“Multi‐infarct” Dementia: A Real Entity?

“Multi-infarct dementia” (MID) is a widely applied concept of a form of dementia in which the cognitive impairment is thought to result from the cumulative effect of repeated episodes of cerebral infarction.1,2 The pathophysiologic assumption that follows is that recurrent cerebral infarction leads to enough cerebral tissue damage to result in loss of a “critical mass” of brain substance beyond that needed for maintenance of normal intellectual performance. As defined, MID is said to represent a distinct entity, clinically, pathologically, and pathophysiologically different from the most common type of dementia, Alzheimer's disease,3 currently referred to as senile dementia of the Alzheimer type (SDAT). The clinical acceptance of MID as a distinct form of dementia has been fostered by the introduction of an “ischemic score,”4 in which a variety of parameters related to the onset, clinical features, and course of the dementia, coupled with historical and physical findings indicative of vascular disease, result in a clinical profile that correlates highly with a “vascular mechanism” of the dementia. Thus, items such as “abrupt onset,” “stepwise deterioration,” and “fluctuating course” of the dementia, in addition to history of hypertension and strokes, evidence of associated atherosclerosis, focal neurologic symptoms and signs, emotional incontinence, depression, somatic complaints, nocturnal confusion, and relative preservation of personality, are each assigned a numerical value, with a possible maximum score of 18 points. It has been estimated that scores of 7 or more correlate with MID, whereas values of 4 or less are indicative of a higher probability of SDAT.2 Such a scoring system, known as the Hachinski Ischemic Score, has been validated pathologically by Rosen et al.,5 who found that high ischemic scores correlated with vascular dementia, whereas low ischemic scores corresponded with pathologic features of Alzheimer's disease. Using these clinical and pathologic criteria, it has been found that MID represents a relatively infrequent, but well established, cause of dementia, accounting for approximately 12 to 20% of the cases, whereas SDAT is responsible for about 50% of the cases, and an additional “mixed group” (showing pathologic features of Alzheimer's disease, as well as multiple cerebral infarcts) amounts to 16 to 20%.6 As defined, the concept of MID is met by a number of problems, primary among them being that of its overuse in the diagnostic evaluation of dementia cases.7 A common clinical scenario is that of an elderly patient with long-standing dementia, who is found not to have elements of a “treatable” form of dementia (chronic subdural hematoma, normal-pressure hydrocephalus, brain tumor, hypothyroidism, neurosyphilis, B12 or folate deficiency), but who shows on computed tomography (CT) scan of the head evidence of one or more old cerebral infarcts, thus being labeled as an example of MID. This clearly represents an instance of an etiologic diagnosis of dementia being made on the basis of insufficient data, because the coexistence of cerebral infarcts and dementia does not imply a cause and effect relationship. Both conditions, dementia and cerebral infarction, are found with high enough frequency in elderly populations that their coincidental association is expected to be a common occurrence. Thus, the question arises, how much pathologic support exists for the concept of MID, as well as for the validity of a clinical diagnosis of MID based on coexistence of dementia and a “high ischemic score” with CT documentation of cerebral infarcts? In this regard, the careful pathologic studies of Tomlinson et al.8,9 are the strongest available evidence for an association between cerebral infarction and dementia. However, their data showed this association to be one somewhat more complex than the mere existence of multiple cerebral infarcts and absence of histopathologic features of Alzheimer's disease in an individual afflicted by progressive dementia. Their study involved, first, the systematic pathologic examination of 28 brains from elderly individuals without evidence of dementia.8 Cerebral infarcts were detected in 20 instances, as minute foci of softening in seven and as larger ones (detectable under naked eye examination) in 13, eight of which corresponded to bilateral basal ganglia infarcts. The volumes of these infarcts varied between the extremes of 2 and 91 mL, with a mean value of 13.2 mL. Histopathologic features of Alzheimer's disease (senile plaques, neurofibrillary change, and granulovacuolar degeneration) were present in small numbers in approximately one-half of the brains in this nondemented group of patients. A second study9 followed, involving the pathologic study of 50 brains of patients with documented dementia, and the findings included: pathologic features diagnostic of Alzheimer's disease in 25 cases, “arteriosclerotic” dementia (corresponding to the current concept of MID) in six cases, and mixed dementia (“senile” and “arteriosclerotic” forms coexisting) in four additional cases. In the group of six cases labeled as arteriosclerotic dementia, large volumes of cerebral softening were detected, all exceeding 100 mL (extremes of 101 and 412 mL, with a mean value of 188 mL); a figure that was significantly higher than the mean volume of infarction detected in their control (ie, nondemented) population.8 These data suggest that multiple cerebral infarcts as a potential cause of dementia are required to amount to a certain critical volume (in the area of 150 to 200 mL), below which there is no clear evidence to implicate cerebral infarction in the cause of dementia. This fact is rarely taken into account when a diagnosis of MID is made on the basis of CT documentation of cerebral infarcts, which are frequently small, in the form of bilateral lacunar infarcts in the basal ganglia and internal capsule region, a finding that was quite common (eight of 28 cases) in the nondemented group studied by Tomlinson et al.8 The latter observation has been corroborated by Fisher10 in studies of cases with bilateral lacunar infarcts affecting the basal ganglia and internal capsule, with the resulting profound motor and behavioral disturbances that characterize the “lacunar state,” but with only mild or no elements of dementia. The next question arises: Is it possible that infarct location (rather than size alone) may also play a major role in the potential for producing dementia? The available pathologic studies5,9 have not addressed this specific point, and the six cases described by Tomlinson et al.9 corresponded to large, multifocal areas of infarction affecting major portions of the cerebral cortex and white matter, as well as variable involvement of the basal ganglia, pons, and cerebellum. Recent pathologic and histochemical studies of brains with evidence of Alzheimer's disease have focused on the apparent role of degeneration of selected subcortical structures, specifically the nucleus basalis of Meynert, in the pathogensis of cortical changes of Alzheimer type.11,12 The possible relationship between selective neuronal loss in this nucleus and SDAT originated from the initial observation of depressed cortical cholinergic activity in brains from patients with Alzheimer's disease. A significant reduction of choline acetyltransferase, the biosynthetic enzyme for acetylcholine, was detected postmortem in the cerebral cortex of Alzheimer's disease patients.13,4 This abnormality was most severe in the temporal cortex, hippocampus, and amygdala, areas that classically show the most marked histopathologic abnormalities in SDAT. These cortical histochemical changes were subsequently found to correlate with the degree of senile plaque formation, both of which in turn correlated with the extent of intellectual impairment.15 This set of findings have led to the current concept of Alzheimer's disease as a form of dementia in which the deficit of cortical cholinergic innervation is of primary etiologic significance. Because the principal source of cortical cholinergic nerve terminals is in the neurons of the nucleus basalis of Meynert, attention to this basal forebrain structure followed, with the finding of significant neuronal loss in Alzheimer's disease cases, as compared with control subjects.13 This finding raises the possibility that small lesions of any kind, including infarcts, could damage this structure, resulting in secondary cortical changes of Alzheimer's disease and a clinical picture of SDAT. Such theoretical combination of bilateral deep small infarcts and cortical histopathologic changes of Alzheimer's disease would fit in the “mixed” category of dementia that shows anatomic evidence of both conditions. Although this association of focal small infarcts in the area of the nucleus basalis with SDAT has not been pathologically documented, future anatomic studies should look into this potential association by specifically correlating clinical dementia with involvement of this subcortical structure by infarction or other types of pathology. In conclusion, the concept of MID as a distinct etiologic variety of dementia has had some limited pathologic support, with a number of qualifying variables, which are largely ignored when diagnosing this condition on clinical grounds. What is required to further test the validity of MID as a real entity? It is proposed that such a task could be accomplished by a variety of approaches, and the combination of their results may eventually solidify a clinicopathologic concept of MID, as well as provide data on its true incidence in the dementia population. These include: 1) further clinicopathologic studies involving dementia populations, in which specific attention is paid to volume and location of cerebral infarcts, to delineate more precisely the anatomic determinants of dementia, and the possibility of high association with involvement of specific structures, such as the nucleus basalis (a strict methodology will have to be used, studying brain specimens systematically for presence and characteristics of infarcts, absence of features of Alzheimer's disease, comparison with control populations, as well as correlating these data with the clinical characteristics of the dementia); and 2) prospective studies of populations at risk for development of stroke and dementia, in an attempt at determining to what degree the two interrelate, thus establishing the possible cause-and-effect relationship between them (this approach can be coupled with physiologic measurements, such as cerebral blood flow (CBF), as recently reported by Rogers et al.,16 in which these authors prospectively followed a population at risk for development of dementia and stroke, finding a significant decline in CBF in those who eventually developed MID, decline that preceded the clinical onset of dementia, whereas such change occurred only late, subsequent to the onset of dementia, in the group categorized as having developed SDAT). This combined clinicopathologic and prospective clinicophysiologic approach may provide the necessary data for the validation of the concept of MID.