Abstract
This study aimed to investigate whether the effect of blood pressure (BP) on mortality differs by levels of cognitive function. The associations of brachial systolic BP, diastolic BP, mean arterial pressure (MAP), and pulse pressure with all‐cause mortality were prospectively explored (follow‐up 7.0±2.2 years) in 660 community‐dwelling individuals (≥60 years) using adjusted Cox models, stratified by cognitive impairment (Mini‐Mental State Examination [MMSE] <24). No association between brachial BP variables and mortality was shown for the total sample in quartiles analysis; however, MAP in the highest quartile, compared with the second, was associated with mortality (hazard ratio, 1.85; 95% confidence intervals, 1.09–3.12) among cognitively impaired individuals. The fractional‐polynomials approach for BP confirmed this finding and further showed, solely in the MMSE <24 subcohort, U‐shaped trends of MAP and systolic BP, with increased mortality risk in extremely low or high values; no such pattern was evident for patients with MMSE ≥24. Elderly individuals with cognitive impairment might be more susceptible to the detrimental effects of low and elevated MAP and systolic BP.
Arterial hypertension, the most common modifiable cardiovascular disease (CVD) risk factor worldwide, increases with age and affects 65% of individuals 70 years and older.1 Despite the strong evidence connecting hypertension to increased mortality,2 the exact effect of blood pressure (BP), especially among elderly individuals, remains unknown. Previous studies report J‐ or U‐shaped associations of BP with mortality,3, 4, 5 and suggest that among the oldest individuals (>80 years), low rather than high BP is a predictor of mortality.4, 5
Cognitive impairment is a common late‐life comorbidity, as 10% of individuals older than 60 years have frank dementia.6 Increasing evidence links BP with cognitive decline,7, 8 attributing the observed associations to the hemodynamic consequences of elevated or low BP and increased pulsatility on cerebral vasculature.9, 10 Subsequent microvascular disease is considered a key player in the pathogenesis of not only vascular dementia, but also Alzheimer's disease.11, 12, 13 Considering this association between BP parameters and cognition14 along with the consistent reports from prospective studies about associations of cognitive impairment with all‐cause and CVD mortality15, 16, 17, 18, 19 it could be assumed that the effects of BP on mortality are differentiated by levels of cognitive function. Indeed, systematic reviews of clinical trials conclude that there is not sufficient evidence for benefit of antihypertensive treatment against mortality among hypertensive patients with dementia.20, 21
Nevertheless, only a few previous studies have examined how cognitive function affects the association of BP with mortality.22, 23, 24, 25 A reinforcement of a detrimental effect of low diastolic BP (DBP) on mortality among individuals with cognitive impairment is the most consistent finding, replicated in three studies,22, 23, 24 whereas results regarding systolic BP (SBP) and pulse pressure (PP) are conflicting.22, 23, 24, 25 Only one study explored mean arterial pressure (MAP).22 In view of the above, we hereby aim with the current study to examine the association of brachial BP components (SBP, DBP, MAP, PP) with 8‐year all‐cause mortality by levels of cognitive function in a community‐dwelling sample of elderly individuals from rural Greece.
Materials and Methods
Study Sample
Velestino is a rural town in Thessaly, central Greece, of around 4000 inhabitants. The recruitment for the second phase of the Velestino study was conducted during the period 2005–2006 and the target population comprised all permanent residents aged 60 years or older. During that time, 1080 individuals 60 years and older were recorded in the town registry; of them, 263 individuals had moved to other residence (nonpermanent inhabitants), leaving a total of 817 potentially eligible participants for the study. An attempt was made to contact all of the patients for inclusion into the study. However, 34 individuals died during the study period, eight were severely physically or mentally disabled and could not provide adequate information for inclusion to the study, and 99 participants refused to participate. Finally, 676 individuals signed the informed consent and participated in the study; all participants were community‐dwelling and noninstitutionalized. Details on study recruitment can be found elsewhere.19, 26, 27 Among the participants, 660 individuals had available BP and cognition measurements. Basic demographic characteristics of the included participants (age, sex) were comparable to those of the entire elderly population. The study protocol was reviewed and approved by the ethics committee of the Athens University Medical School and the Peripheral General Hospital of Volos.
Assessment of Sociodemographic, Lifestyle, and Clinical Characteristics
Using a precoded standardized questionnaire, two local physicians interviewed participants at their residence or workplace. Questions regarding sociodemographic (age, education, and family support) and lifestyle (social activity, smoking habits, and alcohol consumption) characteristics were included.26 Anthropometric measurements were carried out and body mass index (BMI) was calculated as weight divided by height squared. History of previous diagnoses and use of medications were based on self‐reported information, as well as on a thorough review of medical records. Morning blood sample was collected for biochemical assessment, and biochemical measurements combined with medical history allowed for the diagnoses of the diseases of interest. Particularly, hypercholesterolemia was determined by previous diagnosis or low‐density lipoprotein cholesterol levels ≥160 mg/dL28 and diabetes mellitus type 2 by medical history or fasting blood glucose levels ≥126 mg/dL.29 CVD was defined by history of coronary artery disease or cerebrovascular disease.
BP Measurements
Measurement of DBP and SBP was carried out by the two physicians who conducted the interviews. The examinees were in a calm state and a sitting position and were at rest for at least 10 minutes. Two measurements were conducted using manual mercury sphygmomanometers and the mean values of DBP and SBP were recorded. MAP and PP were thereafter calculated using the following respective formulae:30
Evaluation of Cognition and Depressive Symptoms
The validated Greek version of the Mini‐Mental State Examination (MMSE) was administered to participants to assess their global cognitive function. A score lower than 24 of 30 was indicative of cognitive impairment.31 Depressive symptomatology was determined using the 15‐item Geriatric Depression Scale (GDS), which has also been validated in the Greek population,32 with a score higher than six determining presence of depression.
Assessment of Mortality
In collaboration with the Local Registry Office, death certificates referring to the inhabitants of Velestino up to October 31, 2013, were accessed (mean duration of follow‐up: 7.0±2.2 years); vital status and date of death were recorded.
Statistical Analysis
The distribution of study variables by MMSE score (<24 vs ≥24) was derived and P values were subsequently calculated using chi‐square or t test as appropriate. Cox proportional hazard models were designed for all‐cause mortality. Two approaches were followed for BP variables in order to investigate potential nonlinear associations, as previously reported for the detrimental effects of BP:3, 22, 23, 24, 25 (1) in a quartile categorical approach, SBP, DBP, MAP, and PP were treated as categorical variables in quartiles with the second as a reference category; the quartile approach was preferred over tertiles, to more adequately illustrate potential nonlinear associations; and (2) we further examined the effects of SBP, DBP, MAP, and PP on all‐cause mortality using fractional‐polynomial modeling, and the results of the best‐fitting model were graphically presented.
A basic model included age (continuous by 1‐year increment), sex, and BP variables (SBP, DBP, MAP, PP) interchangeably (model 1), whereas in a fully adjusted model (model 2) the following variables were additionally introduced: education level (ordered; <6, 6–8, ≥9 years), social activity (optimal vs suboptimal), family support (adequate vs poor), BMI (continuous by 1 kg/m2 increment), alcohol intake (ordered; no, up to recommended limit, more), diabetes mellitus type 2, hypercholesterolemia, CVD, use of antihypertensive drugs, depression (GDS ≤6 vs >6), and cognitive impairment (MMSE <24 vs ≥24). Smoking (ever vs never) was not introduced in the model because the meager proportion of female smokers led to a high correlation with the sex variable. Similarly, use of lipid‐lowering drugs and glucose‐lowering drugs were not included because of the entering of the related hypercholesterolemia and diabetes mellitus variables. All analyses were performed for the total sample and were additionally stratified by presence of cognitive impairment (MMSE <24 and ≥24). Statistical analysis was conducted using SAS statistical software (version 9.2; SAS Institute Inc, Cary, NC) and STATA (version 12.0, STATA Corporation, College Station, TX).
Results
The mean age among participants at baseline was 73.6±7.9 years, whereas the mean MMSE score was 22.3±4.6. Of 660 participants, 54.8% had an MMSE score <24. Table 1 shows the distribution of the study variables by MMSE score at baseline and serves mostly descriptive purposes, as data are not mutually adjusted. Older age, female sex, lower education and suboptimal social activity, depression, and diabetes were more common in cognitively impaired patients, whereas smoking, alcohol intake, and use of antihypertensive drugs were more common among patients without cognitive impairment. With regards to the BP variables, PP was higher and SBP marginally higher among patients with cognitive impairment; no difference was recorded for DBP and MAP.
Table 1.
Sociodemographic, Lifestyle, and Clinical Characteristics of Study Participants (N=660) by MMSE Score
Variables | MMSE <24 | MMSE ≥24 | P Value |
---|---|---|---|
n=362 | n=298 | ||
Age y | 75.7±7.8 | 71.1±7.4 | <.001 |
Male sex | 131 (36.2) | 155 (52.0) | <.001 |
Education, y | |||
<6 | 271 (74.9) | 134 (45.0) | <.001 |
6–8 | 83 (22.9) | 127 (42.6) | |
≥9 | 8 (2.2) | 37 (12.4) | |
Suboptimal social activity | 148 (40.9) | 75 (25.2) | <.001 |
Poor family support | 56 (15.5) | 32 (10.7) | .08 |
BMI, kg/m2 | 30.4±5.0 | 30.3±5.3 | .72 |
Alcohol intake, glasses per d | |||
0 | 140 (38.7) | 80 (26.9) | .002 |
≤1 (women), ≤2 (men) | 183 (50.6) | 167 (56.0) | |
>1 (women), >2 (men) | 39 (10.8) | 51 (17.1) | |
Ever smoking | 85 (23.5) | 114 (38.3) | <.001 |
Depression (GDS ≤6) | 209 (57.7) | 120 (40.3) | <.001 |
Hypercholesterolemia | 121 (33.4) | 103 (34.6) | .76 |
Total cholesterol, mg/dL | 214.1±48.1 | 214.5±38.4 | .91 |
LDL cholesterol, mg/dL | 133.0±37.7 | 128.8±31.5 | .28 |
HDL cholesterol, mg/dL | 54.5±11.8 | 54.5±11.3 | .98 |
Diabetes mellitus type 2 | 71 (19.6) | 41 (13.8) | .05 |
Cardiovascular disease | 101 (27.9) | 74 (24.8) | .37 |
History of CAD | 75 (20.7) | 50 (19.8) | .77 |
History of cerebrovascular disease | 35 (9.7) | 21 (7.1) | .23 |
Antihypertensive drugs | 260 (71.8) | 192 (64.4) | .04 |
Diuretics | 85 (23.5) | 53 (17.8) | .07 |
Calcium channel blockers | 78 (21.6) | 39 (13.1) | .005 |
β‐Blockers | 54 (14.9) | 67 (22.5) | .01 |
ACEIs/ARBs | 162 (44.8) | 128 (43.0) | .64 |
Other antihypertensive drugs | 20 (5.5) | 14 (4.7) | .63 |
Glucose‐lowering drugs | 55 (15.2) | 34 (11.4) | .15 |
Lipid‐lowering drugs | 61 (16.9) | 63 (21.1) | .16 |
Systolic blood pressure | 148.3±22.2 | 145.3±22.4 | .09 |
Diastolic blood pressure | 81.1±13.0 | 81.8±12.0 | .48 |
Mean arterial pressure, mm Hg | 103.5±14.1 | 103.0±13.9 | .64 |
Pulse pressure, mm Hg | 67.2±18.5 | 63.6±17.8 | .01 |
Data are presented as n (%) or mean ± standard deviation. Abbreviations: ACEIs, angiotensin‐converting enzyme inhibitors; ARB, angiotensin II receptor blockers; BMI, body mass index; CAD, coronary artery disease; GDS, Geriatric Depression Scale; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; MMSE, Mini‐Mental State Examination.
During a mean follow‐up period of 7.0±2.2 years, 196 patients died (29.7%). A larger proportion of participants in the cognitive impairment (MMSE <24) subcohort died (38.4%), compared with cognitively normal (MMSE ≥24) patients (19.1%). Table 2 presents the multivariate Cox proportional hazard regression analysis results for SBP, DBP, MAP, and PP in the total sample and by MMSE subcohorts, according to the quartile approach. In the total sample, none of the BP variables were independently associated with mortality. An increased all‐cause mortality risk (hazard ratio [HR], 1.85; 95% confidence intervals [CI], 1.09–3.12, in model 2) was found for the highest MAP quartile (≥112.0 mm Hg), compared with the second (93.5–101.9 mm Hg) among patients with MMSE score <24. No such effect of MAP was noted for patients with MMSE scores in the normal range. Even though SBP, DBP, and PP were not significantly associated with mortality, a trend for a U‐shaped association for SBP only in the cognitively impaired subcohort was noted; the lowest (≤130.0 mm Hg; HR, 1.50; 95% CI, 0.89–2.55 in model 2) and the highest (>160.0 mm Hg; HR, 1.49; CI, 0.91–2.48 in model 2) quartiles were positively but nonsignificantly associated with mortality relative to the second (130.1–145.0 mm Hg) quartile. Similar to MAP, such a trend was not recorded for the cognitively normal patients. Other factors associated with mortality were increasing age, male sex, presence of cardiovascular disease, diabetes mellitus type 2, and cognitive impairment. In addition, among cognitively impaired patients, poor family support, suboptimal social activity, and depressive symptoms increased all‐cause mortality, whereas among cognitively normal patients, use of antihypertensive medications was negatively associated with mortality (data not shown).
Table 2.
Multivariate Cox Regression Derived HRs and 95% CIs for All‐Cause Mortality by BP Variables Stratified According to MMSE Score
Variable | BP Quartiles | Total Sample (N=660) | MMSE <24 (n=362) | MMSE ≥24 (n=298) | |||
---|---|---|---|---|---|---|---|
HR (95% CIs) | P Value | HR (95% CIs) | P Value | HR (95% CIs) | P Value | ||
Model 1: adjusted for age and sex | |||||||
Systolic blood pressure, mm Hg | ≤130 | 1.36 (0.90–2.05) | .15 | 1.43 (0.86–2.38) | .21 | 1.26 (0.61–2.60) | .53 |
130.1–145.0 | Reference | Reference | |||||
145.1–160.0 | 1.30 (0.85–1.98) | .23 | 1.24 (0.75–2.07) | .40 | 1.35 (0.62–2.94) | .45 | |
>160.0 | 1.31 (0.85–2.01) | .22 | 1.38 (0.84–2.28) | .21 | 0.92 (0.38–2.23) | .86 | |
Diastolic blood pressure, mm Hg | <75.0 | 1.08 (0.76–1.53) | .67 | 0.99 (0.65–1.51) | .95 | 1.09 (0.58–2.05) | .79 |
75.0–80.9 | Reference | Reference | Reference | ||||
81.0–90.0 | 0.93 (0.62–1.39) | .73 | 0.91 (0.55–1.52) | .72 | 0.98 (0.50–1.92) | .94 | |
>90.0 | 1.35 (0.85–2.13) | .20 | 1.47 (0.88–2.46) | .14 | 0.60 (0.20–1.78) | .35 | |
Mean arterial pressure, mm Hg | <93.5 | 1.19 (0.82–1.74) | .37 | 1.30 (0.81–2.10) | .28 | 1.25 (0.63–2.49) | .53 |
93.5–101.9 | Reference | Reference | Reference | ||||
102.0–111.9 | 1.10 (0.73–1.67) | .65 | 1.26 (0.75–2.11) | .39 | 1.69 (0.84–3.41) | .14 | |
≥112.0 | 1.26 (0.83–1.93) | .17 | 1.82 (1.10–3.03) | .02 | 0.65 (0.26–1.62) | .36 | |
Pulse pressure, mm Hg | ≤50.0 | 0.86 (0.57–1.31) | .49 | 0.65 (0.39–1.10) | .11 | 1.76 (0.81–3.83) | .16 |
50.1–61.9 | Reference | Reference | Reference | ||||
62.0–75.0 | 0.90 (0.59–1.37) | .61 | 0.65 (0.40–1.08) | .10 | 1.73 (0.79–3.78) | .17 | |
>75.0 | 1.02 (0.69–1.51) | .93 | 0.92 (0.59–1.44) | .72 | 1.23 (0.53–2.85) | .64 | |
Model 2: fully adjusted | |||||||
Systolic blood pressure, mm Hg | ≤130.0 | 1.33 (0.87–2.04) | .19 | 1.50 (0.89–2.55) | .14 | 1.08 (0.51–2.27) | .85 |
130.1–145.0 | Reference | Reference | Reference | ||||
145.1–160.0 | 1.21 (0.79–1.86) | .38 | 1.33 (0.79–2.25) | .28 | 1.25 (0.55–2.81) | .59 | |
>160.0 | 1.30 (0.84–2.02) | .24 | 1.49 (0.91–2.48) | .12 | 0.85 (0.34–2.09) | .72 | |
Diastolic blood pressure, mm Hg | <75.0 | 0.91 (0.63–1.31) | .60 | 0.79 (0.51–1.24) | .31 | 1.14 (0.58–2.24) | .70 |
75.0–80.9 | Reference | Reference | Reference | ||||
81.0–90.0 | 0.92 (0.61–1.39) | .69 | 0.87 (0.51–1.24) | .59 | 1.00 (0.50–2.01) | .99 | |
>90.0 | 1.23 (0.81–2.04) | .38 | 1.38 (0.81–2.35) | .23 | 0.60 (0.20–1.85) | .38 | |
Mean arterial pressure, mm Hg | <93.5 | 1.01 (0.66–1.55) | .95 | 1.10 (0.66–1.82) | .71 | 1.17 (0.57–2.43) | .66 |
93.5–101.9 | Reference | Reference | Reference | ||||
102.0–111.9 | 1.01 (0.68–1.49) | .96 | 1.25 (0.73–2.13) | .42 | 1.62 (0.76–3.45) | .21 | |
≥112.0 | 1.23 (0.80–1.88) | .34 | 1.85 (1.09–3.12) | .02 | 0.63 (0.25–1.61) | .33 | |
Pulse pressure, mm Hg | ≤50.0 | 0.91 (0.59–1.40) | .66 | 0.65 (0.38–1.12) | .12 | 1.55 (0.68–3.52) | .30 |
51.0–61.0 | Reference | Reference | Reference | ||||
62.0–75.0 | 0.90 (0.58–1.35) | .57 | 0.66 (0.39–1.11) | .12 | 1.51 (0.67–3.40) | .33 | |
>75.0 | 1.00 (0.68–1.50) | .97 | 0.87 (0.55–1.37) | .55 | 1.17 (0.49–2.81) | .72 |
Abbreviations: BP, blood pressure; CI, confidence interval; HR, hazard ratio; MMSE, Mini‐Mental State Examination.
Model 1 is adjusted for age (continuous by 1 year) and sex.
Model 2 is adjusted for age (continuous by 1 year), sex, educational level (ordered; 1 level more), social activity (optimal vs suboptimal), family support (poor vs normal), body mass index (continuous by 1‐year increment), alcohol intake (ordered; three categories), depression (geriatric depression scale score >6 vs ≤6), hypercholesterolemia (yes vs no), diabetes mellitus type 2 (yes vs no), cardiovascular disease (yes vs no), and antihypertensive medications (yes vs no). Analyses for the total sample are further adjusted for cognitive impairment (MMSE <24 vs ≥24).
Figure depicts the best‐fitting models for the associations of BP variables with all‐cause mortality, as derived by the multivariate fractional polynomial analysis. The results showed a U‐shaped effect of SBP and MAP on all‐cause mortality in the total sample (Figure a) with significantly increased risk in both the highest and the lowest values. For MAP, this pattern was observed only among patients with MMSE score <24, whereas for the subcohort with a normal range on the MMSE, no association of MAP with mortality was revealed. Regarding SBP, the U‐shaped trend was observed in both the cognitively impaired and cognitively normal group of individuals, but was stronger and significant only in the former. Nonsignificant trends emerged for DBP and PP in the total sample or the MMSE subcohorts.
Figure 1.
Risk of all‐cause mortality according to systolic blood pressure, diastolic blood pressure, mean arterial pressure, and pulse pressure. Graphs refer to (a) the total sample (n=660), (b) individuals with a Mini‐Mental State Examination (MMSE) score <24 (n=362), and (c) individuals with an MMSE score ≥24 (n=298), and correspond to the best‐fitting models derived from fractional polynomial analysis adjusted for age (continuous by 1 year), sex (males vs females), education level (ordered; 1 level more), social activity (optimal vs suboptimal), family support (poor vs normal), body mass index (continuous), alcohol intake (ordered; three categories), geriatric depression scale score (>6 vs ≤6), hypercholesterolemia (yes vs no), diabetes mellitus type 2 (yes vs no), cardiovascular disease (yes vs no), and antihypertensive medications (yes vs no). The blue lines represent hazard ratios by levels of blood pressure and shaded areas correspond to the respective 95% confidence intervals. Mean values are used as references.
Discussion
In this study, based on a community‐dwelling population of elderly individuals followed for a period of 8 years, a differential by cognitive function association of peripheral BP variables with all‐cause mortality emerged. In fully adjusted analyses, low and high MAP and SBP values were associated with increased all‐cause mortality only among individuals with MMSE scores indicative of cognitive impairment. DBP and PP were not found to independently affect mortality in this study, whereas no significant association was found among patients with MMSE scores in the normal range.
Only four studies have to our knowledge previously examined the BP‐mortality association in view of cognitive performance.22, 23, 24, 25 Two of them, population‐based studies including individuals 65 years and older, showed that only among cognitively impaired patients, low DBP (lowest quartile and tertile compared with second and median, respectively) was associated with mortality22, 24 whereas one of them that also presented MAP and PP22 found that in patients with MMSE <24, significant associations emerged for low MAP and elevated PP. Similarly, a Swedish cohort of a relatively older population (≥75 years)23 reported that low DBP increased mortality in patients with cognitive impairment and not in those with normal cognitive scores; however, this study also found the same pattern for SBP. Lastly, in a recent report of the oldest elderly (>80 years)25 researchers found that both low and high SBP values were associated with increased all‐cause mortality only in the subcohort of patients with the most severe form of cognitive impairment (MMSE 0–10). In our population of elderly (older than 60 years) inhabitants of a rural community characterized by low educational level, our findings on a susceptibility of cognitively impaired individuals to the effects of BP on mortality are in line with past literature. What is differentiated, however, is the profile of this susceptibility. As BP characteristics change with increasing age, age differences in the study populations could cause discrepancies.33 Furthermore, BP variables have been treated categorically in all studies but different categorizations were implemented (dichotomous analysis, tertiles, quartiles, quintiles) corresponding to diverse cutoff points, which could be responsible for the variations in the findings between studies. In this study we have additionally to the traditional quartile analysis, encompassed a fractional polynomial analysis to assess BP nonlinear effects.34 This analysis revealed some significant associations for extreme values that were shadowed in the quartiles approach.
It is possible that this increased vulnerability of patients with cognitive impairment to the detrimental effects of low and high SBP and MAP is mediated by underlying CVD burden. Cognitive decline could be the expression of microvascular or macrovascular cerebral pathology35 possibly in the context of generalized vascular disease.36 CVD is one of the most common causes of death among patients with dementia37 and, as we recently showed in this sample, cognitive impairment is also associated with increased risk of cardiovascular mortality.19 An intriguing explanation would include poor cerebral autoregulation among patients with severe cognitive impairment. Although loss of cerebral autoregulation has been systematically described in animal models of Alzheimer's disease, the evidence in humans is not robust.38, 39, 40 Patients with cognitive impairment in this sample were also characterized by increased PP values, indicating potentially advanced arterial stiffness, which has also been found to contribute to the loss of cerebral vasomotor reactivity.41 If this explanation is true, presence of dementia could indicate susceptibility to hypoperfusion and ischemic changes or uncontrolled flow and hemorrhagic injuries in the cerebral arteriolar bed caused by extremely low or high SBP and MAP values, respectively, especially under presence of arterial stiffness.42, 43, 44 It should be take into account, however, that patients with cognitive impairment who receive suboptimal care are also at higher risk for noncompliance to antihypertensive treatment,45 thus increasing their vulnerability to the negative effects of hypertension. Lastly, regarding the association of low SBP and MBP values with mortality, the potential confounding role of cachexia at terminal stages of diseases should be taken into consideration, as underweight is related to decreased BP.46 Yet, fully adjusted models controlled for BMI levels, whereas the finding was evident only among cognitively normal patients who are less likely to experience cachexia caused by terminal diseases, compared with those with cognitive impairment.
Notably, baseline differences between the two groups of participants in our study (MMSE <24 and ≥24) were evident. Firstly, regarding sociodemographic and lifestyle characteristics, patients with cognitive impairment were more likely women and of older age, whereas smoking and alcohol consumption were more common among individuals without cognitive impairment, possibly as a result of a more active social life. The lower educational level and the higher prevalence of depressive symptoms in patients with cognitive impairment was anticipated, as it is well‐established in previous literature47, 48 and we have shown in the same sample that they both represent independent risk factors for cognitive impairment.26 In terms of comorbidity, a higher rate of diabetes mellitus, as well as a higher proportion of antihypertensive medication use, likely implying increased burden of hypertension, were recorded among patients with cognitive impairment, as has been previously described.49, 50 Even though fully adjusted analyses controlled for the aforementioned factors, it cannot be precluded that these differences partially explain the findings. In particular, older age, female sex, lack of physical activity, depression, diabetes mellitus, and hypertension are all factors that are associated with arterial stiffness in the elderly.51, 52, 53, 54 Therefore, the observed susceptibility of cognitively impaired patients might be attributed to the higher prevalence of arterial stiffness. This is further supported by the fact that among BP parameters, only higher PP values were observed in cognitively impaired patients. After adjusting for age, however, the association between PP and cognitive impairment was completely attenuated (results not shown), reflecting a confounding role of aging. Published literature has shown inconsistent results regarding the PP‐cognition association.7 Under loss of arterial elasticity, impaired microvascular reactivity is observed and thus the negative effects of extremely high or low SBP and MAP values might become more prominent.55
Study Limitations
The study findings should be interpreted in view of certain limitations. Firstly, the use of surrogate formulae for brachial MAP and PP calculation and the lack of assessment of central values should be considered a drawback, and no invasive method was available for their estimation. Secondly, the unavailability of 24‐hour BP monitoring did not allow for the investigation of the effect of BP variability over time, including orthostatic hypotension in the research question. Thirdly, no neuroimaging data were available for assessment of cerebral small vessel disease to investigate the potential underlying pathology of the associations. We should also note the unavailability of other measures of socioeconomic measurements, such as yearly income, which could give more information on the background of the individuals. Additionally, despite the fact that the Greek version of the MMSE has been previously validated in the elderly,31 the potential misclassification of cognitive impairment cases as a result of a lack of clinical assessment cannot be precluded; the extremely low educational level of the participants (61% of the total sample <6 years of education) could possibly explain the high proportion of patients scoring <24 on the MMSE. Lastly, because of a lack of statistical power, no stratification by age and sex could be conducted, whereas despite the availability of cause‐specific mortality, analyses for cardiovascular and noncardiovascular mortality could not be performed because the low number of death events in the subcohorts of cognitively impaired and cognitively normal individuals would hamper the reliability of the analysis. All analyses are, however, adjusted for a wide range of major sociodemographic and medical confounders.
Study Strengths
The strengths of the study include the population‐based design including community‐dwelling individuals, the high participation rates among a homogeneous sample of those born and still residing in Velestino, the relatively long follow‐up period of 8 years, and the high‐quality data collection procedure by physicians familiar with the residents. Furthermore, in contrast to previous studies, the use of fractional polynomial analysis in addition to a categorical analysis for BP allowed the emergence of relationships between extreme values of BP and mortality that were concealed in quartiles.
Conclusions
Our findings show that the association between BP and all‐cause mortality in elderly individuals is dependent on the level of cognitive function. Patients with cognitive impairment seem to be more susceptible to the detrimental effects of low and high SBP and MAP. Future research should focus on identifying the hemodynamic consequences of dementia in cerebral circulation in humans, as well as on unraveling the complex interplay between vascular pathology and cognition. Most importantly, given the lack of evidence on the effects of antihypertensive treatment among individuals with dementia, large prospective studies are needed to confirm the differential effects of BP by cognitive function level to determine ideal BP values for these individuals and better target patients who could benefit from BP‐lowering interventions.
Funding
None.
Conflicts of Interest
None declared.
Acknowledgments
Contributions are acknowledged to Anastasia Anastasiou, MD, PhD, Konstantinos Katsiardanis, MD, and Kalliopi‐Penelopi Katsiardani, MD, in the collection of primary data, as well as Theodoros Michelakos, MD, and Christina Perlepe, MD, in cleaning primary and death certificate data. The authors would also like to thank the Municipality of Velestino for contributing in the planning and data collection phase, the inhabitants of Velestino, and Onassis Public Benefit Foundation.
J Clin Hypertens (Greenwich). 2017;19:161–169. DOI: 10.1111/jch.12880. © 2016 Wiley Periodicals, Inc.
References
- 1. Kearney PM, Whelton M, Reynolds K, et al. Global burden of hypertension: analysis of worldwide data. Lancet. 2005;365:217–223. [DOI] [PubMed] [Google Scholar]
- 2. Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case‐control study. Lancet. 2004;364:937–952. [DOI] [PubMed] [Google Scholar]
- 3. Protogerou AD, Safar ME, Iaria P, et al. Diastolic blood pressure and mortality in the elderly with cardiovascular disease. Hypertension. 2007;50:172–180. [DOI] [PubMed] [Google Scholar]
- 4. van BT, Gussekloo J, Westendorp RG, Blauw GJ. In a population‐based prospective study, no association between high blood pressure and mortality after age 85 years. J Hypertens. 2006;24:287–292. [DOI] [PubMed] [Google Scholar]
- 5. Mattila K, Haavisto M, Rajala S, Heikinheimo R. Blood pressure and five year survival in the very old. Br Med J (Clin Res Ed). 1988;296:887–889. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Prince M, Bryce R, Albanese E, et al. The global prevalence of dementia: a systematic review and metaanalysis. Alzheimers Dement. 2013;9:63–75. [DOI] [PubMed] [Google Scholar]
- 7. van ST, Protogerou AD, Henry RM, et al. Association between arterial stiffness, cerebral small vessel disease and cognitive impairment: a systematic review and meta‐analysis. Neurosci Biobehav Rev. 2015;53:121–130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Taylor C, Tillin T, Chaturvedi N, et al. Midlife hypertensive status and cognitive function 20 years later: the Southall and Brent revisited study. J Am Geriatr Soc. 2013;61:1489–1498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gutierrez J, Elkind MS, Cheung K, et al. Pulsatile and steady components of blood pressure and subclinical cerebrovascular disease: the Northern Manhattan Study. J Hypertens. 2015;33:2115–2122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Allan CL, Zsoldos E, Filippini N, et al. Lifetime hypertension as a predictor of brain structure in older adults: cohort study with a 28‐year follow‐up. Br J Psychiatry. 2015;206:308–315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Nelson L, Gard P, Tabet N. Hypertension and inflammation in Alzheimer's disease: close partners in disease development and progression!. J Alzheimers Dis. 2014;41:331–343. [DOI] [PubMed] [Google Scholar]
- 12. Kling MA, Trojanowski JQ, Wolk DA, et al. Vascular disease and dementias: paradigm shifts to drive research in new directions. Alzheimers Dement. 2013;9:76–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Qiu C, Fratiglioni L. A major role for cardiovascular burden in age‐related cognitive decline. Nat Rev Cardiol. 2015;12:267–277. [DOI] [PubMed] [Google Scholar]
- 14. Novak V, Hajjar I. The relationship between blood pressure and cognitive function. Nat Rev Cardiol. 2010;7:686–698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Wu CY, Chou YC, Huang N, et al. Cognitive impairment assessed at annual geriatric health examinations predicts mortality among the elderly. Prev Med. 2014;67:28–34. [DOI] [PubMed] [Google Scholar]
- 16. Dewey ME, Saz P. Dementia, cognitive impairment and mortality in persons aged 65 and over living in the community: a systematic review of the literature. Int J Geriatr Psychiatry. 2001;16:751–761. [DOI] [PubMed] [Google Scholar]
- 17. O'Donnell M, Teo K, Gao P, et al. Cognitive impairment and risk of cardiovascular events and mortality. Eur Heart J. 2012;33:1777–1786. [DOI] [PubMed] [Google Scholar]
- 18. Bassuk SS, Wypij D, Berkman LF. Cognitive impairment and mortality in the community‐dwelling elderly. Am J Epidemiol. 2000;151:676–688. [DOI] [PubMed] [Google Scholar]
- 19. Georgakis MK, Papadopoulos FC, Protogerou AD, et al. Comorbidity of cognitive impairment and late‐life depression increase mortality: results from a cohort of community‐dwelling elderly individuals in rural Greece. J Geriatr Psychiatry Neurol. 2016;29:195–204. [DOI] [PubMed] [Google Scholar]
- 20. Beishon LC, Harrison JK, Harwood RH, et al. The evidence for treating hypertension in older people with dementia: a systematic review. J Hum Hypertens. 2014;28:283–287. [DOI] [PubMed] [Google Scholar]
- 21. van dWV, Logan P, Conroy S, et al. Antihypertensive treatment in people with dementia. J Am Med Dir Assoc. 2014;15:620–629. [DOI] [PubMed] [Google Scholar]
- 22. Cacciatore F, Abete P, de SD, et al. Mortality and blood pressure in elderly people with and without cognitive impairment. Gerontology. 2005;51:53–61. [DOI] [PubMed] [Google Scholar]
- 23. Guo Z, Viitanen M, Winblad B. Low blood pressure and five‐year mortality in a Stockholm cohort of the very old: possible confounding by cognitive impairment and other factors. Am J Public Health. 1997;87:623–628. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Post HG, Smulders YM, Maier AB, et al. Relation between blood pressure and mortality risk in an older population: role of chronological and biological age. J Intern Med. 2015;277:488–497. [DOI] [PubMed] [Google Scholar]
- 25. Weidung B, Littbrand H, Nordstrom P, et al. The association between SBP and mortality risk differs with level of cognitive function in very old individuals. J Hypertens. 2016;34:745–752. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Michelakos T, Kousoulis AA, Katsiardanis K, et al. Serum folate and B12 levels in association with cognitive impairment among seniors: results from the VELESTINO study in Greece and meta‐analysis. J Aging Health. 2013;25:589–616. [DOI] [PubMed] [Google Scholar]
- 27. Papadopoulos FC, Petridou E, Argyropoulou S, et al. Prevalence and correlates of depression in late life: a population based study from a rural Greek town. Int J Geriatr Psychiatry. 2005;20:350–357. [DOI] [PubMed] [Google Scholar]
- 28. Jellinger PS, Smith DA, Mehta AE, et al. American association of clinical Endocrinologists' guidelines for management of Dyslipidemia and prevention of atherosclerosis. Endocr Pract. 2012;18(suppl 1):1–78. [DOI] [PubMed] [Google Scholar]
- 29. American DA. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2014;37(suppl 1):S81–S90. [DOI] [PubMed] [Google Scholar]
- 30. Gauer O. Kreislauf des blutes. In: Lehrbuch der Physiologie des Menschen, edited by Landois L, Rosemann H‐U. Munich‐Berlin, Germany: Urban & Schwarzenberg, 1960, pp. 95–135. [Google Scholar]
- 31. Fountoulakis K, Tsolaki M, Chantzi H, Kazis A. Mini‐Mental State Examination (MMSE): a validation study in Greece. Am J Alzheimers Dis Other Demen. 2000;15:342–345. [Google Scholar]
- 32. Fountoulakis KN, Tsolaki M, Iacovides A, et al. The validation of the short form of the Geriatric Depression Scale (GDS) in Greece. Aging. 1999;11:367–372. [DOI] [PubMed] [Google Scholar]
- 33. Lernfelt B, Svanborg A. Change in blood pressure in the age interval 70‐90. Late blood pressure peak related to longer survival. Blood Press. 2002;11:206–212. [DOI] [PubMed] [Google Scholar]
- 34. Royston P, Ambler G, Sauerbrei W. The use of fractional polynomials to model continuous risk variables in epidemiology. Int J Epidemiol. 1999;28:964–974. [DOI] [PubMed] [Google Scholar]
- 35. Kapasi A, Schneider JA. Vascular contributions to cognitive impairment, clinical Alzheimer's disease, and dementia in older persons. Biochim Biophys Acta. 2016;1862:878–886. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Gallego J, Martinez‐Vila E. Asymptomatic cerebrovascular disease and systemic diagnosis in stroke, atherothrombosis as a disease of the vascular tree. Cerebrovasc Dis. 2005;20(suppl 2):1–10. [DOI] [PubMed] [Google Scholar]
- 37. Attems J, Konig C, Huber M, et al. Cause of death in demented and non‐demented elderly inpatients; an autopsy study of 308 cases. J Alzheimers Dis. 2005;8:57–62. [DOI] [PubMed] [Google Scholar]
- 38. Claassen JA, Zhang R. Cerebral autoregulation in Alzheimer's disease. J Cereb Blood Flow Metab. 2011;31:1572–1577. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. den Abeelen AS, Lagro J, van Beek AH, Claassen JA. Impaired cerebral autoregulation and vasomotor reactivity in sporadic Alzheimer's disease. Curr Alzheimer Res. 2014;11:11–17. [DOI] [PubMed] [Google Scholar]
- 40. Gommer ED, Martens EG, Aalten P, et al. Dynamic cerebral autoregulation in subjects with Alzheimer's disease, mild cognitive impairment, and controls: evidence for increased peripheral vascular resistance with possible predictive value. J Alzheimers Dis. 2012;30:805–813. [DOI] [PubMed] [Google Scholar]
- 41. Rucka D, Marek J, Rucklova Z, et al. Arterial stiffening contributes to impairment of cerebrovascular reactivity in patients with coronary artery disease without carotid stenosis. Physiol Res. 2015;64:335–343. [DOI] [PubMed] [Google Scholar]
- 42. Sabayan B, Jansen S, Oleksik AM, et al. Cerebrovascular hemodynamics in Alzheimer's disease and vascular dementia: a meta‐analysis of transcranial Doppler studies. Ageing Res Rev. 2012;11:271–277. [DOI] [PubMed] [Google Scholar]
- 43. van BA, Sijbesma JC, Jansen RW, et al. Cortical oxygen supply during postural hypotension is further decreased in Alzheimer's disease, but unrelated to cholinesterase‐inhibitor use. J Alzheimers Dis. 2010;21:519–526. [DOI] [PubMed] [Google Scholar]
- 44. Hajjar I, Goldstein FC, Martin GS, Quyyumi AA. Roles of arterial stiffness and blood pressure in hypertension‐associated cognitive decline in healthy adults. Hypertension. 2016;67:171–175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Salas M, In't Veld BA, van der Linden PD, et al. Impaired cognitive function and compliance with antihypertensive drugs in elderly: the Rotterdam Study. Clin Pharmacol Ther. 2001;70:561–566. [DOI] [PubMed] [Google Scholar]
- 46. Hosseini M, Baikpour M, Yousefifard M, et al. Blood pressure percentiles by age and body mass index for adults. Excli J. 2015;14:465–477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Diniz BS, Butters MA, Albert SM, et al. Late‐life depression and risk of vascular dementia and Alzheimer's disease: systematic review and meta‐analysis of community‐based cohort studies. Br J Psychiatry. 2013;202:329–335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Beydoun MA, Beydoun HA, Gamaldo AA, et al. Epidemiologic studies of modifiable factors associated with cognition and dementia: systematic review and meta‐analysis. BMC Public Health. 2014;14:643. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Ninomiya T. Diabetes mellitus and dementia. Curr Diab Rep. 2014;14:487. [DOI] [PubMed] [Google Scholar]
- 50. Iadecola C. Hypertension and dementia. Hypertension. 2014;64:3–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Prenner SB, Chirinos JA. Arterial stiffness in diabetes mellitus. Atherosclerosis. 2015;238:370–379. [DOI] [PubMed] [Google Scholar]
- 52. Redfield MM, Jacobsen SJ, Borlaug BA, et al. Age‐ and gender‐related ventricular‐vascular stiffening: a community‐based study. Circulation. 2005;112:2254–2262. [DOI] [PubMed] [Google Scholar]
- 53. Endes S, Schaffner E, Caviezel S, et al. Physical activity is associated with lower arterial stiffness in older adults: results of the SAPALDIA 3 Cohort Study. Eur J Epidemiol. 2016;31:275–285. [DOI] [PubMed] [Google Scholar]
- 54. Seldenrijk A, van HH, van MH, et al. Depression, anxiety, and arterial stiffness. Biol Psychiatry. 2011;69:795–803. [DOI] [PubMed] [Google Scholar]
- 55. Mitchell GF, Vita JA, Larson MG, et al. Cross‐sectional relations of peripheral microvascular function, cardiovascular disease risk factors, and aortic stiffness: the Framingham Heart Study. Circulation. 2005;112:3722–3728. [DOI] [PubMed] [Google Scholar]