Clinical and transcranial sonographic correlates of orthostatic hypotension in Parkinson's disease.

OBJECTIVE

Autonomic dysfunction is common across the course of Parkinson's disease (PD). We aimed to delineate clinical and transcranial sonography (TCS) correlates of orthostatic hypotension (OH) in PD and to develop machine-learning-based markers for PD-OH.

METHODS

We consecutively enrolled confirmed PD patients (n = 311) at the Second Affiliated Hospital of Soochow University from September 2017 to March 2023 and classified them by OH status using orthostatic blood pressure testing. Motor symptoms (UPDRS-III) and non-motor burden (NMSQ, FSS, HAMA, HAMD, PDQ-39) were assessed. TCS measured middle cerebral artery (MCA) flow velocities and third-ventricle width. Group differences were compared, and multivariable logistic regression evaluated factors associated with OH. Machine-learning models were trained and validated for OH prediction.

RESULTS

OH prevalence was 36.0 %. Compared with PD without OH, PD + OH showed higher levodopa equivalent daily dose, higher supine and lower standing blood pressures, and more frequent use of antiparkinsonian medications (notably levodopa and dopamine agonists). Cognitive domains differed, with lower orientation scores in PD + OH. In multivariable analysis, higher dopaminergic load (OR 1.001, 95 % CI 1.000-1.001) and lower orientation scores (OR 0.685, 95 % CI 0.534-0.849) independently associated with OH. TCS revealed lower right MCA flow velocity and wider third-ventricle width in PD + OH. A multivariable logistic regression model achieved stable discrimination (AUC 0.726 in training cohort and 0.716 in validation cohort), outperforming single clinical parameters.

CONCLUSION

OH is common in PD and relates to higher dopaminergic exposure and orientation impairment; TCS features provide complementary risk information. A parsimonious multivariable model offers moderate, stable predictive performance and may support clinical decisions. Future studies should prospectively validate these markers across centers and integrate comprehensive autonomic testing, wearable hemodynamics, and multimodal imaging to refine individualized prediction and guide targeted interventions.