Predictors of Mortality in Parkinson's Disease

Abstract & Commentary

By Melissa J. Nirenberg, MD, PhD, Assistant Professor, Neurology and Neuroscience, Weill Cornell Medical College. Dr. Nirenberg reports that she has consulted for Biovail.

Synopsis: Clinical predictors of mortality in Parkinson's disease include age at onset, chronological age, male sex, motor severity, psychosis, and dementia.

Source: Forsaa, EB, et al. What predicts mortality in Parkinson disease? A prospective, population-based long-term study. Arch Neurology 2010;75:1270-1276.

Parkinson's disease (pd) is a clinically and genetically heterogeneous disorder, with striking variability in its presentation and rate of progression. Factors such as young age of onset and tremor predominance have been associated with an overall favorable prognosis. In contrast, features such as later age of onset, dementia, and prominent postural instability-gait disturbance tend to predict a more malignant disease course. Given this clinical heterogeneity, it is not surprising that there is also tremendous variability of life expectancy in PD. For this reason, it is of considerable clinical interest to determine factors associated with increased mortality in PD.

In this study, the authors examined clinical and demographic predictors of increased mortality in PD. The study population was a large (n = 230), prospective, community-based cohort in Norway. Subjects underwent serial motor and non-motor assessments between 1993 and 2005. The vital status of all study subjects was obtained from a national registry in 2009; this included the date of death for subjects who were deceased. Kaplan-Meier analysis was used to examine unadjusted survival after motor onset. Cox proportional hazards models were used to determine the independent predictors of mortality in the cohort.

Within the study period, 211 subjects (92%) died. Survival times ranged from 2.2-36.6 years, with a mean survival of 15.8 years from motor onset. The mean age of death was 81.1 ± 6.3 years (79.8 ± 6.1 for men vs. 82.4 ± 6.3 for women). Independent predictors of mortality included older age at onset, higher chronological age, male sex, higher Unified Parkinson's Disease Rating Scale (PDRS) motor scores (indicative of greater motor impairment), the presence of psychotic symptoms, and the presence of dementia. The use of atypical antipsychotics, however, was not an independent predictor of mortality. The authors conclude that treatments that prevent motor progression, dementia, and psychosis might reduce mortality in PD.


Physicians who treat patients with Parkinson's disease frequently are asked to provide patients with an estimate of their life expectancy and overall prognosis. Although several factors are known to portend a better or worse prognosis in PD, there is no way to predict the clinical course or life expectancy of an individual patient. This reality is reflected in the striking variability of survival times in this study—which ranged from 2.2-36.6 years after onset of motor symptoms.

Although individual life expectancy is highly variable in PD, this study identified a number of clear clinical predictors of mortality on a population basis. Male sex and higher chronological age, both major predictors of mortality in the general population, were not surprisingly also associated with increased mortality in PD. In addition, the authors identified several disease-specific factors associated with mortality in PD. These included older age of PD onset, greater UPDRS motor impairment, the presence of psychotic symptoms, and the presence of dementia. The use of antipsychotics was not independently associated with increased mortality—suggesting that studies associating antipsychotic use with increased morbidity in elderly patients with Alzheimer's disease may not be applicable to those with PD.

Strengths of the study include the large sample size, prospective study design, low dropout rate (for reasons other than death), and availability of complete and accurate mortality data. Limitations include the ethnic and geographical homogeneity of the population that was studied, and the observational nature of the study.

In summary, mortality in PD is associated not only with factors that increase risk in the general population (advanced age, male sex), but also with disease-specific factors such as age at PD onset, severity of motor impairment, and specific non-motor symptoms (dementia and psychosis). These findings may be helpful to physicians in identifying patients at increased risk of mortality.