Informing etiological heterogeneity of mild cognitive impairment and risk for progression to dementia with plasma p-tau217

Background: Mild cognitive impairment (MCI) is a clinical diagnosis representing early symptom changes with preserved functional independence. There are multiple potential etiologies of MCI. While often presumed to be related to Alzheimer's disease (AD), other neurodegenerative and non-neurodegenerative causes are common. Wider availability of relatively non-invasive plasma AD biomarkers, such as p-tau217, can provide invaluable insights into MCI clinico-pathology and the associated implications for symptom etiology, prognosis (e.g., risk for progression to dementia), and treatment options.

Objectives: The main goal of this study was to evaluate differences between individuals with MCI with and without plasma p-tau217 biomarker evidence of AD (MCI_AD+ and MCI_AD-) as well as a control group of clinically normal older adults with negative AD biomarkers (CN_AD-). We evaluated group differences in demographics, recruitment, clinical scales, fluid biomarkers, and brain imaging. We further probed these factors as independent contributors to symptoms among MCI_AD- participants, for whom symptom etiology is most poorly understood. Lastly, in a subset of participants followed longitudinally, we investigated how these factors related to odds of clinical progression to dementia.

Design: We conducted an observational cross-sectional and longitudinal clinical research study. Study groups were compared cross-sectionally on demographics, recruitment, clinical measures, and biomarkers (chi square analyses, analyses of covariance). Contributors to functional changes were evaluated with multiple linear regression. Factors associated with the odds of progression from MCI to dementia longitudinally were evaluated with binary logistic regression.

Setting: 1Florida Alzheimer's Disease Research Center.

Participants: Cross-sectional analyses included 378 older adults classified as CN_AD- (N = 76, age 66.1 ± 7.2, 63.2% female, 23.7% non-Hispanic/White), MCI_AD- (N = 198, age 68.9 ± 7.9, 51.5% female, 29.3% non-Hispanic/White), or MCI_AD+ (N = 104, age 73.9 ± 7.4, 52.9% female, 49.0% non-Hispanic/White). Longitudinal analyses focused on 207 participants with MCI (68.5% of cross-sectional MCI sample) followed for an average of 3 years.

Measurements: Demographics (age, sex, years of education, self-identified race and ethnicity, primary spoken language), National Alzheimer's Coordinating Center-defined clinical phenotypes (Clinically Normal, Impaired - Not MCI, Amnestic MCI, Nonamnestic MCI, Dementia), recruitment source (clinic-based versus community-based), genetics (APOE genotype), functional evaluation (Clinical Dementia Rating scale), global cognition (Mini Mental State Exam), vascular history (Vascular Burden Score), neuropsychiatric symptoms (NPI-Q Total score), plasma biomarkers (ALZPath p-tau217, Quanterix Simoa-based GFAP and NfL), and brain imaging (grey matter volume in select AD-relevant regions of interest, global white matter hyperintensity volume).

Results: Among those with MCI, 104 (34.4%) had plasma biomarker evidence of AD. MCI_AD+ participants were more frequently recruited from clinic-based settings than MCI_AD- (74.8% vs. 47.5%, p<.001). Over half (51.5%) of MCI_AD+ carried at least one APOE e4 allele compared to 26.6% of MCI_AD- and 29.4% of CN_AD- (p<.001). Both MCI_AD+ (p<.001, Cohen's d = 0.93) and MCI_AD- (p<.001, d = 0.75) reported more severe neuropsychiatric symptoms than CN_AD. MCI_AD+ had higher plasma GFAP and NfL than both MCI_AD- (GFAP: p<.001, d = 0.88, NfL: p<.001, d = 0.86) and CN_AD- (GFAP: p<.001, d = 0.80; NfL: p<.001, d = 0.89). For the AD signature region of interest, MCI_AD+ had lower volume than both CN_AD- (p<.001, d = 0.78) and MCI_AD- (p=.018, d = 0.39). For the hippocampus, both MCI_AD+ (p<.001, d = 0.87) and MCI_AD- (p<.001, d = 0.64) had lower volume than CN_AD-. Longitudinally, older age (OR=1.14 [1.06-1.22], p<.001), higher levels of p-tau217 (OR=10.37 [3.00-35.02], p<.001) and higher neuropsychiatric symptoms (OR=1.19 [1.02-1.39], p=.023) were associated with higher odds of progression to dementia.

Conclusions: MCI is etiologically heterogeneous. The presence of Alzheimer's pathology defined by elevated plasma p-tau217 in individuals with MCI significantly worsens prognosis. Neuropsychiatric symptoms may contribute to cognitive complaints and risk for progressive decline irrespective of AD pathology. Plasma p-tau217 can inform our understanding of base rates of different MCI phenotypes on a larger scale. As with other AD biomarkers, frequency of elevated plasma p-tau217 and odds of progression to dementia requires careful consideration of recruitment source (clinic- vs. community-based), especially across ethno-racially diverse older adults. Ongoing integration of emerging neurodegenerative disease biomarkers with detailed clinical evaluations will continue to improve treatment specificity and prognosis.