OBJECTIVE: To evaluate construct validity of the Trunk Impairment Scale (TIS) as a measure of trunk performance in Parkinson's disease (PD). DESIGN: A cross-sectional study of PD patients and healthy subjects. SETTING: University rehabilitation research unit. PARTICIPANTS: Twenty-six PD patients (Hoehn and Yahr stages 2-4) and 26 healthy subjects. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: The TIS and its subscales; static and dynamic sitting balance and trunk coordination. RESULTS: Compared with healthy controls, PD patients showed significantly lower scores on the total TIS, static sitting balance, and coordination subscale. Healthy subjects scored significantly better on the total TIS and coordination subscale compared with patients in the early stage of PD. Patients with PD in the early stage scored significantly higher for the total TIS as well as static and dynamic sitting balance in comparison with PD patients in a later stage. Forward stepwise multiple linear regression analysis showed that trunk impairment in PD patients was significantly related to a combination of older age and a higher score on part III of the Unified Parkinson's Disease Rating Scale, which assesses motor impairments. CONCLUSIONS: Early detection of trunk deficits and the significant relation with PD severity advocates further evaluation and use of the TIS in PD.