Rapid Responses to:

LETTERS: Bryan Williams Guidelines from the British Hypertension Society: Authors' reply BMJ 2004; 329: 570-b-571-b [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

The Sense Of Guidelines

Nigel S de Kare-Silver   (6 September 2004)

Risky Relative Risk

Des Spence   (7 September 2004)

Guidelines: how about all-cause mortality?

Eddie Vos   (8 September 2004)

Grasping the thorns of the Rose Paradox.

Peter Davies   (14 October 2004)

The Sense Of Guidelines 6 September 2004
Nigel S de Kare-Silver, GP Gladstone Medical Centre, 5 Dollis Hill Lane, London NW2 6JH Send response to journal: Re: The Sense Of Guidelines	One must take pick up on Campbell and Murchie’s1 review article, ‘Treating hypertension with guidelines in general practice’ and take issue with William’s ‘authors reply’2 discussion on target cholesterol measurements in the same journal issue on many levels. Campbell and Murchie describe the poor implementation of guidelines and suggest that one key problem may include the guideline authors’ inability to reconcile medication side effects against target measurements and the guideline documentation becoming too long to be digestible. This is undeniable. Guideline authors should remember their target readership are doctors of medicine, one of the most intelligent groups within society, and if their guidelines fail to make impact or sense on this profession it is they who are at fault for their poor writing, not their clinical colleagues for a lack of skill or ability. Perhaps other and more important issues are that general practice is too careful, too critical and too scrutinizing to implement everything it is presented with, in journals or other literature. Campbell and Murchie’s own quotation of Primatesta et al’s 3 assertion that one third of the adult population over 64 are elderly and hypertensive immediately brings the reader to recognise an error and engender a tendency to dismiss the rest of the article. Primatestas’ criteria for diagnosis of hypertension here fail to take account of the British Hypertension Society’s own guidelines and will include large numbers who are not hypertensive. Headline figures such as this may draw attention of the tabloids to the importance of disease and diagnosis but are met with disdain and dismissal by those in practice. Williams is equally at fault for advocating ALLHAT as being of significant importance in the creation of lipid management guidelines. The criticism of this trial is fierce and wide with acknowledged methodological errors and poor compliance in its population. ASCOTT-LLA equally has its faults: of short time span there is just not the evidence there that the targets advocated match up to improved outcomes, the early termination of the trial and its statistical methodology are unequivocally weak 4. The whole concept of ‘reduction of baseline cholesterol’ meets with derision in practice. If after a hundred years we cannot be certain of blood pressure measurement how can we have any confidence in a ‘baseline’ cholesterol. Does this mean a fasting cholesterol if so after, 6, 9 12 hours, or the cholesterol on admission to a coronary unit or a cholesterol measurement taken at 25 years old or what? The reduction of baseline concept fails to take account of the need to have easily available tools for nursing and health care assistants to identify such a figure, calculate a bespoke target measurement and then prescribe medication and advice to achieve the desired reduction. General practice will implement guidelines when these are sensible and pragmatic. General practice is far too wise now, after its own errors, to pick up the prescribing frenzies seen in HRT, gestodene contraception and opren to jump into new attacks on people who are not necessarily ill and therefore possibly not patients to further impose prescriptions on its populations. General practice is wary of the literature it receives and reads be it journal articles, advertising material or documentation defining itself as ‘guidelines’. General practice will embrace guidelines and treatments where these unequivocally relate to improved end points in populations’ health rather than follow a fad in measurements flows. General practice will follow guidelines when guideline authors can intelligently disseminate their recommendations in a way which is pragmatic, evidential, convincing and makes sense to the highly intelligent readership they desire to guide. Campbell N, Murchie P. Treating hypertension with guidelines in general practice. BMJ 2004; 329: 523- 524 Williams B, Author’s reply. BMJ 2004; 329: 570-571 Primatesta P, Poulter NR Hypertension management and control among English adults aged 65 years and older in 2000 and 2001 J Hypertens. 2004 Jun; 22(6):1093-8. Hennekens CH, The ALLHAT-LLT and ASCOT-LLA trials: are the discrepancies more apparent than real? Curr Atheroscler Rep. 2004 Jan;6(1):9-11. Competing interests: None declared
Risky Relative Risk 7 September 2004
Des Spence, GP Glasgow G20 9DR Send response to journal: Re: Risky Relative Risk	I read with interest the response to the Guidelines on hypertension. The response robustly defends the recommendations quote the available evidence. However, this response did not address the very legitimate concerns raised by the GP writers who actually deliver hypertensive care in the UK. All “evidence” is merely modelling and interpretation depends on your standpoint on acceptable risk. The evidence should seen in context and not as simple “fact”. ALLHAT and WOSCOP are important studies that counter the widespread use of statins in primary prevention. Even the ASCOT-LLA study which is quoted as “fully justifies considering statin treatment” with a “10 year cardiovascular disease risk is estimated to be 20%, irrespective of baseline cholesterol values” is not what it seems. Consider that treatment with a statin has a non significant impact on all cause mortality, the NNT per year is 300 for the primary endpoint and lastly the average age of patients was 63 hardly “fully justices” widespread statin use. We in “GP land” may not have the trappings of academia but we know our communities and tickbox guidelines is not medicine despite the “evidence” otherwise. Competing interests: None declared
Guidelines: how about all-cause mortality? 8 September 2004
Eddie Vos, maintains health-heart.org Sutton Qc Canada J0E 2K0 Send response to journal: Re: Guidelines: how about all-cause mortality?	The response formulated by Williams does not answer many questions raised while suggesting that epidemiological studies provide 'evidence' that high blood pressure (BP) is a major cause of cardiovascular disease which such type studies are evidently incapable of proving. Few are the successful placebo controlled trials for BP or cholesterol lowering drugs with all-cause mortality as endpoint.  Dr. Williams cites cholesterol-lowering ASCOT-LLA where the all-cause mortality curves until mean study end (3.2y) may well have been drawn as a single line, while ALLHAT (-LLA) did not find benefit in any department [mortality, or cardio events] from a mean 17% relative decrease in LDL-cholesterol. The referred to ALLHAT BP lowering arm had no placebo group so it provided no answer as to all-cause mortality -while the 7 year mortality curves for the 3 drugs simply continue to overlap. The response proposes "preventive medicine" which is a bit of an oxymoron since there are no drug deficiency diseases, only nutrient deficiency diseases.  So far, we can only hope that the drugs proposed in such number-managing guidelines deal with the underlying causes of cardiovascular diseases. Do they save lives, how many and in which individuals? While epidemiology cannot prove, it can disprove cause, and it is easy to forget the 3.5x difference in coronary mortality in different regions in the Seven Countries study at identical baseline blood pressures; only BP increases were related, either as cause or as a result of artery decline, but certainly not from drug-related treatment effects. Missing from these guidelines and the response are numbers needed to treat and harm  regarding all-cause mortality and a variety of effects and side-effects.  They, therefore, provide little guidance to clinicians or patients that would like to live longer and healthier lives. There is no benefit in dying with low cholesterol or low blood pressure if there is no overall health and all-cause mortality benefit.  Eddie Vos vos@health-heart.org Competing interests: None declared
Grasping the thorns of the Rose Paradox. 14 October 2004
Peter Davies, General Practitioner Mixenden Stones Surgery, Mixenden, Halifax, HX2 8RQ Send response to journal: Re: Grasping the thorns of the Rose Paradox.	Sir, I am grateful to the authors for their reply (1) to my letter (2). I took their advice and worked out the following numbers. A man at a “high” absolute CV risk of 30% over 10 years (or 3 % in one year) treated with a drug that gives him a relative risk reduction of 33% achieves an absolute risk reduction from 30% down to 20% of an event over 10 years. Alternatively 100 people like him have to be treated for 1 year, or 10 people for 10 years to prevent one event. However you look at these numbers most patients individually have a low probability of benefiting personally from their treatment. Similar high numbers needed to treat for other primary prevention interventions can be found on Bandolier (3) (http://www.jr2.ox.ac.uk/bandolier/band50/b50-8.html) The guideline authors state that there is “overwhelming evidence” of benefit from the treatment of hypertension. This statement is true only at the level of public health and populations. The treatment of hypertension is really a public health exercise and it falls under the rules of the Rose Paradox. (4). This states that a small change across a large population at risk (not ill) has a far bigger effect on the public health than a large change in the health status of any one individual in that population. A patient taking treatment for CV risk is mainly doing a service to his community, with possibly some individual benefit as well. For doctors to try and sell anti hypertension treatment as being “of great benefit to individuals” is misleading both to doctors and patients. Some patients are already sensing this as shown by Hunt and Emslie’s (5) paper on the Rose Paradox in lay epidemiology. For us to implement the CV risk tables and these hypertension guidelines (6) will convert nearly all men over 50 and most women over 60 into patients needing treatment for their high CV risk. This treatment will be provided in the form of pharmaceutical products. Under this regime no one will be found healthy, and all will need drugs. We have a real problem if we adopt this guideline’s approach to CV risk management in general practice, and it is not just in terms of numbers and capacity. Can it really be sensible to regard all men over 50 as being ill because of their level of cardiovascular risk? There is an unresolved conflict between the utilitarian agenda of the guideline authors derived from the public health perspective, and the individual duty based focus of individual doctor-patient interactions. GPs operate at this level, as doctors, treating patients one at a time. GPs are not the frontline agents of the public health agenda. Nor should GPs be the guideline compelled distributors of drugs to the population, and profits to the pharmaceutical industry. I am still far from sure whether I help an individual patient very much by lowering their blood pressure. I doubt they take all their tablets, and I doubt they take the tablets long term. I really do not feel comfortable with some of the implications of these guidelines, although I acknowledge their good intentions and that reducing the average level of blood pressure in the population would reduce the incidence and prevalence of cardiovascular disease. 1. Williams, B (2004) Author’s reply BMJ 329: 570-571 2. Davies, P. (2004) Numbers are missing BMJ 329: 570 3. Bandolier Table of NNTs http://www.jr2.ox.ac.uk/bandolier/band50/b50- 8.html 4. Rose, G (1985) Sick individuals and sick populations International Journal of Epidemiology 14: 32-38 5. Hunt, K and Emslie, C (2001) Commentary: The prevention paradox in lay epidemiology- Rose revisited. International Journal of Epidemiology 30: 442-446 6. William,B, Poulter,N.R., Brown,M.J., Davis,M., McInnes,G.T.,Sever,P.S. et al (2004) British Hypertension Society guidelines for hypertension management 2004. (BHS-IV): summary BMJ 328: 634-640 Competing interests: None declared