2nd October 2009
Two Cochrane reviews have addressed these questions. The review of blood pressure (BP) targets for hypertension concluded that treating patients to achieve BPs of less than 140/90mmHg does not prolong survival or reduce the risk of stroke, MI, heart failure, major CV events or end-stage kidney disease. This also seemed to hold true for patients with diabetes or chronic kidney disease, but further systematic reviews in such patients are underway. The review of first-line drugs for hypertension concluded that low dose (but not high dose) thiazides were unsurpassed in reducing mortality and morbidity.
Level of evidence:
Level 1 evidence according to the SORT criteria.
In patients without diabetes or kidney failure, health professionals should continue to follow NICE guidance on management of hypertension, and aim to reduce BP to 140/90mmHg (or less). In patients 55 years and older, and black patients of any age, low-dose thiazide diuretics (e.g. 2.5mg/day bendroflumethiazide) are a good first choice. In younger patients, ACE-inhibitors are a reasonable first choice, adding-in, or replacing with, a low-dose thiazide if BP targets are not achieved. There may be particular reasons for choosing other drugs in particular circumstances, for example people with diabetes (NICE recommends an ACE-inhibitor first choice for people with type 2 diabetes, with additionally a thiazide or a calcium channel blocker for black people). NICE recommends a lower BP target for people with type 2 diabetes and end-organ damage, and in people with chronic kidney disease. Although these are probably still worth pursuing, the Cochrane review reinforces the need to set likely benefits of aggressive BP treatment against possible harms (e.g. hypotension, syncope, the burden of extra medication and associated side effects, etc).
What is the background to this?
Epidemiological studies have shown that there is a continuous relationship between BP and risk of cardiovascular (CV) disease – hence, hypertension diagnostic criteria and treatment targets are to some extent arbitrary. Guidelines differ in both aspects but generally recognise that the lower the blood pressure the better. Randomised controlled trials (RCTs) have shown that in the majority of people in the trials, current targets are not unrealistic to attain. However, to achieve lower BP often requires more drug treatment, with the risk of associated side effects and the burden of more medicines to take, so it is important to weigh these demerits against the benefits of ever lower BP. The HOT study has been the benchmark study in hypertension target studies for some time, along with the UKPDS 38 intensive versus less intensive control of blood pressure arm (in people with type 2 diabetes). The first Cochrane review combines data from the HOT study with other RCTs, but excludes UKPDS 38 because the BP targets in this study were much higher than the targets in other studies.
Having decided on a BP target, the obvious next question is the choice of drug. A review of the ALLHAT study concluded that diuretics were unsurpassed in reducing the risk of patient oriented outcomes. The second Cochrane review addresses this question further.
What does these studies claim?
Data from seven RCTs of 22,089 patients were included. Despite a weighted mean difference of 3.9/3.4mmHg in attained BP, there was no significant difference in total mortality (relative risk [RR] 0.92, 95% confidence interval [95%CI] 0.86 to 1.15), or major CV events (myocardial infarction [MI], stroke, congestive heart failure [CHF] or CV death: RR 0.94, 95%CI 0.83 to 1.07. There was also no significant difference in individual CV outcomes. The net health benefit could not be calculated due to a lack of information on adverse effects and withdrawals from treatment.
Re-analysis including only results for a greater weighted mean difference in target and achieved BPs led to no appreciable changes in RR for any outcome. Sensitivity analysis in people with diabetes also found no significant difference for any outcomes in people randomised to a lower target BP. A second sensitivity analysis in people with chronic renal disease found no significant difference in the two outcomes it was possible to pool: total mortality and end-stage kidney disease
Choice of first drug
Data from 24 RCTS of 58,040 patients were included. Compared to no treatment, first-line low-dose thiazides (e.g. bendroflumethiazide less than 5mg/daily) significantly reduced mortality (RR 0.89, 95%CI0.82 to 0.97), stroke (RR 0.68, 95%CI 0.60 to 0.77), coronary heart disease (CHD, RR 0.72, 95%CI 0.61 to 0.84), and total CV events (stroke, CHD, hospitalisation or death from CHF and other significant vascular deaths, but excluding angina, transient ischaemic attacks [TIAs], surgical or other procedures or accelerated hypertension) RR 0.70, 95%CI 0.64 to 0.76. First-line high dose thiazides (e.g. bendroflumethiazide 5mg/day or greater) significantly reduced the risk of stroke and total CV events, but not CHD or mortality.
First-line beta-blocker therapy significantly reduced the risk of stroke (RR 0.83, 95%CI 0.72 to 0.97) and total CV events (RR 0.89, 95%CI 0.81 to 0.98) but not CHD (RR 0.90, 95% CI 0.78 to 1.03) or mortality (RR 0.96, 95% CI 0.86 to 1.07).
First-line ACE inhibitor therapy significantly reduced mortality (RR 0.83, 95%CI 0.72 to 0.95), stroke (RR 0.65, 95% CI 0.52 to 0.82), CHD (RR 0.81, 95% CI 0.70 to 0.94) and total CV events (RR 0.76, 95% CI 0.67 to 0.85).
Based on one RCT, first-line calcium-channel blocker therapy signficantly reduced the risk of stroke (RR 0.58, 95%CI 0.41 to 0.84) and total CV events (RR 0.71, 95% CI 0.57 to 0.87) but not CHD (RR 0.77, 95% CI 0.55 to 1.09) or mortality (RR 0.86, 95%CI 0.68 to 1.09).
Putting this altogether, it seems that the message for prescribers and patients is to aim for a reasonable blood pressure reduction, to around 140/90mmHg or less. Trying to get much lower than this for most patients is unlikely to be worthwhile. Although blood pressure reduction is of prime importance, choice of drug does seem of some significance, and, all other things being equal, a low dose thiazide is probably first choice, with an ACE-inhibitor coming next in the preference list.
The results of the review looking at BP targets support NICE’s approach for most patients: to aim for 140/90mmHg or below, but to note that drug treatment to lower a patient’s blood pressure is worthwhile even if the BP pressure does not fall this far below on treatment with several drugs (or if the addition of more drugs is inappropriate or declined). In people with diabetes or chronic kidney disease, the case for more rigorous BP targets is not proven. Although the NICE targets are probably still worth pursuing, the Cochrane review reinforces the need to set likely benefits of aggressive BP treatment against possible harms (e.g. hypotension, syncope, the burden of extra medication and associated side effects, etc). Clinicians need to carefully consider whether a fourth, fifth or even sixth drug is really going to have a significantly worthwhile benefit compared to the risks, side effects and the inconvenience to the patient in taking it. A BMJ article has pointed out that guidelines are based on average findings and also that individual patients will have quite differing views on risk and acceptability of side effects. As the authors conclude, a patient’s BP can be lowered until side effects are unacceptable or they wish to avoid further additions or changes to their medicines. NICE reinforces the need for patient centred care.
It is important that BP is measured, and hypertension diagnosed, only using calibrated, validated sphygmomanometers and with immaculate technique. Moreover, as we have blogged, the practice of monitoring and adjusting BP management on the basis of infrequent (e.g. every six months) measurements is an unreliable method of indicating underlying changes in BP. Individual measurements that exceed treatment thresholds would seem inappropriate for adjusting antihypertensive treatment without further evaluation. Owing to variations in measurements resulting from normal day-to-day variations in BP, and instrumental or operator error, readings in excess of values likely to prompt a change in drug treatment are more likely to be erroneous than a true reflection of the change in BP since the last reading.
The review of drug choices is also in line with NICE’s recommendations. The review found that thiazides had the most robust evidence base, more so than ACE-inhibitors or calcium channel blockers, and there was no evidence to support first-line use of angiotensin receptor blockers (known as A2RAs, ARBs or sartans). Most people with hypertension are aged over 55, and in such people, and black people of any age, NICE recommends a thiazide diuretic or a calcium channel blocker. Given the difference in cost between diuretics and calcium channel blockers, and the data from this review, a diuretic is probably the preferred agent of the two, all other things being equal. ACE inhibitors were also shown to have similar benefits to thiazides in this review, although confidence intervals were wider as the results were drawn from fewer data. Again, given the differences in cost, thiazides would probably be preferred. NICE advises using ACE inhibitors first line in hypertensive people younger than 55 years although this is based on pathophysiological principles rather than RCT evidence. However, they should be avoided in pregnant or breast-feeding women. ACE inhibitors are also recommended as first-line treatment in people with diabetes, but many people with diabetes need more than one drug to achieve BP targets and the combination of ACE inhibitor and diuretic seems sensible.
NICE did not recommend beta-blockers for routine use first line, since in head-to-head trials, they were usually less effective than the comparator drug at reducing major CV events, in particular stroke. This Cochrane review suggests that compared to placebo, beta blockers did reduce the risk of stroke and CV events, but not CHD or mortality. Either way, a low dose thiazide or ACE inhibitor would appear to be a better first-choice drug for most people. However, there could be good reasons for including or preferring a beta-blocker, for example in patients with angina or who have suffered a recent MI.
Blood pressure targets
Patients: 22,089 adult patients from 7 RCTs, with hypertension (including patients with co-morbidities such as type 2 diabetes or renal impairment)
Intervention and control: patients were randomised to a “lower” target BP (≤135/85mmHg) or “standard” target BP (≤140–160/90–100mmHg). Follow up ranged from 19 months to 6.4 years (most trials three years or longer)
Outcomes: No trials comparing different systolic BP targets were found. In patients randomised to the “lower targets”, the weighted mean BP was 139.3/81.7mmHg compared to 143.2/85.1mmHg in patients randomised to “standard targets” (≤140–160/90–100mmHg). Despite this weighted mean difference of 3.9/3.4mmHg in attained BP (P< 0.001) , attempting to achieve “lower targets” instead of “standard targets” did not change total mortality (RR 0.99, 95% CI 0.86 to1.15), myocardial infarction (RR 0.90, 95% CI 0.74 to 1.09), stroke (RR 0.99, 95% CI 0.79 to 1.25) , congestive heart failure (RR 0.88, 95% CI 0.59 to1.32), major cardiovascular events (RR 0.94, 95% CI 0.83 to 1.07), or end-stage renal disease (RR 1.01, 95% CI 0.81 to 1.27).
A sensitivity analysis in people with diabetes also found no significant difference for any outcomes in people randomised to lower target BP. The RR for mortality was 0.72 (99%CI 0.47 to 1.10) and for total CV events it was 080 (99%CI 0.59 to 1.09). A second sensitivity analysis in people with chronic renal disease found no difference in the two outcomes it was possible to pool: total mortality (RR 0.89, 99%CI 0.52 to 1.52) and end-stage kidney disease (RR 1.01, 99%CI 0.75 to 1.36).
The net health effect of lower targets cannot be fully assessed due to lack of information regarding all total serious adverse events and withdrawals due to adverse effects in 6 of 7 trials.
Sponsorship: Supported by: Departments of Anesthesiology, Pharmacology & Therapeutics and Medicine, Faculty of Medicine, University of British Columbia, Canada; Canadian Institutes of Health Research, Canada; British Columbia Ministry of Health Grant to the Therapeutics Initiative, Canada; Universidad de Costa Rica, Costa Rica; CONICIT, Ministerio de Ciencia y Tecnologia, Costa Rica.
Choice of drug
Patients: 58,040 patients with hypertension from 24 RCTs with 28 arms, of at least one year duration
Intervention and comparison: Trials compared ACE inhibitors, beta-blockers, calcium channel blockers or thiazide diuretics with a placebo or no treatment. No RCTs were found for angiotensin receptor blockers or alpha-blockers.
Outcomes: Thiazides (19 RCTs) reduced mortality (RR 0.89, 95% CI 0.83, 0.96), stroke (RR 0.63, 95% CI 0.57 to 0.71), CHD (RR 0.84, 95% CI 0.75 to 0.95) and CVE (RR 0.70, 95% CI 0.66 to 0.76). Low-dose thiazides (8 RCTs) reduced CHD (RR 0.72, 95% CI 0.61 to 0.84), but high-dose thiazides (11 RCTs) did not (RR 1.01, 95% CI 0.85 to 1.20).
Beta-blockers (5 RCTs) reduced stroke (RR 0.83, 95% CI 0.72 to 0.97) and CVE (RR 0.89, 95% CI 0.81 to 0.98) but not CHD (RR 0.90, 95% CI 0.78 to 1.03) or mortality (RR 0.96, 95% CI 0.86 to 1.07).
ACE inhibitors (3 RCTs) reduced mortality (RR 0.83, 95% CI 0.72 to 0.95), stroke (RR 0.65, 95% CI 0.52 to 0.82), CHD (RR 0.81, 95% CI 0.70 to 0.94) and CVE (RR 0.76, 95% CI 0.67 to 0.85). Calcium-channel blocker (1 RCT) reduced stroke (RR 0.58, 95% CI 0.41 to 0.84) and CVE (RR 0.71, 95% CI 0.57 to 0.87) but not CHD (RR 0.77 95% CI 0.55 to 1.09) or mortality (RR 0.86 95% CI 0.68 to 1.09).
Sponsorship: Faculty of Medicine, University of British Columbia, Canada; Canadian Institutes of Health Research grant to Hypertension Review Group, Canada
Make sure you are signed up to NPC Email updates — the free email alerting system that keeps you up to date with the NPC news and outputs relevant to you