If a Patient Does Not Require Treatment with Loop Diuretics, Do they Really Have Heart Failure?

Abstract

This article refers to ‘Utility of loop diuretics as a marker of risk in clinical trials of heart failure with preserved ejection fraction: Analysis from the Swedish Heart Failure Registry’ by F. Chouairi et al., published in this issue on pages xxx. Clinical care is about trying to meet a patient's needs. Clinical research is about addressing unmet needs, either by implementing existing solutions or finding new ones. If a trial enrols patients whose needs are already met by existing therapy, then it has little chance of success, and participants are more likely to be harmed than to benefit. Conventionally, trials often apply ‘enrichment’ criteria to ensure inclusion of patients with more severe symptoms or at higher risk. Conceptually, and practically, it is better to exclude patients with few symptoms or at low risk, which increases the average risk of those patients who are included. However, it is modifiable risk that really matters; patients at very high risk might be too sick to benefit and those at low risk have little opportunity to do so.1 Clinical trials should enrol patients with unmet needs for whom there are reasonable prospects of meaningful benefits, either in terms of symptoms or prognosis. This respects patients' time, effort and personal risk.2 Importantly, it is the absolute benefit that really counts for practicing clinicians, service organization, healthcare costs and, most importantly, patients.2 A relative risk reduction of 30% might sound impressive but a reduction from 1.0% to 0.7% per year (three patients per thousand years of treatment) may not be clinically relevant. Similarly, a statistically significant improvement in symptoms or quality of life in a large trial might translate into negligible benefits for individual patients.2 In this issue of the Journal, Chouairi et al.3 report the characteristics, loop diuretic use and outcomes of 25 986 patients with heart failure and a left ventricular ejection fraction (LVEF) ≥40% who were enrolled in the Swedish Heart Failure Registry. Median age was 77 (interquartile range 69–84) years, 43% were women and most were multi-morbid. Only 59% were treated with loop diuretics at any dose (>99% furosemide) and only 25% with furosemide >40 mg/day. Guidelines strongly recommend loop diuretics to treat symptoms and signs of heart failure. However, for patients who were not taking loop diuretics in the Swedish registry, 76% were symptomatic (New York Heart Association functional class II or worse) and 39% had clinical signs of congestion. Of course, problems other than heart failure might have contributed to symptoms and impaired functional capacity. However, patients who were not treated with loop diuretics lacked objective evidence of important cardiac dysfunction, throwing doubt on the diagnosis of heart failure. LVEF was not markedly depressed. Atrial dilation, the hallmark of heart failure on cardiac imaging, was not reported.4 Considering that 52% of patients were in atrial fibrillation, plasma concentrations of N-terminal pro-B-type natriuretic peptide, indicating circulatory congestion,4 were not markedly increased (thresholds for first, median and third quartiles were 326, 869 and 1960 ng/L, respectively). Despite having many clinical characteristics predicting poor outcome, annual mortality was much lower (5%) for those not taking loop diuretics, compared to those taking furosemide ≤40 mg/day (13%) or >40 mg/day (24%); about half of all deaths in each group were cardiovascular. For heart failure hospitalizations, the difference amongst groups was even more striking with 7, 21 and 44 events per 100 patient-years of follow-up, respectively. These data corroborate previous reports suggesting that patients with a diagnostic label of heart failure who are not prescribed loop diuretics have relatively low rates of events that may not be substantially higher than for the general population with a similar age and comorbidity profile.5-7 Such patients are unlikely to receive a substantial benefit on morbidity or mortality, in absolute terms, from an intervention within the 2-to-3-year timeframe of a typical clinical trial of heart failure. Indeed, it is uncertain whether patients who are not treated with loop diuretics should be considered to have heart failure at all (Figure 1A). Congestion is the hallmark of a failing heart. Cardiac dysfunction that does not lead to congestion is just cardiac dysfunction and has a relatively benign prognosis.8 Cardiac dysfunction leading to congestion indicates that the heart is failing, which should suffice for a contemporary universal definition of heart failure.4 However, current guidelines on diagnosis require symptoms and/or signs to be present, which often go unrecognized and unreported.9 Guidelines give a class I recommendation for loop diuretics ‘to reduce the signs and/or symptoms of congestion…’, making loop diuretics the fifth pillar of treatment for patients with heart failure and an LVEF ≤40% and one of only two current pillars for those with an LVEF ≥50%.10, 11 Accordingly, treatment with loop diuretic is, or should be, part of the current definition of heart failure (which requires symptoms and signs), as was suggested by the very first European Society of Cardiology guideline on the diagnosis of heart failure12 (Figure 1A). When loop diuretics are given at a small dose (e.g. furosemide 20 mg/day), this suggests that the clinician is unsure about the presence of congestion, leading to therapeutic prevarication. Unfortunately, the characteristics and outcomes for those taking exactly 40 mg/day, the most common dose in practice, were not reported by Chouairi et al. When loop diuretics are given at standard (~40 mg/day of furosemide) or higher doses (≥80 mg/day), then it is reasonably certain that the clinician is trying to treat congestion, thereby fulfilling the criteria for heart failure. Conversely, if a patient is not receiving a loop diuretic, then either they do not have symptoms and do not fulfil the current definition of heart failure or they have symptoms but are not being managed according to guidelines, which could be considered a deficit in care. Most people prescribed loop diuretics have no diagnostic record of heart failure.9, 13 Only a minority of patients treated with loop diuretics appear to have resistant hypertension or end-stage renal disease, but they do have a prognosis similar to patients with heart failure9 (Figure 1B,C). Either the diagnosis of heart failure (or a disease with a similar prognosis) has been missed or loop diuretics have an adverse effect on prognosis.9 We cannot assume that loop diuretics are safe. If prescribing loop diuretics is an indicator of symptoms and signs of congestion and most of these patients have cardiac dysfunction, which will certainly be the case, then the prevalence of heart failure might be three times higher than current epidemiology suggests.9, 14 For many patients, perhaps most, the diagnosis of heart failure is only made once symptoms and signs are severe enough to precipitate hospitalization.15 The onset of symptoms and prescription of loop diuretics often precedes such admissions by months or years.9 The golden therapeutic opportunity may have already passed. Indeed, most people with heart failure probably die before the diagnosis is ever made. Clinicians should be alert to the possibility of heart failure as a cause of symptoms. All patients treated with loop diuretics should be investigated for cardiac dysfunction, either before diuretics are initiated or at the next opportunity. Worldwide, there might be >100 million people treated with loop diuretics who have not been investigated for or diagnosed with heart failure. Treatment with loop diuretics should also be an integral part of the diagnostic process for heart failure (Figure 1A). Patients with cardiac dysfunction without evidence of congestion do not require loop diuretic therapy and do not have heart failure. Patients with cardiac dysfunction treated with loop diuretics fulfil the definition of heart failure unless loop diuretics can be withdrawn without recurrence of congestion. Patients with end-stage renal disease or resistant hypertension receiving loop diuretics often have cardiac dysfunction and symptoms on exercise (i.e. heart failure). In the absence of congestion, other classes of anti-hypertensive agents that are known to reduce cardiovascular risk, which is not the case for loop diuretics, might be preferred. If loop diuretics are used cosmetically to improve ankle swelling, substitution with other agents such as thiazide diuretics, mineralocorticoid receptor antagonists or sodium–glucose cotransporter 2 inhibitors should be considered. For patients with a previous diagnosis of congestive heart failure, if loop diuretics can be withdrawn without recurrence of congestion, then this might be considered resolution of transient heart failure (Figure 1A).12 Chouairi and colleagues conclude that for heart failure with preserved or only mildly reduced LVEF, prescription of loop diuretics should be used as an enrichment factor to ensure that trials have a high enough event rate to provide statistical power to show treatment effects within a 2-to-3-year timeframe, a standard duration for heart failure trials.3 This is an excellent idea but does not go far enough. Patients with any form of persistent cardiac dysfunction should be considered to have heart failure only if they have evidence of congestion, for which, when symptomatic, guidelines strongly recommend loop diuretics. Patients with cardiac dysfunction who are not receiving loop diuretics will have relatively low rates of hospitalization and death but might still be considered for longer-term trials designed to prevent progression to symptomatic heart failure. Conflict of interest: J.G.F.C. reports grants from British Heart Foundation; personal fees from Abbott, AstraZeneca, Biopeutics, Holosis, Idorsia, Medtronic, and Vectorious; grants and personal fees from Bayer, Bristol Myers Squibb, CSL-Vifor and Pharmacosmos; personal fees and non-financial support from Boehringer Ingelheim and NI Medical; non-financial support from Corvia; stock options from HeartFelt; grants and stock options from Viscardia. All other authors have nothing to disclose.

Topics & Concepts