Letters

Management of hypertension

Ideal body weight is not realistic goal for lifestyle intervention

Patients from ethnic minorities are at greater risk

Guidelines have serious weaknesses

Guidelines do not consider workload implications in primary care

Non-medical remedies should be considered first

Evidence shows that calcium antagonists reduce cardiovascular end points in diabetic patients

Optimal target pressure is not supported by strength A evidence

"Mortality" is preferable to "major disease end points"

Trials showing no reduction in mortality do not receive same exposure

Authors' reply (Psaty and Furberg)

Authors' reply (Ramsay et al)

Ideal body weight is not realistic goal for lifestyle intervention

EDITORRamsay et al have produced a clear and authoritative document with their recent guidelines for the treatment of hypertension.¹ The section on lifestyle modification, however, includes the statement that weight loss to achieve an ideal body weight will lower blood pressure.

Although this is undoubtedly true, it undermines most recent guidelines that recognise the practical near impossibility of achieving ideal body weight in most obese subjects^2-4 and evidence that suggests that more modest (and achievable) reductions in weight of 5-10% of body weight can be effective at lowering systolic and diastolic blood pressure in the range of 4-7 and 3-6 mm Hg respectively.⁵ It should be made explicit in the guidelines that this degree of weight loss is likely to be beneficial in reducing cardiovascular risk, rather than perpetuating the myth that "ideal" body weight is a realistic goal of lifestyle modification in overweight and obese subjects.

John Wilding, senior lecturer in medicine. 
j.p.h.wilding{at}liv.ac.uk

Gareth Williams, professor. 
University Clinical Departments, University Hospital Aintree, Liverpool L9 7AL

Competing interests: Both authors have received honoraria for speaking at lectures, consultancy fees, and grant support from a number of companies that produce, or are developing, pharmacological treatments for obesity.

Patients from ethnic minorities are at greater risk

EDITORThe latest guidelines by the British Hypertension Society aim to address the incomplete detection, treatment, and control of hypertension prevalent across all sections of the community.¹ The emphasis on the assessment and reduction of cardiovascular risk rather than just the maintenance of an optimal blood pressure is to be welcomed.

One group at high risk is, however, hardly mentioned. Sections of the ethnic community, particularly Afro-Caribbeans and South Asians, are at greatly increased risk of end organ damage owing to hypertension. They also exhibit an increased incidence of concomitant cardiovascular risk factors such as diabetes and obesity.

The guidelines note that differences in average response between drug groups are related to ethnic group, but the reader is simply referred to the full report.² Although the British Hypertension Society sees ethnic background as neither a possible nor a compelling indication for any particular class of treatment, it supports the view that such patients are best treated with monotherapy with thiazide or a calcium antagonist.³ This rarely lowers blood pressure below 140/80 mm Hg. It also causes deleterious activation of the renin angiotensin system. Combining a thiazide or calcium antagonist with either an angiotensin converting enzyme inhibitor or a  blocker helps preserve neurohormonal balance. This is also far more efficacious in terms of response rates, blood pressure reduction, and, presumably, reduced cardiovascular risk.⁴

The increasingly recognised benefits of angiotensin converting enzyme inhibitors, both in terms of the reduction of cardiovascular risk and the prevention of the development and progression of diabetes, were highlighted in the heart outcomes prevention evaluation study recently reported at the European Society of Cardiology.⁵ These effects seem independent from and additive to the benefits of blood pressure reduction. Extrapolation of the evidence from the trials of  blockade after myocardial infarction and in heart failure suggests analogous benefits for  blockers. Not to use treatment based around an angiotensin converting enzyme inhibitor or a  blocker in patients at greatest cardiovascular risk is to deprive those who would be expected to benefit most.

The rationale for first line combination treatment in ethnic patients is clear, the evidence in favour persuasive, and the implications in terms of reductions in morbidity and mortality considerable. The guidelines missed an opportunity to address the disparity in treatment and outcomes in an important sector of the community. In this area, at least, it seems that the guidelines will be rapidly superseded.

Garfield Drummond, specialist registrar in clinical pharmacology. 
Department of Clinical Pharmacology, Division of Medicine and Therapeutics, Leicester Royal Infirmary, Leicester University, Leicester LE2 7LX

Competing interests: None declared.

Guidelines have serious weaknesses

EDITORIt is unfortunate that at a time when general practitioners are crying out for clear, practical, evidence based advice the BMJ should publish the British Hypertension Society's guidelines.¹ The guidelines (and their sister document²) contain many references that, taken together, show that it is time to drop the "threshold blood pressure" paradigm and adopt an approach based on estimated cardiovascular risk. The authors ignore the evidence they have amassed and promote the continued use of the "threshold blood pressure" paradigm. They are able to do this only by joining the two paradigms in a forced marriage. The result of this miscegenation is epitomised by the complex and confusing "blood pressure threshold and drug treatment in hypertension" algorithm.

The guidelines have four serious weaknesses. Firstly, they make no reference to patient preferences. Why choose a 10 year risk threshold of 15%? Are we sure that patients are happy to take a treatment for 10 years, knowing that 19 out of 20 will derive no benefit?

Secondly, they make no reference to resources, the most important of which is the workload of general practitioners. Data are available on the distribution of blood pressure and smoking,³ diabetes,⁴ and serum concentrations of cholesterol⁵ in the English population. From these data it is estimated that an average general practitioner list includes 272 patients aged under 75 years who would be eligible for treatment (table). In order to treat these patients, that general practitioner would have to devote about four hours a week to their ongoing management. Alternatively, if each primary care group ran a dedicated hypertension service, it would employ six full time hypertension specialists (data sources and calculations are available from us).

Thirdly, the guidelines provide no estimate of the benefits that might result because of this treatment. Without at least an estimate, it is impossible for a general practitioner to decide whether investing time and energy in hypertension control is worth while. If a general practitioner dedicates about 200 hours a year to managing these 272 patients, he or she will at best prevent two cardiovascular events (myocardial infarction, cardiovascular events, and new episodes of ischaemic heart disease) (table).

Fourthly, no practical advice is given on how general practitioners can identify patients at high risk for further evaluation. Under the most conservative interpretation of the guidelines, 300 patients will require annual serum lipid estimations (table).

These are not guidelines that can help general practitioners through the minefield of diagnosis and management of cardiovascular risk factors; they are guidelines that will lead to a labyrinth of screening and treatment in which even the most seasoned practitioner will get lost.

Tom Marshall, clinical lecturer in public health medicine. 
Andrew Rouse, senior lecturer in public health medicine. 
Department of Public Health and Epidemiology, University of Birmingham, Birmingham B15 2TT

Competing interests: None declared.

Guidelines do not consider workload implications in primary care

EDITORThe new joint British societies' recommendations on prevention of coronary heart disease seem to have been introduced without any consideration of workload implications for general practice. ^{1 2} By lowering the threshold for treatment they have created huge numbers of new hypertensive patients. Each patient will require assessment, workup, and several appointments before control is acceptable to both clinician and patient.

A conservative estimate is that, on average, newly identified hypertensive patients will need four appointments a year, especially in the first 12 months. How many extra patients each general practitioner and practice nurse will be expected to see each week is difficult to quantify. In addition to new hypertensive patients requiring treatment there are established patients who are no longer adequately controlled and new hyperlipidaemic patients.

A partial solution is that practices that are committed to the guidelines be given the extra funding needed (mainly a limited increase in the hours of practice nurses). Otherwise surgeries will spend increasing amounts of time and resources on patients having preventive care at the expense of those who wish to seek advice because they feel ill. The joint societies seem to have insufficiently consulted general practitioners when developing their guidelines. General practitioners tend to blame their excessive and sometimes demoralising workloads on "high patient demand." It may be that on this occasion "high professorial demand" is to blame.

John Eisenberg, general practitioner. 
Macklin Street Surgery, Derby DE1 1JX derby.gpvts{at}virgin.net

Competing interests: None declared.

Non-medical remedies should be considered first

EDITORWith reference to Psaty and Furberg's editorial on managing hypertension in the United Kingdom,¹ I am amazed that there was no attempt to bring down people's blood pressure by reducing salt intake and encouraging daily strengthening, cardiovascular, and stretching exercise. Whereas the medical profession clearly is trained to dish out drugs there is a body of scientific evidencesee www.pritikin.comthat shows that diet and exercise are very effective at reducing hypertension.

I wish Britain's medical profession would become more aggressive in preventive medicine and in treating the whole person. It is not a question of money but one of attitude. A Californian on a somewhat misty visit to the United Kingdom (east London suburbs and Yorkshire), I was the only one out jogging before work, and people looked at me as if I was a freak from another planet. Many people have hypertension.

Angela Hey, consultant. 
Areva International, Portola Valley, CA 94028, USA amhey{at}areva.com

Competing interests: None declared.

Evidence shows that calcium antagonists reduce cardiovascular end points in diabetic patients

EDITORThe British Hypertension Society guidelines for the management of hypertension provide an excellent summary of the recently published randomised controlled trials and provide a template for all doctors who treat patients with hypertension.¹ The accompanying editorial by Psaty and Furberg did no justice to the guidelines.² I take particular issue with the line in the editorial stating that diabetes would have been a compelling or possible contraindication to calcium antagonists.

The studies cited by Psaty and Furberg in diabetic patients include the appropriate blood pressure control in diabetes trial, the premature termination of which was heavily criticised owing to the small number of events on which this decision was based and the possibility that the observations were due to chance.³ The second study cited was the fosinopril versus amlodipine cardiovascular events trial, which also received considerable criticism because its findings have been based on a small number of events. Furthermore, 56.9% of patients in this study required both a calcium antagonist and angiotensin converting enzyme inhibitors to control blood pressure, and fewer events were seen in the patients having dual treatment, suggesting that this combination of treatments is appropriate in diabetic patients.

The results of these studies have been refuted by the studies into systolic hypertension in Europe and optimal treatment for hypertension, both of which have shown highly beneficial effects of treatment based on calcium antagonists for cardiovascular events in diabetic subjects. ^{4 5} The evidence from randomised controlled studies thus shows that calcium antagonists reduce cardiovascular end points in diabetic patients, and the British Hypertension Society guidelines correctly state that they are not contraindicated in diabetic subjects.

Tahseen A Chowdhury, lecturer in medicine. 
Heartlands Diabetes Centre, Birmingham Heartlands Hospital, Birmingham B9 5SS T.A.Chowdhury{at}bham.ac.uk

Competing interests: None declared.

Optimal target pressure is not supported by strength A evidence

EDITORThe British Hypertension Society's guidelines should be commended for taking cost effectiveness into account, for example, in recommending low doses of cheap thiazides as first line drug treatment, or being on the conservative side in recommending the use of expensive statins.¹ Furthermore, it is commendable that the evidence based guidelines for the north of England are used for grading the evidence supporting the recommendations.

The evidence supporting the suggested target blood pressures during antihypertensive treatment has been given the strongest recommendation (A), indicating that the evidence stems from meta-analysis of randomised controlled trials or from at least one randomised controlled trial.

The recommendation that the optimal target blood pressure in non-diabetic people is <140/85 mm Hgstems from the hypertension optimal treatment (HOT) trial.² This particular finding was, however, not the result of a randomised controlled trial and should not be given strength A recommendation. The intention to treat analysis in the HOT trial was negative. The difference in any outcome measure between the three target groups (90 mm Hg, 85 mm Hg, or 80 mm Hg) was not significant. The patients achieving the "optimal" 82.6 mm Hg are not the same as those who were randomised to the lowest diastolic blood pressure but are a mixture of patients from all three groups, probably dominated by those patients who responded most effectively to the intervention. The analysis of the achieved blood pressure is purely observational, treating the total study population as one single cohort, and should therefore be given a strength C recommendation.

Hogne Sandvik, general practitioner. 
Department of Public Health and Primary Health Care, University of Bergen, N-5009 Bergen, Norway hogne.sandvik{at}isf.uib.no

Competing interests: None declared.

"Mortality" is preferable to "major disease end points"

EDITORIn their editorial on the new British Hypertension Society guidelines for managing hypertension, Psaty and Furberg demand evidence that recommended drug treatments reduce major disease end points, such as myocardial infarction, stroke, and heart failure.¹ Although this would be an improvement on evidence that they lowered blood pressure, it is insufficient to justify their use. If a drug were shown to reduce cardiac deaths but had no effect on all cause mortality, I would be reluctant to take it, particularly if the displaced mortality was in the form of increased suicide and homicidepresumably as a consequence of the drug's side effects.

In reading reports of randomised trials of drugs claiming success, I first look for the effect on all cause mortality. If that is not significantly negative or, worse still, not even revealed, I pass on. It saves a lot of reading.

The area of medicine in which treatment has become completely detached from clinical end points is HIV infection. Drugs with the most devastating and frequent side effects are now given to symptomless mothers and babies for no better reason than that they reduce the incidence of scoring positively on an HIV test.

Michael Stewart, statistical consultant. 
2 Lesley Court, Strutton Ground, London, SW1P 2HZ

Competing interests: None declared.

Trials showing no reduction in mortality do not receive same exposure

EDITORThe expert guidelines we are advised to follow are not based on the people I look after as a general practitioner, and the details given in the guidelines are not patient friendly.¹ I would like to know the numbers needed to treat, and when the guidelines say deaths prevented I assume they mean deaths postponed, as we all die. How long are the deaths prevented for, what do the patients die of if it is not a consequence of their hypertension, and could we not have prevented that also?

Around September 1999 a fact sheet from the British Heart Foundation advised that blood pressure should be reduced to below 125/75 mm Hg in a diabetic patient with proteinuria. I believe that if I prescribed sufficient medication the side effects would probably be intolerable. I also question whether the number needed to harm would be lower than the numbers needed to treat. This information is not easily available. All trials seem to add yet more treatment when most of them are conducted on single interventions. The wisdom follows that they all can be added to give "extra" benefit. Sometimes I wonder if after a myocardial infarction the added value of aspirin, a  blocker, a statin, an angiotensin converting enzyme inhibitor, and spirinolactone must make death impossible.

The most telling comment in the editorial accompanying the article was that no declaration of interest was published.² The early hypertension trials conducted by the Medical Research Council showed a moderate benefit in some patients in treating hypertension. Instead we now have a bandwagon of multiple interventions at great cost. In my experience as a general practitioner actively participating in reaching hypertension targets, considerable anxiety and side effects are being generated by this drive. Over the years, there have been trials showing no reduction in mortality with various antihypertensive agents. Why do these not receive the same exposure?

N J Sharvill, general practitioner. 
Balmoral Surgery, Deal, Kent CT14 7AU RoGi11111{at}aol.com

Competing interests: Dr Sharvill is taking part in the PRICCE project, in which hypertension targets are one area of study; if his practice meets all the area targets for diseases there will be a payment from the health authority.

Authors' reply (Psaty and Furberg)

EDITORIn our editorial on the British Hypertension Society guidelines, we noted that the recommendations for drug treatment according to various compelling or possible indications were consensus based rather than evidence based. Although we do not advocate this approach, we did observe that if one were to favour the proliferation of items in a special indications table, diabetes would have been either a compelling or a possible contraindication to calcium channel blockers. We cited two small randomised clinical trials that directly compared calcium channel blockers and angiotensin converting enzyme inhibitors in patients with type 2 diabetes. ^{1 2} In both, the use of calcium channel blockers was associated with a higher risk of cardiovascular events than treatment with angiotensin converting enzyme inhibitors.

Chowdhury responds that the results of these studies have been utterly refuted by the systolic hypertension in Europe (Syst-Eur) and the hypertension optimal treatment (HOT) studies. ^{3 4} We disagree. The Syst-Eur paper, with a total of five stroke and seven cardiac events among diabetic patients in the active treatment group, represents a post hoc subgroup analysis that should be interpreted cautiously. Syst-Eur is a placebo controlled trial. Placebo controlled trials answer the question whether we should treat a condition such as isolated systolic hypertension. The benefits of treating this have been clear since 1991.⁵

Randomised trials that directly compare two treatments are required to determine whether one is better than another. Although cardiovascular events in the small comparative trials ^{1 2} were a secondary end point, both comparative trials suggest that in patients with type 2 diabetes, angiotensin converting enzyme inhibitors are superior to calcium channel blockers in terms of preventing cardiovascular events.

In terms of antihypertensive therapy, the HOT study was robustly null.⁴ The differences in blood pressure among groups were small, and for the primary end point in intention to treat analyses there were no differences among the groups. We agree with Sandvik that the observational analyses related to blood pressure within HOT should be graded C rather than A.

Stewart prefers the outcome of total mortality. Few trials of antihypertensive treatments have been powered for the end point of total mortality. Cardiovascular and cerebrovascular events are often devastating illnesses that affect both quality and duration of life. Interpreting the morbidity outcomes in the context of total mortality is important. When mortality and morbidity outcomes go in different directions, the interpretation becomes problematic. Mortality and morbidity outcomes that go in the same direction are generally reassuring.

Bruce M Psaty, professor. 
Departments of Medicine, Epidemiology and Health Services, University of Washington, Seattle, WA 88101, USA

Curt D Furberg, professor. 
Department of Health Sciences, Wake Forest University School of Medicine, Winston-Salem, NC, USA

Competing interests: The research reported in the editorial was supported in part by grants from the National Heart, Lung, and Blood Institute and the National Institute on Aging. Professor Psaty is a Merck/SER clinical epidemiology fellow (sponsored by the Merck Foundation and the Society for Epidemiologic Research, Baltimore). He has served on the events committee for a trial funded by Wyeth Ayerst. Professor Furberg receives research funding from Pfizer and Wyeth Ayerst and is consultant to Bristol Myers Squibb.

Authors' reply (Ramsay et al)

EDITORWe accept Wilding's proposal that we should have recommended weight loss towards (rather than to achieve) ideal body weight in line with the graded benefits in reduction of blood pressure which result from lesser degrees of weight loss as reported in the full version of the guidelines.¹

Drummond is correct that there is a high prevalence of hypertension and cardiovascular disease in African-Caribbean and South Asian communities, but space constraints prevented discussion of specific subgroups (for example, elderly or diabetic patients) except in the full version of the paper.¹ He was incorrect to say that we recommend monotherapy for any specific group of patients. On the contrary, we emphasised the likelihood that more than one drug was likely to be required to achieve optimal control of blood pressure in most patients. Moreover, Drummond's certainty regarding the benefits of angiotensin converting enzyme inhibition for all patients is unfounded, as is his assertion that the still unpublished benefits of angiotensin converting enzyme inhibition in the heart outcomes prevention evaluation (HOPE) trial are necessarily independent from and additive to the benefits of a reduction in blood pressure. His conclusions are not supported by the results of the Swedish trial in old patients with hypertension 2 (STOP 2 trial).² Furthermore, his alternative proposal that treatment should be based around a  blocker is not commensurate with meta-analysed trial evidence of hypertension management in elderly people.³

Turning to the serious weaknesses identified by Marshall and Rouse: successful application of the guidelines will prevent not only coronary events but also strokes, and treated patients will also develop less atherosclerosis, left ventricular hypertrophy, dementia, and heart failure. We believe that patients, if fully informed rather than confused by inappropriate statistics, would choose treatment under these circumstances.

The implications of the guidelines for resources were discussed under "Implementation," and the British Hypertension Society has established an implementation group with input from general practice and nursing to address this important issue. Nevertheless, we strongly believe that guidelines should reflect best practice according to the best available evidence. These guidelines were developed mindful of the poor record of the NHS in preventing premature cardiovascular morbidity and mortality, and best evidence clearly shows the benefit of the recommended interventions.

To ensure treatment is rationally targeted at those with highest calculated cardiovascular risk, we provided practical advice for the identification of high risk patients in the section headed "Evaluation of hypertensive patients." This approach has been widely applauded and ensures the most effective use of resources. The measurement of lipids is correctly incorporated into this and all other accurate cardiovascular risk assessments.

Eisenberg's assertion that the guidelines have created huge numbers of hypertensive patients is not valid. We have merely highlighted the fact that blood pressure is a continuous variable in the calculation of cardiovascular risk. Those at highest risk will thus benefit from treatment at lower thresholds. We also hold the view that primary prevention of cardiovascular disease is preferable to, if less dramatic than, treating established heart disease and stroke, which all too often is too late. We acknowledged an important role for general practitioners and nurses in achieving this objective and the development of the guidelines benefited from their input.

Hey has apparently not read the guidelines summary since both exercise and salt restriction were discussed and highlighted in the article, although we did not specifically endorse jogging in east London suburbs or Yorkshire as being particularly beneficial.

We support Chowdhury's criticism of the editorial by Psaty and Furberg that accompanied the guidelines, and we agree with Sandvik's careful and incisive comments that the evidence from the hypertension optimal treatment (HOT) trial is less than ideal. We reassert our belief, however, that the HOT trial provided some guidance and reassurance and the best evidence to date on targets for blood pressure.

We hope that Stewart's approach to interpreting trials solely on the basis of all cause mortality results, irrespective of power consideration and a priori hypotheses, will not be emulated by others. Furthermore, in response to Sharvill's report of his experience, we agree that side effects of treatment can be tiresome and troublesome, but in our clinical practice they are perhaps less troublesome than strokes, heart attacks, and heart failure owing in part to suboptimal management of hypertension and associated cardiovascular risk factors. We are not aware of the trials that he says showed no reduction in mortality with various antihypertensive agents.

Lawrence E Ramsay, professor of clinical pharmacology and therapeutics. 
University of Sheffield, Sheffield S10 2TN

Bryan Williams, professor of medicine.
B Williams bw17{at}leicester.ac.uk

John F Potter, professor of medicine for the elderly. 
University of Leicester School of Medicine, Leicester Royal Infirmary, Leicester LE2 7LX

G Dennis Johnston, professor of clinical pharmacology. 
Queen's University of Belfast, Belfast BT7 1NN

Graham A MacGregor, professor of cardiovascular medicine. 
Department of Medicine, St George's Hospital, London SW17 0RE

Lucilla Poston, professor of fetal medicine. 
Department of Obstetrics and Gynaecology, St Thomas's Hospital, London SE1 7EH

Neil R Poulter, director, cardiovascular studies unit. 
Imperial College School of Medicine, London W2 1NY

Gavin Russell, consultant renal physician. 
North Staffordshire Royal Infirmary, Stoke on Trent ST4 7LN

^2∥ B Williams bw17{at}leicester.ac.uk

Competing interests: None declared.