23 Spontaneous oro-facial chorea and tardive dyskinesia

Spontaneous oro-facial chorea is a disorder manifested by repetitive unintentional choreic movements of the face which are usually most prominent around the mouth. It is to be distinguished from cephalic dystonia (Meige's or Breughel's syndrome) in which the movements are of longer duration, have a greater tendency to involve the upper face, and usually are much more distressing to the patient. This latter condition is discussed in Chapter 37. Some of the movements of spontaneous oro-facial chorea may be repetitive, patterned, and hence appear as stereotypies (see later under ‘Tardive dyskinesia’). Some have expressed the view that the majority of spontaneous late onset edentulous dyskinesias are actually a form of idiopathic torsion dystonia (Jankovic 1988, Jankovic and Fahn 1993), but the disability and appearance in this group of patients tends to be different from that seen in Meige's syndrome, so we have chosen to retain the term oro-facial chorea.

Tardive dyskinesia is an involuntary movement disorder which is due to prolonged administration of neuroleptic and other drugs. The commonest manifestation of this disorder is oro-facial chorea, similar, if not identical, to the spontaneously occurring variety.

Dyskinesia involving the face and mouth can occur in most inherited and acquired causes of chorea. In addition, however, there is a spontaneously occurring variety of chorea which is unassociated with other disease and principally involves the cephalic musculature. This disorder has been called by a number of different names, including ‘bucco-lingual dyskinesia’, ‘lingual-facial-buccal dyskinesia’, the ‘bucco-lingual-masticatory syndrome’, and the ‘oral masticatory syndrome’.

Neuropathological and neurochemical studies of this condition are lacking. It is known that brain dopamine decreases with age and this is probably due in part to degeneration of the nigrostriatal system (Carlsson and Windblad 1976, McGeer et al. 1977). In addition, GABA and acetylcholine containing neurons are thought to degenerate with age (Barbeau 1973). These changes may be responsible for the parkinsonian features which many elderly patients develop. It has been suggested that the ventrolateral striatum, which is particularly concerned with oral movements, is especially likely to be affected by the ageing process (Mackay 1981). It has been postulated that this loss of dopamine and alteration in other neurotransmitters may produce supersensitive striatal dopamine receptors, stimulation of which by remaining endogenous dopamine produces the chorea. Although plausible, this pathophysiological mechanism remains speculative. Buccolingual dyskinesia has been reported following bilateral thalamic infarction (Combarros et al. 1990), giving some support to the notion of basal ganglia involvement in the spontaneously occurring condition, and although bruxism has been reported after cerebellar haemorrhage, this was complicated by lacunar disease involving the neostriatum (Pollack and Cwik 1989).

In some cases, edentulous patients with maloccluded dentures may make frequent movements of the lower jaw, tongue, and lips. It has been postulated that these may be due to loss of proprioceptive impulses from the peridontal membrane (Sutcher et al. 1971). Koller (1983) found that 16% of a group of edentulous patients (average age 62 years) had oral dyskinetic movements whereas there were no similar cases in controls with their own teeth. Tooth extraction had been an average of 12 years earlier and preceded the onset of movements in all cases. The movements were restricted to the mouth and seemed less marked than in most cases of spontaneous or drug-induced oro-facial dyskinesia. ‘Edentulous oral dyskinesia’ would appear to constitute a minority subgroup of patients with oro-facial chorea.

From a series of studies it seems that the prevalence of spontaneous oro-facial chorea is wide (Table 23.1), but in a review of 19 studies of 11,000 patients the reported figure was about 5% (Kane and Smith 1982).

The precise figure, however, depends on the mean age of the population surveyed. Spontaneous oro-facial chorea becomes increasingly common with ageing. Klawans and Barr (1982) studied the prevalence in subjects without evidence of neuroleptic or anticonvulsant intake, alcoholism, brain dysfunction, or known movement disorders. Those with full dentures were also excluded. They found a gradually increasing prevalence of spontaneous oro-facial chorea in middle-aged and elderly patients rising from 0.8% between 50 and 59 years to almost 8% between 70 and 79 years. Bougeois et al. (1980) found a prevalence of approximately 23% in a group of patients whose average age was 81 years. Both studies showed the incidence in women was approximately double that in men. Attention, however, has been drawn to the fact that many such reports contain subjects with a variety of medical or neuropsychiatric conditions and, if these are excluded, the prevalence of involuntary movements is very low. In a review of five studies from the literature, only 1.6% of predominantly healthy subjects had involuntary movements (Waddington 1989). When a similar series containing patients with medical but not neuropsychiatric disorders was analyzed the prevalence was 8.9%, and in patients with neuropsychiatric conditions never treated with neuroleptic drugs it was 24% (Waddington 1989). The last group included patients with psychoses, dementia, and mental handicap.

In most of these patients the chorea is largely confined to the mouth (Fig. 23.1) and consists of repetitive tongue movements including protrusion and retraction (‘fly catcher's tongue’) or lateral displacement inside the mouth (‘bon-bon sign’).

Repetitive jaw movements are common with opening and closing of the mouth, and protrusion, retraction, or lateral displacement of the mandible. The lips may purse, smack, or grimace and platysma is frequently tensed. The upper face is involved less frequently, with blepharospasm, blinking, frowning, and wrinkling of the forehead. On occasions movements may extend to the neck and even less commonly there may be a few choreiform movements in the limbs. Unlike oro-facial dystonia, most patients are not disturbed by their movements and many are completely unaware of them. Occasionally the muscular activity is more dystonic in nature (Fig. 23.2) and may be distressing.

In a few cases attention to dentures may be useful, but in the majority reassurance of patients and their relatives is all that is required. Although movements can be lessened by the use of dopamine store depleting and dopamine receptor blocking drugs, these should be avoided as they may eventually aggravate the movements (see later under ‘Tardive dyskinesia’) and can readily produce confusion or depression in the elderly. On occasions, however, movements may be severe enough to cause great distress and difficulty with speech and swallowing. In such cases tetrabenazine or one of the more selective post-synaptic dopamine receptor antagonists may give relief (see later under ‘Tardive dyskinesia’).

Tardive or tardy dyskinesias are literally those that are slow in developing. These were first described by Hall et al. in 1956, 4 years after the introduction of the first neuroleptic drug. Sigwald et al. (1959) drew attention to frequent cephalic involvement, which they called ‘facio-bucco-linguo-masticatory dyskinesia’ – later shortened to bucco-linguo-masticatory (BLM) syndrome. In 1960 Uhrbrand and Faurbye introduced the term tardive dyskinesia. As mentioned in the section on ‘Drug-induced choreas’ in Chapter 24, the use of this term is restricted to dyskinesias that develop after long-term administration of neuroleptic and other medication.

There are many drugs with different chemical and pharmacological properties encompassed by this group of agents. These include phenothiazines, butyrophenones, and thioxanthenes. These drugs have one pharmacological property in common, namely their ability to interfere with dopamine transmission in the brain, and this is thought to be the basis of their common antipsychotic effect.

As Harrison (1999) pointed out, there are drawbacks to the various approaches to the neuropathology of dyskinesias:

As for studies in humans, the most valuable human study of the neuropathological effects of antipsychotics is that by Jellinger (1977) and, overall, there are few well-controlled neuropathological studies of tardive dyskinesia (Jellinger 1977).

Jellinger (1977) who investigated 28 patients treated chronically with antipsychotics, most of whom had a diagnosis of schizophrenia, found alterations largely limited to the rostral part of the caudate nucleus. Changes consisted mainly of swollen large neurons, glial satellitosis, and sometimes a more generalized gliosis. Cerebral cortical atrophy, ventricular dilatation, degeneration in the substantia nigra, midbrain gliosis, and swelling of neurons of the dentate nucleus of the cerebellum have been recorded, but lack of well-matched control material makes the significance of these findings uncertain (Faurbye et al. 1964, Christensen et al. 1970, Arai et al. 1987). Overall, the findings in humans are anatomically congruent with those in rats [see Harrison (1999) for review].

CT brain scan has suggested that patients with tardive dyskinesia have slightly larger lateral ventricles than similar patients without this disorder, but whether this is a repeatable finding and its exact significance remain uncertain (Famuiwa et al. 1974).

Striatal dopamine transporter density binding measured with 123I-FP-CIT SPECT was found normal in schizophrenic patients with tardive dyskinesia (Lavalaye et al. 2001), Yoder et al. (2004), however, found that although subjects with and without tardive dyskinesia were not statistically different in striatal dopamine transporter binding potential, tardive subjects had apparently lower dopamine transporter density than non-tardive schizophrenic subjects. They also reported trend-level inverse correlations, especially with the severity of negative symptom scores, but also cognitive and depression/anxiety scores.

Amphetamine-induced presynaptic dopamine release and D2 receptor binding was normal as measured by 11Craclopride PET (Adler et al. 2002). Yet others had found a positive correlation of the striatal D2 receptor density measured by PET scanning with the severity of facial dyskinesia (Blin et al. 1989). However, another 11C raclopride PET scanning study has confirmed that there is an increase in D2 receptors in vivo even in schizophrenic patients without dyskinesias receiving long-term treatment with antipsychotics (Silvestri et al. 2000) (also see later under ‘Pathophysiological mechanisms’).

The concentration of dopamine metabolites in the cerebrospinal fluid (CSF) has shown no consistent change to suggest an abnormality of dopamine release from presynaptic nerve terminals (Pind and Faurbye 1970, Curzon 1973, Gerlach et al. 1975, Nagao et al. 1979). There have been a number of studies of plasma prolactin concentration and its response to dopamine receptor stimulating agents in an attempt to assess the sensitivity of the tubero-infundibular dopamine receptors in tardive dyskinesia. These have failed to reveal a consistent pattern of abnormality (Tamminga et al. 1980). Goff et al. (1995) found elevated CSF concentrations of alanine in medicated schizophrenia patients with tardive dyskinesia. CSF aspartate concentrations were elevated in patients with tardive dyskinesia when medication status was controlled for.

Studies purporting to show a relationship between platelet monoamine oxidase activity and tardive dyskinesia (Jeste et al. 1979) have not been confirmed (Glazer 1989), and neither have those suggesting increased blood dopamine beta hydroxylase and CSF noradrenaline levels (Glazer 1989). Claims that plasma concentrations of the noradrenaline metabolite 3-methoxy-4-hydroxyphenylglycol are reduced in cases demonstrating exacerbation of dyskinesia on neuroleptic withdrawal (Glazer 1989) seem of doubtful significance.

Autopsy studies of schizophrenics have shown no difference in dopamine receptor binding or in concentrations of dopamine and its metabolite, dihydroxyphenylacetic acid, in the striata between those with and without antemortem tardive dyskinesia. An increase in another dopamine metabolite, homovanillic acid, was noted in the putamen and nucleus accumbens of those with involuntary movements (Crow et al. 1982, Cross et al. 1985). Brain concentrations of cholecystokinin, substance P, neurotensin, somatostatin, vasoactive intestinal polypeptide, choline acetyl transferase, GABA, and cholinergic and GABAergic receptors have also been normal (Cross et al. 1985), although reduced levels of the GABA synthesizing enzyme, glutamate decarboxylase, have been noted in the subthalamus (Anderson et al. 1989) and low CSF GABA levels have been reported. Overall, there is no really convincing evidence of biochemical abnormality that might underlie tardive dyskinesia (Thaker et al., 1987).

Nothing is known about the central neurophysiological changes that underlie tardive dyskinesia. Electromyography has shown both rhythmical and apparently random discharges (Bathien et al. 1984). The former have been grouped into clonic (frequency 1–3 Hz) and tonic (frequency well under 1 Hz and burst length often over 1 sec) types. The latter show great variability in length, frequency, and amplitude of muscle activation. In all of these groups there is no pattern of reciprocal activation and inhibition between agonists and antagonists, which distinguishes the rhythmical movements of tardive dyskinesia from tremor. In fact, there is a tendency for agonists and antagonists to contract simultaneously in tardive dyskinesia (Bathien et al. 1984).

The above measures used to assess neurotransmitter and receptor function are relatively crude and lack of positive results has not invalidated the most popular theory of the pathophysiological mechanism underlying tardive dyskinesia. This is that prolonged dopamine receptor blockade results in development of receptor supersensitivity, which eventually becomes so great that the receptors become overstimulated by the small amount of dopamine available to them. This theory has the advantage of explaining many features of the disorder, including 1) the worsening or development of dyskinesia on reducing or withdrawing the drug; 2) the lessening or disappearance of the dyskinesia on increasing the dose of the drug; 3) aggravation by large doses of l-dopa or dopamine receptor agonists; 4) possible improvement with small doses of l-dopa or dopamine receptor agonists (by preferentially stimulating presynaptic dopamine receptors and decreasing release of dopamine from nerve terminals); 5) aggravation by anticholinergic drugs; and 6) improvement by cholinergic drugs.

Attempts to confirm that dopamine receptor sensitivity underlies tardive dyskinesia using animal models have only been partially successful. Rats maintained on neuroleptics for up to 1 year show altered dopamine receptor sensitivity even while neuroleptic administration continues. The receptors become supersensitive to agonists and subsensitive to antagonists (Jenner and Marsden 1983). This dopamine receptor supersensitivity would seem functionally significant as it is associated with an increase in basal striatal acetylcholine levels. This suggests that the activity of post-synaptic cholinergic neurons, which contain inhibitory dopamine receptors, has been decreased (Jenner and Marsden 1983).

Although such animals do develop spontaneous chewing movements, tongue protrusion, and tremor of the cheeks, pharmacological responses suggest they are more akin to an acute dystonic reaction than tardive dyskinesia. In addition, they disappear rapidly when neuroleptic administration is stopped and most indicies of dopamine receptor supersensitivity return to normal. It is uncertain if this failure to evoke tardive dyskinesia is due to species difference (Jenner and Marsden 1983).

Another finding which cannot readily be explained by the dopamine receptor supersensitivity hypothesis is that an increase in the number of these striatal receptors occurs within 1–2 weeks of neuroleptic administration and is maximal within 4 weeks (Christensen et al. 1976, Clow et al. 1980, Dewey and Fibiger 1983), which is much shorter than the time taken for the appearance of involuntary movements in experimental animals and tardive dyskinesia in patients. Also, as mentioned previously, although functional imaging data are somewhat controversial, 11C raclopride PET scanning study has shown that there is an increase in D2 receptors in vivo in schizophrenic patients without dyskinesias receiving long-term treatment with D2 receptor blocking typical and atypical antipsychotics (Silvestri et al. 2000). Also, there has been no clear correlation between the increase in the amount of dopaminergic receptors in the striatum and the presence or absence of tardive dyskinesia in schizophrenic patients (Crow et al. 1982, Korhuber et al. 1989) and in animal models (Knoble et al. 1994). These and other conflicting data have led some workers to postulate that it may not be a simple matter of dopamine receptor supersensitivity that underlies tardive dyskinesia, but disturbance of the functional interaction between D1 and D2 receptors: namely D1 hypofunction with concurrent D2 hyperfunction (Waddington 1989). In support of this hypothesis are findings that certain selective D1 receptor agonists can induce vacuous movements in rodents and the subtype of the D1 receptor, namely D1A, may have an important role in this regard (Van Kampen and Stoessl 2000). However, other authors have gone even further, suggesting dopamine receptor sensitivity may at best be necessary but not sufficient for tardive dyskinesia to develop, and at the worst to be an unrelated epiphenomenon.

Another suggested theory is that a reduction of inhibitory activity of the neurotransmitter GABA plays a role in tardive dyskinesia. Fibiger and Lloyd (1984) proposed that tardive dyskinesia might result from destruction of a subpopulation of neostriatal GABAergic neurons projecting to the substantia nigra and globus pallidus, and noted that long-term neuroleptic administration in rats was associated with neuronal loss in the ventrolateral striatum (Pakkenberg et al. 1973, Nielsen and Lyon 1978).

As mentioned under ‘Spontaneous oro-facial chorea’, these areas may be involved in control of movements of the mouth and may spontaneously degenerate with age (MacKay 1981). Neuroleptics could be envisaged as facilitating this process, possibly by promoting catecholamine turnover with the increased production of free radicals (Cadet et al. 1986). These drugs have also been shown to decrease the activity of GABA synthesizing enzyme glutamic acid decarboxylase in the substantia nigra, medial globus pallidus, and subthalamic nucleus of animals, including primates (Gunne and Haggstrom 1983, Gunne et al. 1984). They also produce a decrease in GABA turnover (Mao et al. 1977, Itoh 1983) and an increase in GABA receptor sites in substantia nigra pars reticulata (Gale 1980). This is consistent with the decreased concentration of GABA in CSF of patients with tardive dyskinesia, as noted above. It has been suggested that initially the subthalamic nucleus discharges at an excessive rate, as the result of blockage of dopamine receptors, and it causes glutamate mediated neurotoxicity with a functional lesion of the globus pallidus and substantia nigra (De Keyser 1991, Gunne and Andren 1993). This would result in decreased levels of glutamate decarboxylase in the subthalamic nucleus and pallidum. Reduced activity in the subthalamic projection to the medial globus pallidus and substantia nigra reticulata could produce downregulation of the pallidothalamic and nigrothalamic projections with consequent overactivity in the thalamocortical circuits. Experimental primate data support such a conclusion (Mitchell et al. 1992, Gunne and Andren 1993, Delfs et al. 1995), as do reports that GABA-mimetic drugs such as progabide and tiagabine can prevent haloperidol-induced dyskinesias in rat models (Kaneda et al. 1992, Gao et al. 1994).

Saccadic eye movements are known to be modulated by a GABAergic pathway, including the tonically active inhibitory projection from the substantia nigra pars reticulata to the superior colliculi (see under ‘Connections of the substantia nigra and superior colliculus’, including ‘Midbrain and pontine structures related to the basal ganglia’, in Chapter 1). These nigrocollicular fibres have to pause firing in order to allow the superior colliculi to initiate saccadic eye movements. Thus, theoretically, a decrease in nigral GABA activity would increase the ease of such movements and possibly a decrease in the threshold of saccadic distractability. In keeping with this it has been claimed that patients with tardive dyskinesia have twice the distractability of controls (Thaker et al. 1988, 1989).

Serotonin input from the raphe nuclei has been claimed to have an inhibitory action on neostriatal dopamine function and, as such, have a possible role in tardive dyskinesia (Seibyl et al. 1989). The serotonin antagonist activity of the atypical neuroleptic clozapine has been suggested to underlie its low propensity to cause tardive dyskinesia (see under ‘Management’). Similarly, abnormalities of noradrenergic function have been proposed, but evidence for a central role by either of these two transmitters is weak (Glazer 1989). Other mechanisms including upregulation of the adenosine receptors have been postulated, but there are limited data about this so far (Parson's et al. 1995).

Thus, a combination of the effects on the dopaminergic and the indirect striatopallidal and striatonigral pathways via the subthalamus may underlie the development of tardive dyskinesia. A role for other neurotransmitters, however, has not been excluded.

The question of whether neuroleptics cause damage to the basal ganglia and, if so, by what mechanism is a vexed one. These drugs may result in neurodegeneration by binding to neuromelanin (Seeman 1988), or through generation of free radicals with resulting peroxidation of lipid membranes or apoptotic mechanism (Pall et al. 1987, Burkhardt et al. 1993, Galili et al. 2000). Impairment of cellular energy metabolism through mitochondrial dysfunction caused by neuroleptics has been proposed (Andreassen and Jorgensen 1995) and that oxidative injury may have a role has also been suggested by the prevention of neuroleptic-induced oro-facial movement occurring by selegeline (but not vitamin E) in a rat model (Sachdev et al. 1999). Whether such changes actually occur is unknown.

Finally, whether there is a genetic predisposition to developing tardive dyskinesias is being considered and polymorphisms of different receptor genes have been studied.

Advances in genetic techniques have resulted in identification of DNA-sequence variation in genes in relation to serious side effects (tardive dyskinesia, weight gain) of antipsychotics. It is now also possible to relate a major proportion of the interindividual variation of plasma levels of most antipsychotics to specific genetic variants of genes and the encoded proteins, for example CYP P450-enzymes as CYP2D6. Moreover, a long list of mainly functional variants in target protein genes has been explored for their predictive power for the beneficial and adverse treatment outcome [see Maier and Zobel (2008) for review].

With respect to development of tardive dyskinesia, Steen et al. (1997) were the first to investigate the genetic variation of the dopamine D3 receptor as a putative risk factor for tardive dyskinesia in schizophrenic patients. This was a logical step as earlier it had been reported that dopamine receptor 3 gene polymorphisms may be a risk factor for developing schizophrenia itself (Durany et al. 1996). Steen et al. (1994) found a high frequency of homozygosity (22–24%) of a variant allele of the dopamine 3 receptor gene (Ser9Gly) with a serine to glycine variation at position 9 among subjects with tardive dyskinesia in both a cross-sectional and longitudinal evaluation as compared to relative underrepresentation of this genotype in those who did not develop tardive dyskinesia. This result indicated that autosomal inheritance of two polymorphic Ser9Gly (so-called 2-2 genotype) but not homozygosity for the wild type (1-1 genotype) alleles was a susceptibility factor for tardive dyskinesia (Steen et al. 1997). Further, the same group (Lovelie et al. 2000) and others also reported this association with the dopamine D3 receptor gene (Basile et al. 1999, Segman et al. 1999). However, this has not yet been conclusively proven as some have not found the same susceptibility (Rietschel et al. 2000).

The cytochrome P450 enzyme CYP2D6 alleles have also been studied and one report indicated a non-significant tendency for poor metabolizers to have a more severe rating of tardive dyskinesia, but no clear susceptibility was found on this basis (Andreassen et al. 1997). Another study reported an association between certain 5HT2A receptor gene polymorphisms with an increased risk of tardive dyskinesias in schizophrenic patients (Segman et al. 1999), but a further group did not find this association (Basile et al. 2000).

Subsequently, influences of other genes have been found including, for example, genetic variations in the RGS9 gene (Liou et al. 2008). Other genes were also investigated in schizophrenic patients with tardive dyskinesias and are thought not to play a major role, including the N-methyl-D-aspartate (NMDA) receptor subunit 2B (GRIN2B) gene (Liou et al. 2007) and the serotonin-2A receptor gene (HTR2A) (Basile et al. 2001).

Bakker et al. (2008) meta-analyzed pooled data from 1976–2007 for genetic association with tardive dyskinesias. They found: (1) in COMT (val-158-met), using Val-Val homozygotes as reference category, there was a protective effect for Val-Met heterozygotes (OR = 0.63, 95% CI: 0.46–0.86, P = 0.004) and Met carriers (OR = 0.66, 95% CI: 0.49–0.88, P = 0.005); (2) in Taq1A in DRD2, using the A1 variant as reference category, there was a risk-increasing effect for the A2 variant (OR = 1.30, 95% CI: 1.03–1.65, P = 0.026), and A2-A2 homozygotes, using A1-A1 as reference category (OR = 1.80, 95% CI: 1.03–3.15, P = 0.037); (3) in MnSOD Ala-9Val, using Ala-Ala homozygotes as reference category, there was a protective effect for Ala-Val (OR = 0.37, 95% CI: 0.17–0.79, P = 0.009) and for Val carriers (OR = 0.49, 95% CI: 0.24–1.00, P = 0.047). The authors concluded that although associations were small, the effects underlying them may be subject to interactions with other loci that, when identified, may have acceptable predictive power.

Thus, although more work is required in this area, it appears that there is first evidence that genetic factors may have a role in the pathophysiology of this disorder.

Schizophrenic patients may make a variety of abnormal movements which are related to their disease rather than therapy. It is now difficult to obtain sufficient numbers of such patients who have not had neuroleptic therapy to enable a study of these movements. Surveys prior to the use of the first neuroleptic drug in 1952 make it clear that such movements were not uncommon, and although terms such as ‘choreic’, ‘choreiform’, and ‘athetoid’ have been used (Kraeplin 1907, 1919, Bleuler 1911) a careful study of the descriptions suggests they differed from those of tardive dyskinesia. Although there was a spectrum of movements including simple tics, mannerisms, stereotypies, and complex integrated motor patterns, these were probably not the same as athetosis or chorea and did not bear a close resemblance to tardive dyskinesia. They are best described as stereotypies or mannerisms. Stereotypy was defined by Kraeplin (1919) as the persistence of the same movement or action either as the repetition of a movement or the maintenance of a posture. Mannerism was defined as the adornment of a voluntary act resulting in an affected and unnatural expression.

Another problem in interpreting early descriptions of abnormal movements in schizophrenia is the probable inclusion of a variety of other undiagnosed diseases, such as encephalitis lethargica, syphylis, and Huntington's disease, which cause both involuntary movements and mental impairment. It is thus difficult to explain the present-day epidemic of tardive dyskinesia on the basis that the abnormal movements are due to the psychiatric disorders.

None-the-less, more recent studies of patients with a variety of neuropsychiatric disorders, including schizophrenia, who have never been treated with neuroleptics, suggest that in most cases the oro-facial area is affected by the involuntary movements to some extent. In some cases this involvement may be very similar to that seen in typical tardive dyskinesia (Villeneuve et al. 1974, Owens et al. 1982, Waddington 1989). Spontaneous oro-facial dyskinesias were noted in 11% of drug-naïve schizophrenics in the first episode of psychiatric disturbance (Puri et al. 1999). Thus, estimates of the prevalence of tardive dyskinesia in patients with neuropsychiatric disorders are probably inflated by an unknown amount by cases in which the involuntary movements are not due to neuroleptic drugs.

The reported prevalence of tardive dyskinesia in patients on chronic neuroleptic treatment varies from 0.5 to 65% (Task Force 1980). Such widely differing results are probably due to differences in diagnostic criteria as well as patient populations. Some studies have used standardized scales, designed to assess the severity of tardive dyskinesia, as a means of establishing the diagnosis. Such scales, however, cannot reliably separate normal movements from tardive dyskinesia, and the estimated prevalence may show wide variation depending on the strictness of the criteria employed (Table 23.2).

Reviews of published studies suggest that the prevalence of bucco-linguo-masticatory tardive dyskinesia in patients on long-term neuroleptic therapy is between 20 and 40% (Task Force 1980, Kane and Smith 1982, Waddington 1989) (Table 23.3). This is in excess of the reported prevalence of similar spontaneous dyskinesia (see earlier). Prospective studies have suggested an incidence of about 3% for each year of exposure when allowance is made for remissions (Barnes et al. 1983, Kane et al. 1986, Chouinard et al. 1988).

One problem in evaluating the true problem of tardive dyskinesia is that in the majority of reports, most, if not all, subjects are still on neuroleptic therapy, which may be masking expression of the disorder. In an attempt to address this problem, Kane et al. (1988)

withdrew neuroleptic therapy in a group of subjects on long-term medication who had no signs of tardive dyskinesia, and found 67% of chronically ill and 17% of less chronically ill populations developed involuntary movements. In a prospective study in elderly patients with organic mental syndrome (67%) or psychiatric disorder (42%) the prevalence of tardive dyskinesia after 43 weeks of cumulative neuroleptic treatment was 31% (Saltz et al. 1991). Subsequently the same group reported that in a prospective study of 261 patients after a mean follow-up period of 114 weeks, the cumulative rates of tardive dyskinesia were 25%, 34%, and 53% after 1, 2, and 3 years respectively of cumulative antipsychotic treatment (Woerner et al. 1998).

The risk of developing tardive dyskinesia increases with advancing age. This parallels the incidence of spontaneous oro-facial chorea (Woerner et al. 1991, Koshino et al. 1992) (Table 23.4) and supports the notion that denervation of dopamine receptors with age results in increased sensitivity which is enhanced by exposure to neuroleptics.

Children almost never develop persistent tardive dyskinesia, although they may show a choreic syndrome lasting a few days after abrupt neuroleptic withdrawal (Wolf et al. 1993). Some studies have suggested a greater incidence of tardive dyskinesia in women (Miller and Jankovic 1990[a], Woerner et al. 1991). Although hormonal influences could underlie this (see under ‘Oral contraceptives and chorea’, Chapter 24), other investigations, including some that have controlled for age, have not confirmed a sex difference (Jus et al. 1976[a and b], Bell and Smith 1978, Perris et al. 1979, Kane et al. 1986, Glazer and Morgenstern 1988, Koshino et al. 1992). Van Os et al. (1999) suggested that the increased risk for women is only in the older age groups and in the younger age group men may be at more risk of tardive dyskinesia, tying in with the suggested role of oestrogen protecting against the condition (Turrone et al. 2000).

It has been suggested that schizophrenic patients may be somewhat resistant to developing tardive dyskinesia, and those with affective illnesses may be particularly susceptible (Kane et al. 1980, 1986, 1988, Mukherjee et al. 1986), although there is no reliable evidence that antidepressants are involved (Yassa et al. 1987). It has also been claimed that the movements are more severe but remit more rapidly following neuroleptic withdrawal in patients with affective disorders (Glazer and Morgenstern 1988). Some reports have contested the notion that affective disorders predispose to tardive dyskinesia (Schulze et al. 2001) and these observations must be regarded as tentative. The literature on schizophrenic patients reveals that the presence of cognitive dysfunction and negative symptoms, such as flattening of affect, poverty of speech, and social or emotional withdrawal, tend to be associated with tardive dyskinesia (Waddington and Youssef 1986, Waddington 1987, Krausz et al. 1999). A worsening of negative symptoms heralds the onset of tardive dyskinesia, according to Van Os et al. (2000).

The disorder has been reported in patients without psychosis, including those with personality disorders, psychoneuroses, chronic pain, dizziness, and gastrointestinal diseases (Druckman et al. 1962, Faurbye et al. 1964, Evans 1965, Degwitz and Wenzel 1967, Klawans et al. 1974, Sulkava 1984). In one survey only 43% of patients with tardive dyskinesia had evidence of psychosis (Miller and Jankovic 1990[a]). Organic brain damage with ‘soft’ neurological signs and radiological evidence of caudate atrophy or ventricular dilatation have been noted to be more common in patients with this movement disorder (Smith et al. 1978, Gerlach 1979, Perris et al. 1979, Waddington 1987, Waddington 1989) Previous electroconvulsive therapy ((Uhrbrand and Faurbye 1960, Faurbye et al. 1964, Degwitz and Wenzel 1967, Woerner et al. 1998) and leucotomy (Faurbye et al. 1964, Hunter 1964) have been suggested to predispose to tardive dyskinesia, but definite evidence is lacking (Brandon et al. 1971). The most definite risk factor for tardive dyskinesia is advancing years and many of these other putative associations, including deteriorated schizophrenia and evidence of diffuse brain damage, become more common with age. This makes it difficult to establish them as independently related variables. Even if they are related, it has not been prospectively shown that these predispose to tardive dyskinesia rather than being the result of neuroleptics. There is also limited evidence that patients who develop extrapyramidal side effects like parkinsonism, acute akathisia, and acute dystonia shortly after starting neuroleptics may be predisposed to eventual tardive dyskinesia (Barnes and Braude 1985, Kane et al. 1985, Chouinard et al. 1988, Kane et al. 1988, Woerner et al. 1998, Muscettola et al. 1999).

Tardive dyskinesia usually develops only after prolonged neuroleptic therapy (Crane 1973), but occasionally it has been reported after only a few months (Evans 1965, Quitkin et al. 1977, Kane et al. 1980). Apart from this, however, duration of therapy and drug dosage do not convincingly correlate with the incidence of tardive dyskinesia (Jenner and Marsden 1983, Waddington and Youssef 1986, Waddington 1989), although in one prospective study of Woerner et al. (1998) higher mean daily and cumulative antipsychotic doses were said to confer a higher risk.

The disorder has been recorded with virtually all drugs that block dopamine receptors. It had been suggested that neuroleptics with a lesser tendency to produce parkinsonian features, such as thioridazine, are less likely to produce tardive dyskinesia. Conversely it had been argued that the higher anticholinergic activity of such drugs, which decreases their potential to cause parkinsonism, may predispose to tardive dyskinesia. Neither of these proposals has been confirmed (Chien et al. 1980, Jenner and Marsden 1983). It has not been possible to definitely rank dopamine receptor blocking neuroleptic drugs according to their potential to cause this complication (Waddington 1989). It had been thought that more recent drugs which selectively block only one type of dopamine receptor like sulpiride, which blocks D2 receptors, may eventually prove an exception to this. However, this has not been proven to be the case and neither has the suggestion that the incidence of tardive dyskinesia caused by anti-emetics, such as metaclopramide, is low (Wiholm et al. 1984). Of all the neuroleptic drugs, perhaps clozapine has the least potential to cause tardive dyskinesias. However, drugs that delete presynaptic dopamine stores, such as reserpine and tetrabenazine, seem to carry a very low risk of producing tardive dyskinesia.

This complication has been recorded with reserpine, but only in patients who have had exposure to neuroleptics or evidence of brain damage (Wolf 1973). As reserpine has been used very commonly to control hypertension, the paucity of such reports testifies to its relative safety.

There are two major components to the movement disturbance. These are: 1) oro-facial dyskinesia, and 2) limb and axial dyskinesia (Mackay 1981, Barnes et al. 1983[b]). Oro-facial dyskinesia is the commonest manifestation of tardive movement disorders and occurs more frequently in older patients (Marsden et al. 1975, Barnes et al. 1983[b]).

The movements are virtually indistinguishable from those of spontaneously occurring oro-facial chorea (see earlier), although analysis of groups of subjects suggests the drug-induced variety may show relatively more movements of the lip and jaw but not the tongue and neck (Barnes et al. 1983[b]). Involvement of the forehead and eyebrows, as often seen in Huntington's disease, is only rarely present (Jankovic 1995). Some authors have stressed that the majority of these oro-facial movements are repetitive, patterned, and semirhythmic, and are best thought of as being tardive stereotypies, rather than chorea (Stacy et al. 1993, Jankovic 1995). Rarely facial movements may be dystonic and resemble Meige's syndrome (Weiner et al. 1981, Glazer et al. 1983). Speech may be impaired, with deviations especially in loudness and timing (Gerratt et al. 1984). In addition, some patients may show tremor of the lips and hands (Kidger et al. 1980) or rapid (5/sec) lip movements (‘rabbit syndrome’) (Jus et al. 1973). These latter movements represent neuroleptic-induced parkinsonism (they respond pharmacologically like Parkinson's disease) and should not be regarded as tardive dyskinesias.

There may be choreic and athetotic movements of the limbs (Fig. 23.3) with flexion, extension, or spreading of digits, touching, picking, rubbing, tapping of feet, and a variety of other movements. Truncal involvement includes shrugging, swaying, pelvic movements, and shifting weight. Occasionally breathing and swallowing may be affected and there may be a susceptibility to respiratory infections (Youssef and Waddington 1989). Many of these movements are also repetitive and patterned and have been regarded as being stereotypies (Stacy et al. 1993, Jankovic 1995).

Limb or axial dyskinesias tend to be more common in younger subjects, and in children the picture may be dominated by severe dystonia at these sites. Evidence of other topographical localization, however, is lacking and claims that in some groups of patients movements are greater on one side than the other has not been convincing (Waddington 1989). Tardive dystonia is discussed in ‘Acquired secondary (symptomatic) dystonias’, Chapter 43. Akathisia, in which there is continual restless shifting of posture associated with an intense urge to move to gain relief, occurs in a minority of cases and is discussed in ‘Akathisia’, Chapter 46. Tardive tics and myoclonus are mentioned in the sections on ‘Tics’ (Chapter 28) and ‘Myoclonus’ (Chapter 33).

Tardive dyskinesia disappears during sleep and is aggravated by stress (Walters et al. 1990). Movements at one site can be accentuated by using another group of muscles. An essential feature of the movements is that they may be aggravated or first appear with reduction of neuroleptic medication and are improved by increased dosage. If neuroleptics are continued, tardive dyskinesia and parkinsonian features may co-exist in the same patient (Crane 1972, Richardson and Craig 1982). Anticholinergic drugs, which are commonly used in conjunction with neuroleptics to diminish parkinsonian side effects, aggravate the dyskinesia. The involuntary movements may even disappear if anticholinergics are withdrawn (Good 1981). It is uncertain whether concomitant administration of these drugs predisposes to tardive dyskinesia or affects long-term prognosis. In a small number of cases, tardive dyskinesia has fluctuated dramatically with mood swings in manic-depressive patients, and although relationship to mood state has been consistent in individual patients, both aggravation and improvement of movements with depression and mania have occurred (Ashcroft et al. 1972, Davies et al. 1976, Cutler 1980, de Potter et al. 1983).

If neuroleptics are continued most patients with tardive dyskinesia are not distressed by the movements and many may be unaware of them (Macpherson and Collis 1992). This is especially true of minor degrees of stereotypy or chorea, including that involving the mouth and face. Other patients, however, are troubled by their movements and this is particularly likely if they are dystonic. Occasionally there can be major disability, such as dyspnoea in respiratory dyskinesia (Nishikawa et al. 1992) and severe dysphagia (Gregory et al. 1992). Although some movements may not be physically bothersome they give the patient a bizarre appearance and may add to social isolation in patients who already have adjustment problems.

Overall, the disorder seems non-progressive in most cases (Barnes et al. 1983[a], Waddington 1989), although in a small number there may be rapid deterioration with the appearance of a severe dyskinesia (Gordon et al. 1987). In a prospective study of younger patients maintained on neuroleptics, 40% developed involuntary movements over an 8-year period, but in only 22% of these was it apparent for more than 6 months (Kane et al. 1986). Another follow-up investigation, however, found only 19% of cases remitted after 5 years (Waddington and Youssef 1989). There could be expected to be wide differences depending on populations and the precise regimen of maintenance neuroleptics. Thus, in a group of older subjects followed for 11–12 years after onset of dyskinesia and maintained on neuroleptics, it improved in only 18%, worsened in 21%, and did not remit in any (Koshino et al. 1991). Another study with a 14-year follow-up of 53 inpatients continuing to take neuroleptics found that overall there was improvement of the tardive dyskinesia in 66% of cases (with worsening of associated parkinsonism), worsening of tardive dyskinesia in 17%, while in 15% the condition remained the same. Tardive dyskinesia resolved in 33/58 patients (62%) and developed anew in two (Fernandez et al. 2001). There was no association of tardive dyskinesia outcome with age, sex, total number of drugs, use of atypical antipsychotics, or dosage in this study.

Tardive dyskinesia in children has been said to completely remit if neuroleptics are discontinued (McAndrew 1972, Jankovic 1981), but the prognosis in adults is less favourable. Recovery may be more likely in young adults (Degwitz 1969, Jacobsen et al. 1974, Casey et al. 1979) and less common in the elderly (Smith and Baldessarini 1980). Most studies suggest that the longer the period of drug withdrawal, the greater the chances of remission (Table 23.5). A small number of patients recover after delays of 2–5 years (Klawans et al. 1984). Approximately a third to a half show resolution of movements within 1 year and eventually about 60% may recover (Jenner and Marsden 1983, Marsden 1985, Marsden et al. 1986, Jeste et al. 1988), although estimates vary widely.

Because recovery is often delayed and some patients are left with persistent movements, it has been thought that neuroleptics can cause permanent tardive dyskinesia. Assuming the prevalence of tardive dyskinesia among patients on chronic therapy is about 20% (see earlier) and eventual recovery occurs in 60% of these, the prevalence of persistent tardive dyskinesia would be about 8%. This is similar to the prevalence of spontaneous dyskinesia of about 5% (Puri et al. 1999) (see Table 23.1). Although these figures are gross approximations they emphasize the point that it is as yet uncertain that neuroleptics cause permanent dyskinesia.

There are a small number of reports of increased mortality in patients with tardive dyskinesia, but data are conflicting and this remains uncertain (Waddington 1989).

Because of the inherent risks, neuroleptic drugs should only be prescribed for proper psychiatric or neurological indications and their use as non-specific sedatives or tranquillizers is not justified. It has been suggested that frequent ‘drug holidays’ during which the neuroleptic is withdrawn for a period may protect against developing tardive dyskinesia (Jus et al. 1976[b]). The most convincing evidence, however, suggests the converse and patients who have had a greater number of drug-free periods may be more likely to develop dyskinesia, which then becomes permanent (Degwitz 1969, Jeste et al. 1979, Nordic Dyskinesia Study Group 1986). This situation remains to be clarified.

When tardive dyskinesia has developed any anticholinergic that is being co-administered should be stopped or reduced to the smallest dose possible as it is likely to be exacerbating the movements (Jeste et al. 1988). Neuroleptic medication should be withdrawn if clinically possible. If the patient's psychiatric state does not allow this, it is probably best to substitute the neuroleptic with a selective D2 receptor blocker such as sulpiride. This has reasonable antipsychotic activity and seems to have a low potential to cause tardive dyskinesia, although it is not free of this side effect (Achiron et al. 1990, Miller and Jankovic 1990[b]). Tiapride is another selective D2 receptor blocking neuroleptic which may help control tardive dyskinesia, although its potential to cause or continue the disorder is uncertain (Pollack et al. 1985). Both sulpiride and tiapride have been found to be effective in treatment of tardive dyskinesia, in the sense of reducing movements (see Tables 23.6 and 23.7), but the numbers of patients treated have been relatively small and the duration of therapy limited. Risperidone, a potent serotonin and weak D2 antagonist, and remoxipride, an atypical neuroleptic with high D2/D1 specificity, may have potential use in this disorder (Meltzer 1993, Carpenter and Buchanan 1994, Jankovic 1995).

An alternative is to use clozapine, a potent D4 receptor antagonist. This has high antipsychotic potency and a relatively low risk of tardive dyskinesia. The major disadvantages are the 1–2% risk of agranulocytosis and the necessary frequent monitoring that

this entails (Jankovic 1995). It has been found to reduce symptom severity by at least 50% in approximately 43% of patients with tardive dyskinesia (Lieberman et al. 1991, Gardos et al. 1999) and the response is said to be sustained by some (Tegeler 1992) but not others (Yovtcheva et al. 2000). However, there is some disagreement and Modestin et al. (2000) found that clozapine did not reduce tardive dyskinesia. In the latter study, 46 patients treated with clozapine were compared with 127 receiving typical neuroleptics over a 3-year period. The authors reported higher tardive dyskinesia scores in the clozapine group and no reduction of extrapyramidal side effects (Modestin et al. 2000). Despite this report clozapine overall appears to be superior in efficacy and safety compared to other neuroleptics (Chakos et al. 2001) and its major limiting problem is the need to monitor white cell counts regularly. Olanzapine, which has similarities to clozapine, does not have the problem of agranulocytosis. It has also been found to be effective and can be considered instead of clozapine (Raja et al. 1999). Quetiapine, another newer atypical neuroleptic, has also been said to be useful. However, further larger trials are required with these newer agents.

When neuroleptics are withdrawn it initially aggravates the dyskinesia. If the situation is explained to the patient and relatives they may be prepared to tolerate the movements without additional medication. This policy probably carries the best chance of eventual remission. If, however, the movements prove intolerable, other medication may be required. The beta adrenergic blocking drug propranolol and the alpha-1-adrenergic agonist clonidine have been reported to be helpful in up to about two thirds of cases (Jeste et al. 1988), but the number of patients studied by double-blind techniques has been relatively small. A benzodiazepine may provide some relief and clonazepam has been moderately successful in some studies (Bobruff et al. 1981) and not very useful in others (Thaker et al. 1990). Vitamin E (alpha tocopherol) 1200/U daily has been reported to lessen movements (Lohr et al. 1987, Elkashet et al. 1990), but in a placebo-controlled trial it has been ineffective (Shriqui et al. 1992). The calcium channel blockers verapamil (Barrow and Childs 1986) and diltiazem (Ross et al. 1987) and others have been suggested to be helpful, but the evidence is not compelling (Soares and McGrath 2001). If these are ineffective, an acetylcholine precursor, such as choline chloride or lecithin, should be tried (Growdon and Gelenberg 1978), but this is of unproven value. GABAergic drugs including baclofen (Korsgaard 1976), progabide (Ziegler et al. 1987), muscimol, sodium valproate, and tetrahydroisoxazolopyridine (THIP) have been said to be helpful, but controlled trials of baclofen (Glazer et al. 1985) and sodium valproate (Chien et al. 1978) have been negative. The Cochrane database systemic review concluded that although no clear statement could be made about their efficacy, when all the data on GABA agonist drugs were combined they tended to be associated with some degree of improvement of tardive dyskinesia symptoms, but also had side effects such as confusion and drowsiness (Soares et al. 2001). Gabapentin has also been reported to be useful for both its anti-tardive dyskinesia and its mood stabilizing properties in bipolar and schizoaffective patients. It may be used instead of neuroleptics (Hardoy et al. 1999). The selective serotonin 3 receptor antagonist ondansetron has been shown to produce improvement in tardive dyskinesias in an open-label study (Sirota et al. 2000). Another open-label trail reported that branched-chain amino acids decrease the symptoms of tardive dyskinesia (Richardson et al. 1999). The cholecystokinin-8 analogue ceruletide has been found to have a 42% improvement rate compared with 9% in a placebo group (Kogima et al. 1992). The fact that so many different treatments have been tried reflects the fact that no treatment is reliably helpful and most trials have small numbers and are uncontrolled. The Cochrane database systemic review evaluated data for cerulutide and other miscellaneous treatments for tardive dyskinesia including essential fatty acids, oestrogen, lithium, and insulin and found that apart from low-dose insulin therapy, none produced statistically significant improvement (McGrath and Soares 2000).

Should there still be a management problem after trying a range of these measures, it is probably justifiable to introduce a dopamine store depleting drug such as tetrabenazine or reserpine. These have a very low tendency to produce tardive dyskinesia and have theoretical advantage over dopamine receptor blocking drugs. Tetrabenazine is the more effective of the two and has been shown to be well tolerated during long-term administration (Mikkelsen 1983, Ondo et al. 1999). Jankovic (1995) reported that in treatment of over 900 patients no persistent movement disorder was seen that could be attributed to this drug. Care must be taken as these agents may produce depression and have definite risks in this group of patients.

Dopamine receptor blocking drugs should be reintroduced only as a last resort. Although they will usually suppress dyskinesia satisfactorily they may make it irreversible. This plan of action is relatively simple, but it is not free from controversy as a large number of other drugs have been claimed to improve tardive dyskinesia (Table 23.8). The multiplicity of these testifies to their limited effectiveness and at present many of them must be regarded as experimental. A summary of some double-blind studies is given in Table 23.9. Yet another approach is the use of dopamine receptor stimulants. Small doses may act on presynaptic dopamine receptors to decrease dopamine release at synapses and thus diminish involuntary movements (Delwaide and Hurlet 1980). In addition, as dopamine blockade may have produced dopamine supersensitivity it has been postulated that dopamine receptor stimulating drugs may reduce sensitivity and hasten recovery. Although large amounts produce unacceptable aggravation of movements, a small dose may be tolerated and the involuntary movements show improvement on drug withdrawal after a course of therapy (Alpert et al. 1983, Shoulson 1983). Some studies, however, have failed to produce convincing improvement (Hardie et al. 1983, Simpson et al. 1988, Lieberman et al. 1989). Finally, botulinum toxin injections can be used to treat particularly distressing movements and may be especially helpful if there is a dystonic element (Truong et al. 1990, Stip et al. 1992, Yasufuku-Takano et al. 1995).

All of the above serves to reinforce the conclusion of the Task Force on Late Neurological Effects of Antipsychotic Drugs that ‘no treatment to date uniformly benefits dyskinesia in all patients’. To this must be added the warning not to make the cure worse than the complaint as many patients are not bothered by these involuntary movements.