Cognitive frailty: Predementia syndrome and vascular risk factors

Introduction

In older subjects, particularly over age 85 years, the prevalence of vascular factors and other medical conditions that impair cognition increases substantially including chronic cardiovascular, respiratory, and metabolic diseases, and severe sensory deficits [27]. Not all causes of cognitive decline and dementia are known, but some studies have suggested that these conditions may be prevented [47,49]. Previous work has linked cognitive decline to cardiovascular disease, diabetes mellitus, hypertension, and high or low blood pressure [33]. Various genetic and non genetic factors known to increase the risk of vascular disease, including apolipoprotein E (APOE) and angiotensin I converting enzyme 1 (ACE1) genotypes, total cholesterol (TC), lipoprotein(a) [Lp(a)], diabetes mellitus, atrial fibrillation, hypertension, serum APOE levels, and atherosclerosis, have been evaluated as risk factors for Alzheimer's disease (AD) and vascular dementia (VaD) [33]. Subtle cognitive decline can precede the appearance of symptomatic AD by many years [29], so that current clinical research has focused on the identification of early diagnostic indices of dementia [39]. The presence of a transitional phase between early mild cognitive decline and symptomatic dementia raises problems in diagnosis, because it is difficult to establish whether mild cognitive decline will result in dementia within a short time 0197 [19], or in some cases be a distinct and stable clinical entity that will not worsen with time [6].

In the present review we will use the term "predementia syndrome" to identify every age-related deficit in cognitive function supported by the literature, including both concepts used to describe an intermediate stage of cognitive impairment more based on a normality model and concepts considered as a pathological condition and therefore a risk state for dementia. The aims of this review were to examine in depth the possible role of vascular risk factors in modulating the risk of age-related changes in cognitive function, and the progression of predementia syndrome to dementia of degenerative or vascular origin. We reviewed clinical and epidemiological studies from the international literature in English language through keyword (vascular risk factors, progression to dementia, mild cognitive impairment, aging-associated cognitive decline, age-associated memory impairment, age-related cognitive decline, mild cognitive disorder, and cognitive impairment no dementia) and author searches in Medline from January 1986 to August 2004. We reviewed only population-based studies in which vascular risk factors were evaluated in relation to predementia syndrome and its progression to clinically overt dementia.

Predementia syndrome: transitional stage between normal aging and dementia

Different diagnostic criteria and terms have been proposed over the years to describe predementia syndrome in the elderly [3,12] (Table 1). One of the earliest terms introduced in the attempt to discriminate normal from pathological memory decline was "benign senescent forgetfulness" (BSF) proposed by Kral [25]. To characterize age-related changes in memory more precisely, a National Institute of Mental Health (NIMH) Working Group in 1986 proposed research criteria for "age-associated memory impairment" (AAMI) [8]. However, the criteria of AAMI do not address progression, and so do not clearly discriminate between subjects who will develop dementia and those who will remain cognitively stable. The utility of AAMI has therefore been questioned [4]. To address this, Blackford and La Rue proposed three subtypes of objective memory impairment, including an AAMI subtype and two others: age-consistent memory impairment, ACMI, and late-life forgetfulness, LLF, which added an age criterion [4] (Table 1).

In addition to AAMI, various other constructs for the classification of age-related cognitive changes have been introduced. In 1994, a task force of the International Psychogeriatric Association (IPA) in collaboration with the World Health Organization (WHO) proposed diagnostic criteria for "aging-associated cognitive decline" (AACD) [28]. The subject groups identified by the AACD criteria are clinically and experimentally less heterogeneous than those identified by the AAMI criteria. In the AACD criteria there is no age restriction: though the cognitive decline is more prevalent in old age, its onset may occur earlier in life (Table 1).

Other proposed diagnostic classifications include "agerelated cognitive decline" (ARCD) [2] and "mild cognitive disorder" (MCD) [56]. ARCD is defined by the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) as "an objectively identified decline in cognitive functioning consequent to the aging process that is within normal limits given the person's age" [2]. MCD is included in the research criteria for International Classification of Disease-10 (ICD-10) [56]. Although the definition of MCD is quite different from that of Mild Neurocognitive Disorder reported by DSM-IV [2], these two categories share many similarities [7] ( Table 1).

The Canadian Study of Health and Aging (CSHA) introduced the category of "cognitive impairment no dementia" (CIND) to classify all individuals with cognitive impairment that do not meet the criteria of dementia [14,15]. This category includes subjects with impairment caused by medical or psychiatric conditions (delirium, chronic alcohol and drug use, psychiatric illness, depression, and mental retardation), and offers the advantage of categorizing several broad but specific categories of cognitive impairment that include but are not limited to memory (Table 1).

Finally, the currently popular term "mild cognitive impairment" was first operationally defined in 1991 by the New York University group based on the Global Deterioration Scale (GDS) [41] and redefined shortly thereafter at the Max Planck Institute in Germany [57], based on the DSM-III-R (third edition revised) [1] and the ICD-10 criteria [56]. At present, the term Mild Cognitive Impairment and its acronym MCI have been frequently used in studies on the preclinical phases of dementia, although with different and inconsistent definitions [16,32,37,38,46]. MCI is a clinical label that includes nondemented aged persons with memory impairment and no significant functional disability (Table 1).

There is now ample evidence that MCI is often a pathology-based condition [37,40], and, therefore, may be amenable to pharmacological treatment. Thus, factors that can predict the conversion of MCI to clinically diagnosed dementia are sought. In contrast, ARCD and AACD describe normal aging, without progression to dementia, and so lack this utility. In 2001 at the "Current Concepts in MCI Conference" it was suggested to subdivide MCI to encompass the heterogeneity of MCI patients (similar to the CIND concept) [40]. Three subsets of MCI were proposed: amnestic-MCI (aMCI), multiple domains slightly impaired-MCI (mdMCI), and single non-memory domain-MCI (snMCI). The clinical heterogeneity of MCI was intended to parallel the heterogeneity of different types of dementia, but risks disintegrating into the same terminological and nosological chaos as its predecessors. Currently, aMCI as originally proposed by Petersen et al. remains the most widespread and validated "subtype" with high positive predictive value for eventual "conversion" to AD [31,37]. Table 1 Predementia syndrome: an overview of terms, definitions, and operational clinical criteria Term/authors Definition/criteria "Benign senescent forgetfulness" (BSF), Kral [25] Memory deficit in elderly characterized by the inability to recall on certain occasions relatively unimportant experiences of the past, equally frequent in both sexes, expression of senium naturale opposed to senium ex morbo in "malignant senescent forgetfulness"

"Age-associated memory impairment"(AAMI), Crook et al. [8] Form of benign senescent forgetfulness characterized by the following criteria:

Inclusion criteria Age over 50 years; Complaints of memory loss reflected in such everyday problems; Memory test performance at least 1.0 SD below the mean established for young adults on a standardized test of secondary memory (recent memory) with adequate normative data; Adequate intellectual function as determined by a scaled score of at least 9 on the Vocabulary subtest of the WAIS; Absence of dementia as determined by a score of 24 or higher on the MMSE AAMI (modified) Inclusion criteria Age-consistent memory impairment (ACMI)

Age between 50-79 years;

Late-life forgetfulness (LLF), Blackford and La Rue [4] Complaints of memory loss reflected in everyday activities, attested by self-administrated and standardized memory questionnaires.

Memory test performance are characterized by the following criteria: AAMI, test performance at least 1.0 SD below the mean of younger adults in a well-standardized test; ACMI, test performance 1.0 SD below the age appropriate mean in at least 75% of administrated tests; LLF, test performance 1-2 SDs below the age specific mean in at least 50% of administrated; Verbal and performance IQ scores between 90 and 130 on WAIS or WAIS Revised.

"Ageing-associated cognitive decline" (AACD), Levy [28] Inclusion criteria No age restriction; A report by the individual or a reliable informant that cognitive function has declined; Onset of decline must be described as gradual and must have been present for at least 6 months; Difficulties in any one of the following areas: memory and learning, attention and concentration, thinking (e.g., problemsolving, abstraction), language (e.g., comprehension, word finding); visuospatial functioning; Abnormality of performance on quantitative cognitive assessments (for example, neuropsychological tests or mental state evaluations) for which age and educational norms are available for relatively healthy individuals. Performance must be at least 1.0 SD below the mean value for the appropriate population "Age-related cognitive decline" (ARCD), DSM IV [2] It represents cognitive changes that are within normal limits given the person's age.

There are not operational definitions, but the following criteria should be met to make a diagnosis of ARCD (33)

Vascular risk factors in predementia syndrome

Recent evidence on the clinical course of predementia syndrome and particularly MCI, showed that not all MCI patients have memory impairment as the predominant symptom [40], and not all individuals with MCI progress to AD [40,26,50,52]. In a recent longitudinal study with a 3.5-year follow-up, the Italian Longitudinal Study on Aging (ILSA), involving a total of 2963 individuals from age 65 to 84, freeliving or institutionalized, we found a prevalence rate of 3.2% for MCI and a progression rate to dementia of 3.8/100 personyears. [50]. In this study, MCI patients were selected on the basis of the memory loss, and so probably included many patients with aMCI, with a neuropsychological pattern more similar to AD than VaD. The breakdown in AD, VaD, and other types of dementia in the ILSA was 2.3/, 1.3/, and 0.3/100 person-years, and among those who progressed to dementia, 60% progressed to AD and 33% to VaD [50]. This reflects the heterogeneity of MCI [40]. If MCI is essentially incipient AD [32], then factors predictive of AD should bear a similar relationship to MCI and both vascular factors and APOE 4 allele, as a genetic vascular risk factor, have been associated with higher risk of AD [33]. If MCI is, alternatively, an unstable and heterogeneous category, few factors are likely to be found consistently associated with it. Our recent findings suggested the second hypothesis [50], but, at present, this question cannot be definitively resolved using current data. In fact, previous findings suggest that, in contrast with clinicalbased studies [38,40], where progression is more uniform, in population-based studies the MCI classification is unstable [18,50]. Thus, the apparent homogeneity of MCI and its progression to AD pathology in clinical-based samples may reflect referral patterns and selection criteria, suggesting that MCI is a heterogeneous descriptor and that the outcome at follow-up depends on which population is studied and how MCI is defined. In fact, persons who have been referred to a memory clinic are likely to be at a different stage of cognitive decline compared with the cognitively impaired individuals from community samples, because they noticed a memory problem and actively sought help for it.

Recent evidence from population-based studies [9,11,22,30,50,52] and case series [10,17] suggests that cerebrovascular disease (CVD) and vascular risk factors may contribute to the heterogeneity of MCI. In fact, vascular risk factors, such as hypertension, smoking, diabetes mellitus, atrial fibrillation, and the APOE 4 allele are also associated with late-life cognitive decline [33], and so may influence the development of MCI and dementia. Because such vascular risk factors may be modifiable, identification and subsequent management of these possible risk factors may help to prevent, and reduce conversion rates of MCI to dementia [29].

In the ILSA, the multivariable analysis suggested that as in AD only age was a risk factor for incident MCI, while higher levels of education and serum TC appeared to have a protective effect [50]. On the contrary, Kivipelto et al. found that midlife elevated TC serum levels (>6.5 mmol/L) increased the risk for MCI (OR, 1.9), with systolic blood pressure showing a similar trend [22]. In addition, De Carli et al. showed that elevated midlife blood pressure increased the risk for MCI [9]. We did not confirm that elevated TC is a risk factor for incident MCI [50]. This discrepancy can be explained by the fact that both TC and blood pressure were midlife determinations (with average follow-up of 21 and 25 years) in the De Carli et al. [9] and Kivipelto et al. [22] studies, respectively. Furthermore, in a recent study we found lower TC serum levels in AD [48], confirming the data of cross-sectional and prospective studies [24,43].

Additionally, in a cross-sectional evaluation in the same ILSA sample, age and myocardial infarction were positively associated with ARCD, while age, congestive heart failure, and stroke were associated with CIND [11], supporting the role of age and CAD as risk factors for cognitive decline. In the Cardiovascular Health Study (CHS) Cognition Study, the development of MCI was associated with measurements of cognition and depression, racial and constitutional factors, and CVD [30]. In some studies the presence of APOE-4 allele was associated with memory decline in subjects with cognitive impairment [13,21,53]. In the CSH Cognition Study, classifying patients as having either aMCI or mdMCI, the presence of 4 allele was associated to the aMCI only [30]. Other studies suggested that APOE genotype has only little impact on the degree of cognitive impairment [45]. Finally, a recent Finnish population-based study with a 3-year followup showed that higher age, and medicated hypertension are independent risk factors, but high education is a protective factor for MCI, while gender, elevated blood pressure, diabetes mellitus, APOE genotype, or CVD have no significant effect on the incidence of MCI [52]. The considerable amount of evidence above referenced may support the hypothesis that vascular risk factors may have an effect on the development of MCI, but further studies are needed to confirm these findings. The main population-based studies that evaluated vascular risk factors for predementia syndrome are shown in Table 2.

Vascular risk factors and progression to dementia

A number of studies have examined clinical and imaging predictors associated with increased risk to convert from MCI to dementia [40], and in the ILSA also vascular risk factors were evaluated [50]. Stroke was the only vascular risk factor associated with the progression of MCI to dementia [50]. In this study, possibly a follow-up period longer than 3.5 years could have revealed that other vascular factors influence the progression of MCI to dementia. In a recent study, the risk of poststroke dementia (PSD) was high (28.5% in a 3-year follow-up, with most patients during the first 6 months after stroke), with about one-third of patients meeting the criteria for AD and two-thirds meeting the criteria for VaD [20]. CVD is common in the elderly and is known to cause cognitive impairment [34]. However, it is far from clear that cerebral infarction predisposes to AD [44], or even increased one year conversion to dementia [51].

The likelihood of MCI, as conventionally defined [40], may also be increased by clinical and magnetic resonance imaging (MRI) evidence of CVD [9] and in association with pathological evidence of severe cerebral atherosclerosis [42]. The incidence of CIND rises in the first year after cerebral infarction [51]. Furthermore, considerable evidence has doc- umented that CVD is strongly associated with dementia in the absence of AD pathology and also may worse or accelerate cognitive decline in subjects with AD [54]. Given this evidence, it seems reasonable to postulate that the presence of CVD would contribute to incident dementia among individuals with MCI particularly due to vascular pathology. Nonetheless, in a recent case-series study, 52 MCI patients including many with symptomatic CVD were longitudinally evaluated for 3.1 years. Among the clinical and MRI measures, only hippocampal volume was associated with increased risk of progression to dementia. When neuropsychological measures were included in the analysis, memory and executive functions were associated with increased risk of progression to dementia, whereas APOE genotype, cerebrovascular risk factors, clinical stroke, presence or absence of lacunes, and extent of white matter hyperintensities did not predict progression [10]. In another study, 29 patients with MCI of the vascular type were compared with 14 with MCI of degenerative etiology based on the criteria of Petersen et al., following patients over 32 months. MCI patients of the vascular type had a neuropsychological profile characterized by poor performance on tests of executive functions and neurological features of parkinsonism without tremor, supporting the hypothesis that these patients have a distinctive clinical picture that can be identified in a clinical setting [17].

Therefore, whereas considerable epidemiological evidence suggests that cerebrovascular risk factors are associated with various types of dementia, including AD [22,23,33], only limited prospective data exist to support the posited impact of CVD on incident dementia [54,55]. In one of these studies [55], white matter lucencies (leukoaraiosis) assessed with computed tomography were associated with increased risk of progression to dementia in a group of 27 MCI patients. Linear measures of medial temporal lobe were also significantly associated with progression. The authors conclude that both cerebrovascular and AD processes likely contribute to dementia progression among cognitively impaired individuals. In a second study, with a larger cohort (1015 participants) and a 3-year duration of follow-up, a total of 30 individuals converted to dementia (approximately 1%/year) [54]. Prevalent silent lacunar infarction and incident silent lacunar infarctions were both associated with an increased risk to develop dementia or rapid cognitive decline, although this was particularly true for individuals with incident lacunar infarction [10].

Our findings on the role of stroke in the progression of MCI patients to dementia in the ILSA sample [50], in conjunction with these previous studies [54,55], suggest that stroke and CVD play a role in the clinical course of predementia syndrome. In contrast with the results of these studies, the study of DeCarli et al. raises questions about the role of CVD in the progression of MCI to dementia [10]. In fact, as reported above, although it is evident that CVD can be associated with cognitive decline [33,54] and is more prevalent among individuals with MCI [9,30], CVD did not appear to influence progression to dementia when memory impairment was severe [10], suggesting that only important cerebrovascular brain injury may have an impact on cognitive function.

Several reports of neuropsychological testing among individuals with MCI suggest that the severity of memory impairment strongly predicts progression to dementia [38,40], and associated deficits in language or impairments in other cognitive domains may contribute to a more rapid progression [5]. In addition to a neuropsychological assessment, as reported above, different approaches have been undertaken to predict the outcome of MCI, such as assessment of plasma ␤amyloid, structural and functional neuroimaging, and APOE genotyping. In particular, several gene polymorphisms associated with vascular disease have been found to increase susceptibility to sporadic AD [33]. A number of studies have looked at the impact of APOE genotype on the risk of progression of MCI to dementia. In a study carried-out in Rochester, the APOE 4 genotype strongly predicted progression of memory-impaired persons to AD [35]. Furthermore, Petersen et al. reported that APOE 4 allele status appears to be a strong predictor of subsequent development of dementia in patients with MCI [36]. In contrast, three very recent studies failed to find evidence of an association of the presence of the APOE e4 allele with progression to dementia in MCI patients or with incident MCI [9,18,52]. In fact, in a Finnish populationbased study elevated blood pressure, diabetes mellitus, APOE genotype, or CVD had no significant impact on the conversion to MCI [52], confirming the controversial findings on this vascular genetic risk factor in MCI. In the MoVIES study, progression from MCI to AD was associated only with age [18]; lack of power is a possible explanation, given the small numbers of persons who met criteria for MCI in these studies. These findings were confirmed by the case-series study of DeCarli et al. in which APOE genotype, clinical stroke, and other cerebrovascular risk factors hyperintensities did not predict progression [9].

Conclusions

Recently, clinical research has focused on the identification of different clinical conditions that may represent a very early (preclinical) symptomatic stage of dementing illnesses. Several studies have suggested that most of patients who met the MCI criteria will progress to AD, but rates of conversion to AD and dementia vary widely among studies, partly because of the characteristics of the population involved and length of follow-up. Many of the risk factors for CVD and VaD have also been shown to increase the risk of AD. If MCI is essentially incipient AD, then factors predictive of AD should bear a similar relationship to MCI and both vascular factors and APOE 4 allele have been associated with higher risk of AD. Nevertheless, recent evidence from a population-based study suggested that among MCI patients who progressed to dementia, 60% progressed to AD and 33% to VaD [50], reflecting the heterogeneity of MCI. Some recent studies further suggested that CVD or vascular factors increased risk of conversion from MCI to dementia. Although considerable epidemiological evidence indicates that cerebrovascular risk factors are associated with various types of dementia, including AD, only limited prospective information exists regarding the impact of CVD on incident dementia. Recent findings on the role of stroke in the progression of MCI to dementia, suggest a role of stroke and CVD in the natural history of predementia syndrome. Nevertheless, not all studies have shown an effect of CVD on the progression of MCI to dementia, suggesting that perhaps the main effect of cerebral infarction on the risk of post-stroke cognitive impairment is when the infarction itself causes it.