Intended for healthcare professionals

Clinical Review State of the Art Review

Management of mild hypertension in adults

BMJ 2016; 355 doi: (Published 21 November 2016) Cite this as: BMJ 2016;355:i5719

Re: Management of mild hypertension in adults

The authors have done a great deal of work and the review is useful in gathering the information together. However, I had thought that we, as a profession, had accepted that (particularly within the field of primary prevention) absolute measures, such as absolute risk reduction (ARR) or numbers needed to treat (NNT) are absolutely crucial to informed consent for patients (and are extremely helpful for doctors). The main basis of the paper is the meta-analysis in Table 2 in the paper version, or Table 4 in the web version. Neither give a single assessment of absolute benefit - other that a single footnote in Table 4 that points out that the increased odds ratio of death in the treated group in one study of 3.98 (which is slightly worrying) was on the basis of only two deaths (that’s a relief). This is the only absolute measure in the whole, huge meta-analysis. It is also presumably of some importance that this particular 2015 meta-analysis, which showed some statistically significant relative risk reductions and is the basis of most of the recommendations, was made up almost exclusively of patients with diabetes. This (crucial) fact is mentioned once in the table but nowhere in the text. The conclusion drawn by the authors is that: “pooling the results from these trials showed that anti-hypertensive treatment with in patients with mild hypertension likely prevents cardiovascular events, particularly stroke and mortality.” Is this really correct when the 2012 meta-analysis showed no benefit and the 2015 meta-analysis was almost all made up of diabetic patients? The authors discuss the SPRINT study but this has absolutely no relevance to whether low risk patients with mild hypertension should be treated because 90% of the patients enrolled in SPRINT were already on anti-hypertensive drugs, 20% already had cardiovascular disease and 60% had significant risk factors.
If I try and explain this data to a patient on Monday morning would I be correct in saying that there is absolutely no evidence that treating a patient with mild hypertension and no other risk factors with drugs gives any benefit? I am getting to the age where I may need to take anti hypertensive drugs myself. It would be nice to know what my chances are of benefiting (my absolute attributable risk reduction) from taking treatment for, say, 5 years. It would also be useful to know if my risk of acute kidney injury is 4.1% per year (SPRINT intensive treated group) and what my risk of conking out and breaking my hip is.
I have felt uncertain and uncomfortable about treating low risk people with hypertension ever since the last NICE guidance came out. I am sure I am not alone in this. Reverting to meta-analyses like this, that give little indication of absolute levels of risk and risk reduction, does seem to be a bit of a regressive step.

Competing interests: No competing interests

27 November 2016
Kevin Barraclough
Hoyland House, Painswick