Intended for healthcare professionals


Guidelines for managing raised blood pressure

BMJ 1996; 313 doi: (Published 13 July 1996) Cite this as: BMJ 1996;313:64
  1. Rodney T Jackson,
  2. David L Sackett
  1. Associate professor Department of Community Health, University of Auckland, Auckland, New Zealand
  2. Professor NHS Research and Development Centre for Evidence Based Medicine, Oxford Radcliffe NHS Trust, Oxford OX3 9DU

    Evidence based or evidence burdened?

    Randomised controlled trials have consistently shown that lowering blood pressure by about 10-12 mm Hg systolic and 5-6 mm Hg diastolic reduces the relative risk of stroke by about 40% and of coronary disease by about 15%.1 This relative reduction in risk is similar whatever the blood pressure before treatment and the absolute risk of cardiovascular disease. Moreover, the reduction in risk occurs surprisingly quickly; the average time from the start of treatment to a significant impact on major cardiovascular outcomes is only about two and a half years.

    This convincing evidence of a large and rapid relative benefit from treatment has led national and international bodies to recommend that a substantial proportion of adults be considered for long term drug treatment. In 1993 at least one international and three national guidelines on managing raised blood pressure were published.2 3 4 5 All were based on the same data, yet when Fahey and Peters (p 93) applied the various guidelines to a typical population of patients treated for hypertension in British general practice, only about a third of the patients met the treatment criteria of all four guidelines.6

    The authors of all of these guidelines (including one of us, who played a role in the New Zealand guidelines) did their best to make them evidence based. However, we would suggest that the guidelines overemphasise clinically inappropriate relative measures of the benefits of treating raised blood pressure.

    The movement towards evidence based medicine was in part stimulated by the observation of substantial variations in medical practice; Fahey and Peters have now shown similar variations in current guidelines for managing raised blood pressure. Evidence based medicine is “the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients”7 and involves integrating the best external evidence with individual clinical expertise. The best evidence is defined as clinically relevant research, “especially patient centred clinical research into the accuracy and precision of diagnostic tests, the power of prognostic markers, and the efficacy and safety of therapeutic, rehabilitative, and preventive regimens.”7 Evidence based decisions should also be mindful of competing claims on resources. We would suggest that guidelines should provide the necessary evidence based data to empower clinicians to make informed decisions. One of the most helpful tools that integrates the evidence with clinical expertise is the use of “numbers needed to treat to prevent an unwanted event or cause an unwanted side effect.”8 9

    Calculating the number needed to treat requires the integration of external evidence (on the accuracy of diagnosis, the absolute risk before treatment, and the relative benefit of intervention) with clinical expertise in the form of judgment on the applicability of the external evidence to specific patients. We think that clinicians practising evidence based medicine should also explicitly consider the resource implications of their decisions (the cost effectiveness) at least at the level of their individual practice.

    All the guidelines referred to above deal appropriately with the evidence from randomised controlled trials by examining overviews of the data, which have generally provided estimates of relative risks and benefits of treatment.1 10 However, we suggest that, for them to guide evidence based clinical care, they need to include information on the absolute risks and benefits of interventions, as recently reported by Mulrow et al.11 Although all the guidelines acknowledge the importance of the pretreatment prognosis in determining the absolute benefits of treatment, only one of the guidelines provides the necessary information to enable clinicians to estimate the expected event rate for patients and the appropriate numbers needed to treat.2 Moreover, although all of the guidelines mentioned the importance of cost effectiveness, none provided doctors with the information needed to estimate the absolute (rather than relative) cost effectiveness of a treatment. Given that the absolute benefits and risks of treatment (however weighted by the values and resources of patients, practitioners, or payers) are crucial for making informed decisions,12 we suggest that the current guidelines have failed to provide all the evidence necessary for clinical decision making.

    All four guidelines suggest threshold levels of diastolic blood pressure above which treatment should be started in all middle aged patients if the levels are sustained after non-pharmacological treatment. Three different threshold levels are recommended: 90 mm Hg,3 95 mm Hg,5 and 100 mm Hg.2 4 Three of the guidelines do not provide any explicit information, such as the number needed to treat or its precursors, to justify the threshold levels for treatment.

    If we combine estimates of prognosis, based on data from cohort studies,13 14 with estimates of the relative benefits of a reduction in diastolic blood pressure of 5-6 mm Hg, based on data from trials,1 it is possible to calculate numbers needed to treat that are relevant to individual patients. For example, for a 60 year old male smoker with a pretreatment diastolic blood pressure of 90 mm Hg, a ratio of total cholesterol to high density lipoprotein cholesterol of 6.5, and a normal electrocardiogram, the number needed to treat for one year (to prevent one major cardiovascular event) would be 75. In comparison, it would be 320 for a non-smoking 50 year old woman with a considerably higher diastolic blood pressure (100 mm Hg), a ratio of total cholesterol to high density lipoprotein cholesterol of 5, and a normal electrocardiogram. These differences in the number needed to treat reflect the major influence of cardiovascular risk factors other than raised blood pressure on prognosis12 13 14 and therefore on the absolute benefits of treatment.

    Developers of guidelines, ourselves included, have been overburdened by evidence which gives undue emphasis to the relative risks of raised blood pressure and the relative benefits of reducing blood pressure. We think it is time to consider basing guidelines explicitly on clinically more useful absolute measures of the effects of treatment. Indeed, we suggest that the clinical credibility and success of the guidelines process depends on it.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.
    View Abstract