Comparison of Outcomes Between Patients with Different PD Subtypes in the AAN Axon Registry
Iris Chin1, Sarah Kranick2, Heather Moss3, Kathryn Sands1, Peng Jin1, Jeremy Peavey1, Craig Pfeifer1, Youssef Zarrouk1, Zhongdi Chu1, Michael Mbagwu1, Aracelis Torres1
1Verana Health, 2MultiCare Health System, 3Stanford University
Objective:
To compare outcomes between Parkinson’s Disease (PD) subtypes using measures extracted from clinical notes from the AAN Axon Registry, a neurology-specific patient registry of real-world electronic health record data.
Background:
Akinesia/bradykinesia and rigidity-dominant (AR) and tremor-dominant (TD) subtypes have been identified in PD, with the latter demonstrating a milder progression. Leveraging real-world data (RWD) to identify subtypes and distinctions in disease progression enables application of RWD to the study of pharmaceutical targets based on subtype.
Design/Methods:

15,000 PD patients were identified using structured ICD codes and PD related keywords in clinical notes from 2.6 million patients in the Axon Registry. All measures were curated from clinical notes with the exception of deep brain stimulation (DBS), which was curated from linked claims data.

PD subtype was determined by calculating the tremor to rigidity/bradykinesia severity ratio at the earliest encounter within one year of the patient’s first PD ICD code. Patients were classified as TD for ratios >=1 and AR for ratios <0.8 (Kang et al., 2005). 

Subtypes were compared by presence and time to first presence of postural instability, motor fluctuations, dyskinesia, and DBS using chi-square and Mann-Whitney U tests, respectively. 


Results:

1764 and 693 PD patients were identified as AR and TD, respectively. The AR subtype had a higher proportion of patients with postural instability (61% vs. 53%) and dyskinesia (11% vs. 7%), p-values < .01 and a shorter time to first encounter with dyskinesia (mean 17 vs. 24 months) and motor fluctuations (mean 10 vs. 13 months), p-values < .05. No differences in DBS were found between subtypes. 

Conclusions:

Using RWD, we were able to identify AR and TD subtypes and differences in progression profiles consistent with the previous literature, demonstrating that RWD can be leveraged for subtype studies and outcome differences within PD. 

10.1212/WNL.0000000000205563