ABSTRACT
Population-based cohort studies are essential for understanding the pathological basis of dementia in older populations. Previous studies have shown that limbic-predominant age-related TDP-43 encephalopathy neuropathological change (LATE-NC) increases with age, but there have been only a few studies, which have investigated this entity in a population-based setting. Here we studied the frequency of LATE-NC and its associations with other brain pathologies and cognition in a population aged ≥ 85 years. The population-based Vantaa 85+ study cohort includes all 601 individuals aged ≥ 85 years who were living in Vantaa, Finland in 1991. A neuropathological examination was performed on 304 subjects (50.5%) and LATE-NC staging was possible in 295 of those. Dementia status and Mini-Mental State Examination (MMSE) scores were defined in the baseline study and 3 follow-ups (1994-99). The LATE-NC stages were determined based on TDP-43 immunohistochemistry, according to recently updated recommendations. Arteriolosclerosis was digitally assessed by calculating the average sclerotic index of five random small arterioles in amygdala and hippocampal regions, and frontal white matter. The association of LATE-NC with arteriolosclerosis and previously determined neuropathological variables including Alzheimer's disease neuropathological change (ADNC), Lewy-related pathology (LRP), hippocampal sclerosis (HS), and cerebral amyloid angiopathy (CAA), and cognitive variables were analysed by Fisher's exact test, linear and logistic regression (univariate and multivariate) models. LATE-NC was found in 189 of 295 subjects (64.1%). Stage 2 was the most common (28.5%) and stage 3 the second most common (12.9%), whereas stages 1a, 1b and 1c were less common (9.5%, 5.1% and 8.1%, respectively). Stages 1a (P < 0.01), 2 (P < 0.001) and 3 (P < 0.001) were significantly associated with dementia and lower MMSE scores. LATE-NC was associated with ADNC (P < 0.001), HS (P < 0.001), diffuse neocortical LRP (P < 0.002), and arteriolosclerosis in amygdala (P < 0.02). In most cases LATE-NC occurred in combination alongside other neuropathological changes. There were only six subjects with dementia who had LATE-NC without high levels of ADNC or LRP (2% of the cohort, 3% of the cases with dementia), and five of these had HS. In all multivariate models, LATE-NC was among the strongest independent predictors of dementia. When LATE-NC and ADNC were assessed in a multivariate model without other dementia-associated pathologies, the attributable risk was higher for LATE-NC than ADNC (24.2% vs. 18.6%). This population-based study provides evidence that LATE-NC is very common and one of the most significant determinants of dementia in the general late-life aged population.
ABSTRACT
Evolving evidence has supported the existence of two anatomically distinct Lewy-related pathology (LRP) types. Investigation of spinal cord and peripheral LRP can elucidate mechanisms of Lewy body disorders and origins of synuclein accumulation. Still, very few unselected studies have focused on LRP in these regions. Here we analysed LRP in spinal cord, dorsal root ganglion, and adrenal gland in the population-based Vantaa 85 + study, including every ≥ 85 years old citizen living in the city of Vantaa in 1991 (n = 601). Samples from spinal cord (C6-7, TH3-4, L3-4, S1-2) were available from 303, lumbar dorsal root ganglion from 219, and adrenal gland from 164 subjects. Semiquantitative scores of LRP were determined from immunohistochemically stained sections (anti-alpha-synuclein antibody 5G4). LRP in the ventral and dorsal horns of spinal cord, thoracic intermediolateral column, dorsal root ganglion and adrenal gland were compared with brain LRP, previously determined according to DLB Consortium criteria and by caudo-rostral versus amygdala-based LRP classification. Spinal LRP was found in 28% of the total population and in 61% of those who had LRP in the brain. Spinal cord LRP was found only in those subjects with LRP in the brain, and the quantity of spinal cord LRP was associated with the severity of brain LRP (p < 0.001). Unsupervised K-means analysis identified two cluster types of spinal and brain LRP corresponding to caudo-rostral and amygdala-based LRP types. The caudo-rostral LRP type exhibited more frequent and severe pathology in spinal cord, dorsal root ganglion and adrenal gland than the amygdala-based LRP type. Analysis of specific spinal cord regions showed that thoracic intermediolateral column and sacral dorsal horn were the most frequently affected regions in both LRP types. This population-based study on brain, spinal and peripheral LRP provides support to the concept of at least two distinct LRP types.
