Clinical Progression in Alpha‐Synuclein Positive LRRK2‐PD and Sporadic Parkinson's Disease: A Longitudinal Analysis

Abstract

Background

LRRK2‐Parkinson's disease (LRRK2‐PD) is biologically heterogeneous with approximately 30% lacking aggregated alpha synuclein (αSyn) in cerebrospinal fluid by seed amplification assay (SAA). Prior work has suggested slower progression in LRRK2‐PD compared to sporadic PD (sPD).

Objective

We aimed to assess how LRRK2‐PD with αSyn aggregates on SAA (S+ LRRK2‐PD) compares to S+ sPD.

Methods

Data from the Parkinson's Progression Markers Initiative were used to compare S+ LRRK2‐PD and S+ sPD cohorts propensity score‐matched on age, disease duration, sex and levodopa equivalent dose (N = 79 per cohort). Baseline clinical and biological features and 4‐year longitudinal features were assessed.

Results

At baseline, S+ LRRK2‐PD participants had lower motor scores and dopaminergic deficit. Among measures showing within group progression, longitudinal trajectories did not differ significantly between groups.

Conclusions

Longitudinal clinical progression of S+ LRRK2‐PD and sPD in the PPMI study is similar despite differences in baseline features.

The majority of individuals with sporadic Parkinson's Disease (sPD) have evidence of misfolded and aggregated alpha‐synuclein (αSyn) on cerebrospinal fluid (CSF) αSyn seed amplification assay (SAA) (S+). ¹ , ² By contrast, approximately one third of participants with LRRK2‐associated Parkinson's disease (LRRK2‐PD) are negative for αSyn on SAA (S‐), consistent with neuropathological data demonstrating absence of Lewy bodies in 25–60% of LRRK2‐PD brain tissue. ² , ³ , ⁴ , ⁵ , ⁶ , ⁷

Prior reports suggest that LRRK2‐PD is associated with less non‐motor disease burden and slower disease progression as compared to sPD. ⁸ , ⁹ , ¹⁰ Importantly, recent data indicate that there are differences in clinical measures between LRRK2‐PD S+ versus S‐ individuals: at baseline, S‐ LRRK2‐PD participants have lower Movement Disorder Society Unified Parkinson's Disease Rating Scale (MDS‐UPDRS) Part II and III scores as compared to S+ LRRK2‐PD, lesser degree of dopaminergic deficit on imaging, and have a trend towards slower motor and functional decline over time. ⁷ , ¹¹ , ¹²

Considering these findings, we aimed to compare clinical and biomarker characteristics between S+ sPD and S+ LRRK2‐PD in the Parkinson's Progression Markers Initiative (PPMI) study. We hypothesized that baseline characteristics and longitudinal progression would be similar in both groups, consistent with shared underlying proteinopathy.

Methods

The PPMI study protocol, and methods for assessment of CSF biomarkers including αSyn SAA have been published previously. ¹ , ¹³ , ¹⁴ , ¹⁵ The PPMI sample included 165 participants with LRRK2‐PD (carriers of LRRK2 G2019S, R1441C/G/H, N1437H, and I2020T as confirmed by the PPMI Genetic Coordination Core), and 950 participants with sPD for whom genotyping and αSyn SAA status were available. Exclusion criteria included S‐ status, MSA type‐ SAA, and inconclusive, and absence of at least one visit with levodopa equivalent dose (LED) > 0. Individuals with presence of a non‐LRRK2 pathogenic variant were also excluded. Participants with fewer than one year of follow up for LRRK2‐PD or fewer than three years of follow up for sPD were further excluded, initially yielding 96 LRRK2‐PD and 284 sPD participants. Enrollment criteria for the PPMI genetic cohort were different from sPD cohort allowing for longer disease duration (up to 7 years) and use of dopaminergic therapy. ¹⁴ , ¹⁶ , ¹⁷ To account for differences at enrollment between the two groups, a propensity score matching approach was employed where sPD participants were matched 1:1 to LRRK2‐PD participants based on age, sex, disease duration and LED. Participants were aligned on their first treated visit—defined as “baseline” for the purposes of this analysis. The matching process utilized a caliper width of 0.20 times the standard deviation of the logit of the propensity score with a sequential greedy nearest neighbor matching approach.

Differences in baseline characteristics between groups were assessed using the Wilcoxon rank sum test for continuous variables, and Chi‐Square (or Fisher's exact test when at least one expected cell count <5) for categorical measures. All baseline and longitudinal characteristics included in the analysis are presented in the results section and summarized in the accompanying tables.

Longitudinal analysis included clinical measures assessed at each annual visit. Medication OFF Part 3 scores were missing on a proportion of participants, and were therefore not included. Generalized linear mixed effects (LMM) models with random intercept and slope and unstructured working correlation structure were employed to model the linear trajectory of continuous outcomes by subgroup (sPD or LRRK2‐PD) over 4 years, with adjustment for baseline value of the outcome and time‐varying LED. For binary outcomes, only random intercepts were included, with the linear trajectory of the log odds being modeled. To assess the possibility of non‐linear trajectories and differences by sex, tests for differential quadratic and group‐sex‐time effects were also conducted. Interaction and time effects from the models and associated Wald P‐values are reported. Given the exploratory nature of this study, the significance threshold was set to 0.05 (two tailed).

Statistical analyses were performed using SAS v9.4 (SAS Institute Inc., Cary, NC, USA; sas.com; RRID:SCR_008567).

Results

Baseline Characteristics

79 participants were included in the analysis per group after propensity matching. Participant selection and reasons for exclusion are presented in Figure S1. Demographic and disease characteristics of the cohort prior to matching are presented in Tables S1 and S2. Table 1 shows baseline demographic, clinical and biological characteristics of the matched groups. The majority of LRRK2‐PD participants (96%) were G2019S carriers. At baseline, the S+ LRRK2‐PD group had milder degree of functional impairment and motor dysfunction as indicated by higher scores on the Modified Schwab and England scale (median [IQR] 90 [90–100] versus 90 [80–90], P < 0.001), lower MDS‐UPDRS part III ON scores (median [IQR] 16 [10–24] versus 21 [15–31], P = 0.011), lower Tremor score (ON) (median [IQR] 2 [0–5] versus 4 [1–6], P = 0.022), and lower Total MDS‐UPDRS scores (median [IQR] 29 [20–42] versus 34 [27–51], P = 0.027) as compared to sPD. A lesser degree of dopaminergic dysfunction as characterized by lowest putamen ratio was observed in LRRK2‐PD participants compared to sPD (median [IQR] 0.27 [0.24–0.37] versus 0.25 [0.20–0.31], P = 0.016). There were no differences between groups in fluid biomarkers including measures of CSF amyloid beta 1–42 (Aβ_1‐42), total tau, phospho‐tau 181, or serum neurofilament light chain (NfL).

TABLE 1.

Demographics, clinical and biological characteristics of S+ LRRK2‐PD and S+ sporadic PD participants

Variable	LRRK2‐PD (N = 79)	Sporadic PD (N = 79)	P‐value a
Demographics
Age at baseline, years, median (IQR)	62.9 [55.8–68.3]	61.3 [53.0–69.4]	0.707
Age at PD symptom onset, years, median (IQR)	58.1 [51.0–63.5]	58.8 [49.0–64.7]	0.977
Male sex, n (%)	54 (68%)	54 (68%)	1.000
Years of education, median (IQR)	17.0 [15.0–19.0]	16.0 [14.0–18.0]	0.017
Years since PD diagnosis, median (IQR)	1.9 [1.1–3.2]	2.3 [1.4–3.2]	0.317
Race (% White), n (%)	75 (95%)	75 (95%)	1.000
Hispanic, n (%)	11 (14%)	2 (3%)	0.009
LED, median (IQR)	405.0 [260.0–550.0]	400.0 [300.0–581.8]	0.658
LRRK2 variant, n (%)
G2019S	76 (96%)	‐
R1441G	3 (4%)	‐
APOE genotype—number of e4 alleles b , n (%)			0.328
0	65 (82%)	60 (76%)
1	13 (16%)	17 (22%)
2	1 (1%)	2 (3%)
Clinical characteristics
UPSIT percentile c , median (IQR)	7.0 [3.0–19.0]	5.5 [2.0–12.0]	0.069
Hyposmic (UPSIT percentile ≤ 15) c , n (%)	54 (72%)	65 (83%)	0.092
Modified Schwab and England, median (IQR)	90 [90–100]	90 [80–90]	<.001
Hoehn & Yahr stage (>2) (ON), n (%)	0 (0%)	1 (1%)	1.000
MDS‐UPDRS I, median (IQR)	5 [3–9]	6 [4–10]	0.466
MDS‐UPDRS II, median (IQR)	6 [3–9]	8 [4–10]	0.055
MDS‐UPDRS III (OFF), median (IQR)	23 [13–30]	26 [17–36]	0.082
MDS‐UPDRS III (ON), median (IQR)	16 [10–24]	21 [15–31]	0.011
% Change in MDS‐UPDRS Part III (OFF to ON), median (IQR)	−15.1 [−42.1–0.0]	−16.7 [−25.0–0.0]	0.222
Gait (item 3.10) (ON) > 0, n (%)	45 (60%)	51 (68%)	0.307
Freezing of gait (item 3.11) (ON) > 0, n (%)	3 (4%)	2 (3%)	1.000
Tremor score (ON), median (IQR)	2 [0–5]	4 [1–6]	0.022
Total MDS‐UPDRS (ON), median (IQR)	29 [20–42]	34 [27–51]	0.027
TD/PIGD classification (ON), n (%)			0.016
TD	34 (47%)	51 (70%)
PIGD	28 (38%)	15 (21%)
Indeterminate	11 (15%)	7 (10%)
Geriatric depression scale, median (IQR)	2 [1–4]	2 [0–3]	0.259
State–trait anxiety inventory, median (IQR)	63 [52–80]	60 [49–77]	0.266
SCOPA‐AUT, median (IQR)	10 [6–15]	10 [6–17]	0.750
RBDSQ, median (IQR)	3 [2–5]	4 [2–6]	0.186
RBDSQ ≥6, n (%)	14 (18%)	25 (32%)	0.042
Epworth Sleepiness Scale, median (IQR)	6 [4–9]	7 [5–10]	0.323
MoCA, median (IQR)	27 [25–29]	27 [25–29]	0.332
Hopkins Verbal Learning Test Delayed Recall t‐score, median (IQR)	50.5 [38.0–55.0]	48.5 [37.5–54.5]	0.549
Hopkins Verbal Learning Test Immediate/Total Recall t‐score, median (IQR)	48.0 [40.0–54.0]	44.0 [39.0–49.5]	0.079
Benton Judgment of Line Orientation scaled score, median (IQR)	12.3 [10.6–13.8]	12.6 [10.7–14.2]	0.365
Letter Number Sequencing scaled score, median (IQR)	11.0 [10.0–13.0]	11.0 [10.0–13.0]	0.944
Symbol Digit Modalities Test t‐score, median (IQR)	46.6 [40.0–51.0]	46.3 [39.4–51.0]	0.960
Semantic fluency (animals) t‐score, median (IQR)	52.0 [46.0–57.0]	49.0 [46.5–55.0]	0.307
Cognitive summary score, median (IQR)	0.1 [−0.4–0.4]	−0.0 [−0.4–0.4]	0.569
Number of impulse control disorders b , n (%)			0.080
0	51 (65%)	61 (77%)
1	21 (27%)	13 (16%)
≥ 2	7 (9%)	5 (6%)
Biological characteristics
Lowest putamen ratio, median (IQR)	0.27 [0.24–0.37]	0.25 [0.20–0.31]	0.016
Aβ1‐42 (pg/mL), median (IQR)	793.4 [639.9–1030.7]	814.8 [622.3–1137.0]	0.770
Aβ1‐42 ≤ 683 pg/mL, n (%)	22 (31%)	20 (31%)	0.978
Aβ1‐42 ≤ 710 pg/mL, n (%)	26 (37%)	23 (35%)	0.881
Total tau (pg/mL), median (IQR)	140.1 [118.6–193.6]	155.6 [122.7–210.2]	0.227
Total tau ≥ 266 pg/mL, n (%)	4 (6%)	4 (6%)	1.000
Total tau ≥ 112 pg/mL, n (%)	60 (85%)	53 (82%)	0.645
Phospho‐tau181 (pg/mL), median (IQR)	12.3 [10.0–15.5]	12.5 [9.8–17.3]	0.565
Phospho‐tau181 ≥ 24 pg/mL, n (%)	4 (6%)	6 (9%)	0.519
Phospho‐tau181 ≥ 17.6 pg/mL, n (%)	12 (17%)	15 (23%)	0.367
Serum NfL (pg/mL), median (IQR)	10.9 [8.9–14.7]	12.3 [9.6–16.5]	0.156

Note: Baseline demographic, clinical and biological characteristics are summarized for S+ LRRK2 PD and S+ sporadic PD groups, using median (interquartile range, IQR) for continuous measures and frequency (percentage) for categorical measures.

Missing (if more than 5): Hoehn & Yahr stage (>2) (ON) n = 8; MDS‐UPDRS III (OFF) n = 32; MDS‐UPDRS III (ON) n = 11; % Change in MDS‐UPDRS Part III (OFF to ON) n = 40; Gait (item 3.10) (ON) > 0 n = 8; Freezing of gait (item 3.11) (ON) > 0 n = 8; Tremor score (ON) n = 9; Total MDS‐UPDRS (ON) n = 14; TD/PIGD Classification (ON) n = 12; Benton Judgment of Line Orientation scaled score n = 6; Cognitive Summary Score n = 6; Lowest putamen ratio n = 30; Aβ_1‐42, total tau and phospho‐tau₁₈₁ n = 22; Serum NfL n = 36.

Abbreviations: Aβ_1‐42, amyloid beta_1‐42; APOE, apolipoprotein E; LED, levodopa equivalent dose; LRRK2‐PD, alpha‐synuclein positive LRRK2‐Parkinson's disease; MDS‐UPDRS, Movement Disorder Society‐Modified Unified Parkinson's Disease Rating Scale; MoCA, Montreal Cognitive Assessment; NfL, neurofilament light chain; PIGD, postural instability and gait difficulty; RBDSQ, REM Sleep Behavior Disorder Screening Questionnaire; SCOPA‐AUT, Scales for Outcomes in Parkinson's disease—Autonomic Dysfunction; sPD, alpha‐synuclein positive sporadic Parkinson's disease; TD, tremor dominant; UPSIT, University of Pennsylvania Smell Identification Test.

^a Comparisons by group used Chi‐Square or Fisher's Exact tests for categorical variables and Wilcoxon rank sum tests for continuous variables.

^b For the purposes of comparisons, APOE genotype was dichotomized as 0 vs. ≥ 1 e4 alleles, and number of ICDs was dichotomized as 0 vs. ≥ 1 ICDs.

^c UPSIT percentile is carried forward from last available visit.

Longitudinal Analysis

Raw mean and median values of clinical features and available biomarkers at each follow‐up visit are detailed in Table S3. Results of linear mixed effects models are shown in Table S4. Progression slopes are presented in Figure 1. None of the clinical motor or non‐motor measures showed differences in progression between the two groups aside from a marginal difference in the rate of change in the tremor score (ON) (P = 0.049), although neither LRRK2 nor PD groups showed significant change over time (estimate [95% CI] −0.23 [−0.52, 0.06], P = 0.119; 0.16 [−0.12, 0.44], P = 0.257 respectively).

Figure 1.

Predicted longitudinal trajectories (and 95% confidence bands) of selected clinical outcomes over four year follow up. Significance of linear mixed effects models indicated by overall time p‐value and interaction (time × group) P‐value, α = 0.05. CSS, Cognitive Summary Score; GDS, Geriatric Depression Scale; LRRK2 PD, alpha‐synuclein positive LRRK2‐Parkinson's disease; MDS‐UPDRS, Movement Disorder Society—Unified Parkinson's Disease Rating Scale; MoCA, Montreal Cognitive Assessment; Modified S&E, Modified Schwab and England Activities of Daily Living scale; RBDSQ, REM Sleep Behavior Disorder Screening Questionnaire; SCOPA‐AUT, Scales for Outcomes in Parkinson's disease—Autonomic Dysfunction; sPD, alpha‐synuclein positive sporadic Parkinson's disease; STAI, State–Trait Anxiety Inventory.

Discussion

We report an analysis of S+ LRRK2‐PD (predominantly G2019S) compared to S+ sPD with cohorts propensity matched by age, sex, disease duration and LED. Our data indicate that while S+ LRRK2‐PD is characterized by milder motor clinical features and less DAT deficit at baseline, longitudinal trajectories do not differ significantly between the groups. These results, while requiring validation in other cohorts, have implications on our understanding of interaction between LRRK2 and αSyn‐driven pathobiology in PD.

Historically, LRRK2‐PD has been associated with milder phenotype and fewer non‐motor features; however, these observations were made prior to the availability of in vivo biomarkers of αSyn pathology. ⁸ , ⁹ LRRK2‐PD is heterogeneous regarding the underlying proteinopathy even within same families, and in some cases is associated with pure nigrostriatal degeneration in absence of any protein aggregation. ³ , ⁴ Through this unique heterogeneity, LRRK2‐PD serves as a valuable model through which the neurobiological underpinnings of clinical progression in PD—and in particular the role of αSyn—can be examined.

Our recently published data demonstrated that S‐ compared to S+ LRRK2‐PD individuals have milder baseline phenotype and slower functional decline. ⁷ The most striking differences included significant female sex predominance and lower rate of hyposmia. ⁷ We did not elicit other biomarker separation between the groups aside from SAA status. These results suggest that previously reported milder LRRK2‐PD phenotype may be driven by the S‐ subset of LRRK2‐PD. The current analysis supports that hypothesis.

The whole discussion regarding the S‐ LRRK2‐PD subset must be interpreted with caution considering recent reports of presence of oligomeric αSyn in Lewy body negative postmortem brains using proximity ligation assays. ¹⁸ , ¹⁹ If confirmed, the debate will refocus on understanding the implications not of binary αSyn status but of the protein conformation changes and potentially “the load” of pathology for which a quantitative in vivo marker is essential.

Several limitations to this study should be highlighted, particularly regarding sample selection and genetic and ethnic features of the sample.

We used propensity score matching to establish groups matched for age, disease duration, sex, and LED. This approach was taken in an effort to address differences in baseline characteristics of cohorts specific to the PPMI dataset due to known differences in inclusion criteria for participants with sporadic versus genetic PD. ¹⁶ , ¹⁷ While it is recognized that matching procedures reduce the available sample size and can introduce bias, an approach without matching presented risk of comparing groups with differences reflective of recruitment as opposed to true neurobiological difference. Analyses of additional cohorts with a uniform recruitment strategy will be important in confirming our findings. We also acknowledge limitation in interpretation of motor score results using MDS‐UPDRS III ON scores. While statistical analysis adjusted for LED, this similarly underscores the importance of analysis of other cohorts with diverse recruitment and data collection strategies.

Our analysis could be strengthened by comparison between S‐ LRRK2 PD and S+ sPD groups using the same methodology applied in this analysis. The sample size of the S‐ LRRK2 PPMI cohort was not adequate to support an analysis with a valid uniform matching technique and therefore was deferred. This comparison could be considered in different cohorts or with future growth of the LRRK2‐PD PPMI cohort.

LRRK2‐PD is genetically heterogeneous, with multiple associated causal variants including G2019S and R1441C/G along with other less frequently observed variants. ⁷ , ¹⁰ , ²⁰ Over 95% of S+ LRRK2 PPMI participants are p.G2019S carriers of European ancestry, while the S‐ group had greater diversity in variants, including R1441G among others. ⁷ Given this difference, care must be taken in extrapolating the findings presented here acknowledging they reflect primarily features of G2019S LRRK2‐PD and the effect of SAA status on progression of other variant carriers remains to be examined. This is of particular relevance given recent reports that the differences in motor decline between S+ and S‐ LRRK2‐PD cohorts may be driven primarily by R1441C/G + M1646T carriers, a group that was not represented in the present analysis. ¹² We also acknowledge a difference in Hispanic ethnicity between groups which may introduce an additional source of potential bias. Further analysis of other LRRK2 variants in ethnically diverse cohorts is warranted.

Conclusion

Although there are baseline features of LRRK2‐PD that remain distinct from sPD even among the subset of participants with evidence of αSyn aggregation, the results of this analysis suggest that S+ LRRK2‐PD behaves similarly to sPD in terms of clinical progression. This finding has important implications in ongoing efforts to understand biological drivers of PD progression. Furthermore, we emphasize that these findings make an argument for assessment of αSyn biomarkers in clinical trial design, enrollment, and analysis, including genetic cohorts with known proteinopathy heterogeneity like LRRK2‐PD.

Disclosures

Ethical Compliance Statement: This study did not require review by institutional review board or ethics committee. Data used are publicly available as stated in the data availability statement. Each PPMI recruitment site received approval from an institutional review board or ethics committee on human experimentation before study initiation. Written informed consent for research was obtained from all individuals participating in the PPMI study. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Financial Disclosures of all Authors (for the Preceding 12 Months): L.M has received travel expense reimbursement from the Parkinson Study Group. S.H.C. is funded by grants from the Michael J. Fox Foundation for Parkinson's Research. D.E.L. is funded by grants from the Michael J. Fox Foundation for Parkinson's Research. L.M.C. declares consulting fees and research support from the Michael J. Fox Foundation. K.M.M. declares consultancies for Axial Therapeutics, Asceneuron, Calico, Hanall, JnJ, Michael J. Fox Foundation, Nitrase Therapeutics, NuraBio, NRG Therapeutics, Rome Therapeutics, Schrodinger, Ventyx, private equity companies. Serves on the BoD for Envisagenics, Retromer Tx; serves on the SAB for Axial, Nitrase, NRG, Sinopia, Vanqua; received research funding support from Michael J. Fox Foundation and Honoraria from ASAP. T.S. declares consultancies for AcureX, Adamas, AskBio, Amneal, Blue Rock Therapeutics, Critical Path for Parkinson's Consortium, Denali, The Michael J. Fox Foundation, Neuroderm, Roche, Sanofi, Sinopia, Takeda, and Vanqua Bio; on advisory boards for AcureX, Adamas, AskBio, Biohaven, Denali, GAIN, Neuron23 and Roche; on scientific advisory boards for Koneksa, Neuroderm, Sanofi and UCB; and received research funding from Amneal, Biogen, Roche, Neuroderm, Sanofi, Prevail and UCB and an investigator for NINDS, MJFF, Parkinson's Foundation. B.W. is an employee of Berry Consultants, LLC, in which capacity she serves as a consultant to numerous pharmaceutical and device companies on topics of statistical modeling and trial design. The identities of these clients are protected under non‐disclosure agreements. All payments are made to Berry Consultants, LLC.

Authors Roles

(1) Research Project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript Preparation: A. Writing of the First Draft, B. Review and Critique.

L.M.: 1B, 1C, 2A, 2C, 3A, 3B

S.H.C.: 1B, 1C, 2A, 2B, 2C, 3B

D.E.L.: 1B, 1C, 2A, 2B, 2C, 3B

C.G.: 1B, 1C, 2A, 2B, 2C, 3B

L.M.C.: 1A, 1C, 2A, 2C, 3B

K.M.M.: 1A, 2C, 3B

B.W.: 2C, 3B

T.S.: 1A, 1B, 1C, 2A, 2C, 3B

Financial Disclosures and Conflicts of Interest

Author disclosures are available in the Supporting Information.

Supporting information

Figure S1. Flowchart of participant selection for analysis population. LED, levodopa equivalent daily dose; MSA, multiple system atrophy; PD, Parkinson's disease; S−, alpha‐synuclein negative; S+, alpha‐synuclein positive; sPD, sporadic PD.

TABLE S1. LED, levodopa equivalent dose (mg); PD, Parkinson's disease; S+; alpha‐synuclein positive. Baseline was defined to be a participant's first visit on treatment; LRRK2 and sporadic PD participants were required to have at least 1 and 3 years of follow‐up post‐treatment initiation, respectively, to be considered for matching.

^aComparisons by group used Chi‐Square or Fisher's Exact tests for categorical variables and Wilcoxon rank sum tests for continuous variables.

TABLE S2. LED, levodopa equivalent dose; PD, Parkinson's disease; S+; alpha‐synuclein positive. ^aComparisons by group used Chi‐Square or Fisher's Exact tests for categorical variables and Wilcoxon rank sum tests for continuous variables. ^bFor the purposes of comparisons, APOE genotype was dichotomized as 0 vs. ≥ 1 e4 alleles.

TABLE S3. Aβ_1‐42, amyloid beta_1‐42; MDS‐UPDRS, Movement Disorders Society—Modified Unified Parkinson's Disease Rating Scale; MoCA, Montreal Cognitive Assessment; NfL, neurofilament light chain; PD, Parkinson's disease; RBDSQ, REM Sleep Behavior Disorder Screening Questionnaire; S+; alpha‐synuclein positive; SCOPA‐AUT, Scales for Outcomes in Parkinson's disease–Autonomic Dysfunction. In cases where variables are missing less than 10% of values at all visits, the missing values are not shown.

TABLE S4. MDS‐UPDRS, Movement Disorders Society—Modified Unified Parkinson's Disease Rating Scale; MoCA, Montreal Cognitive Assessment; PD, Parkinson's disease; RBDSQ, REM Sleep Behavior Disorder Screening Questionnaire; S+; alpha‐synuclein positive; SCOPA‐AUT, Scales for Outcomes in Parkinson's disease–Autonomic Dysfunction. Models were adjusted for levodopa equivalent dose at each visit. In cases where the model did not converge, the relevant fields are left blank. ^aFor binary outcomes, the time effect is reported as odds ratio (95% CI for OR).

Data S1. COI_disclosure.

Appendix A. PPMI Study Committees, Cores, and Collaborators

A.1. PPMI Executive Steering Committee

Kenneth Marek, MD¹ (Principal Investigator); Tanya Simuni, MD²; Andrew Siderowf, MD³; Caroline Tanner, MD⁴; Thomas F. Tropea, DO¹; Tatiana Foroud, PhD⁵; Lana Chahine, MD⁶; Brit Mollenhauer, MD⁷; Kalpana Merchant, MD²; Douglas Galasko, MD⁸; Christopher Coffey, PhD ⁹; Kathleen Poston, MD¹⁰; Roseanne Dobkin, PhD¹¹; Ethan Brown, MD⁴; Roy Alcalay, MD¹²; Dan Weintraub, MD³; Emily Flagg, BA¹; Kimberly Fabrizio, BA¹.

A.2. PPMI Steering Committee

Susan Bressman, MD¹³; Cornelis Blauwendraat, PhD¹⁴; Paola Casalin, PhD¹⁵; Sonya Dumanis, PhD¹⁴; Raymond James, RN¹⁶; Karl Kieburtz, MD¹⁷; Sneha Mantri, MS¹⁸; Werner Poewe, MD¹⁹; Michael Schwarzschild, MD²⁰; John Seibyl¹, MD; David Standaert, PhD²¹; Duygu Tosun‐Turgut, PhD⁴.

A.3. Michael J. Fox Foundation

Sohini Chowdhury, MA²²; Jamie Eberling, PhD²²; Mark Frasier, PhD²²; Leslie Kirsch, EdD²²; Katie Kopil, PhD²²; Maggie Kuhl, BA²²; Alyssa O'Grady, BA²²; Todd Sherer, PhD²²; Tawny Willson, MBS²².

A.4. PPMI Study Cores

Project Management Core: Emily Flagg, BA¹.

Site Management Core: Tanya Simuni, MD²; Bridget McMahon, BS¹.

Data Strategy and Technical Operations: Craig Stanley, PhD¹; Kim Fabrizio, BA¹.

Data Management Core: Dixie Ecklund, MBA⁹, MSN; Christine Kohnen, PhD⁹.

Screening Core: Tatiana Foroud, PhD⁵; Laura Heathers, BA⁵; Christopher Hobbick, BSCE⁵; Gena Antonopoulos, BSN⁵.

Imaging Core: John Seibyl, MD¹; Kathleen Poston, MD¹⁰.

Statistics Core: Christopher Coffey, PhD ⁹; Chelsea Caspell‐Garcia, MS ⁹; Michael Brumm, MS ⁹.

Bioinformatics Core: Arthur Toga, PhD²³; Karen Crawford, MLIS²³.

Biorepository Core: Tatiana Foroud, PhD⁵; Jan Hamer, BS⁵.

Biologics Review Committee: Brit Mollenhauer, MD⁷; Doug Galasko, MD ⁸; Kalpana Merchant, MD².

Genetics Core: Andrew Singleton, PhD²⁴.

Pathology Core: Tatiana Foroud, PhD⁵; Dirk Keene, MD⁵.

Found: Caroline Tanner, MD⁴; Ethan Brown, MD⁴.

PPMI Online: Carlie Tanner, MD⁴; Ethan Brown, MD⁴; Lana Chahine, MD⁶; Roseann Dobkin, PhD¹¹; Monica Korell, MPH⁴.

A.5. PPMI Site Investigators

Neha Prakash, MD¹; Tanya Simuni, MD²; Nabila Dahodwala, MD³; Caroline Tanner, MD⁴; Lana Chahine MD⁶; Brit Mollenhauer, MD⁷; Sebastian Schade, MD⁷; Douglas Galasko, MD⁸; Anat Mirelman, PhD¹²; Roy Alcalay, MD¹²; Katherine Leaver, MD¹³; Marie Saint‐Hilaire, MD¹⁶; Ruth Schneider, MD¹⁷; Christopher Tarolli, MD¹⁷; Werner Poewe, MD¹⁹; Aleksandar Videnovic, MD²⁰; David Standaert, PhD²¹; Marissa Dean, MD²¹; Sonja Jonsdottir, PhD²⁵; Rejko Krueger, MD²⁵; Claire Pauly, PhD²⁵; Stewart Factor, DO²⁶; Penelope Hogarth, MD²⁶; Robert Hauser, MD²⁸; Amy Amara, PhD²⁹; Michelle Fullard, MD²⁹; Cyrus Zabetian, MD³⁰; Hubert Fernandez, MD³¹; Kathrin Brockmann, MD³²; Isabel Wurster, PhD³²; Yen Tai, PhD³³; Paolo Barone, PhD³⁴; Marina Picillo MD³⁴; Stuart Isaacson, MD³⁵; Alberto Espay, MD³⁶; Eduardo Tolosa, PhD³⁷; Javier Ruiz Martinez, PhD³⁸; Leonidas Stefanis, PhD³⁹; Kelvin Chou, MD⁴⁰; Lorraine Kalia, MD⁴¹; Connie Marras, PhD⁴¹; David Grimes, MD⁴²; Tiago Mestre, PhD⁴²; Rajesh Pahwa, MD⁴³; Mark Lew, MD⁴⁴; Holly Shill, MD⁴⁵; Shyamal Mehta, MD⁴⁶; Giulietta Riboldi, MD⁴⁷; Nikolaus McFarland, PhD⁴⁸; Ron Postuma, MD⁴⁹; Zoltan Mari, MD⁵⁰; David Ledingham, MD⁵¹; Nicola Pavese, PhD⁵¹; Michele Hu, PhD⁵²; Norbert Brueggemann, MD⁵³; Christine Klein, MD⁵³; Bastiaan Bloem, PhD⁵⁴; Cristina Simonet, PhD⁵⁵; Alastair Noyce, PhD⁵⁵; Anette Janzen, PhD⁵⁶; David Pedrosa, MD⁵⁶; Wolfgang Oertel, PhD⁵⁶; Njideka Okubadejo, MD⁵⁷ David Shprecher, DO⁵⁸; Arjun Tarakad, MD⁵⁹; Emile Moukheiber, MD⁶⁰.

A.6. PPMI Site Coordinators

Joy Antala¹; Carla Aranda²; Karen Williams²; Sophia Melton²; Karina Benson²; Ashwini Ramachandran³, Danielle Potts³; Grace LaMoure³; Ritikha Vengadesh³; Ryan Manzler³; Jaime Heller⁴; Primi Ranola⁴; Farah Kausar⁴; Sherri Mosovsky⁶; Diana Willeke⁷; Elizabeth Kalinkara Gomez⁷; Janelle Rodriguez⁸; Nobuko Kemmotsu⁸; May Eshel¹²; Deborah Raymond¹³; Abigail Desrosiers¹⁶; Raymond James¹⁶; Lauren Jackson¹⁷; Iris Egner¹⁹; Wesley Schlett²⁰; Courtney Blair²¹; Lauren Ruffrage²¹; Berenice Sevilla²⁵; Barbara Sommerfeld²⁶; Dustin Le²⁷; Erica Botting²⁸; Gabriella Mazur²⁸; Daniele Derlein²⁹; Evan Doll²⁹; Ying Liu²⁹; Ciera Cobb³⁰; Olivia Masiewicz³⁰; Jennifer Mule³¹; Michael Morsillo³¹;

Ella Hilt³²; Aldazier Jakiran³³; Dominga Valentino³⁴; Lisbeth Pennente³⁵; Bobbie Stubbeman³⁶; Alicia Garrido³⁷; Valeria Ravasi³⁷; Ioana Croitoru³⁸; Christos Koros³⁹; Nikolas Papagiannakis³⁹; Frank Ferrari⁴⁰; Mengyu Zheng⁴¹; Shawna Reddie⁴²; Alicia Alejandra⁴³; Andrea Gray⁴³; Alejandra Valenzuela⁴⁴; Caitlin Goodman⁴⁵; Sara Dresler⁴⁶; Neil Santos46; Fahrial Esha⁴⁷; Kyle Rizer⁴⁸; Nadine Zablith⁴⁹; Liliana. Dumitrescu⁵⁰; Debra Galley⁵¹; Victoria Kate Foster⁵¹; Jamil Razzaque⁵²; Madita Grümmer⁵³; Yara Krasowski⁵⁴; Natalie Donkor⁵⁵; Elisabeth Sittig⁵⁶; Oluwadamilola Ojo⁵⁷; Kelly Clark⁵⁸; Rory Mahabir⁵⁹; Kori Ribb⁶⁰; Shamera Willoughby⁶⁰.

Institutions and affiliations:

1. Institute for Neurodegenerative Disorders; New Haven, CT, USA

2. Northwestern University; Evanston, IL, USA

3. University of Pennsylvania; Philadelphia, PA, USA

4. University of California, San Francisco; San Francisco, CA, USA

5. Indiana University; Indianapolis, IN, USA

6. University of Pittsburgh; Pittsburgh, PA, USA

7. Paracelsus‐Elena Klinik; Kassel, Germany

8. University of California, San Diego, San Diego, CA, USA

9. University of Iowa; Iowa City, IA, USA

10. Stanford University; Stanford, CA, USA

11. Rutgers University; New Brunswick, NJ, USA

12. Tel Aviv Sourasky Medical Center; Tel Aviv, Israel

13. Mount Sinai Beth Israel; New York, NY, USA

14. Coalition for Aligning Science; Chevy Chase, MD, USA

15. BioRep; Milan, Italy

16. Boston University School of Medicine; Boston, MA, USA

17. University of Rochester; Rochester, NY, USA

18. Duke University; Durham, NC, USA

19. University of Innsbruck; Innsbruck, Austria

20. Massachusetts General Hospital; Boston, MA, USA

21. University of Alabama at Birmingham; Birmingham, AL, USA

22. The Michael J. Fox Foundation for Parkinson's Research; New York, NY, USA

23. Laboratory of Neuroimaging (LONI), USC; Los Angeles, CA, USA

24. National Institute on Aging, NIH; Bethesda, MD, USA

25. University of Luxembourg; Esch‐sur‐Alzette, Luxembourg

26. Emory University; Atlanta, GA, USA

27. Oregon Health and Science University; Portland, OR, USA

28. University of South Florida; Tampa, FL, USA

29. University of Colorado; Aurora, CO, USA

30. VA Puget Sound Health System; Seattle, WA, USA

31. Cleveland Clinic; Cleveland, OH, USA

32. University of Tubingen; Tubingen, Germany

33. Imperial College of London; London, UK

34. University of Salerno; Salerno, Italy

35. Parkinson's Disease and Movement Disorders Center; Boca Raton, FL, USA

36. University of Cincinnati; Cincinnati, OH, USA

37. Hospital Clinic of Barcelona; Barcelona, Spain

38. Hospital Universitario Donostia; San Sebastian, Spain

39. University of Athens; Athens, Greece

40. University of Michigan; Ann Arbor, MI, USA

41. Toronto Western Hospital; Toronto, Canada

42. The Ottawa Hospital; Ottawa, Canada

43. University of Kansas Medical Center; Kansas City, KS, USA

44. Keck School of Medicine of the University of Southern California; Los Angeles, CA, USA

45. Barrow Neurological Institute; Phoenix, AZ, USA

46. Mayo Clinic Arizona; Scottsdale, AZ, USA

47. NYU Langone Medical Center; New York, NY, USA

48. University of Florida; Gainesville, FL, USA

49. Montreal Neurological Institute and Hospital/McGill; Montreal, ǪC, Canada

50. Cleveland Clinic‐Las Vegas Lou Ruvo Center for Brain Health; Las Vegas, NV, USA

51. Clinical Aging Research Unit; Newcastle, UK

52. John Radcliffe Hospital Oxford and Oxford University; Oxford, UK

53. University of Luebeck; Luebeck, Germany

54. Radboud University; Nijmegen, Netherlands

55. Ǫueen Mary University of London; London, UK

56. Philipps‐University Marburg; Marburg, Germany

57. University of Lagos; Lagos, Nigeria

58. Banner Sun Health Research Institute; Sun City, AZ, USA

59. Baylor College of Medicine; Houston, TX, USA

60. Johns Hopkins University; Baltimore, MD, USA

Data Availability Statement

Data used in the preparation of this article were obtained on March 31, 2025 from the PPMI database (www.ppmi-info.org/access-data-specimens/download-data), RRID:SCR_006431. For up‐to‐date information on the study, visit www.ppmi-info.org. This analysis was conducted by the PPMI Statistics Core and used actual dates of activity for participants, a restricted data element not available to public users of PPMI data. Statistical analysis codes used to perform the analyses in this article are shared on Zenodo (10.5281/zenodo.16814373). Protocol information for PPMI Clinical—Establishing a Deeply Phenotyped PD Cohort AM 3.2. can be found on protocols.io or by following this link: https://doi.org/10.17504/protocols.io.n92ldmw6ol5b/v2.