CHAPTER 23: EPIDEMIOLOGY OF HYPERTENSION

Elevated blood pressure (BP) is the top attributable cause of mortality globally, responsible for some 7.5 million deaths annually (approximately 12.8% of all deaths) and accounts for 57 million disability-adjusted life years.¹ Global prevalence of elevated BP in 2008 was approximately 40%, being highest (46%) in the World Health Organization African region and lowest in the Americas, around 35% overall for both men and women. Approximately one billion persons throughout the world have uncontrolled hypertension, an increase from 600 million in 1980.

Hypertension is the most prevalent chronic condition in the United States and the most common reason for an office visit to a physician. It accounts for the most drug prescriptions and is one of the most important risk factors for heart disease and stroke.² This chapter reviews the epidemiology of hypertension—how it is defined and classified, the extent of its prevalence, treatment, and control, and its relation to cardiovascular and other consequences.

Hypertension can be quantified on the basis of a large number of epidemiologic studies showing that the distribution of BP in the population is continuous, although the curve is skewed at the higher levels of BP. The unimodal distribution of BP implies that hypertension is unlikely to be the result of a single physiologic process or gene, and perhaps most importantly suggests that any BP level used to define hypertension is arbitrary.

Hypertension can be classified in different ways—helpful for its diagnosis and clinical management (Fig. 23–1). The two principal divisions are severity (the height of the BP) and underlying cause (primary or essential hypertension vs secondary hypertension). A third major component is age: the pathophysiology of hypertension in younger and older people is quite different.

The original subdivision of hypertension according to its severity was benign and malignant. Although malignant hypertension carries a prognosis that is equivalent to that of other malignant diseases (if untreated), the term benign for less severe forms of hypertension is a misnomer and is no longer used. Malignant hypertension is now relatively uncommon in Western countries, but it does still occur and, when present, it requires urgent treatment, which can dramatically alter its natural history.

In hypertensive patients, either or both the systolic and diastolic BP are elevated, but elevations in systolic BP are most common. In particular, the most common circumstance is isolated systolic hypertension where the systolic BP is elevated, but the diastolic BP normal, occurring predominantly in older persons. Isolated diastolic hypertension can also occur and is the most common hypertensive subtype in younger persons. Hypertension can also be the result of the measurement method used; traditional methods of classification have all been based on office or clinic BP measurements, but with the wider use of out-of-office BP measurement, office measurements can be shown to significantly over- or underestimate the BP level during daily life. In most hypertensive patients, the BP is higher in the office than at other times, a phenomenon referred to as the white coat effect and that is usually defined as the difference between the office pressure and the average daytime pressure.

The Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC) 7 classification for hypertension² has continued the definition of hypertension as a BP at or above 140/90 mm Hg for adults aged 18 years or older. The classification is based on the average of two or more seated BPs, properly measured with well-maintained equipment, at each of two or more visits to the office or clinic. Hypertension has been divided into stages 1 and 2, as shown in Table 23–1. JNC 7 has defined normal BP as systolic BP less than 120 mm Hg and diastolic BP less than 80 mm Hg. What JNC-6 previously labeled as normal (120–129 mm Hg systolic BP or 80–84 mm Hg diastolic BP) and high normal (130–139 mm Hg systolic BP or 85–89 mm Hg diastolic BP) are now combined into a single group renamed as prehypertension, to increase awareness to these individuals with an intermediate level of risk that may progress to definite hypertension. The more recent American Heart Association/American College of Cardiology/Centers for Disease Control and Prevention Science Advisory³ “An Effective Approach to High Blood Pressure Control” has retained the JNC-7 cutoff points for stages 1 and 2 hypertension as well as the goal levels that were recommended for treatment, as has the recent clinical practice guidelines from the American Society of Hypertension and International Society of Hypertension.⁴

Prehypertension

It is estimated that approximately 15% of BP–related deaths from coronary heart disease (CHD) occur in individuals with BP in the prehypertensive range.⁵ JNC 7 officially defined prehypertension (systolic BP of 120–139 mm Hg or diastolic BP of 80–89 mm Hg) based on data mainly from the Framingham Heart Study. It showed (1) that people who have BP in the prehypertensive range have a higher risk of heart disease and stroke than do people with BPs below this level, with 18% of those with BPs of 120–129/80–84 mm Hg and 37% of those with BPs of 130–139/85–89 mm Hg progressing to clinical hypertension within only a few years,⁶ and that (2) the lifetime risk of hypertension approaches 90%,⁷ warranting the importance of the concept of prehypertension to help motivate physicians and patients alike to promote lifestyle modifications to prevent or delay the transition to clinical hypertension. Data from the National Health and Nutrition Examination Survey (NHANES) 1999–2006 in persons without cardiovascular disease (CVD) or cancer shows a prevalence of prehypertension of 36.3%, higher in men than in women, and associated with adverse cardiometabolic risk factors.⁸ A recent meta-analysis of 17 prospective cohort studies comprising 591,664 participants showed that, compared with optimal blood pressure (< 120/80 mm Hg), those with prehypertension had a 43% higher risk of incident coronary heart disease (CHD; hazard ratio [HR], 1.43; 95% confidence interval [CI], 1.26–1.63), a risk that was higher in Western subjects (relative risk [RR], 1.70; 95% CI, 1.49–1.94) than in Asian subjects (RR, 1.25; 95% CI, 1.12–1.38), with 24.1% of CHD attributable to prehypertension in Western subjects versus 8.4% in Asian subjects.⁹ A similar analysis by the same group, involving 19 prospective studies, showed prehypertension to be associated with a 66% greater risk of stroke; even those in the lower range of prehypertension (systolic BP 120–129 mm Hg or diastolic BP 80–84 mm Hg) had a significant 44% increased risk.¹⁰ Importantly, prehypertension has been recently shown to be associated with abnormalities of cardiac structure and function, specifically increased left ventricular remodeling and impaired diastolic function¹¹; inflammatory factors including C-reactive protein, interleukin-6, and tumor necrosis factor-alpha¹²; as well as increased arterial stiffness in older persons.¹³

Isolated Systolic Hypertension

In many older adults, systolic BP tends to rise as diastolic BP falls. When the average systolic BP is at least 140 mm Hg and diastolic BP is less than 90 mm Hg, the patient is classified as isolated systolic hypertensive. The increased pulse pressure (systolic-diastolic) and systolic pressure predict risk and determine treatment.¹⁴ The transition of hypertensive subtypes has been described from data from the NHANES: isolated diastolic hypertension is most common in younger persons whereas isolated systolic hypertension is the most common subtype in those aged 60 years and over, accounting for more than 90% of hypertension in those aged 80 years and over (Fig. 23–2).¹⁵

Isolated systolic hypertension can also occur in the young. High systolic pressure in younger persons may be the result of two changes: a high stroke volume and increased arterial stiffness. In contrast to essential hypertension, which raises both systolic and diastolic pressures, peripheral resistance is not increased.¹⁶ In long-term follow-up of 31 years, younger and middle-aged adults with isolated systolic hypertension had higher relative risk than those with high-normal BP.¹⁷

Isolated Diastolic Hypertension

More commonly seen in younger adults, a systolic BP of less than 140 mm Hg accompanied by a diastolic BP of 90 mm Hg or more defines isolated diastolic hypertension. This represents more than half of hypertension under the age of 40 years (see Fig. 23–2).¹⁵ Diastolic pressure is generally thought to be the best predictor of risk in patients aged younger than 50 years¹⁸; an analysis of data from the Framingham Heart Study also concluded that isolated diastolic hypertension may evolve into systolic and diastolic hypertension.¹⁹ The US NHANES survey also showed that obesity was associated with isolated diastolic hypertension in all age groups and both genders—but most frequently in young adults.²⁰ Thus, any patients in whom isolated diastolic hypertension is diagnosed should be carefully followed.

White Coat Hypertension

In approximately 15% to 30% of people with stage 1 hypertension, BP is elevated only persistently in the presence of a health-care worker, particularly a physician. When measured elsewhere, including while at work, the BP is not elevated. When this phenomenon is detected in patients who are not taking medications, it is referred to as white coat hypertension or isolated office hypertension. The commonly used definition is a persistently elevated average office BP of 140/90 mm Hg or more and an average awake ambulatory reading of less than 135/85 mm Hg (Fig. 23–3).²¹ Although it can occur at any age, it is more common in women, older adults, nonsmokers, and those with recently diagnosed hypertension; a misdiagnosis of a person with white coat hypertension as being truly hypertensive can result in his or her being penalized for employment and consideration for insurance, as well as possibly being subjected to lifelong antihypertensive treatment.²² This assertion was recently confirmed by Ogedegbe et al in a study in which patients with white coat hypertension reported significantly higher levels of state anxiety compared with other hypertension diagnostic categories.²³ Its magnitude can be reduced (but not eliminated) by the use of stationary oscillometric devices that automatically determine and analyze a series of BPs over 15 to 20 minutes with the patient in a quiet environment in the office or clinic. Although white coat hypertension has been generally thought to have a benign prognosis, a recent meta-analysis of 14 studies comprising 29,100 patients showed that cardiovascular mortality and morbidity in subjects with white coat hypertension is slightly higher than normotensive controls but well below the cardiovascular risks associated with sustained hypertension.²⁴ All patients should be followed indefinitely with office and out-of-office measurements of BP. Treatment with antihypertensive drugs may lower the office BP but does not change the ambulatory measurement. This pattern of findings suggests that drug treatment of white coat hypertension is less beneficial than treatment of sustained hypertension.

Masked Hypertension

The mirror image of white coat hypertension, masked hypertension is defined as a normal office BP (< 140/90 mm Hg) together with an elevated daytime BP (≥ 135/85 mm Hg) (see Fig. 23–3). It is present in approximately 10% to 40% of patients not receiving antihypertensive treatment. Also, persons with prehypertension are more likely to have masked hypertension and can even develop target organ damage before transitioning to established sustained hypertension. Twenty-four–hour ambulatory BP monitoring is the standard for diagnosing masked hypertension, but home BP monitoring can also be used. Importantly, reliance on in-office BP values to initiate treatment can result in up to a third of persons remaining at increased risk resulting from masked hypertension.²⁵ Masked hypertension is associated both with target organ damage²⁶ and an adverse prognosis,²⁷ and has been detected both in subjects who have not been diagnosed or treated for hypertension and in patients on antihypertensive treatment.²⁸ The Dallas Heart Study also recently showed a two-fold greater risk of future cardiovascular events in those with masked hypertension compared to those who were normotensive.²⁹ Importantly, masked hypertension is more common in older persons, those with mental stress, smokers, and those with metabolic syndrome, diabetes, chronic kidney disease, obstructive sleep apnea. Thus, there is an underdiagnosis of clinical hypertension in such persons, emphasizing an even greater importance for the use of 24-hour ambulatory BP measures in these individuals.²⁸ Furthermore, persons with masked hypertension frequently present with elevated nocturnal BP and a nondipping pattern that predicts worse cardiovascular outcomes; this again can only be diagnosed with 24-hour ambulatory BP monitoring.³⁰

Pseudohypertension

The term pseudohypertension is misleading; it suggests a false elevation of diastolic BP when in actuality it is low. This condition represents wide pulse pressure isolated systolic hypertension in the elderly. It frequently occurs with diabetes and diabetic kidney disease, and is associated with extensive calcification of many large arteries, including the brachial and elastic aorta. These conditions are obviously associated with severe CVD risk. Differential diagnosis must exclude benign white coat hypertension, which is a frequent finding in the elderly and is best diagnosed with ambulatory BP monitoring. Indeed, with the advent of 24-hour ambulatory BP measurement, description of cases of pseudohypertension have largely disappeared from the literature—except for false inclusion in textbooks.³¹

Orthostatic or Postural Hypotension

This condition is defined as a reduction of systolic or diastolic BP of at least 20 or 10 mm Hg, respectively, within 3 minutes of quiet standing.³² An alternative method is to detect a similar decline during head-up tilt at 60 degrees. This may be asymptomatic or accompanied by symptoms of light-headedness, faintness, dizziness, blurred vision, and cognitive impairment.³³ If chronic, the fall of BP may be part of pure autonomic failure or a complication of diabetes. The major life-limiting failure is inability to control the level of BP, especially in those patients with orthostatic hypotension who concomitantly have supine hypertension. In these patients, there are great and swift changes in pressure so that the patients faint as a result of profound hypotension on standing and have very severe hypertension when supine. Often the heart rate is fixed as well. The supine hypertensive person is subject to serious consequences, such as left ventricular hypertrophy³⁴ and stroke.^35,36

Hypertensive Crises: Urgency, Emergency, and Malignant Hypertension

A hypertensive crisis can present as urgency or emergency that requires immediate medical treatment.³⁷ A hypertensive urgency is defined as having two or more readings where the systolic BP is elevated at 180 mm Hg or higher or diastolic BP of 110 mm Hg or higher, without associated organ damage, and may or may not be accompanied by symptoms such as severe headache, nosebleeds, shortness of breath, or severe anxiety. This is distinguished from a hypertensive emergency where target organ damage has occurred (eg, chest pain, shortness of breath, back pain, numbness/weakness, change in vision, difficulty speaking) requiring the individual to seek immediate emergency medical assistance.³⁸ Malignant hypertension is defined as the presence of elevated BP in association with bilateral retinal hemorrhages and/or exudates, with or without papilledema. Although there have been substantial improvements in the general management of hypertension, there is no good evidence of reduction in its prevalence.³⁹ However, in the absence of retinopathy, a hypertensive emergency is based on the criteria of acute elevated BP accompanied by damage to one of many target organs. Its importance lies in the fact that if untreated, it has a 5-year survival rate of 1%.⁴⁰ It is still relatively uncommon in developing countries,⁴¹ but its prevalence may be increasing with the global increase in hypertension prevalence.³⁹ It can result from essential or secondary hypertension. Treatment of malignant hypertension has a dramatic effect on survival, and the first article demonstrating the benefits of antihypertensive drugs was based on patients with malignant hypertension.⁴²

Classification by Cause

In more than 95% of cases of hypertension, no single and reversible cause can be detected, and the terms essential and primary hypertension have been used. The former term was introduced because it was thought that a higher-than-usual level of BP was needed to maintain perfusion of vital organs.

In approximately 5% of cases there is a definable cause of the hypertension.² Table 23–2 shows a list adapted from JNC 7.² From an epidemiologic point of view, the two most important conditions on the list are chronic kidney disease and sleep apnea. Although chronic kidney disease is certainly a major cause of hypertension, it is often difficult to decide whether the hypertension or the kidney disease came first, because a vicious cycle can develop where one condition exacerbates the other. In practice, however, this distinction is of little consequence, because most forms of chronic kidney disease are not reversible. The most common curable form of hypertension is renal artery stenosis, which has two principal causes: (1) fibromuscular disease in children and young adults and (2) atherosclerosis in middle-age and older patients.

Sleep apnea is emerging as one of the major causes of hypertension that is of epidemiologic significance. Support for a causal association between sleep-disordered breathing and hypertension includes physiological mechanisms involving vascular dysfunction secondary to altered sympathovagal balance and insulin resistance. Also, respiratory disturbances accompanied by hypoxemia and sympathetic activation can trigger acute surges in BP.⁴³ A population survey showed that 2% of women and 4% of men have sleep apnea, which was defined as having an apnea-hypopnea index (AHI) score of five or more and daytime sleepiness.⁴⁴ The prevalence of an AHI of 5 or more increases with age, reaching a maximum prevalence in a population at approximately the age of 70 years.⁴⁵ Both sleep apnea and hypertension are common, and unsurprisingly there are many individuals who have both conditions. Furthermore, both are closely linked to obesity (particularly central obesity, as seen in the metabolic syndrome), so there is a cluster of related syndromes: hypertension, sleep apnea, diabetes, and the metabolic syndrome. Thus, approximately 60% of sleep apnea patients are hypertensive⁴⁶ and, conversely, approximately 25% of hypertensive patients have sleep apnea.^47,48 One issue is the causal link between sleep apnea and hypertension. The largest study of this is the Sleep Heart Health Study, a prospective study of the relationship between sleep apnea and cardiovascular morbidity. In an initial cross-sectional study of 6132 subjects aged 40 years or older,⁴⁹ a dose-response relationship existed between the AHI score and the prevalence of hypertension, although some of it was attributable to the effects of increased body mass index. A prospective follow-up of this cohort, restricted to the 2470 individuals without hypertension, showed effect sizes almost identical to that of the cross-sectional analyses: (1) an odds ratio of 2.19 in the age-, sex-, and ethnicity-adjusted models and (2) an odds ratio of 1.51 in a model that was further adjusted for body mass index. However, given the smaller sample in the prospective analysis, this did not quite reach statistical significance.⁵⁰ The Wisconsin Sleep Cohort Study did find a consistent dose-response relationship, even after controlling for age, sex, body mass index, and antihypertensive medications.⁵¹

Classification by Age

There is a fundamental difference between the genesis of hypertension in younger and older patients (Table 23–3). The most obvious difference is that in younger patients, whatever the underlying etiology of the hypertension (with a few exceptions noted below), both systolic and diastolic BPs are raised, whereas in people aged 60 years and older, the diastolic BP starts to fall (Fig. 23–4), but there is a marked increase of systolic BP. The underlying hemodynamics are also different: in younger patients, the characteristic changes are an increased peripheral resistance with a normal cardiac output, whereas in older patients, the reason for the selective increase of systolic BP is increased arterial stiffness.⁵² This has two consequences. First, when the left ventricle pumps into a stiffened aorta, systolic BP will be higher because the stiffer vessel will be less able to accommodate the stroke volume. Second, the velocity of the arterial pulse wave traveling out to the peripheral vessels will be increased. Just like a wave resulting from a stone dropped in a pool, the wave is reflected when it reaches the periphery, so that the pressure waves in the circulation will be a combination of the outgoing and reflected waves. In younger people, where the pulse wave velocity is low, the reflected wave arrives relatively late and coincides with the diastolic downslope of the incident wave, so that it has no effect on the systolic or diastolic pressure. But in older individuals, it returns earlier and forms a second or late systolic peak, which augments the height of the systolic pressure. Another difference concerns the physiologic measurements. In younger patients, the increased peripheral resistance is a result of active vasoconstriction that is mediated hormonally, particularly by the sympathetic nervous system and the renin-angiotensin system. In older patients with systolic hypertension, hormonal mediation is less important and the changes are mostly mechanical (eg, loss of elastin fibers in the media of the arterial wall).⁵² The affected vessels (principally the aorta and central elastic vessels) dilate and stiffen. The effects of increased aortic stiffness may be bidirectional; that is, hypertension will itself increase arterial stiffness, so there may be a vicious cycle.⁵³

Treatment thresholds are also different. In younger patients, it is clearly established that starting drug treatment when the pressure exceeds 140/90 mm Hg is beneficial. This may also be true in older patients, but the clinical trials that have investigated the benefits of treatment have almost all used an initial systolic BP of 160 mm Hg or greater as an entry criterion and have not lowered the pressure to below 140 or 150 mm Hg.⁵⁴ Thus, the benefits of treatment in older patients with systolic pressures below 160 mm Hg remain unproven. Of interest, however, the recently reported Systolic Blood Pressure Intervention Trial (SPRINT) of 9361 persons with diabetes with a mean age of 68 years did show treatment to a target of less than 120 mm Hg systolic BP compared to less than 140 mm Hg (achieved 121.4 vs 136.2 mm Hg) did result in a 25% reduced risk of the composite cardiovascular end point, with a trend toward greater benefit of treatment in those 75 years of age or older. Interpreting the SPRINT study results should be in the context of its use of an automated device for measuring BP that eliminated the white coat effect to obtain a more accurate estimate of cardiovascular risk.⁵⁵

There is some evidence that in the very old (aged 85 years or older), mortality may be higher in patients with the lowest BPs⁵⁶ and that lowering the diastolic pressure with treatment may actually increase mortality.⁵⁷ The benefit or harm of treating very old patients is currently being evaluated.⁵⁸

In the United States, an estimated 80 million adults, or 32.6% of those aged 20 years and older, have hypertension. The overall age-adjusted prevalence in 2009–2012 was 44.9% and 46.1% among non-Hispanic black men and women, respectively; 32.9% and 30.1% among non-Hispanic white men and women, respectively; and 29.6% and 29.9% among Hispanic men and women, respectively.⁵⁹ There is a direct age-related prevalence of hypertension, being less than 10% in those aged 20 to 34 years and increasing to more than 75% in those aged 75 years and older (Fig. 23–5). Since 1988–1994, there has actually been an increase in prevalence of hypertension among most gender-ethnic groups (Fig. 23–6), with non-Hispanic black women having the highest prevalence (42.9%) in 2007–2012. Overall during this time period, 82.7% were aware of their condition and 76.5% were currently being treated, but only 54.1% were under control. Non-Hispanic blacks tend to be most aware of their condition, with Hispanics least likely to be treated or controlled (Fig. 23–7).

Surveys in Europe show much higher rates of hypertension, with an overall prevalence of 44% reported in Europe.⁶⁰ A comparison of seven European countries’ data found the highest rate in Germany (55%) and the lowest in Italy (38%), with France, England, Spain, and Sweden all showing a prevalence between these two extremes. The prevalence in Canada is very similar to that in the United States (27%).⁶¹ The reasons for these wide differences are unknown but do not appear to be because of differences in measurements or sampling rates. In Egypt, the rate is also approximately 25%, whereas in China, where much of the population is still rural, the rate is lower (14%) but is increasing rapidly.⁶¹

More recently, a large analysis in 199 countries examining current levels and trends in systolic BP that involved 5.4 million subjects, a global average systolic BP in 2008 was noted to be 128.1 mm Hg in men and 124.4 mm Hg in women. In this study, women in some East African and West African countries had the highest mean systolic BPs of at least 135 mm Hg, whereas men living in the Baltic and East African and West African countries had mean systolic BP at least 138 mm Hg. Among higher income countries, those in Western Europe had the highest systolic BP. They also noted the greatest decreases in systolic BP between 1980 and 2008 to have occurred among women in Western Europe and Australasia (≥ 3.5 mm Hg) and among men in North America (2.8 mm Hg). Increases in systolic BP during this period, however, were noted in Oceania, East Africa, and southern and Southeast Asia.⁶²

African Americans

In the United States, hypertension is significantly more prevalent in African Americans than in whites. In the most recent NHANES 2009–2012 survey,⁵⁹ the prevalence in African American men was 44.9%, whereas in white men it was 32.9%. In this same survey, in women the prevalence was 46.1% for African Americans and 30.1% for whites. The 40-year incidence of hypertension in someone aged 45 years is also higher for blacks (92.7%) than for whites (86.0%).⁶³ A big issue here is whether the higher prevalence is genetic or environmental. Although the prevalence is higher than in whites in other countries (eg, Brazil),⁶⁴ there is a large body of literature showing that the rates of hypertension in Africans living in traditional rural societies are relatively low but increase markedly when they move to cities.⁶⁵

A huge effort has already been made to try to understand the reasons for the higher prevalence of hypertension in African Americans, almost all of which has made the underlying assumption that there is some genetically determined physiologic difference. However, it appears human hypertension is determined by several genes; genetic factors probably do not account for more than 50% of hypertension.⁶⁶ A recent comparative study involving large surveys of BP in populations of European versus African descent outside the United States indicated a lower range of hypertension prevalence in blacks (14%–44%) than in whites (27%–55%).⁶⁷ One of the most powerful pieces of evidence that environmental factors are important in the development of hypertension is a series of epidemiologic studies showing that when people move from a traditional rural setting to an urban westernized lifestyle, their BP goes up. For example, the Luo study from Kenya⁶⁵ showed those who had moved to Nairobi had higher pressures than people living in the villages, even if they had only been in the city for a month, and others reported the same pattern for other countries in sub-Saharan Africa.⁶⁸ Socioeconomic status, which is strongly associated with ethnicity in the United States, affects the prevalence of hypertension, particularly in African Americans.⁶⁹ Harburg first showed that BP in African Americans living in the inner city was highest in those people living in the worst-off neighborhoods.^70,71 A more recent analysis of the Atherosclerosis Risk in Communities (ARIC) study by Diez-Roux et al⁷² found that “neighborhood” effects were independently related to BP after controlling for other risk factors. In four US communities, people living in the neighborhoods with the lowest median house prices tended to have the highest blood pressure. This was most pronounced in the one community (Jackson, Mississippi) that was largely African American.

Hispanic Americans

One of the fastest-growing segments of the US population is Hispanics (Latinos), who are overtaking African Americans as the largest minority group. Most recent NHANES data from 2009–2012 show a prevalence of hypertension of 29.6% and 29.9% in Hispanic men and women, respectively. Many are recent immigrants, and they tend to be of relatively low socioeconomic status. If psychosocial factors are so important in determining the socioeconomic status gradient of disease, one might expect that Hispanics would also have higher mortality rates than whites. Surprisingly, this does not appear to be the case. For example, death rates in women in 2013 from CHD and stroke are lowest in Hispanics (61.3 and 27.6 per 100,000, respectively) compared to blacks (94.7 and 45.9 per 100,000, respectively) or whites (75.0 and 34.5 per 100,000, respectively).⁵⁹ There are two major subgroups of Hispanics in the United States: those from Mexico and Latin America, who are predominant in states such as Texas and California, and those of Caribbean origin, who are mostly in the Northeast and Florida. However, they can be classified mainly into those of Mexican, South American, Central American, Cuban, Puerto Rican, or Dominican backgrounds. As shown from recently published data from the Hispanic Community Health Study/Study of Latinos, enrolling more than 16,000 participants among four major US communities, the prevalence of hypertension appears to vary greatly according to background and gender, being lowest among South American women (17.2%) and highest in Dominican men (34.3%). Those of Cuban, Puerto Rican and Dominican background had overall higher prevalences of hypertension than those of Mexican, South American, or Central American background.⁷³ Some early studies suggested that, despite higher rates of obesity and diabetes in Mexican Americans, they have a lower all-cause and cardiovascular mortality than whites,⁷⁴ leading to what has been called the Hispanic paradox. However, the highly regarded San Antonio Heart Study found just the opposite—namely, that mortality from CVD is approximately 60% higher in Mexican Americans than whites.⁷⁵ The authors attribute the contrary findings of the earlier studies to underreporting of deaths in Mexican Americans.

Asian Americans

Only recently have national data on the prevalence of hypertension and other risk factors become available in Asian Americans from NHANES. In 2011–2012, 25.6% of non-Hispanic Asian adults aged 20 years and older were defined to have hypertension.⁷⁶ Earlier, the Multiethnic Study of Atherosclerosis showed the prevalence of hypertension to be similar in Chinese (39%) and whites (38%), lower than in African Americans (60%) and Hispanics (42%).⁷⁷ National data on self-reported hypertension from 2004–2006 have previously shown the percentage of people ever being told they had hypertension to be lowest in Chinese (17%) and Koreans (17%), as compared to Filipinos (27%) and Japanese (25%).⁷⁸

Although a dramatic age-related increase in the prevalence of hypertension exists, several important cardiovascular risk factors, including family history and genetic factors, educational/socioeconomic status, greater weight, physical inactivity, tobacco use, psychosocial stressors, diabetes, and diet (including dietary fat, higher sodium intake, lower potassium intake, and excessive alcohol) also relate to the likelihood of hypertension.⁵⁹ In fact, risk prediction models, such as those developed by the Framingham Heart Study, include age, sex, systolic BP, diastolic BP, body mass index, smoking, and parental history of hypertension, and predict the development of hypertension over 25 years,^79,80 although underestimating risk, perhaps because there are additional factors not included in the model that may attribute to development of hypertension.

Age is probably the factor most strongly associated with hypertension and the powerful age-related increase in hypertension prevalence has been described above (see Fig. 23–5). More than 75% of persons older than 75 years of age have hypertension, making it the most predominant risk factor for CVD in older adults. In those free of hypertension at age 55, 93% of men and 91% of women will develop hypertension by age 80.⁷ Systolic BP rises monotonically with age, whereas diastolic BP increases to about age 50, plateaus, and then decreases throughout the remainder of the life span. Increased arterial stiffness and decreased compliance is a major contributor to this increase in systolic BP and corresponding decrease in diastolic BP among most Western nations.¹³

Next to age, increasing weight may be the largest contributor to the development of hypertension; more than 42% of those with a body mass index of 30 kg/m² or greater have hypertension, compared to 28% in those who are overweight (body mass index, 25 to less than 30 kg/m²) and only 15% in those with a body mass index less than 25 kg/m².⁸¹ Of interest, increases in hypertension prevalence of 13% in men and 24% in women among US adults from 1988–1994 to 1999–2004 are almost completely explained by corresponding increases in body mass index, which when adjusted for, nearly completely diminish the increase in hypertension.⁸² In addition, the Coronary Artery Risk Development in Young Adults (CARDIA) study showed those young adults who maintained a stable body mass index did not have increases in blood pressure over 15 years, but those whose body mass index increased by 2 kg/m² or more had significant increases in BP during the same period.⁸³

Key lifestyle factors, in particular cigarette smoking and sodium intake, are important risk factors for hypertension. Cigarette smoking has an acute effect on increasing blood pressure, primarily through stimulation of the sympathetic nervous system with adverse effects on arterial stiffness and wave reflection. Also, hypertensive smokers are more likely to develop severe hypertension, including malignant and renovascular hypertension.⁸⁴ The incidence of hypertension is increased in those who smoke 15 or more cigarettes per day.⁸⁵ A 14-year longitudinal study of more than 5000 Japanese male workers showed smoking to be associated with the onset of both hypertension and systolic hypertension,⁸⁶ and a large Chinese cohort of 38,520 subjects showed both current smoking and previous cigarette smoking to be associated with an increased risk of hypertension, with a log-linear dose response relation found between duration of smoking and hypertension as well.⁸⁷

Dietary sodium is implicated by numerous genetic, epidemiological, migrational, and intervention studies to contribute to increasing BP in the population, and higher salt intake is likely to contribute to CVD mainly through its effects on BP but also independently by increasing arterial stiffness and albuminuria.⁸⁸ The INTERSALT study, probably among the best known of such studies, involved 10,079 persons and found a strong positive relation between sodium intake and BP, with an increase of 6 g/d in salt intake estimated to elevate systolic BP 9 mm Hg over 30 years.⁸⁹ More recently, both the European Prospective Investigation into Cancer (EPIC-Norfolk) and International Population Study on Macronutrients and BP (INTERMAP) studies confirmed these relationships.^90,91 Also, a recent Japanese study of 4523 participants showed higher versus lower sodium intake to be associated with a significant 25% greater risk of developing hypertension over approximately 3 years.⁹² There is also strong evidence that salt reductions of 4.4 g/d for at least 4 weeks can lower systolic and diastolic BP by 4 and 2 mm Hg, respectively.⁹³

Other predisposing factors to hypertension include gender. Men have a higher prevalence until about the sixth decade of life, after which a higher prevalence in women emerges and continues for subsequent decades; more women than men are affected by hypertension, mainly because of their longer life span. Race/ethnicity is also a major factor; as demonstrated above, African Americans have among the highest prevalences of hypertension compared to other major ethnic groups, and Mexican Americans have prevalences that are similar to that of whites. Excessive alcohol use and intake of an omnivorous as opposed to vegetarian diet also appears to be related to an increased likelihood of hypertension.⁵⁹

The Prospective Studies Collaboration data from numerous observational epidemiologic studies provide persuasive evidence of the direct relationship between BP and CVD. This meta-analysis aggregated data across 61 prospective observational studies that together enrolled 958,074 adults with 12.7 million person-years at risk, during which there were about 56,000 vascular deaths⁹⁴; strong, direct relationships existed between average BP and vascular mortality. These relationships were evident in middle-aged and older individuals (Fig. 23–8). Importantly, no evidence of a BP threshold was observed; that is, vascular mortality increased progressively throughout the range of BP, including the prehypertensive range, down to a systolic pressure of 115 mm Hg and a diastolic pressure of 75 mm Hg. Between the ages of 40 and 70 years, each 20 mm Hg increment of systolic pressure (and 10 mm Hg of diastolic pressure) is associated with a doubling in the risk of stroke, with the slope of the lines relating BP and risk attenuated in older people.

Relative Contribution of Blood Pressure Measures to Cardiovascular Risk

A continuing debate in the field of hypertension is the relative importance of the different components of the arterial pressure wave in determining cardiovascular risk. There are four candidates: systolic BP, diastolic BP, pulse pressure, and mean arterial pressure. An additional issue is whether the traditional brachial artery pressure should be used or the central aortic pressure. For many years, the diastolic pressure reigned supreme, and most of the early hypertension treatment trials used a high diastolic pressure as an entry criterion. This was reinforced by the publication of an analysis by MacMahon and coworkers⁹⁵ based on pooled data from 420,000 subjects, which showed a log-linear relationship between diastolic pressure and the risk of stroke and myocardial infarction. Although the importance of systolic pressure was never in doubt, it gained precedence over diastolic pressure with the publication of a series of epidemiologic studies showing not only that a high systolic pressure was the best predictor of risk in the elderly, but also that a low diastolic pressure was associated with increased risk.⁹⁶ An analysis by Franklin and coworkers¹⁸ examined the Framingham Heart Study data and provided an elegant solution to this apparent paradox. In subjects younger than 50 years of age, the best predictor of risk was a high diastolic pressure, but in those older than 60 years of age, systolic pressure was the best predictor, and the relationship between diastolic pressure and risk was now negative, so that a low diastolic pressure was related to higher risk.

Increased Pulse Pressure

It is well known that systolic pressure increases steadily with age, but after the age of 50 years diastolic pressure starts to fall. A number of studies suggest that pulse pressure may be the best predictor of risk in the elderly. In the JNC 7, however, which is still the official guideline for the evaluation and management of hypertension,² there is no mention of pulse pressure as a predictor. Over the past 20 years, there have been multiple observational studies and controlled trials showing the value of pulse pressure as an important risk factor for CVD. This was clearly shown by an analysis of the Multiple Risk Factor Intervention Trial (MRFIT) data, which found that the highest risk group was patients who had a systolic pressure above 160 mm Hg and a diastolic pressure below 80 mm Hg.⁹⁷ The Framingham Heart Study using the combination of mean arterial pressure and pulse pressure together, rather than any single BP component separately (systolic BP, diastolic BP, mean arterial pressure, or pulse pressure), improved the fit monotonically for predicting CVD collectively or CHD, heart failure, and stroke separately.⁹⁸ Indeed, introducing the interaction terms pulse pressure multiplied by mean arterial pressure further improved the fit over the main effects of the two-component models, indicating that the effect of one BP component on risk varied accordingly to the level of the other. When pulse pressure, a measurement of stiffness, was combined with mean arterial pressure, a measurement of resistance, one could relate the two major physiologic components of hydraulic load to clinical outcome; single BP components cannot do this (Fig. 23–9).

Furthermore, there was evidence of a diastolic J-curve of increased CVD risk in individuals with isolated systolic hypertension and low diastolic BP (ie, < 70 mm Hg), which occurred in about 30% of untreated persons of at least 60 years of age in the NHANES survey.⁹⁹ There was a three-fold greater prevalence of CVD events from the highest to the lowest strata in untreated isolated systolic hypertension. Advanced age, female sex, and diabetes mellitus, but not treatment status, were associated with the low diastolic BP and widened pulse pressure. Most recently, in a follow-up of more than 45,000 subjects in the international Reduction of Atherothrombosis for Continued Health (REACH) registry, increasing pulse pressure quartile, after adjustment for mean arterial pressure, BP medication, and other risk factors, was associated with increased risk of myocardial infarction, nonfatal stroke, cardiovascular hospitalization, and the composite CVD end point.¹⁰⁰ Moreover, the European Society of Hypertension has recognized widened pulse pressure as a distinct risk factor that is separate from elevated systolic BP in older individuals.¹⁰¹

Impact of Low Diastolic Blood Pressure

These findings support the point that a very low diastolic pressure in combination with a very high systolic pressure is associated with high risk. There are at least two explanations for this, which we may call direct and indirect. Taking the direct explanation first, it is possible that a very low diastolic pressure might impair coronary artery perfusion, which takes place during diastole. If this were the case, we should expect the adverse outcomes resulting from a low diastolic pressure to be ischemic cardiac events—but not strokes. The indirect explanation is that the low diastolic pressure is not harmful per se but is a marker for generalized CVD or, more particularly, increased arterial stiffness. It is generally accepted that the age-related increase of systolic and decrease of diastolic pressure that is seen in the brachial artery is largely the result of accelerated wave reflection found in stiff arteries.¹⁰² Indeed, the Framingham Heart Study quoted above,⁹⁹ and followed by a second study from the same group, would support stiff arteries as a major cause of low diastolic BP.¹⁰³ This community-based study showed that persons with an initial CVD event and persistent isolated systolic hypertension in combination of diastolic BP of less than 70 mm Hg, versus diastolic BP 70 to 89 mm Hg, had increased risk for recurrent CVD events; this included CHD as well as heart failure and stroke. Moreover, these findings support wide pulse pressure in combination with low diastolic BP as important risk factors, largely independent of antihypertensive treatment status.¹⁰³

Mean Arterial Pressure

Mean arterial pressure is typically defined as diastolic pressure plus one-third of the pulse pressure, measured directly by oscillometric devices, although the numbers are rarely displayed. In an analysis of three intervention studies in older patients, Blacher and coworkers⁹⁶ found that pulse pressure, but not mean arterial pressure, predicted cardiovascular outcomes. However, in younger patients, mean arterial pressure has been shown to be more important, especially in the prediction of stroke¹⁰⁴ in a study using ambulatory BP monitoring. One explanation for this finding may be that the patient with the high pulse pressure has increased arterial stiffness and is effectively being hit with a “double whammy”: the high systolic pressure in the central aorta increases the afterload on the heart, and the low diastolic pressure impairs coronary perfusion. The reason why mean arterial pressure is a better predictor of stroke may be that it is closer to diastolic than to systolic pressure, and it has been shown that the relationship between diastolic pressure and stroke is steeper than for coronary events.⁹⁵

Central versus Peripheral Arterial Pressure

The arterial pressure wave in the brachial artery looks very different from the waves recorded at more proximal sites, where the damage to target organs, such as the heart and brain, occurs. Central aortic pressure can now be estimated indirectly from measurements made noninvasively from peripheral sites, such as the radial artery. In older patients with stiff arteries, the central aortic systolic pressure is similar to the brachial pressure, whereas in younger subjects with compliant arteries, it is substantially lower.¹⁰¹ As shown above, different drugs have different effects on pulse pressure, and recent studies show that similar differences occur in their effects on central pressure.¹⁰⁵ There are, however, many different methods for measuring arterial stiffness,¹⁰⁶ and there is disagreement regarding which is the most reliable. Some have been shown to predict cardiovascular events,^107,108 and it is likely that such measurements may become part of routine clinical practice in the future.

Hypertension promotes complications in the brain, heart, and kidneys through two related mechanisms, both of which involve the effects of increased pressure on the arteries. The first is the effects on the structure and function of the heart and arteries, and the second is the acceleration of the development of atherosclerosis. The former is directly the result of the BP, whereas the latter requires an interaction with other risk factors for CVD, most importantly cholesterol. Thus, strokes are closely related to the direct effects of BP, whereas CHD is related to atherosclerosis, depending more on the combination of the effects of hypertension with other risk factors; therefore, the relationship between BP and events is steeper for stroke than for CHD events.

Stroke, CHD, left ventricular hypertrophy, peripheral arterial disease, heart failure, and chronic kidney disease are principal complications of hypertension. The 36-year follow-up data from the Framingham Heart Study show that while the relative impact of hypertension is greatest for stroke and heart failure (RRs, 2.6–4.0), because overall incidence of CHD is greater than that for stroke or heart failure, the absolute impact of hypertension on CHD is greatest, even though the RR is lower (2.0–2.2).¹⁰⁹ A recent meta-analysis has further emphasized the importance of increasing risk above normotensive levels, showing that starting at a systolic BP/diastolic BP of 115/75 mm Hg, CVD mortality doubles with each increment of 20/10 mm Hg throughout the BP range.¹¹⁰ Even compared to individuals with BPs of less than 120/80 mm Hg, those with BP levels of 130 to 139 mm Hg systolic or 85 mm Hg diastolic have a three-fold greater risk of CVD events.

Coronary Heart Disease

The Prospective Studies Collaboration of 61 studies¹¹⁰ found strong log-linear relationships between both systolic and diastolic pressure and the risk of CHD events in five deciles of age, ranging from 40 to 49 to 80 to 89, such that for each 20 mm Hg increase of systolic pressure, there was a two-fold increase of risk over a range from 115 to 180 mm Hg. For diastolic pressure, the risk doubled with a 10 mm Hg increase over a range of 75 to 100 mm Hg. There is an also an additive effect between various risk factors; for instance, the relation between systolic pressure and CHD risk is much steeper in patients whose cholesterol is high than in patients in whom it is normal.¹¹¹ This relationship has been reported in several countries, although the slope of the line relating BP and risk is shallower in countries where the overall risk of CHD is low, such as Japan.¹¹² The relationship between risk and BP in individuals who have already had an myocardial infarction is different and has been reported to be J-shaped for the first 2 years after the myocardial infarction (ie, there is a paradoxical increase of risk in those with the lowest BP [eg, < 110/70 mm Hg]); however, with longer follow-up, there is a positive relationship.¹¹³ The increased mortality at lower levels of pressure may be an example of reverse causality; that is, the pressure is low because of extensive damage to the heart. A history of hypertension per se does not necessarily lead to an increased mortality after an myocardial infarction, but it does predict reinfarction.¹¹⁴

The official recommendation for the treatment of hypertensive patients with CHD is that the target BP should be 140/90 mm Hg; whether further reduction is beneficial or harmful is controversial. The J-curve hypothesis was postulated on the basis of observations suggesting that if diastolic pressure was lowered below a certain level (~85 mm Hg), there was a paradoxical increase of events.¹¹⁵ The proposed explanation was that because coronary artery perfusion occurs during diastole, excessive reduction in combination with diseased coronary arteries would result in ischemia. Subsequent work supported this idea, because the J-curve was seen for myocardial infarction but not with stroke, where the adage that the lower the BP, the better still holds. A recent analysis of the International Verapamil-Trandolapril Study (INVEST) trial,¹¹⁶ which compared two drug regimens in post–myocardial infarction patients, confirmed that there was an increased risk of all-cause mortality and myocardial infarction. Again, there was no such relationship for strokes. Data from the Framingham Heart Study also support the idea of a J-curve phenomenon, but only in patients with previous myocardial infarctions and in the presence of a high systolic pressure as well as a low diastolic pressure.¹¹⁷ This would be consistent with the results of the Comparison of Amlodipine versus Enalapril to Limit Occurrences of Thrombosis (CAMELOT) trial,¹¹⁸ which randomized normotensive patients with documented CHD to amlodipine, enalapril, or placebo, and found that although both active drugs lowered the BP, amlodipine resulted in a lower rate of recurrent events than either placebo or enalapril. The BPs at entry to the study were 129/78 mm Hg, and there was a 5 mm Hg reduction with treatment. Consequently, at the present time, it would seem prudent to avoid excessive reductions of diastolic pressure in patients with known CHD, particularly if the systolic pressure is high. In practice, however, there is approximately a 3 mm Hg decrease in SBP for each 1 mm decrease in DBP in elderly persons with antihypertensive therapy; thus, there is no evidence of increased cardiovascular risk until DBP reaches levels below 65-70 mm Hg.

Heart Failure

Heart failure is now the leading cause of hospitalization for people aged 65 years and older in the United States, and unlike other complications of hypertension, its prevalence has been increasing over the past 30 years.¹¹⁹ For a 40-year-old man or woman, the “remaining lifetime risk” of developing heart failure is approximately 20%.¹²⁰ BP is a major contributor to heart failure with a two-fold greater risk in hypertensive versus normotensive men and three-fold greater risk in hypertensive versus nonhypertensive women; 90% of people with new cases of heart failure in the Framingham Heart Study had a history of previous hypertension.¹²¹ Heart failure has been estimated to result from hypertension in 59% of women and 39% of men with heart failure. These percentages are substantially higher than the population attributable risks for myocardial infarction resulting from hypertension, which are only 13% and 34%, respectively. This risk is much more strongly related to systolic than diastolic pressure.¹²² Treatment of hypertension in older people reduces the incidence of heart failure by approximately 50%.¹⁰⁵ The good news is that the incidence of heart failure is now decreasing in women (but not men), whereas survival has improved in both sexes.¹²³

Over recent decades, it has become apparent that approximately half of the patients who present with classic signs and symptoms of heart failure appear to have a normal ejection fraction of more than 50% on echocardiography. This group has been described as having "diastolic dysfunction", "diastolic heart failure", or "heart failure with preserved ejection fraction". Diastolic heart failure is thought to be responsible for as many as 74% of cases of heart failure in hypertensive patients.¹²⁴ In a study of 73 heart failure patients,¹²⁵ the Framingham Heart Study showed that half had normal systolic function, and about 75% of those with and without normal systolic function had a history of hypertension. Mortality was high in both groups but higher in those with systolic heart failure. It is generally accepted wisdom that hypertensive patients develop ventricular hypertrophy and heart failure because the increased arterial pressure places an excessive burden on the heart (afterload). Why should the same explanation not apply to diastolic heart failure? Patients with diastolic heart failure have been found to have stiffer aortas than normal,¹²⁶ which may result in accelerated wave reflection from the periphery and a combined increase of central aortic systolic pressure and reduced diastolic pressure. The latter will lead to decreased coronary artery perfusion during diastole and hence impair myocardial relaxation.

Cerebrovascular Disease

Stroke is the third most common cause of death throughout the world after CHD and cancer. Approximately 80% of strokes are ischemic, 15% are hemorrhagic, and 5% are caused by subarachnoid hemorrhage.¹²⁷ As with coronary events, there is a strong log-linear relationship between both systolic and diastolic pressure and stroke, although the relationship is steeper for strokes than for CHD events and much stronger for systolic than diastolic pressure.¹²⁸ Approximately 60% of patients who present with strokes have a past history of hypertension, and in those who are hypertensive, approximately 78% have not had their BP adequately controlled.¹²⁹ Of all the risk factors for stroke, hypertension has the highest relative risk (4.0 at 40–50 years, falling to 2.0 at 70–80 years of age), and the highest population attributable risk (40% at 40–50 years and 30% at 70–80 years of age). Also, stroke incidence is approximately three times greater in persons with stage 2 systolic hypertension (160 mm Hg or greater) and 50% higher in those with stage 1 systolic hypertension (140–159 mm Hg) compared to those with lower BP. In a recent report from the Northern Manhattan Study among 1750 participants aged at least 60 years, after adjustment for diastolic BP and other risk factors, even those with a systolic BP of 140 to 149 mm Hg had a 70% greater risk (HR, 1.7; 95% CI, 1.2–2.6) of stroke compared to those with a systolic BP of less than 140 mm Hg, suggesting a benefit for using the lower cutoff point.¹³⁰ The Prospective Studies Collaboration analyzed data from 45 prospective studies of 450,000 individuals and found a five-fold difference in stroke risk over a range of diastolic pressure from 75 to 102 mm Hg.¹³¹ Global risk algorithms for the prediction of stroke have been developed by the Framingham Heart Study and the Cardiovascular Heath Study¹³² and indicate systolic pressure to be the principal predictor, with other predictors being serum creatinine, diabetes, left ventricular hypertrophy by electrocardiogram, age, atrial fibrillation, and history of heart disease. Moreover, hypertension also indirectly raises risk of stroke through its role as an important risk factor for atrial fibrillation and left atrial enlargement, which both relate directly to stroke risk.^133,134

The different subtypes of stroke have somewhat different relationships with BP. A large proportion of cerebral infarcts are classified as lacunar infarcts, which are the result of lesions in small arteries penetrating the deeper layers of the cerebral cortex, and approximately 70% of patients who experience these are hypertensive.¹³⁵ The characteristic pathology is a destructive lesion of the vessel wall, termed lipohyalinosis, which is distinct from both atherosclerosis and the remodeling of arteries that occurs in other vascular beds. Infarcts of the large cranial and extracranial arteries are more related to atherosclerosis, and only 50% of patients with this type of stroke are hypertensive (despite being the most important risk factor).¹³⁶ Cardioembolic strokes are being increasingly recognized, and atheromatous plaques that protrude from the aorta are an independent risk factor for ischemic strokes.¹³⁷ Cerebral hemorrhage also shows a very strong relationship with hypertension and, particularly concerning, with discontinuation of antihypertensive medications.¹³⁷

The decline in the death rate from strokes since the 1960s has received a lot of publicity, and because BP is such an important risk factor for stroke, it is tempting to think that better hypertension control can take the credit. There are, however, reasons why this may not be the case. First, the decline of stroke mortality preceded the widespread use of antihypertensive medications.¹³⁸ Second, attempts to correlate the decline of mortality with improved BP control rates have been unsuccessful,¹³⁹ suggesting that changes in socioeconomic status might be more important. However, since the publication of the guideline on BP management from committee members of the JNC 8 recommending a higher cutoff point of less than 150/90 mm Hg for control of BP in persons aged 60 years and over, there is renewed concern that this may have a deleterious effect on overall BP control in the very population (older persons) at increased risk of stroke.¹⁴⁰

BP typically rises acutely after a stroke, and it is postulated that this helps to maintain cerebral perfusion in the infarct’s penumbra zone.¹⁴¹ This forms the rationale for avoiding excessive reduction of BP immediately after a stroke.

Treatment of hypertension reduces stroke rates by 35% to 44%¹⁴²; this has been shown in both younger patients with systolic and diastolic hypertension and in older patients with isolated systolic hypertension.¹⁰⁵ Recent data from the Hypertension in the Very Elderly Trial (HYVET) suggests that treatment of patients older than 85 years of age leads to a statistically significant 39% reduction in the risk of death from stroke. Although shy of statistical significance, the primary end point of fatal and nonfatal stroke was reduced by 30% in the active treatment group.⁵⁸ There is general agreement that the reduction of BP is more important than which drugs are used to lower it, but there are some minor differences. β-Blockers are less effective than other drugs for preventing stroke,¹⁴² and there is some evidence that angiotensin-receptor blockers may be more effective.¹⁴³

Chronic Kidney Disease

The Multiple Risk Factor Intervention Trial showed systolic BP of greater than 140 mm Hg to be associated with a five- to six-fold greater risk for end-stage renal disease (ESRD) compared to systolic BP of less than 117 mm Hg. Risk of developing ESRD is graded and continuous; even high normal BP has two-fold greater risk of developing ESRD compared to those with optimal BP.¹⁴⁴ A Japanese study of nearly 100,000 men and women found a progressive relationship between the height of the BP and the risk of developing ESRD during a 17-year follow-up period, such that there was an increased risk even in patients with high normal BP, in comparison with those whose pressure was optimal (< 120/80 mm Hg).¹⁴⁵ Hypertensive patients whose BP is not well controlled are more likely to show a deterioration of renal function.¹⁴⁶ In a large pooled cohort study of more than 500,000 individuals from the Asia-Pacific region followed for a median of 6.8 years, systolic BP was the strongest risk factor for renal death, with each standard deviation increase in systolic BP (19 mm Hg) associated with a more than 80% higher risk (HR, 1.84; 95% CI, 1.60–2.12).¹⁴⁷ A study of post–myocardial infarction patients found that a steep and linear relationship existed between glomerular filtration rate (GFR) and cardiovascular death.¹⁴⁸ In addition, hypertensive patients with mildly impaired renal function (estimated GFR < 60 mL/min) have an increased prevalence of target organ damage, such as left ventricular hypertrophy, increased carotid intima-media thickness, and microalbuminuria.¹⁴⁹ Moreover, compared to those with normal BP, masked hypertension as identified by ambulatory BP readings in the Chronic Renal Insufficiency Cohort was recently shown to be associated with reduced estimated GFR in those with chronic kidney disease.¹⁵⁰ In addition, higher BP variability has also been shown to be associated with poor renal outcomes as indicated by incident ESRD in persons with hypertension in the large Antihypertensive Lipid-Lowering to Prevent Heart Attack Trial (ALLHAT).¹⁵¹ The importance of these considerations to cardiologists is, first, that chronic kidney disease is an important consequence of hypertension and, second, it is associated with a marked increase in cardiovascular risk. In the case of patients on hemodialysis, the risk is 10 to 30 times higher than in the general population, and a task force of the American Heart Association has recommended that patients with chronic kidney disease should be considered as being in the highest risk group for CVD.¹⁵² There are two main effects of chronic kidney disease on the arteries: (1) increased prevalence of atherosclerosis and (2) remodeling of the arteries and increased stiffness, which has been related to increased mortality.¹⁰⁷

As highlighted by JNC 7,² the dramatic decline in death rates from CHD and stroke that we have witnessed in the past 30 years has not been paralleled by any decline of ESRD, which has been showing a relentless increase.² Many of these incident cases have been attributed to poorly controlled hypertension and the others mostly to diabetes. Renal disease and diabetes are the two conditions where JNC 7¹ and other guidelines, such as those from the National Kidney Foundation,¹⁵³ have proposed lower thresholds for treatment (130/80 vs 140/90 mm Hg) than in other hypertensive patients, so one would think that the relationship between BP and clinical outcomes should be particularly tight in these conditions.

Peripheral Arterial Disease

Hypertension is a major risk factor for peripheral vascular disease. This is important for two reasons: first, it causes debilitating symptoms, and second, it is a marker of high risk for cardiovascular events.¹⁵⁴ The accepted definition is an ankle brachial BP index (the ratio between the systolic pressure in the ankle and the brachial artery) of less than 0.9. This test was performed in the latest NHANES examination, and in adults older than 40 years of age, the prevalence was 4.3%.¹⁵⁵ It is strongly associated with the risk factors of atherosclerotic disease—hypertension, smoking, cholesterol, diabetes, and, most importantly, age. In a study of the Framingham Offspring, the strongest risk factor for peripheral vascular disease was age (odds ratio [OR], 2.6/10 y), followed by hypertension (OR 2.2), and smoking (OR 2.0).¹⁵⁶ Peripheral arterial disease (intermittent claudication) also varies three- to four-fold across systolic BP quintiles.¹⁷ Moreover, a recent report involving electronic health records from more than 4 million adults in the United Kingdom noted a 20 mm Hg higher than usual systolic BP to be associated with a 63% higher risk of peripheral arterial disease (HR, 1.63; 95% CI, 1.59–1.66)—more reliably than CHD.¹⁵⁷ The detection of peripheral vascular disease should prompt a search for atherosclerotic disease in other vascular beds, because 60% of patients with it have significant coronary disease, cerebrovascular disease, or both, whereas 40% of patients with coronary or cerebrovascular disease also have peripheral vascular disease.¹⁵⁸

Although current US guidelines suggest treatment for all persons younger than 80 years of age with a BP of at least 140/90 mm Hg, not all persons with hypertension have the same risk for CVD and therefore the appropriateness of a single target level of BP regardless of level of risk could be questioned. In fact, analysis of US adults with hypertension in NHANES using Framingham risk assessment showed that 24% of subjects were actually at low risk (< 10% 10-year risk of CVD), 21% at intermediate risk (10%–20%), 23% at high risk (> 20%), and 32% had prior CVD. These data show 55% of US adults with hypertension to be at high risk (> 20% 10-year risk of CVD or with preexisting CVD), but such an approach may underestimate persons at significant long-term risk. The addition of European algorithms that include those with metabolic syndrome, diabetes, or chronic kidney disease as high risk would raise this proportion to 80%.¹⁵⁹ Current European Society for Hypertension guidelines have such a risk stratification algorithm that identifies persons at successively lower levels of BP to be at high or very high risk according to the accompanying number of risk factors, diabetes, and/or presence of target organ damage (Table 23–4).¹⁶⁰

Awareness

The NHANES surveys have for decades provided estimates of the prevalence of hypertension and have asked participants if they have been diagnosed as having hypertension. The proportion of hypertensive subjects who are aware of their condition has improved since 1988–1991, when it was 69.2%,¹⁶¹ to 82.7% in 2009–2012 in adults aged 20 years and older,¹⁶² with both awareness and treatment greater in older people and in non-Hispanic blacks (87%) compared to Hispanics (77.7%). Awareness, treatment, and control of hypertension have also been shown to vary by region of the United States, being highest in the southeastern United States.¹⁶³

In countries in which health-care systems are less well developed, the level of awareness of hypertension is much lower. In China, 35.6% of city dwellers were aware of their condition in 1991, but only 13.9% of those living in rural areas were aware of their condition.⁶¹

Treatment and Control

The usual definition of BP control is a level less than 140/90 mm Hg. In 2003–2004, only 39.4% of persons with hypertension were under control whereas latest data from 2011–2012 show the control rate now at 51.8%, with a corresponding improvement in treatment rates from 65.0% to 74.5%. Importantly, among those taking prescription medication, BP control rates improved from 65.0% to 74.5% during this period. However, medication use among persons with hypertension is still highly variable, ranging from only 44.5% in those aged 18 to 39 years to 82.2% in those aged 60 years and older. Moreover, treatment rates among non-Hispanic black adults are highest (77.4%) compared to whites (76.7%), Hispanics (73.5%), and Asians (lowest, at 65.2%). Despite the relatively decent treatment rates among non-Hispanic blacks and Hispanics, data show they are 40% more likely to have uncontrolled BP than whites.

In the United States, approximately 52% of hypertensive persons are taking medications, whereas in the Europe the numbers are much lower: in Germany, where the prevalence of hypertension is very high (55%), only 26% of patients are taking antihypertensive medication.⁶⁰ In other countries, the rate is even lower. In China, in 1991, the control rate was 4% in urban areas and 1% in rural areas. In Egypt, the control rate is 8%.⁶¹ However, surveys based on health-care programs rather than the general population give a more optimistic picture. The 2003 Health Plan Employer Data and Information Set (HEDIS) report stated that BP control rates ranged from 58.6% for Medicaid to 62.2% for commercial plans, but studies based on ambulatory care practices have quoted figures in the range of 30% to 50%.⁵⁹ In clinical trials, control rates of up to 70% can be obtained¹⁶⁴; the crucial ingredients here are likely to be the use of rigid treatment protocols and the fact that the participants are likely to be compliant.

The public health implications of poor BP control are enormous. It has been estimated that controlling BP to the JNC target levels could prevent 19% of CHD events in men and 31% in women.¹⁶⁵ Importantly, in a recently published systematic review and meta-analysis, involving 123 clinical trials comprising 613,815 subjects, every 10 mm Hg reduction in systolic BP was noted to significantly reduce the risk of major cardiovascular events by 20%, CHD by 17%, and stroke by 27%, with an overall significant 13% reduction in all-cause mortality. However, there was no benefit seen with renal failure.¹⁶⁶

Although hypertension treatment has increased over the years, treatment still remains suboptimal, with control affected not only by age, severity of hypertension, and presence of comorbidities (eg, resistant patients are often older and have isolated systolic hypertension and CVD), but also by nonadherence, inaccurate measurements, and failure to recognize a white coat effect. Physicians’ therapeutic inertia, involving an insufficient or inappropriate combination of agents, is also an important factor.⁷

Treatment and control rates also vary according to risk group and comorbidity. We have previously demonstrated treatment for hypertension to be lowest (58%) in those categorized as low risk and highest (75%) in those classified as high risk according to Framingham global risk scores, but the reverse is seen with control of BP, being highest (80%) in those at low risk and lowest (< 50%) in high-risk persons.¹⁶⁷ Moreover, in persons with known CVD, control rates are poor with elevations in systolic BP being the main problem, and often being more than 20 mm Hg from goal. Higher risk persons or those with cardiovascular comorbidities are also typically older, and thus isolated systolic hypertension becomes the predominant issue and with increased arterial stiffness is more difficult to control, even with polypharmacy.

Barriers to adequate BP control are also related to other factors involving the patient, the health-care provider, and the health-care system.¹⁶⁸ Table 23–5 lists some of the more important reasons.

Typically, patients are blamed for the poor rates of BP control, on the grounds that they are not taking their prescribed medications. One survey found that 70% of physicians attributed treatment failures to poor compliance by their patients, whereas 81% of patients thought their compliance was good.¹⁶⁹ One recent report¹⁷⁰ notes dropout rates of up to 15% in the first year and 25% in the second year of treatment. By the end of 2 years, 25% to 33% of patients are no longer compliant with the treatment regimen. Thus this report indicates that compliance decreases with years of treatment; because control of hypertension requires a lifelong adherence to medication, this is clearly an important factor in morbidity and mortality caused by sustained BP elevations. It has also been shown that in patients receiving triple therapy, that use of a fixed-dose combination plus one pill results in better adherence and even lower risk of cardiovascular events compared to taking three separate pills.¹⁷¹

Another major factor is therapeutic inertia, a concept that was first put on the map by Berlowitz and coworkers¹⁷² who surveyed Veterans’ Affairs (VA) clinics and found that 40% of hypertensive patients had BPs in excess of 160/90 mm Hg, despite an average of six annual visits. Increases in medications were prescribed in only 7% of visits. This state of affairs is not likely to be a result of physicians’ ignorance about the goals of treatment. In a survey conducted at about the same time as the Veterans’ Affairs study, 97% of physicians knew that the goal BP was 140/90 mm Hg.¹⁷³ Knowledge does not necessarily dictate behavior, however. In another study of 21 primary care physicians in the United States,¹⁷⁴ who were also familiar with the guidelines, the systolic pressure of their hypertensive patients was above 140 mm Hg at 93% of visits. A more recent study¹⁷⁵ surveyed clinical records of 7253 hypertensive patients from family practices in the South and quantified therapeutic inertia as the proportion of visits during which the BP was more than 140 mm Hg systolic or 90 mm Hg diastolic, with no increase in medication. The main finding was a linear relationship between the therapeutic inertia scores and the degree of BP control. For the highest quintile of therapeutic inertia (ie, the least likely to increase antihypertensive medications), the control rate was 25% at the first visit and 6% at the last visit. In contrast, for the lowest therapeutic inertia quintile, the control rate increased from 53% to 75%. In addition, a recent clinical trial showed intervention consisting of an outreach coordinating raising patient and provider awareness of unmet BP goals, as well as arranging BP-focused primary care visits and treatment decision support to reduce clinical inertia, although this did not result in differences in BP changes compared to the usual care group.¹⁷⁶ Also, a large study of 28 US-based physician organizations showed care management processes involving physician education efforts to improve medication fill compliance, systematic processes for screening or assessing patients for hypertension, and maintaining a list of hypertensive patients with clinical data were shown to result in improvement in BP control.¹⁷⁷

In a recent review of BP control rates in the United States, achieved both in different health-care systems and in clinical trials, Krakoff¹⁷⁸ suggested four types of intervention that might improve national rates of control: (1) providing education and feedback to providers (which would include attention to therapeutic inertia), (2) making greater use of nonphysician providers, (3) wider use of electronic medical records, and (4) self-monitoring of BP. Probably the best example of a large health-care system’s implementation of a large-scale hypertension program derives from Kaiser Permanente Northern California. Using a comprehensive hypertension registry, developing and sharing performance metrics, using evidence-based guidelines, medical assistant visits for BP checks, and a single-pill combination therapy, hypertension control improved from 43.6% in 2001 to 80.4% in 2009, in comparison to the national average where control improved from 55.4% to only 64.1%.¹⁷⁹

We would like to thank Dr. Gbenga Ogedegbe and Dr. Thomas G. Pickering, who contributed to the previous version of this chapter in the 13th edition.