Poster Session 1

In recent months, neutral results have been reported for several trials in acute heart failure (AHF): two large phase 3 trials, RELAX-AHF-21 and TRUE-AHF2, as well as a phase 2b study, BLAST-AHF.3 These results add to a long line of disappointing drug development programmes, with the consequence of no new therapies having been developed and approved globally for AHF in more than 30 years. Hence, AHF patients continue to be treated primarily with i.v. loop diuretics, nitrates, and non-invasive ventilation. The first (loop diuretics) were shown in a recent small study to have minor effects on symptoms with no demonstrable effects on clinical outcomes while worsening kidney function,4 whereas i.v. nitrates have never been studied in AHF beyond the first few hours of admission.5, 6 As a result, the short- and long-term outcomes of AHF patients remain dire7 as evidence-based therapeutic options approved for patients with heart failure (HF) have been shown to prevent long-term adverse outcomes only in symptomatic chronic HF patients with reduced ejection fraction (EF), while the majority of patients admitted for AHF have an EF >35%.7 Furthermore, no therapy has ever been shown to improve the dismal short-term (days to months) outcomes of AHF regardless of EF. Hence, for the majority of AHF patients, and especially during the first months after an acute exacerbation, no proven therapeutic options exist. The results of TRUE-AHF follow an extensive programme investigating the effects of i.v. nesiritide, another natriuretic peptide, in AHF. Following a few smaller studies showing some early dyspnoea benefit,8 but associated with creatinine increases and a trend towards increased mortality,9, 10 a larger study was undertaken—ASCEND-HF—in which lower doses of i.v. nesiritide were administered on average as compared with the doses in the earlier studies.11 The ASCEND-HF study found that nesiritide led to minor symptomatic benefit, a small increase in symptomatic hypotension, and no increased risk of 30-day mortality or renal impairment.11 The results of the TRUE-AHF study largely replicate and confirm these observations.2 In this study, ularitide demonstrated some minor symptomatic benefit during infusion [less worsening HF (WHF) at 48 h] which did not persist to day 5. However, this minor transient effect was counterbalanced by a significantly greater mean blood pressure decrease and more patients with hypotension, and a greater mean creatinine increase. From the biomarkers' perspective, a significant reduction in NT-proBNP was observed, with no effect on troponin. However, although not reaching significance, subgroup analysis showed numerically greater benefits on cardiovascular (CV) survival rate in patients with higher baseline systolic blood pressure (SBP) and EF, as well as positive effects on the co-primary endpoint (better, no change, or worse, over the first 48 h) in patients enrolled according to the protocol. The RELAX-AHF-2 study1 followed the results of RELAX-AHF.12 In RELAX-AHF, serelaxin, when administered to patients with AHF, moderate renal impairment, and SBP >125 mmHg, was associated with a reduction in WHF events in the first days of admission and 180-day mortality.12 The results were supported by beneficial effects on early changes in creatinine, NT-proBNP and troponin.13 The RELAX-AHF-2 study1 failed to replicate this finding. In the overall study, no effect was seen on 180-day CV mortality or HF readmission/CV death. A non-significant trend towards benefit was observed with regard to WHF, although WHF occurred in fewer patients than had been observed in previous trials (<7%). In a laboratory sub-study, beneficial effects of serelaxin were suggested on cystatin-C, NT-proBNP, and troponin, largely similar in magnitude to those seen in RELAX-AHF;1 however, in this subpopulation no effects on mortality were observed. The BLAST-AHF study was a large phase 2b study following a small haemodynamic study of TRV027, a biased angiotensin antagonist expected to have both cardiac- and renal-preserving effects on top of vasodilating properties.3 The phase 2b study was changed during an interim analysis to include patients with lower baseline SBP. The main results of the study were neutral,3 although a post hoc analysis found benefits on 30-day renal function and 180-day outcomes in patients enrolled with high baseline SBP.14 TRUE-AHF, RELAX-AHF-2, and BLAST-AHF1-3 join a long line of trials evaluating interventions developed for patients admitted with AHF, where early, phase 2 studies demonstrating haemodynamic benefits, including vasodilatation-driven wedge pressure decrease and cardiac output (CO) increase as well as a substantial decrease in natriuretic peptides, led investigators to assume efficacy only to be disappointed when phase 3 results showed no long-term effects and minimal short-term effects, mostly around weight loss and to a lesser degree symptom relief.1-3, 11, 15-17 In essence, this belief that vasodilatation-driven decreases in wedge pressure and BNP in tandem with CO increase may be beneficial is driven by the assumptions that congestion is related to higher wedge pressure and to higher BNP, and that perfusion can be gauged by CO. Hence, improvements in congestion and perfusion—less 'backward' and 'forward' failure—should follow reductions in wedge pressure and/or natriuretic peptides and/or increases in CO, leading to, at a minimum, better short-term symptom relief and outcomes. However, these assumptions have not held true in AHF research. The plethora of studies1-3, 11, 15-17 showing a disconnect between intervention-induced changes in wedge pressure, BNP, and CO with long-term outcomes in AHF, and to a large degree even symptoms, should convince the scientific community that these measures, although useful in assessing the prognosis of AHF patients in general, are not good surrogates for the efficacy of interventions in AHF and should probably be abandoned. It is therefore possible that at the present time, given the lack of reliable surrogate endpoints, phase 2 trials should employ similar designs and endpoints to phase 3 studies, and simply assess treatment effects across multiple endpoints, utilizing statistical approaches aimed at evaluating which endpoints are most likely to be affected. As suggested more than 15 years ago,18 AHF is not a disease but rather a syndrome comprising multiple phenotypes. Patients enrolled in TRUE-AHF, RELAX-AHF-2, and BLAST-AHF were a mix of AHF patients of multiple phenotypes, and it is possible that interventions with vasodilating agents may work differently in variety of phenotypes. In the TRUE-AHF study, subgroup analyses showed some numerical trends in favour of active treatment in patients with baseline SBP >140 mmHg and EF >40%.2 In the BLAST-AHF study, patients who had baseline SBP >127 mmHg, and especially >140 mmHg, showed trends towards a beneficial effect of TRV027 in longer-term kidney function and 180-day outcomes.14 Hence, consideration may need to be given to further assessing in phase 2 AHF studies not only the dose–response of the intervention but also which patients appear to benefit. Investigators should carefully consider whether patients with low blood pressure, very reduced or preserved renal function, or reduced EF, should be enrolled in studies assessing an intervention with vasoactive effects, as such patients may not benefit (diluting the measured treatment effect) or may even derive harm. On the other hand, such substantial limitations on the patients enrolled will limit the applicability of the interventions in larger AHF populations, as most 'real-life' AHF patients would not meet the criteria for these studies.19, 20 Still, small progress in developing effective interventions, even for a minority of patients, may be better than no new therapies in >30 years. During the last decades the approach to clinical research has changed from tight supervision of clinical studies to one in which monitoring and oversight have been reduced, both to save resources and to enable more 'real-life' studies. In this new approach, named 'risk-based monitoring', 70–80% of onsite monitoring has been replaced with central monitoring guided by statistical flags and a remote offsite review of data by less experienced, less expensive 'central' monitors. Although such approaches may work in simple streamlined studies, such as in hypertension or hyperlipidaemia, this 'risk-based monitoring' approach, which has been implemented in the last few years without any evidence, may be counterproductive in HF research where inclusion criteria of a specific AHF phenotype and outcomes are more subjective and complex. For instance, the only way to ensure that AHF patients, and not advanced chronic HF patients, are enrolled in AHF studies requires tight supervision by an expert team. Failing to do so may lead to significant dilution of the interventions' treatment effect, as interventions developed for AHF might not be expected to benefit advanced chronic HF patients in the same way that thrombolytics will not help patients with severe chronic angina. Indeed, a post hoc analysis of the TRUE-AHF study suggested that the primary endpoint was met in a per-protocol analysis, i.e. when excluding patients identified as ineligible for the trial.2 Thus, it may be in a sponsors' best interests when developing complex HF interventions to invest in more tightly controlled studies than current trends in study conduct suggest. The scientific community should be interested in the way studies are executed, as this may be a critical stepping stone in ensuring that the signal of a new intervention's effect is not drowned in the noise created by sites wrongly interpreting the protocol or incorrectly reporting outcome measures. The attempt to demonstrate an effect on mortality in studies with AHF may also have contributed to our failure to add new therapies to our armamentarium. Given the above-mentioned limitations in AHF research, and especially those relating to study conduct, very large studies may be prohibitively difficult to manage. The repeat case of positive small phase 2 studies not reproduced in phase 3 may well relate to difficulties in executing large studies in AHF. Hence, a less ambitious primary endpoint should be adopted and agreed upon with regulatory agencies such as the short-term 'improved, no change, worsened' clinical condition used in the TRUE-AHF study or a simple 14-day WHF or death endpoint. A small progress towards treating patients with drugs known to be safe and efficacious even on short-term endpoints may be better than the current lack of evidence-based therapy, which results in patients being treated by drugs that are not proved to be either efficacious or safe. Moreover, with such simple endpoints perhaps some sponsors (such as the NIH) will be inclined to perform large enough studies to examine whether the therapies we use now (i.v. loop diuretics and nitrates) are efficacious and safe. Conflict of interest: G.C. and B.D. have received research grants from Amgen, Celyad, Novartis, Singulex, ChanRx, Laguna Pharmaceuticals, Sorbent Therapeutics, and Trevena, Inc.; and personal fees from Novartis. A.C.S. has received research grants from Menarini and Novartis and honoraria from Novartis, Cardiorentis, Amgen, Roche, Servier, CVRx, Boehringer Ingelheim, and Vifor. A.M. received speaker's honoraria from Abbott, Novartis, Orion, Roche, and Servier. A.M. also received fees as member of advisory board and/or steering committee from BMS, Cardiorentis, Adrenomed, MyCartis, Neurotronik, Sanofi, and Sphyngotec.