There is a myth is that the risk of tardive dyskinesia worsens with time, TD is irreversible, acute extrapyramidal symptoms predict increased risk for later TD, and all neuroleptics have been shown to cause it. However, schizophrenia is associated with a spontaneous TD rate in healthy young adults of about 0.5% per year. This contrasts with the normal population and patients with affective disorders, who do not have any notable incidence of spontaneous TD below age 60. After age 60, however, in the non-psychiatrically ill general population, there is a spontaneous incidence of TD that approximates about 0.5% per year. These spontaneous rates probably reflect abnormalities in the extrapyramidal neuronal tracts of the brain. Hence there is a TD risk throughout life in schizophrenia possibly due to abnormal brain structures in that condition in those areas, and there is a TD risk occurring in late life in the general population presumably due to gradual degeneration of brain function in susceptible persons in those areas. Consequently, TD occurs unrelated to neuroleptic medications. We are interested in the risks related to those medications, and thus we must be careful not to attribute spontaneous TD to neuroleptics.
Probably the most carefully conducted long-term study of TD with traditional neuroleptics was conducted at Yale. In that study, 398 patients with psychotic disorders (mostly schizophrenia) were followed prospectively with TD rating scales every three months for 8 years (1985-1993). An average TD incidence of about 5% per year was noted, but the important finding, which conflicts with accepted opinion, was that almost 20% of patients developed TD in the first 3 years of treatment. After the first three years, the TD rate seemed to plateau at a rate of about 1% per year. It is important to remember that the spontaneous TD rate in schizophrenia is about 0.5% per year. Thus, the added risk due to neuroleptics is about 0.5% per year after the first 3 years of treatment. The earlier TD literature seems to suggest overall TD rates of about 40-50% after about 20 years of neuroleptic treatment. The myth is that this risk is linear; the Yale study shows that the risk is asymptotic, with half the TD occurring in the first few years of treatment, and the other half occurring very gradually over the following two decades. In other words, as the Yale study investigators note, in contrast to accepted wisdom, TD risk is highest early in treatment, in the first few years of new treatment with neuroleptics in a person never previously treated with those agents. If a person has taken a neuroleptic for 19 years, they are not very likely to develop TD in year 20. After the first few years of treatment, those who have not developed TD represent a relatively TD-resistant cohort of patients. Such patients are much less at risk for TD than someone who has never taken neuroleptics before and is newly being prescribed them.
It is also important to bear in mind, in relation to the above figures, that not all these cases represent irreversible TD. TD is often temporary, resolving after a time. Thus, patients observed to develop TD in the above reports may not have continued to manifest symptoms years later.
The findings of the Yale study have generally been confirmed by a number of other prospective studies of TD, with the added proviso that the risk of TD in the elderly (above age 60 for the purposes of research) patient with schizophrenia is even higher. In the first year of treatment with a traditional neuroleptic, the risk of TD has been shown to be 25-38%, and after only two years about 34-66% in various studies. Hence, the elderly patient has about as much TD risk in one year as occurs in 5 years in a young adult.
These data directly bear on our ability to know whether atypical neuroleptics possess much risk of TD. One frequently hears from clinicians that we do not have enough experience with atypical neuroleptics to understand their TD risk. These clinicians assume that we need 10-20 years of follow-up to be able to make such assessments. However, based on the data provided above with traditional neuroleptics, 3-5 years of data should provide evidence on the highest risk period, and we do have such data with most atypical neuroleptics. For risperidone, double-blind, controlled data on its use in the first year of treatment in clinical trial conditions (n=3298) indicates a TD incidence of 0.6% vs. 2.7% with haloperidol. With olanzapine, in clinical trials of 1714 patients with schizophrenia, schizoaffective disorder, or schizophreniform disorder treated for up to 2.6 years with either olanzapine or haloperidol in a double-blind fashion, the 1 year risk of TD was 0.52% with olanzapine vs. 7.45% with haloperidol (p=0.002). The risk ratio was 11.86 (95% CI=2.30, 61.14); thus the risk of TD is almost 12 times higher with haloperidol than with olanzapine. The rates with risperidone and olanzapine were equal to the spontaneous rate in schizophrenia. We would expect rates in the 5-10% range as seen with the haloperidol control if atypical neuroleptics were as risky as traditional agents in the first year of treatment. The first year of treatment is the highest risk period for TD. Risk of TD with risperidone has also been studied in the high-risk elderly population with schizophrenia, with a TD rate of about 5% with risperidone at 9 months of treatment versus about 30% with haloperidol (total n=122).
To clairfy: The PL view is not that TD never occurs with atypical neuroleptics nor that it should never be attributed to atypical neuroleptics. However, such TD is extremely rare and there is enough evidence to feel relatively certain that when it occurs it is mild.
It should be noted that in the CATIE study, patients who had pre-existing TD were excluded from assignment to treatment with perphenazine, and thus TD risk comparisons could not be made between atypical in typical neuroleptics in that study.