Alzheimer Disease Cerebrospinal Fluid Biomarkers in a Tertiary Neurology Practice

Abstract

Objective: To evaluate the performance of Alzheimer disease (AD) cerebrospinal fluid (CSF) biomarkers in a tertiary neurology clinic setting with high frequency of non-AD cases, including normal pressure hydrocephalus (NPH).

Methods: There were 534 patients who underwent AD CSF biomarkers (Roche Elecsys Aβ42, p-Tau181, total-Tau) from April 1, 2020, through April 23, 2021. A behavioral neurologist blinded to CSF results assigned a clinical diagnosis retrospectively on the basis of consensus criteria, and a neuroradiologist blinded to the diagnosis and CSF studies graded brain magnetic resonance images for indicators of CSF dynamics disorders. Associations between biomarkers, diagnoses, and imaging were assessed by χ², analysis of covariance, and linear regression methods.

Results: Median age at time of testing was 67 years (range, 19 to 96 years), median symptom duration was 2 years (range, 0.4 to 28 years), and median Short Test of Mental Status score was 30 (range, 0 to 38). Clinical diagnoses significantly correlated with different CSF biomarker values (χ²=208.3; P=10e-4). p-Tau181/Aβ42 ratios above 0.023 positively correlated with Alzheimer dementia (more than individual measures). This ratio also had the best performance for differentiating Alzheimer dementia from NPH (area under the curve, 0.869). Imaging markers supportive of CSF dynamics disorders correlated with low Aβ42, p-Tau181, and total-Tau.

Conclusion: In a heterogeneous clinical population, abnormal p-Tau181/Aβ42 ratios (>0.023) have the strongest association with Alzheimer dementia and probably represent a comorbid AD pathologic component in persons clearly matching non-AD neurodegenerative syndromes. Altered CSF dynamics were associated with lower concentrations of AD CSF biomarkers regardless of clinical diagnosis, but the ratio compensates for these changes. In the appropriate clinical setting, an isolated abnormal Aβ42 should prompt consideration of NPH.

FIGURE 1.

Association between 7 clinical diagnoses and 4 Alzheimer disease cerebrospinal fluid (AD CSF) profiles. Top, Pearson residuals in white text demonstrate statistically significant associations between diagnosis and AD CSF profiles (exceeding the critical z score cutoff of ±2.9). Increasing red circle size and hue indicate an increasingly negative association; increasing blue circle size and hue denote an increasingly positive association. Bottom, Number of patients within each cell. ADem, Alzheimer dementia; aMCI, amnestic mild cognitive impairment; CBS, corticobasal syndrome; DLB, dementia with Lewy bodies; FTLD, frontotemporal lobar degeneration spectrum; NonD, nondegenerative (excluding NPH); NPH, normal pressure hydrocephalus.

FIGURE 2.

Cerebrospinal fluid (CSF) biomarkers and abnormal CSF dynamics. Top, Alzheimer disease (AD) CSF biomarker in AD vs normal pressure hydrocephalus (NPH; with frontotemporal lobar degeneration spectrum [FTLD] and nondegenerative [excluding NPH] as comparison groups). Red dotted line denotes assay cutoff for normal vs abnormal. Alzheimer dementia (ADem) has the lowest Aβ42 and highest p-Tau181, t-Tau, and p-Tau181/Aβ42 ratio. NPH has the second lowest Aβ42 and the lowest p-Tau181 and t-Tau. Ratios for NPH, FTLD, and nondegenerative (excluding NPH) are equivocal. Results of analysis of covariance are displayed under the graphs. Whiskers demark 5% and 95% ranges, the box represents the interquartile interval, the center line represents the median, and outliers are represented by dots (p.perm, P values corrected for multiple comparisons by permutation testing). Middle, Performance of AD CSF biomarkers in differentiating ADem from NPH receiver operating curves. Area under the curve (AUC) of the p-Tau181/Aβ42 ratio is demonstrated (0.869), as are the specificity (0.889) and sensitivity (0.821) using the assay cutoff of >0.023. Bottom, The AUC, confidence interval, positive and negative predictive values, and sensitivity and specificity of all 4 CSF measures using the current cutoffs. Isolated abnormal Aβ42 is sensitive but not specific for NPH.

FIGURE 3.

Alzheimer disease cerebrospinal fluid (AD CSF) measures vs imaging features of abnormal CSF dynamics (in a combined group of patients with clinical Alzheimer dementia, amnestic mild cognitive impairment, and normal pressure hydrocephalus). Left, CSF Aβ42, t-Tau, p-Tau181, and p-Tau181/Aβ42 ratio measures; callosal angles are age-adjusted residuals. Significant associations (p.adj<.01) are denoted by asterisks (p.adj, P values corrected for multiple comparisons by permutation testing). Right, imaging features of abnormal CSF dynamics, representative images taken from patients with clinical normal pressure hydrocephalus in this study. Callosal angle was measured at the level of the posterior commissure perpendicular to the anterior-posterior commissure line. Abnormal features are highlighted by red circles.