Exploring Traditional Medicine Diagnostic Classification for Parkinson’s Disease Using Hierarchical Clustering

Abstract

Introduction: Personalized diagnosis and therapy for Parkinson’s disease (PD) are needed due to the clinical heterogeneity of PD. Syndrome differentiation (SD) in traditional medicine (TM) is a diagnostic method for customized therapy that comprehensively analyzes various symptoms and systemic syndromes. However, research identifying PD classification based on SD is limited. Methods: Ten electronic databases were systematically searched from inception to August 10, 2021. Clinical indicators, including 380 symptoms, 98 TM signs, and herbal medicine for PD diagnosed with SD, were extracted from 197 articles; frequency statistics on clinical indicators were conducted to classify several subtypes using hierarchical clustering. Results: Four distinct cluster groups were identified, each characterized by significant cluster-specific clinical indicators with 95% confidence intervals of distribution. Subtype 2 had the most severe progression, longest progressive duration, and highest association with greater late-stage PD-associated motor symptoms, including postural instability and gait disturbance. The action properties of the herbal formula and original SD presented in the data sources for subtype 2 were associated with Yin deficiency syndrome. Discussion/Conclusion: Hierarchical clustering analysis distinguished various symptoms and TM signs among patients with PD. These newly identified PD subtypes may optimize the diagnosis and treatment with TM and facilitate prognosis prediction. Our findings serve as a cornerstone for evidence-based guidelines for TM diagnosis and treatment.