PASSED: Brain atrophy in non-demented individuals in a long-term longitudinal study from two independent cohorts

Abstract

Introduction: Alzheimer's disease (AD) is indicated by a decrease in amyloid beta 42 (Aβ42) level or the Aβ42/Aβ40 ratio, and by increased levels of Tau with phosphorylated threonine at position 181 (pTau181) in cerebrospinal fluid (CSF) years before the onset of clinical symptoms. However, once only pTau181 is increased, cognitive decline in individuals with subjective or mild cognitive impairment is slowed compared to individuals with AD. Instead of a decrease in Aβ42 levels, an increase in Aβ42 was observed in these individuals, leading to the proposal to refer to them as nondemented subjects with increased pTau-levels and Aβ surge with subtle cognitive deterioration (PASSED). In this study, we determined the longitudinal atrophy rates of AD, PASSED, and Biomarker-negative nondemented individuals of two independent cohorts to determine whether these groups can be distinguished by their longitudinal atrophy patterns or rates.

Methods: Depending on their CSF-levels of pTau 181 (T), total Tau (tTau, N), Aβ42 or ratio of Aβ42/Aβ40 (A), 185 non-demented subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and 62 non-demented subjects from Erlangen AD cohort were assigned to an ATN group (A-T-N-, A-T+N±, A+T-N±and A+T+N±) and underwent T1-weighted structural magnetic resonance imaging (sMRI). Longitudinal grey matter (GM) atrophy patterns were assessed with voxel-based morphometry (VBM) using the cat12 toolbox on spm12 (statistical parametric mapping) of MRI scans from individuals in the ADNI cohort with a mean follow-up of 2 and 5 years, respectively. The annualized atrophy rate for individuals in the Erlangen cohort was determined using region of interest analysis (ROI) in terms of a confirmatory analysis.

Results: In the A-T+N± group, VBM did not identify any brain region that showed greater longitudinal atrophy than the A+T+N±, A+T+N± or biomarker negative control group. In contrast, marked longitudinal atrophy in the temporal lobe was evident in the A+T-N± group compared with A+T-N± and biomarker-negative subjects. The ROI in the angular gyrus identified by VBM analysis of the ADNI cohort did not discriminate better than the hippocampal volume and atrophy rate between AD and PASSED in the confirmatory analysis.

Discussion: In this study, nondemented subjects with PASSED did not show a unique longitudinal atrophy pattern in comparison to nondemented subjects with AD. The nonsignificant atrophy rate compared with controls suggests that increased pTau181-levels without concomitant amyloidopathy did not indicate a neurodegenerative disorder.

Keywords: ATN classification; Alzheimer disease/physiopathology*; SNAP; amyloid beta-peptides/cerebrospinal fluid*; longitudinal studies; magnetic resonance imaging; medial temporal lobe; tau proteins/cerebrospinal fluid*.

Figure 1

Longitudinal voxel-wise analysis revealed regionally increased atrophy rates in Aβ42–T–N– (n = 50), Aβ42–T + N ± (n = 43), Aβ42+T–N–. (n = 32), and Aβ42 + T + N ± (n = 60) non-demented individuals from the ADNI cohort with a mean follow-up of 2 years, p < 0.01 (FWE corrected) (A). Slice overlay of the statistical parametric maps illustrates the linear decrease of grey matter (GM) volume that was significantly larger in the Aβ42 + T + N ± (group-red yellow) or the Aβ42 + T–N ± (blue green) compared to the Aβ42–T + N ± group, p < 0.01 (FWE corrected) (B). The color bars indicate the value of p.

Figure 2

Receiver operating characteristic (ROC) curves illustrate the annualized longitudinal atrophy rates in the unbiased region of interest analysis (ROI; ΔuROI) and the hippocampus (ΔHippocampus) to distinguish A–T + (PASSED) individuals (n = 19) from A + T + (AD) nondemented individuals (n = 17) from the Erlangen Cohort (A). Area under the curve (AUC) values showed no significant difference between groups. Baseline uROI atrophy and hippocampal atrophy also showed no significant difference between PASSED and AD (B). Similarly, when A–T–N– (biomarker-negative) subjects (n = 26) and AD subjects were compared in terms of annualized longitudinal (C) or baseline (D) atrophy, no significant differences in AUC values were observed.