New-Onset Tardive Dyskinesia in Patients With First-Episode Psychosis Receiving Risperidone or Haloperidol

To The Editor: Tardive dyskinesia associated with antipsychotics is an important treatment concern. A recent long-term trial provided a unique opportunity to study tardive dyskinesia in a double-blind evaluation of haloperidol and risperidone in first-episode psychosis (1) . The findings were consistent with previous reports, suggesting lower rates of dyskinesia with atypical antipsychotics compared to conventional agents (1, 2) . However, the a priori definition of tardive dyskinesia is a threshold increase over baseline levels of dyskinesia. This did not distinguish between dyskinesia after withdrawal of prior antipsychotics from new-onset, possibly drug-induced, cases.

We recently defined criteria for tardive dyskinesia using the Extrapyramidal Symptom Rating Scale and based on parameters previously reported (3) . Probable tardive dyskinesia was defined as scores ≥2 (2 or 3) (mild) or ≥1 of 4 (moderate) or greater on any item on the dyskinesia subscale of the Extrapyramidal Symptom Rating Scale. New-onset tardive dyskinesia was identified when patients without baseline dyskinesia (all items: 0 or 1) met this definition at two more consecutive visits (≥4-week duration), with onset beyond week 4 of discontinuing an oral antipsychotic or beyond week 8 of discontinuing a depot antipsychotic (ruling out withdrawal from tardive dyskinesia [4] ). The first application of these Extrapyramidal Symptom Rating Scale criteria to a database was recently reported (3) .

We applied this method in a post hoc analysis of a study by Schooler et al. (1) . Extrapyramidal Symptom Rating Scale dyskinesia data were available for 458 patients; 444 were rated without baseline dyskinesia. Among these patients, 0.87% (2 of 229, 95% confidence interval [CI]=0.00–2.08) receiving risperidone and 2.33% (5 of 215, 95% CI=0.31–4.34) receiving haloperidol met criteria for new-onset tardive dyskinesia (p=0.27, Fisher’s exact test). Annualized rates were 0.72% and 1.87%, respectively. Mean modal doses were 3.4 mg/day (SD=1.8) and 3.2 mg/day (SD=1.7), respectively.

This finding is similar to prior reports for risperidone (4) and atypical antipsychotics (2) . The approximately 2% annual rate of tardive dyskinesia with haloperidol is lower than previously reported and can be explained by the low dose in this study. The lower incidence with risperidone versus haloperidol was not statistically significant owing to the small number of subjects but may be considered clinically relevant given the concerning nature of tardive dyskinesia. Schooler et al. (1) reported higher rates in both groups, at least partly because of methods that did not discriminate withdrawal or existing dyskinesia from newly emergent tardive dyskinesia. The research criteria reported in this letter are consistent with other published criteria for identifying new-onset tardive dyskinesia.