What predicts mortality in Parkinson disease?: a prospective population-based long-term study.

OBJECTIVE: To identify independent risk factors of mortality in a community-based Parkinson disease (PD) cohort during prospective long-term follow-up.
METHODS: A community-based prevalent sample of 230 patients with PD from southwestern Norway was followed prospectively with repetitive assessments of motor and nonmotor symptoms from 1993 to 2005. Information on vital status until October 20, 2009, was obtained from the National Population Register in Norway. Cox proportional hazards models were applied to identify independent predictors of mortality during follow-up. Chronological age, Unified Parkinson's Disease Rating Scale (UPDRS) motor score, levodopa equivalent dose, probable REM sleep behavior disorder, psychotic symptoms, dementia, and use of antipsychotics were included as time-dependent variables, and age at onset (AAO) and sex as time-independent variables.
RESULTS: Of 230 patients, 211 (92%) died during the study period. Median survival time from motor onset was 15.8 years (range 2.2-36.6). Independent predictors of mortality during follow-up were AAO (hazard ratio [HR] 1.40 for 10-years increase, p = 0.029), chronological age (HR 1.51 for 10-years increase, p = 0.043), male sex (HR 1.63, p = 0.001), UPDRS motor score (HR 1.18 for 10-point increase, p < 0.001), psychotic symptoms (HR 1.45, p = 0.039), and dementia (HR 1.89, p = 0.001).
CONCLUSIONS: This population-based long-term study demonstrates that in addition to AAO, chronological age, motor severity, and dementia, psychotic symptoms independently predict increased mortality in PD. In contrast, no significant impact of antipsychotic or antiparkinsonian drugs on survival was observed in our PD cohort. Early prevention of motor progression and development of psychosis and dementia may be the most promising strategies to increase life expectancy in PD.