Acute Heart Failure: Have We Got it All Wrong?

This article refers to ‘Breathlessness at rest is not the dominant presentation of patients admitted with heart failure’ by A. Shoaib et al., on page 1283 in this issue. Much effort and expense has been spent in developing new treatments for acute heart failure with the main or sole purpose of improving dyspnoea caused by pulmonary oedema. The majority of trials have required patients to have breathlessness at rest or on minimum exertion for inclusion. However, is pulmonary oedema causing acute dyspnoea the most important problem in this condition which usually represents worsening of chronic heart failure? In a fascinating and provocative observational study, Shoaib and colleagues suggest that presentation predominantly with acute breathlessness may not be as frequent as commonly portrayed.1 These authors also make the remarkable finding that patients presenting in this way may be at lower risk than those presenting predominantly with increasing peripheral congestion. These observations even question the use of the word ‘acute’ to describe worsening heart failure leading to hospital admission as we know that the often extensive peripheral oedema exhibited by these patients does not develop overnight. In a retrospective case note review of 311 patients admitted to hospital with heart failure, the authors categorized their subjects as either those ‘short of breath at rest’ (SOBAR) or those ‘comfortable at rest but breathless on slight exertion’ (CARBOSE). Patients were said to be SOBAR if they had a respiratory rate >20 breaths/min, were breathless at rest, and received i.v. therapy shortly after admission. The remainder not meeting these criteria were classified as CARBOSE. Only 43% of patients were SOBAR on admission. This observation is consistent with the findings of several hospital registries which reported similar proportions of patients with dyspnoea at rest on admission (Figure 1).2-13 In the Acute De-compensated Heart Failure National Registry (ADHERE)2 and the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients With Heart Failure (OPTIMIZE HF),3 registry only 33% and 43%, respectively, of patients admitted to hospital in the USA with ‘acute’ heart failure suffered from dyspnoea at rest. The EuroHeart heart failure survey reported that 40% had dyspnoea at rest.4 While more patients may become breathless on provocation, e.g. exercising or lying supine, these observations do question targeting improvement in dyspnoea as the primary endpoint in clinical trials.14 The more novel finding in the present report is that CARBOSE patients appeared to be at higher risk of death at 3 and 6 months, compared with SOBAR patients. Mortality was higher, despite similar markers of disease severity in the two groups at baseline, e.g. plasma NT-proBNP concentrations and the proportions of patients with moderate to severe LV dysfunction and severe valve disease. The authors speculate that the CARBOSE group may have been at higher risk because of more frequent right heart dysfunction, although echocardiographic data are not available to support this hypothesis. Alternatively, the SOBAR group, presenting in a more acutely symptomatic fashion, may have received more prompt and intensive management with more careful monitoring. The overall findings of the study of Shoaib et al. and the possible explanations for them clearly require replication and further investigation, in larger, prospective, cohorts. However, our own clinical experience is very consistent with that of the authors. Most patients we see admitted to our hospital with heart failure present with a syndrome of increasing peripheral (and often pulmonary) congestion, and relatively few present with acute pulmonary oedema. Often the history is of days or even weeks of deterioration, i.e. with a subacute rather than an acute presentation. Shoaib and colleagues may also be right about the different prognosis of these two types of patients, given that persistent congestion is related to poor outcomes in patients hospitalized with heart failure.15, 16 The present study also highlights some of the differences between ‘real-world’ patients with heart failure and those recruited into trials, and this may have important implications for future trial design in patients hospitalized with heart failure. The current findings even prompt us to think about how many of these hospitalizations might be avoided, i.e. by identifying and treating worsening congestion before it leads to admission. Comparison of the characteristics of hospitalized heart failure patients in registries with the inclusion/exclusion criteria for trials in ‘acute’ heart failure (Tables 1 and 2) suggests that most would be ineligible for these studies.1, 6-13 Most trials required dyspnoea at rest or on minimal exertion, along with other findings such as an increased respiratory rate and radiological or clinical evidence of pulmonary oedema. Not only that, but, as the authors point out, by recruiting predominantly SOBAR patients, trials may select lower risk individuals, making it harder to show an effect of treatments on mortality. Shoaib and colleagues illustrate these points nicely by citing the recent RELAXin in Acute Heart Failure (RELAX-AHF) trial in which participants were required to have dyspnoea at rest or on minimal exertion, pulmonary congestion on a chest X-ray (CXR), a systolic blood pressure (SBP) >125 mmHg, and to have received a bolus of i.v. diuretic.13 Only 49% of the patients in the cohort of Shoaib and colleagues had congestion on CXR and only 61% had received an i.v. diuretic during the first 24 h of admission. Of the whole cohort, just 43% (57% SOBAR and 33% CARBOSE) had an SBP >125 mmHg between 6 and 12 h after presentation. In other words, many participants in this ‘real-world’ cohort would have been ineligible for RELAX-AHF, and those that were eligible would have been predominantly in the lower risk SOBAR group. We think patients presenting predominantly with increasing peripheral congestion deserve greater attention. Can these patients be identified earlier and admission avoided? Can this be done using simple measures such as daily weighing and prompt adjustment of oral diuretic dosing (or addition of a complementary diuretic)?17 Alternatively, could non-invasive or invasive remote monitoring, or even biochemical surveillance (e.g. with natriuretic peptides) be the answer?18, 19 Do these patients eventually reach the point where gut oedema, renal venous congestion, or some other factor renders them unresponsive to oral diuretic therapy?20, 21 If so, can i.v. diuretic therapy be given in the community and reduce the need for inpatient therapy?22 The artificial division of heart failure into ‘acute’ and ‘chronic’ subtypes has been unhelpful in thinking about the management of what is for most patients a chronic relapsing syndrome. Our goal should be to prevent or minimize relapses, most of which will be manifest as congestion. Whether managing these patients in the ambulatory or hospital setting, how do we define adequate or sufficient decongestion and how important is it to achieve this? Does achievement of ‘optimal’ decongestion lead to better outcomes, i.e. reduced risk of (re-) admission, death, or both? It does seem remarkable to us that although congestion has been recognized as the hallmark of heart failure for centuries we still do not really know how to measure or treat it. R.T.C. is supported by a project grant from the British Heart Foundation, grant number PG/13/17/30050. Conflict of interest: none declared.