Development and internal validation of a prognostic model for loss of balance and falls in mid- to late-stage Parkinson's disease.

Abstract

Mid- to late-stage Parkinson's disease (PD) is often linked with worsened and significant impairment of motor activities, but existing prognostic markers do not adequately capture the risk of loss of balance in PD patients. This study aims to develop a risk prognostic model for mid- to late-stage PD and identify prognostic factors that are indicative of impending loss of balance and falls. The study included 307 participants of which 75 were diagnosed with idiopathic PD and 232 were neurological or non-neurological controls. Among the PD group, 46 were early-stage (Hoehn and Yahr [H&Y] = 1,2) with no significant loss of balance while 29 were mid- to late-stage (H&Y = 3,4,5) which is characterized by loss of balance and falls. Multivariable logistic regression (MLR) was used to develop a prognostic model for mid- to late-stage PD. Model discrimination was assessed by ROC curves. The model was internally validated through bootstrapping and calibration plots. The relevant factors identified and included in the final MLR model were shortness of breath, age, swollen joints, heme oxygenase-1 (HO-1) protein, and total salivary protein. The model had an AUC of 0.82 (95% CI = 0.71-0.92) and was well calibrated (calibration slope = 0.77, intercept = 0.03). The likelihood of shortness of breath (OR = 7.91, 95% CI = 1.63-45.12) was significantly higher among mid- to late-stage PD than early-stage. Age and total salivary protein were also significantly higher among mid- to late-stage PD. The MLR prognostic model for mid- to late-stage PD may assist physicians in identifying patients at high risk for loss of balance and falls.