Association of plasma biomarkers with cognition, cognitive decline, and daily function across and within neurodegenerative diseases: Results from the Ontario Neurodegenerative Disease Research Initiative

Abstract INTRODUCTION We investigated whether novel plasma biomarkers are associated with cognition, cognitive decline, and functional independence in activities of daily living across and within neurodegenerative diseases. METHODS Glial fibrillary acidic protein (GFAP), neurofilament light chain (NfL), phosphorylated tau (p‐tau)181 and amyloid beta (Aβ)42/40 were measured using ultra‐sensitive Simoa immunoassays in 44 healthy controls and 480 participants diagnosed with Alzheimer's disease/mild cognitive impairment (AD/MCI), Parkinson's disease (PD), frontotemporal dementia (FTD) spectrum disorders, or cerebrovascular disease (CVD). RESULTS GFAP, NfL, and/or p‐tau181 were elevated among all diseases compared to controls, and were broadly associated with worse baseline cognitive performance, greater cognitive decline, and/or lower functional independence. While GFAP, NfL, and p‐tau181 were highly predictive across diseases, p‐tau181 was more specific to the AD/MCI cohort. Sparse associations were found in the FTD and CVD cohorts and for Aβ42/40. DISCUSSION GFAP, NfL, and p‐tau181 are valuable predictors of cognition and function across common neurodegenerative diseases, and may be useful in specialized clinics and clinical trials.

greater cognitive decline, and/or lower functional independence.While GFAP, NfL, and p-tau181 were highly predictive across diseases, p-tau181 was more specific to the AD/MCI cohort.Sparse associations were found in the FTD and CVD cohorts and for Aβ 42/40 .DISCUSSION: GFAP, NfL, and p-tau181 are valuable predictors of cognition and function across common neurodegenerative diseases, and may be useful in specialized clinics and clinical trials.

K E Y W O R D S
activities of daily living, amyloid, amyloid beta, blood, blood-based, cognition, dementia, glial fibrillary acidic protein, longitudinal, neurofilament light chain, neuropsychiatric, phosphorylated tau, protein, tau, vascular

BACKGROUND
Cognitive impairment, dementia, and motor dysfunction are increasingly prevalent with pathological forms of aging causing an enormous burden on individuals and society. 1 Previously identified relevant and accurate biomarkers with the ability to help diagnose and to moni- for an overview).7][8][9][10][11] However, more research is needed to determine the extent to which these biomarkers relate to cross-sectional cognitive function, longitudinal cognitive decline, and functional independence in clinically diagnosed patients with neurodegenerative diseases associated with dementia and/or motor dysfunction.This is imperative to understand better the value of plasma biomarkers to predict disease severity before widespread implementation as prognostic tools in specialized clinics and as screening and monitoring tools in clinical trials.3][14][15] It is therefore critical that these biomarkers be characterized and related to baseline and longitudinal changes in cognition not only in AD, but also within other commonly contributing neurodegenerative diseases, such as Parkinson's disease (PD), frontotemporal dementia (FTD) spectrum disorders, and cerebrovascular disease (CVD), and across all these diseases as a whole.
We aimed here to investigate how plasma GFAP, NfL, p-tau181, and

Ontario Neurodegenerative Disease Research Initiative cohort
The characteristics and processes of the ONDRI cohort have been described previously with detailed inclusion and exclusion criteria. 16,17l participants included in this study (n = 480) from the ONDRI cohort had previously been diagnosed with AD or mild cognitive impairment due to AD (AD/MCI; n = 126), PD (n = 140), FTD (n = 53), or CVD (n = 161).The ONDRI amyotrophic lateral sclerosis (ALS) participants were not included in this study due to lack of cognitive impairment and follow-up visits occurring at shorter intervals from the rest of the cohorts due to the more rapid decline observed.Participants were recruited for the ONDRI study through clinicians at tertiary clinics, between July 2014 and March 2017, at 14 academic health science centers in six cities across Ontario, Canada.In total, 584 (excluding ALS) participants were recruited; some did not meet inclusion criteria after screening (n = 91, of which n = 9 were transferred to another diagnostic group), and a few withdrew consent (n = 22), leaving a total of 480 participants enrolled in the study.All participants were screened on general and disease-specific inclusion and exclusion criteria by academically practicing board-certified neurologists, geriatric psychiatrists, or geriatricians depending on clinical diagnostic group.
A board-certified neuroradiologist reviewed the magnetic resonance imaging (MRI) scans to exclude any incidental findings and to assist with identification of supportive imaging features (e.g., for specific diagnostic groups such as stroke and AD/MCI).MRI, blood work (e.g., to exclude contributing factors for cognitive impairment), and clinical history and exam were used to support the ONDRI diagnoses.For the CVD group, a stroke physician determined the severity of white matter hyperintensity burden based on the following checklist included on the Clinical Report Form: none, mild to moderate, or extensive.This checklist is adapted from the Fazekas/age-related white matter changes rating scales. 18,191][22][23][24][25][26][27] Final inclusion into a specific diagnostic group was achieved by consensus review among physician members of that team.
Briefly, the AD/MCI group included typical and atypical AD presentations, in addition to single and multi-domain amnestic MCI due to AD.
The PD group included both cognitively intact participants and those with cognitive impairment or dementia.All patients had presented initially with the typical motor syndrome and none were diagnosed as having dementia with Lewy bodies initially.Ultimately, the composition of the PD group contained patients classified as cognitively unimpaired (34.3%), with MCI (37.1%), and with dementia (18.6%), and others who did not meet criteria for this classification, being between categories (10.0%).This classification is based on the Movement Disorder Society Task Force Level II guidelines, using objective cognitive impairment on neuropsychological evaluation interpreted by a boardcertified neuropsychologist, subjective cognitive decline reported by the participant or study partner, and functional impairment in instrumental ADL (iADL) scale. 28The FTD group was composed of various sporadic FTD spectrum diagnoses: behavioral variant FTD (40.7%), progressive supranuclear palsy (29.6%), progressive non-fluent aphasia (14.8%), semantic dementia (9.3%), and corticobasal syndrome (5.6%).The CVD group included participants who had experienced a mild to moderate ischemic stroke event documented on MRI or computed tomography ≥ 3 months before enrolment and confirmed by radiologist, with or without cognitive impairment but with minimum Montreal Cognitive Assessment (MoCA) cut-off score of 18, and with any level of small vessel disease burden.Participants with history of cognitive impairment or dementia before the vascular event or with large cortical strokes were excluded.
All participants underwent comprehensive clinical and neuropsychological assessments at baseline and annually, in addition to genetic and plasma biomarker assessments at baseline.All participants were required to have a study partner who knew them well, interacted with them regularly, and were able to provide collateral information about their function (in most cases, spouses).A healthy control (HC) group (n = 44), studied using the same ONDRI protocols in addition to brain Aβ PET as part of the Brain-Eye Amyloid Memory (BEAM) study, 29   3. Future directions: This report is an important step for future studies investigating how these plasma biomarkers can be concretely implemented as prognostic tools in specialized clinics and as screening or monitoring tools in clinical trials.also included to be used as a reference for baseline plasma biomarker levels only and not for further analysis.The HC group consisted of cognitively normal individuals who were all amyloid negative on florbetapir PET.Assessment-specific and ONDRI-wide quality control procedures to ensure accurate data are described elsewhere. 30,31

Plasma and genetic biomarkers
Blood samples were drawn from all participants at baseline at the closest LifeLabs Medical Laboratory Services location to their residence, with standardized operating protocols for collection and storage (https://www.lifelabs.com/page-section/specimen-collectionhandling-section-hcp-requisitions-page/).Samples were collected and processed within 24 hours; they were shipped on ice pack (not frozen ONDRISeq 32,33 and used to determine APOE ε4 carrier status.Less than 3% of the ONDRI cohort had monogenic mutations in genes known to be drivers of neurodegenerative diseases. 34

Cognitive domains
All participants underwent the ONDRI neuropsychology protocol, 30 which provides a comprehensive assessment of cognition and behavior, at baseline and at two yearly follow-up visits.Screening was done for English comprehension, visual acuity, and auditory acuity, which may confound test performance.Performance on five cognitive domains (attention and working memory, executive function, language, memory, and visuospatial function) were evaluated with domain-specific composite scores adjusted for age, sex, and education.Detailed methods and the full list of neuropsychological tests by cognitive domain are presented in supporting information Section 1 and Tables S1   and S2.The composite scores were derived from 26 raw test scores from the comprehensive assessment, based on conventions in clinical neuropsychology 35 and consensus agreement among the ONDRI neuropsychologists.

Independence in activities of daily living
Lawton-Brody scales completed by study partners at baseline were used to measure the participant's ability to function independently across activities of daily living. 36The Physical Self Maintenance (bADL) scale includes feeding, dressing, grooming, ambulation, bathing, and toileting.The iADL scale includes telephone use, shopping, food preparation, housekeeping, laundering, use of transportation, managing medications, and financial management.Percentage scores reflecting functional independence, for which higher scores reflect greater independence, were computed for both scales.Study partners had the option to rate iADL items as not applicable, which reduced the maximum obtainable score used in iADL percentage scores calculations.

Sample characteristics
The sample characteristics for the different ONDRI disease groups are described in Table years of education, was considerably worse in most clinical measures, and had the most attrition of all groups.Overall, the cohort consisted of 67.1% men, with a lower proportion in the AD group (54.8%).Women had slightly higher GFAP levels than men across the cohort (t = 5.34, P < 0.001), with no differences for other plasma biomarkers.There were also no differences between men and women in cognitive function or independence in ADLs.The HC group had a slightly lower age range (51 to 77 years old) and a much lower proportion of men (24.4%) than the disease cohorts.
Overall, in reference to the HC group, the AD/MCI group had elevated mean levels of GFAP; the AD/MCI, FTD, and CVD groups had elevated mean levels of NfL; and the AD/MCI, PD, FTD, and CVD groups had elevated mean levels of p-tau181.
No groups were different from HC for levels of Aβ 42/40 .Differences in plasma biomarker levels between all disease groups are displayed in Table 1.Intercorrelations between plasma biomarkers levels are presented in supporting information Section 4 and Table S4.Total attrition in the entire sample was 10.6% at second visit, and

Pooled diseases regardless of given clinical diagnosis
Across all neurodegenerative and cerebrovascular disease participants pooled together, GFAP, NfL, and p-tau181 were found to be associated with most cognitive domains at baseline (Figure 2 S5 and S6. Regarding baseline independence in ADL (Figure 2), higher GFAP levels were significantly associated with greater impairments in iADL function (β = −0.16,P = 0.003), while higher NfL levels were significantly associated with greater impairments in both bADL (β = −0.12,P = 0.022) and iADL function (β = −0.21,P < 0.001).

FTD spectrum disorders
In the FTD group, while Aβ in ADL, no significant associations were found.

CVD
In the CVD group, higher GFAP levels were significantly associated with worse attention and working memory (B = −0.42,P = 0.030) and executive function (B = −0.45,P = 0.044) scores at baseline (Figure 5).A statistically indeterminate (0.05 < P < 0.1) association was also present between higher p-tau181 levels and worse memory (B = −0.32,P = 0.062) scores.No significant associations were found between plasma biomarkers and longitudinal change in any cognitive domains.Regarding baseline independence in ADL, higher NfL levels were significantly associated with greater impairments in iADL function (β = −0.22,P = 0.015; Figure 5).

DISCUSSION
We demonstrated the association of plasma GFAP, NfL, p-tau181, and NfL, and/or p-tau181 across all of these diseases compared to a healthy aging cohort.These results are key to understanding the potential utility of novel plasma biomarkers to characterize participants with diverse clinical presentations or dementia pathology for better prognosis and monitoring in specialized clinics and in clinical trials.

Plasma GFAP and NfL
We found higher levels of both plasma GFAP and NfL to be broadly associated with worse outcomes for most baseline cognitive domains, for cognitive decline, and for loss of functional independence in the pooled cohort of all diseases, which appeared to be driven mostly by the AD/MCI and PD cohorts in which similar associations were found.
In contrast, sparse associations were found for both plasma GFAP and NfL in the CVD cohort, and none in the FTD cohort.3][44][45][46] As such, our broad results are in keeping with the idea that GFAP and NfL are not specific to a single neurodegenerative disease process, and support some neuropathological overlap across these diseases, especially in AD/MCI and PD.Other studies have also investigated the association of GFAP and NfL with cognition, albeit predominantly with screening tools of global cognition (e.g., Mini-Mental State Examination [MMSE], MoCA), and have found elevated GFAP to be associated with worse baseline cognition or cognitive decline in AD/MCI, 10,47 PD or other synucleinopathies, 38 or FTD, 37,48 while others did not. 10,49][48][49]53,55 While the absence of findings with GFAP or NfL in the FTD cohort may be a true finding, the FTD cohort is considerably smaller than the other disease cohorts. Thfore, we would need large effect sizes for associations to be detectable in these multivariable linear models, 16 raising the possibility that false negative findings are present, given that cognitive decline is one of the last observable changes in neurodegenerative trajectories. Furthermore, tFTD spectrum contains immense heterogeneity, both in terms of underlying predicted pathology and clinical presentations.56  While no significant associations were present between the levels of p-tau181 and cognitive outcomes in the CVD cohort, we found a notable statistically indeterminate association between higher p-tau181 levels and worse memory performance, which may be reflecting underlying AD co-pathology in some of these participants.
Additionally, p-tau181 average levels in the CVD cohort were elevated compared to HCs, and for some participants were as elevated as the AD/MCI cohort, aligning with this hypothesis.This also is consistent with the often-reported concomitant presence of AD and cerebrovascular pathology. 12,14We also have to consider that small vessel disease as identified by T2 changes on MRI is heterogeneous, as its etiology can include any combination of cerebrovascular factors, venous collagenosis, 64,65 and cerebral amyloid angiopathy (CAA, an AD-related vasculopathy). 66,67Therefore, the statistically indeterminate association between elevated p-tau181 and greater memory decline in the CVD cohort may be indicative of CVD pathology driven by CAA, which could be more harmful to cognition.Indeed, a previous study hypothesized that CAA might be the most likely etiology of small vessel disease in APOE ε4 carriers and may have a multiplicative detrimental effect on cognition compared to non-carriers, supporting this hypothesis. 66

Strengths and limitations
We note two overall strengths of our study.First is the ONDRI cohort's multiple well-phenotyped disease diagnostic groups, with harmonized testing and processing protocols throughout, which allow valid comparisons among disease cohorts.Second, the assessment of cognitive domains with yearly comprehensive neuropsychological batteries and of valid functional independence assessments by study partners represent a considerable strength compared to screening measures of global cognition that lack granularity and sensitivity, but which have been used much more commonly due to ease of access and administration.
We also acknowledge a few limitations of our study.First, we did not have PET, CSF, or autopsy-confirmed diagnoses of AD pathology, which makes it more probable that participants had unknown mixed pathology contributing to their assigned clinical diagnosis, reflecting a very common clinical reality.As elevated p-tau181 levels were found in non-AD neurodegenerative diseases, although at about half the rate found in AD, it is likely that some AD co-pathology was present in these cohorts.Second, we acknowledge the lack of diversity within the ONDRI cohort, with the majority (>80%) being of European descent, 16 which may potentially limit the generalizability of our findings to other ethnicities.Third, the unbalanced number of diagnoses might also bias the associations found in the pooled cohort analyses.Finally, although our longitudinal results are novel, they are limited by the 2-year follow-up timeframe, which is a relatively narrow window into the disease continuum of dementia, and by asymmetric attrition possibly introducing a positive bias in subsequent yearly cognitive assessments.

CONCLUSION
Altogether, we present a comprehensive assessment of the associations among plasma biomarkers and detailed cognitive domains, longitudinal cognitive decline, and functional independence in ADLs Grant/Award Number: #ALFGBG-71320; Alzheimer Drug Discovery Foundation, Grant/Award Number: #201809-2016862; AD Strategic Fund and the Alzheimer's Association, Grant/Award Numbers: #ADSF-21-831376-C, #ADSF-21-831381-C, #ADSF-21-831377-C; Erling-Persson Family Foundation, Stiftelsen för Gamla Tjänarinnor, Hjärnfonden, Sweden, Grant/Award Number: #FO2022-0270; European Union's Horizon 2020; European Union Joint Programme -Neurodegenerative Disease Research, Grant/Award Number: JPND2021-00694; UK Dementia Research Institute at UCL, Grant/Award Number: UKDRI-1003 tor dementia (e.g., cerebrospinal fluid [CSF] tau and amyloid beta [Aβ] levels, Aβ positron emission tomography [PET]) 2 are expensive, not widely available or easily accessible, and therefore do not have great potential for scalability in the general population.Recent advances in ultra-sensitive blood-based immunoassays now offer accessible and cost-effective quantification of emerging plasma biomarkers with high potential to translate to clinical settings and revolutionize how neurodegenerative diseases are diagnosed and monitored. 3These include glial fibrillary acidic protein (GFAP; a glial neuroinflammatory biomarker), neurofilament light chain (NfL; a neuroaxonal damage biomarker), phosphorylated tau181 (p-tau181; a phospho-tau isoform considered specific to Alzheimer's disease [AD] pathology), and the ratio of Aβ 42 to Aβ 40 (Aβ 42/40 ; a marker of amyloid plaque deposition also considered specific to AD pathology; see Hansson et al. 4

Aβ 42 /
40 are associated with (1) performance on five cognitive domains at baseline, (2) cognitive decline on five cognitive domains over a 2-year follow-up period, and (3) functional independence in activities of daily living (ADL) at baseline, all across and within multiple neurodegenerative and cerebrovascular diseases.Considering what these plasma biomarkers are measuring, we expected outcomes to be generally associated with GFAP and NfL levels across all diseases, but to be mostly associated with p-tau181 in an AD-specific manner.For these aims, we leveraged the Ontario Neurodegenerative Disease Research Initiative (ONDRI), a unique cohort with a wide breadth of longitudinal data and extensive harmonized protocols across sites and diseases.16,17

was RESEARCH IN CONTEXT 1 .
Systematic review: We reviewed studies reporting on plasma biomarkers and cognition in Alzheimer's disease (AD) and other neurodegenerative or cerebrovascular diseases.Studies are mostly limited to a single disease spectrum and in the extent of cognitive assessments completed.Large gaps in the literature exist regarding how plasma biomarkers relate to cognitive performance, cognitive decline, and functional independence.

2 .
Interpretation: We present a comprehensive assessment of plasma biomarkers and their association with performance in five cognitive domains, longitudinal cognitive decline, and functional independence in activities of daily living across AD, Parkinson's disease, frontotemporal dementia spectrum disorders, and cerebrovascular disease.Selected plasma biomarkers show value in predicting cognitive and functional status.
prior to sample processing) overnight to the OBI Biobank Sample Reception at the Robarts Research Institute (Western University, London, ON, Canada) where they were immediately processed upon receipt for plasma isolation by centrifugation at 2000 rpm for 15 minutes at 4 • C, and then stored at −80 • C until shipment to the Clinical Neurochemistry Laboratory (University of Gothenburg, Mölndal, Sweden) for measurement.The concentrations of GFAP, NfL, Aβ 42 , and Aβ 40 were measured using the Neurology 4-plex E ultra-sensitive Single molecule array (Simoa) immunoassays, while p-tau181 was measured using the pTau-181 Advantage kit.The measurements were performed on an HD-X Analyzer according to instructions from the manufacturer (Quanterix).The measurements were performed in one round of experiments using one batch of reagents by board-certified laboratory technicians who were blinded to clinical data.Intra-assay coefficients of variation were below 10% for all analytes.For analyte concentrations below the functional lower limit of quantification (GFAP = 11.6 pg/mL, NfL = 1.6 pg/mL, p-tau181 = 2 pg/mL, Aβ 42 = 1.51 pg/mL, Aβ 40 = 4.08 pg/mL), missing data were imputed with the lower limit divided by two.Across all participants, including HC, there were 17 samples (3.2%) with concentrations below the limit for p-tau181 (4 HC, 3 AD/MCI, 4 PD, 1 FTD, 5 CVD), 25 samples (4.8%) for Aβ 42 (7 AD/MCI, 6 PD, 4 FTD, 8 CVD), and 7 samples (1.3%) for Aβ 40 (3 AD/MCI, 2 PD, 2 FTD).Aβ 42/40 ratio was not calculated for samples with Aβ 42 or Aβ 40 concentrations below the limit.As a general index of directionality, higher GFAP, NfL, and p-tau181 are considered more pathological, while lower Aβ 42/40 is considered more pathological.Apolipoprotein E (APOE) genotypes for each participant were identified using the custom next-generation sequencing-based panel

F I G U R E 1
Cognitive domains at baseline and follow-up visits across diseases.Composite z scores are relative to the mean and standard deviation of pooled disease groups at baseline, with the observed effects of age, sex, and education removed.Truncated violin plots do not extend past maximal values.Language and visuospatial function graphs' y axes are cut at −7 for display purposes, excluding only a few participants for language (FTD, n = 1; 1y-FTD, n = 1; 2y-AD/MCI, n = 1) and visuospatial function (1y-CVD, n = 1).1y, 1-year; 2y, 2-year; AD/MCI, Alzheimer's disease/mild cognitive impairment; CVD, cerebrovascular disease; FTD, frontotemporal dementia; PD, Parkinson's disease.

F I G U R E 4
were significantly associated with worse executive function (B = −0.57,P = 0.039) scores.Higher NfL levels were significantly associated with worse executive function (B = −0.86,P = 0.011) and visuospatial function (B = −0.62,P = 0.012) scores.A statistically indeterminate (0.05 < P < 0.1) association was also present between lower Aβ 42/40 and worse visuospatial function (B = 0.94, P = 0.057) scores.Longitudinally, higher GFAP and NfL levels were both significantly associated with greater decline in attention and working memory (B = −0.016,P = 0.028; B = −0.024,P = 0.011; respectively) and memory (B = −0.025,P = 0.001; B = −0.043,P < 0.001; respectively) scores.F I G U R E 2 Significant plasma biomarker associations with baseline cognitive domains and independence in activities of daily living across pooled diseases regardless of given diagnosis.Age, sex, years of education, and APOE ε4 carrier status were accounted for in statistical models or at initial computing of cognitive domain scores.Raw data are plotted and P values are derived from the combined group linear mixed effect models.(A) Significant GFAP associations.(B) Significant NfL associations.(C) Significant p-tau181 associations.The gray area represents the 95% confidence interval.APOE, apolipoprotein E; GFAP, glial fibrillary acidic protein; iADL, instrumental activities of daily living; NfL, neurofilament light chain; p-tau, phosphorylated tau.F I G U R E 3 Significant plasma biomarker associations with baseline cognitive domains and independence in activities of daily living in AD/MCI.Age, sex, years of education, and APOE ε4 carrier status were accounted for in statistical models or at initial computing of cognitive domain scores.Raw data are plotted and P values are derived from the AD/MCI group linear mixed effect models.(A) Significant GFAP associations.(B) Significant NfL associations.(C) Significant p-tau181 associations.The gray area represents the 95% confidence interval.AD/MCI, Alzheimer's disease/mild cognitive impairment; APOE, apolipoprotein (E) GFAP, glial fibrillary acidic protein; iADL, instrumental activities of daily living; NfL, neurofilament light chain; p-tau, phosphorylated tau.Significant plasma biomarker associations with baseline cognitive domains and independence in activities of daily living in PD.Age, sex, years of education, and APOE ε4 carrier status were accounted for in statistical models or at initial computing of cognitive domain scores.Raw data are plotted and P values are derived from the PD group linear mixed effect models.(A) Significant NfL associations.(B) Significant GFAP associations.(C) Significant Aβ 42/40 associations.The gray area represents the 95% confidence interval.Aβ, amyloid beta; APOE, apolipoprotein (E) GFAP, glial fibrillary acidic protein; iADL, instrumental activities of daily living; NfL, neurofilament light chain; PD, Parkinson's disease; p-tau, phosphorylated tau.
42/40 was associated with language (B = 6.95,P = 0.040) scores at baseline, this association disappeared when excluding extreme cognitive outliers (n = 2).Longitudinally, lower F I G U R E 5 Significant plasma biomarker associations with baseline cognitive domains and independence in activities of daily living in CVD.Age, sex, years of education, and APOE ε4 carrier status were accounted for in statistical models or at initial computing of cognitive domain scores.Raw data are plotted and P values are derived from the CVD group linear mixed effect models.(A) Significant GFAP associations.(B) Significant NfL associations.(C) Significant p-tau181 associations.The gray area represents the 95% confidence interval.APOE, apolipoprotein E; CVD, cardiovascular disease; GFAP, glial fibrillary acidic protein; iADL, instrumental activities of daily living; NfL, neurofilament light chain; p-tau, phosphorylated tau.Aβ 42/40 was significantly associated with less decline in executive function (B = −0.15,P = 0.031) scores.Regarding baseline independence

Aβ 42 /F I G U R E 6
40 with performance on several cognitive domains at baseline, domain-specific cognitive decline over time, and functional independence in ADLs across neurodegenerative and cerebrovascular diseases in the clinical, multi-site ONDRI cohort.A simplified summary of the associative value of these plasma biomarkers in our sample is presented in Figure6.We also described the elevation in levels of GFAP, Simplified summary of the associative value of plasma biomarkers with outcome measures in the ONDRI sample.Black dot: significant associations with two or more cognitive domains, at baseline or longitudinally, or significant associations with any functional independence scale.Crosshatched dot: significant or statistically indeterminate (P < 0.1) association with one cognitive domain, at baseline or longitudinally.The five cognitive domains are: attention and working memory, executive function, language, memory, visuospatial function.The ADLs category includes % of independence in basic and/or instrumental ADLs.Aβ, amyloid beta; ADLs, activities of daily living; AD/MCI, Alzheimer's disease/mild cognitive impairment; CVD, cardiovascular disease; FTD, frontotemporal dementia; GFAP, glial fibrillary acidic protein; iADL, instrumental activities of daily living; NfL, neurofilament light chain; ONDRI, Ontario Neurodegenerative Disease Research Initiative; PD, Parkinson's disease; p-tau, phosphorylated tau.
42/40) were log 10transformed before analysis due to skewness.Chi-square tests or oneway analyses of variance with Tukey post hoc were used to compare descriptive variables (demographic, clinical, plasma biomarker) among all available groups.Independent t tests were used to explore sex dif- Participant ID was entered as a random effect.Time was defined as a continuous variable of months from initial visit at moment of testing, for all three visits included.Plasma biomarker (log) was entered as a continuous variable.Estimates for biomarker associations with baseline cognitive domain composite scores and with longitudinal change in cognitive domain composite scores were derived from this single linear mixed model.Extreme cognitive outliers (greater than −4 standard deviations) were detected for each of language and visuospatial functions (AD/MCI n = 1, FTD n = 2, CVD n = 1) which, while being real values, could affect statistical results.As sensitivity analyses, the models were repeated with these outliers removed.To assess the association between baseline plasma biomarkers and level of independence in activities of daily living at baseline, linear regressions were performed while controlling for age, sex, education, and APOE ε4 carrier status.Associative analyses were first performed in a pooled disease group of all participants to investigate these associations regardless of given diagnosis, and then with stratification to characterize these associations in individual cohorts.Secondary analyses performed in the HC cohort, and also with Aβ 42 , are present in supporting information Sections 2 and 3. Statistical analyses were performed with SPSS Statistics 29 (IBM Corp., 2022).Beta and P-value SPSS outputs (B, linear mixed models; standardized β, linear regressions) are reported.The threshold of statistical significance was set at P < 0.05 (two tailed).
Baseline sample demographic, clinical, biomarker, and genetic characteristics.
23.5% at third visit (Table1includes attrition breakdown by disease group).Relative to baseline, the average interval until second visit was 12.2 ± 0.9 months (range: 10 to 17 months), while the average interval until third visit was 24.5 ± 1.3 months (range: 22 to 32 months).Early Aβ 42/40 ratio and worse cognition.As decreased Aβ 42/40 ratio is a marker of early AD pathology, we can surmise that PD participants with lower Aβ 42/40 ratio may have had early AD co-pathology and were therefore more cognitively impaired.