Is Silent Cerebrovascular Disease an Important Cause of Late‐Life Cognitive Decline?

The report by Ferrucci and co-workers that appears in this issue of the Journal identifies poor scores and declining performance on a mental status questionnaire (MSQ) as potent risk factors for subsequent stroke in older persons without a history of cerebrovascular disease.1 They interpret these observations as evidence that clinically unrecognized cerebrovascular disease is an important cause of aging-related cognitive impairment, as well as a predictor of future strokes. These ideas have major conceptual, medical, and public health implications. The common cognitive declines of late life are usually attributed to a dementing process or to normal aging. Depending on several factors (test score, trajectory, age, education, etc.), some persons who show cognitive decline will receive clinical diagnoses of dementia, while others will not. In most studies conducted in the United States, vascular processes have been considered substantially less important as causes of dementia than Alzheimer's disease (AD), with reported vascular:AD ratios in the range of 0.1 to 0.5.2 By labeling declines not associated with clinically apparent stroke as probably caused by normal aging or a nonvascular dementing disease, it is implied that they are not appropriate targets for aggressive interventions intended to prevent future strokes. If subclinical cerebrovascular disease is a common cause for aging-related cognitive decline, it follows that some or all persons who demonstrate such declines may benefit from vigorous treatment or prevention interventions. Such interventions might limit further cognitive decline while simultaneously preventing future stroke. Although the evidence in support of this idea is still insufficient to warrant any across-the-board recommendation, the potential public health benefits of reduced late-life impairment and dependency require us to examine these issues carefully. There is growing evidence that older persons are more commonly subject to strokes than is indicated by clinical history. Based largely on CT examinations, the prevalence of previously unrecognized strokes has been estimated at levels between 10 and 65% in middle-aged and older persons with co-existing hypertension, depression, transient ischemic attack, first stroke, or carotid stenosis.3–13 These are undoubtedly underestimates, inasmuch as many investigators have found that while no neuroimaging method allows visualization of all strokes, the MRI is substantially more sensitive than CT, perhaps by a factor of 2.14 Preliminary MRI results from the NHLBI's Cardiovascular Health Study (CHS) — one of a limited number of studies of an aging community sample rather than of a panel of disease cases and healthy controls — indicate that the ratio of unrecognized to clinically recognized stroke is in the range of 4:1 or greater.15 Had institutionalized persons been included in the CHS sample, the ratio would likely have been even higher. These startling figures appear to be borne out by past and recent autopsy series.16–18 As might be expected, previously unrecognized strokes found on neuroimaging or autopsy are usually (but not always) small lacunar lesions and/or occur in brain regions not immediately required for an essential motor, sensory, or language capacity. Lacunar strokes are most often found in the basal ganglia, internal capsule, or frontal lobe white matter, usually in areas served by small arteries.16,18,19 Diffuse or multiple white matter vasculopathic lesions, such as those observed in Binswanger's disease or leukoariosis, may also occur with or without clinically recognized cognitive impairment.20,21 Unrecognized large vessel strokes tend to occur in the nondominant hemisphere and/or to involve frontal, parietal, or occipital association cortex, often in regions that serve sensory functions represented bilaterally.16,18,22 Several different pathogenic mechanisms may underlie these diverse conditions, including atherosclerosis, arteriosclerosis, amyloid or lipohyalinoid angiopathy, embolic events, and possibly even chronic or recurrent cerebral hypoperfusion. Risk factors and preventive interventions influencing these diverse vasculopathic mechanisms may be specific or shared. Ferrucci and his co-workers found that the relationship of MSQ score with later stroke was reduced only marginally after controlling for age, history of hypertension, observed blood pressure, smoking, sex, race, education, and history of heart attack. Hypertension and diabetes appeared to be independently predictive of stroke regardless of MSQ score. These results imply that the mechanism underlying the link between poor cognitive function and future stroke is not strongly dependent upon a co-association with hypertension, diabetes, or any of the other controlling variables. This is a somewhat troubling finding because hypertension and diabetes have previously been identified as important risk factors for silent stroke and for lacunar stroke.23–25 It should be noted, however, that diabetes has not consistently been found to be a risk factor for lacunar strokes and that casual office blood pressure measurements may be associated less strongly with silent stroke than ambulatory or nocturnal measurements.26,27 Further, it has now been demonstrated in at least two different longitudinal studies that cross-sectional associations of hypertension with poor cognitive functioning tend to be substantially less strong than associations between mid-life hypertension and late life cognitive functioning.28,29 Although the most direct explanation for low MSQ score predicting future stroke is the one offered by Ferrucci, others are possible. The most difficult of these to examine involves a possible role for intermittent or chronic ischemia. Abnormal fluctuations in brain perfusion could be caused by a structural factor (as carotid stenosis), by changes in pulse rate and pressure over time, or by a failure of the mechanisms that ordinarily control perfusion of metabolically active brain regions.30,31 In early AD the increased metabolic activity ordinarily associated with performance of cognitive tasks is reduced or lacking.32,33 This could be a reflection of reduced neuron or synaptic density, a failure of neurons to undertake the assigned cognitive task, or a failure in the steps by which a cognitive task is assigned. Alternatively, it might reflect an impairment in the mechanisms that would normally maintain fully adequate blood flow to those regions activated metabolically for the performance of a neural tasks. Increased metabolic demand without corresponding increases in the delivery of glucose and oxygen (because of a failure of responsive increases in local perfusion) might lead to a metabolic stress that could in turn aggravate the Alzheimer process. If this hypothesized abnormality of perfusion regulation were also associated with an increased risk of subsequent stroke, it could provide explanation for the findings reported by Ferrucci. Because we ordinarily think of Alzheimer's disease and vascular dementia as pathogenically unrelated conditions, it seems unlikely that poor cognitive function due to the Alzheimer process might predict strokes. However, a substantial part of the apparent independence of the two conditions is based on definitional factors since diagnostic criteria for Alzheimer's disease generally depend on the exclusion of other causes, including cerebrovascular disease. Such diagnostic exclusions better represent the needs of medical taxonomy than any natural force. There is no apparent pathophysiological reason to expect cerebrovascular and Alzheimer's processes to be mutually exclusive; nor is it reasonable to expect the clinical expression of either to be unmodified by the other. Indeed, possible associations between Alzheimer's disease and atherosclerosis have been reported, and there is a growing suspicion that co-existing cardiovascular or cerebrovascular conditions may hasten or aggravate the clinical manifestations of Alzheimer's disease, bringing cases to medical care somewhat earlier.34–36 Thus, it is possible that the subset of participants with low MSQ scores studied by Ferrucci et al. was enriched in individuals with co-existing Alzheimer process and silent stroke. This suggestion (i.e., modulation of the expression of the Alzheimer processes by subclinical cerebrovascular disease) implies that in some cognitively impaired persons without recognized cerebrovascular disease, interventions aimed at vasculopathic processes might ameliorate or delay clinical manifestations of the Alzheimer process. The Ferrucci analysis provides evidence that even in persons without clinically apparent strokes, vascular disease processes may be more important in late-life cognitive decline than usually believed. Their interpretations suggest avenues of preventive intervention that might substantially reduce community burdens of cognitive impairment and dependency. The only certain means for resolving these questions requires prospective, longitudinal epidemiological research in which a rather large sample of older persons receive careful evaluations of cardiovascular and neurological status, cognitive and sociobehavioral functioning, brain structure (neuroimaging during life, autopsy at death), and assessments of environmental and genetic and metabolic risk factors at intervals over a period of several years. Relative to the usual research grant, the cost of such studies is high indeed. The promised benefit is an understanding of how risk factors, vasculopathic processes, and degenerative brain disease processes interact in the pathogenesis of cognitive decline, stroke, dependency, and dementia. Only such understanding can provide a rational basis for effective public health strategies aimed at preventing these tragic conditions, and possibly even for preventing all aging-related cognitive decline. All of the developed and developing nations of the world currently face crises related to a scarcity of the monetary, material, and human resources needed to care for a growing population of older citizens who have lost their abilities to care for themselves, mostly because of strokes and/or dementia. To paraphrase A. Whitney Griswald, a past President of Yale University, we must compare the cost of learning how to prevent these tragic conditions with the cost of continued ignorance.