Hypertension strategies in the third millennium: conservatism, evidence and the folly of speculation

Conventional wisdom has recently been found wanting in major studies of both coronary artery disease and diabetes. We have learnt that the mechanical pursuit of the perfect coronary lumen may not improve outcome (1) and that the relentless pharmacological pursuit of lower Hba1c in diabetes may cause harm (2). A recent perspective in the New England Journal of Medicine (3) challenges us to be cautious when our assumptions assume the mantle of fact. The authors observe; ‘it is increasingly apparent, that the specific strategies used to modify (vascular risk factors) make a critical difference in patient outcomes’ and note that ‘guidelines and performance measures (should) reflect the evidence about interventions that are known to be beneficial’. Such simple words beguile. What is being said is that if you do not have the trial evidence then you fundamentally do not know if the treatment strategy works. Furthermore, as many recent examples have shown [hormone replacement therapy (HRT), aprotinin in cardiac surgery, high-density lipoprotein cholesterol rising] assuming that a treatment or intervention is beneficial, because observational studies or meta-analyses suggest this may be so, risks hubris. Considered from this perspective, the results of recent studies in hypertension suggest that we need to re-examine recommendations on the choice of first-line agents, particularly diuretics, used to lower blood pressure. In addition, the equivalence of effect, but disadvantage of the combination, in a definitive comparison of angiotensin-converting enzyme inhibition and angiotensin receptor blockade, casts doubt on our understanding of how these medications work. Finally and probably more importantly, we have for the first time, evidence for the benefit of blood pressure lowering in a truly elderly population that contradicts preceding meta-analyses, and provides a clear reminder of the dearth of evidence for the management of vascular risk in older and elderly patients in whom cardiovascular disease (CVD) is increasingly common (4). Controversy exists over the choice of first-line anti-hypertensives. Until recently, guidelines endorsed the primacy of blood pressure lowering per se, and favoured diuretics or beta-blockers as first-line mono-therapy. Born of research conducted in the 1950s, which highlighted the importance of sympathetic over-activity in hypertension, beta-blockade has, until recently, been the mainstay of anti-hypertensive therapy. As newer agents have become available, however, concerns over metabolic side effects have accumulated, with recent re-analysis of older studies suggesting lack of effect in preventing stroke (5). In addition, surrogate outcome data from the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) study (below), which found that beta blockade was less effective at reducing central blood pressure (6), have resulted in the rapid fall of beta blockade from first to ‘fourth-line’ status in hypertension (7). Although debate persists over whether this decision is justified (8) given the lack any reliable large-scale monotherapy data, attention has now shifted to the other ‘old’ class of anti-hypertensives – the thiazide diuretics. All major guidelines endorse diuretics as possible first line anti-hypertensive agents notably the US Joint National Committee (JNC) VII guidelines, which specify thiazide or thiazide-type diuretic as a single first-line choice (9). The more circumspect National Institute of Health and Clinical Excellence/British Hypertension Society (NICE/BHS) guidelines in the UK recommend a diuretic as first-line therapy in ‘older’ patients using a cut-off of 55 years (http://www.nice.org.uk/CG034). Other guidelines from Europe, Canada and Australia also show no consensus as they recommend a number of first-line options. Several major studies of thiazide diuretics (TD) in the 1980s and 1990s, in older although not elderly populations (70–76 years of age), found a significant reduction in stroke with lesser reduction in myocardial infarction (MI) and mortality. Importantly, however, these studies compared thiazide diuretic with placebo, in patients with very high baseline blood pressures (185/91 mmHg in the largest study) (10). Furthermore, the dose of diuretic studied (typically hydrochlorthiazide 25 mg daily) was higher than those currently used (hydrochlorthiazide 12.5 mg or bendroflumethiazide 2.5 mg daily). In addition, and contrary to the assumptions made in meta-analyses, not all diuretics are identical. Indapamide, an indoline, although commonly considered a thiazide may have a more favourable metabolic profile. The only recent study available for diuretic monotherapy, which compared indapamide with placebo 5665 Chinese patients in the Post-Stroke Anti-hypertensive Treatment Study, found that a 5 mmHg reduction in blood pressure was associated with a 29% reduction in recurrent stoke after 2 years (11). The JNC VII authors argue that the more recent Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) study (12), which reported significant benefit of chlorthalidone when compared with both amlodipine and lisinopril, provides contemporary evidence of the ongoing primacy of diuretics. Unfortunately, this study, as with several others which compared active treatment groups (i.e. no placebo group), also found differences in achieved blood pressure levels among the groups (see below). Furthermore, in seeking to compare four major classes of anti-hypertensive (the doxazocin arm was terminated early), in patients with blood pressure of > 160/100 mmHg, ALLHAT also required that additional blood pressure lowering medication was limited to second-line therapy with atenolol, followed by methyl-dopa and reserpine. As 67% of patients required two or more medications, the study is predominantly that of dual therapy with a prominent beta-blocker background, which further limits not only interpretation but also implementation, given the subsequent concerns raised over beta-blockade. The NICE/BHS guidelines justify their recommendation for diuretics on economic calculations that assume ‘equal quality of life loss for all drugs’; however, beyond a 4% tolerance, any cost benefit is lost (9). Simple prescription-based studies suggest that adherence to medication at 1 year is poorer with diuretics than, for example, angiotensin-converting enzyme inhibitors (ACEI) and angiotensin receptor blockers (ARB) (13), casting doubt on such economic modelling. The side effect profile of thiazide diuretics is well recognised, indeed the ALLHAT study found a significant excess of diabetes over the 5-year period of the study in the chlorthalidone group (12). In addition, impotence, gout and electrolyte disturbance are common. Given the availability of other effective agents, it is important that we continue to re-appraise the (often long-term) use of these agents. Over familiarity with an older medication should not be allowed to cloud thorough assessment of newer options. What evidence points to an alternative strategy? The ASCOT study (14) asked whether the combination of the calcium-channel blocker (CCB) amlodipine and ACEI perindopril was superior to the beta-blocker (BB) atenolol combined with the thiazide diuretic hydrochlorthiazide, in preventing coronary artery disease in patients with high-risk hypertension. The average age of participants was 63 years, 78% required two or more agents. In finding no difference in the primary end-point of ‘non-fatal MI (including silent MI) and fatal CHD’, the study failed to prove its underlying hypothesis; however, several secondary end-points including all stroke, total cardiovascular events and procedures and total mortality were significantly reduced. Adjustment for the 2.7/1.9 mmHg in-trial difference in blood pressure achieved (in favour of the CCB/ACEI combination) and for difference in lipid profile from the concomitant lipid-lowering component of the study negated the significance of these findings (15) and thus the weight that can be attached to the study’s conclusions. The amlodipine/perindopril combination was associated with less diabetes – itself an important finding. Major support for a CCB/ACEI strategy comes from preliminary data recently reported from the Avoiding Cardiovascular Events in Combination Therapy in Patients Living with Systolic Hypertension (ACCOMPLISH) study (16). Over 11,000 patients aged 55 or more on anti-hypertensive medication or with untreated systolic BP > 160 mmHg, of whom 60% were diabetic, received either a fixed-dose combination of hydrochlorthiazide plus benazepril or of amlodipine plus benazepril. At 36 months, blood pressure levels were significantly improved, with more than 75% of patients in both treatment arms having blood pressure levels < 140/90 mmHg. Treatment with the CCB/ACEI combination was associated with a 20% reduction in the combined end-point of cardiovascular death, fatal/non-fatal MI, fatal/non-fatal stroke, hospitalisation for unstable angina. As yet unpublished, we do not know the between-group differences in achieved BP; however, these results, if confirmed, challenge not only the primacy of diuretics as first-line therapy, but also the wisdom of initiating monotherapy per se. For some time, it had been argued that both ACEIs and ARBs ‘needed’ a small dose of diuretic to activate the renin–angiotensin system increasing potency. Both ACCOMPLISH and ASCOT suggest that a CCB/ACEI combination produced similar if not slightly greater blood pressure reduction than either BB/TD or TD/ACEI combinations. Secondly, if the metabolic profile of diuretics is a cause for concern, then it becomes more difficult to justify even second line use particularly when patients take medication for 15 or 20 years. Indeed, given the current attention paid to the ‘epidemic’ of diabetes, perhaps our circumspection should be even greater to avoid any sense of self-contradiction. Finally, we have no evidence to support the use of low-dose thiazide diuretics, with the only study to have done so – the ACCOMPLISH study, which used hydrochlorthiazide 12.5–25 mg – finding less benefit in the diuretic arm. Should we not heed The New England Journal of Medicine editorialists warning: that ‘guidelines and performance measures (should) reflect the evidence about interventions that are known to be beneficial’. This is not to deny a place for diuretics in the control of blood pressure. The decision lies in how best to use the evidence base available to us to improve patient care. Angiotensin-converting enzyme inhibitors undoubtedly provide benefit in CVD when compared with placebo in patients with moderate or high risk (17,18). Furthermore, their benefit in the management of heart failure is undisputed (19). The evidence base in hypertension is less robust, in part, because ACEI entered a market where the benefit of thorough blood pressure lowering was already well established. This necessitated trials with actively treated comparator groups (as it would have been unethical to use a placebo) with many studies designed to show ‘equivalence’; i.e. that ‘new’ treatments were of equal efficacy to ‘old’. Two major obstacles to interpretation arise in such studies. First, statistical analysis, in seeking to show equivalence, usually relies upon the results being within one standard deviation of the comparator group. There is thus a risk of over-estimating the effect, as the boundary of ‘equivalence’ with an active comparator is the same as of ‘significant difference’ in a placebo-based study. Secondly, interpretation of results is hindered if blood pressure control differs between treatment groups, as it is then impossible to assess whether any treatment effect is because of advantages of the agent studied or of the additional blood pressure lowering provided. This applies to even small between-group differences in BP, as the effect is being examined at a population level. Such study design also presupposes that blood pressure lowering with the active comparator is itself of definite benefit. The most frequently cited evidence base for ACEI: the Swedish Trial in Old patients with Hypertension-2 (STOP2) (20), Captopril Prevention Project (CAPP) (21), ALLHAT (12) and second Australian National Blood Pressure (ANBP2) (22) comprises studies that compared treatment regimes and produced ambivalent results. Comparison of ACEI-based and conventional therapy strategies in both STOP2 and CAPP found no difference in the respective primary end-points of ‘fatal stroke, fatal MI or fatal CVD death’, and ‘non-fatal MI, stroke and CVD death’ respectively; however, there was an unexplained 25% excess of stroke (p = 0.044) in the captopril group in CAPP. Uniquely, among the ACEI studies, comparison with placebo was made possible in the Perindopril Protection Against Recurrent Stroke Study (PROGRESS) (23), by examining the effect of the combination of the thiazide-type diuretic indapamide with the ACEI perindopril, in patients following stroke or transient ischaemic attack (TIA), where the evidence base for treatment was unexamined. Active treatment with the combination was associated with a 9/4 mmHg reduction in blood pressure and 28% fall in total stroke (p < 0.0001) and 26% reduction in major vascular events. In a prescient foreshadowing the ASCOT study hypothesis and results (14), the large, widely publicised ALLHAT study (12) was designed to establish whether ‘a calcium-channel blocker or an angiotensin-converting enzyme inhibitor, lowers the incidence of coronary heart disease or other CVD vs. treatment with a diuretic’ (hence: Antihypertensive and Lipid Lowering treatment to prevent Heart Attack Trial). The study found no difference between treatment groups in the primary end-point, of ‘fatal CHD or -fatal MI’, over 6 years (chlorthalidone – 11.5%, amlodipine – 11.3% and lisinopril – 11.4%). All-cause mortality also did not differ between groups. Five-year systolic blood pressure values were significantly higher in the amlodipine (0.8 mmHg) and (lisinopril 2 mmHg) groups. In contrast, open-label use of enalapril vs. hydrochlorthiazide in the ANBP2 produced a barely significant reduction of 11% (p = 0.05) in all cardiovascular events or death from any cause, but this was only seen in male trial participants, in a study criticised for the lack of restriction on concomitant therapy choices. At best, these studies represent an uncertain evidence base, with the later ASCOT study (2005) providing little the study blood pressure lowering with a low-dose combination of perindopril and indapamide major vascular events in a population when compared with placebo in the recent Study blood pressure in the active group at mmHg was mmHg lower than ARBs the For reduction in hypertension Antihypertensive and Study on and in the have studied angiotensin receptor blockade in patients with blood pressure. – one of the more active comparator studies, compared and strategies in patients with hypertension and reduction in blood pressure achieved in the two treatment arms mmHg treatment with the was associated with a (p = reduction in the primary end-point of cardiovascular stroke and MI and a 25% (p = reduction in fatal and after for the blood pressure In high-risk in the study, both and strategies vascular events primary end-point blood pressure was significantly lower in the amlodipine group the study particularly so in the mmHg at There was a significant excess of MI in the group. Uniquely, the study to the effect of treatment of hypertension on and vascular events. to assess the effect of compared with placebo in older of agents in the placebo group resulted in an active comparator trial and statistical although which produced greater BP significantly In patients with very high vascular the in Trial and of in with the studies compared the effect of angiotensin receptor blockade, with active treatment in patients with diabetes, hypertension and similar blood pressure treatment with an was associated with a significant reduction in of however, there was no on mortality or cardiovascular Finally, the combination of ACEI and has not been studied in hypertension. In patients following MI, and in those with heart no benefit has been shown when the combination was compared with ACEI therapy use of the combination has been associated with side effects In in the management of hypertension, inhibition of the system at as as ‘older’ agents, the for inhibition to such ‘equivalence’ be their benefit in in patients with vascular cardiac and diabetes as well as and side effect profile – particularly for the of both ACEI and ARBs to the of diabetes, which has been seen in several studies with comparator groups assume Two to be – which is the an ACEI or and the combination Although not a study of blood pressure the and in combination with Trial Study the Heart Outcomes Prevention study upon which was of ACEI and groups, over patients with vascular disease or high-risk diabetes, were given mg or for a of with from a baseline of mmHg, blood pressure by mmHg and by mmHg in the and combination groups Although significant in the of and and in in the group. and were all significantly more in the combination group. There was no difference in the primary end-point of death, MI, stroke or hospitalisation for heart or in any secondary end-point among groups, with take for the results of this study to be the perhaps this that is as effective as – an important in The study also us for example, over the of and that truly a evidence In addition, of the of ACEI and have been by the finding that use of the combination, greater BP did not improve outcome and was associated with significantly outcome Indeed, although a for the this study the central importance of trials that an that hypertension is most in older yet it is in this very group that the evidence base is well is that blood pressure management should be seen as a in older with US guidelines in those over years of A suggest that hypertension in patients should be seen as the as in this hypertension is it the that it is in patients that trials are conducted The benefit of lowering blood pressure in patients the age of years has been A of elderly in studies of blood pressure that a significant reduction in stroke was by an excess in mortality In addition, the of an the between blood pressure and CHD in patients with MI, in data from the study in raised concerns over the wisdom of BP reduction in older has been argued that this the of pressure in older as a of the of systolic hypertension with age recent of over patients in the study, which examined the effect of in patients with hypertension and coronary artery found evidence of at lower pressures with a at mmHg, which mortality on such observational and studies is by the lack of and may only to deny benefit to patients and a significant in the reported of older patients with hypertension in this the Hypertension in the Trial Study is of major Blood pressure with a combination of indapamide and perindopril required two or more was compared with placebo in patients with a age years. At 2 years, a reduction in blood from to mmHg with active was associated with a reduction in the primary end-point of all stroke (p = a reduction in death (p = and a reduction in heart failure (p of this study should be a of CVD at baseline Furthermore, stroke mortality was high at perhaps the older age of the but also with of patients from and from The of any with blood pressure may reflect the of baseline This study as the first at the truly elderly – an increasing for whom the evidence base for all cardiovascular including and lipid-lowering is Such as currently is upon or from age groups. The primary outcome of in itself is – the that a significant reduction in total mortality was achieved over a period vs. in this age group is both as it the and Indeed, many had that studies in truly elderly populations were either or The study has shown that this is not the and thus casts an upon the lack of evidence for treatments in this age group. is when assumptions the status of fact. In we that should not yet the more data we the less to This is of our because we of is the best agent in is a A more – is the strategy – the of disease and the need for agents seen in most studies. The evidence base that this is by studies in the which challenge the or use of diuretics to lower blood pressure and increasingly support fixed-dose combination The use of low-dose thiazide diuretics bendroflumethiazide or is now cost not an evidence The of the evidence with currently available agents, the strategy of CCB/ACEI efficacy and benefit and may a strategy for where should reflect the evidence base in studies. of a thiazide-type diuretic to the most evidence in the of particularly if combined with understanding of how the renin–angiotensin system has also been found ACEIs and ARBs are on of the with yet now the of inhibition with ACEI or of angiotensin levels with ARBs of the inhibition of the central to the treatment of cardiovascular however, we do not know which component of the system the benefit. The benefit of inhibition with and particularly given their increasing in blood pressure inhibition now also possible with there is a need for studies to the of these agents, both and in combination with other We are with The first is for hypertension guidelines to this of are a that favour the over the or the of where evidence is This to be as beta-blockers thiazide diuretics now for first or even second-line in blood pressure lowering Secondly, populations are yet we have allowed to assume that trials in the truly elderly populations are to This with age as a in guidelines, now be seen as and perhaps even it possible that we are other treatment to the older There is a need to the results in other studies. we not now also need the study of lipid lowering in the very elderly and beyond this trials at of vascular risk for the of and Finally, and most importantly, it is clear from recent data that we challenge our studies and may but they do not fact. As in coronary artery disease and diabetes, so in hypertension and that you do not know and is a place to studies we do not the economic and may well prove less to such studies in the is a cause for has received and from and and

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.