Blood pressure targets for type 2 diabetes should be lower, say researchersBMJ 2015; 350 doi: https://doi.org/10.1136/bmj.h749 (Published 11 February 2015) Cite this as: BMJ 2015;350:h749
All rapid responses
The convention of blood pressure reduction has for years relied on the principle of ‘less than’ given numerical values (e.g.<140/80), presumably because clinicians were initially indifferent to low values and subsequently because systolic pressure alone seemed to tell the prognostic tale. It seems easy to overlook the tiny ‘less than’ sign in both medical articles and clinical practice, so that treatment ‘targets’ become expressed as the surviving single numerical value, on the page and in the surgery. There is also an uncanny tendency for any declared numerical value (regardless of qualification) to become the mean of study outcome distributions in RCTs. The loss of ‘less than’ during the translation of the terms used in clinical science and the distortions of translated statistical jargon are casualties of the shorthand of medical publishing1. Other common terms are also prone to migrate.
The irresistible attraction of ‘targeting’ outcomes, a modern linguistic hyperbole to infer quasi-military precision and power, has moved from the practical notion of ‘aiming at’ the value of a clinical variable to merely expressing the desire to reach a particular outcome value. It is apparent that outcome variables can only be described in statistical distributions, and these, because of biology and the logistical challenges of RCTs, must be actively manoeuvred towards desired means or pre-defined ranges. This must be achieved through the introduction of treatment algorithms using empirically chosen thresholds of intervention and default. Whether it is sensible to talk of ‘targeted ‘ algorithms remains to be seen. The validation and statistical properties of treatment algorithms has become a core, and sometimes obscure, feature of RCTs. It might be hoped that successful algorithms could be made available for routine practice. In the event, the romantic vigour of ‘targeting’ is lost2.
This all has a bearing on the head-scratching of guideline authors and begs a more coherent approach to what might be called the ‘ballistics’ of hypertension (early cannoneers ignored the mathematical ballistics of Tartaglia, 1537, to their probable detriment in front of unbreached city walls). Conceptually, study algorithms are part of such ballistics. A recent BMJ paper made the point strongly by its mere existence and there is a small literature on clinical procrastination, for example3. Whatever the inevitably blunt RCT evidence from large, defined, patient groups under various algorithms, without a clearer idea of treatment ballistics the real world implications must be quite uncertain. They might well expose phenomena like dose-targeting bias, in which the effort of achieving given outcome targets appears to reflect survival characteristics 4,5. This is something hidden in studies of blood pressure control and may yet explain some of the unintended consequences of hypotensives deployed at any recommended ‘target’ values.
1. WILL EJ. Surfing the wave of misrepresented hospital death rates—a reticence of the well informed? BMJ 2013;347:f5367
2. Will E. Intention and outcome in guideline-based nephrology practice: a suitable place for ‘clinical technology’. Nephrol Dial Transplant 2007;22:3110-3114
3. Xu W, Goldberg SI, Shubina M, Turchin A. Optimal systolic blood pressure target, time to intensification and time to follow-up in treatment of hypertension: population based retrospective cohort study. BMJ 2015;350:h158
4. Greene T, Daugirdas J, Depner T. Hemodialysis Study Group: 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
5. Patrick S. Parfrey. Should Hemoglobin Targets for Anemic Patients with Chronic Kidney Disease Be Changed? Am J Nephrol 2010;31:565–566
Competing interests: No competing interests