Dyskinesia and FAB score predict future falling in Parkinson's disease

Research output: Contribution to journalArticle


A growing body of research highlights the importance of cognition for prediction of falls in Parkinson's disease (PD). However, a previously proposed prediction model for future near falls and falls in PD, which includes history of near falls, tandem gait, and retropulsion, was developed without considering cognitive impairment. Therefore, by using a sample of 64 individuals with relatively mild PD and not excluding those with impaired cognition we aimed to externally validate the previously proposed model as well as to explore the value of additional predictors that also consider cognitive impairment. Since this validation study failed to support the proposed model in a PD sample including individuals with impaired global cognition, extended analyses generated a new model including dyskinesia (item 32 of Unified PD Rating Scale) and frontal lobe impairment (Frontal Assessment Battery—FAB) as significant independent predictors for future near falls and falls in PD. The discriminant ability of this new model was acceptable (AUC, 0. 80; 95% CI 0.68-0.91). Replacing the continuous FAB scores by a dichotomized version of FAB with a cut-off score ≤14 yielded slightly lower but still acceptable discriminant ability (AUC, 0. 79; 95% CI 0.68-0.91). Further studies are needed to test our new model and the proposed cut-off score of FAB in additional samples. Taken together, our observations suggest potentially important additions to the evidence base for clinical fall prediction in PD with concomitant cognitive impairment.


External organisations
  • Skåne University Hospital
  • Karolinska Institutet
  • Jönköping University
Research areas and keywords

Subject classification (UKÄ) – MANDATORY

  • Neurology
  • Physiotherapy


  • dyskinesia, falls, frontal lobe impairment, near falls, Parkinson's disease, prediction
Original languageEnglish
Pages (from-to)512-518
JournalActa Neurologica Scandinavica
Issue number6
Early online date2019 Mar 1
Publication statusPublished - 2019
Publication categoryResearch