Objective: To assess the discriminative power of the PREDIGT formula for incident Parkinson disease (PD).

Background: The inability to predict the incidence rate of PD in neurologically healthy adults limits its prevention and future therapy. We recently designed an incidence prediction model founded on the concept that PD pathogenesis is multifactorial. We postulated 5 factors to determine cumulative incidence rates: DNA variants (D); Exposure to environmental factors (E); Gene–environment interactions that initiate pathological responses (I); Gender (G); and Time (T) to encompass effects of ageing and latency of illness. The proposed formula for calculating the PD incidence rate (PR; %) in an individual is PR=(E+D+I)*G*T.

Method: We began to validate this mathematical model using enrollment data from two longitudinal, nested-control cohorts: (1) the multicentre Parkinson’s Progression Marker Initiative (PPMI; PD subjects, n=492; neurologically healthy controls (HC), n=241); and (2) the single-centre De Novo Parkinson’s Study (DeNoPa; PD, n=159; HC, n=110). Known risk and protective factors were selected based on published meta-analyses and assigned positive and negative values, respectively. The PREDIGT formula was applied to each subject based on the aggregate score in each category (E,D,I,G,T). Score distribution in both groups (PD vs. HC) and receiver operating characteristic curves were plotted, and the area under the curve (AUC) was calculated to evaluate the PREDIGT model’s discriminative power.

Results: Pooled results from the analyses of PPMI and DeNoPa data, based on variables recorded in both cohorts, achieved an AUC of 0.821 (sensitivity: 0.772; specificity: 0.737) to distinguish PD patients from HC. Individual analyses, using shared and cohort-specific variables that were available, generated an AUC of 0.845 (sensitivity: 0.760; specificity: 0.780) for PPMI and an AUC of 0.782 (sensitivity: 0.873; specificity: 0.564) in DeNoPa.

Conclusion: Our results suggest a promising, early validation of the original PREDIGT formula. Work is currently ongoing to interrogate additional cohorts (and new epidemiological data) to refine existing variables and their assigned values. We will also expand variables and calibrate them within the five risk categories. This, to test PREDIGT’s predictive power for incident PD.

« Back to 2019 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/validation-of-the-predigt-score-for-the-incidence-rate-of-parkinson-disease/