wave at the vertex in a patient described by Neville and Boyd (1995), but these were not associated with clinical features.

Movement is an occasional precipitant. The cases described by Gowers (1881) had attacks induced by movement or change in position. Falconer et al. (1963) described tonic stiffening of ipsilateral limbs with abduction of that arm and arching of the trunk, precipitated by movement of that leg. Attacks were relieved by removal of a small scar from the superior border of the contralateral hemisphere, 2 cm rostral to the central sulcus. Allen (1945) reported a patient with ipsilateral tonic seizures caused by catching one foot on the ground when walking. In the patient of Neville and Boyd (1995), mentioned above, putting the foot to the floor when attempting to walk, especially if it had a slapping quality, resulted in flexion of the left leg with inversion of the foot and extension of the great toe. This was felt to be epileptic but responded to prednisolone and not carbamazepine.

A peculiar variety of movement-induced epilepsy has been described in association with non-ketotic hyperglycaemia (Gabor 1974, Aquino and Gabor 1980, Brick et al. 1989, Hennis et al. 1992). Reported cases have been mainly middle aged or elderly. Attacks have been the initial manifestation of diabetes and have usually involved an upper limb, the face, or neck. The leg has occasionally been affected. Sometimes Jacksonian progression has occurred and the whole of one side may be involved. In one patient attacks were also contraversive (Gabor 1974). Seizures arise spontaneously or are precipitated by repetitive active movement of the affected part. This has included elevation of the eyelids and movement of the hand or arm. Attacks may be tonic, clonic, or a mixture of these. They usually last less than a minute but are sometimes longer and can be followed by a refractory period of about a quarter to half an hour. Postictal paresis may be present (Gabor 1974).

It has been postulated that hyperglycaemia leads to a reduction in the inhibitory neurotransmitter GABA with a resultant tendency to seizures (Hennis et al. 1992). Ketosis, however, causes intracellular acidosis, a secondary rise in glutamic acid decarboxylase, and an increase in brain GABA (Roberts et al. 1958).

EEG has revealed repetitive spikes and sharp waves over the contralateral hemisphere, and in a patient whom movements were mainly tonic the focus centred around the rolandic fissure rather than the supplementary motor area (Gabor 1974). In this patient paralysis of the affected limb by nerve block did not prevent the appearance of an epileptic paroxysm in the EEG, with associated contraversion when the subject attempted active movements of the hand, suggesting the seizure was not triggered by proprioceptive feedback (Gabor 1974).

Severe hyperosmolarity has not generally been a feature and plasma glucose has usually been less than 35 mmol/L, although higher values are recorded (Hennis et al. 1992). Seizures have responded to anticonvulsants, but insulin administration and correction of electrolyte abnormality has been said to be more effective (Hennis et al. 1992).

Transient hemiballismus due to non-ketotic hyperglycaemia is mentioned in Chapters 24 and 48.

Episodic movement disturbances in multiple sclerosis may be choreic and athetotic (see Chapter 49), ataxic (Chapter 51), or tonic. About 10% of patients with multiple sclerosis have tonic spasms (Solaro and Tanganelli 2004). Tonic attacks may presumably be cerebral or spinal in origin, but unless the face is involved it may not be possible to distinguish between these. In the subsequent sections the use of the terms ‘cerebral’ and ‘spinal’ should be viewed as presumptive. While it has not been shown that the pathophysiological mechanisms underlying these two movements are different, there are clinical features that tend to separate them.

Painful tonic spasms in demyelinating disease were first described by Guillain et al. (1928) in necrotic neuromyelitis optica. Unilateral ‘tetanoid’ spasms were accompanied by flexion of the opposite leg and precipitated by movement or hyperventilation. Although several lesions were present at autopsy, the attacks were attributed to involvement of the contralateral subthalamic nucleus and cerebral peduncle. The following year Redlich (1929) reported hemispasms in two patients with multiple sclerosis. Storring (1940) included two similar cases in a series of patients with epilepsy secondary to multiple sclerosis. It was not, however, until Matthews (1958) reported four cases that the entity became widely recognized. Subsequently there have been a number of other reports (Joynt and Green 1962, Kreindler et al. 1962, Kuroiwa and Araki 1963, Lance 1963, Castaigne et al. 1970, Osterman and Westerberg 1975). There seems to be overlap between these spasms and the paroxysmal dystonia which has been reported in multiple sclerosis and it is doubtful if they are really different.

Attacks are characterized by brief, often painful tonic spasms which affect the limbs, trunk, or face on one side. Multiple areas of demyelination make it difficult to identify causative sites and definite clinico-pathological correlation has not been possible. In a reported autopsy case both midbrain and spinal cord lesions were present (Kreindler et al. 1962). Similar lesions were noted in another case (Shibaski and Kuroiwa 1974), but there are reasons to suspect that the spinal disease may have been causative (see later). Subcortical localization was favoured by Matthews (1958) and, on the basis of associated discomfort, Lance (1963) suggested involvement of the thalamus or specific thalamo-cortical pathways. Osterman and Westerberg (1975) argued that many such tonic attacks probably arise in the spinal cord. CT and MRI scanning, however, suggest demyelination of the contralateral cerebral peduncle (Fig. 50.2) (Verheul et al. 1990, Rose et al. 1993), internal capsule (Watson and Chiu 1979, Honig et al. 1988, Maimone et al. 1991, Aladro et al. 1993), thalamus (Burguera et al. 1991), globus pallidus and putamen (Roos et al. 1991), and even the medulla (Gatto et al. 1996) may be responsible in some patients. In other words, lesions at multiple sites in the brain have been postulated as being responsible for these spasms [see Tranchant et al. (1995) for a review].

Pathophysiological mechanisms are also obscure. A peculiar form of epilepsy (Matthews 1958), partially demyelinated hyperirritable axons, and ephaptic transmission are possibilities. Matthews (1958) suggested central lesions might increase excitability of lower motor neurons to metabolic fluctuations and thus precipitate a form of tetany. This has not been supported by subsequent workers.

The onset of these tonic spasms appears unrelated to the duration or severity of multiple sclerosis and they may be the first sign of the disease or commence in patients with chronic disability (Matthews 1958, Joynt and Green 1962). Although spasms may arise spontaneously they are frequently precipitated by movement, including putting the foot to the ground or turning in bed. Sometimes attacks only occur when turning onto one side and can awaken patients from sleep (Matthews 1958, Joynt and Green 1962). Passive movements (Matthews 1958, Kreindler et al. 1962) and startle (Roos et al. 1991) may also trigger episodes. Hyperventilation frequently provokes spasms (Matthews 1958, Joynt and Green 1962, Espir et al. 1970, Sethi et al. 1992). The paroxysms are rapid in onset and last between 30 seconds and 2 minutes. They may be very frequent, occurring 50 or more times daily. Attacks often commence with an unusual, unpleasant, tight, or painful sensation at one site, which rapidly spreads ipsilaterally. One of Matthews (1958) patients

had a tickling sensation in the palm, causing her to laugh. Tonic spasm typically results in abduction of the shoulder, flexion of the elbow, and extension of the leg, with or without inversion of the foot. The pattern, however, is variable and the upper limb may extend while the leg flexes (Matthews 1958, Joynt and Green 1962, Lance 1963). Occasionally involuntary finger movements have been reported (Joynt and Green 1962) but frequently they are tonicly flexed. Although voluntary movement may be suspended, some patients are able to move the affected limbs, particularly at proximal joints (Matthews 1959). The face may be affected ipsilaterally and while most can talk, occasional patients are unable to speak (Joynt and Green 1962). Formal neurological examination during an attack is seldom possible, but transient hyper-reflexia and an extensor plantar response has been noted on the affected side (Joynt and Green 1962). Confusion, tongue biting, incontinence, and postictal weakness do not occur (Joynt and Green 1962). Matthews (1958) suggested an inconstant relationship with spino-thalamic sensory impairment in the same site as the spasms. This, however, has not been a feature of most other cases.

A bout of attacks usually lasts between several weeks and 2 months before spontaneously disappearing and occasionally it may be recurrent (Joynt and Green 1962). The duration of bouts is similar to that of many other symptoms caused by episodes of demyelination and it has been postulated they are due to formation of new plaques (Matthews 1958, Joynt and Green 1962).

EEG, even during attacks, has shown no specific abnormality, although intermittent, scattered, slow activity has occasionally been reported (Matthews 1958, Joynt and Green 1962, Lance 1963). Similarly, angiography, pneumoencephalography, and myelography have been unremarkable, apart from showing possible mild ventricular dilatation, consistent with multiple sclerosis (Joynt and Green 1962). CT and MRI have suggested that this paroxysmal movement may be caused by demyelination at a number of different sites, as mentioned above, but multiplicity of lesions in individual cases makes it difficult to be certain which, if any, is responsible for the attack.

Anticonvulsants have been useful in suppressing tonic attacks. Phenytoin (Matthews 1958, Joynt and Green 1962), carbamazepine (Castaigne et al. 1970), phenobarbitone (Matthews 1958), and tiagabine, a selective inhibitor of the gamma-aminobutyric acid (GABA) transporter, GAT-1 (Solaro and Tanganelli 2004), have all been effective in normal anticonvulsant doses. Gabapentin and baclofen may be useful (Solaro et al. 2000). Attacks usually cease promptly and therapy can be cautiously withdrawn after 6–10 weeks. Acetazolamide has also suppressed the movements (Sethi et al. 1992, Nardocci et al. 1995), but relapse while still on this therapy may occur (Sethi et al. 1992). Botulinum toxin may also improve tonic spasms of multiple sclerosis (Restivo et al. 2003).

Reflex-induced, flexion and extension limb spasms are common in spinal disease of varying aetiology. In addition, however, a different type of painful, tonic spasm due to spinal disease has been reported in multiple sclerosis (Kuroiwa and Araki 1963, Kuroiwa and Shibasaki 1967, 1968, Shibasaki and Kuroiwa 1974, Osterman and Westerberg 1975). These attacks seem more common in Japan and have been reported in 17% of Japanese multiple sclerosis patients. The question as to whether painful tonic spasms are generally more frequent in multiple sclerosis sufferers of oriental or South-East Asian descent is raised by the observation that they occur in 44% of Thais with this disorder (Jitpimolmard and Vejjajiva 1994).

Spasms may arise spontaneously but are frequently precipitated by passive or active movements, postural change, or tactile stimulation. Occasionally noise or hyperventilation may be precipitants. Cutaneous trigger zones are often at, or below, the clinically estimated segment of spinal involvement and frequently accompanied by an area of parasthesiae or cutaneous sensory impairment. Attacks commonly start in a limb and rapidly spread proximally or distally, before radiating to the trunk or other limbs. Involvement may be unilateral or bilateral, but the head is not affected. The pattern of attacks relatively stereotyped for each patient. Dysaesthesiae or parasthesiae rapidly sweep over this area and painful tonic spasm follows immediately or is delayed by up to a quarter of a minute. Contraction usually, although not necessarily, affects the areas of parasthesiae and dysasthesiae. A variety of postures may be adopted and limbs may be flexed or extended. Tendon reflexes have been reported increased in the affected limbs during an attack (Shibasaki and Kuroiwa 1974).

Osterman and Westerberg (1975) termed these paroxysms ‘spinal sensorimotor seizures’ and described patients in whom unilateral tonic spasm has been preceded or followed by contralateral sensory disturbance. Such cases have been regarded as showing a ‘positive’ Brown-Sequard-type of symptomatology. They stressed that a sensation of warmth or heat precedes the pain and suggested this is ‘explained by the spatial arrangement of pain and temperature fibres in the lateral funiculus of the spinal cord’ (Fig. 50.3). These authors regarded ephaptic transmission in spinal cord lesions as being the most likely mechanism for spinal ‘seizures’.

Autonomic features may occur during a spasm and include flushing, sweating, pilo erection, and hyperactive bowel sounds. There is no impairment of consciousness, even in bilateral attacks. These attacks do not correlate closely with the presence of spasticity or weakness, but all patients have evidence of spinal involvement, particularly cervical or thoracic. Sixty-four percent of Shibasaki and Kuroiwa's (1974) patients had Lhermitte's symptom, compared with 13% of multiple sclerosis patients without such attacks.

Arguments supporting separating this entity from spinal flexor and extensor spasms include preceding, radiating dysaesthesiae, stereotyped spread of the tonic spasm, and lack of correlation between the presence of attacks and severity of pyramidal signs.

Spasms are rapidly abolished in most cases by carbamazepine 600 mg daily or phenytoin 300–400 mg daily (Shibasaki and Kuroiwa 1974). Carbamazepine has been suggested to be the more effective (Kuroiwa and Shibasaki 1967, Kuroiwa and Shibasaki 1968). Sodium channel blockers lidocaine (plasma level 12.4 pg/ml) and its oral analogue mexiletine (300–400 mg/day) have been reported to be useful in treating all positive symptoms of multiple sclerosis particularly tonic spasms and Lhermitte's symptoms (Sakurai and Kanazawa 1999).

In addition to the well-known tonic extensor spasms caused by coning, which are generally bilateral and fairly symmetrical, unilateral involuntary movements can occur in fully conscious patients due to focal brainstem lesions. Al-Shahwan et al. (1994) reported two children with gangliogliomas in the cerebello-pontine angle who had ipsilateral spasms of the face and limbs. These involved flexion of the upper limb, flexion or extension of the leg, and rotation of the neck. They noted that four out of seven previously published cases of children with posterior fossa ganglioglioma had similar bouts of spasm (Langston and Tharp 1976, Flueler et al. 1990, Bills and Hanieh 1991). Retention of consciousness, lack of EEG abnormality, and failure to respond to anticonvulsants suggested the attacks were non-epileptic. It was hypothesized that the hemifacial spasms might arise due to compression of the root entry zone or genu of the facial nerve and that the limb movement could be caused by encroachment on the rubrospinal tract, which would cross the midline just proximal to the usual tumour site.

Other brainstem lesions can cause tonic or dystonic-like movement and these have been reported to be synchronous with the ocular tilt reaction in cavernous haemiangioma involving the meso-encephalic region (Paine et al. 1996).

In addition to the common flexion spasms caused by disinhibition of the spinal reflex arc as the result of spinal lesions, other types of tonic and dystonic-like postures can occur. They may include extension movements and may be unilateral or bilateral. While such activity can result from a variety of pathologies (Previdi and Buzzi 1992), multiple sclerosis is perhaps the most common.

Tonic spasm resulting in dystonic-like posturing of an extremity may be provoked by activities involving the affected part in patients with peripheral nerve or plexus lesions (Scherokman et al. 1986). Such spasms can be quite painless and last seconds to minutes (Fig. 50.4). They may occur before or after the onset of the other features of the nerve lesions and be relieved by carbamazepine or anticholinergics (Scherokman et al. 1986). Nerve conduction studies and electromyography may be required to confirm the peripheral nature of the problem.

The causes and neurological features of hypocalcaemia and hypomagnesaemia are discussed in Chapter 24. Episodic movement disorders include 1) chorea, athetosis, and dystonia, which are discussed in Chapter 48, and 2) tetany, clonic jerks, and seizures, which are discussed here.

Tetany is characterized by numbness and parasthesiae of the extremities and perioral area, distal cramps, carpopedal spasms, and laryngeal stridor. Some authors have included convulsions (Albright and Reifenstein 1948), but it seems best to regard these

as a separate entity. Tetany is precipitated by low plasma calcium or magnesium concentrations (see Chapter 24). As the literature related to tetany occurring as an episodic movement disorder is largely restricted to calcium abnormalities, further discussion is limited to this.

Tetany can be produced in normal subjects by alkalosis, which results in a fall in plasma ionized calcium concentration. This is commonly due to transiently lowered plasma carbon dioxide, secondary to hyperventilation occurring in circumstances of excitement and anxiety. It causes spontaneous repetitive discharges in the proximal part of peripheral nerves, particularly the longest nerves supplying the most distal structures. Over-breathing is followed by circumoral and distal numbness or tingling with painful carpopedal spasm. Laryngeal stridor occasionally occurs. Although the full-blown syndrome is characteristic, hyperventilation may go unrecognized and spasm may be unilateral or focal, making diagnosis difficult. Careful history taking and reproduction of typical attacks by voluntary hyperventilation are essential in confirming the cause. With the patient's awareness of the problem, voluntarily slowing the respiration or rebreathing into a paper bag usually controls the problem.

Disorders causing sustained hypocalcaemia may produce paroxysms of tetany. Attacks may be spontaneous or precipitated by exercise, sudden movement, change in position, or anxiety (Arden 1953, Robinson et al. 1954). They can be focal, segmental, or generalized and last seconds to hours (Simpson 1952, Robinson et al. 1954). Attacks range from brief cramping of a hand or foot to sustained, generalized, tonic spasm (Arden 1954, Robinson et al. 1954). The pattern of abnormality is often stereotyped for an individual.

Cephalic involvement includes spasms of ocular movement, with or without diplopia (Robinson et al. 1954, McKinney 1962, Muenter and Whisnat 1968). There may be facial tightness or spasm (Muenter and Whisnat 1968) and trismus (Robinson et al. 1954). Some patients are unable to talk (Simpson 1952, Wise and Hart 1952). Laryngeal stridor can occur in isolation or in association with other features. Occasionally it is severe (Albright et al. 1942, Sutphin 1943) and may cause respiratory embarrassment with unconsciousness (Drake et al. 1939). Torticollis, retrocollis, or spasms of the trunk are reported (Robinson et al. 1954, McKinny 1962). Abnormal posturing of limbs including carpopedal spasm with the ‘main d'accoucheur’ position of the hand and similar deformity of the foot are common. At times spasms are accompanied by slow involuntary movements. Generalized attacks or those involving the legs may result in falling (Drake et al. 1939, Robinson et al. 1954).

Occasionally patients show involuntary jumping, clonic, or twitching movements of the limbs (Robinson et al. 1954, Meunter and Whisnat 1968). Focal or generalized epileptic seizures may occur. At times tetany may precipitate or follow an epileptic convulsion (Simpson 1952, Robinson 1954, Willison and Whitty 1957). Structural intracranial abnormality may predispose to the focal peripheral effects of hypocalcaemia (Willison and Whitty 1957).

The electromyogram in tetany can show repetitive firing of single or grouped motor unit potentials in early spasm, but in the more advanced cases it resembles a full interference pattern (Kugelberg 1948).

Tetany is controlled by correcting hypocalcaemia or hypomagnesaemia. Epilepsy may require interim anticonvulsant therapy.

Tetanus is a common disease caused by a ubiquitous organism and may even have been known in pre-historic times. It was clearly recorded in the 5th century bc by the Greek physician Hippocrates. Perhaps the best description is from Cappadocious Aretaeus which concludes: ‘This is an awful disease, horrible to watch and incurable. Even the closest friends withdraw from the victim and the – otherwise all but pious – wish of his people appears to be justified; the patient should die soon, in order to finish pain and wretched evil together with life. The physician has no means to sustain life, to diminish pain, to correct distortion, because if he would stretch the extremities, he would have to dissect or break the victim – This is the physician's agony’ (Habermann 1978). It was not until the end of the 19th century, however, that it was realized tetanus was transferred by infective agent. Before the turn of the century, passive immunization had been used in man and early in the 20th century tetanus toxoid became available (Table 50.1).

The name tetanus comes from the Greek ‘tetanos’ which means to stretch. It clearly relates to the main clinical feature of the disease, which is due to the action of the soluble toxin tetanospasmin (or tetanus toxin; molecular weight 150 kDa) on the nervous system. This large molecule has several antigenic sites and is controlled from a single plasmid (Laird et al. 1980). This extremely potent neurotoxin is formed by Clostridium tetani. The Clostridia are exceptional amongst the bacteria for the frequency and diversity of toxin production, the most lethal being produced by Clostridum botulinum (for the medical uses of this toxin see Section 8 on dystonia). Interestingly both toxins have mechanisms of action at the cellular level. Tetanus toxin (TeNT) is a heterodimeric protein antigen, whose light chain (L) is translocated in the cytosol of neuronal target cells and (like certain botulinum toxin serotypes) specifically cleaves its substrates, vesicle-associated (VAMP-3) membrane protein-2 (VAMP-2, or synaptobrevin) or cellubrevin (Pellizzari et al. 1999). Clostridium tetani can form spores and thus survive for long periods, even in dry conditions (Schofield 1986). It is particularly common in the faeces of herbivores but grows freely in soil, especially if this is wet, warm, and well manured. The organism is, however, widely found in dust and even road dirt from large cities. Lowbury and Lilly (1958) found it was common in unfiltered air, including in an operating theatre, although this was not associated with surgical infections. It has also been found as a contaminant of heroin (Weinsterein 1973).

Clostridium tetani, a motile gram positive bacillus, enters the body through contaminated wounds, including lacerations, burns, open fractures, or injections. As it is an anaerobic organism it is more likely to grow in sites that contain necrotic material, particularly if there are foreign bodies. As the organism multiplies an increasing amount of tetanospasmin is formed. The main pathological effect of this is on the central nervous system and it may enter the brain and spinal cord in several ways. As early as 1903 it was found that the toxin could be transported to the central nervous system via motor axons of peripheral nerves (Meyer and Ransom 1903). It has been demonstrated, however, that entry is via nerve endings (Wernig et al. 1977) and axonal transport does not occur when the toxin is directly injected into nerves (Price et al. 1975). Such direct transport in peripheral nerves supplying the site of the infected wound is thought to be the basis of localized tetanus. This is, however, relatively uncommon and it has been postulated the low incidence may be due to associated damage to nerves in the area of the wound (Habermann 1978).

Another possibility is that tetanospasmin may enter the central nervous system from the blood, with resulting generalized tetanus. Penetration through the blood-brain barrier, however, is extremely poor and although about 1% of a dose of toxin eventually finds its way into the central nervous system following intravenous injection in an experimental animal (Habermann and Dimpfel 1973), this is thought to be by way of ascending to the central nervous system via many peripheral nerves and has thus been regarded as ‘multiple local tetanus’ (Habermann 1978). Ligation of the sciatic nerve in generalized tetanus leads to accumulation of toxin at the level of the ligature and diminution of the amount in the corresponding half segment of the spinal cord (Seib et al. 1973).

In peripheral nerves toxin appears to travel both within axons and within the interstitial space, but entry into the spinal cord through the ventral roots may be by way of the axons alone (Erdmann et al. 1975). There is also evidence of spread of toxin along sensory nerve fibres, but this does not seem to enter via the dorsal nerve roots (Kryzhanovsky 1973). There is also some spread along some autonomic nerves (Hensel et al. 1973) and involvement of the intermediolateral column of the spinal cord can be associated with sympathetic overactivity.

The toxin binds to most central nervous system sites when applied directly so that the specific patterns of toxin accumulation seen in animals with tetanus probably reflect the mode of entry rather than variation in binding affinities (Haberman 1978).

After entering the spinal cord via ventral root axons the toxin concentrates in the anterior horn cell motor neurons. It then migrates transynaptically and accumulates in spinal inhibitory interneurons, where it blocks the release of their inhibitory transmitters glycine and gamma-aminobutyric acid (Duchen 1973[a], Dasta et al. 1981). Direct inhibition of motor neurons mediated via inhibitory interneurons, driven by Ia muscle afferents and recurrent inhibition via Renshaw cells, is decreased, leading to hyperactivity of both alpha and gamma motor neurons (Weinstein 1973, Haberman 1978). Thus, increasing muscle rigidity and overactivity of spinal reflexes develop. The same type of process occurs in the brainstem and cranial muscles. The results are not dissimilar to the action of strychnine. There is experimental evidence to suggest that tetanospasmin can induce blockage at the neuromuscular junction by impairing acetylcholine release (Duchen 1973[a], [b], Kryzhanovsky 1973). There are also data suggesting that there may be a degree of axonal damage to both motor and sensory nerves in some cases (Shahani et al. 1979).

The electromyogram shows intermittent or continuous discharge of motor unit potentials. Single fibre electromyogram has been reported as showing increased jitter and blocking which improves on higher firing rates, suggestive of a presynaptic defect of neuromuscular transmission (Fernadez et al. 1983). F-waves are increased in size in keeping with hyperexcitability of the motor neurons (Risk et al. 1981). The silent period of the stretch reflex is lost, reflecting the absence of central reflex inhibition. Blink reflex abnormalities suggesting brainstem dysfunction and abnormalities of the cortical silent period have also been described in generalized tetanus (Warren et al. 1999, Poncelet 2000). Transcranial magnetic brain stimulation in a patient with generalized tetanus revealed enlarged electromyographic responses and absence or reduction of the late phase of the electromyographic silence following the motor evoked potential in sternomastoid and biceps brachii muscles (Warren et al. 1999). Thus, there is impairment of inhibitory mechanisms at multiple levels of the nervous system, including the cortex, in generalized tetanus. Following clinical recovery, the silent period returns to normal. Direct injection of tetanus toxin into cerebral cortex in experimental animals produces clinical and EEG epileptic seizures (Mellanby and George 1975).

Although tetanus is uncommon in developed countries it is a major cause of death in underdeveloped societies and it has been estimated that it causes approximately 1 million deaths annually (Furste 1982). The epidemiology of this disorder shows dramatic differences. In third world countries the disorder usually occurs in neonates, while in industrialized societies it presents mainly in the elderly and drug addicts (Humbert et al. 1972, Weinstein 1973, Blake and Feldman 1975) and new fads, for example tongue piercing (Dyce et al. 2000). It has been estimated that in parts of Africa up to 1% of neonates die of this disorder (Woodruff et al. 1984), while in parts of India the rate of neonatal infections has been put at 5.3 per 1000 live births (Smucker et al. 1984).

The decreased rate of death in developed countries is due to a lower rate of infection, better primary care of wounds and widespread immunization with tetanus toxoid. Nowhere has the decline in incidence of tetanus been more dramatic than on the battlefield. In the USA the incidence of tetanus decreased by a factor of 10 between 1947 and 1982–84, when it was 0.036 per 100,000; 95% of cases during these most recent years received primary immunization against the disease (Immunization Practices Advisory Committee 1985). Surveillance data for the United States for 1998–2000 revealed an average of 43 tetanus cases annually, an annual incidence rate of 0.16 cases per million population. Fifteen percent were among injection drug-users (Pascual et al. 2003) (Fig. 50.5A).

Tetanus may follow apparently trivial wounds and it is possible that by the time the illness develops the injury may have been forgotten. Nonetheless, in some series no portal of entry has been found in up to 20% of cases (Adams 1968). In neonatal tetanus the site of entry is usually the umbilicus. Traditional application of dirt or dung to the umbilical stump is causative in some societies (Rey and Diopmar 1967) as is ritual, ear piercing, and circumcision (Rey et al. 1967). Non-medical abortion is not only associated with a high risk of infection but also the prognosis is particularly poor.

The latent period between the time of entry of the organism and the first symptom shows considerable variation but is usually less than a fortnight and probably never longer than a month (Haberman et al. 1978). Although the mean incubation time in neonates is about a week it can occur as early as the second day of life (Martin-Bouyer 1967, Newel et al. 1967). In all forms of tetanus the shorter the latent period of the disease the more fulminant the course of the illness and the worse the prognosis (Table 50.2 and Fig. 50.5B).

The clinical presentation of tetanus can range from mild to severe (see Table 50.3) and had been divided into three types. The first is localized tetanus which constitutes less than 20% of all cases. It consists of rigidity and spasms in the muscles supplied by the nerves which terminate in the region of the infection. The area involved is thus restricted to the general site of the injury. The disorder may last for several weeks before subsiding. More frequently, however, it progresses to generalized tetanus.

The second clinical variety of this disease which is distinguished is cephalic tetanus. This occurs in about 1% of patients (Biglan et al. 1978) and is really localized tetanus affecting the head. It usually follows head injury or middle-ear infection (Alfery and Rauscher 1979) and presents with involvement of facial or cervical muscle (Weinstein 1973). One patient has been reported who developed it after ‘tongue piecing’, having had a ring inserted into her tongue (Duce et al. 2000). As with other varieties of the localized form of tetanus this may gradually subside or become generalized.

The third clinical picture is that of generalized tetanus, which constitutes in excess of 80% of all cases. Muscular rigidity commences in the head or neck in the vast majority of cases, possibly because the nerves supplying them are short and transportation time to the motor neurons is correspondingly reduced. Rigidity in the masseter muscles initially restricts jaw opening and this becomes progressively more marked. It leads on to involuntary spasms of jaw closure. Thus, speaking and chewing become difficult or impossible. This classical picture of trismus or lock-jaw is the initial complaint in 75% of patients (Garnier et al. 1975, Edmondson and Flowers 1979). Involvement of facial muscles is usually also early and consists of grimacing (risus sardonicus) or eye closure.

Pharyngeal rigidity may result in a sensation of tightness or a ‘sore throat’. There can be dysphagia with choking. The cervical muscles are usually affected along with the face, and neck stiffness may be interpreted as meningitis. The trunk and limbs become involved and palpation reveals undue firmness of muscles and the abdomen. Tendon reflexes are increased.

As the disease deteriorates this stage of generalized stiffness gives way to bouts of painful spasm. These are precipitated by a number of different stimuli, including light and sound, but particularly by touching or moving the patient. Flexors and extensors are usually affected equally and the resulting sustained positions are due to the relative strength of the different muscle groups. The legs tend to extend with flexion of the toes and feet, while the hands form fists, the elbows flex, and the shoulders adduct. Involvement of the trunk produces marked opisthotonus and neck extension (Fig. 50.6). These ‘spinal convulsions’ are frequently accompanied by pharyngeal or pharyngeal spasms. This, along with respiratory muscle contraction, may produce asphyxia. These painful spasms last 10–20 seconds and may be so severe that they result in compression fractures of thoracic vertebrae. In the initial stages patients are fully conscious during these bouts.

Autonomic involvement is manifest by profuse sweating, pallor, tachycardia, irregularities of cardiac rhythm, and hypertensive and hypotensive crises. These are not solely the result of muscular

spasms and can be seen even in paralyzed and ventilated patients (Corbett and Harris 1973). Fever may result from muscle over-activity or coincident infection.

The patient may succumb from choking caused by aspiration or spasm-induced asphyxia. There may be cardiovascular collapse due to autonomic involvement, fluid loss, acidosis, and electrolyte imbalance. Disseminated intravascular coagulation and infection may also be terminal events (Haberman et al. 1978). Although autonomic involvement can occur in up to 60% of serious cases (Trujillo et al. 1980, Wesley et al. 1983) and sudden cardiac arrest has been reported in 40% of these patients, cardiac resuscitation is relatively easy (Trujillo 1980).

The total picture in neonatal tetanus is not dissimilar. Inability to suck is usually an early feature and the upper lip may feel stiff. The jaw may clamp firmly on a finger put into the mouth. With deterioration there is involvement of the facial, truncal, and limb musculature. Generalized spasms and opisthotonus complete the picture.

Prevention is the cornerstone of control of this disease and immunization is central to this. The tendency to combine tetanus, diphtheria, and pertussis immunization has meant that there has sometimes been confusion as to the exact cause of adverse reactions. It is now known, however, that the majority result from the pertussis component. The overall rate of significant reaction to tetanus immunization is in the order of about 2 per 100,000 doses (Edmondson and Flowers 1975) and most of these have been acute anaphylaxis (Blumstein and Kreithen 1966). Local reactions, however, with swelling and redness may occur in 1–2% of patients and are most likely to occur in those who have been given boosters less than 10 years apart (Jacobs et al. 1982, Kiwit 1984). Very rarely brachial plexis neuropathy, cranial nerve palsies, polyradiculopathy, and polyradiculomyelopathy have occurred, but recovery is almost invariable (Quast et al. 1979, Fenichel 1983, Holliday and Bauer 1983, Kiwit 1984.)

Prevention of tetanus includes adequate surgical toilet of any potentially infected wound with appropriate antibiotic cover and a booster tetanus immunization, if one has not been given in the previous 2 years (Fulde 1988).

When infection with tetanus has occurred wound toilet and administration of penicillin is recommended, although extensive mutilating surgery is pointless (Haberman et al. 1978). Human hyperimmune antitetanus globulin is frequently given to bind any unabsorbed toxin and it has displaced the use of equine antitetanus serum, which was associated with a high incidence of allergic reactions. There are, however, relatively few controlled studies of the use of such hyperimmune globulin and a number of these have been negative (Haberman et al. 1978).

Benzodiazepines such as diazepam act at both brain and spinal level to inhibit the reflex activity and produce marked improvement. Intravenous doses of 5 mg/kg/day of diazepam have been used (Habermann 1978). Benzodiazepines may be just as effective. Although long-acting barbiturates, such as phenobarbitone, have been used, their effects on respiration and conscious level make them less appropriate. Ironically, Botulinum toxin injections into the facial muscles have been used successfully in a case with cephalic tetanus to relieve trismus (Andrade and Brucki 1994).

Severe cases require the use of muscle relaxants, artificial respiration, and full intensive care in addition to sedation. Sympathetic over-activity must be watched for and is usually treated with beta-blockers, and morphine may be helpful for this in certain situations (Buchanan et al. 1974)

Tetanus remains a serious disease with a case fatality rate in excess of 50% (Faust et al. 1976, Alfery and Rouscher 1979, Trujillo et al. 1980 and Schofield 1986). In a study of 150 patients from India, Udwadia et al. (1987) reported that with intense management in an intensive care unit, with proper nutrition, early tracheostomy, and ventilator support, overall mortality was 12% and the mortality in severe tetanus was 23%.

Successful treatment results in complete recovery, and neurological sequellae, unless they result from hypoxia, are rare. Full recovery usually takes many weeks or months and there may be difficulty in mobilization due to contractures or persisting neuropathy (Shahani et al. 1979).

Sandifer's syndrome is one cause of torticollis in infancy/childhood. Other causes may include contractures of the sternocleidomastoid, ocular torticollis caused by eye muscle weakness, neural axis abnormalities, atlantoaxial rotatory displacement resulting from trauma or oropharyngeal inflammation (Grisel's syndrome), infectious causes such as retropharyngeal abscesses and pyogenic cervical spondylitis, tumors of the posterior fossa or the upper cervical spine, and benign paroxysmal torticollis (Herman 2006). Torticollis resulting from cervical dystonia is rare in children.

Sandifer (Sutcliffe 1969) was the first to recognize abnormal posturing of the head, neck, and upper trunk in children due to hiatus hernia. Probably less than 1% of children with hiatus hernia show these movements (Sutcliffe 1969), although it may occur in up to 4% of brain-damaged infants with gastroesophageal reflux (Wesley et al. 1981). Such patients may have a history of vomiting in infancy or early childhood, but this is often inconspicuous or absent when movements commence. They usually start after the first year and may be delayed until the fourth or fifth year of life, although involvement as early as 2 months has been reported (Williams and Frias 1987). To some extent, abnormal movements or posturing may be present most of the time, but they are greatly exacerbated or may only appear after eating. The neck can be deviated in almost any direction, but extreme lateral tilting and torticollis are common (Fig. 50.7). The upper trunk is usually tilted in the same direction, so the patient may sit or stand with the vertex pointing towards the floor. This position may be maintained during a variety of activities, including walking, reading, washing, and lying in bed (Kinsbourne 1964, Sidaway 1965, Sutcliffe 1969). Frequently bizarre, dystonic-looking movements of the neck and upper trunk occur as well. Movements and abnormal postures are absent during sleep. Although some have claimed these postures are spontaneous, they do not seem to be due to involuntary tonic spasm like other conditions included in this section. They have been interpreted as a way of relieving abdominal or thoracic discomfort, but not all children admit to this symptom (Sutcliffe 1969). Surprisingly, these postures seem to increase or precipitate displacement of the stomach into the thorax, and oesophageal manometry shows they increase peristalsis (Puntis et al. 1989). In eight cases barium swallow, 24 h pH-metering, manometry, endoscopy, and biopsy were carried out (Gorrotxategi et al. 1995). The barium swallow was pathologic in 62% of cases and the pH-metering was abnormal in 37%. The lower oesophageal sphincter pressures were decreased in 37%, with oesophageal motility alteration in 75%. Signs of macro- and/or microscopic oesophagitis were found in 62%. The authors suggested that the most frequently found alteration is oesophageal dysmotility. This was supported by another study in which real-time ultrasonography showed delayed gastric emptying time, which returned to normal when the patient was asymptomatic (Cardi et al. 1996). The importance of gastric motility investigations in Sandifer syndrome was stressed by this report. Movements usually rapidly disappear following surgical repair of the hiatus hernia (Kinsbourne 1964, Sutcliffe 1969) or sometimes with medical drug treatment. Three of eight patients reported by Gorrotxategi and colleagues (1995) received surgical treatment and the rest medical treatment, with improvement of the neck contortion in all their cases. The disorder may be confused with benign paroxysmal infantile torticollis (Menkes and Ament 1988). Other differential diagnoses are listed above.

The neuroanatomical basis was speculated on by Cerimagic et al. (2008) who believe that neurological manifestations of the syndrome are the consequence of the vagal reflex with the reflex centre in the nucleus tractus solitarii. They provide three models for the hypothetic reflex. In the first, the hypothetic reflex arc is based on the classic hypothesis of two components: nervus accessorius (n. XI) – radix cranialis and radix spinalis. The nervous impulses are transmitted by nervus vagus (n. X) general visceral afferent fibres to the nucleus tractus solitarii situated in medulla oblongata, then by interneuronal connections on the nucleus ambiguus and nucleus dorsalis nervi vagi. Special visceral efferent fibre impulses from the nucleus ambiguus are in part transferred to n. XI ramus externus (carrying the majority of general somatic efferent fibres) via hypothetic anastomoses in the region of foramen jugulare. This leads to contraction of the trapezius and sternocleidomastoideus muscles and the occurrence of intermittent torticollis. In the second model, the hypothetic reflex arc is organized in the absence of the radix cranialis of n. XI, the efferent part of the reflex arc continues as the nucleus ambiguus, which is the motor nucleus of nervus glossopharyngeus (n. IX) and n. X in this case, while distal roots of n. XI that appear at the level of the olivary nucleus lower edge represent n. X roots. Finally, in the third presented model the hypothetic reflex arc includes no jugular transfer and could be realized via interneuronal connections directly from the nucleus tractus solitarii to the spinal motoneurons within nucleus radicis spinalis nervi accessorii or from the nucleus ambiguus to nucleus radicis spinalis nervi accessorii. The afferent segment of the postulated reflex arc in all three models is mediated via n. X. Which or if any of these models is the underlying mechanism remains uncertain.

Recurrent paroxysmal attacks of torticollis in infancy and early childhood have been reported (Snyder 1969, Gourley 1971, Sanner and Bergstrom 1979, Bratt and Menelaus 1992). Onset is usually in the first year of life, although it may occur up to the fifth year, and has been reported to be more common in females (Hanukogklu et al. 1984). However, in a recently reported series of 22 cases there were 10 female and 12 male cases, although putting together all the reported cases with theirs the authors did note a slight female preponderance (63%) (Drigo et al. 2000). In this series the patients were aged between 1 and 18 years of age (at follow up), although in 95% onset was before 9 months of age. Symptoms occurred suddenly on waking in about 50% of cases and one or more trigger factors (teething, cold, concurrent illness) were identified in 36% of cases.

The head is tilted to one side and the face may be slightly rotated in the opposite direction (Fig. 50.8). In the series of Drigo et al. (2000) in 27% of cases the head turned always to the same side, while in the remainder it turned randomly to either side. The frequency of the episodes in this series initially ranged from once every 2 days to once every 45 days.

Sometimes onset is accompanied by pallor, vomiting, and agitation, but often the child appears unconcerned, although attempts to straighten the head produce distress. Drigo et al. (2000) reported that one of their patients also presented with hypotonia of the ipsilateral lower limb and two cases had had upward deviated gaze, apart from photophobia, sleepiness, asthenia, and headache in one. Paroxysms usually last 2 or 3 days but may be as short as 10 minutes or as long as 2 weeks. In the series of Drigo et al. (2000), the torticollis lasted a few hours to a few days in 59%, but in 41% it lasted more than a week.

The pathogenesis of this disease has not been determined. During episodes in one patient, continuous electrical discharges over the sternocleidomastoid muscle on the same side as the torticollis were observed on surface electromyogram, similar to that seen in dystonia (Kimura and Nezu 1998). Investigations including EEG and brain ultrasound were performed in a third of patients in the series of Drigo et al. (2000) and were normal in all.

Spontaneous recovery is usual by 2–3 years of age (Drigo et al. 2000). Sanner and Bergstrom (1979) reported that bouts may be modified in the later stages of the disorder, with ataxia being the main feature.

Snyder (1969) suggested these attacks may be due to benign paroxysmal vertigo as some children complain of dizziness or rotation. Caloric testing is sometimes abnormal and audiometry later in life has shown diminished hearing in a proportion of cases. In line with this Drigo et al. (2000) reported that after disappearance of episodes of torticollis, six of their patients later presented with periodic symptoms compatible with benign paroxysmal vertigo, recurrent abdominal pain, and cyclic vomiting and/or migraine. A family history of migraine is present in 40% of cases (Drigo et al. 2000). It has also been suggested that the disorder is related to benign idiopathic dystonia of infancy or may represent a self-limiting form of paroxysmal dystonic dyskinesia (Fahn 1994) (see Chapter 36 and Table 1 in the Introduction to Section 11). The existence of familial cases (Lipson and Robertson 1978, Sanner and Bergstrom 1979) and the association of benign positional torticollis with other paroxysmal disorders, such as migraine, suggests that the condition could also be a ion channel disorder or channelopathy (see Chapter 49). In fact, Giffin et al. (2003) presented four patients with benign paroxysmal torticollis of infancy, two of whom came from a kindred with familial hemiplegic migraine linked to CACNA1A mutation.

It is also possible that at least some of these children may really have Sandifer's syndrome as there has not been adequate investigation to exclude this possibility in all cases.