Losing ALTITUDE? How should ASTRONAUT Launch into ATMOSPHERE

Direct renin inhibitors (DRIs) have become established therapy in the treatment of systemic hypertension and are currently being explored in a number of disease states including heart failure.1 The putative advantages of these agents are that they block the renin–angiotensin–aldosterone system (RAAS) effectively at its source, i.e. blocking generation of angiotensin I via its conversion from angiotensinogen. This is particularly relevant in conditions such as heart failure where the RAAS is maximally activated, and downstream RAAS blockers such as ACE inhibitors, ARBs, and mineralocorticoid receptor antagonists lead to further renin activation, thus potentially overcoming the downstream blockade.1 This is evidenced by increases in angiotensin II and aldosterone during chronic ACE inhibition.2 A potential role for the (pro)renin receptor and the impact of DRIs on this is also currently being explored.3 On this background, a number of large-scale outcome trials have been conducted with DRIs (Table 1), with mixed results. As the Executive Committee of the Aliskiren Trial to Minimize OutcomeS in Patients with HEart failure (ATMOSPHERE) study of the DRI, aliskiren, in systolic chronic heart failure, the findings of these large-scale studies have been of particular interest to us. The first of the major clinical outcome trials to report was the Aliskiren Trial In Type 2 Diabetes Using Cardio-Renal Disease Endpoints (ALTITUDE)4 study of aliskiren in patients with type 2 diabetes mellitus and nephropathy, the majority of which had hypertension. Very few had LV dysfunction or heart failure. Importantly, patients were all treated with a background ACE inhibitor or ARB, but not both. The main finding of this trial was that there was no difference between the effect of aliskiren and placebo on the primary composite endpoint comprising cardiovascular death, resuscitated sudden death, non-fatal myocardial infarction, non-fatal stroke, and unplanned hospitalization for heart failure. This was also true for both the cardiovascular and renal secondary endpoint, as well as all-cause death. One finding of interest was an excess of strokes (fatal and non-fatal) which led to a change in the product information and consent form for aliskiren studies going forward. The second major study of interest to ATMOSPHERE has been the recently published Aliskiren Trial on Acute Heart Failure Outcomes (ASTRONAUT)5 study. This study is of particular relevance because the patient population (comprising those in the period immediately after an acute heart failure admission with low LVEF) are much closer to an ATMOSPHERE-type patient than the patients with diabetes in ALTITUDE. The primary endpoint of ASTRONAUT was cardiovascular death or heart failure re-hospitalization (the same primary endpoint as in ATMOSPHERE) over the first 6 months. However, patients were all followed for 12 months. Overall, there was a non-significant 8% relative risk reduction favouring aliskiren vs. placebo. Pre-defined subgroup analyses did not reveal any heterogeneity across a range of subgroups. Although there was no evidence of an interaction between the effect of aliskiren and baseline diabetes status for the primary outcome (at 6 months), there appeared to be one at 12 months whereby patients with diabetes seemed to fare worse on aliskiren than on placebo (with the opposite trend in patients without diabetes). The P-value for interaction was 0.03 for cardiovascular death or heart failure rehospitalization and <0.0003 for all-cause mortality, both at 12 months. Interestingly, fatal or non-fatal stroke, which was the major cause for concern amongst patients with diabetes in ALTITUDE, was not found to be increased with aliskiren in ASTRONAUT. Indeed, overall there was a 37% (non-significant) trend to a reduction in the risk of stroke with aliskiren. This example points to the dangers of attaching too much significance to secondary endpoints, especially when the number of events is relatively small and the mechanistic plausibility underlying such an unexpected finding is lacking. For that reason, the 53% reduction (P = 0.009) in myocardial infarction with aliskiren in ASTRONAUT has not been emphasized. Following ASTRONAUT, some health authorities requested that patients with diabetes in ongoing aliskiren trials including ATMOSPHERE should have study drug discontinued and that no more patients with diabetes should be enrolled in these studies. In light of this requirement, the Executive Committee (together with the sponsor) felt that it would be confusing and impracticable to have different rules related to patients with diabetes in the same trial in different parts of the world and recommended a single global strategy. The Executive Committee feel strongly that ATMOSPHERE remains a scientifically valid study, notwithstanding withdrawal of diabetics. Having said that, our preference would have been for the study to continue unchanged. This is based on a number of considerations. First, both ALTITUDE and ASTRONAUT studied dual (or even triple) RAAS blockade—in both studies all patients were taking background ACE inhibitor or ARB therapy. In contrast, there are three equally sized study arms in ATMOSPHERE (aliskiren monotherapy, enalapril monotherapy, and aliskiren and enalapril), i.e. two-thirds of patients in this study are not per protocol exposed to dual background RAAS blockade.6 Secondly, and perhaps most importantly, the only group with access to unblinded data is of course the Data Monitoring Committee (DMC) for ATMOSPHERE (which was also the DMC for ASTRONAUT), comprising five of the most experienced and astute heart failure clinical trialists and statisticians worldwide. Importantly, the DMC have consistently (before, during and after withdrawal of patients with diabetes) recommended that the study continue unchanged. Ample evidence suggests that it is unwise to second-guess independent data and safety monitoring boards.8 There has also been a vastly greater exposure of patients to alsikiren in ATMOSPHERE compared with ASTRONAUT. As of January 2013, almost 10 000 patient-years of exposure had occurred in ATMOSPHERE compared with only 1100 in ASTRONAUT. Thus, there has been a nine-fold greater exposure to aliskiren in ATMOSPHERE without raising data and safety monitoring board concerns. Additionally, earlier studies of RAAS blockade in diabetics have consistently demonstrated the absence of major concerns regarding excess clinical events. This has been shown by Shekelle et al.9 with regard to ACE inhibitors, and individual ARB studies10,11 also support this contention. Furthermore, earlier (albeit much smaller) systolic heart failure and LV dysfunction studies with aliskiren itself have not shown any evidence of particular efficacy or safety concerns in diabetics.12,13 Finally, in both ALTITUDE and ASTRONAUT, the major safety concerns were renal dysfunction and hyperkalaemia, both of which are recognized problems with dual RAAS blockade and of which patients with diabetes are at particular risk. These risks are minimized in ATMOSPHERE by a key design feature which is exposure of all patients to dual aliskiren and enalapril therapy during the active run-in (sequentially after an enalapril monotherapy active run-in period; Figure 1).6 Patients developing worsening renal function or hyperkalaemia (or hypotension) are excluded according to strict protocol-driven criteria. Indeed, 20.3% of screened patients have been withdrawn during the run-in, 11.8% failing during the ACE inhibitor run-in and 8.3% during the aliskiren plus enalapril run-in. In conclusion, the Executive Committee for ATMOSPHERE believes that neither ALTITUDE nor ASTRONAUT (both of which tested the combination of aliskiren with an ACE inhibitor or ARB) is particularly relevant to the majority of patients in ATMOSPHERE in which two of the three arms are testing either enalapril or aliskiren monotherapy. Moreover, the Executive Committee believes that use of two RAAS-blocking agents has a solid track record of benefit in patients with chronic systolic heart failure. This has been shown in studies combining ARBs with ACE inhibitors and either of these agents combined with a mineralocorticoid receptor antagonist. The benefits seen were consistent in patients with and without diabetes. In marked contrast to both ALTITUDE and ASTRONAUT, neither of which had an active run-in period, all patients randomized in ATMOSPHERE have tolerated a period of treatment with enalapril followed by a period of treatment with both enalapril and aliskiren. We believe that this has identified and excluded patients most vulnerable to the adverse effects of dual RAAS blockade. Most importantly, ATMOSPHERE is overseen by a highly experienced DMC who have recommended that ATMOSPHERE continue, even after the publication of the results of both ALTITUDE and ASTRONAUT. Finally, it is important to note that the above views are exclusively those of the ATMOSPHERE Executive Committee and not necessarily shared by other parties. Conflict of interest: B.M. received modest payments (approximately US$5000) from Novartis, the trial sponsor, for his time spent designing and overseeing the trial as Co-PI. The employers of J.J.V.M, Glasgow University, is paid by the sponsor of ATMOSPHERE, Novartis, for his time spent as a member of the trial Executive Committee.