Abbreviations and acronyms
Appropriate Blood pressure Control in Diabetes
ambulatory blood pressure monitoring
Acute Candesartan Cilexetil Therapy in Stroke Survival
Avoiding Cardiovascular Events in Combination Therapy in Patients Living with Systolic Hypertension
Action to Control Cardiovascular Risk in Diabetes
- ACTIVE I
Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events
Action in Diabetes and Vascular Disease: Preterax and Diamicron-MR Controlled Evaluation
Action for HEAlth in Diabetes
Antihypertensive and Lipid-Lowering Treatment to Prevent Heart ATtack
ALiskiren Trial In Type 2 Diabetes Using Cardio-renal Endpoints
ANgioTensin II Antagonist In Paroxysmal Atrial Fibrillation
A Randomized Controlled Trial of Aliskiren in the Prevention of Major Cardiovascular Events in Elderly People
angiotensin receptor blocker
Atherosclerosis Risk In Communities
aldosterone renin ratio
Anglo-Scandinavian Cardiac Outcomes Trial
Anglo-Scandinavian Cardiac Outcomes Trial—Lipid Lowering Arm
Angioplasty and STenting for Renal Artery Lesions
body mass index
body surface area
coronary artery bypass graft
CAPtopril Prevention Project
CAndesartan in the Prevention of Relapsing Atrial Fibrillation
coronary heart disease
Controlling Hypertension and Hypertension Immediately Post-Stroke
chronic kidney disease
Chronic Kidney Disease—EPIdemiology collaboration
Controlled ONset Verapamil INvestigation of CV Endpoints
Dietary Approaches to Stop Hypertension
diastolic blood pressure
Diabetes Control and Complications Study
DIabetic REtinopathy Candesartan Trials
dipeptidyl peptidase 4
European Atherosclerosis Society
European Association for the Study of Diabetes
estimated glomerular filtration rate
European Lacidipine Study on Atherosclerosis
European Society of Cardiology
European Society of Hypertension
end-stage renal disease
Amlodipine–Valsartan Combination Decreases Central Systolic Blood Pressure more Effectively than the Amlodipine–Atenolol Combination
U.S. Food and Drug Administration
Felodipine EVent Reduction study
Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico-Atrial Fibrillation
home blood pressure monitoring
Heart Outcomes Prevention Evaluation
Hypertension Optimal Treatment
hormone replacement therapy
HYpertension in the Very Elderly Trial
Irbesartan in Heart Failure with Preserved Systolic Function
Effect of Potentially Modifiable Risk Factors associated with Myocardial Infarction in 52 Countries
INternational VErapamil SR/T Trandolapril
Isolated systolic hypertension
Joint National Committee
Justification for the Use of Statins in Primary Prevention: an Intervention Trial Evaluating Rosuvastatin
left atrial volume index
Losartan Intervention For Endpoint Reduction in Hypertensives
left ventricle/left ventricular
left ventricular hypertrophy
left ventricular mass
Modification of Diet in Renal Disease
Multiple Risk Factor Intervention Trial
magnetic resonance imaging
The Nordic Diltiazem Intervention study
ONgoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial
peripheral artery disease
Prevention And Treatment of Hypertension Study
percutaneous coronary intervention
peroxisome proliferator-activated receptor
Prevention of REnal and Vascular ENdstage Disease
Prevention Regimen for Effectively Avoiding Secondary Strokes
Perindopril Protection Against Recurrent Stroke Study
pulse wave velocity
Quality adjusted life years
randomized controlled trials
Randomized Olmesartan And Diabetes MicroAlbuminuria Prevention
systolic blood pressure
Angiotensin-Receptor Blocker Candesartan for Treatment of Acute STroke
Study on COgnition and Prognosis in the Elderly
Systematic COronary Risk Evaluation
Systolic Hypertension in the Elderly Program
Swedish Trials in Old Patients with Hypertension
The second Swedish Trial in Old Patients with Hypertension
SYSTolic Hypertension in the Elderly: Chinese trial
SYSTolic Hypertension in Europe
transient ischaemic attack
Trials Of Hypertension Prevention
Telmisartan Randomised AssessmeNt Study in ACE iNtolerant subjects with cardiovascular Disease
United Kingdom Prospective Diabetes Study
Veterans' Affairs Diabetes Trial
Valsartan Antihypertensive Long-term Use Evaluation
World Health Organization
The 2013 guidelines on hypertension of the European Society of Hypertension (ESH) and the European Society of Cardiology (ESC) follow the guidelines jointly issued by the two societies in 2003 and 2007.1,2 Publication of a new document 6 years after the previous one was felt to be timely because, over this period, important studies have been conducted and many new results have been published on both the diagnosis and treatment of individuals with an elevated blood pressure (BP), making refinements, modifications and expansion of the previous recommendations necessary.
The 2013 ESH/ESC guidelines continue to adhere to some fundamental principles that inspired the 2003 and 2007 guidelines, namely (i) to base recommendations on properly conducted studies identified from an extensive review of the literature, (ii) to consider, as the highest priority, data from randomized, controlled trials (RCTs) and their meta-analyses, but not to disregard—particularly when dealing with diagnostic aspects—the results of observational and other studies of appropriate scientific calibre, and (iii) to grade the level of scientific evidence and the strength of recommendations on major diagnostic and treatment issues as in European guidelines on other diseases, according to ESC recommendations (Tables 1 and 2). While it was not done in the 2003 and 2007 guidelines, providing the recommendation class and the level of evidence is now regarded as important for providing interested readers with a standard approach, by which to compare the state of knowledge across different fields of medicine. It was also thought that this could more effectively alert physicians on recommendations that are based on the opinions of the experts rather than on evidence. This is not uncommon in medicine because, for a great part of daily medical practice, no good science is available and recommendations must therefore stem from common sense and personal clinical experience, both of which can be fallible. When appropriately recognized, this can avoid guidelines being perceived as prescriptive and favour the performance of studies where opinion prevails and evidence is lacking. A fourth principle, in line with its educational purpose, is to provide a large number of tables and a set of concise recommendations that could be easily and rapidly consulted by physicians in their routine practice.
The European members of the Task Force in charge of the 2013 guidelines on hypertension have been appointed by the ESH and ESC, based on their recognized expertise and absence of major conflicts of interest [their declaration of interest forms can be found on the ESC website (www.escardio.org/guidelines) and ESH website (www.eshonline.org)]. Each member was assigned a specific writing task, which was reviewed by three co-ordinators and then by two chairmen, one appointed by ESH and another by ESC. The text was finalized over approximately 18 months, during which the Task Force members met collectively several times and corresponded intensively with one another between meetings. Before publication, the document was also assessed twice by 42 European reviewers, half selected by ESH and half by ESC. It can thus be confidently stated that the recommendations issued by the 2013 ESH/ESC guidelines on hypertension largely reflect the state of the art on hypertension, as viewed by scientists and physicians in Europe. Expenses for meetings and the remaining work have been shared by ESH and ESC.
Because of new evidence on several diagnostic and therapeutic aspects of hypertension, the present guidelines differ in many respects from the previous ones.2 Some of the most important differences are listed below:
Epidemiological data on hypertension and BP control in Europe.
Strengthening of the prognostic value of home blood pressure monitoring (HBPM) and of its role for diagnosis and management of hypertension, next to ambulatory blood pressure monitoring (ABPM).
Update of the prognostic significance of night-time BP, white-coat hypertension and masked hypertension.
Re-emphasis on integration of BP, cardiovascular (CV) risk factors, asymptomatic organ damage (OD) and clinical complications for total CV risk assessment.
Update of the prognostic significance of asymptomatic OD, including heart, blood vessels, kidney, eye and brain.
Reconsideration of the risk of overweight and target body mass index (BMI) in hypertension.
Hypertension in young people.
Initiation of antihypertensive treatment. More evidence-based criteria and no drug treatment of high normal BP.
Target BP for treatment. More evidence-based criteria and unified target systolic blood pressure (SBP) (<140 mmHg) in both higher and lower CV risk patients.
Liberal approach to initial monotherapy, without any all-ranking purpose.
Revised schema for priorital two-drug combinations.
New therapeutic algorithms for achieving target BP.
Extended section on therapeutic strategies in special conditions.
Revised recommendations on treatment of hypertension in the elderly.
Drug treatment of octogenarians.
Special attention to resistant hypertension and new treatment approaches.
Increased attention to OD-guided therapy.
New approaches to chronic management of hypertensive disease.
Relationship of blood pressure to cardiovascular and renal damage
The relationship between BP values and CV and renal morbid- and fatal events has been addressed in a large number of observational studies.3 The results, reported in detail in the 2003 and 2007 ESH/ESC guidelines,1,2 can be summarized as follows:
Office BP bears an independent continuous relationship with the incidence of several CV events [stroke, myocardial infarction, sudden death, heart failure and peripheral artery disease (PAD)] as well as of end-stage renal disease (ESRD).3–5 This is true at all ages and in all ethnic groups.6,7
The relationship with BP extends from high BP levels to relatively low values of 110–115 mmHg for SBP and 70–75 mmHg for diastolic BP (DBP). SBP appears to be a better predictor of events than DBP after the age of 50 years,8,9 and in elderly individuals pulse pressure (the difference between SBP and DBP values) has been reported to have a possible additional prognostic role.10 This is indicated also by the particularly high CV risk exhibited by patients with an elevated SBP and a normal or low DBP [isolated systolic hypertension (ISH)].11
A continuous relationship with events is also exhibited by out-of-office BP values, such as those obtained by ABPM and HBPM (see Section 3.1.2).
Definition and classification of hypertension
The continuous relationship between BP and CV and renal events makes the distinction between normotension and hypertension difficult when based on cut-off BP values. This is even more so because, in the general population, SBP and DBP values have a unimodal distribution.14 In practice, however, cut-off BP values are universally used, both to simplify the diagnostic approach and to facilitate the decision about treatment. The recommended classification is unchanged from the 2003 and 2007 ESH/ESC guidelines (Table 3). Hypertension is defined as values ≥140 mmHg SBP and/or ≥90 mmHg DBP, based on the evidence from RCTs that in patients with these BP values treatment-induced BP reductions are beneficial (see Sections 4.1 and 4.2). The same classification is used in young, middle-aged and elderly subjects, whereas different criteria, based on percentiles, are adopted in children and teenagers for whom data from interventional trials are not available. Details on BP classification in boys and girls according to their age and height can be found in the ESH's report on the diagnosis, evaluation and treatment of high BP in children and adolescents.15
aThe blood pressure (BP) category is defined by the highest level of BP, whether systolic or diastolic. Isolated systolic hypertension should be graded 1, 2, or 3 according to systolic BP values in the ranges indicated.
Prevalence of hypertension
Limited comparable data are available on the prevalence of hypertension and the temporal trends of BP values in different European countries.16 Overall the prevalence of hypertension appears to be around 30–45% of the general population, with a steep increase with ageing. There also appear to be noticeable differences in the average BP levels across countries, with no systematic trends towards BP changes in the past decade.17–37
Owing to the difficulty of obtaining comparable results among countries and over time, the use of a surrogate of hypertension status has been suggested.38 Stroke mortality is a good candidate, because hypertension is by far the most important cause of this event. A close relationship between prevalence of hypertension and mortality for stroke has been reported.39 The incidence and trends of stroke mortality in Europe have been analysed by use of World Health Organization (WHO) statistics. Western European countries exhibit a downward trend, in contrast to eastern European countries, which show a clear-cut increase in death rates from stroke.40
Hypertension and total cardiovascular risk
For a long time, hypertension guidelines focused on BP values as the only- or main variables determining the need for—and the type of—treatment. In 1994, the ESC, ESH and European Atherosclerosis Society (EAS) developed joint recommendations on prevention of coronary heart disease (CHD) in clinical practice,41 and emphasized that prevention of CHD should be related to quantification of total (or global) CV risk. This approach is now generally accepted and had already been integrated into the 2003 and 2007 ESH/ESC guidelines for the management of arterial hypertension.1,2 The concept is based on the fact that only a small fraction of the hypertensive population has an elevation of BP alone, with the majority exhibiting additional CV risk factors. Furthermore, when concomitantly present, BP and other CV risk factors may potentiate each other, leading to a total CV risk that is greater than the sum of its individual components. Finally, in high-risk individuals, antihypertensive treatment strategies (initiation and intensity of treatment, use of drug combinations, etc.: see Sections 4, 5, 6 and 7), as well as other treatments, may be different from those to be implemented in lower-risk individuals. There is evidence that, in high-risk individuals, BP control is more difficult and more frequently requires the combination of antihypertensive drugs with other therapies, such as aggressive lipid-lowering treatments. The therapeutic approach should consider total CV risk in addition to BP levels in order to maximize cost-effectiveness of the management of hypertension.
Assessment of total cardiovascular risk
Estimation of total CV risk is easy in particular subgroups of patients, such as those with antecedents of established cardiovascular disease (CVD), diabetes, CHD or with severely elevated single risk factors. In all of these conditions, the total CV risk is high or very high, calling for intensive CV risk-reducing measures. However, a large number of patients with hypertension do not belong to any of the above categories and the identification of those at low, moderate, high or very high risk requires the use of models to estimate total CV risk, so as to be able to adjust the therapeutic approach accordingly.
Several computerized methods have been developed for estimating total CV risk.41–48 Their values and limitations have been reviewed recently.49 The Systematic COronary Risk Evaluation (SCORE) model has been developed based on large European cohort studies. The model estimates the risk of dying from CV (not just coronary) disease over 10 years based on age, gender, smoking habits, total cholesterol and SBP.43 The SCORE model allows calibration of the charts for individual countries, which has been done for numerous European countries. At the international level, two sets of charts are provided: one for high-risk and one for low-risk countries. The electronic, interactive version of SCORE, known as HeartScore (available through www.heartscore.org), is adapted to also allow adjustment for the impact of high-density lipoprotein cholesterol on total CV risk.
The charts and their electronic versions can assist in risk assessment and management but must be interpreted in the light of the physician's knowledge and experience, especially with regard to local conditions. Furthermore, the implication that total CV risk estimation is associated with improved clinical outcomes when compared with other strategies has not been adequately tested.
Risk may be higher than indicated in the charts in: In SCORE, total CV risk is expressed as the absolute risk of dying from CVD within 10 years. Because of its heavy dependence on age, in young patients, absolute total CV risk can be low even in the presence of high BP with additional risk factors. If insufficiently treated, however, this condition may lead to a partly irreversible high-risk condition years later. In younger subjects, treatment decisions should better be guided by quantification of relative risk or by estimating heart and vascular age. A relative-risk chart is available in the Joint European Societies' Guidelines on CVD Prevention in Clinical Practice,50 which is helpful when advising young persons.
Sedentary subjects and those with central obesity; the increased relative risk associated with overweight is greater in younger subjects than in older subjects.
Socially deprived individuals and those from ethnic minorities.
Subjects with elevated fasting glucose and/or an abnormal glucose tolerance test, who do not meet the diagnostic criteria for diabetes.
Individuals with increased triglycerides, fibrinogen, apolipoprotein B, lipoprotein(a) levels and high-sensitivity C-reactive protein.
Individuals with a family history of premature CVD (before the age of 55 years in men and 65 years in women).
Further emphasis has been given to identification of asymptomatic OD, since hypertension-related asymptomatic alterations in several organs indicate progression in the CVD continuum, which markedly increases the risk beyond that caused by the simple presence of risk factors. A separate section (Section 3.7) is devoted to searching for asymptomatic OD,51–53 where evidence for the additional risk of each subclinical alteration is discussed.
For more than a decade, international guidelines for the management of hypertension (the 1999 and 2003 WHO/ International Society of Hypertension Guidelines and the 2003 and 2007 ESH/ESC Guidelines)1,2,54,55 have stratified CV risk in different categories, based on BP category, CV risk factors, asymptomatic OD and presence of diabetes, symptomatic CVD or chronic kidney disease (CKD), as also done by the 2012 ESC prevention guidelines.50 The classification in low, moderate, high and very high risk is retained in the current guidelines and refers to the 10-year risk of CV mortality as defined by the 2012 ESC prevention guidelines (Figure 1).50 The factors on which the stratification is based are summarized in Table 4.
BMI = body mass index; BP = blood pressure; BSA = body surface area; CABG = coronary artery bypass graft; CHD = coronary heart disease; CKD = chronic kidney disease; CV = cardiovascular; CVD = cardiovascular disease; EF = ejection fraction; eGFR = estimated glomerular filtration rate; HbA1c = glycated haemoglobin; IMT = intima-media thickness; LVH = left ventricular hypertrophy; LVM = left ventricular mass; PCI = percutaneous coronary intervention; PWV = pulse wave velocity.
aRisk maximal for concentric LVH: increased LVM index with a wall thickness/radius ratio of >0.42.
All currently available models for CV risk assessment have limitations that must be appreciated. The significance of OD in determining calculation of overall risk is dependent on how carefully the damage is assessed, based on available facilities. Conceptual limitations should also be mentioned. One should never forget that the rationale of estimating total CV risk is to govern the best use of limited resources to prevent CVD; that is, to grade preventive measures in relation to the increased risk. Yet, stratification of absolute risk is often used by private or public healthcare providers to establish a barrier, below which treatment is discouraged. It should be kept in mind that any threshold used to define high total CV risk is arbitrary, as well as the use of a cut-off value leading to intensive interventions above this threshold and no action at all below. Finally, there is a strong effect of age on total CV risk models. It is so strong that younger adults (particularly women) are unlikely to reach high-risk levels even when they have more than one major risk factor and a clear increase in relative risk. By contrast, many elderly men (e.g. >70 years) reach a high total risk level whilst being at very little increased risk relative to their peers. The consequences are that most resources are concentrated in older subjects, whose potential lifespan is relatively short despite intervention, and little attention is given to young subjects at high relative risk despite the fact that, in the absence of intervention, their long-term exposure to an increased risk may lead to a high and partly irreversible risk situation in middle age, with potential shortening of their otherwise longer life expectancy.
Summary of recommendations on total cardiovascular risk assessment
CKD = chronic kidney disease; CV = cardiovascular; CVD = cardiovascular disease; OD = organ damage; SCORE = Systematic COronary Risk Evaluation
aClass of recommendation.
bLevel of evidence.
cReference(s) supporting recommendation(s).
The initial evaluation of a patient with hypertension should (i) confirm the diagnosis of hypertension, (ii) detect causes of secondary hypertension, and (iii) assess CV risk, OD and concomitant clinical conditions. This calls for BP measurement, medical history including family history, physical examination, laboratory investigations and further diagnostic tests. Some of the investigations are needed in all patients; others only in specific patient groups.
Bood pressure measurement
Office or clinic blood pressure
At present, BP can no longer be estimated using a mercury sphygmomanometer in many—although not all—European countries. Auscultatory or oscillometric semiautomatic sphygmomanometers are used instead. These devices should be validated according to standardized protocols and their accuracy should be checked periodically through calibration in a technical laboratory.56 Measurement of BP at the upper arm is preferred and cuff and bladder dimensions should be adapted to the arm circumference. In the event of a significant (>10 mmHg) and consistent SBP difference between arms, which has been shown to carry an increased CV risk,57 the arm with the higher BP values should be used. A between-arms difference is meaningful if demonstrated by simultaneous arm measurement; if one gets a difference between arms with sequential measurement, it could be due to BP variability. In elderly subjects, diabetic patients and in other conditions in which orthostatic hypotension may be frequent or suspected, it is recommended that BP be measured 1 min and 3 min after assumption of the standing position. Orthostatic hypotension—defined as a reduction in SBP of ≥20 mmHg or in DBP of ≥10 mmHg within 3 min of standing—has been shown to carry a worse prognosis for mortality and CV events.58,59 If feasible, automated recording of multiple BP readings in the office with the patient seated in an isolated room, though providing less information overall, might be considered as a means to improve reproducibility and make office BP values closer to those provided by daytime ABPM or HBPM,60,61. BP measurements should always be associated with measurement of heart rate, because resting heart rate values independently predict CV morbid or fatal events in several conditions, including hypertension.62,63 Instructions for correct office BP measurements are summarized in Table 5.
BP = blood pressure.
Out-of-office blood pressure
The major advantage of out-of-office BP monitoring is that it provides a large number of BP measurements away from the medical environment, which represents a more reliable assessment of actual BP than office BP. Out-of-office BP is commonly assessed by ABPM or HBPM, usually by self-measurement. A few general principles and remarks hold for the two types of monitoring, in addition to recommendations for office BP measurement:64–67
The procedure should be adequately explained to the patient, with verbal and written instructions; in addition, self-measurement of BP requires appropriate training under medical supervision.
Interpretation of the results should take into account that the reproducibility of out-of-office BP measurements is reasonably good for 24-h, day and night BP averages but less for shorter periods within the 24 hs and for more complex and derived indices.68
ABPM and HBPM provide somewhat different information on the subject's BP status and risk and the two methods should thus be regarded as complementary, rather than competitive or alternative. The correspondence between measurements with ABPM and HBPM is fair to moderate.
Office BP is usually higher than ambulatory and home BP and the difference increases as office BP increases. Cut-off values for the definition of hypertension for home and ambulatory BP, according to the ESH Working Group on BP Monitoring, are reported in Table 6.64–67
Devices should have been evaluated and validated according to international standardized protocols and should be properly maintained and regularly calibrated; at least every 6 months. The validation status can be obtained on dedicated websites.
BP = blood pressure.
188.8.131.52 Ambulatory blood pressure monitoring
184.108.40.206.1 Methodological aspects
A number of methodological aspects have been addressed by the ESH Working Group on Blood Pressure Monitoring.64,65 ABPM is performed with the patient wearing a portable BP measuring device, usually on the non-dominant arm, for a 24–25 h period, so that it gives information on BP during daily activities and at night during sleep. At the time of fitting of the portable device, the difference between the initial values and those from BP measurement by the operator should not be greater than 5 mmHg. In the event of a larger difference, the ABPM cuff should be removed and fitted again. The patient is instructed to engage in normal activities but to refrain from strenuous exercise and, at the time of cuff inflation, to stop moving and talking and keep the arm still with the cuff at heart level. The patient is asked to provide information in a diary on symptoms and events that may influence BP, in addition to the times of drug ingestion, meals and going to- and rising from bed. In clinical practice, measurements are often made at 15 min intervals during the day and every 30 min overnight; excessive intervals between BP readings should be avoided because they reduce the accuracy of 24-h BP estimates.69 It may be recommended that measurements be made at the same frequency during the day and night—for example every 20 min throughout. The measurements are downloaded to a computer and a range of analyses can be performed. At least 70% of BPs during daytime and night-time periods should be satisfactory, or else the monitoring should be repeated. The detection of artifactual readings and the handling of outlying values have been subject to debate but, if there are sufficient measurements, editing is not considered necessary and only grossly incorrect readings should be deleted. It is noteworthy that readings may not be accurate when the cardiac rhythm is markedly irregular.70
220.127.116.11.2 Daytime, night-time and 24-hour blood pressure
In addition to the visual plot, average daytime, night-time and 24-h BP are the most commonly used variables in clinical practice. Average daytime and night-time BP can be calculated from the diary on the basis of the times of getting up and going to bed. An alternative method is to use short, fixed time periods, in which the rising and retiring periods—which differ from patient to patient—are eliminated. It has, for example, been shown that average BPs from 10 am to 8 pm and from midnight to 6 am correspond well with the actual waking and sleeping BPs,71 but other short, fixed time periods have been proposed, such as from 9 am to 9 pm and from 1 am to 6 am. In the event of different measurement intervals during the day and the night, and to account for missing values, it is recommended that average 24-h BP be weighted for the intervals between successive readings or to calculate the mean of the 24 hourly averages to avoid overestimation of average 24-h BP.72
The night-to-day BP ratio represents the ratio between average night-time and daytime BP. BP normally decreases during the night—defined as ‘dipping’. Although the degree of night-time dipping has a normal distribution in a population setting, it is generally agreed that the finding of a nocturnal BP fall of >10% of daytime values (night–day BP ratio <0.9) will be accepted as an arbitrary cut-off to define subjects as ‘dippers’. Recently, more dipping categories have been proposed: absence of dipping, i.e. nocturnal BP increase (ratio >1.0); mild dipping (0.9 <ratio ≤1.0); dipping (0.8 <ratio ≤0.9); and extreme dipping (ratio ≤0.8). One should bear in mind that the reproducibility of the dipping pattern is limited.73,74 Possible reasons for absence of dipping are sleep disturbance, obstructive sleep apnoea, obesity, high salt intake in salt-sensitive subjects, orthostatic hypotension, autonomic dysfunction, chronic kidney disease (CKD), diabetic neuropathy and old age.
18.104.22.168.3 Additional analyses
A number of additional indices may be derived from ABPM recordings.75–81 They include: BP variability,75 morning BP surge,76,77,81 blood pressure load,78 and the ambulatory arterial stiffness index.79,80 However, their added predictive value is not yet clear and they should thus be regarded as experimental, with no routine clinical use. Several of these indices are discussed in detail in ESH position papers and guidelines,64,65 including information on facilities recommended for ABPM software in clinical practice, which include the need for a standardized clinical report, an interpretative report, a trend report to compare recordings obtained over time and a research report, offering a series of additional parameters such as those listed above.
22.214.171.124.4 Prognostic significance of ambulatory blood pressure
Several studies have shown that hypertensive patients' left ventricular hypertrophy (LVH), increased carotid intima-media thickness (IMT) and other markers of OD correlate with ambulatory BP more closely than with office BP.82,83 Furthermore, 24-h average BP has been consistently shown to have a stronger relationship with morbid or fatal events than office BP.84–87 There are studies in which accurately measured office BP had a predictive value similar to ambulatory BP.87 Evidence from meta-analyses of published observational studies and pooled individual data,88–90 however, has shown that ambulatory BP in general is a more sensitive risk predictor of clinical CV outcomes, such as coronary morbid or fatal events and stroke, than office BP. The superiority of ambulatory BP has been shown in the general population, in young and old, in men and women, in untreated and treated hypertensive patients, in patients at high risk and in patients with CV or renal disease.89–93 Studies that accounted for daytime and night-time BP in the same statistical model found that night-time BP is a stronger predictor than daytime BP.90,94 The night–day ratio is a significant predictor of clinical CV outcomes but adds little prognostic information over and above 24-h BP.94,95 With regard to the dipping pattern, the most consistent finding is that the incidence of CV events is higher in patients with a lesser or no drop in nocturnal BP than in those with greater drop,89,91,92,95,96 although the limited reproducibility of this phenomenon limits the reliability of the results for small between-group differences.89,91,92,95 Extreme dippers may have an increased risk for stroke.97 However, data on the increased CV risk in extreme dippers are inconsistent and thus the clinical significance of this phenomenon is uncertain.89,95
126.96.36.199 Home blood pressure monitoring
188.8.131.52.1 Methodological aspects
The ESH Working Group on Blood Pressure Monitoring has proposed a number of recommendations for HBPM.66,67 The technique usually involves self-measurement of BP but, in some patients, the support of a trained health provider or family member may be needed. Devices worn on the wrist are currently not recommended but their use might be justified in obese subjects with extremely large arm circumference. For diagnostic evaluation, BP should be measured daily on at least 3–4 days and preferably on 7 consecutive days; in the mornings as well as in the evenings. BP is measured in a quiet room, with the patient in the seated position, back and arm supported, after 5 min of rest and with two measurements per occasion taken 1–2 min apart: the results are reported in a standardized logbook immediately after each measurement. However, BP values reported by the patient may not always be reliable, which can be overcome by storage in a memory-equipped device. Home BP is the average of these readings, with exclusion of the first monitoring day. Use of telemonitoring and smartphone applications for HBPM may be of further advantage.98,99 Interpretation of the results should always be under the close guidance of the physician.
When compared with office BP, HBPM yields multiple measurements over several days, or even longer periods, taken in the individual's usual environment. Compared with ambulatory BP, it provides measurements over extended periods and day-to-day BP variability, is cheaper,100 more widely available and more easily repeatable. However, unlike ABPM, it does not provide BP data during routine, day-to-day activities and during sleep, or the quantification of short-term BP variability.101
184.108.40.206.2 Prognostic significance of home BP
Home BP is more closely related to hypertension-induced OD than office BP, particularly LVH,82,83 and recent meta-analyses of the few prospective studies in the general population, in primary care and in hypertensive patients, indicate that the prediction of CV morbidity and mortality is significantly better with home BP than with office BP.102,103 Studies in which both ABPM and HBPM were performed show that home BP is at least as well correlated with OD as is the ambulatory BP,82,83 and that the prognostic significance of home BP is similar to that of ambulatory BP after adjustment for age and gender.104,105
White-coat (or isolated office) hypertension and masked (or isolated ambulatory) hypertension
Office BP is usually higher than BP measured out of the office, which has been ascribed to the alerting response, anxiety and/or a conditional response to the unusual situation,106 and in which regression to the mean may play a role. Although several factors involved in office or out-of-office BP modulation may be involved,107 the difference between the two is usually referred to—although somewhat improperly—as the ‘white-coat effect’,107,108 whereas ‘white-coat-’ or ‘isolated office-’ or ‘isolated clinic hypertension’ refers to the condition in which BP is elevated in the office at repeated visits and normal out of the office, either on ABPM or HBPM. Conversely, BP may be normal in the office and abnormally high out of the medical environment, which is termed ‘masked-’ or ‘isolated ambulatory hypertension’. The terms ‘true-’ or ‘consistent normotension’ and ‘sustained hypertension’ are used when both types of BP measurement are, respectively, normal or abnormal. Whereas the cut-off value for office BP is the conventional 140/90 mmHg, most studies in white-coat or masked hypertension have used a cut-off value of 135/85 mmHg for out-of-office daytime or home BP and 130/80 mmHg for 24-h BP. Notably, there is only moderate agreement between the definition of white-coat or masked hypertension diagnosed by ABPM or HBPM.101 It is recommended that the terms ‘white-coat hypertension’ and ‘masked hypertension’ be reserved to define untreated individuals.
220.127.116.11 White-coat hypertension
Based on four population studies, the overall prevalence of white-coat hypertension averaged 13% (range 9–16%) and it amounted to about 32% (range 25–46%) among hypertensive subjects in these surveys.109 Factors related to increased prevalence of white-coat hypertension are: age, female sex and non-smoking. Prevalence is lower in the case of target OD or when office BP is based on repeated measurements or when measured by a nurse or another healthcare provider.110,111 The prevalence is also related to the level of office BP: for example, the percentage of white-coat hypertension amounts to about 55% in grade 1 hypertension and to only about 10% in grade 3 hypertension.110 OD is less prevalent in white-coat hypertension than in sustained hypertension and prospective studies have consistently shown this to be the case also for CV events.105,109,112,113 Whether subjects with white-coat hypertension can be equalled to true normotensive individuals is an issue still under debate because, in some studies, the long-term CV risk of this condition was found to be intermediate between sustained hypertension and true normotension,105 whereas in meta-analyses it was not significantly different from true normotension when adjusted for age, gender and other covariates.109,112,113 The possibility exists that, because white-coat hypertensive patients are frequently treated, the reduction of clinic BP leads to a reduced incidence of CV events.112 Other factors to consider are that, compared with true normotensive subjects, in white-coat hypertensive patients, (i) out-of-office BP is higher,105,109 (ii) asymptomatic OD such as LVH may be more frequent,114 and (iii) this is the case also for metabolic risk factors and long-term risk of new-onset diabetes and progression to sustained hypertension.115,116 It is recommended that the diagnosis of white-coat hypertension be confirmed within 3–6 months and these patients be investigated and followed-up closely, including repeated out-of-office BP measurements (see Section 6.1).
18.104.22.168 Masked hypertension
The prevalence of masked hypertension averages about 13% (range 10–17%) in population-based studies.109 Several factors may raise out-of-office BP relative to office BP, such as younger age, male gender, smoking, alcohol consumption, physical activity, exercise-induced hypertension, anxiety, job stress, obesity, diabetes, CKD and family history of hypertension and the prevalence is higher when office BP is in the high normal range.117 Masked hypertension is frequently associated with other risk factors, asymptomatic OD and increased risk of diabetes and sustained hypertension.114–119 Meta-analyses of prospective studies indicate that the incidence of CV events is about two times higher than in true normotension and is similar to the incidence in sustained hypertension.109,112,117 The fact that masked hypertension is largely undetected and untreated may have contributed to this finding. In diabetic patients masked hypertension is associated with an increased risk of nephropathy, especially when the BP elevation occurs mainly during the night.120,121
3.1.4 Clinical indications for out-of-office blood pressure
It is now generally accepted that out-of-office BP is an important adjunct to conventional office BP measurement, but the latter currently remains the ‘gold standard’ for screening, diagnosis and management of hypertension. The time-honoured value of office BP, however, has to be balanced against its important limitations, which have led to the increasingly frequent suggestion that out-of-office BP measurements play an important role in hypertension management. Although there are important differences between ABPM and HBPM, the choice between the two methods will depend on indication, availability, ease, cost of use and, if appropriate, patient preference. For initial assessment of the patient, HBPM may be more suitable in primary care and ABPM in specialist care. However, it is advisable to confirm borderline or abnormal findings on HBPM with ABPM,122 which is currently considered the reference for out-of-office BP, with the additional advantage of providing night-time BP values. Furthermore, most—if not all—patients should be familiarized with self-measurement of BP in order to optimize follow-up, for which HBPM is more suitable than ABPM. However, (self-measured) HBPM may not be feasible because of cognitive impairment or physical limitations, or may be contra-indicated because of anxiety or obsessive patient behaviour, in which case ABPM may be more suitable. Conditions considered as clinical indications for out-of-office BP measurement for diagnostic purposes are listed in Table 7.
ABPM = ambulatory blood pressure monitoring; BP = blood pressure; CKD = chronic kidney disease; CV = cardiovascular; HBPM = home blood pressure monitoring.
3.1.5 Blood pressure during exercise and laboratory stress
BP increases during dynamic and static exercise, whereby the increase is more pronounced for systolic than for diastolic BP.123 Exercise testing usually involves dynamic exercise, either on a bicycle ergometer or a treadmill. Notably, only SBP can be measured reliably with non-invasive methods. There is currently no consensus on normal BP response during dynamic exercise testing. A SBP of ≥210 mmHg for men and ≥190 mmHg for women has been termed ‘exercise hypertension’ in a number of studies, but other definitions of an exaggerated BP response to exercise have also been used.124,125 Furthermore, the increase of SBP at fixed submaximal exercise is related to pre-exercise BP, age, arterial stiffness and abdominal obesity and is somewhat greater in women than in men and less in fit than in unfit individuals.123–127 Most—but not all—studies have shown that an excessive rise of BP during exercise predicts the development of hypertension in normotensive subjects, independently from BP at rest.123,124,128 However, exercise testing to predict future hypertension is not recommended because of a number of limitations, such as lack of standardization of methodology and definitions. Furthermore, there is no unanimity on the association of exercise BP with OD, such as LVH, after adjustment for resting BP and other covariates, as well in normotensives as in hypertensive patients.123,124 Also the results on the prognostic significance of exercise BP are not consistent,125 which may be due to the fact that the two haemodynamic components of BP change in opposite directions during dynamic exercise: systemic vascular resistance decreases whereas cardiac output increases. It is likely that the decisive prognostic factor is a blunted reduction of systemic vascular resistance during exercise, compatible with structural pathophysiological changes in arteries and arterioles.123,129 Whether or not the impaired arterial dilatation is translated into an excessive rise of BP may at least partly depend on cardiac output. In normotensive subjects and in mild hypertensive patients with adequate increase of cardiac output, an exaggerated BP response predicts a poorer long-term outcome.125,130 In the case of normal resting BP, exercise-induced hypertension can be considered an indication for ABPM because of its association with masked hypertension.131 On the other hand, when hypertension is associated with cardiac dysfunction and blunted exercise-induced increase of cardiac output, the prognostic significance of exercise BP may be lost.129 Finally, a higher BP during exercise may even carry a better prognosis, such as in 75-year-old individuals,132 in patients with suspected cardiac disease,133 or with heart failure,134 in whom a higher exercise BP implies relatively preserved systolic cardiac function.125 In conclusion, the overall results question the clinical utility of BP measurements during exercise testing for diagnostic and prognostic purposes in patients with hypertension. However, exercise testing is useful as a general prognostic indicator using exercise capacity and electrocardiogram (ECG) data and an abnormal BP response may warrant ABPM.
A number of mental stress tests have been applied to evoke stress and increase BP via a problem of mathematical, technical, or decisional nature.123 However, these laboratory stress tests in general do not reflect real-life stress and are not well standardized, have limited reproducibility, and correlations between BP responses to the various stressors are limited. In addition, results on the independent relationships of the BP response to mental stressors with future hypertension are not unanimous and, if significant, the additional explained variance is usually small.123,135 A recent meta-analysis suggested that greater responsiveness to acute mental stress has an adverse effect on future CV risk status—a composite of elevated BP, hypertension, left ventricular mass (LVM), subclinical atherosclerosis and clinical cardiac events.136 The overall results suggest that BP measurements during mental stress tests are currently not clinically useful.
Central blood pressure
The measurement of central BP in hypertensive patients raises increasing interest because of both its predictive value for CV events and the differential effect of antihypertensive drugs, compared with brachial BP. The arterial pressure waveform is a composite of the forward pressure wave created by ventricular contraction and a reflected wave.137 It should be analysed at the central level, i.e. in the ascending aorta, since it represents the true load imposed on heart, brain, kidney and large arteries. The phenomenon of wave reflection can be quantified through the augmentation index—defined as the difference between the second and first systolic peaks, expressed as a percentage of the pulse pressure, preferably adjusted for heart rate. Owing to the variable superimposition of incoming and reflected pressure waves along the arterial tree, aortic systolic and pulse pressures may be different from the conventionally measured brachial pressure. In recent years several methods, including applanation tonometry and transfer function, have been developed to estimate central systolic BP or pulse pressure from brachial pressure wave. They have been critically reviewed in an expert consensus document.138
Early epidemiological studies in the 2000s showed that central augmentation index and pulse pressure, directly measured by carotid tonometry, were independent predictors of all-cause and CV mortality in patients with ESRD.139 A recent meta-analysis confirmed these findings in several populations.140 However, the additive predictive value of central BP beyond brachial BP was either marginal or not statistically significant in most studies.140
Thus the current guidelines, like previous ones,2,141 consider that, although the measurement of central BP and augmentation index is of great interest for mechanistic analyses in pathophysiology, pharmacology and therapeutics, more investigation is needed before recommending their routine clinical use. The only exception may be ISH in the young: in some of these individuals increased SBP at the brachial level may be due to high amplification of the central pressure wave, while central BP is normal.142
The medical history should address the time of the first diagnosis of arterial hypertension, current and past BP measurements and current and past antihypertensive medications. Particular attention should be paid to indications of secondary causes of hypertension. Women should be questioned about pregnancy-related hypertension. Hypertension translates into an increased risk of renal and CV complications (CHD; heart failure; stroke; PAD; CV death), especially when concomitant diseases are present. Therefore, a careful history of CVDs should be taken in all patients, to allow assessment of global CV risk, including concomitant diseases such as diabetes, clinical signs or a history of heart failure, CHD or PAD, valvular heart disease, palpitations, syncopal episodes, neurological disorders with an emphasis on stroke and transient ischaemic attack (TIA). A history of CKD should include the type and duration of kidney disease. Nicotine abuse and evidence for dyslipidaemia should be sought. A family history of premature hypertension and/or premature CVD is an important first indicator of familial (genetic) predisposition to hypertension and CVD and may trigger clinically indicated genetic tests. Details on family and medical history are summarized in Table 8.
BP = blood pressure; CKD = chronic kidney disease; CVD = cardiovascular disease; TIA = transient ischaemic attack.
Physical examination aims to establish or verify the diagnosis of hypertension, establish current BP, screen for secondary causes of hypertension and refine global CV risk estimation. BP should be measured as summarized in Section 3.1.1 and should be repeated to confirm the diagnosis of hypertension. On at least one occasion, BP needs to be measured at both arms and differences between the two arms in SBP >20 mmHg and/or in DBP >10 mmHg—if confirmed—should trigger further investigations of vascular abnormalities. All patients should undergo auscultation of the carotid arteries, heart and renal arteries. Murmurs should suggest further investigation (carotid ultrasound, echocardiography, renal vascular ultrasound, depending on the location of the murmur). Height, weight, and waist circumference should be measured with the patient standing, and BMI calculated. Pulse palpation and cardiac auscultation may reveal arrhythmias. In all patients, heart rate should be measured while the patient is at rest. An increased heart rate indicates an increased risk of heart disease. An irregular pulse should raise the suspicion of atrial fibrillation, including silent atrial fibrillation. Details on physical examination are summarized in Table 9.
BP = blood pressure; BMI = body mass index.
Summary of recommendations on blood pressure measurement, history, and physical examination
ABPM = ambulatory blood pressure monitoring; BP = blood pressure; CV = cardiovascular; CVD = cardiovascular disease; HBPM = home blood pressure monitoring; OD = organ damage.
aClass of recommendation.
bLevel of evidence.
cReference(s) supporting recommendation(s).
Laboratory investigations are directed at providing evidence for the presence of additional risk factors, searching for secondary hypertension and looking for the absence or presence of OD. Investigations should progress from the most simple to the more complicated ones. Details on laboratory investigations are summarized in Table 10.
BP = blood pressure; ECG = electrocardiogram; GFR = glomerular filtration rate.
A positive family history is a frequent feature in hypertensive patients,143,144 with the heritability estimated to vary between 35% and 50% in the majority of studies,145 and heritability has been confirmed for ambulatory BP.146 Several rare, monogenic forms of hypertension have been described, such as glucocorticoid-remediable aldosteronism, Liddle's syndrome and others, where a single gene mutation fully explains the pathogenesis of hypertension and dictates the best treatment modality.147 Essential hypertension is a highly heterogeneous disorder with a multifactorial aetiology. Several genome-wide association studies and their meta-analyses point to a total of 29 single nucleotide polymorphisms, which are associated with systolic and/or diastolic BP.148 These findings might become useful contributors to risk scores for OD.
Searching for asymptomatic organ damage
Owing to the importance of asymptomatic OD as an intermediate stage in the continuum of vascular disease, and as a determinant of overall CV risk, signs of organ involvement should be sought carefully by appropriate techniques if indicated (Table 10). It should be pointed out that a large body of evidence is now available on the crucial role of asymptomatic OD in determining the CV risk of individuals with and without high BP. The observation that any of four markers of OD (microalbuminuria, increased pulse wave velocity [PWV], left ventricular hypertrophy [LVH] and carotid plaques) can predict CV mortality independently of SCORE stratification is a relevant argument in favour of using assessment of OD in daily clinical practice,51–53 although more data from larger studies in different populations would be desirable. It is also noteworthy that the risk increases as the number of damaged organs increases.51
A 12-lead ECG should be part of the routine assessment of all hypertensive patients. Its sensitivity in detecting LVH is low but, nonetheless, LVH detected by the Sokolow-Lyon index (SV1 + RV5 >3.5 mV), the modified Sokolow-Lyon index (largest S-wave + largest R-wave >3.5 mV), RaVL >1.1 mV, or Cornell voltage QRS duration product (>244 mV*ms) has been found in observational studies and clinical trials to be an independent predictor of CV events.149 Accordingly, the ECG is valuable, at least in patients over 55 years of age.150,151 Electrocardiography can also be used to detect patterns of ventricular overload or ‘strain’, which indicates more severe risk,149,150,152 ischaemia, conduction abnormalities, left atrial dilatation and arrhythmias, including atrial fibrillation. Twenty-four-hour Holter electrocardiography is indicated when arrhythmias and possible ischaemic episodes are suspected. Atrial fibrillation is a very frequent and common cause of CV complications,153,154 especially stroke, in hypertensive patients.153 Early detection of atrial fibrillation would facilitate the prevention of strokes by initiating appropriate anticoagulant therapy if indicated.
Although not immune from technical limitations, echocardiography is more sensitive than electrocardiography in diagnosing LVH and is useful to refine CV and renal risk.155-157 It may therefore help in a more precise stratification of overall risk and in determining therapy.158 Proper evaluation of the left ventricle (LV) in hypertensive patients includes linear measurements of interventricular septal and posterior wall thickness and internal end-diastolic diameter. While LVM measurements indexed for body size identify LVH, relative wall thickness or the wall-to-radius ratio (2 × posterior wall thickness/end-diastolic diameter) categorizes geometry (concentric or eccentric). Calculation of LVM is currently performed according to the American Society of Echocardiography formula.159 Although the relation between LVM and CV risk is continuous, thresholds of 95 g/m2 for women and 115 g/m2 (body surface area [BSA]) for men are widely used for estimates of clear-cut LVH.159 Indexation of LVM for height, in which height to the allometric power of 1.7 or 2.7 has been used,160,161 can be considered in overweight and obese patients in order to scale LVM to body size and avoid under-diagnosis of LVH.159 It has recently been shown that the optimal method is to scale allometrically by body height to the exponent 1.7 (g/m1.7) and that different cut-offs for men and women should be used.160 Scaling LVM by height exponent 2.7 could overestimate LVH in small subjects and underestimate in tall ones.160 Concentric LVH (relative wall thickness >0.42 with increased LVM), eccentric LVH (relative wall thickness ≤0.42 with increased LVM) and concentric remodelling (relative wall thickness >0.42 with normal LVM) all predict an increased incidence of CVD, but concentric LVH is the strongest predictor of increased risk.162–164
Hypertension is associated with alterations of LV relaxation and filling, globally defined as diastolic dysfunction. Hypertension-induced diastolic dysfunction is associated with concentric geometry and can per se induce symptoms/signs of heart failure, even when ejection fraction (EF) is still normal (heart failure with preserved EF).165 The Doppler transmitral inflow pattern can quantify filling abnormalities and predict subsequent heart failure and all-cause mortality,166,167 but is not sufficient to completely stratify the hypertensive clinical status and prognosis.166,167 According to recent echocardiographical recommendations,168 it should therefore be combined with pulsed Tissue Doppler of the mitral annulus. Reduction of the Tissue Doppler-derived early diastolic velocity (e') is typical of hypertensive heart disease and, often, the septal e' is reduced more than the lateral e'. Diagnosis and grading of diastolic dysfunction is based on e' (average of septal and lateral mitral annulus) and additional measurements including the ratio between transmitral E and e' (E/e' ratio) and left atrial size.168 This grading is an important predictor of all-cause mortality in a large epidemiological study.169 The values of e' velocity and of E/e' ratio are highly dependent on age and somewhat less on gender.170 The E/e' ratio is able to detect an increase of LV filling pressures. The prognostic value of e' velocity is recognized in the hypertensive setting,171 and E/e' ratio ≥ 13168 is associated with increased cardiac risk, independent of LVM and relative wall thickness in hypertensive patients.171 Determination of left atrial dilatation can provide additional information and is a prerequisite for the diagnosis of diastolic dysfunction. Left atrial size is best assessed by its indexed volume or LAVi.159 LAVi ≥34 mL/m2 has been shown to be an independent predictor of death, heart failure, atrial fibrillation and ischaemic stroke.172
Normal ranges and cut-off values for hypertensive heart disease for key echocardiographic parameters are summarized in Table 11. The most used scaling for evaluating LVH in hypertension is to divide LVM by BSA, so that the effects on LVM of body size and obesity are largely eliminated. Despite largely derived from control study populations with the obvious possibility for bias, these parameters recommended by the American Society of Echocardiography and the European Association of Echocardiography are used in the majority of laboratories for echocardiography. Data from large general populations in different ethnicities will be available soon.
LA = left atrium; LV = left ventricle; RWT = relative wall thickness.
To assess subclinical systolic dysfunction, speckle tracking echocardiography can quantify longitudinal contractile function (longitudinal strain) and help to unmask early subclinical systolic dysfunction of newly diagnosed hypertensive patients without LVH.173,174 However, assessment of LV systolic function in hypertensive heart disease does not add prognostic information to LVM, at least in the context of a normal EF.
In clinical practice, echocardiography should be considered in hypertensive patients in different clinical contexts and with different purposes: in hypertensive patients at moderate total CV risk, it may refine the risk evaluation by detecting LVH undetected by ECG; in hypertensive patients with ECG evidence of LVH it may more precisely assess the hypertrophy quantitatively and define its geometry and risk; in hypertensive patients with cardiac symptoms, it may help to diagnose underlying disease. It is obvious that echocardiography, including assessment of ascending aorta and vascular screening, may be of significant diagnostic value in most patients with hypertension and should ideally be recommended in all hypertensive patients at the initial evaluation. However, a wider or more restricted use will depend on availability and cost.
22.214.171.124 Cardiac magnetic resonance imaging
Cardiac magnetic resonance imaging (MRI) should be considered for assessment of LV size and mass when echocardiography is technically not feasible and when imaging of delayed enhancement would have therapeutic consequences.175,176
126.96.36.199 Myocardial ischaemia
Specific procedures are reserved for diagnosis of myocardial ischaemia in hypertensive patients with LVH.177 This is particularly challenging because hypertension lowers the specificity of exercise electrocardiography and perfusion scintigraphy.178 An exercise test, demonstrating a normal aerobic capacity and without significant ECG changes, has an acceptable negative predictive value in patients without strong symptoms indicative of obstructive CHD. When the exercise ECG is positive or uninterpretable/ambiguous, an imaging test of inducible ischaemia, such as stress cardiac MRI, perfusion scintigraphy, or stress echocardiography is warranted for a reliable identification of myocardial ischaemia.178–180 Stress-induced wall motion abnormalities are highly specific for angiographically assessed epicardial coronary artery stenosis, whereas myocardial perfusion abnormalities are frequently found with angiographically normal coronary arteries associated with LVH and/or coronary microvascular disease.177 The use of dual echocardiographic imaging of regional wall motion and transthoracic, Doppler-derived coronary flow reserve on the left anterior descending artery has recently been suggested to distinguish obstructive CHD (reduced coronary reserve plus inducible wall motion abnormalities) from isolated coronary microcirculatory damage (reduced coronary reserve without wall motion abnormalities).180 A coronary flow reserve ≤1.91 has been shown to have an independent prognostic value in hypertension.181,182
188.8.131.52 Carotid arteries
Ultrasound examination of the carotid arteries with measurement of intima media thickness (IMT) and/or the presence of plaques has been shown to predict the occurrence of both stroke and myocardial infarction, independently of traditional CV risk factors.51,183–186 This holds true, both for the IMT value at the carotid bifurcations (reflecting primarily atherosclerosis) and for the IMT value at the level of the common carotid artery (reflecting primarily vascular hypertrophy). The relationship between carotid IMT and CV events is a continuous one and determining a threshold for high CV risk is rather arbitrary. Although a carotid IMT >0.9 mm has been taken as a conservative estimate of existing abnormalities in the 2007 Guidelines,2 the threshold value for high CV risk was higher in the elderly patients of the Cardiovascular Health Study and in the middle-aged patients of the European Lacidipine Study on Atherosclerosis (ELSA) study (1.06 and 1.16 mm, respectively).184,186 Presence of a plaque can be identified by an IMT ≥1.5 mm or by a focal increase in thickness of 0.5 mm or 50% of the surrounding carotid IMT value.187 Although plaque has a strong independent predictive value for CV events,51,183–185,188 presence of a plaque and increased carotid IMT added little to each other for predicting CV events and re-classifying patients into another risk category in the Atherosclerosis Risk In Communities (ARIC) study.185 A recent systematic review concluded that the added predictive value of additional carotid screening may be primarily found in asymptomatic individuals at intermediate CV risk.189
184.108.40.206 Pulse wave velocity
Large artery stiffening and the wave-reflection phenomenon have been identified as being the most important pathophysiological determinants of ISH and pulse pressure increase with ageing.190 Carotid-femoral PWV is the ‘gold standard’ for measuring aortic stiffness.138 Although the relationship between aortic stiffness and events is continuous, a threshold of >12 m/s has been suggested by the 2007 ESH/ESC Guidelines as a conservative estimate of significant alterations of aortic function in middle-aged hypertensive patients.2 A recent expert consensus statement adjusted this threshold value to 10 m/s,191 by using the direct carotid-to-femoral distance and taking into account the 20% shorter true anatomical distance travelled by the pressure wave (i.e. 0.8 × 12 m/s or 10 m/s). Aortic stiffness has independent predictive value for fatal and non-fatal CV events in hypertensive patients.192,193 The additive value of PWV above and beyond traditional risk factors, including SCORE and Framingham risk score, has been quantified in a number of studies.51,52,194,195 In addition, a substantial proportion of patients at intermediate risk could be reclassified into a higher or lower CV risk, when arterial stiffness is measured.51,195,196
220.127.116.11 Ankle–brachial index
Ankle–brachial index (ABI) can be measured either with automated devices, or with a continuous-wave Doppler unit and a BP sphygmomanometer. A low ABI (i.e. <0.9) signals PAD and, in general, advanced atherosclerosis,197 has predictive value for CV events,198 and was associated with approximately twice the 10-year CV mortality and major coronary event rate, compared with the overall rate in each Framingham category.198 Furthermore, even asymptomatic PAD, as detected by a low ABI, has prospectively been found to be associated in men with an incidence of CV morbid and fatal events approaching 20% in 10 years.198,199 However, ABI is more useful for detecting PAD in individuals with a high likelihood of PAD.
18.104.22.168 Other methods
Although measurements of carotid IMT, aortic stiffness or ABI are reasonable for detecting hypertensive patients at high CV risk, several other methods, used in the research setting for detecting vascular OD, cannot be supported for clinical use. An increase in the wall–lumen ratio of small arteries can be measured in subcutaneous tissues obtained through gluteal biopsies. These measurements can demonstrate early alterations in diabetes and hypertension and have a predictive value for CV morbidity and mortality,199–202 but the invasiveness of the method makes this approach unsuitable for general use. Increase in coronary calcium, as quantified by high-resolution cardiac computed tomography, has also been prospectively validated as a predictor of CVD and is highly effective in re-stratifying asymptomatic adults into either a moderate or a high CVD risk group,203,204 but the limited availability and high cost of the necessary instrumentations present serious problems. Endothelial dysfunction predicts outcome in patients with a variety of CVDs,205 although data on hypertension are still rather scant.206 Furthermore, the techniques available for investigating endothelial responsiveness to various stimuli are laborious, time consuming and often invasive.
The diagnosis of hypertension-induced renal damage is based on the finding of a reduced renal function and/or the detection of elevated urinary excretion of albumin.207 Once detected, CKD is classified according to estimated glomerular filtration rate (eGFR), calculated by the abbreviated ‘modification of diet in renal disease’ (MDRD) formula,208 the Cockcroft-Gault formula or, more recently, through the Chronic Kidney Disease EPIdemiology Collaboration (CKD-EPI) formula,209 which require age, gender, ethnicity and serum creatinine. When eGFR is below 60 mL/min/1.73 m2, three different stages of CKD are recognized: stage 3 with values between 30–60 mL/min/1.73 m2; and stages 4 and 5 with values below 30 and 15 mL/min/1.73 m2, respectively.210 These formulae help to detect mild impairment of renal function when serum creatinine values are still within the normal range.211 A reduction in renal function and an increase in CV risk can be inferred from the finding of increased serum levels of cystatin C.212 A slight increase (up to 20%) in serum creatinine may sometimes occur when antihypertensive therapy—particularly by renin-angiotensin system (RAS) blockers—is instituted or intensified but this should not be taken as a sign of progressive renal deterioration. Hyperuricaemia is frequently seen in untreated hypertensive patients (particularly in pre-eclampsia) and has been shown to correlate with a reduced renal blood flow and nephrosclerosis.213
While an elevated serum creatinine concentration or a low eGFR point to diminished renal function, the finding of an increased rate of urinary albumin or protein excretion points, in general, to a derangement in glomerular filtration barrier. Microalbuminuria has been shown to predict the development of overt diabetic nephropathy in both type 1 and type 2 diabetic patients,214 while the presence of overt proteinuria generally indicates the existence of established renal parenchymatous disease.215 In both diabetic and non-diabetic hypertensive patients, microalbuminuria, even below the threshold values usually considered,216 has been shown to predict CV events,217–225 and continuous relationships between CV, as well as non-CV mortality and urinary albumin/creatinine ratios >3.9 mg/g in men and >7.5 mg/g in women, have been reported in several studies.224,226 Both in the general population and in diabetic patients, the concomitance of an increased urinary protein excretion and a reduced eGFR indicates a greater risk of CV and renal events than either abnormality alone, making these risk factors independent and cumulative.227,228 An arbitrary threshold for the definition of microalbuminuria has been established as 30 mg/g of creatinine.228
In conclusion, the finding of an impaired renal function in a hypertensive patient, expressed as any of the abnormalities mentioned above, constitutes a very potent and frequent predictor of future CV events and death.218,229–233 Therefore it is recommended, in all hypertensive patients, that eGFR be estimated and that a test for microalbuminuria be made on a spot urine sample.
The traditional classification system of hypertensive retinopathy by fundoscopy is based on the pioneering work by Keith, Wagener and Barker in 1939 and its prognostic significance has been documented in hypertensive patients.234 Grade III (retinal haemorrhages, microaneurysms, hard exudates, cotton wool spots) and grade IV retinopathy (grade III signs and papilloedema and/or macular oedema) are indicative of severe hypertensive retinopathy, with a high predictive value for mortality.234,235 Grade I (arteriolar narrowing either focal or general in nature) and grade II (arteriovenous nicking) point to early stage of hypertensive retinopathy and the predictive value of CV mortality is controversially reported and, overall, less stringent.236,237 Most of these analyses have been done by retinal photography with interpretation by ophthalmologists, which is more sensitive than direct ophthalmoscopy/fundoscopy by general physicians.238 Criticism with respect to the reproducibility of grade I and grade II retinopathy has been raised, since even experienced investigators displayed high inter-observer and intra-observer variability (in contrast to advanced hypertensive retinopathy).239,240
The relationship of retinal vessel calibre to future stroke events has been analysed in a systematic review and individual participant meta-analysis: wider retinal venular calibre predicted stroke, whereas the calibre of retinal arterioles was not associated with stroke.241 Retinal arteriolar and venular narrowing, similarly to capillary rarefaction in other vascular beds,242,243 may be an early structural abnormality of hypertension but its additive value to identify patients at risk for other types of OD needs to be defined.243–244 The arteriovenous ratio of retinal arterioles and venules predicted incident stroke and CV morbidity, but criticism that concomitant changes of the venule diameters may affect this ratio and the methodology (digitized photographs, need of core reading centre) prohibited its widespread clinical use.245–248 New technologies to assess the wall–lumen ratio of retinal arterioles that directly measure the vascular remodelling in early and later stages of hypertensive disease are currently being investigated.249
Hypertension, beyond its well-known effect on the occurrence of clinical stroke, is also associated with the risk of asymptomatic brain damage noticed on cerebral MRI, in particular in elderly individuals.250,251 The most common types of brain lesions are white matter hyperintensities, which can be seen in almost all elderly individuals with hypertension250 – although with variable severity – and silent infarcts, the large majority of which are small and deep (lacunar infarctions) and the frequency of which varies between 10% and 30%.252 Another type of lesion, more recently identified, are microbleeds, seen in about 5% of individuals. White matter hyperintensities and silent infarcts are associated with an increased risk of stroke, cognitive decline and dementia.250,252–254 In hypertensive patients without overt CVD, MRI showed that silent cerebrovascular lesions are even more prevalent (44%) than cardiac (21%) and renal (26%) subclinical damage and do frequently occur in the absence of other signs of organ damage.255 Availability and cost considerations do not allow the widespread use of MRI in the evaluation of elderly hypertensives, but white matter hyperintensity and silent brain infarcts should be sought in all hypertensive patients with neural disturbance and, in particular, memory loss.255–257 As cognitive disturbances in the elderly are, at least in part, hypertension related,258,259 suitable cognitive evaluation tests may be used in the clinical assessment of the elderly hypertensive patient.
Clinical value and limitations
Table 12 summarizes the CV predictive value, availability, reproducibility and cost-effectiveness of procedures for detection of OD. The recommended strategies for the search for OD are summarized in the Table.
Scores are from + to ++ + +.
Summary of recommendations on the search for asymptomatic organ damage, cardiovascular disease, and chronic kidney disease
See ‘Search for asymptomatic organ damage, cardiovascular disease, and chronic kidney disease’ on page 21.
CV = cardiovascular; ECG = electrocardiogram; GFR = glomerural filtration rate; LVH = left ventricular hypertrophy; MRI = magnetic resonance imaging; PAD = peripheral artery disease; PWV = pulse wave velocity.
aClass of recommendation.
bLevel of evidence.
cReference(s) supporting recommendation(s).
dThe MDRD formula is currently recommended but new methods such as the CKD-EPI method aim to improve the accuracy of the measurement.
Searching for secondary forms of hypertension
A specific, potentially reversible cause of BP elevation can be identified in a relatively small proportion of adult patients with hypertension. However, because of the overall high prevalence of hypertension, secondary forms can affect millions of patients worldwide. If appropriately diagnosed and treated, patients with a secondary form of hypertension might be cured, or at least show an improvement in BP control and a reduction of CV risk. Consequently, as a wise precaution, all patients should undergo simple screening for secondary forms of hypertension. This screening can be based on clinical history, physical examination and routine laboratory investigations (Tables 9, 10, 13). Furthermore, a secondary form of hypertension can be indicated by a severe elevation in BP, sudden onset or worsening of hypertension, poor BP response to drug therapy and OD disproportionate to the duration of hypertension. If the basal work-up leads to the suspicion that the patient is suffering from a secondary form of hypertension, specific diagnostic procedures may become necessary, as outlined in Table 13. Diagnostics of secondary forms of hypertension, especially in cases with a suspicion of endocrine hypertension, should preferably be performed in referral centres.
CT = computed tomography; GFR = glomerular filtration rate; MRI = magnetic resonance imaging; RAA = renin–angiotensin–aldosterone.
Evidence favouring therapeutic reduction of high blood pressure
Evidence favouring the administration of BP-lowering drugs to reduce the risk of major clinical CV outcomes (fatal and non-fatal stroke, myocardial infarction, heart failure and other CV deaths) in hypertensive individuals results from a number of RCTs—mostly placebo-controlled—carried out between 1965 and 1995. Their meta-analysis260 was referred to in the 2003 edition of ESH/ESC Guidelines.1 Supportive evidence also comes from finding that a BP-induced regression of OD, such as LVH and urinary protein excretion, may be accompanied by a reduction of fatal and non-fatal outcomes,261,262 although this evidence is obviously indirect, being derived from post-hoc correlative analyses of randomized data.
Randomized trials based on hard clinical CV outcomes do, however, also have limitations, which have been considered in previous ESH/ESC Guidelines:2 (i) to limit the number of patients needed, trials commonly enrol high-risk patients (old age, concomitant or previous disease) and (ii) for practical reasons, the duration of controlled trials is necessarily short (in best cases between 3 and 6 years, with an average time to an endpoint of only half of this)—so that recommendations for life-long intervention are based on considerable extrapolation from data obtained over periods much shorter than the life expectancy of most patients. Support for the belief that the benefits measured during the first few years will continue over a much longer term comes from observational studies of a few decades duration.263
The recommendations that now follow are based on available evidence from randomized trials and focus on important issues for medical practice: (i) when drug therapy should be initiated, (ii) the target BP to be achieved by treatment in hypertensive patients at different CV risk levels, and (iii) therapeutic strategies and choice of drugs in hypertensive patients with different clinical characteristics.
When to initiate antihypertensive drug treatment
Recommendations of previous Guidelines
The 2007 ESH/ESC Guidelines,2 like many other scientific guidelines,54,55,264 recommended the use of antihypertensive drugs in patients with grade 1 hypertension even in the absence of other risk factors or OD, provided that non-pharmacological treatment had proved unsuccessful. This recommendation also specifically included the elderly hypertensive patient. The 2007 Guidelines,2 furthermore, recommended a lower threshold for antihypertensive drug intervention in patients with diabetes, previous CVD or CKD and suggested treatment of these patients, even when BP was in the high normal range (130–139/85–89 mmHg). These recommendations were re-appraised in a 2009 ESH Task Force document141 on the basis of an extensive review of the evidence.265 The following now summarizes the conclusions for the current Guidelines.
Grade 2 and 3 hypertension and high-risk grade 1 hypertension
RCTs providing incontrovertible evidence in favour of antihypertensive therapy,260 as referred to in Section 4.1, were carried out primarily in patients with SBP ≥160 mmHg or DBP ≥100 mmHg, who would now be classified as grade 2 and 3 hypertensives—but also included some patients with grade 1 high-risk hypertension. Despite some difficulty in applying new classifications to old trials, the evidence favouring drug therapy in patients with marked BP elevation or in hypertensive patients at high total CV risk appears overwhelming. BP represents a considerable component of overall risk in these patients and so merits prompt intervention.
Low-to-moderate risk, grade 1 hypertension
The evidence favouring drug treatment in these individuals is scant because no trial has specifically addressed this condition. Some of the earlier trials on ‘mild’ hypertension used a different grading of hypertension (based on DBP only)266–268 or included patients at high risk.268 The more recent Felodipine EVent Reduction (FEVER) study switched patients from pre-existing therapies to randomized treatments and, therefore, could not precisely define baseline hypertension grade; it also included complicated and uncomplicated hypertensives.269 Further analyses of FEVER have recently confirmed a significant benefit attached to more-intensive lowering of BP after exclusion of all patients with previous CVD or diabetes, and in patients with randomization SBP below the median (153 mmHg).270 Because, at randomization, all patients were on a 12.5 mg daily dose of hydrochlorothiazide only, it is likely that these individuals—if untreated—would be within or very close to the SBP range defining grade 1 hypertension. Overall, a number of trials have shown significant reductions of stroke in patients at low-to-moderate CV risk (8–16% major CV events in 10 years) with baseline BP values close to, even if not exactly within, the range of grade 1 hypertension.266,267,270 Also a recent Cochrane Collaboration meta-analysis (2012-CD006742) limited to patients strictly responding to grade 1 low risk criteria finds a trend towards reduction of stroke with active therapy, but the very small number of patients retained (half of those in 266, 267) makes attainment of statistical significance problematic.
Recent guidelines have also underlined the paucity of data for treating grade 1 hypertension,271 recommending treatment only after confirming hypertension by ABPM and restricting treatment to grade 1 hypertensive patients with signs of OD or at high total CV risk. The advantage of systematically excluding white-coat hypertensives from the possible benefit of treatment is unproven. Further arguments in favour of treating even low-moderate risk grade 1 hypertensives are that: (i) waiting increases total risk, and high risk is often not entirely reversible by treatment,272 (ii) a large number of safe antihypertensive drugs are now available and treatment can be personalized in such a way as to enhance its efficacy and tolerability, and (iii) many antihypertensive agents are out of patent and are therefore cheap, with a good cost–benefit ratio.
Isolated systolic hypertension in youth
A number of young healthy males have elevated values of brachial SBP (>140 mmHg) and normal values of brachial DBP (<90 mmHg). As mentioned in section 3.1, these subjects sometimes have normal central BP. No evidence is available that they benefit from antihypertensive treatment; on the contrary there are prospective data that the condition does not necessarily proceed to systolic/diastolic hypertension.142 On the basis of current evidence, these young individuals can only receive recommendations on lifestyle, but because available evidence is scanty and controversial they should be followed closely.
Grade 1 hypertension in the elderly
Although the 2007 ESH/ESC and other guidelines recommended treating grade 1 hypertensives independently of age,2,273 it has been recognized that all the trials showing the benefits of antihypertensive treatment in the elderly have been conducted in patients with SBP ≥160 mmHg (grades 2 and 3).141,265
High normal blood pressure
The 2007 ESH/ESC Guidelines suggested initiation of antihypertensive drug treatment when BP is in the high normal range (130–139/85–89 mmHg) in high- and very high-risk patients because of diabetes or concomitant CV or renal disease.2 The 2009 re-appraisal document pointed out that evidence in favour of this early intervention was, at best, scanty.141,265 For diabetes, the evidence is limited to: (i) the small ‘normotensive’ Appropriate Blood Pressure in Diabetes (ABCD) trial, in which the definition of normotension was unusual (<160 mmHg SBP) and benefit of treatment was seen only in one of several secondary CV events,274 and (ii) subgroup analyses of two trials,275,276 in which results in ‘normotensives’ (many of whom were under treatment) were reported not to be significantly different from those in ‘hypertensives’ (homogeneity test). Furthermore, in two studies in pre-diabetic or metabolic syndrome patients with a baseline BP in the high normal range, administration of ramipril or valsartan was not associated with any significant improvement in morbid and fatal CV events, compared with placebo.277,278
Of two trials showing CV event reduction by lowering of BP in patients with a previous stroke, one included only 16% normotensives,279 while, in a sub-analysis of the other, significant benefits were restricted to patients with baseline SBP ≥140 mmHg (most already under baseline antihypertensive therapy).280 A review of placebo-controlled trials of antihypertensive therapy in coronary patients showed dissimilar results in different studies.265 In most of these trials, randomized drugs were added on a background of antihypertensive drugs, therefore it is inappropriate to classify these patients as normotensive.265 This consideration also applies to recent large meta-analyses showing the benefits of BP-lowering therapy also in individuals with baseline SBP above and below 140 mmHg, since the great majority of the individuals had been involved in trials in which antihypertensive agents were present at baseline.281–284 It is true that two studies have shown that a few years' administration of antihypertensive agents to individuals with high normal BP can delay transition to hypertension,285,286 but how far the benefit of this early intervention lasts—and whether it can also delay events and be cost-effective—remains to be proven.
Summary of recommendations on initiation of antihypertensive drug treatment
Recommendations on initiation of antihypertensive drug treatment are summarized in Figure 2 and below.
BP = blood pressure; CKD = chronic kidney disease; CV = cardiovascular; CVD = cardiovascular disease; OD = organ damage; SBP = systolic blood pressure.
aClass of recommendation.
bLevel of evidence.
cReference(s) supporting recommendation(s).
Blood pressure treatment targets
Recommendations of previous Guidelines
The 2007 ESH/ESC Guidelines,2 in common with other guidelines, recommended two distinct BP targets, namely <140/90 in low-moderate risk hypertensives and <130/80 mmHg in high-risk hypertensives (with diabetes, cerebrovascular, CV, or renal disease). More recently, the European Guidelines on CVD Prevention recommended a target of <140/80 mmHg for patients with diabetes.50 A careful review of the available evidence,265 however, leads to a re-appraisal of some of these recommendations,141 as detailed below.
Low-to-moderate risk hypertensive patients
In three trials,266,268,269 reducing SBP below 140 mmHg compared with a control group at >140 mmHg was associated with a significant reduction in adverse CV outcomes. Although, in two of these trials,268,269 CV risk in the less-intensively treated group was in the high-risk range (>20% CV morbidity and mortality in 10 years), a recent sub-analysis of FEVER has shown, over ten years, CV outcome reduction through lowering SBP to 137 rather than 142 mmHg in patients free of CVD and diabetes with CV risk of about 11% and 17%.270
Hypertension in the elderly
In the large number of randomized trials of antihypertensive treatment in the elderly (including one in hypertensive patients aged 80 years or more)287 all showing reduction in CV events through lowering of BP, the average achieved SBP never attained values <140 mmHg.265 Conversely, two recent Japanese trials of more- vs. less-intensive BP lowering were unable to observe benefits by lowering average SBP to 136 and 137 mmHg rather than 145 and 142 mmHg.288,289 On the other hand, a subgroup analysis of elderly patients in the FEVER study showed reduction of CV events by lowering SBP just below 140 mmHg (compared with 145 mmHg).270
The re-appraisal of ESH/ESC Guidelines carried out in 2009141 has adopted the results of an extensive review of RCT evidence,265 showing that the recommendation of previous Guidelines,2 to lower BP to <130/80 mmHg in patients with diabetes or a history of CV or renal disease, is not supported by RCT evidence.
22.214.171.124 Diabetes mellitus
Lowering BP was found to be associated with important reductions in CV events: (i) in patients with diabetes included in a number of trials,270,275,290–292 (ii) in two trials entirely devoted to these patients,276,293 and (iii) in a recent meta-analysis.294 In two trials,290,293 the beneficial effect was seen from DBP reductions to between 80–85 mmHg, whereas in no trial was SBP ever reduced below 130 mmHg. The only trial in patients with diabetes that achieved SBP values just lower than 130 mmHg in the more intensively treated group, was the ‘normotensive’ ABCD study, a very small study in which CV events (only a secondary endpoint) were not consistently reduced.274 Although being somewhat underpowered, the much larger Action to Control Cardiovascular Risk in Diabetes (ACCORD) study was unable to find a significant reduction in incidence of major CV events in patients with diabetes whose SBP was lowered to an average of 119 mmHg, compared with patients whose SBP remained at an average of 133 mmHg.295
126.96.36.199 Previous cardiovascular events
In two studies of patients who had experienced previous cerebrovascular events,279,296 more aggressive lowering of BP, although associated with significant reductions in stroke and CV events, did not achieve average SBP values lower than 130 mmHg; a third much larger study was unable to find outcome differences between groups achieving SBP of 136 vs. 140 mmHg.297 Among several trials in patients who had previous coronary events, SBP values lower than 130 mmHg were achieved by more intensive treatment in five trials, but with inconsistent results (a significant reduction of CV events in one,298 a significant reduction by one antihypertensive agent, but not by another, in a second trial,299 and no significant reduction in hard CV outcomes in three other studies).300–302
188.8.131.52 Renal disease
In patients with CKD—with or without diabetes—there are two treatment objectives: (i) prevention of CV events (the most frequent complication of CKD) and (ii) prevention or retardation of further renal deterioration or failure. Unfortunately, evidence concerning the BP target to be achieved in these patients is scanty and confused by the uncertainty about the respective roles of reduction of BP and specific effects of RAS blockers.303 In three trials in CKD patients, almost exclusively without diabetes,304–306 patients randomized to a lower target BP (125–130 mmHg) had no significant differences in ESRD or death from patients randomized to a higher target (<140 mmHg). Only in a prolonged observational follow-up of two of these trials was there a trend towards lower incidence of events, which was more evident in patients with proteinuria.307,308 The two large trials in patients with diabetic nephropathy are not informative on the supposed benefit of a SBP target below 130 mmHg,309,310 since the average SBPs achieved in the groups with more intensive treatment were 140 and 143 mmHg. Only a recent co-operative study has reported a reduction in renal events (GFR reduction and ESRD) in children randomized to a BP target below—rather than above—the 50th percentile,311 but these values in children can hardly be compared with adult values. Furthermore it should be considered that, in ACCORD, although eGFR at baseline was in the normal range, more intensive lowering of BP (119/67 vs. 134/73 mmHg) was associated with a near-doubling of cases with eGFR <30 mL/min/1.73 m2.295 Finally, recent meta-analyses of trials investigating different BP targets in patients with CKD failed to demonstrate definite benefits from achieving lower BP goals in terms of CV or renal clinical events.312,313
The ‘lower the better’ vs. the J-shaped curve hypothesis
The concept that ‘the lower the SBP and DBP achieved the better the outcome’ rests on the direct relationship between BP and incident outcomes, down to at least 115 mmHg SBP and 75 mmHg DBP, described in a large meta-analysis of 1 million individuals free of CVD at baseline and subsequently followed for about 14 years3—not the usual situation for hypertension trials. The concept assumes that the BP/outcome relationship down to the lowest BP values is also seen when the BP differences are induced by drug therapy and that the relationship in patients with CVD can be superimposed on that described in individuals free of CV complications. In the absence of trials that have specifically investigated low SBP ranges (see above), the only available data in favour of the ‘lower the better’ concept are those of a meta-analysis of randomized trials, showing that reduction of SBP to a mean of 126 mmHg, compared with 131 mmHg, had the same proportional benefits as reduction to a mean of 140 mmHg, compared with 145 mmHg.281 Of course, this was a post-hoc analysis, in which randomization was lost because the splitting of the patients into the BP categories was not considered at the randomization stage. Demonstration of the ‘lower the better’ hypothesis is also made difficult by the fact that the curve relating BP and adverse CV events may flatten at low BP values, and therefore demonstration of benefits requires much larger and longer studies than those yet available. This is consistent with the semi-logarithmic nature of the relationship shown in observational studies,3 but it may also raise the question of whether a small benefit is worth large effort.
An alternative to the ‘lower the better’ concept is the hypothesis of a J-shaped relationship, according to which the benefits of reducing SBP or DBP to markedly low values are smaller than for reductions to more moderate values. This hypothesis continues to be widely popular for several reasons: (i) common sense indicates that a threshold BP must exist, below which survival is impaired, (ii) physiology has shown that there is a low (as well as a high) BP threshold for organ blood-flow autoregulation and this threshold can be elevated when there is vascular disease, and (iii) there is a persistent hang-over from an old belief viewing high BP as a compensatory mechanism for preserving organ function (the ‘essential’ nature of hypertension).314 Correct investigation of the J-curve requires randomized comparison of three BP targets, only attempted in the Hypertension Optimal Treatment (HOT) study but in low-risk hypertensives and using DBP targets.290 Owing to the lack of direct evidence, recourse has been made to the indirect observational approach of relating outcomes to achieved BP. A number of trials have been so analysed and their results recently reviewed.314 Some of the trial analyses have concluded that no J-curve exists,280,290,315 while others have concluded in favour of its existence,316–319 although in some trials it was also seen in placebo-treated patients.320,321 Furthermore, two recent trials investigating more- or less-intensive low-density lipoprotein cholesterol lowering by statins also found a J-curve relating BP to adverse CV events, although protocols did not include BP-lowering interventions.322,323 The approach used to investigate the J-curve raises important hypotheses, yet has obvious limitations: (i) it changes a randomized study into an observational one, (ii) the numbers of patients and events in the lowest BP groups are usually very small, (iii) patients in the lowest BP groups are often at increased baseline risk and, despite statistical adjustments, reverse-causality cannot be excluded; and (iv) the ‘nadir’ SBP and DBP values (the values at which risk starts to increase) are extremely different from trial to trial, even when baseline CV risk is similar.314 Some trial analyses have also raised the point that a J-curve may exist for coronary events but not for strokes—but this is not a consistent finding in various trials.317,318,324–326 Whether or not the underlying high risk to patients is more important than the excessive BP reduction should be considered. The limitations of the current approach for investigating the J-curve obviously also apply to their meta-analyses.327 Yet the J-curve hypothesis is an important issue: it has a pathophysiological rationale and deserves to be investigated in a correctly designed trial.
Evidence on target blood pressure from organ damage studies
It would be of some interest to receive guidance about target BP from OD studies, but unfortunately this information must be judged with great caution. Indeed, trials using OD as an endpoint often do not have sufficient statistical power to safely measure effects on CV outcome and the data they provide on fatal and non-fatal CV events are subject to the effects of chance. For example, a study of 1100 non-diabetic hypertensive patients, followed for 2 years, showed that the incidence of electrocardiographic LVH is reduced by tighter (about 132/77 mmHg) vs. less-tight BP control (about 136/79 mmHg) and reported a parallel reduction in CV events (although there were only about 40 hard outcome events).328 On the other hand, the recent Randomized Olmesartan And Diabetes MicroAlbuminuria Prevention (ROADMAP) study329 in diabetic patients showed a significant reduction of new-onset microalbuminuria in more intensively treated patients (olmesartan vs. placebo), but the more intensively treated group also had a higher incidence of CV outcomes.329 Because of the small number of CV events in the two trials, it is likely that both their reduction and their increase are due to chance effects. Furthermore, when analyses of OD and event effects are made in large trials, dissociation of the two types of effects has been reported: in the Losartan Intervention For Endpoint Reduction in Hypertensives (LIFE) study, LVH regression was linearly related to the treatment-induced BP changes (the lower the better),330 whereas, in the same trial, achieved BP and morbid and fatal CV events were related in a J-shaped manner.319 In ONngoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial (ONTARGET), the lowest BP achieved by the ramipril–telmisartan combination was associated with reduced proteinuria, but with a greater risk of acute renal failure and a similar CV risk.331 The clinical significance of treatment-induced changes in OD is further discussed in Section 8.4.
Clinic vs. home and ambulatory blood pressure targets
No direct evidence from randomized outcome studies is yet available about BP targets when home or ambulatory BP measurements are used,332 although some evidence is available that differences with office BP may not be too pronounced when office BP is effectively reduced.333 Out-of-office measurements should always be evaluated together with measurements at the clinic. Notably, however, the adjustment of antihypertensive therapy on the basis of a similar target ambulatory or home BP led to less-intensive drug treatment, without a significant difference in OD.334–336 The lower cost of medications in the out-of-office BP groups was partially offset by other costs in the home BP groups.335,336
Summary of recommendations on blood pressure targets in hypertensive patients
Recommendations on BP targets are summarized in Figure 2 and below.
CHD = coronary heart disease; CKD = chronic kidney disease; CV = cardiovascular; DBP = diastolic blood pressure; SBP = systolic blood pressure; TIA = transient ischaemic attack.
aClass of recommendation.
bLevel of evidence.
cReference(s) supporting recommendation(s).
Appropriate lifestyle changes are the cornerstone for the prevention of hypertension. They are also important for its treatment, although they should never delay the initiation of drug therapy in patients at a high level of risk. Clinical studies show that the BP-lowering effects of targeted lifestyle modifications can be equivalent to drug monotherapy,337 although the major drawback is the low level of adherence over time—which requires special action to be overcome. Appropriate lifestyle changes may safely and effectively delay or prevent hypertension in non-hypertensive subjects, delay or prevent medical therapy in grade I hypertensive patients and contribute to BP reduction in hypertensive individuals already on medical therapy, allowing reduction of the number and doses of antihypertensive agents.338 Beside the BP-lowering effect, lifestyle changes contribute to the control of other CV risk factors and clinical conditions.50
The recommended lifestyle measures that have been shown to be capable of reducing BP are: (i) salt restriction, (ii) moderation of alcohol consumption, (iii) high consumption of vegetables and fruits and low-fat and other types of diet, (iv) weight reduction and maintenance and (v) regular physical exercise.339 In addition, insistence on cessation of smoking is mandatory in order to improve CV risk, and because cigarette smoking has an acute pressor effect that may raise daytime ambulatory BP.340–342
There is evidence for a causal relationship between salt intake and BP and excessive salt consumption may contribute to resistant hypertension. Mechanisms linking salt intake and BP elevation include an increase in extracellular volume—but also in peripheral vascular resistance, due in part to sympathetic activation.343 The usual salt intake is between 9 and 12 g/day in many countries and it has been shown that reduction to about 5 g/day has a modest (1–2 mmHg) SBP-lowering effect in normotensive individuals and a somewhat more pronounced effect (4–5 mmHg) in hypertensive individuals.339,344,345 A daily intake of 5–6 g of salt is thus recommended for the general population. The effect of sodium restriction is greater in black people, older people and in individuals with diabetes, metabolic syndrome or CKD, and salt restriction may reduce the number and doses of antihypertensive drugs.345,346 The effect of reduced dietary salt on CVD events remains unclear,347–350 although the long-term follow-up of the Trials of Hypertension Prevention (TOHP) trial showed a reduced salt intake to be associated with lower risk of CV events.351 Overall there is no evidence that reducing sodium from high- to moderate intakes causes harm.352
At the individual level, effective salt reduction is by no means easy to achieve. Advice should be given to avoid added salt and high-salt food. A reduction in population-wide salt intake remains a public health priority but requires a combined effort by the food industry, governments and the public in general, since 80% of salt consumption involves ‘hidden salt’. It has been calculated that salt reduction in the manufacturing processes of bread, processed meat and cheese, margarine and cereals will result in an increase in quality-adjusted life-years.353
Moderation of alcohol consumption
The relationship between alcohol consumption, BP levels and the prevalence of hypertension is linear. Regular alcohol use raises BP in treated hypertensive subjects.354 While moderate consumption may do no harm, the move from moderate to excessive drinking is associated both with raised BP and with an increased risk of stroke. The Prevention And Treatment of Hypertension Study (PATHS) investigated the effects of alcohol reduction on BP. The intervention group had a 1.2/0.7 mmHg greater reduction in BP than the control group at the end of the 6-month period.355 No studies have been designed to assess the impact of alcohol reduction on CV endpoints. Hypertensive men who drink alcohol should be advised to limit their consumption to no more than 20–30 g, and hypertensive women to no more than 10–20 g, of ethanol per day. Total alcohol consumption should not exceed 140 g per week for men and 80 g per week for women.
Other dietary changes
Hypertensive patients should be advised to eat vegetables, low-fat dairy products, dietary and soluble fibre, whole grains and protein from plant sources, reduced in saturated fat and cholesterol. Fresh fruits are also recommended—although with caution in overweight patients because their sometimes high carbohydrate content may promote weight gain.339,356 The Mediterranean type of diet, especially, has attracted interest in recent years. A number of studies and meta-analyses have reported on the CV protective effect of the Mediterranean diet.357,358 Patients with hypertension should be advised to eat fish at least twice a week and 300–400 g/day of fruit and vegetables. Soy milk appeared to lower BP when compared with skimmed cows' milk.359 Diet adjustment should be accompanied by other lifestyle changes. In patients with elevated BP, compared with the Dietary Approaches to Stop Hypertension (DASH) diet alone, the combination of the DASH diet with exercise and weight loss resulted in greater reductions in BP and LVM.360 With regard to coffee consumption, a recent systematic review found that most of the available studies (10 RCTs and 5 cohort studies) were of insufficient quality to allow a firm recommendation to be given for or against coffee consumption as related to hypertension.361
Hypertension is closely correlated with excess body weight,362 and weight reduction is followed by a decrease in BP. In a meta-analysis, the mean SBP and DBP reductions associated with an average weight loss of 5.1 kg were 4.4 and 3.6 mmHg, respectively.363 Weight reduction is recommended in overweight and obese hypertensive patients for control of risk factors, but weight stabilisation may be a reasonable target for many of them. In patients with established CVD manifestations, observational data indicate a worse prognosis following weight loss. This seems to be true also in the elderly. Maintenance of a healthy body weight (BMI of about 25 kg/m2) and waist circumference (<102 cm for men and <88 cm for women) is recommended for non-hypertensive individuals to prevent hypertension and for hypertensive patients to reduce BP. It is noteworthy, however, that the optimal BMI is unclear, based on two large meta-analyses of prospective observational population-based outcome studies. The Prospective Studies Collaboration concluded that mortality was lowest at a BMI of about 22.5–25 kg/m2,364 whereas a more recent meta-analysis concluded that mortality was lowest in overweight subjects .365 Weight loss can also improve the efficacy of antihypertensive medications and the CV risk profile. Weight loss should employ a multidisciplinary approach that includes dietary advice and regular exercise. Weight-loss programmes are not so successful and influences on BP may be overestimated. Furthermore, short-term results are often not maintained in the long term. In a systematic review of diabetic patients,366 the mean weight loss after 1–5 years was 1.7 kg. In ‘pre-diabetic’ patients, combined dietary and physical activity interventions gave a 2.8 kg extra weight reduction after 1 year and a further 2.6 kg reduction after 2 years: while not impressive, this is sufficient to have a protective effect against the incidence of diabetes.367 In established type 2 diabetes mellitus (DM), intentional weight loss—according to the Action for HEalth in Diabetes (AHEAD) study—did not reduce CV events, so that a general control of risk factors is probably more important than weight loss per se. Weight loss can also be promoted by anti-obesity drugs, such as orlistat and, to a greater degree, by bariatic surgery, which appears to decrease CV risk in severely obese patients.368 Details are available in a recent document by the ESH and the European Association for the Study of Obesity.368
Regular physical exercise
Epidemiological studies suggest that regular aerobic physical activity may be beneficial for both prevention and treatment of hypertension and to lower CV risk and mortality. A meta-analysis of randomized controlled trials has shown that aerobic endurance training reduces resting SBP and DBP by 3.0/2.4 mmHg overall and even by 6.9/4.9 mmHg in hypertensive participants.369 Even regular physical activity of lower intensity and duration has been shown to be associated with about a 20% decrease in mortality in cohort studies,370,371 and this is also the case for measured physical fitness.372 Hypertensive patients should be advised to participate in at least 30 min of moderate-intensity dynamic aerobic exercise (walking, jogging, cycling or swimming) on 5–7 days per week.373 Aerobic interval training has also been shown to reduce BP.374 The impact on BP values of other forms of exercise, such as isometric resistance training (muscular force development without movement) and dynamic resistance exercise (force development associated with movement) has been reviewed recently.375,376 Dynamic resistance training was followed by significant BP reduction, as well as improvements in other metabolic parameters, and performance of resistance exercises on 2–3 days per week can be advised. Isometric exercises are not recommended, since data from only a few studies are available.
Smoking is a major risk factor for atherosclerotic CVD. Although the rate of smoking is declining in most European countries (in which a legalized smoking ban is effective) it is still common in many regions and age groups, partly due to education-related inequalities in cessation of smoking.377 There is evidence also on the ill-health effects of passive smoking.378 Smoking causes an acute increase in BP and heart rate, persisting for more than 15 minutes after smoking one cigarette,340 as a consequence of stimulation of the sympathetic nervous system at the central level and at the nerve endings.379 A parallel change in plasma catecholamines and BP, plus an impairment of the baroreflex, have been described that are related to smoking.379–381 Studies using ABPM have shown that both normotensive and untreated hypertensive smokers present higher daily BP values than non-smokers.341,342,382 No chronic effect of smoking has been reported for office BP,383 which is not lowered by giving up smoking. Beside the impact on BP values, smoking is a powerful CV risk factor and quitting smoking is probably the single most effective lifestyle measure for the prevention of CVDs including stroke, myocardial infarction and peripheral vascular disease.384–386 Therefore tobacco use status should be established at each patient visit and hypertensive smokers should be counselled regarding giving up smoking.
Even in motivated patients, programmes to stop smoking are successful (at 1 year) in only 20–30%.387 Where necessary, smoking cessation medications, such as nicotine replacement therapy, bupropion, or varenicline, should be considered. A meta-analysis of 36 trials comparing long-term cessation rates using bupropion vs. control yielded a relative success rate of 1.69 (1.53–1.85),388 whereas evidence of any additional effect of adding bupropion to nicotine replacement therapy was inadequate.389 The partial nicotine-receptor agonist varenicline has shown a modest benefit over nicotine replacement therapy and bupropion,388 but the U.S. Food & Drug Administration (FDA) has recently issued a warning regarding the safety profile of varenicline (http://www.fda.gov/Drugs/DrugSafety/ucm330367.htm). Although these drugs have been shown to be effective in clinical trials, they are underused due to adverse effects, contra-indications, low acceptance, high cost and lack of reimbursement in many countries. Relapse prevention is a cornerstone in fighting nicotine addiction but the field is inadequately studied and existing evidence is disappointing.388 There is insufficient evidence to support the use of any specific behavioural intervention; some positive results can be expected from interventions focussing on identifying and resolving temptation situations, as well as from strategies steering patients towards changes in behaviours, such as motivational interviews. Extended treatment with varenicline may prevent relapse but studies of extended treatment with nicotine replacement are not available.390
Summary of recommendations on adoption of lifestyle changes
The following lifestyle change measures are recommended in all patients with hypertension to reduce BP and/or the number of CV risk factors.
BMI = body mass index.
aClass of recommendation.
bLevel of evidence.
cReference(s) supporting recommendation(s).
dBased on the effect on BP and/or CV risk profile.
eBased on outcome studies.
Choice of antihypertensive drugs
In the 2003 and 2007 versions,1,2 the ESH/ESC Guidelines reviewed the large number of randomized trials of antihypertensive therapy and concluded that the main benefits of antihypertensive treatment are due to lowering of BP per se and are largely independent of the drugs employed. Although meta-analyses occasionally appear, claiming superiority of one class of agents over another for some outcomes,391–393 this largely depends on the selection bias of trials and the largest meta-analyses available do not show clinically relevant differences between drug classes.284,394,395 Therefore the current Guidelines reconfirm that diuretics (including thiazides, chlorthalidone and indapamide), beta-blockers, calcium antagonists, angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers are all suitable for the initiation and maintenance of antihypertensive treatment, either as monotherapy or in some combinations. However, some therapeutic issues that have recently been raised are discussed below.
The reasons why, at variance from some guidelines,271 beta-blockers were maintained as a possible choice for antihypertensive treatment were summarized in the 2007 ESH/ESC Guidelines and further discussed in the 2009 re-appraisal document.2,141 Although acknowledging that the quality of the evidence was low, a Cochrane meta-analysis (substantially reproducing a 2006 meta-analysis by the same group)396,397 has reported that beta-blockers may be inferior to some—but not all—other drug classes for some outcomes. Specifically, they appear to be worse than calcium antagonists (but not diuretics and RAS blockers) for total mortality and CV events, worse than calcium antagonists and RAS blockers for stroke and equal to calcium antagonists, RAS blockers and diuretics for CHD. On the other hand, the large meta-analysis by Law et al. has shown beta-blocker-initiated therapy to be (i) equally as effective as the other major classes of antihypertensive agents in preventing coronary outcomes and (ii) highly effective in preventing CV events in patients with a recent myocardial infarction and those with heart failure.284 A similar incidence of CV outcomes with beta-blockers and/or diuretics or their combinations compared with other drug classes has also been reported in the meta-analysis of the BP-lowering treatment trialists' collaboration.394
A slightly lower effectiveness of beta-blockers in preventing stroke284 has been attributed to a lesser ability to reduce central SBP and pulse pressure.398,399 However, a lower effectiveness in stroke prevention is also shared by ACE inhibitors,284 although these compounds have been reported to reduce central BP better than beta-blockers.398 Beta-blockers also appear (i) to have more side-effects (although the difference with other drugs is less pronounced in double blind studies)400 and (ii) to be somewhat less effective than RAS blockers and calcium antagonists in regressing or delaying OD, such as LVH, carotid IMT, aortic stiffness and small artery remodelling.141 Also, beta-blockers tend to increase body weight401 and, particularly when used in combination with diuretics, to facilitate new-onset diabetes in predisposed patients.402 This phenomenon may have been overemphasized by the fact that all trial analyses have been limited to patients free of diabetes or with glucose <7.0 mmol/L, ignoring the fact that a noticeable number of patients with a diagnosis of diabetes at baseline do not have this diagnosis reconfirmed at study end, which obviously reduces the weight of treatment-induced diabetes and raises doubts about the precision of the definition of diabetes used in the above analyses.403 Some of the limitations of traditional beta-blockers do not appear to be shared by some of the vasodilating beta-blockers, such as celiprolol, carvedilol and nebivolol—more widely used today—which reduce central pulse pressure and aortic stiffness better than atenolol or metoprolol404–406 and affect insulin sensitivity less than metoprolol.407,408 Nebivolol has recently been shown not to worsen glucose tolerance compared with placebo and when added to hydrochlorothiazide.409 Both carvedilol and nebivolol have been favourably tested in RCTs, although in heart failure rather than arterial hypertension.410 Finally, beta-blockers have recently been reported not to increase, but even reduce, the risk of exacerbations and to reduce mortality in patients with chronic obstructive lung disease.411
Diuretics have remained the cornerstone of antihypertensive treatment since at least the first Joint National Committee (JNC) report in 1977412 and the first WHO report in 1978,413 and still, in 2003, they were classified as the only first-choice drug by which to start treatment, in both the JNC-7264 and the WHO/International Society of Hypertension Guidelines.55 The wide use of thiazide diuretics should take into account the observation in the Avoiding Cardiovascular Events in Combination Therapy in Patients Living with Systolic Hypertension (ACCOMPLISH) trial,414 that their association with an ACE inhibitor was less effective in reducing CV events than the association of the same ACE inhibitor with a calcium antagonist. The interesting findings of ACCOMPLISH will be discussed in Section 5.2.2 but need replication, because no other randomized study has shown a significant superiority of a calcium antagonist over a diuretic. Therefore, the evidence provided by ACCOMPLISH does not appear to bear sufficient weight to exclude diuretics from first-line choice.
It has also been argued that diuretics such as chlorthalidone or indapamide should be used in preference to conventional thiazide diuretics, such as hydrochlorothiazide.271 The statement that ‘There is limited evidence confirming benefit of initial therapy on clinical outcomes with low doses of hydrochlorothiazide’271 is not supported by a more extensive review of available evidence.332,415 Meta-analyses claiming that hydrochlorothiazide has a lesser ability to reduce ambulatory BP than other agents, or reduces outcomes less than chlorthalidone,416,417 are confined to a limited number of trials and do not include head-to-head comparisons of different diuretics (no large randomized study is available). In the Multiple Risk Factor Intervention Trial (MRFIT), chlorthalidone and hydrochlorothiazide were not compared by randomized assignment and, overall, chlorthalidone was used at higher doses than hydrochlorothiazide.418 Therefore no recommendation can be given to favour a particular diuretic agent.
Spironolactone has been found to have beneficial effects in heart failure419 and, although never tested in RCTs on hypertension, can be used as a third- or fourth-line drug (see Section 6.14) and helps in effectively treating undetected cases of primary aldosteronism. Eplerenone has also shown a protective effect in heart failure and can be used as an alternative to spironolactone.420
184.108.40.206 Calcium antagonists
Calcium antagonists have been cleared from the suspicion of causing a relative excess of coronary events by the same authors who had raised the question. Some meta-analyses suggest that these agents may be slightly more effective in preventing stroke,284,394,421 although it is not clear whether this can be ascribed to a specific protective effect on the brain circulation or to a slightly better or more uniform BP control with this class of drugs.141 The question of whether calcium antagonists may be less effective than diuretics, beta-blockers and ACE inhibitors in preventing incipient heart failure is still an open one. In the largest available meta-analysis,284 calcium antagonists reduced new-onset heart failure by about 20% compared with placebo but, when compared with diuretics, beta-blockers and ACE inhibitors were inferior by about 20% (which means a 19% rather than 24% reduction). The lower effectiveness of calcium antagonists on the onset of heart failure may also be a consequence of the design of the trials pointing to this conclusion, which required prevention or withdrawal of agents essential in heart failure therapy such as diuretics, beta-blockers and ACE inhibitors in patients randomized to calcium antagonists.422 In fact, in all trials in which the design permitted or prescribed the simultaneous use of diuretics, beta-blockers or ACE inhibitors,269,299,301,423 calcium antagonists were not inferior to comparative therapies in preventing heart failure. Calcium antagonists have shown a greater effectiveness than beta-blockers in slowing down progression of carotid atherosclerosis and in reducing LV hypertrophy in several controlled studies (see sections 6.11.4 and 6.12.1).
220.127.116.11 Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers
Both classes are among those most widely used in antihypertensive therapy. Some meta-analyses have suggested that ACE inhibitors may be somewhat inferior to other classes in preventing stroke284,395,421 and that angiotensin receptor blockers may be inferior to ACE inhibitors in preventing myocardial infarction424 or all-cause mortality.393 The hypothesis raised by these meta-analyses has been undermined by the results of the large ONTARGET, directly comparing outcomes under treatment with the ACE inhibitor ramipril and the angiotensin receptor blocker telmisartan (section 18.104.22.168). ONTARGET has shown telmisartan not to be statistically inferior to ramipril as far as incidence of major cardiac outcomes, stroke and all-cause death is concerned. ONTARGET has also disproved the hypothesis that the peroxisome proliferator-activated receptor (PPAR) activity of telmisartan may render this compound more effective in preventing or delaying onset of diabetes: incidence of new diabetes was non-significantly different between telmisartan and ramipril in ONTARGET.
Most recently, the hypothesis has been raised of an association of angiotensin receptor blockers with cancer onset.425 A much larger meta-analysis, including all major randomized trials investigating all major compounds of the class, has subsequently found no evidence of increased cancer incidence,426 for which there is also no basis from a mechanistic standpoint.427 Among the well-known ancillary properties of ACE inhibitors and angiotensin receptor blockers, are their peculiar effectiveness in reducing proteinuria (see section 6.9) and improving outcomes in chronic heart failure (section 6.11.2).
22.214.171.124 Renin inhibitors
Aliskiren, a direct inhibitor of renin at the site of its activation, is available for treating hypertensive patients, both as monotherapy and when combined with other antihypertensive agents. To date, available evidence shows that, when used alone, aliskiren lowers SBP and DBP in younger and elderly hypertensive patients;428 that it has a greater antihypertensive effect when given in combination with a thiazide diuretic, a blocker of the RAS at other sites, or a calcium antagonist;429,430 and that prolonged administration in combination treatment can have a favourable effect (i) on asymptomatic OD, such as urinary protein excretion431 or (ii) on prognostic biomarkers for heart failure, such as B-type natriuretic peptides.432
No trial is available on the effect of aliskiren on CV or renal morbid and fatal events in hypertension. A large-scale trial on diabetic patients, ALiskiren Trial In Type 2 Diabetes Using Cardio-renal Endpoints (ALTITUDE), in which aliskiren was administered on top of an RAS blocker, has recently been stopped because, in these patients at high risk of CV and renal events, there was a higher incidence of adverse events, renal complications (ESRD and renal death), hyperkalaemia and hypotension.433 This treatment strategy is therefore contra-indicated in such specific conditions, similar to the contra-indications for the ACE inhibitor–angiotensin receptor blocker combination resulting from the ONTARGET trial (see Section 5.2.2).331 Another large-scale trial, A Randomized Controlled Trial of Aliskiren in the Prevention of Major Cardiovascular Events in Elderly People (APOLLO), in which aliskiren was used alone or in combination with a thiazide diuretic or a calcium channel blocker, has also been stopped, despite no evidence of harm in the aliskiren-treated group. No aliskiren-based antihypertensive trials with hard endpoints are expected in the near future. No beneficial effect on mortality and hospitalization has recently been shown by adding aliskiren to standard treatment in heart failure.434
126.96.36.199 Other antihypertensive agents
Centrally active agents and alpha-receptor blockers are also effective antihypertensive agents. Nowadays, they are most often used in multiple drug combinations. The alpha-blocker doxazosin has effectively been used as third-line therapy in the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT). This will be further discussed in the section on resistant hypertension (6.14).
188.8.131.52 Antihypertensive agents and visit-to-visit blood pressure variability
Attention has recently been drawn to the association of visit-to-visit variability of intra-individual BP during antihypertensive treatment and the incidence of CV events (particularly stroke) in high-risk patients.435 In coronary hypertensive patients, consistency of BP control between visits is accompanied by less-frequent CV morbidity and mortality, independently of the mean BP level.436 However, in the mild hypertensive, low-CV-risk patients of the ELSA trial, mean on-treatment BP, rather than visit-to-visit BP variations, predicted both the progression of carotid atherosclerosis and the incidence of CV events.437 Thus the clinical importance of visit-to-visit BP variability within treated individuals, vis-a-vis the achieved long-term average BP level, is not yet indisputably proven.
An analysis of the ASCOT trial has suggested that visit-to-visit BP variability may be lower with the combination of a calcium antagonist and an ACE inhibitor, than with the combination of a beta-blocker and a diuretic.438 Additionally, from a meta-analysis of several trials, the conclusion has been reached that visit-to-visit BP variability is more pronounced in patients under beta-blockade than with other drug classes.439,440 Yet, the underlying cause of visit-to-visit BP variability is not known—whether it is really pharmacologically driven or, rather, a marker of treatment adherence. Also, the abovementioned meta-analyses based their results on inter-individual BP variability (i.e. the range of the BP effects of treatment in the whole group of patients) rather than intra-individual variability. The use of inter-individual BP variability as a surrogate of intra-individual variability to classify antihypertensive agents as associated with greater or lesser visit-to-visit BP variations or more or less consistent BP control439,440 seems unjustified, since discrepancies have been reported between the two measures.441 Furthermore, despite any possible correlations, the two types of variability are unlikely to measure the same phenomena.442 In practical terms, until intra-individual visit-to-visit BP variability from new large-scale trials is analysed, inter-individual visit-to-visit variability should not be used as a criterion for antihypertensive drug choice. It remains, however, an interesting subject for further investigation.
184.108.40.206 Should antihypertensive agents be ranked in order of choice?
Once it is agreed that (i) the major mechanism of the benefits of antihypertensive therapy is lowering of BP per se, (ii) the effects on cause-specific outcomes of the various agents are similar or differ by only a minor degree, (iii) the type of outcome in a given patient is unpredictable, and (iv) all classes of antihypertensive agents have their advantages but also contra-indications (Table 14), it is obvious that any all-purpose ranking of drugs for general antihypertensive usage is not evidence-based.141,443 Rather than indulging in an all-purpose ranking, the Task Force decided to confirm (with small changes) the table published in the 2007 ESH/ESC Guidelines,2 with the drugs to be considered in specific conditions, based on the fact that some classes have preferentially been used in trials in specific conditions or have shown greater effectiveness in specific types of OD (see Mancia et al. for detailed evidence)2 (Table 15). However, no evidence is available that different choices should be made based on age or gender (except for caution in using RAS blockers in women with child bearing potential because of possible teratogenic effects).444,445 In any case, physicians should pay attention to adverse drug effects—even those purely subjective—as they are powerful deterrents to treatment adherence. If necessary, doses or drugs should be changed in order to combine effectiveness with tolerability.
A-V = atrio-ventricular; eGFR = estimated glomerular filtration rate; LV = left ventricular.
ACE = angiotensin-converting enzyme; ARB = angiotensin receptor blocker; BB = beta-blocker; BP = blood pressure; CV = cardiovascular; ESRD = end-stage renal disease; ISH = isolated systolic hypertension; LVH = left ventricular hypertrophy.
Monotherapy and combination therapy
220.127.116.11 Pros and cons of the two approaches
The 2007 ESH/ESC Guidelines underlined that, no matter which drug is employed, monotherapy can effectively reduce BP in only a limited number of hypertensive patients and that most patients require the combination of at least two drugs to achieve BP control.2 Therefore, the issue is not whether combination therapy is useful, but whether it should always be preceded by an attempt to use monotherapy, or whether—and when—combination therapy may be the initial approach.
The obvious advantage of initiating treatment with monotherapy is that of using a single agent, thus being able to ascribe effectiveness and adverse effects to that agent. The disadvantages are that, when monotherapy with one agent is ineffective or insufficiently effective, finding an alternative monotherapy that is more effective or better tolerated may be a painstaking process and discourage adherence. Additionally, a meta-analysis of more than 40 studies has shown that combining two agents from any two classes of antihypertensive drugs increases the BP reduction much more than increasing the dose of one agent.446 The advantage of initiating with combination therapy is a prompter response in a larger number of patients (potentially beneficial in high-risk patients), a greater probability of achieving the target BP in patients with higher BP values, and a lower probability of discouraging patient adherence with many treatment changes. Indeed, a recent survey has shown that patients receiving combination therapy have a lower drop-out rate than patients given any monotherapy.447 A further advantage is that there are physiological and pharmacological synergies between different classes of agents, that may not only justify a greater BP reduction but also cause fewer side-effects and may provide larger benefits than those offered by a single agent. The disadvantage of initiating with drug combinations is that one of the drugs may be ineffective.
On the whole the suggestion, given in the 2007 ESH/ESC Guidelines,2 of considering initiation with a drug combination in patients at high risk or with markedly high baseline BP can be reconfirmed.
When initiating with monotherapy or with a two-drug combination, doses can be stepped up if necessary to achieve the BP target; if the target is not achieved by a two-drug combination at full doses, switching to another two-drug combination can be considered or a third drug added. However, in patients with resistant hypertension, adding drugs to drugs should be done with attention to results and any compound overtly ineffective or minimally effective should be replaced, rather than retained in an automatic step-up multiple-drug approach (Figure 3).
18.104.22.168 Preferred drug combinations
Only indirect data are available from randomized trials giving information on drug combinations effective in reducing CV outcomes. Among the large number of RCTs of antihypertensive therapy, only three systematically used a given two-drug combination in at least one arm: the ADVANCE trial compared an ACE inhibitor and diuretic combination with placebo (but on top of continued background therapy),276 FEVER compared a calcium antagonist and diuretic combination with diuretic alone (plus placebo)269 and ACCOMPLISH compared the same ACE inhibitor in combination with either a diuretic or a calcium antagonist.414 In all other trials, treatment was initiated by monotherapy in either arm and another drug (and sometimes more than one drug) was added in some patients. In some trials, the second drug was chosen by the investigator among those not used in the other treatment arms, as in Antihypertensive and Lipid-Lowering Treatment to Prevent Heart ATtack (ALLHAT).448
With this important reservation, Table 16 shows that, with the exception of an angiotensin receptor blocker and a calcium antagonist (never systematically used in an outcome trial), all combinations were used in at least one active arm of placebo-controlled trials in which the active arm was associated with significant benefit.269,276,287,296,449–454 In trials comparing different regimens, all combinations have been used in a larger or smaller proportion of patients, without major differences in benefits.186,445,448,455,456,458–461 The only exceptions are two trials in which a large proportion of the patients received either an angiotensin receptor blocker–diuretic combination or a calcium antagonist–ACE inhibitor combination,423,457 both of which were superior to a beta-blocker–diuretic combination in reducing CV events. Admittedly, a beta-blocker–diuretic combination was as effective as other combinations in several other trials,448,455,460,461 and more effective than placebo in three trials.449,453,454 However, the beta-blocker–diuretic combination appears to elicit more cases of new-onset diabetes in susceptible individuals, compared with other combinations.462
ACE-I = angiotensin-converting-enzyme inhibitor; ARB = angiotensin receptor blocker; BB = beta-blocker; CA = calcium antagonist; CHD = coronary heart disease; CV = cardiovascular; D = diuretic; ISH = isolated systolic hypertension; LVH = left ventricular hypertrophy; NS = not significant; RAS = renin angiotensin system; TIA = transient ischaemic attack.
The only trial directly comparing two combinations in all patients (ACCOMPLISH)414 found significant superiority of an ACE inhibitor–calcium antagonist combination over the ACE inhibitor–diuretic combination despite there being no BP difference between the two arms. These unexpected results deserve to be repeated, because trials comparing a calcium antagonist-based therapy with a diuretic-based therapy have never shown superiority of the calcium antagonist. Nonetheless, the possibility that ACCOMPLISH results may be due to a more effective reduction of central BP by the association of an RAS blocker with a calcium antagonist deserves to be investigated.398,399,464
The only combination that cannot be recommended on the basis of trial results is that between two different blockers of the RAS. Findings in ONTARGET,331,463 that the combination of an ACE inhibitor and an angiotensin receptor blocker are accompanied by a significant excess of cases of ESRD, have recently been supported by the results of the ALTITUDE trial in diabetic patients.433 This trial was prematurely interrupted because of an excess of cases of ESRD and stroke in the arm in which the renin inhibitor aliskiren was added to pre-existing treatment using an ACE inhibitor or an angiotensin receptor blocker. It should be noted, however, that BP was less closely monitored for hypotension in ALTITUDE. Two-drug combinations most widely used are indicated in the scheme shown in Figure 4.
22.214.171.124 Fixed-dose or single-pill combinations
As in previous guidelines, the 2013 ESH/ESC Guidelines favour the use of combinations of two antihypertensive drugs at fixed doses in a single tablet, because reducing the number of pills to be taken daily improves adherence, which is unfortunately low in hypertension, and increases the rate of BP control.465,466 This approach is now facilitated by the availability of different fixed-dose combinations of the same two drugs, which minimizes one of its inconveniences, namely the inability to increase the dose of one drug independently of the other. This holds also for fixed-dose combinations of three drugs (usually a blocker of the RAS, a calcium antagonist and a diuretic), which are increasingly becoming available. Availability extends to the so-called polypill (i.e. a fixed-dose combination of several antihypertensive drugs with a statin and a low-dose aspirin), with the rationale that hypertensive patients often present with dyslipidaemia and not infrequently have a high CV risk.12,13 One study has shown that, when combined into the polypill, different agents maintain all or most their expected effects.467 The treatment simplification associated with this approach may only be considered, however, if the need for each polypill component has been previously established.141
Summary of recommendations on treatment strategies and choice of drugs
ACE = angiotensin-converting enzyme; BP = blood pressure; CV = cardiovascular; OD = organ damage; RAS = renin-angiotensin system.
aClass of recommendation.
bLevel of evidence.
cReference(s) supporting recommendation(s).
6 Treatment strategies in special conditions
If the evidence favouring drug treatment in grade 1 hypertensives at low-to-moderate risk is scant (see Section 4.2.3), evidence is even weaker in white-coat hypertensives. In these individuals, no randomized trial has ever investigated whether administration of BP-lowering drugs leads to a reduction in CV morbid and fatal events. To date, information is largely limited to a subgroup analysis of the SYSTolic Hypertension in Europe (SYSTEUR) trial, which concluded that drug treatment reduces ambulatory BP and CV morbidity and mortality less in white-coat than in sustained hypertensive individuals, based on a small number of events.468
The following considerations may help orientating the therapeutic decision in individual cases. Subjects with white-coat hypertension may frequently have dysmetabolic risk factors and some asymptomatic OD (see Section 3.1.3), the presence of which raises CV risk. In these higher-risk individuals with white-coat hypertension, drug treatment may be considered in addition to appropriate lifestyle changes. Both lifestyle changes and drug treatment may be considered also when normal ambulatory BP values are accompanied by abnormal home BP values (or vice versa) because this condition is also characterized by increased CV risk.105 In the absence of additional CV risk factors, intervention may be limited to lifestyle changes only, but this decision should be accompanied by a close follow-up of the patients (including periodical out-of-office BP monitoring) because, in white-coat hypertensive subjects, out-of-office BP is often higher than in truly normotensive subjects and white-coat hypertensives have a greater risk of developing OD and to progress to diabetes and sustained hypertension (see Section 3.1.3). Consideration should also be given to the fact that, because of its high prevalence (particularly in mild-to-moderate hypertension), white-coat hypertension was presumably well represented in antihypertensive drug trials that have established clinic BP reduction as the guidance for treatment. Recommendations on treatment strategies in white-coat hypertension are listed below.
Isolated ambulatory or masked hypertension is infrequently diagnosed because finding a normal clinic BP only exceptionally leads to home or ambulatory BP measurements. When this condition is identified, however, both lifestyle measures and antihypertensive drug treatment should be considered because masked hypertension has consistently been found to have a CV risk very close to that of in-office and out-of-office hypertension.109,112,117,469 Both at the time of treatment decision and during follow-up, attention to dysmetabolic risk factors and OD should be considered since these conditions are much more common in masked hypertension than in normotensive individuals. Efficacy of antihypertensive treatment should be assessed by ambulatory and/or home BP measurements.
Summary of recommendations on treatment strategies in white-coat and masked hypertension
CV = cardiovascular; OD = organ damage.
aClass of recommendation.
bLevel of evidence.
In previous sections (4.2.5 and 4.3.3) we mentioned that there is strong evidence of benefits from lowering of BP by antihypertensive treatment in the elderly, limited to individuals with initial SBP of ≥160 mmHg, whose SBP was reduced to values <150 but not <140 mmHg. Therefore the recommendation of lowering SBP to <150 mmHg in elderly individuals with systolic BP ≥160 mmHg is strongly evidence-based. However, at least in elderly individuals younger than 80 years, antihypertensive treatment may be considered at SBP values >140 mmHg and aimed at values <140 mmHg, if the individuals are fit and treatment is well tolerated.
Direct evidence of the effect of antihypertensive treatment in elderly hypertensives (older than 80 years) was still missing at the time the 2007 ESH/ESC Guidelines were prepared. The subsequent publication of the HYpertension in the Very Elderly Trial (HYVET) results,287 comparing active treatment (the diuretic indapamide supplemented, if necessary, by the ACE inhibitor perindopril) with placebo in octogenarians with entry SBP ≥160 mmHg, reported a significant reduction in major CV events and all-cause deaths by aiming at SBP values <150 mmHg (mean achieved SBP: 144 mmHg). HYVET deliberately recruited patients in good physical and mental condition and excluded ill and frail individuals, who are so commonplace among octogenarians, and also excluded patients with clinically relevant orthostatic hypotension. The duration of follow-up was also rather short (mean: 1.5 years) because the trial was interrupted prematurely by the safety monitoring board.
RCTs that have shown beneficial effects of antihypertensive treatment in the elderly have used different classes of compounds and so there is evidence in favour of diuretics,287,449,454,470,471 beta-blockers,453,454 calcium antagonists,451,452,460 ACE inhibitors,460 and angiotensin receptor blockers.450 The three trials on isolated systolic hypertension used a diuretic449 or a calcium antagonist.451,452
A prospective meta-analysis compared the benefits of different antihypertensive regimens in patients younger or older than 65 years and confirmed that there is no evidence that different classes are differently effective in the younger vs. the older patient.444
Summary of recommendations on antihypertensive treatment strategies in the elderly
SBP = systolic blood pressure.
aClass of recommendation.
bLevel of evidence.
cReference(s) supporting recommendation(s).
In young adults with moderately high BP it is almost impossible to provide recommendations based directly on evidence from intervention trials, since outcomes are delayed by a period of years. The results of an important observational study on 1.2 million men in Sweden, initially investigated at a mean age of 18.4 years at the time of military conscription examination and followed-up for a median of 24 years, have recently been reported.472 The relationship of SBP to total mortality was U-shaped with a nadir at approximately 130 mmHg, but the relationship with CV mortality increased monotonically (the higher the BP the higher the risk). In these young men (without stiff, diseased arteries) the relationship of DBP to total and CV mortality was even stronger than that of SBP, with an apparent threshold around 90 mmHg. Approximately 20% of the total mortality in these young men could be explained by th