40 Angiotensin receptor blockers

Angiotensin receptor blockers (ARBs) were developed to provide an alternative approach to interrupting the renin–angiotensin system (RAS), which was pharmacologically distinct from that of angiotensin converting enzyme (ACE) inhibitors.¹ By selectively blocking the angiotensin II type 1 (AT₁) receptor, ARBs were thought to offer the potential for more complete blockade of the RAS. Evidence had accumulated that angiotensin II could be generated through enzymes other than ACE (e.g. chymase) and ARBs block the action of angiotensin II, regardless of its source, at the receptor level (Fig. 40.1). Another fundamental difference between the two approaches is the lack of effect of ARBs on bradykinin which is metabolized by ACE (also known as kininase II). Accumulation of bradykinin with ACE inhibitors is responsible for cough and angioedema. Consequently, it was hoped that ARBs would be better tolerated than ACE inhibitors. Bradykinin, however, may also have beneficial effects, including vasodilatation, antithrombotic activity, and inhibition of pathological tissue remodelling. This latter perspective, that bradykinin might have beneficial actions, underpinned the hypothesis that the combination of an ACE inhibitor and ARB might be better than an ACE inhibitor alone in heart failure (HF). In other words, the ARB would maximize RAS inhibition and the ACE inhibitor would augment bradykinin.

Less tangible differences between ARBs and ACE inhibitors were also postulated. Because an ARB displaces angiotensin II from the AT₁ receptor, ligand is available to bind to other AT receptor subtypes, the number and activity of which remain uncertain in humans (Fig. 40.1). One theory is that stimulation of an AT₂ receptor leads to biological actions that oppose those resulting from AT₁ receptor activation, i.e. actions that might be beneficial in HF.

These considerations led to the hypothesis that ARBs might be more effective, better tolerated or both, compared with ACE inhibitors. The challenge, however, was to prove this in randomized controlled trials, a challenge made much more difficult by the fact that the clinical development of ARBs did not begin until about 15 years after that of ACE inhibitors. By that time, ACE inhibitors were established as the cornerstone of treatment of patients with HF and a low ejection fraction. Consequently, only four options were available: (1) demonstrate superiority or ‘comparability’ (noninferiority) of the new treatment to an ACE inhibitor in a head-to-head comparison, (2) demonstrate additional benefit when an ARB is added to an ACE inhibitor, (3) demonstrate benefit from ARBs in patients intolerant of an ACE inhibitor, and (4) demonstrate benefit from ARBs in new patient groups in which ACE inhibitors had not been shown to be of value.

This approach was pursued in the second Evaluation of Losartan in the Elderly (ELITE-2) trial which compared losartan 50 mg once daily with captopril 50 mg three times daily (Table 40.1).² The design and dosing used in ELITE-2 was based on the smaller ELITE pilot study designed to compare the renal tolerability of the same two drugs used in the same doses.³ Although the proportion of patients experiencing a rise in creatinine of 26.5 µmol/L (0.3 mg/dL) was similar in the two treatment arms, another, unexpected, finding in ELITE caused enormous excitement. Although there were few deaths overall in this relatively small and short-term study, their distribution between treatment groups was unequal, with 17/353 (4.8%) deaths in patients assigned to losartan and 32/370 (8.7%) deaths in those assigned to captopril, giving a relative risk reduction in death for losartan compared with captopril of 46% (95% CI 5–69%, p = 0.035).

With 3152 patients and over 530 deaths (greater than 10 times ELITE 1), occurring over a median follow-up of 18 months, ELITE-2 was statistically powered to provide a much more reliable comparison of losartan with captopril. In ELITE-2 the mortality difference between the two treatments was not significant, although there was a trend in favour of the ACE inhibitor with 250/1574 (15.9%) deaths in the captopril group and 280/1578 (17.7%) deaths in the losartan group, hazard ratio 1.13 (97.5% CI 0.95–1.35, p = 0.16) (Fig. 40.2).² With the trend toward better survival in the captopril group it was not even possible to conclude that losartan was ‘as good as’ (not inferior to) an ACE inhibitor in this patient population. With hindsight, it is possible to speculate that the design of ELITE-2 and a sister study, Optimal Therapy in Myocardial Infarction with the Angiotensin II Antagonist Losartan (OPTIMAAL), in acute myocardial infarction (MI) using the same drugs and dosing strategy, was fatally flawed because the dose of losartan used was too low (although this would not have been obvious at the time, given the apparent benefit of 50 mg daily in ELITE). That 50 mg daily may have been an inadequate dose was suggested by the two subsequent trials in patients with hypertension and diabetic nephropathy where losartan used in a target dose of 100 mg daily led to a statistically significant clinical benefit.^4,5

To test this hypothesis, a further trial in HF was set up. In that trial, Heart failure Endpoint evaluation of Angiotensin II Antagonist Losartan (HEAAL), patients with HF and a low ejection fraction who were intolerant of ACE inhibitors, were randomly assigned to treatment with losartan 50 mg once daily or losartan 150 mg once daily (Table 40.1).⁶ After a median follow-up of 4.7 years, patients assigned to the higher dose of losartan were significantly less likely to have died or been admitted to hospital for worsening HF (828/1927, 43%) than those randomized to the lower dose (889/1919. 46%), hazard ratio 0.90 (95% CI 0.82–0.99), p = 0.027 (Fig. 40.3). The numbers experiencing a cardiovascular death or HF hospitalization were 698 and 771, respectively, giving a hazard ratio of 0.88 (95% CI 0.79–0.97), p = 0.011. Of interest was the finding that the incremental benefits of higher-dose losartan were obtained with few additional serious adverse effects.

The story of losartan in HF illustrates many of the pitfalls in drug development including the importance of dose selection and the potential for misleading findings (or the misinterpretation of findings) from small studies. Although we shall never know, it is interesting to speculate what the result of ELITE-2 (and the history of ARBs) might have been if 150 mg of losartan had been used instead of 50 mg.

This was the second approach to the use of ARBs tested in a large outcome trial and was clearly an approach testing a completely different hypothesis than ELITE-2. The first of the two trials to adopt the strategy was the Valsartan Heart Failure Trial (Val-HeFT) (Table 40.1).⁷

Val-HeFT was designed with the coprimary outcomes of death from any cause and the combination of death or any of hospitalization for heart failure, administration of intravenous inotropic or vasodilator drugs for 4 h or more without hospitalization or cardiac arrest with resuscitation. The ‘alpha’ was split between the two endpoints. With a total of 979 deaths occurring during a mean follow-up of 23 months, survival was not improved with the addition of the ARB to conventional therapy: 495/2511 (19.7%) deaths in the valsartan group and 484/2499 (19.4%) deaths in the placebo group, relative risk 1.02; (98% CI 0.88–1.18, p = 0.80) (Fig. 40.4). On the other hand, the rate of composite coprimary outcome was reduced by valsartan (RR 0.87; 97.5% CI 0.77–0.97, p = 0.009) (Fig. 40.5). The major contributor to the reduction was fewer hospitalizations for HF in the valsartan group (13.8% valsartan vs 18.2% placebo). The benefit was obtained at the cost of a small increase in renal dysfunction and hyperkalaemia.

A small subgroup of patients (7%, n = 366) were not treated with an ACE inhibitor at baseline. In this subgroup, the reduction in the mortality/morbidity composite outcome with valsartan was 44%.⁸

A second trial also tested the strategy of adding an ARB to an ACE inhibitor, this time candesartan (in a target daily dose of 32 mg once daily). This Candesartan in Heart failure: Assessment of Reduction in Mortality and morbidity (CHARM)-Added, differed from Val-HeFT in several important respects (Table 40.1).⁹ First, by design, patients were at higher risk because those in NYHA class II had to have been hospitalized for a cardiac reason in the previous 6 months. Secondly, and also by design, all patients had to be taking an ACE inhibitor, and investigators were encouraged to optimize patients’ dose of ACE inhibitor based on evidence-based guidelines and tolerability. Because CHARM-Added began enrolling after Val-HeFT, use of β-blockers had become more prevalent and treatment with these drugs (and spironolactone) was considerably more frequent in CHARM-Added than Val-HeFT.

In CHARM-Added, after a median follow-up of 41 months, 483/1276 (38%) patients in the candesartan group experienced a cardiovascular death or HF hospitalization compared with 538/1272 (42%) patients in the placebo group, hazard ratio 0.85 (95% CI 0.75–0.96), p = 0.011 (Fig. 40.6). As in Val-HeFT, there was also a more striking reduction in both the total number of patients hospitalized (17%) and in the number of admissions, for worsening HF (27%).

The results of Val-HeFT and CHARM-Added led to much debate about adequacy of background ACE inhibitor dosing, and the US Food and Drug Administration requested subgroup analyses of CHARM-Added to examine the effect of candesartan according to baseline ACE inhibitor dose.¹⁰ These analyses showed that the effect of candesartan was consistent, irrespective of baseline ACE inhibitor dose, and even patients taking very large doses of ACE inhibitor seemed to obtain the same benefit from candesartan as that observed in the overall CHARM-Added trial population (Fig. 40.7). However, as this conclusion is based on retrospective subgroup analyses, it is open to criticism and, with hindsight, a better design for both of these trials would have been to mandate treatment with an evidence-based dose of a proven ACE inhibitor for all patients.

Such as design was used in subsequent trials in patients with acute infarction (Valsartan in Acute Myocardial Infarction Trial Investigators. VALsartan In Acute myocardial iNfarcTion trial, VALIANT) and in patients with atherosclerotic arterial disease (Ongoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial, ONTARGET).^11,12 In both of the latter trials, addition of an ARB to a ‘full dose’ of ACE inhibitor resulted in no additional clinical benefit (but did cause more adverse effects). These findings have led many to doubt the value of adding an ARB in patients with HF taking a full dose of ACE inhibitor. While it is possible that Val-HeFT and CHARM-Added might not have shown any benefit from adding an ARB if designed like VALIANT and ONTARGET, it is also possible that HF is different and that greater RAS activation occurs in this condition and that more intense RAS blockade leads to greater clinical benefit. One piece of evidence supports this latter hypothesis. The Randomized Evaluation of Strategies for Left Ventricular Dysfunction (RESOLVD) pilot study was a precursor of CHARM in which patients were randomized to various doses of candesartan, enalapril 20 mg per day, or the combination of enalapril and candesartan (and also subsequently rerandomized to placebo or metoprolol).¹³ In RESOLVD, addition of candesartan to full-dose enalapril led to greater reverse left ventricular remodelling and neurohumoral suppression, compared with placebo, providing mechanistic support for the findings of CHARM-Added.

Another question that is frequently asked is much more difficult to answer, namely what should be added next after an ACE inhibitor (and β-blocker)—an ARB or an aldosterone antagonist? Until recently, for patients with milder heart failure, the only direct evidence was for an ARB as aldosterone antagonists had not been studied in mild chronic heart failure. At the time of writing, however, a trial filling this gap, the Eplerenone in Mild Patients Hospitalization And SurvIval Study in Heart Failure (EMPHASIS-HF) has just been reported, showing an impressive effect of eplerenone on death from any cause and all-cause hospitalization. These new findings make aldosterone antagonism the preferred next treatment-step after use of an ACE inhibitor and beta-blocker; i.e, an ACB should only be the third neurohumoral antagonist if a patient cannot tolerate an aldosterone blocker.¹⁴ Given the very different mechanism of action of the two types of treatment, perhaps the most pertinent question now is whether both drugs should be added, although there is no evidence to support this suggestion and such an approach must carry a significant risk of renal dysfunction and hyperkalaemia.

As an aside, it is worth mentioning that early termination of trials for benefit is likely to exaggerate the effect of treatment. Indeed, if CHARM-Alternative or Added (or the pooled low ejection fraction cohort) had been stopped early, highly statistically significant and substantial reductions in the primary endpoint and death from any cause would have been reported.¹⁵ Both the CHARM-Programme and the SOLVD-Treatment trial¹⁶ before it, show that with long-term follow-up, the early benefit of treatment becomes less prominent over time.

A minority of patients cannot tolerate an ACE inhibitor because of cough, and in such patients an ARB may be a useful alternative. Proof of this is provided by the findings of the CHARM-Alternative trial in which 334/1013 (33%) of patients with low-ejection-fraction HF assigned to candesartan experienced a cardiovascular death or HF hospitalization over a median follow-up of 33.7 months compared with 406/1015 (40%) of patients assigned to placebo, hazard

ratio 0.77 (95% CI 0.67–0.89), p = 0.0004 (Table 40.1 and Fig. 40.8).¹⁷ Similarly large benefits were seen with valsartan in the small subset (7%, n = 366) of patients not treated with an ACE inhibitor at baseline in Val-HeFT.⁸

It is important to point out that a ‘full’, evidence-based, dose of ARB is just as likely to cause hypotension and renal dysfunction as an ACE inhibitor—the best evidence for this comes from VALIANT.¹¹ Consequently, if patients have genuinely been unable to tolerate an ACE inhibitor because of these adverse effects, they are unlikely to tolerate an ARB either. It is only cough and angio-oedema that are less likely with an ARB than an ACE inhibitor.

The last strategy adopted during the development of ARBs was to examine the potential role of the drugs in patient populations where ACE inhibition had not been tested. For heart failure, this meant patients with preserved ejection fraction (HeFPEF) or so called ‘diastolic’ heart failure. Two trials addressed this question.^18,19 The first was one component trial of the CHARM-Programme (Table 40.1). During screening of patients with heart failure, CHARM investigators could enrol patients with LVEF over 40% into CHARM-Preserved.¹⁸ In this trial, 333/1514 (22%) patients assigned to candesartan experienced a cardiovascular death or HF hospitalization over a median follow-up of 36.6 months compared with 366/1509 (24%) patients assigned to placebo, hazard ratio 0.89 (95% 0.77–1.03), p = 0.118 (Fig. 40.9). Although the primary outcome did not differ significantly between treatments, there were reductions in the proportion of patients admitted to hospital with worsening HF (15%) and in the number of hospitalizations for this problem (29%) with candesartan. So, while not a ‘positive’ trial, CHARM-Preserved did suggest that RAS blockade might be of benefit in patients with HeFPEF.

A second study sought to confirm or refute this possibility. I-PRESERVE, randomized patients to placebo or 300 mg of irbesartan, a dose previously used successfully in a trial in patients with diabetic nephropathy.²⁰ Patients were required to have an ejection fraction of 45% or above (Table 40.1).¹⁹ The primary outcome in I-PRESERVE was the composite of death or hospitalization for a cardiovascular cause (heart failure, myocardial infarction, unstable angina, arrhythmia, or stroke). During a mean follow-up of 49.5 months, 742/2067 (35.9%) patients in the irbesartan group and 763/2061 (37.0%) patients in the placebo group experienced the primary outcome, hazard ratio 0.95 (95% CI 0.86–1.05), p = 0.35 (Fig. 40.10). No other outcome was favourably affected either.

These disappointing findings led to much discussion about the similarities and differences of CHARM-Preserved and I-PRESERVE and why the two trials appeared to give a different result (of course, strictly speaking, CHARM-Preserved showed no between treatment-difference in its primary outcome and, therefore, interpreted rigorously, did not differ from I-PRESERVE). If there is a difference, there are at least two possible explanations. First, the dose of irbesartan used was probably only half as effective as blocking the action of angiotensin II as 32 mg of candesartan (and the blood pressure reduction in I-PRESERVE was about half that in CHARM-Preserved).²¹ On the other hand, CHARM-Preserved enrolled a segment of patients with what might be called ‘mild systolic dysfunction’, i.e. those with a LVEF of 41–45% who might have had more to gain from RAS blockade. Whatever the correct perspective is, it is certain that ARBs are of unproven value in HeFPEF, unlike low-LVEF heart failure.

A prespecified analysis of the CHARM-Programme that received less attention than it might have done was the planned pooling of the two low-LVEF trials (CHARM-Alternative and -Added).²² This was done because, although neither had the statistical power to look at all-cause mortality individually, when combined there was sufficient power to look at this outcome. Among patients randomly assigned to candesartan, followed for a median of 40 months, 642/2289 (28.0%) died compared with 708/2287 (31.0%) in the placebo group, hazard ratio 0.88 (95% CI 0.79–0.98), p = 0.018 (Fig. 40.11).

It is interesting to compare this finding to the SOLVD-Treatment trial in which enalapril 20 mg daily given for a mean of 41.4 months reduced mortality (452/1285, 35.2%) compared with placebo (510/1284, 39.7%), relative risk reduction 16% (95% CI 5–26%), p = 0.0036. The approximate relative risk reduction with candesartan 12% (95% CI 2–21%) compares favourably with (and completely overlaps) that of enalapril, although, of course, 56% of patients in the CHARM low EF population were taking an ACE inhibitor (compared with 0% in the placebo group of SOLVD-T) and 55% were taking a β-blocker (compared with 8% in SOLVD-T).¹⁶

The most current guidelines show some inconsistency in their recommendations about the role of ARBs in the treatment of low-LVEF HF (Table 40.2 and Table 40.3).^23,–27 Although most guidelines give a level A recommendation for the use of ARBs in patients intolerant of an ACE inhibitor, this level of evidence requires data from ‘multiple randomized trials or meta-analyses’ (Table 40.2): multiple trials do not exist—unless the subset of Val-HeFT patients not treated with an ACE inhibitor at baseline is equated to a second trial, in addition to CHARM-Added.

There is more heterogeneity in the recommendations regarding use of ARBs in addition to an ACE inhibitor and the recommendation also varies between the summary and full guideline in certain cases (and has changed between the 2006 and 2009 American College of Cardiology/American Heart Association guideline). It is also not clear what outcome the recommendation refers to. The European Society of Cardiology (ESC) guideline is more precise in this respect, indicating that with both Val-HeFT and CHARM-Added showing a reduction in admissions for worsening heart failure, the level of evidence of clinical benefit is greater than for reducing cardiovascular mortality: Val-HeFT did not show a mortality benefit but it was observed in CHARM-Added.

The use of an ARB should follow the same principles as that of an ACE inhibitor and the starting doses, target doses, and approach to dose up-titration should be guided by the relevant clinical trials, as summarized in Box 40.1.²⁸

ARBs provide an alternative to ACE inhibitors in patients intolerant of the latter because of cough or angio-oedema. When added to an ACE inhibitor in patients with low ejection fraction HF, ARBs reduce hospital admission for worsening HF and candesartan has also been shown to reduce death from cardiovascular causes when used in this way. Use of both an ACE inhibitor and ARB together, however, necessitates careful biochemical surveillance for renal dysfunction and hyperkalaemia and is no longer the preferred next treatment-step after an ACE inhibitor and beta-blocker (an aldosterone antagonist is now preferred).