Background

According to the author Paul Valéry, ‘everything simple is false, everything complex is unusable’. This maxim sums up the difficulty facing any team tasked with developing a guideline: if they aim for simplicity, they will probably leave out many special cases; and if all cases are covered, the result becomes so complex that nobody can understand how it works. In other words, there are no miracle solutions or magic formulae. Guidelines on the management of hypertension are no exception. The 2-year period from 2013 to 2014 saw the publication of new hypertension guidelines in France by the French Society of Hypertension (Société Française d’Hypertension Artérielle [SFHTA]) , in Europe by the European Society of Hypertension (ESH) , and in the USA by the Eighth Joint National Committee (JNC-8) . In the UK, the guidelines were updated in 2011 by the National Institute for Health and Care Excellence (NICE) . These guidelines differ in both content and form. As the leading cause of death worldwide, the management of this condition is as important as ever, yet the advice differs between countries and continents. This critical appraisal summarizes the key recommendations of these guidelines, highlighting points on which they agree and disagree, as well as unique features.

A leaflet or a book?

The ideal length for a guideline is a thorny issue. A short document is easy to read and use, but inevitably simplistic, while an intentionally exhaustive document that exceeds 50 pages is difficult to use in routine practice. The French and European guidelines sit at opposite ends of this spectrum: the French guideline is just four pages long and contains 39 references, while the European document runs to 77 pages and contains 735 references. The USA and UK guidelines are 14 and 27 pages long, respectively. The SFHTA, JNC-8 and NICE had evidently taken the view that their guidelines should be short enough to ensure that they are read. The European guideline, on the other hand, is intended as a reference work, and addresses every situation encountered in the initial management, workup and treatment of hypertension, including techniques currently under evaluation, such as renal denervation. This detailed document nevertheless includes many tables summarizing the positions adopted. It is clearly intended for specialists. In contrast, the French guideline targets generalists and its simplicity is a stated aim; it is divided into short, deliberately didactic subsections. This guideline does not address specific situations, but proposes a general plan of action for managing hypertension in adults. Thus, for example, resistant hypertension is dealt with in a separate document that differs in form and length.

Should recommendations be graded?

As clinical practice is nowadays based on evidence and as little as possible on opinion, the selection of documents and articles on which to base recommendations is a crucial stage; it makes it possible to weigh up expert opinion and naturally results in a grading of recommendations.

For the USA guideline, a panel of methodologists selected the articles most relevant to the questions posed, and passed them on to the writing panel. The selection criteria were extremely restrictive, in that only publications referring to clinical trials in hypertension were retained.

The French guideline lists few references, including former guidelines (issued in 2005 by the French National Authority for Health [Haute Autorité de santé; HAS]) , guidelines from other countries or scientific societies (NICE, ESH) and meta-analyses. The USA and European recommendations are graded in the usual manner, ranging from strong recommendations, based on clinical trials of high methodological quality that included a large number of patients, down to more empirical expert opinions, with a series of intermediate levels in between. The French and UK recommendations are not graded, as the intention was to convey a straightforward message to clinicians.

Do levels of evidence matter to typical prescribers who are not hypertension specialists? It would be useful to answer this question by conducting a survey among the clinicians targeted by the guideline. The European guideline strongly recommends renin-angiotensin system blockers in unilateral renal artery stenosis (class I recommendation), but tempers it with a B grading for its level of evidence, signaling that it is supported by clinical trials of limited size. The possibility of adjusting recommendations through grading appears better suited to evidence-based rather than opinion-based medicine. The danger, however, is that the message can become overly complex and convoluted.

Diagnosing hypertension: out-of-office measurements are best!

A paradox highlighted by all of the guidelines is that office blood pressure measurement is undoubtedly the worst method for diagnosing and monitoring hypertensive patients. Alternative methods – ambulatory blood pressure monitoring (ABPM) or self-measurement – must now be considered to have an important if not essential role. The SFHTA recommends out-of-office measurements to confirm the diagnosis of hypertension, but does not specify whether this refers to self-measurement or ABPM. However, the SFHTA and the French Committee for the Prevention and Control of Hypertension (CFLHTA) have been promoting the use of self-measurement for several years. In fact, the French National Health Insurance Fund (CNAM) is going to offer home blood pressure monitors to primary care physicians as an aid to the diagnosis of hypertension. The aim of this initiative is to distinguish true hypertension from white coat syndrome, which leads to overdiagnosis, and to reduce the cost to society of unnecessary treatment. However, recent studies have revealed that the conditions under which self-measurement is performed, and therefore its reliability, can be far from satisfactory. Although patients generally master the technical aspects of self-measurement, the major advantages of this method are undermined by failure to take enough readings, to take them at the required frequency or to record and submit them, which ultimately means that the results are unusable and have little impact on blood pressure control. If self-measurement is to make a positive contribution to the management of hypertensive patients, physicians must therefore receive training and patients must be properly informed.

The UK guideline recommends that suspected hypertension should be confirmed by 24-hour ABPM. ABPM appears to be a more sensitive and specific method for diagnosing hypertension than multiple office blood pressure measurements at repeated appointments. NICE also claims that ABPM is more cost-effective than home self-measurement, and that self-measurement is more effective than office measurements, but less effective than ABPM. It is thought, however, that 5–10% of patients do not tolerate ABPM, and that no automated systems are suitable for patients with atrial fibrillation. According to NICE, the use of ABPM results in fewer visits to the physician and earlier initiation of treatment for patients who require therapy. Including the cost of buying the equipment, the NICE expert committee calculated that it would be cost neutral in 2 years and produce savings after 3 years.

According to the ESH, alternatives to office blood pressure measurement (self-measurement or ABPM) are a useful adjunct to conventional blood pressure measurement, which remains the gold standard for screening for, diagnosing and monitoring hypertension.

The USA guideline does not address the diagnosis of hypertension. This document focuses on a limited number of objectives: to define the blood pressure thresholds above which therapy should be initiated, treatment goals and treatment strategies. A recap of the conditions under which blood pressure is measured, particularly the standard sphygmomanometer method, is included in the supplemental content.

Blood pressure goals

The USA guideline revives the old debate about the blood pressure thresholds above which treatment is justified, and the blood pressure targets to achieve through treatment, which in reality amount to the same thing. This guideline, drafted by a panel of experts led by a professor of family medicine, focuses on three questions: what are the blood pressure values above which treatment initiation leads to a reduction in cardiovascular events; what blood pressure targets should be achieved in order to reduce cardiovascular events; and what is the harm-benefit balance of the various antihypertensive drug classes in the treatment of hypertension? The rationale for developing this guideline is as follows: the optimal target blood pressure for antihypertensive treatment is unclear, including in subgroups (black patients, patients with diabetes or chronic kidney disease); setting blood pressure goals too low has led to overmedication of hypertensive patients and an increase in adverse effects; blood pressure goals should probably be raised, in order to limit treatment intensification. New treatment targets were defined, based on publications selected by a group of methodologists, using strict criteria, and then forwarded to the writing panel. Following the recent publication of two Japanese clinical trials showing that reducing systolic blood pressure (SBP) to 136–137 mmHg was no better than a goal of < 142–145 mmHg in terms of reducing the incidence of the complications of hypertension, the USA guideline recommends a target of < 150/90 mmHg for patients aged > 60 years. The blood pressure goal for patients aged < 60 years remains < 140/90 mmHg. In patients with diabetes or chronic kidney disease, the blood pressure goal has been increased from ≤ 130/80 mmHg to < 140/90 mmHg. A recent study estimated the proportion of hypertensive adults in the USA affected by the change in these recommendations . The new blood pressure goals would reduce the number of persons eligible for antihypertensive therapy by 5.8 millions. The proportion of hypertensives meeting blood pressure goals would increase from 40.6% (based on the JNC-7 targets) to 56.5% (based on the newly determined targets) .

The French guideline gives more discretion to clinicians. The target blood pressure it recommends, including for patients with diabetes or chronic kidney disease, is an SBP between 130 and 139 mmHg and a diastolic blood pressure (DBP) < 90 mmHg, but lower goals can be proposed for certain patients, after consulting a specialist.

The European guideline draws a distinction between ‘high normal’ blood pressure and three grades of hypertension. A two-way table is used to decide at which stage patients should be treated, where the columns correspond to blood pressure and the rows to the number of risk factors, the presence of target organ disease, chronic kidney disease or diabetes and the patient’s status (e.g. secondary prevention). The different colours indicate the overall cardiovascular risk level, where green corresponds to the lowest risk and red to the highest. This table may appear complicated and relatively unusable in practice. However, it highlights the value of taking into account the intensity of the risk factor considered – in this case hypertension severity – and of estimating overall cardiovascular risk when deciding whether to treat patients. This observation is based on the results of clinical trials conducted in patients at ‘high cardiovascular risk’, in whom it was shown that a statin or renin-angiotensin system blocker significantly reduced the incidence of cardiovascular complications. For patients categorized as having a high cardiovascular risk (e.g. a diabetic with a reduced glomerular filtration rate), it is only logical to prescribe treatment proven to reduce overall cardiovascular risk, specifically a renin-angiotensin system blocker, regardless of whether they are also hypertensive. It is worth mentioning that the HOPE study , which, among other things, lent support to the ESH’s paradigm of the benefit of treating patients with high cardiovascular risk, was rejected by the JNC-8 because the patients enrolled in this trial were not all hypertensive.

France’s 2005 HAS guideline also contained a two-way table, a similar (but less complicated) version of the ESH table, but it was omitted from the brief overview of the 2013 SFHTA recommendations. The concept was no doubt considered unhelpful for generalists.