|Home | About | Journals | Submit | Contact Us | Français|
Stroke risk factors have been increasingly implicated in the development of age-related cognitive decline, the spectrum of vascular cognitive impairment, and, more recently, Alzheimer’s disease (AD). In addition, depression and the apolipoprotein (APOE) ε4 allele have been reported to influence the association between stroke risk and cognition. However, few studies have described the relations among stroke risk, cognition, and APOE genotype in AD, and the findings have been equivocal.
Thirty cognitively normal older adults, 30 AD patients with depression, and 30 AD patients without depression were administered a comprehensive neuropsychological battery measuring several domains including memory, attention, language, visuospatial skills, executive functions, and speed of information processing. The Framingham Stroke Risk Profile (FSRP), a validated scale that was developed to predict 10-year probability of stroke, was used to quantify stroke risk burden.
AD patients with depression demonstrated greater stroke risk burden relative to the cognitively normal group and, across all participants, increased stroke risk was associated with poorer performance on memory and processing speed measures. Moreover, stroke risk accurately predicted AD diagnosis. Notably, there were no significant differences in stroke risk or cognitive performance between the AD participants with depression and those without depression.
Given that many markers of stroke risk are modifiable or treatable, our findings have implications for assessment, prevention, and treatment of cognitive decline.
Increasing attention has focused on the contributions of cardiovascular risk factors for stroke to the pathogenesis and exacerbation of Alzheimer’s disease (AD). Research in this area has shown that stroke risk factors, such as diabetes (Ott et al., 1999) and hypertension (Skoog et al., 1996), are associated with increased risk for developing AD. Stroke risk factors are also associated with an accelerated rate of decline after establishment of an AD diagnosis (Mielke et al., 2007). Although the precise mechanisms linking stroke risk factors and AD are not fully understood, proposed theories suggest that (1) vascular pathology may have an additive effect by increasing the overall burden of pathology (Chui et al., 2006), (2) AD and vascular disease may interact to worsen pathologic effects (Honig et al., 2003), and (3) AD itself may be conceptualized as a vascular disorder with amyloid deposition linked to a breakdown in the blood-brain barrier and alterations in brain perfusion (Kalaria, 1997; for review, see Thal et al., 2008).
Stroke risk factors are also associated with cognitive decline in older adults who are free of dementia and clinical stroke. This relationship has been shown for executive functions (Brady et al., 2001), attention, visuospatial memory (Elias et al., 2004), and verbal episodic memory (Bangen et al., 2007). Although the mechanism underlying the relationship between increased stroke risk and cognitive decline remains unclear, greater stroke risk, in general, is associated with smaller brain volume in older adults without evidence of clinical stroke (Bangen et al., 2007). Moreover, stroke risk factors are associated with the development of white matter lesion pathology that is related to cognitive impairment across the aging spectrum (Delano-Wood et al., 2008a).
In addition to the relationship between the presence of stroke risk factors and cognitive decline, there is some evidence that vascular disease predisposes nondemented older adults to depressive symptoms (the ‘vascular depression’ hypothesis; Thomas et al., 2004). Depression is associated with neuropsychological impairment and increased risk of developing dementia (Thomas and O’Brien, 2008), and vascular disease may be a mechanism linking depression and AD (Luchsinger et al., 2008). However, the relationship among these variables is complex, and it is possible that vascular disease and depression contribute independently or synergistically to the development of cognitive decline and dementia (Alexopoulos, 2003).
Results from studies examining the associations among stroke risk, cognition, and depressive symptoms have been equivocal. For instance, two studies of adults with major depressive disorder (MDD) reported that greater stroke risk was associated with poorer cognitive performance (Smith et al., 2007; Steffens et al., 2007). In contrast, a longitudinal study by Luchsinger et al. (2008) demonstrated greater AD risk with increased depressive symptomatology in older adults who were not demented at baseline. Their results were unchanged after adjustment for vascular risk, indicating that the link between depressive symptoms and AD is likely not explained by a history of stroke risk factors.
Given these equivocal findings, there remains debate whether stroke risk or depression appreciably modify cognition either in mild cognitive impairment (MCI) or dementia. We therefore examined stroke risk and cognitive functioning in AD patients with and without depression, as well as in cognitively normal adults free of clinical stroke, in an effort to further clarify the associations among stroke risk, cognition, and depression. Given that the apolipoprotein E (APOE) ε4 allele is associated with increased risk for developing AD (Corder et al., 1993) and greater stroke risk burden (Ilveskoski et al., 1999), we also assessed the association between APOE genotype and stroke risk. We hypothesized that greater stroke risk would be related to poorer cognitive performance across AD and cognitively normal participants, and that stroke risk would be highest in AD patients with depression, intermediate in AD patients without depression, and lowest in cognitively normal older adults.
Sixty individuals with clinically diagnosed probable AD and 30 cognitively normal older adults (NC) were selected for this study from the larger cohort of research volunteers of the Alzheimer’s Disease Research Center (ADRC) at the University of California, San Diego (UCSD). The AD participants included 30 consecutively enrolled AD participants with depression who were selected for the present study from the larger cohort based solely on their diagnoses of AD and depression (see below for details). Thirty AD participants characterized as nondepressed and 30 cognitively normal older adults were selected from the larger cohort based on their demographic similarity to the depressed AD group on age, education, and sex distributions as well as study enrollment date. In addition, the two AD groups were matched on overall level of cognitive impairment as measured by the total score on the Mattis dementia rating scale (DRS; Mattis, 1976). Participants were selected without regard to race, ethnicity, socioeconomic status, or APOE genotype. The study was approved by the UCSD institutional review board and written informed consent was obtained from all participants or from caregivers.
The three groups did not significantly differ in terms of age (F(2,87) = 0.15, p = 0.86, ), education (F(2,87) = 2.52, p = 0.09, ), or women/men ratio (χ2 = 0.09, p = 0.96). As expected, the groups significantly differed in APOE genotype distribution (χ2 = 24.63, p <0.001). The NC group included five ε4 carriers (16.7%) and 25 non-ε4 carriers (83.3%). The nondepressed AD group included 17 ε4 carriers (56.7%) and 13 non-ε4 carriers (43.3%). The depressed AD group included 24 ε4 carriers (80.0%) and six non-ε4 carriers (20.0%; see Table 1).
Participants underwent a clinical evaluation that included neurological, neuropsychological, psychiatric, and medical examinations, as well as laboratory tests and APOE genotyping. Patients received a diagnosis of probable AD by two independent, senior staff neurologists according to the criteria developed by the National Institute of Neurological and Communicative Disorders and Stroke (NINCDS) and the Alzheimer’s Disease and Related Disorders Association (McKhann et al., 1984). Individuals were excluded from this study if they had causes of dementia other than AD, a history of severe head injury, substance abuse, or psychiatric disorders (with the exception of depressive disorders). In addition, participants with a history of stroke or transient ischemic attack (TIA) or who had a modified Hachinski score greater than 4 were excluded. Thus, individuals whose dementia may have had a significant vascular component were not included in this study.
All participants were evaluated for depression according to the protocol described elsewhere (Delano-Wood et al., 2008b). Briefly, the Diagnostic Interview Schedule (DIS), a DSM-based semi-structured interview for psychiatric diagnoses, was used. The DIS was administered by trained nurse practitioners who were highly experienced in the assessment of older adults with cognitive impairment. In addition, two board certified geriatric psychiatrists supervised the nurse practitioners’ evaluations and made the determination regarding the diagnosis of a depressive disorder. Responses to the DIS were obtained from both the patient and an informant/caregiver and any inconsistencies were resolved by medical chart review, clinical assessment, and consensus judgment. Of the 30 AD patients with a current diagnosis of co-occurring depression, 19 were diagnosed with dementia with depression, six were diagnosed with major depressive disorder, four were diagnosed with adjustment disorder with depressed mood, and one was diagnosed with atypical depression. Based on self or informant reports, 12 of the 30 depressed AD patients had a history of depression prior to their enrollment in the study. None of the 30 nondepressed AD patients had a prior history or current diagnosis of depression.
All participants received global cognitive screening (DRS) and the ADRC Core Neuropsychological Battery (for a detailed description of the tests that comprise this battery, see Salmon and Butters, 1992). Neuropsychological tests of interest included measures from six cognitive domains: memory, attention, language, visuospatial functioning, executive functioning, and speed of information processing. The conceptually-based inclusion of each measure within one of the six domains was supported by six separate principal components analyses, each of which yielded a single factor solution (eigenvalues ranged from 1.17 to 3.36). The raw scores for each of the 15 measures listed below were z-transformed using the means and standard deviations of the NC group. Then the z-scores for the individual tests were averaged to create a summary score for each cognitive domain. The specific measures that comprise each domain were as follows: (1) Memory was assessed with the Wechsler Memory Scale—Revised (Wechsler, 1987) Logical Memory Subtest (WMS-R; immediate and delayed free recall) and the California Verbal Learning Test (CVLT; Trials 1–5 total recall and long delay free recall; Delis et al., 1987); (2) Attention was measured with the Attention subscale of the DRS and the Digit Span subtest of the Wechsler Adult Intelligence Scale—Revised (WAIS-R; Wechsler, 1981); (3) Language was measured with the Boston Naming Test (Kaplan et al., 1983) and the WAIS-R Vocabulary subtest; (4) Visuospatial skills were assessed with the Block Design subtest of the Wechsler Intelligence Scale for Children—Revised (WISC-R; Wechsler, 1974) and the Construction sub-scale of the DRS; (5) Executive functioning was measured with the Wisconsin Card Sorting Test (WCST-48-card version; number of perseverative errors; Lineweaver et al., 1999), letter fluency (FAS), and Trail Making Test, Part B (Reitan and Wolfson, 1985); (6) Speed of information processing was assessed with the Trail Making Test, Part A (Reitan and Wolfson, 1985) and the Digit Symbol subtest of WAIS-R (Wechsler, 1981).
Using methods previously described by D’Agostino et al. (1994) stroke risk was quantified using the Framingham Stroke Risk Profile (FSRP). The FSRP is a validated scale that was developed to predict 10-year probability of stroke. The FSRP provides gender-corrected scores based on the following risk factors which were identified from 36 years of longitudinal data from the Framingham Heart Study: age, systolic blood pressure, diabetes mellitus, cigarette smoking, history of cardiovascular disease, atrial fibrillation, left ventricular hypertrophy as identified by electrocardiogram, and use of antihypertensive medications. Total allocated points ranged from 1 (a 10-year stroke probability of 1% in women and 3% in men) to 30 (a 10-year stroke risk of 84% in women and 88% in men). Given that no electrocardiograms were performed on participants, the presence of left ventricular hypertrophy could not be determined and was, therefore, not included in any statistical analyses. In addition to the variables composing the FSRP score, total cholesterol was also measured.
Correlational analyses were conducted to examine the associations among APOE genotype (participants were classified as either ε4 carriers or non-ε4 carriers), stroke risk, and cognition. Analyses of variance (ANOVAs) and χ2 tests were conducted to examine group differences for variables of interest. In cases where the omnibus F-test demonstrated statistical significance, Tukey’s Honestly Significant Differences (HSD) post hoc tests were conducted to determine which specific groups differed. Logistic regression was used to assess the contribution of stroke risk to diagnostic group classification. All analyses were performed using SPSS.
Correlational analyses revealed that, across the entire sample (i.e., NC subjects and AD patients), greater stroke risk was significantly associated with poorer performance on memory (r = −0.22, p = 0.02) and processing speed (r = −0.22, p = 0.02) measures (see Figure 1). Stroke risk was not significantly correlated with performance on measures of global cognitive functioning (i.e., DRS total score; r = −0.15, p = 0.08), attention (r = −0.01, p = 0.45), language (r = −0.03, p = 0.39), visuospatial skills (r = 0.02, p = 0.44), or executive functioning (r = −0.04, p = 0.38). In addition, stroke risk was not associated with the presence of the APOE ε4 allele (r = 0.05, p = 0.62). The same correlational analyses were conducted separately for each of the three participant groups as well as for the two AD groups combined. There were no significant correlations between stroke risk, cognition, and the presence of the APOE ε4 allele (all p-values >0.05).
ANOVAs demonstrated that the NC, nondepressed AD, and depressed AD groups significantly differed in stroke risk (F(2,87) = 3.33, p = 0.04, ). Post hoc pairwise comparisons using Tukey’s HSD tests showed that stroke risk was significantly greater in the depressed AD patients than in NC participants (p = 0.03). There were no significant differences in stroke risk scores between the NC participants and the nondepressed AD patients (p = 0.36), or between depressed and non-depressed AD patients (p = 0.45; see Figure 2).
The three participant groups were compared on each of the individual variables composing the FSRP score. χ2 analyses demonstrated that the groups significantly differed in terms of frequency of cardiovascular disease (χ2 = 7.73, p = 0.02). Follow-up post hoc χ2 analyses demonstrated that the depressed AD (χ2 = 7.50, p = 0.006) group had a significantly greater frequency of cardiovascular disease as compared to the NC group; however, there were no significant differences between the nondepressed AD and NC groups or between the two AD groups (p-values >0.05). In addition, a one-way ANOVA revealed a trend toward significant differences among the NC, nondepressed AD, and depressed AD groups in terms of systolic blood pressure (F(2,87) = 3.04, p = 0.05, ). In contrast, there were no significant between-group differences in terms of frequency of anti-hypertensive medication use (χ2 = 2.10, p = 0.35), diabetes mellitus (χ2 = 0.42, p = 0.81), cigarette smoking (χ2 = 2.22, p = 0.33), or history of atrial fibrillation (χ2 = 0.31, p = 0.86), or in total cholesterol (F(2,87) = 1.74, p = 0.18, ). Not surprisingly, given its role in cholesterol transport (e.g., Mahley, 1988), APOE genotype was significantly associated with total cholesterol level such that APOE ε4 carriers had higher cholesterol than non-ε4 carriers (r = 0.26, p = 0.01).
The three participant groups differed on DRS total score (F(2,87) = 51.08, p <0.001, ) and on the memory (F(2,87) = 245.22, p <0.001, ), attention (F(2,87) = 11.10, p <0.001, ), language (F(2,87) = 26.30, p <0.001, ), visuospatial function (F(2,87) = 19.27, p <0.001, ), executive function (F(2,87) = 39.01, p <0.001, ), and speed of information processing (F(2,87) = 38.16, p <0.001, ) summary scores. Post hoc pair-wise comparisons showed that both depressed and nondepressed AD patients performed significantly worse than NC participants on the DRS total score and each of the six cognitive domain scores (all p-values <0.05). The two AD groups did not differ significantly on any of these cognitive measures (all p-values >0.10).
Binary logistic regression analyses showed that when NC participants and nondepressed AD patients were included in the analyses and stroke risk was the only predictor variable, it did not produce a significant improvement in the model relative to the null model (χ2 = 2.29, p = 0.13, Nagelkerke R2 = 0.05). The model accurately classified 51.7% of the participants as either nondepressed AD or cognitively normal. In a second model, both stroke risk and APOE genotype were added as predictors. This model was a significant improvement over the null model (χ2 = 15.13, p = 0.001, Nagelkerke R2 = 0.30) and accurately classified 75.0% of participants as nondepressed AD or cognitively normal.
A second set of binary logistic regression analyses included NC participants and depressed AD patients. When stroke risk was the only predictor variable, it produced a significant improvement in the model relative to the null model (χ2 = 7.11, p = 0.008, Nagelkerke R2 = 0.15). The model accurately classified 56.7% of the participants as either belonging to the depressed AD or NC group. When both stroke risk and APOE genotype were added as predictors, the model was a significant improvement over the null model (χ2 = 30.21, p <0.001, Nagelkerke R2 = 0.53) and accurately classified 83.3% of participants as depressed AD or cognitively normal.
Our results indicate that AD patients with depression have a greater degree of stroke risk than cognitively normal individuals, even after excluding those with cerebrovascular disease. Stroke risk for AD patients without depression fell between that of depressed AD patients and cognitively normal individuals, but did not differ significantly from either group. Correspondingly, logistic regression analyses showed that stroke risk was more effective at differentiating depressed AD patients than nondepressed AD patients from cognitively normal individuals. These results highlight the complex and subtle nature of the relations among AD, stroke risk, and depression. The group differences in overall stroke risk appeared to be driven by differences in the rate of cardiovascular disease and possibly also in blood pressure.
The finding that stroke risk was not significantly higher in the depressed AD group than in the nondepressed AD group is consistent with other studies reporting that stroke risk alone does not appear to explain the link between depression and AD (Luchsinger et al., 2008). As Luchsinger et al. (2008) note, there are several mechanisms other than stroke risk factors that could explain the link between depression and AD, including a psychological reaction to cognitive decline, damage to catecholaminergic circuits during AD development and progression, or degeneration of limbic structures leading to vulnerability to depression.
The present results demonstrated that the frequency of the APOE ε4 allele was significantly different across the three groups. APOE ε4 allele frequencies were 80.0%, 56.7%, and 16.7% for the depressed AD, nondepressed AD, and NC groups, respectively. The observed ε4 allele frequencies are consistent with previously published reports of approximate frequencies of 72% (Delano-Wood et al., 2008b), 50% (Saunders et al., 1993), and 16% (Strittmatter et al., 1993) in these populations, respectively.
Given its role in cholesterol transport and neuronal repair (e.g., Mahley, 1988), it is proposed that the APOE ε4 allele may be related to stroke risk factors which, in turn, may be linked to depressive symptoms. However, stroke risk burden did not differ between depressed and nondepressed AD patients in the present study, suggesting that other APOE ε4-related mechanisms may be involved. In a population-based study of older adults, the presence of the APOE ε4 allele was associated with greater risk of severe depression even after adjusting for both cardiovascular conditions and lipid profile (Yen et al., 2007). Their findings suggest that the APOE ε4 allele and stroke risk factors are independent risk factors for the development of depression. In the present study, an association between stroke risk and APOE genotype was not demonstrated.
In conjunction with previous findings indicating that stroke risk is associated with increased rate of decline in individuals with MCI (Reitz et al., 2007) and AD (Mielke et al., 2007), the present findings suggest that assessment of stroke risk may aid in the early detection and monitoring of individuals who are vulnerable to future cognitive decline. It is well known that depression is common in stroke patients and is associated with decreased quality of life and increased mortality, and as noted by Robinson (2005), it is reasonable to expect that depression associated with vascular disease is also associated with a high mortality rate. Fortunately, evidence suggests that treatment with antidepressants reduces mortality in post-stroke depression (Jorge et al., 2003).
A second finding from the present study is that across all participants stroke risk was associated with decrements in memory and speed of information processing. Notably, the pattern of neuropsychological test performance in vascular dementia and AD dementia can be similar (Desmond, 2004). Indeed, the moderately demented AD patients in the present study exhibited deficits in executive functioning, speed of information processing, and attention that were similar to those observed in vascular dementia, but the cognitive domain most strongly associated with vascular risk factors was episodic memory, the cognitive ability that is often first to decline during the course of AD (Salmon and Bondi, 2009).
An important aspect of the present study was the careful characterization of the participant groups and the exclusion of individuals with a history of TIA, clinical stroke, or significant cerebrovascular disease. These measures were taken in an effort to exclude any nondemented older adults with cognitive impairment from participating in the study and to exclude individuals whose dementia may have had a significant vascular component. However, by excluding individuals with greater stroke risk burden, it is possible that the link between stroke risk and cognition was underestimated.
Despite these limitations, our findings highlight the possibility that prevention and treatment of stroke risk factors may lead to the preservation of cognitive functioning or the slowing of cognitive decline (Bowler and Hachinski, 2003). AD and stroke are growing public health concerns due to increasing longevity in the US population, and both have devastating psychological and financial effects. The present findings emphasize the importance of assessing and monitoring stroke risk factors not only in individuals with AD but also in healthy, cognitively normal older adults with no history of stroke or significant cerebrovascular disease.
This work was supported by grants from the National Institutes of Health (F31 NS059193 [KJB], P50 AG05131 [MWB, DPS], K24 AG026431, RO1 AG012674, and by grant IIRG-07-59343 from the Alzheimer’s Association [MWB]). The content is solely the responsibility of the authors and does not necessarily represent the official views of the Alzheimer’s Association or the National Institutes of Health. The authors gratefully acknowledge the assistance of staff, patients, and volunteers of the UCSD Alzheimer’s Disease Research Center.
Conflict of interest