| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Wolfson Unit for Prevention of Peripheral Vascular Diseases (S.B.R., F.B.S., F.G.R.F.), Public Health Sciences, Medical School, University of Edinburgh, Edinburgh, Scotland; Department of Psychology (I.J.D., M.C.W.), School of Philosophy, Psychology and Language Sciences, University of Edinburgh, Edinburgh, Scotland.
Address correspondence and reprint requests to Snorri B. Rafnsson, Wolfson Unit for Prevention of Peripheral Vascular Diseases, Public Health Sciences, Medical School, University of Edinburgh, Teviot Place, Edinburgh EH8 9AG, UK. E-mail: S.B.Rafnsson{at}ed.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: The Edinburgh Artery Study is a population cohort study of men and women who were recruited to a baseline survey in 1987 and 1988. From the time of study entry, the participants have been invited to two follow-up clinical examinations and continuously monitored for major fatal and nonfatal vascular events. All alive and eligible subjects were invited for cognitive testing in two study years when the mean age of the sample was 73.1 (standard deviation = 5.0) years. A follow-up cognitive assessment was performed in 2002 and 2003 on 452 survivors.
Results: In multivariate analyses controlling for demographic characteristics, depression, and major atherosclerotic risk factors, stroke was associated with a significantly worse performance on tests of verbal memory (p = .02) and letter fluency (p = .002). In addition, stroke was related to a significantly steeper 4-year decline in verbal memory performance (p = .04). Among the subjects who had not had an overt stroke, those with symptomatic peripheral arterial disease experienced a significantly greater 4-year decline in verbal memory functioning (p = .04).
Conclusions: In older people, stroke is associated with both worse performance on cognitive tests and progressive verbal memory decline. Elderly individuals with vascular diseases other than stroke may also be vulnerable to a greater decline in verbal memory function. A relationship between vascular diseases and verbal memory decline may exist independently of depressed mood and major atherosclerotic risk factors.
Key Words: cognitive decline • cardiovascular diseases • elderly population • atherosclerosis • aging
Abbreviations: CVD = cardiovascular disease; PAD = peripheral arterial disease; IC = intermittent claudication; MI = myocardial infarction; WHO = World Health Organization; ECG = electrocardiogram; MRI = magnetic resonance imaging; LMT = logical memory test; RPM = Raven’s progressive matrices; VFT = verbal fluency test; DST = digit symbol test; NART = national adult reading test; GCF = general cognitive factor; SD = standard deviation; MMSE = mini-mental state examination; CHD = coronary heart disease; SBP = systolic blood pressure.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
The term "vascular cognitive impairment" was proposed to describe mild-to-severe cognitive impairment of a primarily vascular basis (9). To date, however, few have investigated comprehensively neuropsychological functioning in cognitively intact individuals with vascular diseases. Specifically, longitudinal data on the pattern of cognitive decline in elderly people with different clinical types of CVD are lacking. For example, cross-sectional data show reduced performance on both global (10,11) and domain-specific tasks (12) in stroke patients. Stroke may also be associated with an increased decline in global cognitive function (13,14). In particular, in the Cardiovascular Health Study, stroke at baseline was related to a steeper decline in performance on both mental status and information processing speed tasks after incident events during follow-up were controlled (15). At present, however, it is not clear whether stroke is associated with a progressive decline in either global or specific cognitive abilities, and whether such a relationship exists independently of concomitant vascular risk factors (16).
Comprehensive data on cognitive functioning of individuals with major vascular pathology other than stroke are currently limited. However, widespread cognitive deficits have been reported in hospitalized patients with advanced cardiac disease (17,18). In population studies, coronary heart disease (CHD) has cross-sectionally been associated with worse performance on both mental status tests (19,20) and measures of specific cognitive functions (21), although findings are conflicting with respect to the role of CHD in progressive cognitive decline (15,22). Moreover, generalized cognitive deficits similar in severity to that observed for patients who have suffered a stroke have been noted in selected samples of patients with advanced peripheral arterial disease (PAD) (23,24). In the elderly population, PAD is associated with poorer performance on domain-specific cognitive measures (19,20). Other studies have linked PAD with a progressive decline in global mental status but longitudinal data on specific cognitive functions have so far been scarce (25,26).
To further investigate the pattern of cognitive decline in elderly people with CVD, we present data from the Edinburgh Artery Study, a population-based cohort study. The current investigation has the following features: a) a random population sample, b) validated measures for the determination of CVD, c) assessments of major domains of cognition, d) the collection of longitudinal data on these cognitive functions, and e) the collection of data on potential confounders.
The aim of the present study was to examine cognitive performance and 4-year change in cognitive function in relation to different indices of CVD. In particular, we were interested in the neuropsychological performance of subjects with vascular diseases other than stroke. A further objective was to determine whether an association between atherosclerotic CVD and cognitive decline would exist independently of major vascular risk factors.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Each subject underwent two sets of procedures previously described (27). A questionnaire was administered, asking about sociodemographic (28), medical (29), and lifestyle characteristics. A blood sample from a fasting subject was drawn for biochemical analysis. Height and weight were measured, and a 12-lead electrocardiogram (ECG) and rhythm strip were recorded and coded using the Minnesota Code (30). Subjects were asked about a doctor’s diagnosis of diabetes and the use of injections or tablets for treatment of the condition. Nondiabetic subjects underwent a fasting oral glucose tolerance test. Blood pressure was measured using a random zero sphygmomanometer.
All alive and eligible participants were invited to 5- and 12-year follow-up examinations where most of the above procedures were repeated (31). Cognitive testing was first held in 1998 and 1999 and repeated 4 years later using the same protocol (Figure 1). On each occasion, all eligible subjects (those alive, deemed fit to participate by their general practitioners (GPs), and whose addresses were known) were invited to participate. A Lothian Health Ethics Committee granted approval for the project.
|
From baseline to cognitive testing in 1998 and 1999, 383 subjects died. A total of 1209 subjects were available for testing: 1103 subjects were invited, 59 withdrew, 19 others died, and 28 were excluded by their GPs. Of the 1103 subjects invited, 933 replied, 151 did not reply, and 19 letters were returned by the post office. Of those who replied, 740 accepted the invitation and 193 refused. Of those who accepted, 717 kept their appointments; seven did not attend; and 16 subsequently refused to be tested. In total, 717 (59.3%) subjects were cognitively tested. By midyear 2002, 601 (83.8%) of the 717 subjects were still alive; of these, 23 were excluded by their GPs. Four more subjects could not be located. Therefore, 574 (95.5%) subjects were found eligible and were sent an invitation. Of these, 101 refused participation and 13 failed to reply. In total, 460 (76.5%) subjects accepted the invitation and received an appointment. An additional seven participants withdrew from the study and one could not be contacted. As a result, 452 subjects (75.2%) were tested; 280 at the University and 172 at home.
Ascertainment of Cardiovascular Disease
CVD was determined on the basis of data collected at baseline, at the follow-up examinations, and through continuous follow-up (Figure 1). At baseline, subjects were asked if they had ever had a doctor’s diagnosis of angina, myocardial infarction (MI), or stroke. In addition, the World Health Organization (WHO) questionnaire for CVD was administered (29). Participants were then followed up for 5 years. As described previously (31), the following methods were used to ascertain incident CVD: subjects were sent annual questionnaires, general practices flagged patient records, and annual printouts of hospital discharges were obtained from the Scottish Office Home and Health Department. Finally, information on patient attendance to a peripheral vascular clinic in Edinburgh was supplied. At the end of the 5-year follow-up, subjects were invited to a second clinical examination where they were asked about new CVD in the preceding year as well as completing the WHO questionnaire. Subjects were followed up for an additional 7 years during which time the same methods were used to detect incident CVD. At the end of this period, subjects were invited to a 12-year follow-up examination.
The following criteria were used to establish CVD at any point from baseline to a cut-off date 6 months before cognitive testing in 2002:
For the present analysis, the following CVD categories were constructed: any CVD, angina, IC, MI, and stroke. In comparison, subjects not meeting study criteria for any of the above during the study period were considered free of CVD.
Assessment of Cognitive Function
The following cognitive tests were administered: a) The logical memory test (LMT) was used for assessing verbal declarative memory (32). It comprises two stories that are read out to the subject. Immediate and delayed recall was assessed. Subjects were asked to recall the two stories. In the present study, the total of immediate and delayed recall was used. b) The Raven’s standard progressive matrices (RPM), which was used for assessing nonverbal reasoning, comprises logical pattern problems of increasing difficulty where a part has been removed (33). One point was given for each correctly completed pattern problem. c) A phonemic verbal fluency test (VFT), used for assessing executive functioning, is based on the number of words produced orally in response to a stimulus (34). Three 1-minute word-naming trials were used whereby subjects were asked to name as many words as they could think of beginning with the letters C, F, and L. d) the digit symbol-coding test (DST) was administered for the assessment of processing speed (35). Subjects were asked to complete as many number-symbol pairs as possible in 90 seconds.
Adjustment Variables
The following were considered as possible confounders of the relationship under study:
a) Age at the time of cognitive testing in 2002 and 2003, and gender.
b) Highest level of education completed at baseline.
c) Depressed mood in 2002 and 2003, which was assessed using the Hospital Anxiety and Depression Scale (36).
d) Peak prior cognitive ability, which was estimated in 2002 and 2003 using the national adult reading test (37). The NART requires the subject to pronounce 50 single words that are irregular with respect to common rules of pronunciation stress in order to minimize the likelihood of reading by phonemic decoding rather than word recognition.
e) Baseline levels of smoking, alcohol intake, and exercise. Pack-years of smoking were estimated by multiplying the number of 20-cigarette packs smoked per day with the number of years as a smoker. A zero value was entered for life-long nonsmokers. Alcohol intake was based on the number of alcohol units consumed in a typical week (38). Physical activity was grouped into four levels of intensity according to type of activity.
f) Baseline body mass index was defined as weight in kilograms divided by the square of the height in meters.
g) Diabetes at baseline was based on the subject’s recall of doctor’s diagnosis of diabetes, current treatment for diabetes, or 2-hour blood glucose concentration 
11.1 mmol/L. Glucose levels ranging from 7.8 to 11.0 mmol/L were used to denote impaired glucose tolerance.
h) Systolic blood pressure (SBP) at baseline.
i) Blood lipids in fasting subjects at baseline (serum total cholesterol and triglycerides).
Statistical Analysis
A n + 1 logarithmic transformation of depression scores was used in multivariate analyses. The distributions of pack-years of smoking and triglyceride levels were skewed, and as a result, the square root was used throughout.
Principal axis factoring was applied to identify a general cognitive factor (GCF), representing the cognitive test variance common to the following neuropsychological measures: LMT, RPM, VFT, and DST. Scores were saved on the first unrotated factor. Based on the presence of eigenvalues 1.0 and an examination of a scree plot, one factor which accounted for approximately 57% of the variance was found, thus validating the use of a general factor.
Pearson’s r correlation coefficients were computed to assess the strength and direction of associations between cognitive tests. The t test was used statistically to evaluate the 4-year change in cognitive performance in each vascular category. Multiple linear regression analysis was used for comparing cognitive test scores in 2002 and 2003 in diseased and nondiseased subjects, and in assessing the effects of CVD on the 4-year raw change in cognitive test scores (from 1998 and 1999 to 2002 and 2003), at the same time adjusting for possible confounders. Specifically, in the modeling approach used, adjustment variables were introduced into the regression models in two cumulative steps. In the first step, age and gender were controlled for in all models. Then, in a subsequent step, peak prior ability as indexed by the NART, education level, depression symptoms, and selected vascular risk factors were additionally adjusted for in the modeling of cognitive performance in 2002 and 2003. In the analysis of the 4-year change in cognitive function, the fully adjusted multivariate model included all of the above covariables apart from the NART. All tests were two-tailed and a two-sided p < .05 was taken to indicate statistical significance. All analyses were performed using SPSS 13.0 for Windows (39).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
|
Of the 452 subjects tested in 2002 and 2003, 441 had complete data on the LMT in 1998 and 1999 and 2002 and 2003. Complete data on the RPM, VFT, and the DST were provided by 427, 442, and 407 subjects, respectively. Noncompleteness was mainly due to hearing and visual difficulties rather than test refusal. In 1998 and 1999, the following mean ± standard deviation scores were obtained for those who had complete data on both occasions: LMT = 37.1 ± 13.5; RPM = 35.5 ± 8.9; VFT = 39.6 ± 12.7; DST = 41.7 ± 10.7. Similarly, in 2002 and 2003, the following results were observed: LMT = 38.5 ± 15.7; RPM = 32.6 ± 10.0; VFT = 37.8 ± 12.9; DST = 36.9 ± 11.4. Bivariate correlations showed high stability of individual differences in test performance over the 4-year interval, with coefficients ranging from 0.71 (p < .001) for the LMT to 0.85 (p < .001) for the DST.
In total, 169 (37.4%) tested subjects met the study criteria for one or more CVD. More specifically, 113 (66.9%) subjects met the diagnostic criteria for angina (40 subjects were diagnosed after baseline cognitive testing), 62 (36.7%) for IC (13 cases were diagnosed after baseline cognitive testing), 53 (31.4%) for MI (12 subjects had a first diagnosis of MI after baseline cognitive testing), and 20 (11.8%) for stroke (five subjects had a first diagnosis of stroke after baseline cognitive testing). Over half of all the cases had more than one CVD manifestation. This was reflected in the nonmutual exclusiveness of some of the CVD groups constructed for the present analysis. An exception was angina, which was kept distinct from the more severe CVD conditions, allowing the study of cognitive function in relation to a relatively mild, nonacute form of atherosclerotic CVD.
All groups performed worse on most measures in 2002 and 2003 relative to 1998 and 1999 (Table 2). In particular, statistically significant decline on the RPM, VFT, and DST, occurred in both CVD and non-CVD subjects. Significant decline on the RPM was found for all groups apart from angina. All CVD groups declined in DST performance and in general cognitive function (not in those with angina). Similar patterns, yet mostly nonsignificant, were also noted for the LMT and the VFT.
|
The relationship of CVD with cognitive performance in 2002 and 2003 was assessed in a series of regression analyses (Table 3). Only the main effects of CVD are reported because there was no evidence of gender by disease interaction in cognitive function in the present sample (data not shown). After adjusting for age and gender, a trend toward worse performance was observed for most cognitive measures across the majority of disease categories relative to subjects without CVD. In particular, lower LMT scores were seen in all vascular groups although only among subjects with stroke did the difference reach statistical significance (p = .02). Similarly, lower RPM scores were found in those with evidence of any CVD (p = .004), IC (p = .005), and MI (p = .001). Only subjects with IC (p = .02) and stroke (p = .008) had lower VFT scores. Apart from the angina category, all the CVD groups demonstrated significantly lower DST and GCF scores. Furthermore, most of the associations between vascular disease and cognitive performance were attenuated so they no longer remained significant when further adjusted for prior cognitive ability, current depression symptoms, education level, and major vascular risk factors. However, stroke continued to be associated with worse performance on two cognitive measures, LMT (p = .02) and VFT (p = .002), respectively.
|
Results from the analyses of change in cognitive test performance in relation to vascular disease are presented in Table 4. In the age- and gender-adjusted models, subjects with evidence of any CVD (p = .01), and those with PAD (p = .01) and stroke (p = .02) experienced a significantly greater decline in LMT performance over the 4-year follow-up period compared with subjects without any major vascular disease. Moreover, a significant decline in performance on the RPM was observed in subjects with MI (p = .03) relative to nonvascular controls. In contrast, no associations with other cognitive outcomes were noted. After further adjustment for depression symptoms at the time of follow-up cognitive testing, level of education, and CVD risk factors, most associations were attenuated so they no longer proved significant. On the other hand, both IC (p = .04) and stroke (p = .04) continued to exert independent effects on the rate of 4-year decline in verbal memory.
|
In a final step to determine if prevalent stroke was related to progressive cognitive decline, we repeated the above analyses of 4-year change in cognitive test scores after excluding the five subjects who were diagnosed with stroke during follow-up (data not shown). After the exclusion of incident cases, stroke no longer remained significantly associated with cognitive decline although the presence of negative regression coefficients in most instances suggested that this might be the case, despite the inability of the present analysis to show this statistically.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Of the individual vascular categories examined, stroke turned out to be independently related to both worse verbal memory and letter fluency performance, and further predicted an increased 4-year decline in verbal memory. In subsequent analyses, where incident stroke cases were deliberately excluded, the results pointed in the direction of negative effects of stroke on cognitive decline although these turned out to be statistically nonsignificant. These observations did, however, hint at the possibility that prevalent stroke could lead to progressive, ongoing cognitive decline, a notion which also has received some support in previous studies, although further research using larger samples is needed. For example, the Zutphen Elderly Study (13) noted a >3-year decline in mental status (mini-mental state examination (MMSE)) during follow-up in elderly men with stroke at baseline but domain-specific tests data were not provided. Also, the Cardiovascular Health Study reported a progressive yet modest decline in both MMSE performance and psychomotor speed over a 7-year period in relation to baseline stroke (21). Because each of the cognitive measures used in the above studies is likely to reflect the integrity of the underlying neural substrate in a nonspecific way, only the use of detailed cognitive testing can assist in outlining the pattern of cognitive decrements after stroke.
In the present analysis, angina was neither related to cognitive performance nor decline, although there was a tendency toward greater decline in several tests. In contrast, the Whitehall II study reported cross-sectional effects of angina on several measures of cognition in middle-aged and elderly civil servants (20). Possibly, the almost four times as many subjects with angina in this latter study provided enough statistical power to detect subtle cognitive decrements associated with such milder CHD. On the other hand, significant age- and gender-adjusted effects of MI on decline in abstract reasoning were observed, which continued albeit nonsignificantly in the fully adjusted models. In an earlier report, ECG-diagnosed MI was associated with a shift in the distribution of MMSE scores in the Rotterdam Study (18). In the Whitehall II study, and in contrast to other manifestations of CVD, the effects of MI were restricted to verbal reasoning and measures of executive functioning (20).
Further to the above, an association of PAD with steeper decline in verbal memory was found independently of major vascular risk factors. However, judging from the regression coefficients for some of the other cognitive measures, it is possible that additional aspects of cognition might also be affected despite our inability to demonstrate this statistically. Diffuse deficits in cognitive functioning have been observed in the context of symptomatic PAD, such as in short-term memory, verbal and nonverbal reasoning, word recognition and knowledge, and executive functioning (20). Similarly widespread, yet nonsignificant, cross-sectional patterns were also noted in older men in the Caerphilly study (19).
In general, our findings add to those from a growing number of studies. To date, however, few studies have presented longitudinal data on different aspects of cognition in relation to different indices of vascular diseases. Moreover, our data suggest nonspecific involvement of cognitive function but intercorrelations of neuropsychological tests make it challenging to pin down what abilities in particular may be disproportionately affected in vascular cognitive decline. For this reason, we included a factor representing the variance common to the individual cognitive tests in our analyses. In the age- and gender-adjusted models of cognitive performance, all vascular manifestations except angina were associated with lower general factor scores. Also, we further minimized the likelihood of residual confounding by peak prior ability by further controlling for performance on the NART rather than on only level of education as most previous studies have done.
The participants in our study were tested on two separate occasions. The observed decline in cognitive performance might be an underestimate of the true extent of change, given the possibility of practice effects. In particular, this might be the case for the memory task we used. However, this does not affect the individual differences in change over time, which is the data that were modeled here. Finally, the tested sample had a slightly more favorable demographic and vascular risk factor profile, suggesting that our findings may be an underestimate of the true relationship between CVD and cognitive decrements in the elderly.
In contrast to previous studies, we further controlled for major atherosclerotic risk factors. Our relatively conservative analytic strategy also took account of mood at the time of cognitive testing, which may adversely affect performance on cognitive tests (40). As a result, our findings show that, at least to some extent, the relationship under study may actually depend on the underlying vascular pathology. In stroke, the effects on the brain are direct but the ability to either tolerate or recover from such an injury may be negatively affected by preexisting risk factors (41). Although the progressive verbal memory decline associated with stroke in the present study might have occurred from the infarction itself, it could also have been attributed to a slow progressive damage as a result of cerebral vascular pathology of which stroke would be but one marker. In the absence of stroke, vascular cognitive decline might be explained by several separate or overlapping mechanisms. For example, CVD and cognitive dysfunction may both be manifestations of systemic atherosclerosis. The diffuse nature of atherosclerosis is well recognized (42) and both significant intra- and extracranial disease increases the risk of acute stroke, clinically silent cerebral infarcts, and other vascular-related cerebral pathology (43). Specifically, in patients with symptomatic peripheral vascular disease, evidence of atherosclerosis in both the coronary and cerebral circulation can be found in as many as 50% to 90% of cases (44). In addition, the prevalence of carotid atherosclerotic disease may be increased several fold. Even in the absence of overt stroke, atheromatous lesions in either the carotid or intracerebral arteries may be the source of multiple, clinically silent emboli that could induce progressive ischemic changes in neurocognitive functioning.
Several methodological issues need discussion. Specifically, the number of subjects with CVD in our study was modest, notably those with stroke. As a result, we may have lacked power to detect statistically significant associations between stroke and change in some cognitive outcomes. The cases comprised those who at any point from 1987 to 1988 onwards were found to meet the predetermined criteria for major CVD and stroke. Information on CVD came from a combination of sources, including baseline and follow-up questionnaires and clinical examinations. As an example, nine subjects with stroke were diagnosed on the basis of a previous doctor’s diagnosis of the event at either baseline or 5- or 12-year clinical examination. Another nine subjects met the criteria for a definite stroke (onset of symptoms<48 hours with duration of >24 hours during follow-up or a positive brain scan), and an additional two subjects had a possible stroke based on hospital discharge diagnosis of the event. The use of different diagnostic criteria at different points in the study might have increased the potential for misclassification under certain circumstances. For example, where the diagnosis of stroke was solely based on clinical symptoms or on a subject’s own recall of a doctor’s diagnosis of stroke, some subjects who did not have true acute cerebrovascular disease despite symptoms might have been falsely identified as cases. On the other hand, it is also possible that some persons falsely recalled having ever been told by a doctor that they had suffered a stroke. In any case, any misclassification whereby a truly healthy individual had been categorized as having stroke would have lowered the overall morbidity level of the stroke group.
A limitation of the current study is the lack of brain imaging data on our participants. As a result, information on different pathological aspects of the stroke, including etiological type, lesion site, and severity were not available to us. However, our estimation is that the majority of strokes in our study were thromboembolic in origin (45). Because the severity of stroke may be strongly related to early mortality and nonparticipation, the current sample is moreover likely to comprise survivors with mild-to-moderate stroke pathology. Also, the fact that we used an interval of 6 months between the diagnosis of stroke and follow-up cognitive testing may further have reduced the possibility of severe strokes being included in the study.
Similarly, we were unable to determine the presence of asymptomatic cerebral lesions in our participants. Magnetic resonance imaging-based deep subcortical lesions are common in stroke patients and are also frequently observed in neurologically intact older people (46). These are correlated with both cognitive impairment and dementia. To the extent that such unrecognized lesions existed in our nonvascular controls, any differences between these and the vascular groups in the level of cognitive decline might have been reduced.
A final limitation is the fact that only 452 participants of the original 1592 cohort provided longitudinal data on cognitive function. This group most definitely comprises healthy survivors who also might be expected to show less cognitive morbidity. Even so, the present study underscores the importance of assessing cognitive functioning of older survivors with vascular diseases and the need for halting or preventing additional cognitive deterioration in this population. However, further research is needed where the mechanisms underlying cognitive decline in elderly vascular patients are better clarified and the structural changes that may be responsible for the observed patterns of cognitive decline are examined. In this context, the administration of a number of tests for each major cognitive domain, particularly executive function, must be considered as opposed to the use in the present study of only a single marker of each principal domain. Moreover, the impact of subtle cognitive deficits on activities of daily living needs assessing, as must the extent to which decline in specific cognitive abilities may be predictive of further cognitive decrements in this population.
We acknowledge the financial support from the British Heart Foundation, and all subjects, staff, and general practitioners involved in the Edinburgh Artery Study.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
DOI:10.1097/psy.0b013e318068fce4
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
4 in modulating effects of other risk factors for cognitive decline in elderly persons. JAMA 1999;282:40–6.4 are independent risk factors for cognitive decline. A population-based study. Stroke 2000;31:2431–6.This article has been cited by other articles:
 |
|
 |
|
  M. Luciano, R. E. Marioni, A. J. Gow, J. M. Starr, and I. J. Deary Reverse Causation in the Association Between C-Reactive Protein and Fibrinogen Levels and Cognitive Abilities in an Aging Sample Psychosom Med, May 1, 2009; 71(4): 404 - 409. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Johnson, I. J. Deary, M. McGue, and K. Christensen Genetic and Environmental Links Between Cognitive and Physical Functions in Old Age J Gerontol B Psychol Sci Soc Sci, February 10, 2009; (2009) gbn033v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. B Rafnsson, I. J Deary, and F. Fowkes Peripheral arterial disease and cognitive function Vascular Medicine, February 1, 2009; 14(1): 51 - 61. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
