Premium
Longitudinal change in CSF biomarkers, especially NPTX2, in non‐demented elderly predicts cognitive decline and conversion to dementia
Author(s) -
Galasko Doug R.,
Smirnov Denis S.,
Salmon David P.
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.046475
Subject(s) - dementia , cognitive decline , proportional hazards model , medicine , biomarker , oncology , longitudinal study , hazard ratio , cognition , effects of sleep deprivation on cognitive performance , cognitive reserve , psychology , gerontology , confidence interval , disease , biology , psychiatry , pathology , genetics
Background We have previously shown that markers of synaptic dysfunction (e.g. NPTX2) measured cross‐sectionally in CSF improve prediction of current cognitive status and future cognitive decline above and beyond amyloid (i.e., Aβ1‐42) alone (Galasko et al , 2019). Here we assess if longitudinal change in these and other candidate CSF biomarkers (NPTX2, VGF, SCG2, CMGA, FAPBH), with total tau (T‐Tau), correlate with cognitive decline and predict conversion to dementia. Methods Longitudinal CSF samples collected over 4 years from 193 non‐demented elderly from ADNI (81 NC, 45 EMCI, 67 MCI, age = 72.6±7, 45% Female, MMSE = 28.4±1.7) were analyzed by mass spectrometry for absolute concentration of NPTX2, VGF, SCG2, CMGA, and FAPBH. Elecsys data for Aβ1‐42 and T‐Tau were also available. All biomarker values were log transformed. Rates of change (Δ) were calculated from regression on age. Linear mixed effects models assessed the ability of the slope of each marker over the same time period to predict the rate of longitudinal cognitive decline on the cognitive tests of the ADNI cognitive battery, adjusted for age, sex, education, baseline cognitive performance, and baseline CSF Aβ1‐42 levels. Cox Proportional Hazard models were fit for progression to dementia (CDR >= 1.0), adjusting for the same covariates. Results Although the candidate markers were all highly intercorrelated (r= .42‐.95), their change over time was differentially associated with amyloid status (Figure 1) and to longitudinal decline on cognitive test score and functional measures (Table 1, e.g. Figure 2). Slopes of NPTX2 (p=0.0008), SCG2 (p=0.03), CMGA (p=0.04) and VGF (p=0.01) each predicted conversion to dementia (Table 2, e.g. Figure 3) and MMSE decline. Change in T‐tau only predicted MMSE decline. Conclusions Longitudinal change of candidate markers, especially NPTX2, has utility for predicting degree of prodromal/early cognitive decline and transition to dementia, beyond the ability of amyloid alone or change in CSF T‐tau. Thus, these candidate CSF biomarkers may have utility as indicators of cognitive impairment and predictors of cognitive decline.