Introduction

The number of elderly people with dementia is increasing, especially in developed countries.1 Many studies have considered the association between hypertension and cognitive impairment, including mild cognitive impairment and dementia, in various generations.2, 3, 4 According to several longitudinal studies, high blood pressure (BP) in midlife significantly increases the risk of cognitive impairment and dementia in later life.5, 6, 7, 8 Several previous studies have also noted that in later life, around age 70 years, hypertension and/or elevated BP may be significant predictors of a progressive decline in cognitive performance. However, the influence of BP on the progression of cognitive impairment has not been aggressively investigated in the general population, including those over age 80 years.

The association between later-life BP and cognitive impairment may be because of the very strong effect of aging on cognitive functioning, hypertensive treatment status and the emergence of comorbidities, such as cardiovascular diseases.3, 10, 11, 12 Additionally, hypertension is associated with body mass index (BMI), smoking, alcohol consumption and salt intake.13, 14, 15, 16, 17 Furthermore, cognitive functioning in later life is associated with social factors, especially the level of activity, including the frequency of going outdoors.18, 19 In most cross-sectional studies, there are inconclusive data regarding the association between hypertension and cognitive functioning in later life.20, 21

Our primary objective was to investigate differences in the influence of aging on the association between cognitive functioning and BP in later life in a narrow age range cohort study. This is known as the SONIC (Septuagenarians, Octogenarians, Nonagenarians Investigation with Centenarians) study, a longitudinal cohort study targeting individuals of older ages (70, 80, 90 and over 100 years) of the general Japanese population and including medical and social variables.

Methods

Study population

This study was a cross-sectional examination conducted as a baseline assessment of a prospective study called SONIC, which investigated health and longevity. The participants were all volunteers living independently, and included 1000 individuals (excluding institutionalized individuals) aged 69–71 years (479 men, 521 women) and 973 individuals aged 79–81 years (456 men, 517 women) from four areas of both western and eastern, as well as urban and rural parts of Japan: Itami City, Hyogo (Western urban); Asago City, Hyogo (Western rural); Itabashi ward, Tokyo (Eastern urban); and Nishitama County, Tokyo (Eastern rural). Participants were randomly selected from the local resident registry and enrolled between 2010 and 2011.22 We used the alternative narrow age range cohort approach to a cross-sectional design, in which covariances among age-related variables in cross-sectional studies are highly confounded with regard to inferences of association among rates of change within individuals. This is because covariances can result from a number of sources, including average population age-related differences (fixed age effects) and initial individual differences and effects of aging (random age effects).23

The study protocol was approved by the Institutional Review Board of Osaka University Graduate School of Medicine, Dentistry and Human Sciences and the Tokyo Metropolitan Geriatric Hospital and Institute of Gerontology (approval number 266, H22-E9, 22 018 and 38, respectively). All participants provided written informed consent to participate.

Measurements

BP measurement and classification

BP of the left and right arm was measured two times by a physician or trained nurse using a mercury sphygmomanometer or electronic monitor while the participant was seated after at least 1 min of rest. The mean of the two measurements of both arms was used in the analysis. BP measurements were classified into two categories according to the criteria of the Japanese Society of Hypertension guidelines for the management of hypertension (JSH 2009) (Supplementary Figure S1).24 Based on these guidelines, the uncontrolled BP group was defined as those with a systolic blood pressure (SBP) of 140 mm Hg or greater and a diastolic BP (DBP) of 90 mm Hg or greater, and the controlled BP group was defined as those with an SBP lower than 140 mm Hg and a DBP lower than 90 mm Hg. Diagnosis of hypertension was based on BP values greater than 140/90 mm Hg and/or receiving antihypertensive treatment at the first contact.

Other factor measurements

Participants were interviewed at enrollment using questionnaires covering demographics, clinical information and psychosocial characteristics. Blood samples were drawn after overnight fasting. Levels of total cholesterol, high-density lipoprotein cholesterol, triglycerides, serum albumin and fasting/casual blood glucose were determined by biochemical examinations. Diabetes mellitus (DM) was defined by fasting blood glucose concentrations ⩾7.0 mmol l−1 (126 mg dl−1), casual blood glucose concentrations ⩾11.1 mmol l−1 (200 mg dl−1), HbA1c (National Glycohemoglobin Standardization Program) ⩾6.5% or taking medications for diabetes according to the World Health Organization criteria for epidemiologic studies of DM. Dyslipidemia was defined as a low-density lipoprotein cholesterol ⩾3.62 mmol l−1 (140 mg dl−1), high-density lipoprotein cholesterol <1.03 mmol l−1 (40 mg dl−1), triglyceride ⩾1.69 mmol l−1 (150 mg dl−1) and/or medications for dyslipidemia. Smoking behavior was based on a questionnaire that classified subjects into three categories: never, past and current. Alcohol consumption was also classified into three categories by ethanol units: never, current (1–3 days and <3 units per week) and excessive current (⩾3 days and ⩾3 units per week). Salt and fat intake was assessed with a Brief-Type Self-Administered Diet History Questionnaire, which was completed by each participant at home and checked by a trained research staff member during the medical examination.

Assessment of cognitive functioning

We used the Japanese version of the Montreal Cognitive Assessment (MoCA-J) as a general index of the cognitive status.25 MoCA is a brief cognitive screening tool for detecting mild cognitive impairment in elderly people.26 MoCA consists of a one-page, 30-point test administered by trained geriatricians and psychologists, with higher scores reflecting more favorable cognitive functioning, and assesses the following domains of cognition: visuospatial ability (3 points), naming task (3 points), attention task (6 points), language (3 points), abstraction task (2 points), delayed recall (5 points) and orientation (6 points). The MoCA-J showed favorable reliability and better validity for predicting mild cognitive impairment than conventional cognitive tests.25 We used the MoCA-J total score as a predictor of cognitive functioning.

Statistical analyses

In this cross-sectional examination, descriptive data were summarized using the mean±s.d. and proportions. Student’s t-test for independent groups was used, and the comparison of proportions among groups was performed using a χ2 test. Multiple linear regression models were produced for each group to calculate the standardized regression coefficients (β) expressing independent statistical associations between variables. In these analyses, data were stratified by age, BP category and antihypertensive treatment. All statistical analyses were performed with SPSS Statistics 21 (IBM Japan, Tokyo, Japan). All reported P-values are two-tailed, and P<0.05 was considered significant.

Results

Clinical and social characteristics between participants aged 70 and 80 years

The characteristics of the participants aged 70 and 80 years are shown in Table 1. A total of 1973 participants answered the questionnaire and underwent the examinations and were analyzed after we excluded missing BP data (Supplementary Figure S1).

Table 1 Characteristics of study population aged between 70 and 80 years
Full size table

SBP and PP at age 80 years were higher compared with that at age 70 years. In contrast, DBP at 80 years was significantly lower compared with that at age 70 years. The proportions of those with hypertension and taking medication for hypertension were also higher at age 80 years. Specifically, the proportions of major antihypertensive drugs, including calcium channel blockers (CCB), angiotensin receptor blockers (ARB) and diuretics, were higher at age 80 years compared with that at age 70 years. The proportions of those with DM and dyslipidemia at age 70 years exhibited no significant differences between the age groups. Serum total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglyceride, serum albumin and BMI at age 70 years were also higher compared with that at age 80 years. Salt and fat intake evaluated by brief-type self-administered diet history questionnaire showed no significant differences between the age groups. Smoking habits, alcohol consumption and frequency of going outdoors were significantly different between the age groups. The MoCA-J total score at age 70 years was higher compared with that at age 80 years. A similar trend was observed in subscale scores, except for delayed recall (Supplementary Table S1).

Participant characteristics by sex are shown in Supplementary Table S2. At age 70 years, SBP, DBP, the proportion with hypertension and those taking medication for hypertension were higher in men compared with that in women, whereas at age 80 years, SBP and PP were higher in women compared with that in men. The proportion with DM was higher in men compared with that in women at both age 70 and 80 years. The proportion with dyslipidemia was lower in men compared with that in women at both 70 and 80 years. The MoCA-J total score was lower in men compared with that in women at age 70 years but not at age 80 years. Smoking habits and alcohol consumption were significantly different between men and women. The frequency of going outdoors was significantly higher in women compared with that in men at both age 70 and 80 years.

Comparison of participant characteristics divided by BP category

Table 2 displays the clinical characteristics according to BP categories within each age group. In participants at both age 70 and 80 years, significant differences between BP categories can be observed. SBP, DBP, and PP in the uncontrolled BP group were higher than those in the controlled BP group. The proportion with hypertension in the uncontrolled BP group was also higher than that in the controlled BP group. The proportion with DM in the uncontrolled BP group was higher than that in the controlled BP group at age 70 years but not at 80 years. Dyslipidemia and low-density lipoprotein cholesterol showed no significant differences at either 70 or 80 years. On the other hand, high-density lipoprotein cholesterol in the controlled BP group was higher compared with that in the uncontrolled BP group at age 70 years, whereas at age 80 years, the uncontrolled BP group showed higher values compared with the controlled BP group. BMI was higher in the uncontrolled BP group compared with that in the controlled BP group at both 70 and 80 years. Salt and fat intake showed no significant differences between BP categories. Smoking and alcohol consumption were significantly different between BP categories at age 70 years but not at age 80 years.

Table 2 Characteristics of study population according to BP categories within each age group
Full size table

The MoCA-J total score and subscales were significantly different between BP categories within each age group. More specifically, at age 70 years, the MoCA-J total score and two subscales, attention and language tasks, in the uncontrolled BP group were lower compared with those in the controlled BP group, whereas the MoCA-J total score showed no significant difference at age 80 years, except in the two subscales of visuospatial/executive functioning and naming tasks (Supplementary Table S1).

Supplementary Table S3 shows BP and cognitive functioning according to the treatment status within each age group. At age 70 years, SBP, DBP and PP in the treated group were significantly higher compared with that in the untreated group. In contrast, SBP, DBP and PP showed no significant differences at age 80 years. The MoCA-J total score was significantly different at age 80 years; in other words, the score in the treated group was significantly higher compared with that in the untreated group, although there was no significant difference in SBP between the treated and untreated groups at age 80 years (Supplementary Table S3). Similar trends were observed among subjects who were or were not taking ARB or CCB in both age groups (Supplementary Figure S3).

Effects of BP on cognitive functioning

The regression coefficients are shown in Table 3 and Supplementary Table S4. To examine the main effect of SBP on the cognitive functioning, we stratified the BP control status and antihypertensive treatment for each age group and used the same model in the stratified analysis. Because this study used a narrow age range design, there was no SBP × age interaction for cognitive functioning within each age group.

Table 3 Standardized multiregression coefficients (β) of age and BP category as predictors of MoCA-J total score
Full size table

At age 70 years, SBP, BMI, smoking, sex and frequency of going outdoors were correlated with cognitive functioning. In the case of uncontrolled or treated BP, a negative association between SBP and cognitive functioning was found. However, in the controlled BP or untreated BP cases, there was no significant correlation between SBP and cognitive functioning; in contrast, DM was independently correlated with cognitive functioning in controlled BP cases at age 70 years. At age 80 years, on the other hand, SBP and other clinical factors were not significantly correlated with cognitive functioning. Serum albumin and frequency of going outdoors were significantly correlated with cognitive functioning. Additionally, in the case of controlled or treated BP, clinical factors were not correlated with cognitive functioning, except for the no-hypertensive treatment model.

DBP and PP were also examined in each age group, and we used the same model for ease of comparison with the subsequent BP analysis (Supplementary Table S5). Although DBP was not a significant predictor of cognitive functioning, the pattern of association between PP and cognitive functioning was similar to that observed with SBP.

Discussion

A high BP in middle age has been suggested to be a predictive indicator of cognitive functioning and dementia.5, 6, 7 We aimed to clarify the association between high BP and cognitive functioning in a general Japanese elderly population using the SONIC study, a large-scale longitudinal cohort study of elderly Japanese individuals in the general population.22 In general, age is one of the crucial predictors of an elevated SBP and of cognitive functioning. Therefore, we investigated a cohort study involving participants within a narrow age range (the SONIC study) to specify predictors under conditions that minimize the influence of age.23 This narrow age approach in the cross-sectional study may clarify differences in the influence of aging on the association between cognitive functioning and BP among those aged 70 and 80 years, accounting for important potential confounders of not only age but also medication for hypertension and dyslipidemia, BMI, smoking, excessive alcohol consumption, serum albumin as a predictor of the nutritional status and frequency of going outdoors as a predictor of activities.23

The most important finding of this study was a high SBP at age 70 years, especially in the case of uncontrolled BP (SBP ⩾140 mm Hg or DBP ⩾90 mm Hg), but not in the case of controlled BP, and this result may be independently associated with reduced cognitive functioning. This association indicates that high BP in later life is a risk factor for cognitive decline at age 70 years. Several cross-sectional studies suggested that uncontrolled BP was associated with reduced cognitive functioning compared with normal BP around the age of 70 years.17, 27 Our study supported these findings of an apparent age-related distinction. This is consistent with our primary hypothesis, which is attributed to the effect of aging on the association between BP and cognitive functioning. When we stratified the effect of antihypertensive treatment (Supplementary Table S4), a higher SBP was associated with reduced cognitive functioning in the treated BP group at age 70 years compared with the untreated group. In the previous randomized controlled trial, patients who had previously been treated with antihypertensive medication had lower mini-mental state examination scores compared with their untreated counterparts; however, they had a higher baseline SBP.28 Our findings suggest similar results, whereby the level of control of SBP may be more important than taking a class on antihypertensive medication for the prevention of cognitive decline.

In contrast, no such associations were observed in participants aged 80 years. Although previous cross-sectional studies indicated a conflicting association between high BP and cognitive functioning, with positive and negative associations, most studies varied in terms of the ages and ranges of BP values of the participants; therefore, age may strongly influence the association between BP and cognitive functioning.3, 8, 21, 28 Several prospective studies demonstrated that middle-aged participants with hypertension still had a higher risk of cognitive decline and dementia in later life.5, 6, 7 Additionally, one of the previous randomized controlled trials, which assessed the effects of antihypertensive intervention on cognition in patients aged 80 years or older, did not reveal a significant influence of BP on cognition.29 Our study clearly demonstrated a positive association between high SBP and reduced cognitive functioning at age 70 years, but not at age 80 years, in a general Japanese population. However, when an individual reaches his or her 80s, BP is controlled for many reasons other than to prevent cognitive decline. Also noteworthy is that our study does not suggest that BP control is unimportant after age 80 years because the beneficial effect of taking medication for hypertension (i.e., CCB and ARB) on cognitive functioning has been found only at age 80 years; nevertheless, the cognitive functioning of SONIC participants appeared less likely to be associated with BP at age 80 years. Recently, another Japanese study that included middle-aged participants reported the positive association of home BP with cognitive decline, but only in participants who were not prospectively treated with antihypertensive medication.9 The age range of this study was relatively younger than that of the SONIC participants; therefore, medication adherence in this study may be better than in the SONIC study population (which recruited late-life volunteers). Because this Japanese study longitudinally investigated the relationship between home BP and cognitive decline,9 the evidence level in this study was very high. Our present study confirmed the aforementioned study results in an elderly population. The SONIC study will also provide causal information about the association in those over 70 years of age based on future follow-up data collected every 3 years.

This study also demonstrated that DBP was not associated with cognitive functioning, whereas PP may have an important role in the progression of functional disability at age 70 years. Previous studies suggested that DBP and PP were also associated with a higher risk of cognitive decline; however, no such associations were observed at age 80 years.10, 30 These findings may suggest that SBP control at 70 years is more important in predicting cognitive functioning than DBP levels.

Frequency of going outdoors was the only variable to exhibit a unique association with cognitive functioning at both 70 and 80 years. Several previous studies suggested that a high frequency of going outdoors was directly linked to the levels of physical and cognitive functioning among community-dwelling older Japanese; therefore, high activity levels, such as going outdoors, may actually help preserve cognitive functioning in later life.18, 19 Our study may provide support for the suggestion that a high frequency of going outdoors exerts a protective effect against cognitive functioning decline by way of increased physical activity and social engagement.18, 19 Furthermore, serum albumin was significantly correlated with cognitive functioning at age 80 years. In previous case–control studies, serum albumin was reported to be decreased in patients with Alzheimer's disease,31 and several cross-sectional studies showed that higher serum albumin was associated with higher cognitive performance.32, 33 Our study may indicate that a sufficient nutritional status and level of activity, such as the frequency of going outdoors, may prevent cognitive decline in both age groups.

The strength of our SONIC study was that the MoCA-J score, exhibiting a comparatively normal distribution, was sensitive compared with the mini-mental state examination (Supplementary Figure S2).25 Only a few previous studies used the MoCA score, so we provided additional outcomes to examine the association between BP and cognitive functioning.2

On the other hand, there are several limitations of this study. First, most of our study participants voluntarily participated in the SONIC study. Thus, it is also probable that the study participants are relatively healthy and show lower rates of mild cognitive impairment and dementia than the normal population, including nonparticipants. Therefore, there is potential inconsistency between these results and previous studies at age 80 years, in that previous studies did not generally include this age group.8, 34 Study subjects at age 80 years did not demonstrate a positive association between SBP and cognitive decline in the present study. We speculate that the reason for this inconsistency with previous studies is mainly due to the age ranges of the study subjects. HYVET-COG (Hypertension in the Very Elderly Trial cognitive function assessmen)29 indicates the negative association between high BP and cognitive function in study subjects over age 80 years. The result of HYVET-COG is consistent with our study results, although HYVET-COG is an randomized controlled trial design. However, the mean SBP, mean DBP and prevalence of hypertension observed in our SONIC study were similar compared with both Japanese national survey data from the Ministry of Health, Labor and Welfare, and with data from previous studies.5, 10, 35, 36 In fact, similar regression coefficients were observed compared with the previous studies.28 Second, because of the study’s cross-sectional design, our findings do not clarify whether high BP directly reduces cognitive functioning. Although we confirmed the absence of a correlation with cognitive functioning in the case of controlled BP, prospective findings are essential to clarify this issue. Third, the differences in the MoCA-J did not exceed the clinical cutoff point. Although this difference would not emerge in a clinical setting, the significance of the difference is noteworthy and contribute to important evidence of the association between SBP and cognitive functioning at age 70 years. Our study did not clarify the period of hypertension, considering midlife BP in relation to late-life cognitive decline, as assessed prospectively in the previous studies. Consequently, our study helps to clarify the association between late-life onset of hypertension and cognitive decline over the age of 70 years. Fourth, we could not clarify the influence of medication adherence in the present study, although SBP was strongly associated with cognitive functioning in participants treated with antihypertensive medication. Poor medication adherence caused by impaired cognitive functioning may affect these results. We have now begun examining medication adherence and will therefore be able to adjust for confounders, including medication adherence, in the follow-up analysis. Finally, we used self-reported questionnaires about the frequency of going outdoors as the indicator of activity. This may be a crude measure compared with energy expenditure; therefore, we need to use objective measures, such as digital devices that automatically measure activity level, walking speed and periods of activity.

In conclusion, our study clarified the difference in the influence of aging on the association between cognitive functioning and BP in the elderly. Because uncontrolled, high BP at 70 years is significantly correlated with cognitive decline, strict BP management from midlife according to high BP guidelines may protect against cognitive decline. Furthermore, an adequate nutritional status and daily activities, such as the frequency of going outdoors, are very important for preventing cognitive decline in old age, especially over the age of 80 years.