Combined neuropathological pathways account for age-related risk of dementia

Abstract

Objective: Our objectives were to characterize the inter-relation of known dementia-related neuropathologies in one comprehensive model and quantify the extent to which accumulation of neuropathologies accounts for the association between age and dementia.

Methods: We used data from 1,362 autopsied participants of three community-based clinicopathological cohorts: the Religious Orders Study, the Rush Memory and Aging Project, and the Minority Aging Research Study. We estimated a series of structural equation models summarizing a priori hypothesized neuropathological pathways between age and dementia risk individually and collectively.

Results: At time of death (mean age, 89 years), 44% of our sample had a clinical dementia diagnosis. When considered individually, our vascular, amyloid/tau, neocortical Lewy body, and TAR DNA-binding protein 43 (TDP-43)/hippocampal sclerosis pathology pathways each accounted for a substantial proportion of the association between age and dementia. When considered collectively, the four pathways fully accounted for all variance in dementia risk previously attributable to age. Pathways involving amyloid/tau, neocortical Lewy bodies, and TDP-43/hippocampal sclerosis were interdependent, attributable to the importance of amyloid beta plaques in all three. The importance of the pathways varied, with the vascular pathway accounting for 32% of the association between age and dementia, wheraes the remaining three inter-related degenerative pathways together accounted for 68% (amyloid/tau, 24%; the Lewy body, 1%; and TDP-43/hippocampal sclerosis, 43%).

Interpretation: Age-related increases in dementia risk can be attributed to accumulation of multiple pathologies, each of which contributes to dementia risk. Multipronged approaches may be necessary if we are to develop effective therapies. Ann Neurol 2018;84:10-22.

Figure 1. Hypothesized pathways by which neuropathology may mediate the association of age and dementia

Figure represents our a priori hypothesized model, with the addition of a modification allowing midtemporal tangles to contribute to the neocortical tangles latent variable. Squares or rectangles represent manifest variables and ellipses represent latent variables. Each single-headed arrow denotes a hypothesized unidirectional effect of one variable on another. Abbreviations: Arter. Scler: arteriolosclerosis, CVDA: atherosclerosis, Gross: gross infarcts, Hippo. Sclerosis: hippocampal sclerosis; Mesio Tangles: Mesiotemporal tangles, Micro: microinfarcts, Neocort Tangles: neocortical neurofibrilary tangles, TDP: TAR DNA-binding protein 43, T-EC: neurofibrillary tangles in the entorhinal cortex, T-Hip: neurofibrillary tangles in the hippocampus, T-IP: neurofibrillary tangles in the inferior parietal cortex, T-MF: neurofibrillary tangles in the midfrontal cortex, T-MT: neurofibrillary tangles in the midtemporal cortex.

Figure 2. Results of path analysis considering mediation by the vascular pathology pathway of the effect of age on dementia risk

Squares or rectangles represent manifest variables and ellipses represent latent variables. Each single-headed arrow denotes a hypothesized unidirectional effect of one variable on another. Numbers associated with effects are standardized regression coefficients (e.g., from age to dementia) or standardized factor loadings (i.e., from a latent variable to its indicators). Paths that were statistically significant at p<0.05 are represented by solid lines. Paths that were hypothesized but were not statistically significant at p<0.05 are denoted by dashed lines. Abbreviations: Arter. Scler: arteriolosclerosis, CVDA: atherosclerosis (cerebrovascular disease of the artery), Gross: gross infarcts, Micro: microinfarcts.

Figure 3. Results of path analysis considering mediation by the amyloid/tau pathway of the effect of age on dementia risk

Squares or rectangles represent manifest variables and ellipses represent latent variables. Each single-headed arrow denotes a hypothesized unidirectional effect of one variable on another. Numbers associated with effects are standardized regression coefficients (e.g., from age to dementia) or standardized factor loadings (i.e., from a latent variable to its indicators). Paths that were statistically significant at p<0.05 are represented by solid lines. Paths that were hypothesized but were not statistically significant at p<0.05 are denoted by dashed lines. Abbreviations:, Neocort Tangles: neocortical neurofibrilary tangles, T-EC: neurofibrillary tangles in the entorhinal cortex, T-Hip: neurofibrillary tangles in the hippocampus, T-IP: neurofibrillary tangles in the inferior parietal cortex, T-MF: neurofibrillary tangles in the midfrontal cortex, T-MT: neurofibrillary tangles in the midtemporal cortex.

Figure 4. Results of path analysis considering mediation by the neocortical Lewy body pathology pathway of the effect of age on dementia risk

Squares or rectangles represent manifest variables and ellipses represent latent variables. Each single-headed arrow denotes a hypothesized unidirectional effect of one variable on another. Numbers associated with effects are standardized regression coefficients (e.g., from age to dementia) or standardized factor loadings (i.e., from a latent variable to its indicators). Paths that were statistically significant at p<0.05 are represented by solid lines. Paths that were hypothesized but were not statistically significant at p<0.05 are denoted by dashed lines.

Figure 5. Results of path analysis considering mediation by the TDP43/hippocampal sclerosis pathway of the effect of age on dementia risk

Squares or rectangles represent manifest variables and ellipses represent latent variables. Each single-headed arrow denotes a hypothesized unidirectional effect of one variable on another. Numbers associated with effects are standardized regression coefficients (e.g., from age to dementia) or standardized factor loadings (i.e., from a latent variable to its indicators). Paths that were statistically significant at p<0.05 are represented by solid lines. Paths that were hypothesized but were not statistically significant at p<0.05 are denoted by dashed lines. Abbreviations: Hippo. Sclerosis: hippocampal sclerosis; TDP: TAR DNA-binding protein 43.

Figure 6. Results of path analysis of combined pathologic pathways mediating the effect of age on dementia risk

Squares or rectangles represent manifest variables and ellipses represent latent variables. Each single-headed arrow denotes a hypothesized unidirectional effect of one variable on another. Numbers associated with effects are standardized regression coefficients (e.g., from age to dementia) or standardized factor loadings (i.e., from a latent variable to its indicators). Paths that were statistically significant at p<0.05 are represented by solid lines. Paths that were hypothesized but were not statistically significant at p<0.05 are denoted by dashed lines. Abbreviations: Arter. Scler: arteriolosclerosis, CAA: cerebral amyloid angiopathy; CVDA: atherosclerosis, Gross: gross infarcts, Hippo. Sclerosis: hippocampal sclerosis; Micro: microinfarcts; Mesio. Tangles: mesiotemporal neurofibrilary tangles; Neocort Tangles: neocortical neurofibrilary tangles, TDP: TAR DNA-binding protein 43, T-EC: neurofibrillary tangles in the entorhinal cortex, T-Hip: neurofibrillary tangles in the hippocampus, T-IP: neurofibrillary tangles in the inferior parietal cortex, T-MF: neurofibrillary tangles in the midfrontal cortex, T-MT: neurofibrillary tangles in the midtemporal cortex.