Motor function in Parkinson's disease and supranuclear palsy: simultaneous factor analysis of a clinical scale in several populations

BACKGROUND In order to better understand the similarities and differences in the motor behaviour of different groups of patients, their scores on the Motor Examination section of the Unified Parkinson's Disease Rating Scale (UPDRS) were analysed simultaneously. The three groups consisted, respectively, of patients with Parkinson's disease (PD) on medication, patients with Parkinson's disease withdrawn from anti-parkinsonian medication for at least 12 hours, and patients diagnosed with a specific Parkinsonism syndrome: Progressive Supranuclear Palsy (PSP). METHODS A total of 669 consecutively sampled patients from three separate hospital-based clinics participated (294 PD on medication; 200 PD off medication: 175 PSP). The Motor Examination section of the UPDRS was administered by neurologists at the three participating clinics. The patient scores on each item were recorded. To assess similarities and differences among the components of the UPDRS in these samples, we performed simultaneous or multigroup factor analysis on the covariance matrices of the three groups. In addition, it was investigated whether a single model for the Motor Examination section of the UPDRS could be developed which would be valid for all three groups at the same time. RESULTS A single six-dimensional factor solution was found that fitted all groups, although this was not straightforward due to differences between the tremor-at-rest variables. The factors were identified as Tremor-at-rest, Postural Tremor, Axial Dysfunctioning, Rigidity, Left Bradykinesia and Right Bradykinesia. The analysis also pointed to a somewhat lower lateralization in bradykinesia for PSP patients. The groups differed in intensity of motor impairment, especially with respect to Tremor-at-Rest, but the overall relationships between the variables were shared by the three groups. In addition, apart from the common factor structure evidence of differences in body part-specific and motor-specific variances was found. CONCLUSION From a clinical point of view, the analyses showed that using the Motor Examination section of the UPDRS is also appropriate for patients with PSP, because the correlational structure of the items is directly comparable to that of Parkinson's patients. Methodologically, the analysis of all groups together showed that it is possible to evaluate similarities and differences between factor structures in great detail.