Blood pressure targets in primary care
BMJ 2016; 352 doi: https://doi.org/10.1136/bmj.i813 (Published 25 February 2016) Cite this as: BMJ 2016;352:i813All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
There is too much that needs to be said about the elements of this thoughtful paper to fit a letter-scale response. The ‘Targeting’ of clinical variables was considered extensively in renal medicine in the past and remains a common rubric. 1 My recent BMJ response to a Diabetic Guideline gave a summary of the semantic pitfalls and some earlier references. 2 Even when the variable is more stable than blood pressure, say haemoglobin under ESA, the declaration of a target value used as a clinical aiming point results operationally in the mean of the outcome range close to or equivalent to the target value, with the implication that half of any treated group will fall above and below.
There are probably several causes for this, but it is important to recognise that a declared target used as an aiming point and any subsequent outcome are not going to be identical. The narrowing of any distribution (leptokurtosis) is difficult to achieve, and seasonal and other influences conspire to shift constantly the outliers of any group.
With blood pressure the style of guideline statements as a vague ‘less than’ declared value has probably not helped. The adequate separation of groups by outcome within RCTs has also proved difficult for renal Haemoglobin, for example. It was learned that because of systematic underachievement it was necessary to aspire somewhat beyond the value of interest, but that carries hazard for, say, blood glucose and blood pressure (as an aside - how much better to use ‘intensive’ instead of the wishful thinking of ’strict’). When prescription is based on threshold methods the practical instruction to intervene had to be quite different from the desired goal (aka target) to achieve a largely compliant population (using, say, a 4mmol/l threshold of total cholesterol for statins to achieve high population compliance with <5). 3
The relation between target and outcome in each study depends on the ‘treatment algorithm’ (explicit or implicit) being deployed.
For Haemoglobin and Dialysis dose the achieved values seemed to give less information about prognosis than the quantified therapeutic difficulty of achieving the ‘target’. The clinical outcomes were reflected in ESA dose, not the achieved Haemoglobin, for example, and there was a Dose Targeting Bias effect in dialysis. 4 5 Similar findings are reported for Blood Pressure. 6 [This phenomenon can be compared to the crude but accurate comparison of premises heated to a stable constant temperature by examination of their fuel bill.] The clinical effort necessary to control blood pressure is likely to be related to an illness potential that is otherwise obscure, a possibly useful corollary of any patient management.
Guideline authors have been always much clearer about where to go than how to get there. Their language has not helped. The use of target to mean desirable outcome as well as a default aiming point creates systematic confusion and underachievement. Targets are not a substitute for goals, and that has had more than semantic consequences.
1. Will E. Intention and outcome in guideline-based nephrological practice: a suitable space for ‘clinical technology’. Nephrol Dial Transplant 2007;22: 3110–3114. doi:10.1093/ndt/gfm516
2. Response to : Beckles ZL et al. Diabetes in children and young people. BMJ 2016;352:i139 (dx.doi.org/10.1136/bmj.i139)
Minding other words: targets and goals. 30th January 2016
3. Garthwaite EA, Will EJ, Bartlett C, Richardson D, Newstead CG.
Patient-specific prompts in the cholesterol management of renal transplant outpatients: results and analysis of underperformance. Transplantation 2004;78 (7):1042-1047
4. Tom Greene, John Daugirdas, Thomas Depner et al. Association of Achieved Dialysis Dose with Mortality in the Hemodialysis Study: An Example of “Dose-Targeting Bias” J Am Soc Nephrol 2005;16: 3371–3380, doi: 10.1681 /ASN.2005030321
5. Parfrey PS. Should Hemoglobin Targets for Anemic Patients with Chronic Kidney Disease Be Changed? Am J Nephrol 2010;31:565–566
doi: 10.1159/000313894
6. Davis EA, Appel LJ, Xuelei W etal. Limitations of analyses based on achieved blood pressure: Lessons from the AASK trial.
Hypertension 2011; 57(6): 1061–1068. doi:10.1161/HYPERTENSIONAHA.111.169367.
Competing interests: No competing interests
Re: Blood pressure targets in primary care
As a patient subject to the guidelines in practice one begins to wonder at the “targets” so defined for various parameters. Taking the statement in the editorial:
“Use of the more intensive target led to a statistically significant but modest difference of about 3 mm Hg in SBP between groups”
one wonders at precisely what a change of 3 mm Hg actually accomplishes both in terms of “risk” and in terms of adverse reactions in individual patients. What in fact is the increased drop out rate actually responding to? It apparently of no concern to the medical statisticians.
Furthermore, a paper by Port (Lancet. 2000 Jan 15;355(9199):175-80) has shown that, to quote:
“The Framingham data contradict the concept that lower pressures imply lower risk and the idea that 140 mm Hg is a useful cut-off value for hypertension for all adults. There is an age-dependent and sex-dependent threshold for hypertension. A substantial proportion of the population who would currently be thought to be at increased risk are, therefore, at no increased risk.”
a point that has been studiously ignored by those who support the paradigm of a linear response between SBP and risk of CHD and/or stroke.
Another interesting quote from a comment to the Lancet (2000) is:
It is a curious fact that simply fitting a statistical model to data can give the impression that the assumptions of the model are supported by the data. Sidney Port and colleagues reassess the Framingham data, in which a linear-logistic model has previously been used to relate all-cause mortality to systolic blood pressure: this model assumes that the odds ratio for death for a 10 mmHg increase in systolic blood pressure is the same at all levels of systolic blood pressure. Port and colleagues show convincingly that this assumption is false, and they propose a better model in which the odds ratio for a 10 mmHg increase in systolic blood pressure is larger at systolic blood pressures above the 70th centile.
While I am a keen supporter of statistics (my interest goes back to 1953), these days I get the impression that they takes precedence over all; odds ratios are prime but unfortunately as a ratio of two ratios, which, while helping to demonstrate significance they also conceal the real numbers and benefits and non-benefits. What seems always to be forgotten is the patient and the probability that he or she will benefit (or not benefit as the case may be) from the therapy. Unfortunately, in many cases these days, while therapies may be statistically beneficial, the actual benefit may be trivial and improbable.
Competing interests: No competing interests