Intended for healthcare professionals

Editorials

Thresholds for normal blood pressure and serum cholesterol

BMJ 2005; 330 doi: https://doi.org/10.1136/bmj.330.7506.1461 (Published 23 June 2005) Cite this as: BMJ 2005;330:1461
  1. Steinar Westin, general practitioner and professor of social medicine (steinar.westin{at}ntnu.no),
  2. Iona Heath, general practitioner
  1. Department of Public Health and General Practice, Norwegian University of Science and Technology, Medical Technical Research Centre, N-7489 Trondheim, Norway
  2. Caversham Group Practice, London NW5 2UP

    Lower thresholds mean that 90% of people over 50 years are identified as patients

    For several years, disagreement has been growing in health services about ever lower thresholds for “normal” blood pressure and serum cholesterol, the most common biological risk factors for cardiovascular disease. This was most publicly expressed in 1999, when more than 800 doctors, pharmacists, and scientists from 42 countries signed an open letter to the then director general Gro Harlem Brundtland, outlining fears that the World Health Organization's new hypertension guidelines would result in increased use of antihypertensive drugs, at great expense and for little benefit.1 2 The simplistic linear structuring of many research questions and the extrapolation of research results over prolonged and unstudied time periods produce guidelines that make many doctors, and particularly general practitioners, feel uneasy about the high proportion of their patients who are being labelled as sick.

    General practitioners are aware of the side effects of undue medicalisation and tend to question the external validity of randomised controlled trials under experimental conditions.3 They also have to consider the opportunity costs of intervening to alter the risk profile of large numbers of healthy people and the time and resources that this takes away from people who are currently sick.4 This does not mean that general practitioners question the efficacy or cost effectiveness of drug treatment for persons with overt atherosclerotic disease or at unquestionably high risk. The uneasiness is about primary prevention being conceived increasingly as a strategy implying individual risk identification and questionable labelling of disease.

    The unrest provoked by the earlier WHO guidelines intensified with the publication of the latest European guidelines on prevention of cardiovascular disease in clinical practice in 2003.5 These suggest blood pressure above 140/90 mm Hg, with no age correction, and serum cholesterol of 5 mmol/l as the appropriate thresholds for intervention. The guidelines consider other risk factors and recommend a range of lifestyle changes alongside drug treatment but the bottom line is that the doctor is expected to inform the patient that these measurements mean that he or she is at increased cardiovascular risk regardless of the management proposed. In other words, a disease label is to be attached to the patient.

    A paper by Getz et al outlines the results of applying these European guidelines to a total county population in Norway.6 The Nord Tröndelag health survey provides data on blood pressure and cholesterol for some 62 000 adults aged 20-79 in 1995-7. The figure from that paper, reproduced here, shows the proportion of each age group that would be identified as being at “increased risk.” The proportions are disturbingly high even if only the recommended thresholds for blood pressure are applied. However, if the threshold for normal cholesterol is also applied half of the population would be considered at risk by the early age of 24 years. By the age of 49, this proportion rises to 90%. As much as 76% of the total adult population would be considered at “increased risk.” The current life expectancy at birth in Norway is 78.9 years, making it one of the longest living populations ever.7 This compares with a life expectancy at birth in the United Kingdom of 78.1 years, which implies that even higher proportions may be found if the study is repeated in other populations.

    Fig 1
    Fig 1

    Point prevalence of individuals aged 20-79 (men and women combined) with unfavourably high blood pressure or cholesterol concentrations, as defined by the 2003 European guidelines on cardiovascular disease prevention in clinical practice6

    Several issues need to be considered if such a large part of the population is to become a target for individual and lifelong risk interventions. Firstly, the potential benefits for treated patients become less at lower risk levels, indicated by increasing numbers needed to treat,8 whereas rates of side effects remain similar regardless of the level of risk. Secondly, evidence for the long term effectiveness of treatment is lacking, since data from short term studies are being extrapolated over the whole of the remaining lifespan. We have recently had sobering news regarding drugs that were thought to reduce cardiovascular events.9 Thirdly, the side effects of drugs tend to be under-reported as well as under-published, as with some of the cholesterol lowering drugs, and α receptor blockers. Most randomised controlled trials are powered for efficacy end points and therefore grossly underpowered to detect anything more than common adverse events. Fourthly, we have very limited evidence on the effects of preventive drug treatment when several drugs are used to treat different risk factors simultaneously, and unfavourable drug interactions are an increasing problem. Fifthly, we have far too little understanding of the psychological impact and the wider health consequences of being labelled at risk.10 Finally, the huge cost of pharmaceutical interventions for an ever greater proportion of the population has the potential to destabilise publicly funded healthcare systems in even the richest nations. Such considerations are urgent as the guidelines from the European Society of Cardiology are in the process of being implemented and the quality and outcomes framework of the new general practitioner contract in the UK can be seen as part of this implementation.

    Footnotes

    • Competing interests None declared.

    References

    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    View Abstract