Too much blood pressure?

This editorial is about our (anaesthetists’) preoccupation with pre-operative blood pressure. In it, I propose that abrupt unilateral cancelation or postponement of operations due to raised blood pressure readings is wrong, that the pre-operative management of blood pressure measurement is often careless, and that management of patients, particularly the postponement of surgery, has been driven by concerns that might be unfounded rather than by considered assessment of the evidence. Furthermore, communication of this evidence with patients may be rushed, incomplete or absent. Postoperative strokes might or might not be the anaesthetist's fault, whether or not he/she has postponed surgery to reduce blood pressure, whilst a patient's overall risk of stroke (i.e. over his/her lifetime) might be unaffected even if the anaesthetist reduces the risk of postoperative stroke. The postoperative period is too narrow a window to see whether what we do benefits the panorama of a patient's life. You will be less likely to have a stroke or a myocardial infarction, and you will probably live longer, by reducing your blood pressure, if raised 1-4. At least, you will if you're well enough to tolerate a reduction in your blood pressure: for instance, reduction of blood pressure might increase mortality and morbidity in patients with peripheral arterial disease and in patients with heart failure 5-12. You might also need a higher blood pressure as you get older 13, 14. You also have to be ‘dying’, or running the risk of cardiovascular catastrophe at a high enough rate, that the absolute reduction in bad outcomes is worthwhile: randomised controlled trials have been designed to detect an effect and therefore they have recruited people older than 60 years, often with existing cardiovascular morbidity and pre-existing treatment with antihypertensive drugs. Direct evidence from randomised controlled trials is that as long as you're dying at a high enough rate, either by being old enough or sick enough, then reducing blood pressure from at least 160/95 mmHg is a good thing, with greater benefit the closer one can get to 140/90 mmHg. The treatment of blood pressures less than 160/95 mmHg is more contentious and is based predominantly on indirect evidence 15, 16. Treatment targets less than 140/90 mmHg are also contentious, with more recent guidelines abandoning targets below this reading. When drugs are used to reduce blood pressure, their good effects are achieved solely through the reduction in blood pressure 17, 18, whilst the good effects of exercise are mediated through more complex mechanisms 19. Most drugs don't make you feel better but exercise does (or at least, it does when you're not actually exercising!). The interplay between risks and benefits can be studied using simulation. Take, for example, the effects of a 10% reduction in systolic blood pressure, which would usually require either two antihypertensive drugs given at full dose or regular exercise 1. Let us use a cohort with a high five-year rate of cardiovascular events so that the effects of treatment are more apparent e.g. a 33% mortality rate at five years (> 16% is considered ‘high risk’) 1. Let's favour treatment over no treatment and let's also assume that: i) the untreated cohort remains untreated; ii) treatment is beneficial for the whole treated cohort, in whom cardiovascular events are reduced by 36% (the magnitude of effect expected in the risk group); iii) everyone takes their drugs and that the drugs are equally effective in everyone; iv) drugs immediately reduce both the blood pressure and the associated rate of cardiovascular events; and v) reductions in cardiovascular events are accompanied by a reduction in all-cause mortality of 20% (the magnitude of effect expected in this risk group). Figure 1 shows the effect of treatment compared with no treatment on survival, combined cardiovascular events and stroke. So what happens when these two cohorts have a scheduled open abdominal aortic aneurysm repair? I have used this operation as an example because the patients who present for this surgery are generally at a high risk of postoperative mortality and morbidity. Let's assume that this cohort has an untreated monthly mortality rate of 0.40% without surgery, which is similar to the cohorts recruited to the UK Small Aneurysm Trial (UK SAT) and the US Aneurysm Detection and Management (ADAM) trial in the early 1990s 20, 21. Figure 2 illustrates events following surgery in cohorts whose hypertension is either treated or untreated: it is similar to Fig. 1, other than the survival curves start below 100% because of the 3-4% of deaths caused by surgery, which have been modelled as if they happen immediately. The conclusion from Figs. 1, 2 is that a reduction in blood pressure reduces mortality and cardiovascular morbidity. The cohort that presents for scheduled repair of abdominal aortic aneurysm nowadays has a mortality rate half that depicted in Figs. 1, 2, owing to a fall in population mortality rates since the UK SAT and ADAM trials, contributed to by improving health, less smoking and more drugs that reduce cardiovascular risk. For most patients presenting for scheduled surgery, the absolute benefit of blood pressure reduction is substantially less than depicted in these graphs – i.e. the red and black lines are much closer together. I have not yet addressed the main issue that has exercised anaesthetists when faced with a patient with a pre-operative blood pressure reading in excess of 160/90 mmHg: is it better to: i) postpone surgery to confirm or refute in primary care a provisional diagnosis of ‘hypertension’, and then offer treatment, or ii) proceed with surgery and offer antihypertensive treatment later? This, too, can be addressed using simulation. I have again used cohorts with a high untreated monthly mortality rate of 0.40%. If one is going to see an effect of delaying surgery for the reduction of blood pressure, it will be in such a cohort with a high absolute rate of events. (I would like to emphasise here the importance of absolute risk: anaesthetists should gauge their concern for a patient with raised blood pressure by the calculated mortality rate or rate of cardiovascular events for that patient and not by the blood pressure measurement per se). For these simulations I have used a calculator that I developed, which is described elsewhere 22-24. In both cohorts, blood pressure is lowered three months after pre-operative assessment. In one cohort, scheduled open repair of abdominal aortic aneurysm proceeds without antihypertensive treatment, which is initiated (in those who survive) three months postoperatively. In the other cohort, surgery is postponed to reduce blood pressure; scheduled abdominal aortic aneurysm repair proceeds three months later than in the other cohort (Fig. 3). The striking conclusion from this simulation is that there is no difference in survival, or in the rate of cardiovascular events, between these management strategies, even in this ‘high-risk’ cohort. This simulation assumes that antihypertensive treatment generated the same 36% reduction in the rate of cardiovascular events and 20% reduction in mortality in the months following surgery, in both groups. The reduction in events that one would expect during the three months that one cohort was treated and the other cohort was not was counterbalanced by first, some patients’ dying during the three months after antihypertensive treatment started and before surgery, and second, treated patients’ dying at a slightly higher postoperative rate than one might expect because they were three months older at the time of surgery than the cohort who proceeded directly to aneurysm repair. Different patients might prefer one or other strategy; clinicians should avoid the temptation to recommend one or other. Clinicians should also avoid the temptation to feel virtuous if they initiate pre-operative blood pressure reduction in a patient who does not have a postoperative cardiovascular event; it may be nothing to do with them. Similarly, they should also resist the temptation to feel guilty if a patient has a cardiovascular event without initiating pre-operative treatment. You might believe that postponing surgery for some pathologies is merely inconvenient for a patient and would justify the commencement of antihypertensive treatment. Most populations have a lower mortality rate than the aortic aneurysm cohort that I have used, so delaying surgery in these cohorts would be accompanied by fewer deaths in the interim and a smaller increase in mortality (due to ageing). One might therefore think that reductions in peri-operative mortality and morbidity can be achieved at little additional cost. However, these cohorts will have a much smaller benefit from antihypertensive treatment for exactly the same reason – their rates of untreated cardiovascular events are much lower (than the aneurysm cohort). The smaller ‘harm’ in delaying surgery is balanced by a smaller ‘benefit’ in operating on a patient who has had antihypertensive medication commenced. In addition, operations for which delay might be viewed (by clinicians) as merely inconvenient are also associated with lower relative and absolute increases in peri-operative mortality and morbidity, for instance hernia repair, joint replacement or cholecystectomy. So neither the type of patient nor the type of surgery alter the conclusion that postponing surgery for antihypertensive treatment does not alter a patient's overall survival or rate of morbidity. In these simulations, I have assumed that there is no special harmful interaction between having surgery and having raised blood pressure, and that postoperative deaths and cardiovascular events are the consequence of a general interaction between pre-operative mortality rate and surgery. One might plead a special association between pre-operative blood pressure and postoperative mortality and morbidity, although no one has been sufficiently motivated to run an appropriately powered randomised controlled trial to determine the case 25-27. Perhaps I have not persuaded you to discontinue your special relationship with pre-operative blood pressure. You might like to consider that I have been generous attributing immediate and absolute efficacy to drugs to reduce blood pressure. In fact, only one in four patients prescribed antihypertensive medication achieve a blood pressure of 140/90 mmHg, and it takes time: a few weeks to achieve a stable reduction in blood pressure but an uncertain (probably longer) time to achieve a stable reduction in cardiovascular event rates 28, 29. One reason that blood pressure has become a pre-operative issue is that it is the only cardiovascular risk factor, apart from fitness, that can be easily measured anywhere. But, ironically, blood pressure is the only variable that systematically overestimates risk when measured in hospital. The blood pressure measurements that correlate best with cardiovascular events are those taken outside hospital 30. The other reason that blood pressure has become a pre-operative issue above all other cardiovascular risk factors is that it is measured intra-operatively, and anaesthetic-induced hypotension is more likely in hypertensive patients. This may be a good reason to alter peri-operative management for patients with a history of hypertension, particularly those with higher absolute risks of mortality and morbidity, but it is not a good reason to cancel elective surgery. In summary, relinquish your fear of raised blood pressure: give it the same care and attention that you might to equivalent modifiable risk factors, such as inactivity, smoking and raised cholesterol. Take care to consider an individual's overall risk of morbidity and mortality i.e. incorporate risk factors that are not modifiable, such as sex, age and cardiovascular events, and investigate symptoms or signs consistent with important disease that have not been explained. Develop a pre-operative service that does not measure blood pressure in hospital; instead, rely upon accurate measurements taken by the patient's general practitioner in the preceding 12 months. Avoid peri-operative hypoxaemia and hypotension 31! No external funding declared. I am a member of the Association of Anaesthetists of Great Britain & Ireland's guideline working party on pre-operative measurement of adult blood pressure and management of hypertension. The opinions expressed above are my own and are not intended to represent the views of the working party.