Objective: To identify subclinical changes of balance and gait in patients with X-linked dystonia-parkinsonism (XDP) and non-manifesting carriers of the XDP-causing mutation.

Background: XDP is a hereditary neurodegenerative disorder indigenous to the Philippines and caused by a founder SINE-VNTR-Alu (SVA)-type retrotransposon insertion in the TAF1 gene. The disorder is characterized by rapidly progressive adult-onset dystonia in the majority of cases. At the time of the clinical diagnosis, up to 40% of striatal volume is already degenerated. This implies that subtle motor abnormalities could precede the onset of XDP and serve as a biomarker for future interventional trials.

Method: We investigated 13 male non-manifesting SVA insertion carriers (NMC), 17 XDP patients (XDP), and 24 matched controls negative for the disease-causing mutation (HC). Standardized video-taped neurological examinations were performed by two movement disorder specialists. Sensor-based posturography and gait analyses were carried out under different conditions with increasing difficulty. Participants and investigators were kept blinded to the genetic status. Motor features were extracted from raw sensor data followed by an application of the Mobility Lab software routine. Classification analyses were performed within Rapid-Miner Studio version 9.8. We employed gradient-boosted trees (GBT) methodology to classify groups of interest.

Results: BFMDRS and MDS-UPDRS-III did not differ between NMCs and HCs (all p ≥0.579), and NMCs could not be identified on clinical grounds. The best GBT-based model on posturography measurements showed a classification accuracy of 90% in the comparison of NMC vs. HC. The exploratory analysis of feature weights showed that sway area, path length and ellipse axis contributed to the classification for the condition ‘feet apart with eyes closed on a foam surface’. In patients, path length, mean velocity of sway, sway area, frequency dispersion, and ellipse axis of sway correlated with MDS-UPDRS-III (all p≤0.031; all r≥0.312) and BFMDRS scores (all p≤0.032; all r≥0.316).For the gait analysis, the best-performing GBT-based model showed a balanced accuracy of 95% (NMC vs. HC; walking with maximum speed).

Conclusion: Our findings support the hypothesis of previously unidentified, prodromal motor changes among non-manifesting TAF1 mutation carriers who will develop XDP with a very high likelihood in the future.

« Back to MDS Virtual Congress 2021

MDS Abstracts - https://www.mdsabstracts.org/abstract/wearable-sensors-are-able-to-identify-individuals-in-the-prodromal-phase-of-x-linked-dystonia-parkinsonism/