If Midlife Brain Injury Is a Risk Factor for Alzheimer’s Disease and Related Dementias,What Is the Neuropathologic Mechanism?

It is now widely accepted that repeated head injuries in middle life sometimes leads to dementia. While this issue has most immediate relevance to military and sport-related injuries and to a specific type of brain pathology (chronic traumatic encephalopathy), one cannot avoid wondering if childhood or adult brain injuries might increase the risk of developing Alzheimer’s disease or a related dementia (ADRD) in later life. Several analyses addressing this important question have now been carried out by eminent researchers and reviewers employing impeccable data [1–4]. The answers have been mixed and the issue remains unresolved despite exhaustive meta-analyses, all providing extensive bibliographies. The problem is especially challenging because of difficulties identifying and quantifying: 1) a reliable, unbiased measure of prior brain injury, and 2) a solid endpoint, i.e., unbiased and unambiguous neuropsychologic measures and/or demonstrable abnormalities of brain structures already accepted as neuropathologic criteria for ADRD.

A report in this issue of the Journal of Alzheimer’s Disease (JAD), “Association of Traumatic Brain Injury with Late Life Neuropathological Outcomes in a Community-Based Cohort” authored by Gibbons et al. [5], addresses these problems effectively. They employed multiple types of complementary head injury information and used these in searches for possible linkages with specific ADRD brain lesions in comprehensive brain autopsies of a generally representative and unbiased community sample of older decedents. This approach is based on the idea that if a head injury were to initiate or aggravate the development of dementia, it might do so by contributing to the clinical manifestations or development of any of the accepted neuropathologic markers of ADRD, i.e., amyloid plaques, tau-positive neurofibrillary tangles, infarcts associated with small vessel disease, cortical Lewy bodies, hippocampal sclerosis, or limbic-associated TDP-43 encephalopathy. One or a combination of these five types of dementia-associated neuropathologic change (NC) are currently viewed by the neuropathology research community as responsible for the majority of ADRD cases [6–8]. Establishing an association of a history of brain injury with any of these lesions, with or without concomitant cognitive impairment, would represent a “smoking gun” implicating prior trauma as a contributing risk factor for ADRD. In this report none of these five types of dementia-associated NC were linked to prior head injury. Instead, a significant association was found with a different abnormality: primary brain atrophy.

Identifying brain atrophy as “primary” simply means that it was not attributed to something else, such as any of the five principal types of NC named above. Two indicators of atrophy from the gross autopsy report were considered by Gibbons and colleagues [5], one based on a comparison of weights of brain parts above and below the tentorium (a method for assessing cerebral atrophy described by Alvord half a century ago) and the second a single dichotomized assessment of ventricular width [9]. Neither is a particularly robust or reliable indicator of brain atrophy if head or skull size is not taken into account. Nonetheless the association with prior injury was consistent and solid.

A co-author of the Gibbons report (MCP, a respected neuroepidemiologist) was also a co-author of a related report in a recent JAD by Grasset and colleagues [10]. Analyses In that report were carried out to identify associations of a brain injury history with subsequent cognitive impairment and with changes assessed by neuroimaging. While they failed to identify an association with cognitive impairment, infarcts, or hyperintensities, they did observe a significant association with lesser white matter volumes. Obviously, a substantial loss of either gray or white matter represents brain atrophy.

Assessments of brain atrophy implicitly involve a comparison of current with prior known or imputed brain size. Neuroimagers have long appreciated the critical importance of intracranial volume or related internal skull measures as primary determinants of regional and general brain volumes, and of the importance of longitudinally repeated assessments to appreciate the dynamic aspects of atrophy. Most analyses of MR or CT data are currently provided with adjustments for individual differences in intracranial measures. Most autopsy reports do not provide any means for such adjustments. To address this weakness intracranial volumes may be extracted from available neuroimaging data. Alternatively, intracranial volume may be confidently estimated based on head circumference and internal skull widths determined at autopsy. Using one or both methods with our own autopsy and neuroimaging data we have determined that the ratio of total brain weight (in grams) to intracranial volume (in ml) ranges from about 0.7 to nearly 1.0. A low ratio provides strong, quantified support for generalized brain atrophy. Unfortunately, few or no skull or head measurements are part of most gross brain autopsy protocols, thereby diminishing the meaning of a simple brain weight or subjective assessment of ventricular dilation. In contrast, neuroimaging methods provide measures of both intracranial volume and regional and total brain volumes— and by extension, of atrophy. However, currently available neuroimaging methods cannot yet identify and measure the specific histopathologic changes required by neuropathologic diagnostic criteria for identifying the several different varieties of ADRD. This inherent weakness of neuroimaging sharply diminishes its utility for appraising specific candidate risk factors likely to operate through an association with a correspondingly specific pathogenic mechanism responsible for ADRD.

Using different study designs these two JAD reports point to the same pathologic endpoint. Taken together they suggest that head injury in childhood or adult life may well leave a legacy of focal or distributed reductions in gray and/or white matter which might reasonably increase vulnerability to late life cognitive decline. These and other accruing observations suggest a need for innovative research to better understand the causes and consequences of brain atrophy above and beyond those attributable to Alzheimer or other disease processes currently linked to ADRD. The scrupulous epidemiologic designs of both JAD reports exemplify the kind of research needed to discern critical childhood and adult experiences and exposures that may explain why some of us will escape while one third to half of us will fall victim to dementia in the final years of ourlives.