Abstract
This cross‐sectional study aimed to evaluate the associations of characteristics of hypertension, including hypertension status, duration, blood pressure (BP), and pulse pressure (PP), with two cognitive functions—episodic memory and executive function, in people aged over 45 years. Using 2013 survey of the China Health and Retirement Longitudinal Study (CHARLS) and weighted multiple linear regression, data from 6,732 participants were utilized. After fully adjusted in full sample, a significantly (P < 0.05) negative association was found between treated but uncontrolled hypertension and cognition. In people aged 45‐59 years, there was no significant association between hypertension and cognition. However, in people aged ≥60 years, the systolic blood pressure (SBP) and PP showed significantly adverse correlations to cognition. The negative association of untreated, treated but uncontrolled hypertension, and elevated PP with cognition increased with aging. In conclusion, this study shows the correlation between hypertension and cognition was age‐dependent with greater correlation in older people; uncontrolled hypertension and PP may be used as predictors of the cognitive decline in people ≥75 years.
Keywords: cognition, diastolic blood pressure, hypertension, pulse pressure, systolic blood pressure
1. INTRODUCTION
Hypertension, as the leading preventable cause of premature death (death before age 70 years),1 is highly prevalent all over the world. One‐third of the adults and two‐thirds of adults over 65 years old are hypertensive.2 According to a recent national cohort study conducted in China, the incidence of hypertension is now over 40% in Chinese adults.3 Hypertension has already been well recognized as a risk factor for cardiovascular and cerebrovascular diseases.4 Recently, there is evidence showing that it may also play a role in cognitive dysfunction, increasing risks of related diseases such as Alzheimer's disease (AD) and vascular dementia (VaD).5 The global prevalence of dementia had reached 24 million by 2005 and was predicted to quadruple by the year 2050.6 About 3%‐5% of Chinese population was reported to suffer from dementia, the incidence of which is different in various regions but has brought heavy economic burden to the whole country.7
Hypertension is a modifiable condition, especially in the early stage, and therefore, it has been hypothesized that antihypertensive treatments might help to prevent early cognitive decline, thus reducing the risk of further neurodegenerative diseases. However, the results on the role that hypertension plays in cognitive decline have been quite inconsistent so far. Several clinical trials failed to draw a conclusion on the effects of controlling blood pressure (BP) on inhibiting cognitive decline due to short study period and low power to detect treatment effects.8, 9 Numerous epidemiologic studies showed that hypertension in midlife increased the risk of cognitive damage that occurred 20‐30 years later,10, 11, 12 but the outcomes in regards to late‐life hypertension were inconsistent.13, 14, 15 As both AD and VaD have long preclinical phases which are present as mild cognitive decline, a better understanding of the effects of hypertension on cognition in different life stages, especially during and after middle age, is important for preventing both diseases.
The form of hypertension is distinct in middle‐aged people (aged 45‐59) as compared to in the elderly (aged ≥60). According to Smulyan et al,16 midlife hypertension is systolic/diastolic hypertension (elevation in both SBP and DBP caused by a raised total peripheral resistance), whereas most of the aged hypertensives present systolic hypertension only (a steady rise in SBP with normal or low DBP due to aortic stiffening). In addition, people over 75 years old tend to have a higher incidence of dementia.17 Based on above, participants in the current study were divided into three groups: middle‐aged people (aged 45‐59), the young elderly (aged 60‐74), and the old elderly (aged ≥75). We aimed to investigate the relationship between hypertension and cognition by comprehensively evaluating the age‐dependent effects of hypertension status, duration, BP, and pulse pressure (PP) on two cognitive measures—episodic memory and executive function.
2. MATERIALS AND METHODS
2.1. Data source and samples
This study utilized data from Chinese Health and Retirement Longitudinal Survey (CHARLS), the participants of which were ≥45 years old and from various regions in China. Administered by the National School of Development at Peking University, CHARLS was initiated in 2011 and was designed to obtain social, economic, and health status of community residents. The health status was evaluated through both self‐reports and physical examinations. The data are available online at https://charls.pku.edu.cn/zh-CN. Participants were selected from both urban and rural areas in 28 provinces using multistage probability sampling. The interviewers were trained by CHARLS staff members, and they used computer‐assisted personal interviewing (CAPI) to conduct a face‐to‐face interview with participants. Follow‐up has been conducted every 2 years, with some questions added each time.
For this study, we used the 2013 follow‐up data from CHARLS with 18, 605 participants.18 There were 13 modules of questions in the survey. To be included in the analysis, participants must have completed the modules of demographic background, health status, and functioning and biomarker (n = 13, 169). Exclusion criteria were having memory‐related diseases including VaD, AD, or Parkinson's disease; severe chronic diseases such as stroke, malignant tumor, and kidney diseases; or emotional, nervous, or psychiatric problems. Any related missing value or outlier more than three standard deviation (SD) from the mean was excluded as well. After applying the exclusion criteria, 6, 732 participants were qualified for the analysis.
2.2. Cognition
Cognition of the participants was the explained variable of this study. Based on the American Health and Retirement Study, CHARLS designed a questionnaire evaluating two cognition measures. The first one is episodic memory through immediate and delayed recall. Ten unrelated Chinese words were read to each participant, and the memory ability was evaluated by counting how many words could be recalled immediately (immediate word recall) and 4 minutes later (delayed word recall). Episodic memory was calculated as the mean of immediate and delayed recall scores based on method used by Lei et al.19 The second measure is executive function, which were evaluated through items in the Telephone Interview of Cognitive Status (TICS) as detailed below as well as figure drawing. TICS is a reliable and valid method as Mini‐Mental State Examination (MMSE) used to screen cognitively impaired elderly.20 Items in TICS include naming the day of the week, today's date (day, month, and year), the current season, and successively subtracting 7 from 100 for five times. Figure drawing lets participants repaint a picture presented to him/her. All these measures were combined to obtain the evaluation of a participant's executive function.21 The total score for episodic memory and for executive function is 10 and 11, respectively. Cronbach's coefficient alpha was calculated to evaluate the reliability of each scale testing immediate word recall, delayed word recall, or executive function.
2.3. Socio‐demographic characteristics
The socio‐demographic information of participants including age, sex, type of residency, education level, and marital status was collected in CHARLS. For our basic analysis, age was divided into 3 groups: 45‐59, 60‐74, and 75 years or older. Type of residency was classified into rural, nonrural, unified residency, and do not have legal residency. Education level was categorized into elementary school or below, middle school, and high school or above. Marital status was grouped into married, cohabitating and unmarried, divorced, separated, widowed, and never married.
2.4. Hypertension
The following aspects in regards to blood pressure/hypertension were analyzed as independent variables: hypertension status, hypertension duration, systolic blood pressure (SBP), diastolic blood pressure (DBP), and PP. The interviewers of CHARLS 2013 follow‐up went to each participant's home and measured the SBP and DBP on the left arm approximately 1‐2 cm above the elbow three times (approximately 45 seconds apart) using an electronic monitor (Omron model HEM‐7112). The mean of the three readings was calculated as the BP value of each participant. PP was calculated as follows: PP = SBP − DBP. Hypertension was defined as a SBP ≥140 mm Hg and/or a DBP ≥90 mm Hg and/or current use of antihypertensive medication.22 The 2013 follow‐up also asked three questions about hypertension including: “When was hypertension first diagnosed or noticed by yourself?”, “Do you take any treatment to control your hypertension?” and “Is your blood pressure generally under control?” According to the answers and the results of BP measurements, hypertension duration was calculated, and all participants (aged ≥45) were classified into four groups of hypertension status: normotensives (70.17%: SBP <140 mm Hg and DBP <90 mm Hg, without treatment of hypertension); controlled hypertensives (10.32%: treated for hypertension, SBP <140 mm Hg and DBP <90 mm Hg); untreated hypertensives (5.10%: SBP ≥140 mm Hg or DBP ≥90, no treatment of hypertension); and treated but uncontrolled hypertensives (14.41%: SBP ≥140 mm Hg or DBP ≥90 although receiving treatment). Hypertension duration, SBP, DBP, and PP levels were included in analysis in continuous form.
2.5. Other risk factors
In addition to hypertension and socio‐demographic characteristics, the following risk factors were also included in analysis, and participants were divided into different groups based on their self‐reports: (a) diabetes or high blood sugar—participants were classified into no diabetes, controlled hyperglycemia, untreated hyperglycemia, and treated but uncontrolled hyperglycemia; (b) dyslipidemia (elevation of low‐density lipoprotein, triglycerides (TGs) and total cholesterol, or a low high‐density lipoprotein level)—participants were divided into having or not having dyslipidemia; (c) heart problems (heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems)—participants were divided into having or not having heart problems; (d) smoking status—in CHARLS, smoking was defined as having smoked more than 100 cigarettes in life, and participants were classified into current smokers, nonsmokers, and ex‐smokers; and (e) drinking status—in CHARLS, drinking was defined as ever drank alcoholic beverage in the year before the time of interview, and participants were classified into current drinkers, nondrinkers, and ex‐drinkers.
2.6. Statistical analysis
Descriptive statistics (means and standard deviations, SD, for continuous data and percentages for categorical data) were used to report basic characteristics. One‐way analysis of variance (ANOVA) or Kruskal‐Wallis test for continuous variables and chi‐square test for categorical variables were used for analyzing the basic characteristics. We used 2013 cross‐sectional weights with nonresponse adjustment to account for multistage sampling and nonresponse. Weighted multiple linear regression was used to explore relationships between hypertension characteristics and the two measures of cognition. First, baseline model was conducted only including socio‐demographic characteristics and “other risk factors” as mentioned above to filter covariates. Second, unadjusted analysis was performed to examine the relationship between hypertension and cognition. Finally, adjusted analysis using significant covariates selected from baseline model (see the next paragraph) was conducted to examine the effects of hypertension status, hypertension duration, SBP, DBP, and PP on episodic memory and executive function, respectively. The above analysis was first conducted in the full sample and then in people aged 45‐59, 60‐74, and ≥75 years, respectively, to explore whether the relation is age‐dependent. All statistical analyses were conducted with Stata 15.1. All reported probabilities (P) were two‐sided, and P < 0.05 was considered as statistically significant.
Since compositions of the full sample and subgroups were different and the two measures of cognition had different relationships with various covariates, covariates selected from baseline models for analyses in different models were not the same. For full sample, age, residency type, educational level, marital status, and drinking status were adjusted for episodic memory; age, sex, residency type, educational level, marital status, diabetes status, dyslipidemia, smoking status, and drinking status were adjusted for executive function. For people aged 45‐59, sex, residency type, educational level, and drinking status were adjusted for episodic memory; sex, residency type, educational level, smoking status, and drinking status were adjusted for executive function. For people aged 60‐74, residency type, educational level, and marital status were adjusted for episodic memory; sex, residency type, educational level, smoking status, and drinking status were adjusted for executive function. For people aged ≥75 years, residency type and educational level were adjusted for episodic memory; residency type, educational level, and diabetes status were adjusted for executive function.
3. RESULTS
3.1. Basic characteristics
Basic characteristics of the participants including socio‐demographic and health information are summarized in Table 1. A number of 6, 732 participants were included in the analysis. The mean age of the overall population was 59 ± 9 years. More than half (65.98%) were females. Most (80.29%) were from rural area, and 69.13% received low levels of education (elementary or below). Of all the participants (aged ≥45), 29.83% were hypertensives and they were significantly (P < 0.001) older than normotensives (70.17%) and were significantly more likely to have dyslipidemia (P < 0.001) and heart problems (P < 0.001). Among hypertensives, 34.61% had BP under control. Their SBP and DBP were both significantly (P < 0.05) lower than those of untreated and treated but uncontrolled hypertensives. Untreated hypertensives had the highest number of current smokers and current drinkers. The mean cognitive scores were 3.4 ± 1.9 for episodic memory and 6.8 ± 3.4 for executive function, with statistically significant differences among people with different hypertension status, which will be reported later.
Table 1.
Basic characteristics of participants (N = 6,732)
| Characteristics | Status of hypertension | ||||
|---|---|---|---|---|---|
| Normotensives | Controlled hypertensives | Untreated hypertensives | Treated but uncontrolled hypertensives | P value | |
| Socio‐demographic variables | |||||
| No. participants | 4724 | 695 | 343 | 970 | |
| Age, mean(SD) | 58 (9) | 61 (9) | 61 (9) | 63 (9) | <0.001[Link] |
| Women, % | 65.69 | 67.77 | 58.31 | 68.87 | 0.002[Link] |
| Rural, % | 81.88 | 75.11 | 80.76 | 76.08 | <0.001[Link] |
| Educational level, % | |||||
| Elementary school or below | 67.82 | 71.51 | 69.68 | 73.61 | <0.001[Link] |
| Middle school | 20.98 | 16.26 | 19.53 | 16.39 | |
| High school or above | 11.20 | 12.23 | 10.79 | 10.00 | |
| Health status | |||||
| SBP, mean(SD), mm Hg | 118 (12) | 126 (10) | 155 (19) | 156 (19) | <0.001[Link] |
| DBP, mean(SD), mm Hg | 71 (9) | 74 (8) | 89 (13) | 88 (14) | <0.001[Link] |
| Diabetes, % | 2.69 | 12.09 | 9.04 | 11.13 | <0.001[Link] |
| Dyslipidemia, % | 4.76 | 21.65 | 10.26 | 15.47 | <0.001[Link] |
| Heart problems, % | 6.84 | 21.13 | 12.87 | 22.77 | <0.001[Link] |
| Smoking status, % | |||||
| Nonsmokers | 78.37 | 81.87 | 71.72 | 79.90 | <0.001[Link] |
| Ex‐smokers | 4.57 | 6.04 | 6.12 | 6.80 | |
| Current smokers | 17.06 | 12.09 | 22.16 | 13.30 | |
| Drinking Status, % | |||||
| Nondrinkers | 62.74 | 64.03 | 58.31 | 65.36 | <0.001[Link] |
| Ex‐drinkers | 7.32 | 11.08 | 8.16 | 12.78 | |
| Current drinkers | 29.93 | 24.89 | 33.53 | 21.86 | |
| Cognition, mean (SD) | |||||
| Episodic memory | 3.5 (1.9) | 3.3 (1.9) | 3.3 (1.9) | 3.0 (1.9) | <0.001[Link] |
| Executive function | 6.9 (3.4) | 6.7 (3.5) | 6.9 (3.4) | 6.2 (3.6) | <0.001[Link] |
DBP, diastolic blood pressure; SBP, systolic blood pressure; SD, standard deviation.
Unweighted summarizations and comparisons on participants’ basic characteristics.
P values were derived from one‐way ANOVA.
P values were derived from chi‐square test.
P values were derived from Kruskal‐Wallis test.
3.2. Reliability of cognition scales
Cronbach's coefficient alpha of scales testing immediate word recall, delayed word recall, and executive function were 0.650, 0.690, and 0.858, respectively. The total Cronbach's coefficient alpha was 0.810, indicating an excellent internal consistency reliability of cognition scales used in this study.
3.3. Associations between variables and cognition
As shown in Table 2, people who had lower scores of episodic memory and executive function tended to be the elderly in rural area with lower educational level and unfavorable marital status. Being female, having treated but uncontrolled hyperglycemia, having dyslipidemia, and being current smoker were negatively associated with executive function (P < 0.05), while current drinking was a potential protective factor in both measures of cognition. We used these significant variables as covariates in the weighted full sample multivariate linear model for further analysis of the correlation between hypertension and cognition.
Table 2.
Associations between covariates and cognition in full sample (n = 6,732)
| Characteristics |
Episodic memory β (P) |
Executive function β (P) |
|---|---|---|
| Socio‐demographic variables | ||
| Age | ||
| 45‐59 (Reference) | ||
| 60‐74 | −0.576 (<0.001a) | −0.528 (<0.001a) |
| ≥75 | −1.713 (<0.001a) | −2.328 (<0.001a) |
| Sex (“male” as reference) | 0.138 (0.104) | −0.938 (<0.001a) |
| Residency type (“rural” as reference) | 0.531 (<0.001a) | 1.197 (<0.001a) |
| Educational level | ||
| Elementary school or below (reference) | ||
| Middle school | 0.986 (<0.001a) | 1.945 (<0.001a) |
| High school or above | 1.377 (<0.001a) | 2.461 (<0.001a) |
| Marital status (“married” as reference) | −0.092 (<0.001a) | −0.106 (0.009a) |
| Health status | ||
| Diabetes status | ||
| No diabetes (reference) | ||
| Controlled hyperglycemia | −0.307 (0.242) | −0.308 (0.475) |
| Untreated hyperglycemia | 0.002 (0.990) | 0.134 (0.615) |
| Treated but uncontrolled hyperglycemia | 0.054 (0.746) | −0.509 (0.038a) |
| Dyslipidemia (“no” as reference) | −0.018 (0.832) | −0.304 (0.026a) |
| Heart problems (“no” as reference) | −0.087 (0.308) | −0.149 (0.278) |
| Smoking status | ||
| Nonsmokers (reference) | ||
| Ex‐smokers | −0.010 (0.9309) | −0.034 (0.865) |
| Current smokers | −0.064 (0.470) | −0.542 (<0.001a) |
| Drinking Status | ||
| Nondrinkers (reference) | ||
| Ex‐drinkers | 0.071 (0.474) | −0.084 (0.587) |
| Current drinkers | 0.234 (0.001a) | 0.340 (0.005a) |
Weighted multiple linear models for episodic memory and executive function including all covariates.
P is statistically significant (P < 0.05).
3.4. The association between hypertension and cognition in the full sample
As shown in Table 3, the deterioration of hypertension status, elevated SBP, DBP, and PP (per 10 mm Hg) were significant (P < 0.05) potential factors in the decline of episodic memory and executive function. After adjusting with covariates as mentioned before, treated but uncontrolled hypertension remained as a significant potential risk factor for the decline of episodic memory (β = −0.207; P = 0.017) and for that of executive function (β = −0.278; P = 0.037).
Table 3.
Weighted β (P)s of episodic memory and executive function by explanatory variables in full sample (n = 6,732)
| Explanatory variables | Episodic memory | Executive function | ||
|---|---|---|---|---|
| Unadjusted model β (P) | Adjusted model[Link] β (P) | Unadjusted model β (P) | Adjusted model[Link] β (P) | |
| Status of hypertension | ||||
| Normotensives (Reference) | ||||
| Controlled hypertensives | −0.265 (0.009a) | −0.104 (0.212) | 0.060 (0.729) | 0.065 (0.700) |
| Untreated hypertensives | −0.418 (0.025a) | −0.136 (0.224) | −0.325 (0.292) | −0.096 (0.681) |
| Treated but uncontrolled hypertensives | −0.458 (0.001a) | −0.207 (0.017a) | −0.559 (0.006a) | −0.278 (0.037a) |
| Duration of hypertension, 5 years | −0.130 (0.007a) | −0.011 (0.763) | 0.036 (0.635) | 0.141 (0.068) |
| SBP, 10 mm Hg | −0.093 (<0.001a) | −0.020 (0.280) | −0.126 (<0.001a) | −0.046 (0.117) |
| DBP, 10 mm Hg | 0.066 (0.029a) | −0.004 (0.859) | 0.119 (0.014a) | −0.022 (0.591) |
| PP, 10 mm Hg | −0.193 (<0.001a) | −0.031 (0.144) | −0.277 (<0.001a) | −0.067 (0.063) |
DBP, diastolic blood pressure; SBP, systolic blood pressure.
Weighted multiple linear model adjusted for covariates: age, residency type, educational level, marital status, and drinking status.
Weighted multiple linear model adjusted for covariates: age, sex, residency type, educational level, marital status, diabetes status, dyslipidemia, smoking status, and drinking status.
P is statistically significant (P < 0.05).
Age dependency of the correlation between hypertension and cognition.
The population was divided into three subgroups: people aged 45‐59, 60‐74, and ≥75 years. The correlation between hypertension and cognition was analyzed in each age‐group. Table 4 shows that, in people aged 45‐59 years (n = 3, 755), after fully adjusted (sex, type of residency, educational level, and drinking status for episodic memory; sex, residency type, educational level, drinking status, and smoking status for executive function), there was no significant association between any hypertension characteristic and dimension of cognition.
Table 4.
Weighted β (P)s of episodic memory and executive function by explanatory variables in people aged 45‐59 years (n = 3,755)
| Explanatory variables | Episodic memory | Executive function | ||
|---|---|---|---|---|
| Unadjusted model β (P) | Adjusted model[Link] β (P) | Unadjusted model β (P) | Adjusted model[Link] β (P) | |
| Status of hypertension | ||||
| Normotensives(Reference) | ||||
| Controlled hypertensives | −0.074 (0.584) | −0.107 (0.378) | 0.052 (0.819) | −0.007 (0.970) |
| Untreated hypertensives | 0.237 (0.180) | 0.176 (0.271) | 0.510 (0.064) | 0.329 (0.225) |
| Treated but uncontrolled hypertensives | 0.056 (0.816) | −0.055 (0.724) | 0.239 (0.477) | 0.032 (0.881) |
| Duration of hypertension, 5 years | 0.141 (0.043a) | 0.089 (0.149) | 0.253 (0.017a) | 0.164 (0.086) |
| SBP, 10 mm Hg | 0.014 (0.706) | 0.017 (0.624) | 0.026 (0.639) | 0.012 (0.809) |
| DBP, 10 mm Hg | 0.084 (0.049a) | 0.045 (0.232) | 0.171 (0.006a) | 0.060 (0.268) |
| PP, 10 mm Hg | −0.043 (0.339) | −0.007 (0.872) | −0.093 (0.188) | −0.026 (0.692) |
DBP, diastolic blood pressure; SBP, systolic blood pressure.
Weighted multiple linear model adjusted for covariates: sex, residency type, educational level, and drinking status.
Weighted multiple linear model adjusted for covariates: sex, residency type, educational level, smoking status, and drinking status.
P is statistically significant (P < 0.05).
For people aged 60‐74 years (n = 2, 505), however, Table 5 indicates that, in adjusted models (type of residency, educational level, and marital status for episodic memory; sex, type of residency, educational level, smoking status, and drinking status for executive function), the significantly negative associations of treated but uncontrolled hypertensives with episodic memory (β = −0.257; P = 0.016) and executive function (β = −0.390; P = 0.029) were found. An increase of 10 mm Hg in SBP was associated with a decreased score of 0.044 in episodic memory (β = −0.044; P = 0.017) and 0.074 in executive function (β = −0.074; P = 0.030). PP (per 10 mm Hg) was negatively associated with executive function (β = −0.097; P = 0.032). DBP was not associated with either measure of cognition.
Table 5.
Weighted β (P)s of episodic memory and executive function by explanatory variables in people aged 60‐74 years (n = 2,505)
| Explanatory variables | Episodic memory | Executive function | ||
|---|---|---|---|---|
| Unadjusted model β (P) | Adjusted model[Link] β (P) | Unadjusted model β (P) | Adjusted model[Link] β (P) | |
| Status of hypertension | ||||
| Normotensives (Reference) | ||||
| Controlled hypertensives | −0.066 (0.576) | −0.137 (0.201) | 0.212 (0.367) | 0.072 (0.713) |
| Untreated hypertensives | −0.239 (0.117) | −0.244 (0.105) | 0.146 (0.650) | 0.045 (0.867) |
| Treated but uncontrolled hypertensives | −0.245 (0.033a) | −0.257 (0.016a) | −0.379 (0.070) | −0.390 (0.029a) |
| Duration of hypertension, 5 years | 0.028 (0.590) | −0.050 (0.302) | 0.226 (0.017a) | 0.065 (0.437) |
| SBP, 10 mm Hg | −0.044 (0.029a) | −0.044 (0.017a) | −0.058 (0.136) | −0.074 (0.030a) |
| DBP, 10 mm Hg | −0.044 (0.191) | −0.048 (0.125) | −0.015 (0.823) | −0.053 (0.365) |
| PP, 10 mm Hg | −0.050 (0.070) | −0.047 (0.062) | −0.091 (0.071) | −0.097 (0.032a) |
DBP, diastolic blood pressure; SBP, systolic blood pressure.
Weighted multiple linear model adjusted for covariates: residency type, educational level, and marital status.
Weighted multiple linear model adjusted for covariates: sex, residency type, educational level, smoking status, and drinking status.
P is statistically significant (P < 0.05).
The results in people ≥75 years (n = 472) were similar to those in people aged 60‐74 years generally, but the influence of SBP attenuated while PP became more influential. Table 6 shows that, in adjusted models (type of residency and educational level for episodic memory; type of residency, educational level, and diabetes status for executive function), the significantly negative relation of untreated hypertensives (β = −0.462; P = 0.040) and treated but uncontrolled hypertensives (β = −0.582; P = 0.003) to episodic memory was found. PP (per 10 mm Hg) was negatively associated with score of episodic memory (β = −0.095; P = 0.044). But there were no significant associations between SBP or DBP and episodic memory. In regard to executive function, untreated hypertensives (β = −1.400; P = 0.022) and treated but uncontrolled hypertensives (β = −1.038; P = 0.008) were significantly associated with on it. Executive function also significantly decreased with elevated SBP (per 10 mm Hg) (β = −0.213 P = 0.007) and PP (per 10 mm Hg) (β = −0.205; P = 0.037), but had no association with DBP (per 10 mm Hg). There was no significant correlation between duration of hypertension and cognition in fully adjusted models among either full sample or three subgroups.
Table 6.
Weighted β (P)s of episodic memory and executive function by explanatory variables in people ≥75 years (n = 472)
| Explanatory variables | Episodic memory | Executive function | ||
|---|---|---|---|---|
| Unadjusted model β (P) | Adjusted model[Link] β (P) | Unadjusted model β (P) | Adjusted model[Link] β (P) | |
| Status of hypertension | ||||
| Normotensives(reference) | ||||
| Controlled hypertensives | 0.048 (0.916) | −0.090 (0.832) | 0.752 (0.485) | 0.390 (0.616) |
| Untreated hypertensives | −0.321 (0.156) | −0.462 (0.040a) | −0.526 (0.260) | −1.400 (0.022a) |
| Treated but uncontrolled hypertensives | −0.635 (0.003a) | −0.582 (0.003a) | −0.868 (0.053) | −1.038 (0.008a) |
| Duration of hypertension, 5 years | −0.096 (0.459) | −0.173 (0.144) | 0.509 (0.134) | 0.220 (0.461) |
| SBP, 10 mm Hg | −0.072 (0.059) | −0.070 (0.053) | −0.181 (0.047a) | −0.213 (0.007a) |
| DBP, 10 mm Hg | −0.051 (0.507) | −0.033 (0.665) | −0.306 (0.116) | −0.297 (0.095) |
| PP, 10 mm Hg | −0.090 (0.076) | −0.095 (0.044a) | −0.151 (0.173) | −0.205 (0.037a) |
DBP, diastolic blood pressure; SBP, systolic blood pressure.
Weighted multiple linear model adjusted for residency type and educational level.
Weighted multiple linear model adjusted for residency type, educational level, and diabetes status.
P is statistically significant (P < 0.05).
4. DISCUSSION
To our knowledge, the current study is the first of its kind to examine the individual effects of hypertension status, BP, and PP on cognition in different age‐groups in a large Chinese population. In this cross‐sectional study, we found that uncontrolled hypertension may negatively impaired the cognitive function (episodic memory and executive function) in people over 45 years old. Many epidemiologic studies have confirmed that high BP in 4th and 5th decades of life, especially uncontrolled hypertension, leads to higher possibility of cognition decline 20‐30 years later.23, 24 Given that the participants in the current study had a wide age distribution, we conducted further analysis to estimate effects of hypertension characteristics on cognition in three subgroups of target population—people aged 45‐59, 60‐74, and ≥75 years, respectively—and found that the cognitive function showed different correlations to hypertension status, SBP, DBP, and PP in different age‐groups.
In people aged 45‐59 years (subgroup 1; middle‐aged), there was no association between hypertension characteristics and cognition. There has been limited research exploring the relationship between hypertension and cognition in midlife, and they gave inconsistent outcomes. Vasilopoulos et al25 reported no significant differences in any cognitive function among people in 51‐60 years with different hypertension status, whereas other two studies found inverse association between BP and cognitive function.12, 26 It should be noted that the mean age of our subgroup 1 was similar to those in Vasilopoulos et al’s study, yet had a narrower range than those in the other two studies, which might explain the conflicting results. Because of the change of hypertension form from systolic/diastolic hypertension to systolic hypertension and the aortic stiffening, we hypothesized that SBP and PP may become more influential in cognition as people age.
Consistently with our hypothesis, in people ≥60 years, treated but uncontrolled hypertension, and elevated SBP and PP (per 10 mm Hg) were negatively associated with cognition. As for BP in the elderly, DBP showed no correlation with cognition. In contrast, an increase of 10 mm Hg in SBP was associated with a decreased score in both episodic memory and executive function in people aged 60‐74, but the association in episodic memory became insignificant in people ≥75 years. Hypertension's negative effect on cognition in people aged 60‐74 has been shown by lots of studies.27 However, one interesting note was that studies on people ≥75 years gave protective, U‐shaped, or not significant effect of BP on cognition.23, 28 A cross‐sectional study in Xi'an, China, also gave a similar result to ours that the relationship between BP and cognition in the elderly was age‐dependent.29
The present study showed that untreated, treated but uncontrolled hypertension, and elevated PP (per 10 mm Hg) started to be negatively associated with executive function in people aged 60‐74 and progressively became a significantly potential risk factor of both episodic memory and executive function in people ≥75 years. It suggests that uncontrolled hypertension and elevated PP may be used as predictors for dementia in the elderly, especially people ≥75 years. Tzourio C et al conducted a 4‐year cohort of elderly ≥60 years, reporting that the untreated hypertensives had lower MMSE scores compared with the normotensives.30 Obisesan TO et al conducted a study in 6, 163 participants, showing that higher PP predicted worse cognitive performance at ages over 70 years.31 Nation DA et al followed 877 participants and found that, in very old participants (≥80 years), increasing PP was observed in people with more severe AD.32 Yeung SE et al examined PP in adults ≥60 years, reporting consistently to ours that higher PP predicted worse cognitive function.33 The mechanism of the negative effect of uncontrolled hypertension and PP on cognition has been explored in some studies. Chronic uncontrolled hypertension could increase vascular resistance and vessel wall stiffening, and further increase the PP.34 Elevated PP, as the symbol of arterial stiffening, is the risk factor of white matter lesions and transient self‐limited hemorrhage, which are both related to cognitive decline.35 Therefore, uncontrolled hypertension and elevated PP may play more and more important role in cognitive dysfunction with aging of the elderly.36 However, it should be noted that although the association of SBP and PP with cognitive function in adjusted models was found to be significant in people ≥45, 60‐74, and ≥75 years (P < 0.05), the β values for the associations were quite low (<0.3). Only the β value for the adjusted correlation for executive memory in treated but uncontrolled hypertensive patients was >0.3. The best predictor of change in cognition appears to be the category of treated but uncontrolled hypertensives.
The current study has several strengths. First, a number of 6,732 participants were enrolled in the study. As far as we know, the sample size is larger than that in any cross‐sectional study of interactions between age and hypertension in relation to cognition. Second, the age dependency of hypertension's effect on cognition was studied. Many early studies 37 have shown larger effects of hypertension in young people than middle‐aged adults, but have not examined old participants. This study will contribute to the literature by providing data on appropriately large numbers of participants in different age cohorts. Another strength of this study was the comprehensive measurements of hypertension characteristics, including hypertension status, SBP, DBP, and PP. Also, both successfully and not successfully controlled hypertensive individuals were included in the analyses.
Limitations should also be noted in this study. First, a cross‐sectional design precludes testifying causal relations between hypertension and cognition. Second, the database did not differentiate participants who refused to do the cognition tests and who failed to do the tests, which might lead to the misclassification of the participants. In addition, hypertension status was obtained both via self‐reports and blood pressure calculations based on three readings collected in a single day, which might compromise our results due to the recall and detection bias. Furthermore, a large percentage of low education level (elementary or below; 69.13%) might bias the results because low education is likely to be associated with poor self‐perceived health status and self‐reported chronic morbidity. Moreover, this study involves limited tests due to the shortage of data in database that episodic memory and executive function are only two among many cognitive abilities related to hypertension. Interpreting and generalizing the results should also be with caution considering the homogeneous ethnic background of our study population. Lastly, the lack of examinations of biomarkers of hypertension or cognition or neuroimaging precluded the possibility of more robust evaluation of the correlation, which could be investigated in future studies.
In conclusion, our study shows that the correlation between hypertension and cognition was age‐dependent in Chinese population aged 45 and above. People aged 45‐59 might be able to compensate the cognitive deficits brought by hypertension. Treated but uncontrolled hypertension and elevated SBP may be risk factors for cognitive decline in people aged 60‐74 while untreated, treated but uncontrolled hypertension, and increased PP may predict cognitive degeneration in people ≥75 years. Relevant strategies and measures should be put forward in the future to control BP of the hypertensives, especially of the older hypertensives, thus reducing the risk of cognitive decline.
ACKNOWLEDGMENTS
This research used data from the China Health and Retirement Longitudinal Study (CHARLS). The authors thank the CHARLS team for providing the data.
Wei J, Yin X, Liu Q, Tan L, Jia C. Association between hypertension and cognitive function: A cross‐sectional study in people over 45 years old in China. J Clin Hypertens. 2018;20:1575–1583. 10.1111/jch.13393
Funding information
The authors report no specific funding in relation to this research and no conflict of interests to disclose.
REFERENCES
- 1. Muller DC, Murphy N, Johansson M, et al. Modifiable causes of premature death in middle‐age in Western Europe: results from the EPIC cohort study. BMC Med. 2016;14:87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Mills KT, Bundy JD, Kelly TN, et al. Global disparities of hypertension prevalence and control: a systematic analysis of population‐based studies from 90 countries. Circulation. 2016;134(6):441‐450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Li W, Gu H, Teo KK, et al. Hypertension prevalence, awareness, treatment, and control in 115 rural and urban communities involving 47 000 people from China. J Hypertens. 2016;34(1):39‐46. [DOI] [PubMed] [Google Scholar]
- 4. Kintscher U. The burden of hypertension. EuroIntervention: journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of. Cardiology. 2013;9(R):R12–R15. [DOI] [PubMed] [Google Scholar]
- 5. Akinyemi RO, Mukaetova‐Ladinska EB, Attems J, Ihara M, Kalaria RN. Vascular risk factors and neurodegeneration in ageing related dementias: Alzheimer's disease and vascular dementia. Curr Alzheimer Res. 2013;10(6):642‐653. [DOI] [PubMed] [Google Scholar]
- 6. Reitz C, Mayeux R. Alzheimer disease: epidemiology, diagnostic criteria, risk factors and biomarkers. Biochem Pharmacol. 2014;88(4):640‐651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Liu J, Wang LN, Tan JP. Dementia in China: current status. Neurology. 2013;81(12):1077‐1078. [DOI] [PubMed] [Google Scholar]
- 8. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension. Final results of the Systolic Hypertension in the Elderly Program (SHEP). SHEP Cooperative Research Group. JAMA. 1991;265(24):3255‐3264. [PubMed] [Google Scholar]
- 9. Anderson C, Teo K, Gao P, et al. Renin‐angiotensin system blockade and cognitive function in patients at high risk of cardiovascular disease: analysis of data from the ONTARGET and TRANSCEND studies. Lancet Neurol. 2011;10(1):43‐53. [DOI] [PubMed] [Google Scholar]
- 10. Elliott WJ. Association of higher diastolic blood pressure levels with cognitive impairment. Yearbook of Cardiology. 2010;2010:29‐30. [Google Scholar]
- 11. Pavlik VN, Hyman DJ, Doody R. Cardiovascular risk factors and cognitive function in adults 30–59 years of age (NHANES III). Neuroepidemiology. 2005;24(1–2):42‐50. [DOI] [PubMed] [Google Scholar]
- 12. Singh‐Manoux A, Marmot M. High blood pressure was associated with cognitive function in middle‐age in the Whitehall II study. J Clin Epidemiol. 2005;58(12):1308‐1315. [DOI] [PubMed] [Google Scholar]
- 13. Pandav R, Dodge HH, DeKosky ST, Ganguli M. Blood pressure and cognitive impairment in India and the United States: a cross‐national epidemiological study. Arch Neurol. 2003;60(8):1123‐1128. [DOI] [PubMed] [Google Scholar]
- 14. Reitz C, Tang MX, Manly J, Mayeux R, Luchsinger JA. Hypertension and the risk of mild cognitive impairment. Arch Neurol. 2007;64(12):1734‐1740. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Morris MC, Scherr PA, Hebert LE, et al. Association between blood pressure and cognitive function in a biracial community population of older persons. Neuroepidemiology. 2002;21(3):123‐130. [DOI] [PubMed] [Google Scholar]
- 16. Smulyan H, Mookherjee S, Safar ME. The two faces of hypertension: role of aortic stiffness. J Am Soc Hypertens. 2016;10(2):175‐183. [DOI] [PubMed] [Google Scholar]
- 17. Scheltens P, Blennow K, Breteler MM, et al. Alzheimer's disease. Lancet (London, England). 2016;388(10043):505‐517. [DOI] [PubMed] [Google Scholar]
- 18. Qian J, Li N, Ren X. Obesity and depressive symptoms among Chinese people aged 45 and over. Sci Rep. 2017;7:45637. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Lei X, Hu Y, McArdle JJ, Smith JP, Zhao Y. Gender differences in cognition among older adults in China. J Hum Resour. 2012;47(4):951‐971. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Seo EH, Lee DY, Kim SG, et al. Validity of the telephone interview for cognitive status (TICS) and modified TICS (TICSm) for mild cognitive impairment (MCI) and dementia screening. Int J Geriatr Psychiatry. 2011;52(1):e26–e30. [DOI] [PubMed] [Google Scholar]
- 21. Li J, Cacchione PZ, Hodgson N, et al. Afternoon napping and cognition in chinese older adults: findings from the china health and retirement longitudinal study baseline assessment. J Am Geriatr Soc. 2017;65(2):373‐380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Organization WH . Global status report on noncommunicable diseases 2014. Geneva, Switzerland: World Health Organization; 2015. [Google Scholar]
- 23. Launer LJ, Masaki K, Petrovitch H, Foley D, Havlik RJ. The association between midlife blood pressure levels and late‐life cognitive function. The Honolulu‐Asia Aging Study. JAMA. 1995;274(23):1846‐1851. [PubMed] [Google Scholar]
- 24. Kilander L, Nyman H, Boberg M, Hansson L, Lithell H. Hypertension is related to cognitive impairment : a 20‐year follow‐up of 999 men. Hypertension. 1998;31(3):780‐786. [DOI] [PubMed] [Google Scholar]
- 25. Vasilopoulos T, Kremen WS, Kim K, et al. Untreated hypertension decreases heritability of cognition in late middle age. Behav Genet. 2012;42(1):107‐120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Knecht S, Wersching H, Lohmann H, Berger K, Ringelstein EB. How much does hypertension affect cognition?: explained variance in cross‐sectional analysis of non‐demented community‐dwelling individuals in the SEARCH study. J Neurol Sci. 2009;283(1–2):149‐152. [DOI] [PubMed] [Google Scholar]
- 27. Qiu C, Winblad B, Fratiglioni L. The age‐dependent relation of blood pressure to cognitive function and dementia. Lancet Neurol. 2005;4(8):487‐499. [DOI] [PubMed] [Google Scholar]
- 28. Guo Z, Fratiglioni L, Winblad B, Viitanen M. Blood pressure and performance on the Mini‐Mental State Examination in the very old. Cross‐sectional and longitudinal data from the Kungsholmen Project. Am J Epidemiol. 1997;145(12):1106‐1113. [DOI] [PubMed] [Google Scholar]
- 29. Shang S, Li P, Deng M, Jiang Y, Chen C, Qu Q. The age‐dependent relationship between blood pressure and cognitive impairment: a cross‐sectional study in a rural area of Xi'an, China. PLoS One. 2016;11(7):e0159485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Tzourio C, Dufouil C, Ducimetiere P, Alperovitch A. Cognitive decline in individuals with high blood pressure: a longitudinal study in the elderly. EVA Study Group. Epidemiology of Vascular Aging. Neurology. 1999;53(9):1948‐1952. [DOI] [PubMed] [Google Scholar]
- 31. Obisesan TO, Obisesan OA, Martins S, et al. High blood pressure, hypertension, and high pulse pressure are associated with poorer cognitive function in persons aged 60 and older: the Third National Health and Nutrition Examination Survey. J Am Geriatr Soc. 2008;56(3):501‐509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Nation DA, Edmonds EC, Bangen KJ, et al. Pulse pressure in relation to tau‐mediated neurodegeneration, cerebral amyloidosis, and progression to dementia in very old adults. JAMA Neurol. 2015;72(5):546‐553. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Yeung SE, Loken Thornton W. “Do it‐yourself”: Home blood pressure as a predictor of traditional and everyday cognition in older adults. PLoS One. 2017;12(5):e0177424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Muller M, van der Graaf Y, Visseren FL, Mali WP, Geerlings MI. Hypertension and longitudinal changes in cerebral blood flow: the SMART‐MR study. Ann Neurol. 2012;71(6):825‐833. [DOI] [PubMed] [Google Scholar]
- 35. Mitchell GF. Effects of central arterial aging on the structure and function of the peripheral vasculature: implications for end‐organ damage. J Appl Physiol (1985). 2008;105(5):1652‐1660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Waldstein SR, Rice SC, Thayer JF, Najjar SS, Scuteri A, Zonderman AB. Pulse pressure and pulse wave velocity are related to cognitive decline in the baltimore longitudinal study of aging. Hypertension. 2008;51(1):99‐104. [DOI] [PubMed] [Google Scholar]
- 37. Waldstein SR, Manuck SB, Ryan CM, Muldoon MF. Neuropsychological correlates of hypertension: review and methodologic considerations. Psychological Bull. 1991;110(3):451‐468. [DOI] [PubMed] [Google Scholar]