Blood Pressure Trajectories Across the Life Course

Abstract

High blood pressure (BP) is a strong modifiable risk factor for cardiovascular disease (CVD). Longitudinal BP patterns themselves may reflect the burden of risk and vascular damage due to prolonged cumulative exposure to high BP levels. Current studies have begun to characterize BP patterns as a trajectory over an individual’s lifetime. These BP trajectories take into account the absolute BP levels as well as the slope of BP changes throughout the lifetime thus incorporating longitudinal BP patterns into a single metric. Methodologic issues that need to be considered when examining BP trajectories include individual-level vs. population-level group-based modeling, use of distinct but complementary BP metrics (systolic, diastolic, mean arterial, mid, and pulse pressure), and potential for measurement errors related to varied settings, devices, and number of readings utilized. There appear to be very specific developmental periods during which divergent BP trajectories may emerge, specifically adolescence, the pregnancy period, and older adulthood. Lifetime BP trajectories are impacted by both individual-level and community-level factors and have been associated with incident hypertension, multimorbidity (CVD, renal disease, cognitive impairment), and overall life expectancy. Key unanswered questions remain around the additive predictive value of BP trajectories, intergenerational contributions to BP patterns (in utero BP exposure), and potential genetic drivers of BP patterns. The next phase in understanding BP trajectories needs to focus on how best to incorporate this knowledge into clinical care to reduce the burden of hypertensive-related outcomes and improve health equity.

High blood pressure (BP) is a major modifiable risk factor for cardiovascular disease (CVD), including coronary heart disease, heart failure, and stroke. Every year over 7.6 million deaths worldwide are attributed to high BP.¹ In the United States, the burden of cardiovascular mortality related to hypertension is growing with increases in age-adjusted mortality rates for cardiovascular deaths related to hypertension between 2000 and 2018.² Similarly, worldwide, the prevalence of hypertension is expected to increase 60% from 2000 to 2025 for a total of 1.56 billion individuals.³ To date, the vast majority of research examining the relationship of BP with clinical outcomes has focused on a single measurement in time. With the advances in statistical methods as well as a growing understanding of the physiology of BP, recent studies have characterized the effects of longitudinal BP patterns. Among most individuals, BP changes over the life course with individuals experiencing increasing BP with age. This age-related rise in BP is not solely a result of aging and, in fact, in some studies of indigenous populations among the INTERSALT study⁴ and additional follow-up among younger Yanomami peoples of South America^5,6 have shown that among populations with low salt intake there is little to no association of BP with age. It is likely that increases in BP with age result from an interaction of age with accumulation of lifestyle and behavioral risk factors.

Longitudinal BP patterns themselves may reflect the burden of risk and vascular damage due to prolonged cumulative exposure to high BP levels. Current studies have begun to characterize BP patterns as a trajectory over an individual’s lifetime. These BP trajectories take into account the absolute BP levels as well as the slope of BP changes throughout the lifetime thus incorporating longitudinal BP patterns into a single metric. The goal of this narrative review article is to provide an overview of methodologic considerations of trajectory analysis, summarize data on predictors and outcomes associated with BP trajectories (Figure 1), and highlight unanswered questions in key periods across the life course.

Figure 1.

Methodologic inputs and outcomes related to cumulative blood pressure exposure modeled as trajectories. Abbreviations: BP, blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; PP, pulse pressure; SBP, systolic blood pressure.

METHODOLOGIC CONSIDERATIONS

Individual vs. population BP trajectories

There are 3 main methodologic ways of conceptualizing BP trajectories. The first is that each individual in the population has their own unique trajectory. In this type of study, we include random intercepts and slopes to incorporate individual variation within a multilevel regression model. The second approach, on the opposite end of the spectrum is to assume that there is one underlying population trajectory. While individuals will vary around this population trajectory, those variations represent individual physiologic responses. Lastly, there are group-based methods that assume multiple different underlying trajectories exist within the population and that individuals belong to one of those groups. The choice of which way to approach the development of BP trajectories depends on the scientific question under study as well as the choice between balancing the fit of the data with the ability to pragmatically identify groups of individuals of interest.

BP metric included in trajectory modeling

BP, even at a single point in time, represents a complex physiologic phenomenon that can be difficult to capture with a single number. Thus, the choice of BP metric is highly variable across studies. BP changes throughout the day and night within each individual and in response to particular stressors. Current clinical consensus uses systolic and diastolic BP to guide treatment decisions and make assessments on clinical risk.

In light of these considerations, the vast majority of studies examining BP have focused on systolic BP and diastolic BP with a primary focus on systolic BP trajectories given its stronger association with risk in adulthood.⁷ The classic view has been that diastolic BP is a more robust predictor of future CVD risk in young adulthood, with systolic BP becoming more important in middle age. The studies that have examined both systolic BP and diastolic BP trajectories have generally found a stronger relationship of systolic BP trajectories with outcomes of interest likely due to the ages of the participants.⁷ However, a recent study done among Black and White participants of the Coronary Artery Risk in Young Adults (CARDIA) Study found racial differences in which metric was more strongly associated with premature CVD events with systolic BP being more predictive in young-adult black individuals and diastolic BP being more predictive of future CVD in young-adult white individuals.⁸ Interestingly, by middle age, systolic BP was the best predictors of incident CVD for both races.

There has been relatively less work on metrics that combine the information on diastolic BP and systolic BP in order to simultaneously examine their effects. Given the longitudinal nature of BP trajectories across the life course and the differential effects of diastolic BP and systolic BP with age, mid-BP (the mean of the systolic BP and diastolic BP) offers a way to take into account systolic BP and diastolic BP jointly. Mid-BP may reflect pathophysiologic differences in the underlying health of the heart and vasculature (e.g., cardiac output, peripheral vascular resistance, and arterial stiffness). The association of systolic and diastolic BP on cardiovascular outcomes is different with age and thus mid-BP may represent a more consistent measure of future risk when examining BP trajectories across the life course. An alternative strategy was demonstrated in a recent study which examined clusters of systolic BP and diastolic BP trajectories thus providing groups of individuals with shared patterns in both systolic BP and diastolic BP.⁹ In this study of older adults in the Cardiovascular Health Study (CHS), there was clustering of systolic BP and diastolic BP trajectories which were significantly related to mortality.⁹ Alternatively, pulse pressure (PP: systolic minus diastolic blood pressure), which reflects an index of the pulsatile component of the cardiac cycle, may be important to examine. PP has been associated with adverse clinical outcomes¹⁰ (e.g., myocardial infarction, heart failure¹¹) in general and high-risk populations (e.g., heart failure,¹² end-stage renal disease¹³). However, conflicting results have been published regarding the predictive utility of PP with a lack of association in younger, healthier populations where diastolic BP tends to be more predictive.^14,15

Accuracy and precision of various BP measurement techniques

A third important methodologic consideration when examining BP trajectories is the accuracy and precision of the BP measures being assessed longitudinally. Measurement error may potentially influence trajectory patterns and may itself be influenced by individual and study characteristics.¹⁶ Much remains unknown about how methods for assessment of BP and the frequency of measurements influence long-term BP trajectories. A rigorous approach to the accurate and precise measurement of BP is essential.¹⁶ While direct intra-arterial measurement of BP is the gold standard, this is not practical, and noninvasive approaches with manual or automated devices are routinely utilized. BP machines should be properly calibrated, appropriate cuff size selected for each individual, and standardized techniques utilized for high-quality assessment of BP.¹⁷ BP measurements can be obtained in the clinic as part of routine care or as part of a research study, which may be automated or manually obtained. Measurements should be obtained in both arms at the initial visit.¹⁸ Typically, the first BP reading is higher than subsequent readings and guidelines recommend measurement of BP at least 2 times in the office to average if not using automated office BP monitors.¹⁹ While the upper arm (over the brachial artery) is the standard location for monitoring, many home monitoring devices use alternative locations (wrist, finger), but this may lead to inaccurate values due to poor reproducibility. Measurement error can be related to individual-level factors (recent caffeine intake), device-related (noncalibrated), or be procedure-related (incorrect cuff size). Because BP trajectories by definition incorporate multiple BP measurements over time they may be less sensitive to nondifferential misclassification and therefore stronger predictors of future outcomes simply because they incorporate more information than a single measurement at a single time point. However, there has been little methodologic assessment about the influence of how measurement error associated with specific individual and study factors may impact BP trajectories.

Alternatively, BP may be measured at home, either by the individual participant or through ambulatory BP monitoring (ABPM). ABPM uses a device that is worn for 24–48 hours with repeated measurements every 15–20 minutes during the daytime and 30–60 minutes at nighttime to examine the average day (diurnal) and night (nocturnal) BPs. In addition, the percentage of BP readings that are above the threshold for hypertension can be tracked and compared over time. One of the only studies to utilize ABPM data to create long-term BP trajectories was done among youth (n = 663) which had on average of five 24-hour ABPMs over 15 years of follow-up.²⁰ Interestingly, they created trajectories for daytime and nighttime BP levels jointly and identified racial differences in youth.²⁰ Use of remote monitoring and mobile health devices is rapidly emerging and may enhance development of pattern-based BP with the ability to include a higher frequency of measurements and rapid transmission for data analysis and determination of BP trajectories.²¹ BP trajectories using home BP monitoring are unknown.

STATE OF THE SCIENCE REGARDING BP TRAJECTORIES

To summarize the current state of the science on BP trajectories and highlight key findings, we conducted a narrative review consisting of a search of PubMed using key terms related to BP and trajectories across the life course (Figure 2).

Figure 2.

Conceptual figure outlining key periods across the life course with 3 theoretical blood pressure trajectories.

CRITICAL LIFE PERIODS FOR BP TRAJECTORIES

Adolescence

There appear to be very specific developmental periods during which divergent BP trajectories may emerge (Figure 3). The influence of adverse childhood experiences as well as puberty on BP contribute to the development of hypertension later in life. BP trajectories early in life are often highly variable and set the individual on their adult BP trajectory. Among children in the Bogalusa Heart Study, they found that childhood BP trajectories were associated with the risk for hypertension as an adult.²² In particular, BP at the time of puberty was critical in the development of hypertension. Even after adjustment, BP slopes during the period of puberty (ages 13–19 years) were significantly positively associated with adult hypertension.²² Adolescence is also the period where we start to see sex differences emerge such that the slope of systolic BP increases for boys during puberty setting them up for higher BP trajectories throughout adulthood.^23,24 The slopes for systolic BP among girls then start to rise in²⁵ early adulthood, and by the seventh decade of life, gender differences in BP are less evident.

Figure 3.

Conceptual figure highlighting the heterogeneity of blood pressure trajectories from adolescence to midlife.

Pregnancy

Pregnancy represents an important life period in BP trajectories for women. Physiologic changes are evident in BP trajectories during pregnancy with studies reporting drops in BP during the first and second trimesters of pregnancy. Studies with prepregnancy and pregnancy measures are limited, and thus, it is unclear how prepregnancy BP may modify the pregnancy BP trajectory. It remains unknown whether a woman’s long-term BP trajectory is altered by adverse events in pregnancy or whether pregnancy BP trajectories simply unmask preexisting BP risk. In a recent study from the HUNT cohort, declines in BP during pregnancy persisted after giving birth and accumulated over pregnancies (with lower drops in BP for subsequent pregnancies) such that parous women had lower mean systolic BP and diastolic BP at age 50 years compared with nulliparous women. In contrast, studies of women with preeclampsia, high BP during pregnancy associated with proteinuria have been associated with higher burden of CV risk factors, coronary artery calcium in midlife, and CVD.²⁶

Older adulthood

As mentioned above, BP trajectories become more variable in the latter decades of life with many papers demonstrating a decline in BP during the 80s and 90s years of age. In part this may be due to increasing proportion of the population on antihypertensive medication as well as the fact that mortality rates are higher for individuals with higher BP levels and in greater risk trajectories. However, there is some indication that this decline in BP may be related to declines in overall health.²⁷ Whether these BP declines signal future health risk or rather they represent the first markers of underlying declines in health remains unknown.

PREDICTORS OF BP TRAJECTORIES

Contextual and environmental influences

Lifetime BP trajectories are impacted by a broad range of environmental and societal influences. As countries have experienced changes so to we have seen temporal trends in population BP trajectories. In China, socioeconomic changes were associated with increases in BP trajectories such that BP levels rose and the increases in BP occurred at early ages.²⁸ Future research is needed to understand the temporal changes in population BP trajectories as countries experience demographic transitions.

Both the Institute of Medicine²⁹ and the American Heart Association³⁰ have described the central role of social determinants of health (SDoH) in BP, cardiovascular health, and CVD. Addressing SDoH, which include “structural determinants and conditions in which people are born, grow, live, work, and age” that affect health, functioning, and quality of life³¹ likely offers the most significant opportunities for altering BP trajectories beginning early in the life course, especially among disadvantaged populations. Indeed, disparities in BP patterns and BP-related morbidity and mortality significantly contribute to healthy and overall years of life lost. This may be, in part, related to the role of structural and systemic racism on BP. Among 407 African American and European American preterm children born in 1985, the rate of increases in systolic BP and diastolic BP between 2 and 7 years of age was greater for African American children.³² Neighborhood socioeconomic status (SES) explained 29% of the racial difference in the BP trajectories while family SES did not account for racial differences.³² These findings point to the critical impact that neighborhood environment can have on long-term BP and its role in racial disparities in BP trajectories.

Demographic and SES

An individual’s BP trajectory is highly influenced by their individual demographic and SES characteristics. There are important gender differences in lifetime BP trajectories. BP trajectories for women and men diverge at specific critical periods (adolescence and pregnancy) as noted above. In general, girls experience lower BP levels in childhood than boys, however, the rate of BP increases is steeper for females by early adulthood^24,25 and may be compounded by adverse pregnancy outcomes during childbearing years. By the seventh decade of life, women and men have similar BP levels and rates of change throughout older adulthood.²⁴

In the United States, African American individuals bear a disproportionate burden of hypertension-related morbidity.³³ These racial disparities in BP appear early in life. Already by age 8 years significant differences in diastolic BP trajectories through age 18 years for Black vs. non-Black participants of the Project HeartBeat! Study were identified.²³ In that study, no significant racial differences were found for systolic BP in childhood; however, in the Bogalusa Heart Study significant racial differences were noted starting at age 15 years onwards.²² Studies of BP trajectories in adulthood have consistently noted that African American participants are more likely to be in high-risk BP trajectories.⁷ Recent data using multiple ABPM measurements over 15 years have showed important racial differences in daytime vs. nighttime systolic BP trajectories by race.²⁰ African Americans experienced higher nighttime systolic BP and greater increases in nighttime systolic BP during young adulthood than the European American individuals²⁰; thus supporting cross-sectional studies demonstrating racial differences in ABPM phenotypes.

Individual-level SES has been shown to be associated with BP trajectories across the life course.

In an analyses of 7 UK cohorts (n = 30,372) spanning from age 7 years through 80 years the one occupational cohort, made up of civil servants, as compared with the broader population-based cohorts was found to have lower mean systolic BP levels, shallower rises in systolic BP in midlife and later acceleration of systolic BP increases thus leading to lower cumulative exposure to high BP.²⁴ It was hypothesized that these differences were due to higher SES status of occupational cohorts however, there was limited ability to disentangle the SES effects. In a single UK birth cohort study, childhood social class was a stronger mediator than adult SES of systolic BP trajectories in midlife and individuals in a lower social class as children had higher systolic BP levels throughout middle age (36–53 years).³⁴ Within the US participants of the Framingham Heart Study with 17 of more years of education had significantly lower systolic BP levels over 30 years of follow-up with a mean systolic BP 3.3 mm Hg lower than individuals with lower education.³⁵

Early childhood factors

Early childhood exposures are highly influential on adult BP trajectories. Exposure to adverse childhood experiences (including abuse, neglect, and household dysfunction) has been associated with the slope of BP increases in early adulthood that are independent of negative health behaviors such as illicit drug use.³⁶ In addition, as discussed above SES in early life seems to be even more influential than adult SES on lifetime BP trajectories. Our understanding of the prenatal influences on the offspring’s BP trajectories is just starting to be explored. Among 1,370 children included in Project Viva and recruited prior to birth, maternal factors including gestational hypertension/preeclampsia, chronic hypertension, gestational diabetes, and smoking during pregnancy were related to the offspring’s BP throughout childhood.³⁷ Children who were fed formula milk during infancy had higher systolic BP levels and greater increases in systolic BP from childhood through adolescence.³⁷ Many of these relationships between adverse pregnancy outcomes and offspring BP trajectory have been consistent in other cohorts.³⁸ However, in a very different population, the Birth to Twenty Plus (BT20) Cohort, in South Africa no association was found between prenatal and breastfeeding factors within offspring BP.³⁹ Further work is needed to disentangle these conflicting findings with particular consideration to the confounding effects of SES and cultural context.

Lifestyles and behaviors

At an individual level, health behaviors and lifestyle choices are important predictors of BP levels. Epidemiologic studies have found that physical activity and fitness are associated with lower BP levels at specific points in time.⁴⁰ Increasing cardiorespiratory fitness has been shown to lower BP levels in intervention trials.⁴⁰ However, data on the relationship between physical activity, fitness, diet, and smoking on BP trajectories are scarce. Recent data from the Aerobics Center Longitudinal Study of 13,953 men followed for up to 36 years demonstrated that men with higher fitness levels experienced a slower increase in BP over time after 20 years of follow-up and had a later onset of prehypertension or hypertension.⁴¹ Men who were in the highest tertile of fitness crossed the systolic BP threshold of 120 mm Hg at an average 54 years of age compared with 46 years in the lowest tertile of fitness, leading to a delay in hypertension by 6 years.⁴¹ Little is known about the impact of diet on BP trajectories. Limited information from an older cohort of Russian participants found no association between fruit and vegetable intake with BP trajectories over 12 years of follow-up.⁴² However, no association was seen with cross-sectional BP levels either which conflicts with prior epidemiologic studies demonstrating an association between diet and BP.¹⁹ Importantly, there is no information on overall diet quality or other important dietary components such as salt intake on BP trajectories. Understanding the lifestyle and behavioral influences of BP trajectories will be critical and could help to identify targets for interventions focused on improving long-term BP trajectories and subsequent outcomes.

Antihypertensive treatment

There is strong and consistent clinical trial evidence that treating high BP reduces the risk for future outcomes including CVD, renal disease, and death.¹⁹ With clinical advances in BP management and hypertension treatment, a growing proportion of individuals are on antihypertensive medications which increases with age. Initiation of antihypertensive medication alters an individual’s BP trajectory and influences our understanding of population-wide patterns in BP trajectories individuals with the highest BP levels and trajectories will be most likely to be treated. Of particular interest may be an individual’s BP trajectory while on treatment. Their response to antihypertensive medication and control of BP while taking treatment may involve both the patient’s adherence to the medication regime and physiologic response to treatment. In a post hoc analysis among patients in the SPRINT Trial BP trajectory mediated the relationship between intensive BP lowering and cardiovascular outcomes such that only the systolic BP trajectory group which maintained on target systolic BP in the intensive treatment arm had significantly lower CVD outcomes.⁴³ In general population, however, medication adherence may not be the key driver of on-treatment BP trajectory. Among patients in an HMO with hypertension and incident coronary heart disease, Maddox et al. found that treatment intensity was associated better BP trajectories in the year after CAD diagnosis while medication adherence was not.⁴⁴ Better 1-year BP trajectories, reflecting good BP control, were associated with lower rates of myocardial infarction and revascularization.⁴⁴ A better understanding of how treatment alters an individual’s BP trajectory and specifically how treatment decisions impact and are reflected in shorter-term BP trajectories is needed.

OUTCOMES RELATED TO BP TRAJECTORIES

BP has long been known as a major modifiable risk factor for a variety of diverse outcomes. Over the past decade, studies have demonstrated a significant association of BP trajectories the diagnosis of hypertension,⁴⁵ subclinical CVD,^7,46 cardiovascular events,^9,43 stroke,^47,48 renal disease,⁴⁹ dementia,⁵⁰ frailty,^27,51,52 and ultimately death.^9,51,52 What is of particular importance is whether these BP trajectories add to our understanding of an individuals’ risk for events above and beyond single (or short-term) BP measures. Relatively few studies have specifically focused on this question, although, those that have found a significant relationship of BP trajectories with outcomes even after adjustment for cross-sectional BP measurements.⁷ Little is known about how long individuals need to be followed in order to generate BP trajectories and identify groups at high risk for outcomes. Current studies span several months to over 50 years. It is likely that the number of BP measures and length of follow-up vary by life stages with more information needed during periods of life in which BP levels are more dynamic (e.g., during adolescence, childbearing, and late in life) and less needed during periods of more stable changes such as during middle age. Future studies will need to be done to specify best practices in terms of data needed to accurately quantify BP trajectories. These findings have important implications for the future of risk prediction. Understanding whether the inclusion of BP trajectories in risk prediction algorithms will improve risk prediction remains unknown. In addition, future implementation work is needed to determine how to capture, identify, and communicate an individual patients’ BP trajectory within the electronic health record system which is now possible given recent advances in the technologic capabilities of electronic health records. Lastly, clinical trials specifically designed to consider BP trajectories as the outcome will be needed to guide clinical care and identify the most effective strategies for improving long-term BP trajectories.

CONCLUSIONS AND FUTURE DIRECTIONS

Research on BP trajectories has been accumulating over the past decade. BP trajectories allow us to incorporate a large amount of information into a single summary metric, which accounts for the complex interplay between BP levels and changes in BP over the life course. Importantly, the individual components of BP trajectories such as levels at specific ages, cumulative BP exposure, rate of increase (i.e., slope), and critical periods can all be analyzed within the framework of BP trajectories. We have strong evidence demonstrating the association of BP trajectories with a variety of cardiovascular, renal, and aging outcomes. As we move away from considering a single BP measurement in isolation, we will need to take advantage of advances in other fields to more accurately and precisely characterize BP trajectories. There is a variety of exciting work utilizing remote monitoring and passive data collection, which offers the possibility to have daily or even hourly information about an individuals’ BP. In contrast to the work we do in cohorts and electronic health records where we have BP levels separated by weeks or years these advances offer the opportunity to create much more nuanced and detailed BP trajectories. Secondly, the role of genetics to inform risk for adverse BP trajectories remains unknown. Identification of individuals who have or may develop a higher risk BP trajectory across the life course may facilitate treatment, screening, and prevention. Like most traits, BP is highly heritable and is consistent with polygenic inheritance whereby many common genetic variants (single nucleotide polymorphisms) each individually contribute a small effect and may, in aggregate, contribute to risk. However, it is unclear if genome-wide polygenic scores (GPS) will identify individuals at significantly increased risk for worse BP trajectories. Finally, our next phase in understanding BP trajectories needs to focus on how best to incorporate this knowledge into clinical care to reduce the burden of hypertensive-related outcomes and improve health equity.

FUNDING

Research reported in this publication was supported, in part, by the National Institutes of Health’s National Institute on Aging, Grant Numbers: P30AG059988 (SSK), R61NS120245 (NBA), R01HL149869 (NBA), R01HL148661 (NBA), R01AG058969 (NBA), HHSN268201800003I (NBA), and 75N92020D00004 (NBA). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was additionally supported by grants from the American Heart Association: 17SFRN33660752 and 18SRG343600001.

DISCLOSURE

The authors declared no conflict of interest.