Skip main navigation

Blood Pressure and Cognitive Decline Over 8 Years in Middle-Aged and Older Black and White Americans

Originally publishedhttps://doi.org/10.1161/HYPERTENSIONAHA.118.12062Hypertension. 2019;73:310–318

    Abstract

    Although the association between high blood pressure (BP), particularly in midlife, and late-life dementia is known, less is known about variations by race and sex. In a prospective national study of 22 164 blacks and whites ≥45 years without baseline cognitive impairment or stroke from the REGARDS cohort study (Reasons for Geographic and Racial Differences in Stroke), enrolled 2003 to 2007 and followed through September 2015, we measured changes in cognition associated with baseline systolic and diastolic BP (SBP and DBP), as well as pulse pressure (PP) and mean arterial pressure, and we tested whether age, race, and sex modified the effects. Outcomes were global cognition (Six-Item Screener; primary outcome), new learning (Word List Learning), verbal memory (Word List Delayed Recall), and executive function (Animal Fluency Test). Median follow-up was 8.1 years. Significantly faster declines in global cognition were associated with higher SBP, lower DBP, and higher PP with increasing age (P<0.001 for age×SBP×follow-up-time, age×DBP×follow-up-time, and age×PP×follow-up-time interaction). Declines in global cognition were not associated with mean arterial pressure after adjusting for PP. Blacks, compared with whites, had faster declines in global cognition associated with SBP (P=0.02) and mean arterial pressure (P=0.04). Men, compared with women, had faster declines in new learning associated with SBP (P=0.04). BP was not associated with decline of verbal memory and executive function, after controlling for the effect of age on cognitive trajectories. Significantly faster declines in global cognition over 8 years were associated with higher SBP, lower DBP, and higher PP with increasing age. SBP-related cognitive declines were greater in blacks and men.

    Introduction

    Cognitive impairment and dementia (CID) increases the risk of disability, burden of illness, and costs.1 CID affects about 8.6 million Americans,2 a number expected to triple by 2050 as the Baby Boomer generation ages,3 with higher prevalence in blacks compared with whites and in women compared with men.4 The identification of modifiable risk factors that reduce disparities and prevent CID are top public health priorities.5 Exciting preliminary results from the SPRINT MIND trial (Systolic Blood Pressure Intervention Trial Memory and Cognition in Decreased Hypertension) show that high blood pressure (BP) is a modifiable risk factor to reduce CID risk.6

    Although it is known that high BP, particularly in midlife, is associated with late-life CID,7 less is known about variations by race and sex. Blacks are more likely to develop high BP and to have greater severity of high BP than whites.8 Not only are blacks more likely to have worse BP control than whites,9 but they also seem more likely to have detrimental brain effects from high BP.10–13 A recent study14 suggests that blacks have greater BP-related cognitive declines, whereas, another study15 suggests that whites do.

    Sex differences in the association between BP and cognitive decline are less clear. Men are more likely than women to have high BP before age 50, but this disparity narrows and reverses at older ages. A recent study16 suggests that midlife high BP increases late-life CID risk in women, but not in men. It is uncertain whether the effect of high BP on late-life CID risk is greater in blacks than in whites or in women than in men, after accounting for the effect of age on the BP-cognition relationship.17

    We assessed the association between systolic and diastolic BP (SBP and DBP) levels, as well as levels of mean arterial pressure (MAP) and pulse pressure (PP) and cognitive trajectories, over 8 years in a large, nationally representative cohort of middle-aged and older black and white adults and tested whether race and sex modified the associations. We hypothesized that (1) higher SBP, lower DBP, and higher PP at baseline each are associated with faster cognitive decline and (2) that the effects of SBP and DBP on the slope of cognitive decline are greater in blacks compared with whites and in women compared with men.

    Methods

    Study Design, Participants, and Measurements

    Study protocol is available at http://www.regardsstudy.org. Statistical code is available through written agreement with authors from D.A. Levine. Dataset is available through a data use agreement with University of Alabama at Birmingham (email: [email protected]).The REGARDS study (Reasons for Geographic and Racial Differences in Stroke) is a prospective cohort study of 30 239 non-Hispanic black and white individuals examining regional and racial influences on stroke mortality.18 Details are described elsewhere.18

    Participants or their proxies were followed every 6 months by telephone with retrieval of medical records for reported hospitalizations. For this study, we followed participants through September 30, 2015. We required participants to have the first measurement of cognition by study design. We excluded participants with baseline cognitive impairment, defined as a Six-item Screener (SIS) score <5.19 We also excluded participants who self-reported a baseline history of stroke or with an incident stroke before the first measurement of each cognitive outcome. The institutional review boards of all participating institutions approved the study, and all participants provided written informed consent.

    Cognitive Function Assessments

    Trained REGARDS interviewers administered cognitive function tests longitudinally by telephone including (1) the SIS (primary outcome; scores range, 0–6) assessed global cognition annually beginning in 2003; and (2) a battery of 3 cognitive tests (secondary outcomes) measured biennially starting in 2006 that included the Consortium to Establish a Registry for Alzheimer’s Disease Word List Learning (new learning; scores range, 0–30), Word List Delayed Recall (verbal memory; scores range, 0–10), and Animal Fluency Test (executive function; scores range >0).20,21 These tests can be measured reliably and precisely over the telephone,22–24 are used in Vascular Cognitive Impairment Harmonization Standards,25 and have been validated in blacks and whites.19,26

    Measurement of BP

    At baseline, after a 5-minute rest, BP was measured twice in the left arm with a standard aneroid sphygmomanometer and the participant seated in a chair with 30 s between measurements.27 REGARDS uses the average of the 2 BP measurements. We centered BP at 120/80 mm Hg because this was close to the mean BP for the study cohort and is considered normal BP.28 We excluded 1 participant with an extreme value of BP (DBP ≤20; BP was 122/20) because this value is likely to be a result of incorrect measurement.

    Statistical Analysis

    Each outcome measure was treated as a continuous variable and analyzed separately. Linear mixed-effects models measured changes in cognitive function over time by BP and included random effects for intercept and slopes, as well as the covariates, in Table 1. Time was expressed as the years from the date of the first measurement of each cognitive outcome. We rescaled the outcomes by dividing the cognitive scores by the SD of the distribution of the baseline cognitive measurement of each test to facilitate interpretation. For each outcome, all available cognitive observations were used in the analysis except observations after the time of first expert-adjudicated incident stroke during follow up because we have previously shown that incident stroke alters the intercept and trajectory of cognition.29 To examine assumptions of linear mixed-effects models, for example, linearity of the studied relationships and normality of residual errors, we inspected residual plots. There was no evidence of nonlinear effects of SBP or DBP on cognitive slopes.

    Table 1. Adjusted Changes of Global Cognitive Function Associated With BP Levels and Age Over Time: REGARDS Study, 2003 to 2015

    Six-Item Screener Score (n=22 164)
    SBP and DBPMAP and PP
    CoefficientModel AModel BModel AModel B
    Estimate (95% CI)P ValueEstimate (95% CI)P ValueCoefficientEstimate (95% CI)P ValueEstimate (95% CI)P Value
    Slope per year of follow up0.26 (0.23 to 0.29)<0.0010.26 (0.23 to 0.29)<0.001Slope per year of follow up0.32 (0.30 to 0.35)<0.0010.32 (0.30 to 0.35)<0.001
    Change in slope (per year of follow-up) associated with 10 y increase in age at baseline (age×follow-up time interaction)−0.04 (−0.05 to −0.04)<0.001−0.04 (−0.05 to −0.04)<0.01Change in slope (per year of follow-up) associated with 10 y increase in age at baseline (age×follow-up time interaction)−0.05 (−0.06 to −0.05)<0.001−0.05 (−0.06 to −0.05)<0.001
    Change in slope (per year of follow-up) associated with 20 mm Hg increase in SBP (SBP×follow-up time interaction)0.12 (0.08 to 0.16)<0.0010.12 (0.09 to 0.16)<0.001Change in slope (per year of follow-up) associated with 10 mm Hg increase in MAP (MAP×follow-up time interaction)−0.01 (−0.04 to 0.01).23−0.01 (−0.04 to 0.01)0.36
    Change in slope (per year of follow-up) associated with 10 mm Hg increase in DBP (DBP×follow-up time interaction)−0.08 (−0.11 to −0.04)<0.001−0.07 (−0.11 to −0.04)<0.001Change in slope (per year of follow-up) associated with 10 mm Hg increase in PP (PP×follow-up time interaction)0.07 (0.05 to 0.09)<0.0010.07 (0.05 to 0.09)<0.001
    Effect of age at baseline (per 10 y increase) on slope (per year of follow-up) associated with 20 mm Hg increase in SBP (age×SBP×follow-up time interaction)−0.02 (−0.03 to −0.01)<0.001−0.02 (−0.03 to −0.01)<0.001Effect of age at baseline (per 10 y increase) on slope (per year of follow-up) associated with 10 mm Hg increase in MAP (age×MAP×follow-up time interaction term)0.002 (−0.002 to 0.01).260.002 (−0.002 to 0.01)0.29
    Effect of age at baseline (per 10 y increase) on slope (per year of follow-up) associated with 10 mm Hg increase in DBP (age×DBP×follow-up time interaction)0.012 (0.01 to 0.02)<0.0010.012 (0.01 to 0.02)<0.001Effect of age at baseline (per 10 y increase) on slope (per year of follow-up) associated with 10 mmm Hg increase in PP (age×PP×follow-up time interaction)−0.011 (−0.014 to −0.008)<0.001−0.011 (−0.014 to −0.008)<0.001
    Change in slope (per year of follow-up) associated with black race (black race×follow-up time interaction)NANA−0.005 (−0.01 to 0.004)0.26Change in slope (per year of follow-up) associated with black race (black race×follow-up time interaction)NANA−0.008 (−0.02 to −0.002)0.01
    Effect of black race on slope (per year of follow-up) associated with 20 mm Hg increase in SBP (black race×SBP×follow-up time interaction)NANA−0.01 (−0.02 to −0.002)0.02Effect of black race on slope (per year of follow-up) associated with 10 mm Hg increase in MAP (black race×MAP×follow-up time interaction)NANA−0.007 (−0.01 to 0.00)0.04
    Effect of black race on slope (per year of follow-up) associated with 10 mm Hg increase in DBP (black race×DBP×follow-up time interaction)NANA−0.001 (−0.01 to 0.01)0.79Effect of black race on slope (per year of follow-up) associated with 10 mm Hg increase in PP (black race×PP×follow-up time interaction)NANA−0.004 (−0.01 to 0.002)0.22
    Change in slope (per year of follow-up) associated with female sex (female sex×follow-up time interaction)NANA−0.001 (−0.009 to 0.008)0.89Change in slope (per year of follow-up) associated with female sex (female sex×follow-up time interaction)NANA0.002 (−0.004 to 0.009)0.49
    Effect of female sex on slope (per year of follow-up) associated with 20 mm Hg increase in SBP (female sex×SBP×follow-up time interaction)NANA0.006 (−0.004 to 0.02)0.23Effect of female sex on slope (per year of follow-up) associated with 10 mm Hg increase in MAP (female sex×MAP×follow-up time interaction)NANA0.001 (−0.01 to 0.01)0.86
    Effect of female sex on slope (per year of follow-up) associated with 10 mm Hg increase in DBP (female sex×DBP×follow-up time interaction)NANA−0.002 (−0.01 to 0.01)0.57Effect of female sex on slope (per year of follow-up) associated with 10 mm Hg increase in PP (female sex×PP×follow-up time interaction)NANA0.003 (−0.003 to 0.01)0.31

    Interpretative key: The SIS measures global cognition (scores range 0–6). Higher scores indicate better performance. The SIS was rescaled by dividing the SIS scores by the SD of the distribution of the baseline SIS measurement. Consequently, estimates of the effects (ie, coefficients from our models) are expressed relative to the SD of the outcome at baseline. Linear mixed-effects models for a continuous included time since baseline, and baseline values of SBP, DBP, SBP×time, DBP×time, age, race, sex, age×SBP, age×DBP, race×SBP, race×DBP, current hypertension treatment, education, region, income, cigarette smoking, waist circumference, diabetes mellitus, alcohol use, depressive symptoms, self-reported health status, body mass index, age×time, race×time, age×SBP×time, age×DBP×time, and a subject-specific random intercept and slope. SBP and DBP were centered at 120 and 80 mm Hg, respectively. MAP and PP were centered at 93 and 50 mm Hg, respectively. Model A tested the hypothesis that higher SBP and lower DBP at baseline are associated with faster cognitive decline in older adults by including age×SBP and age×DBP interaction effects on the slope of global cognition. Model B tested the hypothesis that the effects of SBP and DBP on cognitive decline are greater in blacks compared with whites and in women compared with men by including race×SBP, race×DBP, sex×SBP, and sex×DBP interaction effects on the slope of global cognition. Model A for SBP and DBP is equivalent to model A for PP and MAP. Model B for SBP and DBP is equivalent to model B for PP and MAP. BP indicates blood pressure; DBP, diastolic BP; MAP, mean arterial pressure; NA, not applicable; PP, pulse pressure; REGARDS, Reasons for Geographic and Racial Differences in Stroke; SBP, systolic BP; and SIS, Six-Item Screener.

    To test the effect of baseline BP on the slopes of cognitive trajectories (first hypothesis), model A included SBP×follow-up time and DBP×follow-up-time interaction terms, as well as SBP, DBP, follow-up time, and covariates (Table 1 and online-only Data Supplement). We examined whether age17 and hypertension treatment status30 modified the effect of BP level on the slopes of cognitive trajectories (eg, by investigating age×SBP×follow-up time and age×DBP×follow-up-time interaction terms). To test whether race and sex modified the effect of BP level on the slopes of cognitive trajectories (second hypothesis), model B added race×BP×follow-up time and sex×BP×follow-up-time interaction terms to model A. We stratified the data by age at baseline (45–64 versus ≥65 model years) and fitted model B for the primary outcome, SIS. The parameter estimates from the age-stratified analyses were similar to the parameter estimates from the analysis of all ages (data not shown), so we present the results from the total sample.

    We also performed analyses replacing SBP and DBP with MAP ([(DBP×2)+SBP]/3) and PP (SBP–DBP). We calculated participant-specific predicted values for each cognitive score over time for an exemplar subject, for example, a 75-year old residing in the Stroke Belt with typical values of all covariates slopes of cognitive decline was estimated for 20-mm Hg differences in SBP and 10-mm Hg differences in DBP.31 We also estimated the difference in predicted cognitive scores for different BPs (SBP, 160 versus 120 mm Hg; DBP, 60 versus 80 mm Hg) at different intervals (study time 0 versus year 8, the median follow-up time) for the exemplar subject.

    Statistical significance for all analyses was set as P<0.05 (2-sided). We performed all analyses using STATA version 14.2 (Stata Corporation, College Station, TX).

    Results

    The study sample included 22 164 participants. Median follow-up was 8.1 years (interquartile range, 5.0–10.1 years). Follow-up time was 7.4 years in the 13 670 white participants and 6.8 years in the 8494 black participants (absolute difference, 0.63 years; 95% CI, 0.54–0.72 years; P<0.001). Follow-up time was 7.2 years in the 12 436 female participants and 7.1 years in the 9728 male participants (absolute difference, 0.05 years; 95% CI, −0.04 to 0.13 years; P=0.3).

    Figure S1 in the online-only Data Supplement presents derivation of cohort. Table S1 presents baseline characteristics of study participants. Mean age was 64.2 (SD, 9.2) years, mean SBP was 127 (SD, 16) mm Hg, and mean DBP was 76 (SD, 10) mm Hg. Of the 22 164 participants, 849 had incident stroke (727 ischemic, 78 hemorrhagic, and 44 of undeterminable type) during follow up. The frequency of incident stroke did not differ by race (521 strokes [3.8%] in whites and 328 [3.9%] strokes in blacks [absolute difference, 0.05%; 95% CI, −0.4% to 0.5%; P=0.85]) but was greater in men compared with women (428 [4.4%] strokes in men and 421 [3.4%] strokes in women [absolute difference, 1.0%; 95% CI, 0.5%–1.5%; P<0.001]).

    Participants had a median of 7 SIS tests (interquartile range, 4–10 tests) and a median of two 3-test batteries (interquartile range, 2–4 tests). Because the secondary outcome measures were introduced during follow up and performed less frequently, the Word List Learning analysis included 12 178 participants, the Word List Delayed Recall analysis included 11 984 participants, and the Animal Fluency Test analysis included 12 902 participants.

    Change in Global Cognition Associated With BP, Race, and Sex

    Age significantly modified the effect of SBP, DBP, and PP levels, but not MAP, on the slope of global cognition trajectories (P<0.001 for age×SBP×follow-up-time interaction, P<0.001 for age×DBP×follow-up-time interaction, P<0.001 for age×PP×follow-up-time interaction, and P=0.26 for age×MAP×follow-up-time interaction; model A for SBP/DBP and PP/MAP; Table 1). With increasing age, higher SBP, lower DBP, and higher PP were each associated with faster declines in global cognition. The effect of SBP and DBP on the slope of global cognition trajectories was not significantly different by hypertension treatment status (P=0.53 for hypertension treatment×SBP×follow-up-time interaction and P=0.26 for hypertension treatment×DBP×follow-up-time interaction).

    In models replacing SBP and DBP with PP and MAP, PP (P<0.001), but not MAP (P=0.23), was significantly associated with changes in global cognition (model A for PP and MAP; Table 1). Independent of MAP, higher PP was associated with significantly faster declines in global cognition controlling for age.

    Blacks had significantly faster declines in global cognition associated with SBP compared with whites (P=0.02 for race×SBP×follow-up-time interaction), but race did not modify the effect of DBP on the slope of global cognition (P=0.79 for race×DBP×follow-up-time interaction; model B for SBP and DBP; Table 1). Black race also modified the effect of MAP on the slope of global cognition (P=0.04 for race×MAP×follow-up-time interaction) but not the effect of PP (P=0.22 for race×PP×follow-up-time interaction; model B for PP and MAP; Table 1). Blacks, compared with whites, had significantly faster declines in global cognition associated with higher MAP. The effect of BP on the slope of global cognition trajectories was not significantly different by sex (P=0.23 for sex×SBP× follow-up-time interaction, P=0.57 for sex×DBP×follow-up-time interaction, P=0.31 for sex×PP× follow-up-time interaction, and P=0.86 for sex×MAP×follow-up-time interaction; model B for SBP/DBP and PP/MAP; Table 1).

    We illustrate results of models B using slope estimates of global cognition changes by age, race, sex, and the 4 BP measures (SBP, DBP, PP, and MAP) and by calculating participant-specific predicted values of cognition (Table 2). Black women aged 75 years with common values of all covariates, compared with white women, had faster declines in global cognition associated with higher SBPs by 0.012 SD points-per-year-per 20-mm Hg increase (95% CI, 0.002–0.022; P=0.02). Declines in global cognition associated with lower DBPs were similar in black women (decrease in slope, 0.015 SD points-per-year-per-10-mm Hg decrease; 95% CI, 0.005–0.024; P=0.002) and white women (decrease in slope, 0.016 SD points-per-year-per-10-mm Hg decrease; 95% CI, 0.008–0.024; P<0.001; Table 2). Figure 1 shows slopes of global cognition by SBP, race, and sex (DBP fixed at 80 mm Hg). Figure 2 shows slopes of global cognition by DBP, race, and sex (SBP fixed at 120 mm Hg). At age 75 years, the difference in global cognition at year 8 was significantly greater than that at year 0 for baseline SBP 160 versus 120 mm Hg (difference of difference for white women, −0.12 points; 95% CI, −0.05 to −0.19; P<0.001) and nonsignificantly greater for baseline DBP 60 versus 80 mm Hg (difference of difference: −0.05 points; 95% CI, −0.11 to 0.01; P=0.08).

    Table 2. Estimated Changes in Slopes of Global Cognition Associated With Baseline Blood Pressure, Age, and Race-Sex: REGARDS Study, 2003 to 2015

    Age at baseline, yChange in Slope for Each 20 mm Hg Increase in SBP per Year (DBP Fixed at 80 mm Hg), Points per Year (95% CI)P ValueChange in Slope for Each 10 mm Hg Decrease in DBP per Year (SBP fixed at 120 mm Hg), Points per Year (95% CI)P ValueChange in Slope for Each 10 mm Hg Increase in MAP per Year (Pulse Pressure Fixed at 40 mm Hg), Points per Year (95% CI)P ValueChange in Slope for Each 15 mm Hg Increase in Pulse Pressure per Year (MAP Fixed at 93.3 mm Hg), Points per Year (95% CI)P Value
    White women
     550.019 (0.008 to 0.028)<0.0010.009 (0.0002 to 0.017)0.040.0003 (−0.007 to 0.008)0.930.013 (0.005 to 0.022)0.003
     65−0.002 (−0.01 to 0.006)0.57−0.004 (−0.01 to 0.003)0.310.002 (−0.005 to 0.01)0.51−0.003 (−0.01 to 0.004)0.43
     75−0.023 (−0.032 to −0.013)<0.001−0.016 (−0.024 to −0.008)<0.0010.004 (−0.004 to 0.013)0.30−0.019 (−0.028 to −0.011)<0.001
    White men
     550.012 (0.001 to 0.023)0.030.006 (−0.002 to 0.015)0.16−0.0003 (−0.009 to 0.008)0.940.009 (0 to 0.02)0.05
     65−0.009 (−0.017 to 0)0.05−0.006 (−0.013 to 0.001)0.090.002 (−0.005 to 0.009)0.63−0.007 (−0.015 to 0.0002)0.05
     75−0.029 (−0.038 to −0.019)<0.001−0.018 (−0.026 to −0.01)<0.0010.004 (−0.005 to 0.012)0.37−0.024 (−0.032 to −0.015)<0.001
    Black women
     550.006 (−0.004 to 0.016)0.260.01 (0.001 to 0.019)0.03−0.007 (−0.015 to 0.001)0.08−0.008 (−0.001 to 0.017)0.08
     65−0.014 (−0.023 to −0.005)0.002−0.002 (−0.01 to 0.005)0.55−0.005 (−0.013 to 0.003)0.22−0.008 (−0.016 to −0.0004)0.04
     75−0.035 (−0.045 to −0.024)<0.001−0.015 (−0.024 to −0.005)0.002−0.003 (−0.012 to 0.006)0.56−0.025 (−0.034 to −0.015)<0.001
    Black men
     55−0.002 (−0.012 to 0.012)0.980.007 (−0.002 to 0.017)0.14−0.008 (−0.016 to 0.001)0.090.004 (−0.007 to 0.014)0.49
     65−0.021 (−0.031 to −0.01)<0.001−0.005 (−0.014 to 0.004)0.29−0.005 (−0.014 to 0.003)0.20−0.013 (−0.022 to −0.004)0.007
     75−0.041 (−0.053 to −0.029)<0.001−0.017 (−0.027 to −0.007)0.001−0.003 (−0.013 to 0.006)0.50−0.029 (−0.039 to −0.019)<0.001

    These results are from model B shown in Table 1. Participant-specific estimated changes in slopes for participants with the common values of all covariates at baseline (some college education, Stroke Belt residence, income $20 000–$34 999, never smoker, no alcohol use, no diabetes mellitus, no hypertension treatment, waist circumference 95 cm, overweight BMI, 4-item CES-D score of 0.9 points, and fair health status). Global cognition was measured by SIS. SIS scores range 0–6. SBP and DBP were centered at 120 and 80 mm Hg, respectively. Similarly, pulse pressure and mean arterial pressure were centered at 40 and 93.33 mm Hg. The SIS was rescaled by dividing the SIS scores by the SD of the distribution of the initial SIS measurement. The changes in slopes (ie, coefficients from our models) are expressed relative to the SD of the outcome at baseline. BMI indicates body mass index; CES-D, Center for Epidemiologic Studies Depression Scale; DBP, diastolic BP; MAP, mean arterial pressure; REGARDS, Reasons for Geographic and Racial Differences in Stroke; SBP, systolic BP; and SIS, Six-Item Screener.

    Figure 1.

    Figure 1. Conditional predicted values of global cognition over time by baseline systolic blood pressure (BP), race, and sex: REGARDS study (Reasons for Geographic and Racial Differences in Stroke), 2003–2013. Participant-specific predicted values of global cognition were calculated for 75-y old adults with the common values of all covariates at baseline (college education, Stroke Belt residence, income $20 000–$34 999, never smoker, no alcohol use, no diabetes mellitus, no hypertension treatment, waist circumference 95 cm, overweight body mass index, 4-item CES-D score (Center for Epidemiologic Studies Depression Scale) of 0.9 points, and fair health status) and model B. Global cognition was measured by Six-Item Screener. Six-Item Screener scores range from 0 to 6 with higher scores indicating better performance. Diastolic BP is fixed at 80 mm Hg. The brown line indicates the slopes for participants with systolic BP of 110 mm Hg. The green line indicates the slopes for participants with systolic BP of 120 mm Hg. The blue line indicates the slopes for participants with systolic BP of 130 mm Hg. The black line indicates the slopes for participants with systolic BP of 140 mm Hg. The red line indicates the slopes for participants with systolic BP of 150 mm Hg.

    Figure 2.

    Figure 2. Conditional predicted values of global cognition over time by baseline diastolic blood pressure (BP), race, and sex: REGARDS study (Reasons for Geographic and Racial Differences in Stroke), 2003–2013. Participant-specific predicted values of global cognition were calculated for 75-y old adults with the common values of all covariates at baseline (college education, Stroke Belt residence, income $20 000–$34 999, never smoker, no alcohol use, no diabetes mellitus, no hypertension treatment, waist circumference 95 cm, overweight body mass index, 4-item CES-D score (Center for Epidemiologic Studies Depression Scale) of 0.9 points, and fair health status) and model B. Global cognition was measured by Six-Item Screener. Six-Item Screener scores range from 0 to 6 with higher scores indicating better performance. Systolic BP is fixed at 120 mm Hg. The brown line indicates the slopes for participants with diastolic BP of 70 mm Hg. The green line indicates the slopes for participants with diastolic BP of 80 mm Hg. The blue line indicated the slopes for participants with diastolic BP of 90 mm Hg. The black line indicates the slopes for participants with diastolic BP of 100 mm Hg. The red line indicates the slopes for participants with diastolic BP of 110 mm Hg.

    Changes in New Learning, Verbal Memory, and Executive Function Associated With BP, Race, and Sex

    Higher SBP, lower DBP, and higher PP were significantly associated with faster declines in new learning, verbal memory, and executive function in partially adjusted models that did not include the effect of age on cognitive slopes (age×follow-up-time interaction; Tables S2 and S3). However, the effects of SBP, DBP, and PP on the slopes of these cognitive functions did not remain significant after further adjustment for the age-by-time interaction effect except men, compared with women, had faster declines in new learning associated with SBP (P=0.04; Tables S4 and S5).

    Sensitivity Analyses

    Results were similar in analyses imputing missing values of baseline covariates, adjusting for death, adding history of myocardial infarction and glomerular filtration rate to models, and in analyses limited to participants with ≥2 follow-up cognitive measures (Tables S6 and S9).

    Discussion

    In this national cohort of black and white Americans 45 years or older, faster declines in global cognition were associated with higher SBP, lower DBP, and higher PP in older adults. Faster declines in global cognition were associated with higher SBP and lower DBP, as well as higher PP, but not MAP. Blacks, compared with whites, had faster declines in global cognition associated with SBP and MAP. Men, compared with women, had faster declines in new learning associated with SBP. BP was not associated with slopes of verbal memory and executive function, after controlling for the effect of age on cognitive trajectories.

    Our data suggest that BP-related cognitive declines are greater in blacks compared with whites and in men compared with women. Our results are consistent with a recent study14 of older adults showing that blacks, compared with whites, have faster cognitive declines associated with high BP. Some studies have suggested that the effect of high BP on CID risk is greater in whites compared with blacks15 and in women compared with men.16,32,33 Our results may differ because we used different cognitive measures, we studied older adults, and we controlled for the age-dependent effect of BP on cognition.17 The results of the analysis of secondary outcomes may differ from those of the SIS because the former have fewer observations, longer measurement intervals, and reduced statistical power.

    The declines in global cognition associated with high SBP and low DBP in older adults are likely clinically meaningful. Declines of ≥0.5 SD from baseline have been defined as clinically meaningful decline34 and are correlated with clinically meaningful decline in adults with normal cognition and dementia.35,36 A 0.5-SD decrease from the baseline score is ≈0.2 points for the SIS. The 95% CIs for the 8-year differences in global cognition for baseline SBP of 160 compared with that of 120 mm Hg for exemplar individuals aged 75 years or older include declines of this magnitude or approach them for white women (95% CI, 0.05–0.19).

    Declines in global cognition significantly increase the risk of death, dementia, and functional decline.19,37–39

    High or low BP may cause cognitive decline through several mechanisms. Hypertension may cause ongoing inflammation, oxidative stress, and cerebrovascular injury.40 Although we excluded cognitive measures occurring after the time of clinically apparent stroke, hypertension may cause subclinical vascular brain injury and damage to white matter integrity that contributes to subsequent cognitive decline.41,42 Hypertension may also induce or exacerbate neurodegenerative disease.43,44 Low BP may cause cognitive decline because of cerebral hypoperfusion. Blacks and men may be more likely to have detrimental brain effects from high BP because of early age of onset of high BP and worse BP control over the life course leading to greater arterial stiffness and more severe atherosclerosis.10–13 Our data suggest a scientific need to determine how the timing, duration, and intensity of BP lowering interventions over the life course, as well as race and sex, influence the risk of cognitive decline.

    Our study has several strengths. We had longitudinal cognitive assessments in a cohort of sufficient size to estimate BP-related changes in cognitive decline and to examine potential effect modification by race, sex, age, and hypertension treatment. REGARDS systematically measured cognitive domains commonly affected by vascular factors like hypertension: global cognition, learning, memory, and executive function.25 We had repeated measures over time up to 14 years of follow up.

    Our study has limitations. Results are generalizable only to community-dwelling adults not requiring a proxy respondent. Although excluded participants had higher prevalence of dementia risk factors than included participants, these differences would reduce the ability to detect the cognitive effects of BP. BP was measured only at baseline. Although selective attrition may lead to underestimation of cognitive decline because participants with worse cognition at baseline or during follow up die, drop out, or require a proxy, analyses that accounted for loss to follow-up or death did not change our results, consistent with prior research.45 Fewer cognitive observations potentially limited statistical power to detect associations between BP and the secondary outcomes. The slight increases in global cognition, new learning, and verbal memory over time may be because of selective attrition of cognitively impaired participants and practice effects associated with repeated testing.46,47 We did not have data on functional impairments, brain imaging, daily medication use, or incident dementia.

    Our study has clinical and policy implications. Preliminary results of the SPRINT MIND trial suggest that aggressive lowering of SBP reduces CID risk in older adults at high cardiovascular risk.6 Our results extend this work by suggesting that treatment of high SBP will reduce CID risk in a broader population of older adults, particularly blacks and men. Our results also suggest that avoidance of low DBP is important to prevent cognitive decline in older adults. This is clinically important because SBP tends to increase and DBP tends to decrease as adults age. Blacks are more likely to have worse BP control than whites. Our results suggest that eliminating the black-white disparity in BP control has the potential to reduce racial disparities in CID risk. Our data also show that failure to account for the effect of age on cognitive slopes can overestimate the effect of BP on cognitive slopes and even cause a spurious association between BP and cognitive decline. Moreover, our results suggest that PP rather than MAP is a potential target for preventing CID.

    Perspectives

    Faster declines in global cognition over 8 years were associated with higher SBP, lower DBP, and higher PP in older adults. SBP-related cognitive declines were greater in blacks and men. Our results suggest that BP may contribute to black-white disparities in CID risk.

    Acknowledgments

    We thank the other investigators, the staff, and the participants of the REGARDS study (Reasons for Geographic and Racial Differences in Stroke) for their valuable contributions. A full list of participating REGARDS investigators and institutions can be found at http://www.regardsstudy.org.

    Footnotes

    The online-only Data Supplement is available with this article at https://www.ahajournals.org/doi/suppl/10.1161/HYPERTENSIONAHA.118.12062.

    Correspondence to Deborah A. Levine, Division of General Medicine, University of Michigan, NCRC, 2800 Plymouth Rd, Bldg 16-430W, Ann Arbor, MI 48109. Email

    References

    • 1. Langa KM, Levine DA. The diagnosis and management of mild cognitive impairment: a clinical review.JAMA. 2014; 312:2551–2561. doi: 10.1001/jama.2014.13806CrossrefMedlineGoogle Scholar
    • 2. Plassman BL, Langa KM, Fisher GG, Heeringa SG, Weir DR, Ofstedal MB, Burke JR, Hurd MD, Potter GG, Rodgers WL, Steffens DC, McArdle JJ, Willis RJ, Wallace RB. Prevalence of cognitive impairment without dementia in the United States.Ann Intern Med. 2008; 148:427–434.CrossrefMedlineGoogle Scholar
    • 3. Hebert LE, Weuve J, Scherr PA, Evans DA. Alzheimer disease in the United States (2010-2050) estimated using the 2010 census.Neurology. 2013; 80:1778–1783. doi: 10.1212/WNL.0b013e31828726f5CrossrefMedlineGoogle Scholar
    • 4. Alzheimer's Association. 2014 Alzheimer’s disease facts and figures.Alzheimers Dement. 2014; 10:e47–e92. doi: 10.1016/j.jalz.2014.02.001.CrossrefMedlineGoogle Scholar
    • 5. National Institute of Neurological Disorders and Stroke Council. Alzheimer’s Disease-Related Dementias: Research Challenges and Opportunities; Bethesda, MD; September 13, 2013. https://wayback.archive-it.org/1170/20161005013551/https://www.ninds.nih.gov/news_and_events/proceedings/adrd2013.htm. Accessed July 18, 2018.Google Scholar
    • 6. Alzheimer’s Association. Study shows intensive blood pressure control reduces risk of mild cognitive impairment (MCI) and the combined risk of MCI and dementia.Paper presented at: Alzheimer’s Association International Conference 2018; Chicago, IL; July 25, 2018. https://www.alz.org/aaic/downloads2018/Wed-am-briefing-developing-topics.pdf. Accessed September 4, 2018.Google Scholar
    • 7. Elias MF, Goodell AL, Dore GA. Hypertension and cognitive functioning: a perspective in historical context.Hypertension. 2012; 60:260–268. doi: 10.1161/HYPERTENSIONAHA.111.186429LinkGoogle Scholar
    • 8. Levine DA, Lewis CE, Williams OD, Safford MM, Liu K, Calhoun DA, Kim Y, Jacobs DR, Kiefe CI. Geographic and demographic variability in 20-year hypertension incidence: the CARDIA study.Hypertension. 2011; 57:39–47. doi: 10.1161/HYPERTENSIONAHA.110.160341LinkGoogle Scholar
    • 9. Flack JM, Sica DA, Bakris G, et al; International Society on Hypertension in Blacks. Management of high blood pressure in Blacks: an update of the International Society on Hypertension in Blacks consensus statement.Hypertension. 2010; 56:780–800. doi: 10.1161/HYPERTENSIONAHA.110.152892LinkGoogle Scholar
    • 10. Birns J, Morris R, Jarosz J, Markus H, Kalra L. Ethnic differences in the cerebrovascular impact of hypertension.Cerebrovasc Dis. 2008; 25:408–416. doi: 10.1159/000121341CrossrefMedlineGoogle Scholar
    • 11. Brickman AM, Schupf N, Manly JJ, Luchsinger JA, Andrews H, Tang MX, Reitz C, Small SA, Mayeux R, DeCarli C, Brown TR. Brain morphology in older African Americans, Caribbean Hispanics, and whites from northern Manhattan.Arch Neurol. 2008; 65:1053–1061. doi: 10.1001/archneur.65.8.1053CrossrefMedlineGoogle Scholar
    • 12. Howard G, Lackland DT, Kleindorfer DO, Kissela BM, Moy CS, Judd SE, Safford MM, Cushman M, Glasser SP, Howard VJ. Racial differences in the impact of elevated systolic blood pressure on stroke risk.JAMA Intern Med. 2013; 173:46–51. doi: 10.1001/2013.jamainternmed.857CrossrefMedlineGoogle Scholar
    • 13. Marcus J, Gardener H, Rundek T, Elkind MS, Sacco RL, Decarli C, Wright CB. Baseline and longitudinal increases in diastolic blood pressure are associated with greater white matter hyperintensity volume: the northern Manhattan study.Stroke. 2011; 42:2639–2641. doi: 10.1161/STROKEAHA.111.617571LinkGoogle Scholar
    • 14. Hajjar I, Rosenberger KJ, Kulshreshtha A, Ayonayon HN, Yaffe K, Goldstein FC. Association of JNC-8 and SPRINT systolic blood pressure levels with cognitive function and related racial disparity.JAMA Neurol. 2017; 74:1199–1205. doi: 10.1001/jamaneurol.2017.1863CrossrefMedlineGoogle Scholar
    • 15. Gottesman RF, Schneider AL, Albert M, Alonso A, Bandeen-Roche K, Coker L, Coresh J, Knopman D, Power MC, Rawlings A, Sharrett AR, Wruck LM, Mosley TH. Midlife hypertension and 20-year cognitive change: the atherosclerosis risk in communities neurocognitive study.JAMA Neurol. 2014; 71:1218–1227. doi: 10.1001/jamaneurol.2014.1646CrossrefMedlineGoogle Scholar
    • 16. Gilsanz P, Mayeda ER, Glymour MM, Quesenberry CP, Mungas DM, DeCarli C, Dean A, Whitmer RA. Female sex, early-onset hypertension, and risk of dementia.Neurology. 2017; 89:1886–1893. doi: 10.1212/WNL.0000000000004602CrossrefMedlineGoogle Scholar
    • 17. Qiu C, Winblad B, Fratiglioni L. The age-dependent relation of blood pressure to cognitive function and dementia.Lancet Neurol. 2005; 4:487–499. doi: 10.1016/S1474-4422(05)70141-1CrossrefMedlineGoogle Scholar
    • 18. Howard VJ, Cushman M, Pulley L, Gomez CR, Go RC, Prineas RJ, Graham A, Moy CS, Howard G. The reasons for geographic and racial differences in stroke study: objectives and design.Neuroepidemiology. 2005; 25:135–143. doi: 10.1159/000086678CrossrefMedlineGoogle Scholar
    • 19. Callahan CM, Unverzagt FW, Hui SL, Perkins AJ, Hendrie HC. Six-item screener to identify cognitive impairment among potential subjects for clinical research.Med Care. 2002; 40:771–781. doi: 10.1097/01.MLR.0000024610.33213.C8CrossrefMedlineGoogle Scholar
    • 20. Morris JC, Heyman A, Mohs RC, Hughes JP, van Belle G, Fillenbaum G, Mellits ED, Clark C. The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer’s disease.Neurology. 1989; 39:1159–1165.CrossrefMedlineGoogle Scholar
    • 21. Morris JC, Mohs RC, Rogers H, Fillenbaum G, Heyman A. Consortium to establish a registry for Alzheimer’s disease (CERAD) clinical and neuropsychological assessment of Alzheimer’s disease.Psychopharmacol Bull. 1988; 24:641–652.MedlineGoogle Scholar
    • 22. Unverzagt FW, Monahan PO, Moser LR, Zhao Q, Carpenter JS, Sledge GW, Champion VL. The Indiana University telephone-based assessment of neuropsychological status: a new method for large scale neuropsychological assessment.J Int Neuropsychol Soc. 2007; 13:799–806. doi: 10.1017/S1355617707071020CrossrefMedlineGoogle Scholar
    • 23. Rapp SR, Legault C, Espeland MA, Resnick SM, Hogan PE, Coker LH, Dailey M, Shumaker SA; CAT Study Group. Validation of a cognitive assessment battery administered over the telephone.J Am Geriatr Soc. 2012; 60:1616–1623. doi: 10.1111/j.1532-5415.2012.04111.xCrossrefMedlineGoogle Scholar
    • 24. Wilson RS, Leurgans SE, Foroud TM, Sweet RA, Graff-Radford N, Mayeux R, Bennett DA; National Institute on Aging Late-Onset Alzheimer’s Disease Family Study Group. Telephone assessment of cognitive function in the late-onset Alzheimer’s disease family study.Arch Neurol. 2010; 67:855–861. doi: 10.1001/archneurol.2010.129CrossrefMedlineGoogle Scholar
    • 25. Hachinski V, Iadecola C, Petersen RC, et al. National Institute of Neurological Disorders and Stroke-Canadian Stroke Network vascular cognitive impairment harmonization standards.Stroke. 2006; 37:2220–2241. doi: 10.1161/01.STR.0000237236.88823.47LinkGoogle Scholar
    • 26. Lucas JA, Ivnik RJ, Smith GE, Ferman TJ, Willis FB, Petersen RC, Graff-Radford NR. Mayo’s older African Americans Normative Studies: norms for Boston naming test, controlled oral word association, category fluency, animal naming, token test, WRAT-3 reading, trail making test, stroop test, and judgment of line orientation.Clin Neuropsychol. 2005; 19:243–269. doi: 10.1080/13854040590945337CrossrefMedlineGoogle Scholar
    • 27. Levine DA, Calhoun DA, Prineas RJ, Cushman M, Howard VJ, Howard G. Moderate waist circumference and hypertension prevalence: the REGARDS Study.Am J Hypertens. 2011; 24:482–488. doi: 10.1038/ajh.2010.258CrossrefMedlineGoogle Scholar
    • 28. Mozaffarian D, Benjamin EJ, Go AS, Arnett DK, Blaha MJ, Cushman M, et al. Heart disease and stroke statistics-2016 update: a report from the American Heart Association.Circulation. 2016; 133:e38–e360.LinkGoogle Scholar
    • 29. Levine DA, Galecki AT, Langa KM, Unverzagt FW, Kabeto MU, Giordani B, Wadley VG. Trajectory of cognitive decline after incident stroke.JAMA. 2015; 314:41–51. doi: 10.1001/jama.2015.6968CrossrefMedlineGoogle Scholar
    • 30. Köhler S, Baars MA, Spauwen P, Schievink S, Verhey FR, van Boxtel MJ. Temporal evolution of cognitive changes in incident hypertension: prospective cohort study across the adult age span.Hypertension. 2014; 63:245–251. doi: 10.1161/HYPERTENSIONAHA.113.02096LinkGoogle Scholar
    • 31. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R; Prospective Studies Collaboration. Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies.Lancet. 2002; 360:1903–1913. doi: 10.1016/S0140-6736(02)11911-8CrossrefMedlineGoogle Scholar
    • 32. Qiu C, Winblad B, Viitanen M, Fratiglioni L. Pulse pressure and risk of Alzheimer disease in persons aged 75 years and older: a community-based, longitudinal study.Stroke. 2003; 34:594–599. doi: 10.1161/01.STR.0000060127.96986.F4LinkGoogle Scholar
    • 33. 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; 58:1308–1315. doi: 10.1016/j.jclinepi.2005.03.016CrossrefMedlineGoogle Scholar
    • 34. Wolinsky FD, Unverzagt FW, Smith DM, Jones R, Stoddard A, Tennstedt SL. The ACTIVE cognitive training trial and health-related quality of life: protection that lasts for 5 years.J Gerontol A Biol Sci Med Sci. 2006; 61:1324–1329.CrossrefMedlineGoogle Scholar
    • 35. Rossetti HC, Munro Cullum C, Hynan LS, Lacritz LH. The CERAD neuropsychologic battery total score and the progression of Alzheimer disease.Alzheimer Dis Assoc Disord. 2010; 24:138–142. doi: 10.1097/WAD.0b013e3181b76415CrossrefMedlineGoogle Scholar
    • 36. Unger JM, van Belle G, Heyman A. Cross-sectional versus longitudinal estimates of cognitive change in nondemented older people: a CERAD study. Consortium to Establish a Registry for Alzheimer’s Disease.J Am Geriatr Soc. 1999; 47:559–563.CrossrefMedlineGoogle Scholar
    • 37. Bennett HP, Corbett AJ, Gaden S, Grayson DA, Kril JJ, Broe GA. Subcortical vascular disease and functional decline: a 6-year predictor study.J Am Geriatr Soc. 2002; 50:1969–1977.CrossrefMedlineGoogle Scholar
    • 38. Clark LJ, Gatz M, Zheng L, Chen YL, McCleary C, Mack WJ. Longitudinal verbal fluency in normal aging, preclinical, and prevalent Alzheimer’s disease.Am J Alzheimers Dis Other Demen. 2009; 24:461–468. doi: 10.1177/1533317509345154CrossrefMedlineGoogle Scholar
    • 39. Cosentino S, Scarmeas N, Albert SM, Stern Y. Verbal fluency predicts mortality in Alzheimer disease.Cogn Behav Neurol. 2006; 19:123–129. doi: 10.1097/01.wnn.0000213912.87642.3dCrossrefMedlineGoogle Scholar
    • 40. Coleman CG, Wang G, Faraco G, Marques Lopes J, Waters EM, Milner TA, Iadecola C, Pickel VM. Membrane trafficking of NADPH oxidase p47(phox) in paraventricular hypothalamic neurons parallels local free radical production in angiotensin II slow-pressor hypertension.J Neurosci. 2013; 33:4308–4316. doi: 10.1523/JNEUROSCI.3061-12.2013CrossrefMedlineGoogle Scholar
    • 41. Faraco G, Iadecola C. Hypertension: a harbinger of stroke and dementia.Hypertension. 2013; 62:810–817. doi: 10.1161/HYPERTENSIONAHA.113.01063LinkGoogle Scholar
    • 42. Wang R, Fratiglioni L, Laukka EJ, Lövdén M, Kalpouzos G, Keller L, Graff C, Salami A, Bäckman L, Qiu C. Effects of vascular risk factors and APOE ε4 on white matter integrity and cognitive decline.Neurology. 2015; 84:1128–1135. doi: 10.1212/WNL.0000000000001379CrossrefMedlineGoogle Scholar
    • 43. Sparks DL, Scheff SW, Liu H, Landers TM, Coyne CM, Hunsaker JC. Increased incidence of neurofibrillary tangles (NFT) in non-demented individuals with hypertension.J Neurol Sci. 1995; 131:162–169.CrossrefMedlineGoogle Scholar
    • 44. Launer LJ, Ross GW, Petrovitch H, Masaki K, Foley D, White LR, Havlik RJ. Midlife blood pressure and dementia: the Honolulu-Asia aging study.Neurobiol Aging. 2000; 21:49–55.CrossrefMedlineGoogle Scholar
    • 45. Salthouse TA. Selectivity of attrition in longitudinal studies of cognitive functioning.J Gerontol B Psychol Sci Soc Sci. 2014; 69:567–574. doi: 10.1093/geronb/gbt046CrossrefMedlineGoogle Scholar
    • 46. Stein J, Luppa M, Luck T, et al. The assessment of changes in cognitive functioning: age-, education-, and gender-specific reliable change indices for older adults tested on the CERAD-NP battery: results of the German Study on Ageing, Cognition, and Dementia in Primary Care Patients (AgeCoDe).Am J Geriatr Psychiatry. 2012; 20:84–97. doi: 10.1097/JGP.0b013e318209dd08CrossrefMedlineGoogle Scholar
    • 47. Gross AL, Benitez A, Shih R, Bangen KJ, Glymour MM, Sachs B, Sisco S, Skinner J, Schneider BC, Manly JJ. Predictors of retest effects in a longitudinal study of cognitive aging in a diverse community-based sample.J Int Neuropsychol Soc. 2015; 21:506–518. doi: 10.1017/S1355617715000508CrossrefMedlineGoogle Scholar

    Novelty and Significance

    What Is New?

    • It is unclear whether hypertension contributes to racial/ethnic and sex differences in dementia risk.

    • We assessed the association between blood pressure (BP) levels and cognitive trajectories over 8 years and tested whether race and sex modified the associations, in a large, nationally representative cohort of middle-aged and older black and white adults.

    • Our results suggest that lowering systolic BP in older adults, particularly blacks and men, will reduce dementia risk. This is critical because blacks have double the risk of dementia as whites.

    What Is Relevant?

    • Hypertension is a modifiable risk factor associated with late-life dementia.

    Summary

    Higher systolic BP, lower diastolic BP, and higher pulse pressure were associated with faster declines in global cognition over 8 years in older adults. Systolic BP-related cognitive declines were greater in blacks and men.

    eLetters(0)

    eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

    Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.