Authors: Swati Sakhuja (1University of Alabama at Birmingham, Birmingham, AL), Byron C. Jaeger (2Wake Forest School of Medicine, Winston-Salem, NC), Yuichiro Yano (3Department of Family Medicine and Community Health, Duke University, Durham, NC; 4Yokohama City University, Kanagawa, Japan), Daichi Shimbo (5Columbia University, New York, NY), Cora E. Lewis (1University of Alabama at Birmingham, Birmingham, AL), Donald Clark, III (6University of Mississippi Medical Center, Jackson, MS), Gabriel S. Tajeu (7Temple University, Philadelphia, PA), Shakia T. Hardy (1University of Alabama at Birmingham, Birmingham, AL), Norrina B. Allen (8Northwestern University, Chicago, IL), James M. Shikany (1University of Alabama at Birmingham, Birmingham, AL), Joseph E. Schwartz (5Columbia University, New York, NY; 9Stony Brook University, Stony Brook, NY), Anthony J. Viera (3Department of Family Medicine and Community Health, Duke University, Durham, NC), Paul Muntner (1University of Alabama at Birmingham, Birmingham, AL)
Categories: Article, Ambulatory blood pressure monitoring, race differences, blood pressure load, peak blood pressure, peak increase in blood pressure, blood pressure variability
Source: Blood pressure monitoring
Authors: Swati Sakhuja, Byron C. Jaeger, Yuichiro Yano, Daichi Shimbo, Cora E. Lewis, Donald Clark, Gabriel S. Tajeu, Shakia T. Hardy, Norrina B. Allen, James M. Shikany, Joseph E. Schwartz, Anthony J. Viera, Paul Muntner
Mean systolic and diastolic blood pressure (SBP and DBP) on ambulatory blood pressure (BP) monitoring (ABPM) are higher among Black compared with White adults. With 48 to 72 BP measurements obtained over 24 hours, ABPM can generate parameters other than mean BP that are associated with increased risk for cardiovascular events. There are few data on race differences in ABPM parameters other than mean BP.
To estimate differences between White and Black participants in ABPM parameters, we used pooled data from five US-based studies in which participants completed ABPM (n=2,580). We calculated measures of SBP and DBP level, including mean, load, peak, and measures of SBP and DBP variability, including average real variability (ARV) and peak increase.
There were 1,513 (58.6%) Black and 1,067 (41.4%) White participants with mean ages of 56.1 and 49.0 years, respectively. After multivariable adjustment, asleep SBP and DBP load were 5.7% (95% CI: 3.5%,7.9%) and 2.7% (95% CI: 1.1%,4.3%) higher, respectively, among Black compared with White participants. Black compared with White participants also had higher awake DBP ARV (0.3 [95%CI: 0.0,0.6] mmHg) and peak increase in DBP (0.4 [95% CI: 0.0,0.8] mmHg). There was no evidence of Black:White differences in awake measures of SBP level, asleep peak SBP or DBP, awake and asleep measures of SBP variability or asleep measures of DBP variability after multivariable adjustment.
Asleep SBP load, awake DBP ARV and peak increase in awake DBP were higher in Black compared to White participants, independent of mean BP on ABPM.
The large number of BP readings obtained over 24 hours of ambulatory blood pressure (BP) monitoring (ABPM) allows for assessment of measures other than mean BP including measures of level and variability in BP. Measures of BP level that can be obtained from an ABPM recording include BP load (i.e., percentage of BP readings in hypertensive range) and peak or maximum BP. Measures of BP variability include average real variability (ARV) and peak increase in BP (i.e., difference between peak BP and mean BP). Prior studies have provided evidence that these parameters may also be associated with subclinical CVD and, therefore, may identify people at increased risk for CVD events [1–5].
Prior studies have reported that Black adults have higher mean asleep systolic BP (SBP) and diastolic BP (DBP) on ABPM as compared to White adults [6,7]. Black adults are also more likely to have arterial stiffness, endothelial dysfunction, abnormal baroreflex function, and altered sodium excretion [8–13]. These biological mechanisms are implicated in higher out-of-office BP and increased BP variability [14–16]. In addition, Black adults are more likely to live in disadvantaged neighborhoods, experience discrimination, and high levels of stress that increase the risk for hypertension.[17] Therefore, race differences in measures of out-of-office BP level and BP variability may exist. Awake and asleep mean BP on ABPM are also associated with increased risk for CVD, however, data on race differences in other measures of awake and asleep BP level and variability are limited [18,19]. Identifying differences in BP level and variability while awake and asleep between Black and White adults can provide information on interventions to reduce racial disparities in CVD risk.
The goal of the current study was to examine if differences exist in measures of BP level, including BP load and peak BP and measures of BP variability, including ARV and peak increase in BP, between Black and white adults. Also, we assessed race differences in mean BP on ABPM. To accomplish the study goals, we pooled data from five US population- and community-based studies in which ABPM was performed.
Data were pooled from five cohort the Jackson Heart Study (JHS), Coronary Artery Risk Development in Young Adults (CARDIA) study, Masked Hypertension (MHT) study, Improving Detection of Hypertension (IDH) study and North Carolina Masked Hypertension (NCMH) study. Details of the design and methods of each study have been described previously [3,20–23]. In the JHS, 1,148 participants volunteered to undergo ABPM at the baseline visit between 2000 and 2004 [20]. In the CARDIA study, ABPM was conducted as part of an ancillary study among 831 participants at the Year 30 Exam in 2015 and 2016 at the Birmingham and Chicago field centers [21]. In the MHT study, 892 participants completed ABPM between 2005 and 2012 [22]. The IDH enrolled and conducted ABPM among 408 adults between 2011 and 2013.[23] From 2012 to 2014, the NCHM study enrolled 420 participants who underwent ABPM [3]. The analysis was restricted to participants who had a complete ABPM recording (defined below; n=3,028). We further restricted the study population to include only White and Black adults resulting in the final sample size of 2,580 participants (Supplemental Figure 1). Each study’s protocol was approved by an institutional review board and all participants provided written informed consent. The study protocol for the analysis of de-identified data from the five studies was approved by the institutional review board at the University of Alabama at Birmingham.
Each study used standardized protocols in accordance with their quality control procedures. For all five studies, information on age, race and sex were assessed by self-report. Information on education, medication history (i.e., glucose-lowering and antihypertensive medication use) and health behaviors (i.e., smoking status and alcohol intake) were collected using questionnaires. Height and weight were measured using calibrated equipment following standardized protocols from which body mass index (BMI) was calculated. Blood and urine specimens were collected for biochemical testing and used for defining diabetes and albumin-creatinine ratio. Details on assessment and definitions of the variables are presented in Supplemental Table 1.
Standardized protocols were followed for the assessment of office BP and ABPM in each study (Supplemental Table 2). Office BP was measured two times in JHS participants and three times in CARDIA, MHT, IDH and NCMH study participants. An appropriately sized BP cuff was placed on participants’ right arm in the JHS and CARDIA studies, and on the non-dominant arm in the MHT, IDH and NCMH studies and BP was measured at ≥ 30-second intervals after participants had been seated in a quiet room for 5 minutes. This was done during the baseline visit in the JHS and Exam Year 30 in the CARDIA study, and at multiple study visits for the IDH, MHT and NCMH studies. For consistency across studies, office BP from the first visit for the MHT, IDH and NCMH studies was used. The average of the two office BP measurements for JHS participants and the second and third office BP measurements for CARDIA participants and the three office BP measurements for IDH, MHT and NCMH participants were used to define mean office BP.
Participants were fitted with a validated ABPM device on their non-dominant arm and provided instructions for wearing it for 24 hours. The awake and asleep periods during the ABPM procedure were determined using actigraphy supplemented by sleep diaries in the CARDIA, MHT, and IDH studies. Self-reported awake and asleep times were used in the JHS and NCMH studies. A complete ABPM recording was defined as having at least 70% of the total number of planned readings over 24 hours [24]. On average, participants had 35 awake, 17 asleep and 52 total systolic and diastolic BP readings. Definitions of measures of level of BP including mean BP, BP load and peak BP, and measures of BP variability including ARV and peak increase in BP on ABPM are presented in Table 1. Each ABPM parameter was calculated using BP measurements collected while participants were awake and asleep, separately. The values for the ABPM parameters over the full 24-hour monitoring period were calculated as a weighted average of the awake and asleep parameters. These weights were proportional to the amount of the 24-hour period that a participant spent awake and asleep.
Summary statistics for characteristics of participants by race and by study cohort were calculated. Additionally, the means and standard deviations (SD) of the measures of SBP and DBP level (load, peak and mean), and the measures of SBP and DBP variability (ARV and peak increase) during the awake, asleep and 24-hour period were calculated by race and by study cohort. We used pooled data from the five studies for the remainder of the analyses. The analysis conducted for awake SBP load is described below. Identical analyses were conducted for awake DBP load and each of the other measures for the awake, asleep and 24-hour ABPM periods, separately. Linear regression with sequential multivariable adjustment was used to estimate adjusted differences in awake SBP load for Black participants compared to White participants. The initial model included adjustment for age, sex and study cohort. A second model included these variables and having less than a high-school education, BMI, smoking, alcohol consumption, diabetes, albumin-creatinine ratio ≥30 mg/g and antihypertensive medication use. In a third model, for systolic ABPM parameters we additionally adjusted office SBP and for diastolic ABPM parameters, we additionally adjusted for office DBP. In a final model, for awake, asleep and 24-hour SBP load, we included additional adjustment for mean awake, asleep or 24-hour SBP, respectively, and for 24-hour DBP load, we included additional adjustment for mean awake, asleep or 24-hour DBP for awake, asleep, respectively. Also, we examined Black versus White differences in ABPM parameters within categories of office SBP and DBP defined by thresholds in the 2017 ACC/AHA BP Guideline (i.e., <120 and <80 mm Hg, ≥120 to <130 and <80 mm Hg, ≥130 to <140 and/or ≥80 to <90 mm Hg, and ≥140 and/or ≥90 mm Hg, respectively) adjusting for variables in the final model as described above. By design, participants in the IDH, MHT and NCMH were not taking antihypertensive medication. The only White participants in the pooled cohort taking antihypertensive medication (n=60) were from the CARDIA study. Due to the low number of White participants taking antihypertensive medication, analyses were not conducted stratified by antihypertensive medication use.
Random forests comprising 500 decision trees were applied to impute missing values for each variable, separately [25]. Ten imputed datasets were generated, analyzed, and their results pooled [26]. Data on albumin-creatinine ratio were missing among 237 (9.2%) participants. Less than 5% of participants had missing data for any of the other variables included in the analysis. All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC) and R version 4.1.0 (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined by a two-sided P value <0.05.
The mean age of White and Black participants in the pooled cohort was 49.0 (SD=10.6) and 56.1 (SD=10.5) years, respectively (Table 2). Compared with White participants, Black participants were less likely to be male and more likely to have less than a high-school education, be a current smoker and not consume alcohol. Additionally, Black participants were more likely than White participants to have diabetes and an albumin-creatinine ratio ≥30 mg/g and report taking antihypertensive medication. Mean BMI and office SBP were higher in Black participants as compared with White participants. Participant characteristics by study cohort are presented in Supplemental Table 3.
Among measures of BP level, Black as compared with White participants had higher awake, asleep and 24-hour mean, load and peak SBP and, higher asleep and 24-hour mean, load and peak DBP (Supplemental Table 4). For the measures of BP variability, Black as compared with White participants had lower awake, asleep and 24-hour SBP ARV, higher awake and lower asleep DBP ARV, higher peak increase in SBP over asleep and 24-hour period, and higher peak increase in DBP over awake, asleep and 24-hour periods. Supplemental Table 5 presents the measures of SBP and DBP level and variability by study cohort.
Black as compared with White participants had higher awake mean and peak DBP after multivariable adjustment including office DBP (Table 3-upper panel; model 3). After further adjustment for mean awake DBP on ABPM (model 4), there was no evidence of association between peak awake DBP and race. Asleep mean, load and peak SBP and DBP were higher in Black as compared with White participants after multivariable adjustment including office SBP or DBP (Table 3-middle panel; model 3). After further adjustment for mean asleep SBP or DBP on ABPM (model 4), asleep SBP and DBP load were higher in Black compared with White participants, whereas there was no evidence that peak asleep SBP or DBP were higher among Black compared with White participants. Mean SBP load over the 24-hour period was also higher in Black participants as compared with White participants (Table 3-lower panel).
Among those with office SBP and DBP <130 mm Hg and <80 mm Hg, Black participants had higher asleep SBP and DBP load as compared with White participants after full multivariable adjustment (Supplemental Table 6). Among those with office SBP ≥130 to <140 mm Hg and/or DBP ≥80 to <90 mm Hg, Black as compared with White participants had higher awake SBP load and peak SBP, and higher asleep SBP and DBP load.
Among measures of BP variability, there was no evidence of race differences between awake or asleep SBP ARV and peak increase in SBP (Table 4). Awake ARV in DBP and peak increase in DBP and 24-hour ARV in DBP were higher in Black participants as compared with White participants after multivariable adjustment including mean office DBP and awake DBP. The results were similar when blood pressure variability was defined as standard deviation or coefficient of variation in a sensitivity analysis (data not shown). Among participants with office SBP ≥130 to <140 mm Hg and/or DBP ≥80 to <90 mm Hg, Black as compared with White participants had higher peak increase in awake SBP [Supplemental Table 7].
In the current study, Black compared with White participants had higher asleep SBP and DBP load, asleep peak SBP and DBP and asleep mean SBP and DBP after multivariable adjustment including office SBP and DBP, respectively. The differences in awake BP measures between black and white participants were smaller than the differences in asleep BP measures. After further adjustment for mean asleep SBP or DBP, asleep SBP and DBP load were higher for Black compared with White participants. Awake ARV and peak increase in DBP were higher in Black participants as compared with White participants after multivariable adjustment including mean office and awake DBP. There was no evidence of race differences in awake measures of level of SBP or DBP, and awake and asleep measures of variability in SBP after multivariable adjustment.
High mean BP on ABPM is associated with increased risk for CVD events, independent of office-measured BP [27,28]. Other ABPM parameters have been associated with increased risk for CVD. For example, higher BP load and peak BP have been associated with increased prevalence and risk for end organ damage, coronary events and stroke [1,2,4]. A prior study suggested that BP load may be useful in monitoring the effect of antihypertensive medication for controlling BP.[29] Black as compared to White adults are also at higher risk for CVD and mortality [30]. However, there are few data examining differences in BP load and peak BP between Black and White adults or assessing these ABPM parameters separately when a person is asleep or awake. In the current study, Black participants had 5.7% more asleep SBP readings and 2.7% more asleep DBP readings in the hypertensive range as compared with White adults after multivariable adjustment. In addition, asleep peak SBP and DBP were higher in Black compared with White participants. High BP load may represent chronic pressure overload that may induce myocardial and vascular damage leading to increased risk for CVD events [31]. Similarly, peak BP may be a result of baroreceptor dysfunction and can damage the vasculature [32]. As Black compared to White adults are at higher risk for CVD, BP load during sleep may be useful for CVD risk stratification in this high-risk group.
BP variability may be relevant to the pathophysiology of hypertension and hypertension-related CVD. Black compared with White adults experience more daily stressors that can impact vascular resistance and BP regulation [33]. In a meta-analysis of 8,938 participants from 11 studies, ARV of SBP and DBP were associated with incident CVD, stroke, all-cause mortality and cardiovascular mortality [5]. In addition, diastolic ARV was a stronger predictor of CVD events as compared to systolic ARV. This could be due to increased arterial stiffness with age that may lead to less increase or even fall in DBP with concurrent increase in SBP, resulting in greater variability in BP [34]. Consistent with results from prior studies, Black adults in the current study had higher 24-hour ARV of DBP as compared with White adults after adjustment for mean DBP in office or ABPM [6,35]. However, there was no evidence of race differences in awake or asleep SBP variability parameters in the current study, although, they may hold prognostic significance in predicting the risk for future CVD events. In a prior study, a peak SBP that was 20% or more above the 24-hour mean SBP on ABPM was associated with an increased risk for stroke or coronary events [1]. Data on peak increase in BP in Black adults are scarce. In the current study, there was no evidence of race differences in awake or asleep peak increase in SBP, but Black participants had higher peak increase in awake DBP as compared with White participants after multivariable adjustment.
Various exogenous factors including physical activity, stress, sleep, and endogenous regulatory mechanisms consisting of humoral and neural pathways are implicated in changes in BP profile during awake and asleep periods [36,37]. In the current analyses of race differences in ABPM parameters stratified by office-measured BP categories, Black as compared with White participants with elevated office BP (i.e., SBP and DBP ≥120 to <130 and <80 mm Hg, respectively) had higher asleep SBP load. In addition, among those with stage 1 hypertension (office SBP ≥130 to <140 mm Hg and/or DBP ≥80 to <90 mm Hg), Black as compared with White participants had higher awake and asleep SBP load and higher awake peak SBP. Prior studies have shown that cumulative exposure to elevated BP over a long time may be associated with increased risk for CVD [38,39]. In addition, there is a substantial risk for misclassification of BP status among those with borderline hypertension based on office BP alone [40]. Therefore, awake and asleep SBP load on ABPM may be useful tool for identifying individuals at high risk for CVD.
The current study has several strengths. We pooled data from five population- and community-based US studies that included good representation of Black and White adults. Office BP and ABPM were conducted following standardized protocols. The results should also be interpreted in the context of known and potential limitations. Office BP measurements from a single visit were used in the analysis. As MHT, IDH, and NCMH studies did not include adults who were taking antihypertensive medication, only a small proportion of White participants in the pooled analysis were taking antihypertensive medication. The associations could be under-estimated as some of the covariates in the fully adjusted models may be causally associated with the ABPM parameters. There were differences in characteristics between Black and White participants. We controlled for these differences by performing multivariable adjustment. Finally, the estimates for the analyses stratified by office-measured BP categories should be interpreted with caution because of small sample size in each category of office-measured BP.
In conclusion, among asleep measures of level in SBP and DBP, mean and load SBP and DBP were higher in Black compared with White participants. In addition, Black participants had higher awake ARV of DBP and peak increase in DBP as compared with White participants. These findings suggest that Black compared with White adults have higher levels of several ABPM parameters, with larger Black:White differences being present during sleep versus while awake.