Abstract
Introduction. The aim of the present study was to examine the power of B-type natriuretic peptide (BNP) and mild cognitive impairment as independent predictors of total and cardiovascular mortality in combination with established cardiovascular risk markers in an elderly general population without severe cognitive impairment.
Methods. A total of 499 individuals, aged more than 75 years, were examined and followed up for a median of 7.9 years in a prospective population-based stratified cohort study carried out in eastern Finland. The Cox proportional hazards regression model was used to determine the impact of multiple factors on total and cardiovascular mortality.
Results. In a multivariable model including established cardiovascular risk factors and conditions, both continuous BNP (adjusted hazard ratio (HR) 1.44 for a 1-SD change; 95% confidence interval (CI) 1.22–1.77; P < 0.001) and continuous MMSE score (HR 0.81 for a 1-SD change; 95% CI 0.70–0.94; P = 0.007) were independently associated with all-cause mortality. In a multivariable model, BNP remained a significant predictor of cardiovascular mortality, while MMSE score lost its significance.
Conclusions. BNP, a measure of cardiovascular burden, and MMSE score 18–23, an indicator of mild cognitive impairment, are both independent predictors of total mortality. BNP and MMSE score may potentially be useful in screening elderly patients for elevated risk of mortality.
Key messages
B-type natriuretic peptide (BNP) and Mini-Mental State Examination (MMSE) score are both independent predictors of total mortality in an elderly general population free of dementia.
BNP also predicts cardiovascular mortality in an elderly population even after adjusting for traditional cardiovascular risk markers and measurement of cognition by the MMSE score.
Introduction
Traditional cardiovascular risk factors such as hypertension, dyslipidemia, and obesity seem to lose their value in predicting mortality for an elderly population and are often associated with a neutral or even better prognosis (Citation1–3). Regarding newer risk markers, natriuretic peptides are powerful predictors of mortality for populations with a variety of cardiovascular diseases or in a general population (Citation4). They have also performed well as a risk marker of mortality in several high-risk groups including the elderly, whereas C-reactive protein (CRP), another recently established cardiovascular risk marker for the general population, has failed to provide the same accuracy as a harbinger of mortality among aged populations (Citation5,Citation6).
Severe cognitive decline and dementing illness have a remarkable impact on mortality (Citation7,Citation8), but the role of modest impairment is less clear (Citation9–12). Previous studies examining the impact of modest cognitive impairment on mortality have often relied solely on the Mini-Mental State Examination (MMSE) score to exclude those with different forms of dementing illnesses (Citation13,Citation14). Many participants with undiagnosed dementia and, therefore, compromised prognosis but who have a relatively preserved MMSE score have most probably been included in those studies.
Accumulating evidence suggests that factors predisposing to cardiovascular disease are also pivotal in the development of cognitive dysfunction (Citation15,Citation16). Furthermore, cognitive impairment frequently develops in elderly patients with heart failure (Citation17), and cognitive impairment is an independent prognostic marker in older patients with heart failure (Citation18). B-type natriuretic peptide (BNP) level has been shown to associate with the MMSE score in patients with congestive heart failure (Citation19). In our recent study based on an elderly general population, we showed the independent association of B-type natriuretic peptide (BNP) and the MMSE score, decline of the MMSE score, and the occurrence of new cases of dementia during a 5-year follow-up (Citation20). To our knowledge, the MMSE score and BNP, both powerful predictors of mortality among the elderly, have not been studied simultaneously as predictors of mortality among an elderly general population.
The aim of the present study was to examine the power of BNP and mild cognitive impairment, defined as MMSE 18–23, as independent predictors of total and cardiovascular mortality together with established cardiovascular risk markers in an elderly general population free of severe cognitive impairment.
Methods
Study population
This study is a part of a larger population-based, multidisciplinary Kuopio 75+ health study focusing on the clinical epidemiology of diseases, medication, and functional capacity in elderly persons aged 75 years or older. The target population was a stratified random sample (n = 700) of all residents of the city of Kuopio in eastern Finland who were aged 75 years or more on 1 January 1998 (n = 4,518).
The cohort included 700 participants (). Five persons could not be contacted, 79 refused to take part in the study, and 15 died before the examination. The remaining 601 participants attended a structured clinical examination and an interview conducted by a geriatrician and a trained nurse during the year 1998. As a part of the diagnostic process, brain imaging either by computer tomography (CT) or magnetic resonance imaging (MRI) was carried out for all participants with a suspicion of a dementing illness but no brain imaging in the medical history. Dementia was diagnosed as Alzheimer's disease, vascular dementia, dementia with Lewy bodies, or dementia due to other medical conditions by an experienced neurogeriatrician (R.S.) according to the Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV criteria and consensus guidelines for the clinical and pathological diagnosis of dementia with Lewy bodies (Citation21). A clinical diagnosis of dementia was established, and the type of the dementia was determined in consensus meetings using all the data available. Participants with 18 MMSE points or more were included (n = 499) in this substudy.
Figure 1. Flow chart of the study.
A geriatrician and a trained nurse interviewed and examined the participants at the out-patient clinic of the municipal hospital about their medical history and use of medicines. If a participant was unable to visit the study site, the nurse and the geriatrician made a home visit to perform the interview and examination. Medical records from the municipal health center, home nursing service, local hospitals, and the Kuopio University Hospital were also available. Base-line clinical and demographic data were also recorded. Diabetes was defined as a previous diagnosis of diabetes or a fasting plasma glucose level of 7.0 mmol/L or more. Data on other cardiovascular conditions were obtained from medical records. Blood systolic and diastolic pressures were measured twice, and the average of the measurements was recorded. Depression was screened using Zung's self-rating depression scale (Citation22).
Written informed consent was obtained from the study participants or their relatives as stipulated in the Declaration of Helsinki. The study was approved by the ethics committees of the Hospital District of Northern Savo and the Kuopio University Hospital.
Laboratory analyses
Basic blood count, creatinine, cholesterol profile, and fasting blood glucose were measured once at the Kuopio University Hospital after a 12-hour fast without interruption in the medication of the subjects. All serum total cholesterol assays were analyzed at the Kuopio University Hospital laboratory using standard enzymatic techniques. Creatinine clearance was calculated using creatinine, age, and body weight according to Cockcroft-Gault's formula. Estimation of low-density lipoprotein (LDL) cholesterol was calculated using Friedewald's formula. The blood samples for the natriuretic peptide analysis were drawn with other blood samples into chilled tubes containing 1.5 mg K2-EDTA per mL blood after the patient had been in a supine position for 30 min at 8 a.m. Whole blood was centrifuged and plasma immediately frozen and stored at −70°C. BNP was extracted from plasma using Sep-Pak C18 cartridges. The radioimmunoassay employed for BNP has been described previously (Citation23). The sensitivity of the BNP assays was 0.5 pmol/L. The within- and between-assay coefficients of variation were <10% and <15%, respectively. With this method, BNP plasma levels of 6.25 ± 2.12 pmol/L (mean ± SD) have been detected in healthy adults aged 20–55 years. All laboratory data were analyzed in random order and blinded to clinical data.
Main outcome measures
Mortality data were obtained from Statistics Finland, which is the national health register authority in Finland. There were no losses during follow-up. All deaths that occurred between March 1998 and November 2006 were recorded. Life-span was calculated from the date of examination in 1998 to 30 November 2006. The causes of deaths were classified according to the Tenth International Classification of Disease (ICD-10), and codes I00–99 were classified as cardiovascular deaths.
Statistical analyses
Base-line characteristics are given for participants divided into two groups by the survival status (). Comparisons between the groups were performed with the aid of the t test for independent samples or the Mann-Whitney U test for continuous variables, based on whether the distribution was Gaussian. The chi-square test was used for categorical data. Base-line BNP and MMSE score levels were highly skewed and therefore logarithmically transformed in all analyses, as were other variables when appropriate. We used Cox proportional hazards model with penalized splines (Citation24) adjusted for age and sex to examine whether the association between mortality and independent variables BNP and MMSE score changes in a non-linear fashion across the full range of these independent parameters. A priori, we selected three degrees of freedom based on biologic plausibility. Non-linear models did not differ from linear models with statistical significance using analysis of deviance table with chi-square test, and for the rest of the tests we applied linear models. Cox proportional hazards regression models for total and cardiovascular mortality were constructed to calculate hazard ratios (HR) with 95% confidence intervals (CI) for BNP, the MMSE score, and the variables listed in . The proportionality assumption was checked for the main analyses based on correlations of survival rankings with Schoenfeld residuals; all covariates fulfilled this criterion. BNP and the MMSE score were tested both as continuous and categorical variables. Regarding the cut-points for categorical formulations, we used the pre-specified < 24 points for MMSE, which was approximately the 20th percentile. Similarly, the 80th percentile for BNP (79.7 pg/mL) was used as a cut-point to facilitate comparisons with earlier BNP studies (Citation12,Citation25). Cardiovascular risk markers and previous illnesses () with significant or near-significant (P < 0.100) stratifier capacity for mortality and cardiovascular mortality, separately, were used as covariates in multivariable Cox proportional hazards models (). The analyses were repeated in pre-specified subgroups: participants with no diagnosed dementia (n = 454) and participants with no previously diagnosed heart failure (n = 376). HDL and total cholesterol were highly collinear and both associated with mortality, but only the one with the stronger association, HDL, was used in the Cox proportional hazards models. All tests were two-sided, and P < 0.05 was considered significant. The data were analyzed with SPSS release 15.0 for Windows (SPSS Inc., Chicago, Illinois).
Table I. Base-line characteristics of the participants by survival status.
Table II. Hazard ratios for total and cardiovascular mortality during the median follow-up of 7.9 years according to base-line BNP and MMSE score (n = 499).
Results
Base-line data
During a median follow-up period of 8.4 years (interquartile range 8.2–8.6) for survivors and 7.9 years (2.7–8.4) for the whole study population, 248 (49.6%) participants expired, producing an annual mortality rate of 7.9%. The cause of death was cardiovascular in 139 (53.9%) cases—malignancy with 44 (17.1%) and infection with 22 (8.5%) deaths being the other common causes.
The base-line data of the participants according to their survival status are presented in . Compared to survivors, the non-survivors were older and more likely to have a history of heart failure, atrial fibrillation, or stroke. In a comparison of individuals who expired and those who survived, the base-line median (interquartile range) BNP was 48.6 pmol/L (24.0–86.5) versus 27.2 pmol/L (16.2–47.9), respectively (P < 0.001). The MMSE score among the non-survivors was 26 points (24–29 points) and among survivors 28 points (25–29 points) (P < 0.001).
Predictors of total mortality
When tested separately or in sex- and age-adjusted models, both the MMSE score and BNP as continuous and dichotomous variables were significantly linked to mortality (). The age- and sex-adjusted mortality risk increased with the increase in the quintile of BNP (P < 0.001) (), with a hazard ratio (HR) of 2.75 between the lowest and highest quintile. Similarly, the age- and sex-adjusted risk for the MMSE score was elevated with the decrease in the quintile of the MMSE score (P < 0.001) (), with an HR of 2.01 between the highest and lowest quintile.
Figure 2. Age- and sex-adjusted Cox proportional hazards survival curves for total and cardiovascular mortality divided by B-type natriuretic peptide (BNP) and Mini-Mental State Examination test (MMSE) quintiles (I–V), separately, in the entire study population (n = 499). A: BNP and total mortality (P for trend < 0.001). B: MMSE and total mortality (P < 0.001). C: BNP and cardiovascular mortality (P < 0.001). D: MMSE and cardiovascular mortality (P = 0.067). The BNP quintiles were 3.2–17.0, 17.2–28.0, 28.2–45.3, 46.2–78.8, and 79.7–500 pg/mL. The MMSE score quintiles were 18–23, 24–25, 26–27, 28, and 29–30.
In a fully adjusted multivariable model (used in ), age (HR 1.71; 95% confidence interval (CI) 1.42–2.11; P < 0.001), systolic blood pressure (HR 0.80; 95% CI 0.68–0.98; P = 0.029), diabetes (HR 1.65; 95% CI 1.14–2.38; P = 0.008), continuous MMSE score (HR 0.81; 95% CI 0.70–0.94; P = 0.007), and continuous BNP (HR 1.44; 95% CI 1.22–1.77; P < 0.001) were independently associated with mortality. Regarding cardiovascular morbidities, only a history of stroke showed a trend towards greater mortality (HR 1.53; 95% CI 0.98–2.36; P = 0.07). A history of heart failure, myocardial infarction, or atrial fibrillation was not independently associated with total mortality (P > 0.100 for all). BNP was a clear and the MMSE score a border-line prognostic factor as dichotomous variables in multiparameter models ().
Predictors of cardiovascular mortality
BNP and the MMSE score, as both continuous and dichotomous variables, were associated with cardiovascular mortality when tested separately or in an age- and sex-adjusted model (). The age- and sex-adjusted risk of cardiovascular mortality was elevated with increasing quintile of BNP (P < 0 .001) (), with an HR of 3.29 between the lowest and highest quintile. A similar model for the MMSE score showed no significant connection to cardiovascular mortality (P = 0.067) (), the HR being 1.49.
Age (HR 1.71; 95% CI 1.42–2.75; P < 0.001), systolic blood pressure (HR 0.74; 95% CI 0.58–0.95; P = 0.016), New York Heart Association (NYHA) class III–IV versus I–II (HR 1.64; 95% CI 1.02–2.64; P = 0.044), diabetes (HR 1.71; 95% CI 1.05–2.80; P = 0.033), and BNP (HR 1.72; 95% CI 1.37–2.15; P < 0.001) were independently associated with cardiovascular mortality in a multivariable Cox model. None of the cardiovascular conditions () with an association to mortality (P < 0.100) was an independent predictor of cardiovascular mortality when tested in the multivariable model (P > 0.100 for all). Importantly, the MMSE score showed no association with cardiovascular mortality either as a continuous variable or in a dichotomous formulation () when the known risk factors were fitted into the same model. Removing BNP from the model did not materially change the result for the MMSE score.
Predictive value of the MMSE score and BNP in subgroup analyses
BNP and the MMSE score were tested in the pre-specified subgroups both as continuous and dichotomous variables in similarly constructed multivariable models. In the subgroup of participants with no previously diagnosed heart failure (n = 376; 75.4%), the results for BNP and the MMSE score remained essentially the same as in the entire study population.
In a multivariable model constructed of the subgroup with no diagnosed dementia at the base-line visit (n = 454; 90.1%), the results for BNP remained significant and unchanged for both end-points (data not shown). MMSE score 18–23 points was a significant predictor of total mortality; but for cardiovascular mortality, the same applied only in a univariable model.
Discussion
Main findings
This is the first study to examine the predictive capacity of BNP and the MMSE score simultaneously in combination with established cardiovascular risk markers. As in earlier studies conducted with the elderly, BNP was a powerful, independent predictor of total and cardiovascular mortality (Citation5,Citation6). Several recent studies have suggested that cardiovascular morbidity and cognitive dysfunction are closely connected among the elderly (Citation20,Citation26,Citation27). In the present study—even after adjusting for cognitive function measured by the MMSE score—BNP remained a robust predictor of total and cardiovascular mortality. Considering the known association with heart failure, BNP was also studied in the subgroup of participants who had no previously diagnosed heart failure, with no material changes in the results.
Altered levels of BNP are known to be associated with several cardiovascular conditions, such as hypertension, atrial fibrillation, history of myocardial infarction, and, particularly, heart failure caused by systolic or diastolic cardiac dysfunction (Citation28). In elderly populations, especially diastolic heart failure and asymptomatic cardiac dysfunction are exceedingly common (Citation29), and they are likely to explain some of the good predictive value of BNP on mortality in an elderly population. We used clinical manifestations of the above-mentioned diseases as covariates in our multivariable analysis, but BNP consistently remained a significant variable in the models. Furthermore, the median BNP level in our study was remarkably low as compared to patients with clinical heart failure. These findings underscore the predictive importance of even modest increases of BNP.
Predictive value of BNP versus traditional cardiovascular risk markers
Traditional cardiovascular risk markers such as hypertension, dyslipidemia, and obesity are known to lose much if not all of their prognostic power when measured in the elderly (Citation2,Citation3). Low blood pressure and low cholesterol have often served as predictors of mortality when studied among the aged population. Both of these parameters also had the inverse association with mortality in our study, even though low total cholesterol only in a univariable model. Systemic blood pressure is attenuated by several conditions frequently found in the elderly population, such as heart failure, atrial fibrillation, aortic stenosis, and dehydration. All of these derangements increase mortality, deteriorating the prognostic value of hypertension amongst the aged, while high blood pressure remains a marker of cardiovascular morbidity and mortality in younger populations. Thus, it is possible that the heart and cardiovascular system is under stress—as evident in the relative increase in the BNP level—even in the case of normal or even low blood pressure. This is also a putative mechanism for antihypertensive medication possibly being protective against dementia even in the presence of lower blood pressures.
In line with previous studies, a high NYHA class and diabetes were associated with cardiovascular mortality. Although a history of atrial fibrillation, symptomatic heart failure, myocardial infarction, and stroke was more common among the individuals who expired, only a history of stroke had any connection with mortality in the multivariable models. This corresponds with earlier mortality studies conducted in the elderly: the quantitative characteristic of disease severity, such as NYHA class, BNP, or the MMSE score, are stronger determinants of prognostic impact than the sheer existence of the condition (Citation11). CRP, a marker of inflammation and an extensively studied cardiovascular risk marker, has also performed worse in prognostic studies conducted in the elderly (Citation5,Citation30), while BNP as a direct marker of left ventricular stretch and cardiovascular stress has kept its impact as a prognostic marker also in various studies among the aged population (Citation5,Citation6).
The MMSE score and mortality
Severely reduced cognitive function, defined as MMSE score less than 18 points, and known dementing illness are associated with severely compromised prognosis (Citation12,Citation25). The results have been contradictory in studies with patients having mild cognitive impairment. Studies using an MMSE score of 18–23 as a definition of mild cognitive impairment (Citation12,Citation25) have found this criterion to be a significant predictor of mortality, while a study employing another measure of cognition, the Short Portable Mental Status Questionnaire (SPMSQ), showed no predictive value for mild cognitive impairment (Citation9). One study suggested that the association is apparent only in individuals aged less than 80 years (Citation25). The present data suggest an association with the MMSE score and mortality also among those beyond 80 years of age (data not shown).
In our population, the MMSE score, both as a continuous variable and using the cut-off point of 24, was a significant predictor of total mortality in all the models. In a model including other factors associated with mortality, its individual impact on mortality was only of moderate value. The predictive value of the MMSE score of less than 24 points, as studied separately and in an age- and sex-adjusted model, agreed with recently published data by Strandberg et al. (Citation12) in a similar type of setting. In a multivariable model, the predictive value of this variable was somewhat reduced in our data, possibly reflecting the more extensive use of confounding variables in the survival model.
To avoid the strong prognostic implications of a diagnosis of dementia, we performed a subanalysis after excluding the 45 individuals with dementia (all with MMSE ≥ 18). The value of the MMSE score was somewhat attenuated but not abolished with regard to total mortality, when the patients with an established compromised prognosis were excluded. This suggests that the prognostic importance of the MMSE score is also conveyed by other pathways than dementia. These additional mechanisms may include, for example, difficulties in engaging in health promoting activities, seeking medical advice, or using the prescribed medication.
The MMSE score and cardiovascular mortality
Mid-regional proatrial natriuretic peptide, a stable form of the N-terminal fragment of proatrial natriuretic peptide, has been suggested to hold additional value in diagnosing Alzheimer's disease, the most common cause of dementia (Citation27). In our recent study we also reported BNP to associate with declining cognitive function and future dementia in a population free of dementing illness at base-line (Citation20). Cognitive dysfunction has been associated with heart failure due to systolic and diastolic dysfunction (Citation17,Citation31), and cognitive decline has been shown to predict mortality in heart failure patients (Citation18). The failure of traditional cardiovascular risk markers for the elderly and accumulating evidence of the association between cardiovascular burden and cognitive dysfunction have led to the hypothesis that the MMSE score might hold additional predictive value for cardiovascular mortality (Citation12). Our data suggest that mild cognitive impairment defined by MMSE 18–23 is associated with cardiovascular mortality when studied separately or in a sex- and age-adjusted model. In a multivariable model with known cardiovascular risk markers, with or without BNP, the MMSE score did not provide any additional prognostic information.
Study limitations and strengths
In our study population, the prevalence of heart failure (24.6%) and other cardiovascular conditions was markedly greater than in most of the studies focused on the epidemiology of heart failure (Citation29,Citation32). The diagnosis of heart failure was obtained from the medical records, which may have caused some level of over-diagnosis. Nevertheless, eastern Finland is known for an exceptionally high cardiovascular disease burden (Citation33). Our study was carried out in an almost exclusively white Caucasian population. Extrapolation of our results to non-Caucasian populations should, therefore, be done with caution.
The cognitive characterization of the participants is one of the particular strengths of the present study. Compared to earlier studies examining the power of mild cognitive impairment as a risk factor for mortality, the participants of our study were systematically examined by an experienced geriatrician (R.S.) for possible dementia. Moreover, CT or MR imaging of the brain was performed for all participants with a suspicion of dementia.
Conclusions
BNP and MMSE score as measures of cardiovascular burden and cognitive dysfunction, respectively, are both independent predictors of total mortality. Unlike the MMSE score, however, BNP is also an independent harbinger of cardiovascular mortality. Suppression of BNP with cardiovascular medication offers an interesting target for future randomized studies. BNP and MMSE score may potentially be useful to screen elderly patients for elevated risk of mortality.
Declaration of interest: This work was supported by the Finnish Cultural Foundation, Paolo Foundation and Paavo Nurmi Foundation.
References
- Schupf N, Costa R, Luchsinger J, Tang MX, Lee JH, Mayeux R. Relationship between plasma lipids and all-cause mortality in nondemented elderly. J Am Geriatr Soc. 2005;53:219–26.
- Oates DJ, Berlowitz DR, Glickman ME, Silliman RA, Borzecki AM. Blood pressure and survival in the oldest old. J Am Geriatr Soc. 2007;55:383–8.
- Takata Y, Ansai T, Soh I, Akifusa S, Sonoki K, Fujisawa K, . Association between body mass index and mortality in an 80-year-old population. J Am Geriatr Soc. 2007;55:913–7.
- Wang TJ, Larson MG, Levy D, Benjamin EJ, Leip EP, Omland T, . Plasma natriuretic peptide levels and the risk of cardiovascular events and death. N Engl J Med. 2004; 350:655–63.
- Kistorp C, Raymond I, Pedersen F, Gustafsson F, Faber J, Hildebrandt P. N-terminal pro-brain natriuretic peptide, C-reactive protein, and urinary albumin levels as predictors of mortality and cardiovascular events in older adults. JAMA. 2005;293:1609–16.
- Wallen T, Landahl S, Hedner T, Nakao K, Saito Y. Brain natriuretic peptide predicts mortality in the elderly. Heart. 1997;77:264–7.
- Kelman HR, Thomas C, Kennedy GJ, Cheng J. Cognitive impairment and mortality in older community residents. Am J Public Health. 1994;84:1255–60.
- Larson EB, Shadlen MF, Wang L, McCormick WC, Bowen JD, Teri L, . Survival after initial diagnosis of Alzheimer disease. Ann Intern Med. 2004;140:501–9.
- Stump TE, Callahan CM, Hendrie HC. Cognitive impairment and mortality in older primary care patients. J Am Geriatr Soc. 2001;49:934–40.
- Nguyen HT, Black SA, Ray LA, Espino DV, Markides KS. Cognitive impairment and mortality in older Mexican Americans. J Am Geriatr Soc. 2003;51:178–83.
- Fried LP, Kronmal RA, Newman AB, Bild DE, Mittelmark MB, Polak JF, . Risk factors for 5-year mortality in older adults: the Cardiovascular Health Study. JAMA. 1998; 279:585–92.
- Strandberg TE, Pitkala KH, Tilvis RS. Predictors of mortality in home-dwelling patients with cardiovascular disease aged 75 and older. J Am Geriatr Soc. 2009;57:279–84.
- Ganguli M, Dodge HH, Shen C, Pandav RS, DeKosky ST. Alzheimer disease and mortality: a 15-year epidemiological study. Arch Neurol. 2005;62:779–84.
- Jagger C, Andersen K, Breteler MM, Copeland JR, Helmer C, Baldereschi M, . Prognosis with dementia in Europe: A collaborative study of population-based cohorts. Neurologic Diseases in the Elderly Research Group. Neurology. 2000;54(11 Suppl 5):16–20.
- Kloppenborg RP, van den Berg E, Kappelle LJ, Biessels GJ. Diabetes and other vascular risk factors for dementia: which factor matters most? A systematic review. Eur J Pharmacol. 2008;585:97–108.
- Fitzpatrick AL, Kuller LH, Lopez OL, Diehr P, O'Meara ES, Longstreth WT Jr, . Midlife and late-life obesity and the risk of dementia: cardiovascular health study. Arch Neurol. 2009;66:336–42.
- Cacciatore F, Abete P, Ferrara N, Calabrese C, Napoli C, Maggi S, . Congestive heart failure and cognitive impairment in an older population. Osservatorio Geriatrico Campano Study Group. J Am Geriatr Soc. 1998;46:1343–8.
- Zuccalà G, Pedone C, Cesari M, Onder G, Pahor M, Marzetti E, . The effects of cognitive impairment on mortality among hospitalized patients with heart failure. Am J Med. 2003;115:97–103.
- Feola M, Rosso GL, Peano M, Agostini M, Aspromonte N, Carena G, . Correlation between cognitive impairment and prognostic parameters in patients with congestive heart failure. Arch Med Res. 2007;38:234–9.
- Kerola T, Nieminen T, Sulkava R, Hartikainen S, Vuolteenaho O, Kettunen R. B-type natriuretic peptide as a predictor of declining cognitive function and dementia—a cohort study of an elderly general population with a 5-year follow-up. Ann Med. 2010;42:207–15.
- McKeith IG, Galasko D, Kosaka K, Perry EK, Dickson DW, Hansen LA, . Consensus guidelines for the clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international workshop. Neurology. 1996;47:1113–24.
- Zung WW. A self-rating depression scale. Arch Gen Psychiatry. 1965;12:63–70.
- Ala-Kopsala M, Magga J, Peuhkurinen K, Leipälä J, Ruskoaho H, Leppäluoto J, . Molecular heterogeneity has a major impact on the measurement of circulating N-terminal fragments of A- and B-type natriuretic peptides. Clin Chem. 2004;50:1576–88.
- Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology. 1995;6:356–65.
- Bassuk SS, Wypij D, Berkman LF. Cognitive impairment and mortality in community-dwelling elderly. Am J Epidemiol. 2000;151:677–88.
- Iadecola C, Davisson RL. Hypertension and cerebrovascular dysfunction. Cell Metab. 2008;7:476–84.
- Buerger K, Ernst A, Ewers M, Uspenskaya O, Omerovic M, Morgenthaler NG, . Blood-based microcirculation markers in Alzheimer's disease—diagnostic value of midregional pro-atrial natriuretic peptide/C-terminal endothelin-1 precursor fragment ratio. Biol Psychiatry. 2009;65: 979–84.
- Vuolteenaho O, Ala-Kopsala M, Ruskoaho H. BNP as a biomarker in heart disease. Adv Clin Chem. 2005;40:1–36.
- Redfield MM, Jacobsen SJ, Burnett JC Jr, Mahoney DW, Bailey KR, Rodeheffer RJ. Burden of systolic and diastolic ventricular dysfunction in the community: appreciating the scope of the heart failure epidemic. JAMA. 2003;289: 194–202.
- Strandberg TE, Tilvis RS. C-reactive protein, cardiovascular risk factors, and mortality in a prospective study in the elderly. Arterioscler Thromb Vasc Biol. 2000;20: 1057–60.
- Suwa M, Ito T. Correlation between cognitive impairment and left ventricular diastolic dysfunction in patients with cardiovascular diseases. Int J Cardiol. 2009;136:351–4.
- Allender S, Peto V, Scarborough P, Rayner M. Coronary heart disease statistics. British Heart Foundation; London 2008.
- Pyörälä K, Salonen JT, Valkonen T. Trends in coronary heart disease mortality and morbidity and related factors in Finland. Cardiology. 1985;72:35–51.