Tardive Dyskinesia :

Daniel E. Casey

Tardive dyskinesia (TD), a syndrome of potentially irreversible, involuntary hyperkinetic dyskinesias that occurs during long-term neuroleptic treatment, is a major limitation of chronic antipsychotic drug therapy. This has stimulated substantial research to identify the underlying pathophysiological mechanisms of TD, with the hope that these results will lead to new drugs that maintain antipsychotic efficacy but are free of TD risk. Data from multiple lines of investigation in both the preclinical and clinical realms have been brought to bear on the topic, but have not yet produced a parsimonious explanation. Hypotheses span a wide range of concepts that incorporate neurochemical and/or structural abnormalities associated with the underlying psychoses as well as the drug therapies for these disorders. Each hypothesis utilizes multiple models to marshal data relevant to specific issues. The aim of this chapter is to review and critique the results that are in support of, or in conflict with, the major hypotheses about the pathophysiology of TD (see Psychotropic Drug Metabolism in old Age: Principals and Problems of Assessment).

The hypothesis of dopamine hypersensitivity has dominated the conceptual approaches to studying TD. It proposes that the nigrostriatal dopamine system develops increased sensitivity to dopamine as a consequence of chronic dopamine receptor blockade induced by neuroleptic drugs (45, 68). This derives from the classic studies of denervation-induced hypersensitivity seen in peripheral muscles. The very large database generated by investigating this hypothesis can be interpreted as limited support for, or ample evidence against, the dopamine hypersensitivity hypothesis (12).

Behavioral evidence of dopamine hypersensitivity following neuroleptic treatment is seen in many different models in several species. The most common is exaggerated oral stereotyped behavior in response to acute dopamine agonist challenges after discontinuing neuroleptic drugs. Such increased behavior can be seen after a single dose, a few days, several weeks, or 1 year of neuroleptics in rodents (18, 20, 45, 64, 68) and nonhuman primates (10). Biochemical changes of increased numbers of dopamine D2 receptors are also usually found after repeated neuroleptic treatment (8), but this is not the case in all studies (73).

Though this model has served as the bedrock of the dopamine D2 receptor hypersensitivity hypothesis of TD, there are several fundamental flaws. The principal problems center around the incompatibility between features of the model and the clinical TD syndrome. In patients, the symptoms gradually develop over many months or years of treatment and continue to be present without the requirement of dopamine agonist provocation. In contrast, all animals rapidly develop behavioral and biochemical measures of hypersensitivity, which quickly resolve following neuroleptic discontinuation, whereas only a subgroup of patients develop TD, and it persists for long periods of time (14), except in the case of withdrawal dyskinesias (3). Perhaps these behavioral and biochemical models in rodents are better conceptualized as characterizing the acute and subacute effects of neuroleptics in extrapyramidal syndromes (EPS) or antipsychotic actions (12).

The model of spontaneous vacuous chewing movements (VCMs) that increase with chronic neuroleptic treatment has been proposed as a closer fit to the time course of clinical TD (24). In this setting, traditional neuroleptics show a wide range of inducing VCMs that correspond to the milligram potency continuum (i.e., haloperidol produces more and chlorpromazine produces fewer movements) (33). However, this is not a clinical feature of TD, because all neuroleptics used in the clinic appear to have similar TD liability, with the exception of clozapine (13, 42). These VCMs increase with age and neuroleptic treatment (60, 73) but may or may not correlate with changes in dopamine receptor numbers (73). Reversibility of VCMs is also unclear, because one study noted prompt resolution when discontinuing neuroleptics or giving anticholinergics (60), but another observed persistence for at least 2.5 months (73). Perhaps this VCM model also reflects more of the neuroleptic effects in acute EPS than in TD. Alternatively, it may be modeling both early and late aspects of neuroleptics in rodents, but those differ in some important features from the clinical findings. It is also possible that the chewing movements, which are variously described by different investigators, are actually different phenomena seen in these different studies.

The model of TD in nonhuman primates much more closely fits the clinical features (10, 11, 34) but provides only limited support for the dopamine hypersensitivity hypothesis. Dopamine turnover was significantly decreased in the caudate and substantia nigra in monkeys with TD 2 months after neuroleptics were discontinued (35), though no receptor quantification was done to compare and contrast findings with other species.

A modification of the D2-receptor hypersensitivity hypothesis incorporates a role for D1 dopamine receptors. The core of this hypothesis is that clozapine has a unique biochemical profile at D1 and D2 receptors that may be associated with atypical behavioral and clinical features. Clozapine in rodents does not produce biochemical or behavioral indications of basal ganglia-related D2 hypersensitivity (21). Similarly, clozapine produces very little EPS and TD in patients (13) and has relatively high D1 and moderate D2 occupancy rates, as measured by positron emission tomography (PET) (25). These observations have led to the proposal that TD develops from an imbalance between D1- and D2-mediated effects in the basal ganglia (29, 30). This possibility is supported by the wealth of data documenting the functional interactions of D1 and D2 receptors (72), but it is too soon to know if there is a critical role for D1-D2 interactions in TD.

Data from the clinic are also difficult to fully reconcile with the dopamine-receptor hypersensitivity hypothesis. Efforts at finding direct evidence of support for this proposal have been unsuccessful. Receptor binding studies of human postmortem brain tissue comparing schizophrenic patients with and without TD found no significant differences in either D1 or D2 receptors (22, 23). However, these data must be considered in the context of concern about (a) the time of, and conditions at, death and (b) current and past drug treatment that may obscure the role of dopamine in TD. Neuroimaging with PET to assess functional evidence of dopamine-receptor hypersensitivity has not been systematically applied to the question.

Biochemical data from the clinic also fail to support dopamine dysfunction as the explanation of TD. When comparing TD patients with non-TD patients, there were no consistent differences in (a) dopamine-mediated endocrine measures of prolactin or growth hormone or (b) cerebrospinal fluid (CSF), plasma, or urinary homovanillic acid assessments (40, 44, 66).

Clinical pharmacological observations suggest an important role for dopamine in TD. There is an increased incidence and prevalence of involuntary hyperkinetic dyskinesias in patients receiving dopamine antagonists in most (42, 43, 62), but not all, reports (71, 74). Additionally, dopamine antagonists suppress TD, whereas dopamine agonists usually aggravate TD symptoms (14). However, even in these findings, observations are conflicting, because potent direct dopamine agonists such as bromocriptine fail to markedly worsen TD (37), and neuroleptics may paradoxically increase TD in a few patients (16, 31). These clinical data, which in part seem to support the dopamine hypersensitivity hypothesis, may best be incorporated into an understanding of TD if dopamine plays a secondary or modulatory role. If the primary pathophysiology lies outside the dopamine system but is indirectly influenced by dopaminergic mechanisms, the clinical pharmacological impact of perturbations from dopamine agonists and antagonists could be parsimoniously combined (12, 14, 71), with the absence of any direct evidence of dopamine hypersensitivity.

An alternate hypothesis recently receiving considerable interest is the proposal that TD is due to neurotoxic effects of free radical byproducts from catecholamine metabolism. The basal ganglia, by virtue of their high oxidative metabolism, would be particularly vulnerable to membrane lipid peroxidation as a result of the increased catecholamine turnover induced by neuroleptic drugs (9, 50, 51). This concept has also been proposed as an explanation for several dyskinetic syndromes referable to the basal ganglia (50). While this hypothesis is speculative, there are data to support the proposal.

Animal studies in rodents receiving acute or subacute neuroleptic treatment show a mitigating effect of vitamin E (a-tocopherol) on behavioral measures of dyskinesias (50). This benefit is purportedly on the basis of vitamin E serving as a free radical scavenger, thus reducing the potentially cytotoxic effects of free radicals.

Clinical studies have produced conflicting data in this area. Some studies have found increased levels of lipid peroxidation byproducts in blood or CSF of TD compared to non-TD patients (52, 58). In clinical trials with TD, the majority of patients with TD did not show major (>50%) improvement (1, 51, 65). However, there was a more favorable outcome in those patients who were younger and had shorter durations of TD (1, 51, 65)—both of which are factors associated with a greater likelihood of spontaneous recovery from, or improvement in, TD (15). Other possible explanations for these findings that are yet to be explored include alterations in neuroleptic metabolism or brain drug levels from vitamin E, which could have the effect of raising the amount of neuroleptic antagonism at dopamine receptors and thus mask or suppress TD.

Another competing hypothesis involves gamma-aminobutyric acid (GABA) insufficiency in the neuroanatomical loop controlling movement (26, 27, 63). Again the data are conflicting. In rodent models there may (27, 63) or may not (55, 61) be consistent alterations in GABA function associated with acute and chronic neuroleptic treatment. Studies in nonhuman primates indicate a GABA association with TD because decreased glutamic acid decarboxylase (GAD), the GABA synthesizing enzyme, was noted in the substantia nigra, medial globus pallidus, and subthalamic nuclei in dyskinetic monkeys compared to similarly neuroleptic treated non-TD monkeys (35, 36). These were the same animals that also have abnormalities in dopamine parameters listed above (35).

Clinical biochemical data suggest that GABA may play a role, whereas clinical treatment trials are not as supportive. A small postmortem study found a trend toward decreased GAD in the medial globus pallidus, but normal levels in the nigra in TD in patients (2), and a separate study reported decreased CSF levels in TD patients (70). However, treatment trials with drugs that enhance GABA have been disappointing. Single doses of muscimol partially decreased TD but produced unacceptable toxicity and sedation (67), and trials with other agonists produced negative (17, 46, 47) or mixed results (69). Thus, it has not been possible to effectively treat TD with GABA acting agents, which raises important questions about a primary role for GABA in this disorder.

Because most neuroleptics also antagonize many other receptor types besides dopamine, it is possible that these play an important role in the pathophysiology of TD. A noradrenergic dysfunction theory derives some support from findings of greater dopamine b-hydroxylase activity in TD patients compared to non-TD patients (40) and a positive CSF norepinephrine correlation with TD severity (44). However, noradrenergic agents have not been successful treatments for TD (32, 41).

Serotonin may modulate dopamine activity and thus be involved with TD. However, efforts to find consistent abnormalities of serotonin parameters or effective serotonin treatments for TD (41) offer little support for this hypothesis.

There is some evidence that nutritional or metabolic issues may play a role in TD, but these concepts are currently more heuristic than well-developed. Meals which alter the phenylalanine/large amino acid ratio temporarily decrease TD movements (59). Another interesting observation is the increased risk of TD in diabetics (28, 56, 76), suggesting that alterations in glucose and/or insulin parameters may interact with neuroleptics.

Cholinergic hypofunction has been proposed as a cause of TD, but this is not adequately supported. Several other possible avenues of exploration include (a) roles for neuropeptides, such as cholecystokinin, substance P, neurotensin, and somatostatin (71), and (b) disorders of mineral metabolism, such as iron (4).

Electrophysiological measures of decreased activity in the nigrostriatal dopamine system (depolarization inactivation) following chronic treatment in rodents with traditional neuroleptics but not with clozapine (7) indicate this mechanism may be involved with TD. In contrast, some drugs known to produce TD, such as thioridazine, do not cause striatal depolarization block (75), and adding anticholinergics to traditional neuroleptics in rodents makes them appear atypical in this model (7), when this is not the case in patients. Thus, the role of depolarization inactivation in TD is unknown.

Attempts to identify a structural basis for TD have involved animals and patients. Electron-microscopic studies have recently shown that treatment for 2 weeks or more with traditional neuroleptics such as haloperidol increases perforated postsynaptic densities in the head of the caudate nucleus, but this is not seen with clozapine (54). The D1 antagonist SCH 23390 also produces this effect, but when combined with haloperidol to make D1 and D2 antagonism similar to clozapine, the combination does not alter structure (53). Thus, there are selective drug-induced neuroanatomical alterations that are site-specific and may be related to TD, but require much more investigation. Whether long-term neuroleptic treatment causes cell loss is unclear. Brains from rodents receiving several months of neuroleptics show either cell loss or no change (57) or cytoarchitectural alterations (5).

Postmortem studies in humans have also produced conflicting results. One light-microscopic study noted higher rates of nigral degeneration and gliosis in patients with TD (19), but another suggested that these findings may be nonspecific age-related changes (39). Neuroimaging investigations have similarly found variable results. Computerized tomography studies often find abnormalities, but the wide range of different methodologies makes it impossible to conclude if there are consistent structural defects associated with TD (38). The few reports of magnetic resonance imaging in TD also find several different questionable lesions that do not yet lead to a conclusion (4).

Up to this point in the search for the underlying pathophysiology of TD, considering neuroleptics as the sole etiological agent has not been fruitful. Expanding the issue to incorporate important patient parameters (e.g., disease, age) in addition to drug factors seems prudent (11).

Kraepelin (48) and Bleuler (6) observed abnormal movements of the orofacial and limb regions when they were describing the fundamental aspects of schizophrenia. Unfortunately, it is not possible to know the true prevalence of these phenomena or whether the movements described were the same as, or fundamentally distinct from, TD. A more recent report found that the type and severity of dyskinesias were similar in treated and untreated patients, but the prevalence was significantly higher in the neuroleptic group when age was controlled (23). Additionally, several studies indicate that TD patients often have greater cognitive dysfunction, negative symptoms, poor premorbid function, and poor prognosis (71).

Age may make a similarly important contribution to the expression of TD. Spontaneously occurring orofacial dyskinesias occur more often in aging humans and monkeys (11) and in patients with more neuromedical illnesses (49). The long observed positive correlation between age and neuroleptic-associated TD risk further emphasizes the role of age.

Thus, an important coalescence of critical factors may best explain TD. The underlying syndromes of psychosis, particularly schizophrenia, advancing age, and some unknown component of neuroleptic action, may combine to convert a covert vulnerability to dyskinesias to overt symptomology in the clinical constellation of TD (11). The central challenge which has not yet been fully overcome is to accurately attribute the relative contribution of each of these factors to the pathogenesis and pathophysiology of TD.

The pathophysiology of TD has been actively pursued, but remains elusive. Human studies and animal models have been a rich but incomplete source of data about the effects of neuroleptic treatment. A major limitation is that most of the animal models utilize short-term paradigms, which may more accurately assess the effects of acute or subacute treatment. To date, no direct evidence of any pathophysiological process has been identified, though indirect evidence suggests that dopamine, and perhaps other neurotransmitters, play a modulating or secondary role. A cogent argument can be made for a critical interaction between patient and drug parameters, but the specific contribution of these factors is unknown. In the current state of conflicting findings, it is best to keep an open mind about the pathogenesis and pathophysiology of the complex syndrome known as TD.

This work was supported, in part, by funds from the Veterans Administration Research Program and by NIMH grant 36657.

published 2000