Results of this study showed that neuropsychological test scores were most strongly associated with HV and cGM volume regardless of the specific neuropsychological domain being assessed. Abnormal white matter was weakly associated with cognitive performance, particularly for timed measures and measures of executive function. This pattern of results was essentially the same in participants with lacunes and without lacunes. Subcortical lacunes were weakly associated with cognitive performance.
Results from categorical data showed that hippocampal atrophy and cortical atrophy made independent and additive contributions to cognitive impairment and dementia. Dementia was much more likely in the presence of both cortical and hippocampal atrophy (82%), but was very uncommon in the absence of both (3%). Cognitive impairment not meeting criteria for dementia was observed with similar frequencies in all four combinations of normal vs abnormal HV and normal vs abnormal cGM volume. HV was most effective in differentiating normal and cognitively impaired cases, whereas cGM was a more important determinant of impaired vs demented.
These results converge with a previous report13
to suggest that subcortical cerebrovascular disease has an indirect relation with cognitive impairment and dementia. Our results suggest that, as in AD, cognitive impairment in ischemic vascular dementia results from associated atrophy in the cGM and hippocampus. Thus, clearly different pathologic entities, such as AD and cerebrovascular disease, may both impact similar cortical structures and produce similar dementia syndromes, with the symptom profile and severity determined by the location and amount, rather than qualitative type, of pathology.
An obvious explanation for the relation between cortical and hippocampal volume and cognition observed among patients with cerebrovascular disease is that these patients have concomitant AD.27
The prevalence of both AD and cerebrovascular disease in the older population makes coincidence of the two pathologies likely in a significant proportion of dementia cases. Further, some have argued that cerebrovascular disease might be a risk factor for AD.28
Results from autopsied cases from this project suggest that concomitant AD pathology is unlikely as the sole explanation for cortical and hippocampal atrophy in patients with subcortical vascular lesions.13
That study described three autopsied cases with dementia and lacunes. All three had a primary pathologic diagnosis of ischemic vascular disease, and none had neurofibrillary tangles in the neocortex. All three cases had both HV and cGM volume below the 10th percentile of the normal group. The number of neuropathologically diagnosed cases is small and replication of this study with neuropathologically diagnosed cases is critical, but results indicate that dementia with associated cortical and hippocampal atrophy can exist in the absence of significant AD.
There are several potential mechanisms that could explain structural changes in the cortex and hippocampus associated with SIVD. First, damage to subcortical neurons and demyelination of axons could result in secondary structural changes in cortical or hippocampal neurons. Second, the cerebral cortex and hippocampus might suffer subclinical ischemic brain injury.29
Microvascular changes in the cortex have been described as an important pathologic feature of vascular dementia.30
Severe WML has been associated with deficient autoregulatory response31
and increased oxygen extraction fraction32-34
that could result in ischemic brain injury. The strong relation in this study between abnormal white matter and cortical volume loss supports a hypothesized link between white matter lesion and cortical ischemic injury, and may indicate that white matter change provides a useful marker for more extensive cerebrovascular disease. This was a robust relation that did not significantly differ across cognitive impairment groups, nor across those with and without lacunes.
WML clearly was a better indicator of cognitive impairment than subcortical lacunes. Current diagnostic criteria for vascular dementia emphasize the presence of radiologically documented infarcts. White matter disease is formally recognized only in the California ADDTC criteria and is not incorporated in either the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) or the NINDS and Association Internationale pour la Recherche et l’Enseignement en Neurosciences (AIREN) diagnostic criteria for vascular dementia. These results suggest that white matter pathology may actually be a better marker for vascular dementia due to SIVD.
The finding that HV was the most sensitive discriminator of normal vs impaired cognitive status, and cGM volume best discriminated impaired and demented is of particular interest. In AD, there is an established progression of neurofibrillary pathology from medial temporal lobes to neocortex.35
This helps to explain the early memory impairment as a result of hippocampal changes, and subsequent global dementia that presumably relates to broader neocortical changes. As previously discussed, there is also a rationale to explain the importance of cortical changes for dementia in cases where vascular disease is the primary etiology. The hippocampal influence on milder impairment associated with cerebrovascular disease is not as easy to explain and merits further study, especially given that HV was minimally correlated with lacunes and WML.
This study has a number of limitations. The study design was essentially correlational, which limits interpretations about causal mechanisms underlying cognitive impairment associated with vascular disease. MRI volumetric measures of more specific regions of cortical gray matter would be valuable, and measures of frontal lobe volume would appear to be particularly relevant given the theoretical relevance of frontal-subcortical circuits in ischemic vascular dementia. Similarly, more detailed examination of effects of lacunes in specific locations and of more specific types, and locations of WML might provide valuable information. The study sample was essentially a clinical sample of convenience and might not be representative of broader populations of patients with AD and cerebrovascular disease. Future studies using representative sampling methods are essential to enhance the generalizability and interpretation of studies of ischemic vascular disease with respect to dementia. Careful longitudinal studies with neuropathologically verified diagnoses will be important.
There is a growing body of literature examining additive or interactive effects of AD and cerebrovascular disease. A better understanding of how these common problems for older persons contribute alone and in combination to cognitive impairment and dementia is likely to be very important for a better understanding of dementia, and might also yield important clues for prevention and treatment of dementia.