OBJECTIVE The objective of this study was to evaluate the relationship between antipsychotic-induced tardive dyskinesia (TD) and estimated dopamine D2 receptor occupancy levels in patients with schizophrenia, using the dataset from the Clinical Antipsychotic Trials in Intervention Effectiveness (CATIE). METHODS The dataset from 218 subjects (risperidone, N=78; olanzapine, N=100; ziprasidone, N=40) who presented with a score of zero on the Abnormal Involuntary Movement Scale (AIMS) at baseline in Phase 1 of the CATIE study, and remained for ≥6months, was used. Peak and trough dopamine D2 receptor occupancy levels on the day of the AIMS assessment at the endpoint were estimated from plasma antipsychotic concentrations, using population pharmacokinetic analysis and our D2 prediction model. The estimated dopamine D2 receptor occupancy levels were compared between patients who presented an AIMS score of ≥2 at endpoint and those with a score of zero, using the Mann-Whitney U test. RESULTS Estimated dopamine D2 receptor occupancy levels at trough were significantly higher in subjects who developed involuntary movements (N=23) than those who did not (N=195) (71.7±14.4% vs. 64.3±19.3%, p<0.05) while no significant difference was found in the estimated peak D2 receptor occupancy between them (75.4±8.7% vs. 72.1±9.9%, p=0.07). When the analyses were separately conducted for the three drugs, there were no significant differences in estimated peak or trough D2 occupancy although the values were consistently numerically higher among those developing involuntary movements. CONCLUSIONS Greater dopamine D2 receptor blockade with antipsychotics at trough might increase the risk of tardive involuntary movements although this finding needs to be replicated in larger trials.