Incretin-based therapy for heart failure with reduced ejection fraction: why we need a trial

While obesity is common in patients with heart failure (HF) and preserved ejection fraction (HFpEF), affecting half or more of individuals with this HF phenotype, it is less well recognized that obesity is also quite frequent in patients with HF and reduced ejection fraction (HFrEF).1 Indeed, in a syndrome where cardiac cachexia was once a concern, the prevalence of obesity has been steadily rising to the point where, in recent trials, around a third of patients with this HF phenotype now have obesity (Graphical Abstract). This proportion is much greater in certain regions such as North America, and applying body mass index (BMI) categories derived from Western populations may underestimate the prevalence of this comorbidity in Asia, where body habitus is different, and visceral adiposity may be elevated despite lower BMIs.2 In PARADIGM-HF, the alternative anthropometric index, waist-to-height ratio, which better accounts for sex- and race-based differences in stature and distribution of body fat, showed that 43% of patients had a ratio >0.6 (corresponding to ‘obesity’) compared with 32% of patients with a BMI of ≥30 kg/m2.1 It has also become clear that obesity is associated with much worse symptoms and quality of life (‘health status’) in patients with HFrEF and with higher rates of hospitalization for worsening HF.1,3,4 Although unadjusted analyses continue to suggest an ‘obesity-survival’ paradox, it is improbable that obesity does not reduce life expectancy in patients with HFrEF. The apparent lack of association between higher BMI and higher mortality may be partly explained by some more favourable prognostic characteristics of patients with obesity e.g. younger age, higher systolic blood pressure, and better kidney function.1,3–5 Even if mortality rates are not increased in patients with obesity, the burden of symptoms and hospitalization undoubtedly is.1,3,4 The question of anti-obesity treatment therefore clearly arises, especially given the promising preliminary data in patients with HF and mildly reduced or preserved ejection fraction (HFmrEF/HFpEF).6,7 The striking improvements in health status and 6-minute walk distance, coupled with possible reductions in episodes of worsening HF, have led, appropriately, to the design of larger, longer-term, trials powered to test the effect of these new therapies on clinical outcomes, including cardiovascular and all-cause mortality. However, to date, we know of no such trial planned in patients with HFrEF. One reason for this is a potential concern about the safety of incretin-based therapies in these individuals.8,9 In part, this reflects prior experience in small trials in patients with HFrEF and a retrospective analysis of the EXCSCEL trial, as well as the known increase in heart rate with these compounds, likely due to a direct effect on the sinus node, which may be undesirable in HFrEF.8,9 How valid are these concerns? Tables 1 and 2 show a summary of these data plus the more recent findings from the SELECT and FLOW trials.9–13 Because most of the trials shown were not weight-loss trials, the glucagon-like peptide 1-receptor agonist (GLP-1 RA) used and doses employed led to only small reductions in weight (around 2 kg) although in SELECT, which was a weight-loss trial, semaglutide up to 2.4 mg weekly led to an average reduction of 8.5 kg.9–13 In the two small, dedicated, HF trials, there was no change in left ventricular ejection fraction (LVEF) or N-terminal pro-B-type natriuretic peptide levels although neither was large enough (and possibly long enough) to allow a robust evaluation of either of these outcomes.9–13 Most of the trials shown in Table 2 did show an increase in heart rate (except for FIGHT), although the average increase was modest in the recent semaglutide trials, even with a higher dose.9–13 Characteristics of glucagon-like peptide-1 receptor antagonist trials including patients heart failure and reduced ejection fraction BMI, body mass index; CKD, chronic kidney disease; CVD, cardiovascular disease; EXCEL, Exenatide Study of Cardiovascular Event Lowering; FIGHT, Functional Impact of GLP-1 for Heart Failure Treatment; FLOW, Evaluate Renal Function with Semaglutide Once Weekly; HF, heart failure; HFpEF; heart failure with preserved ejection fraction; HFrEF; heart failure with reduced ejection fraction; LVEF, left ventricular ejection fraction; N/A = not applicable; N/R = not reported; NT-proBNP, N-terminal pro-B-type natriuretic peptide; NYHA, New York Heart Association; SELECT, Semaglutide Effects on Cardiovascular Outcomes in People with Overweight or Obesity; T2D, type 2 diabetes. aInvestigators were asked to report LVEF in 4 categories <25%, 25%–39%, 40%–55%, and >55%. bInvestigators were asked to report LVEF in 3 categories: <40%, 40%–49%, and ≥50%. After the first 1783 patients were enrolled, investigators were asked to report LVEF in 3 categories: <40%, 40%–49%, and ≥50%. cMeasured in 43 patients only. Characteristics of glucagon-like peptide-1 receptor antagonist trials including patients heart failure and reduced ejection fraction BMI, body mass index; CKD, chronic kidney disease; CVD, cardiovascular disease; EXCEL, Exenatide Study of Cardiovascular Event Lowering; FIGHT, Functional Impact of GLP-1 for Heart Failure Treatment; FLOW, Evaluate Renal Function with Semaglutide Once Weekly; HF, heart failure; HFpEF; heart failure with preserved ejection fraction; HFrEF; heart failure with reduced ejection fraction; LVEF, left ventricular ejection fraction; N/A = not applicable; N/R = not reported; NT-proBNP, N-terminal pro-B-type natriuretic peptide; NYHA, New York Heart Association; SELECT, Semaglutide Effects on Cardiovascular Outcomes in People with Overweight or Obesity; T2D, type 2 diabetes. aInvestigators were asked to report LVEF in 4 categories <25%, 25%–39%, 40%–55%, and >55%. bInvestigators were asked to report LVEF in 3 categories: <40%, 40%–49%, and ≥50%. After the first 1783 patients were enrolled, investigators were asked to report LVEF in 3 categories: <40%, 40%–49%, and ≥50%. cMeasured in 43 patients only. Main findings from glucagon-like peptide-1 receptor antagonist trials including patients with heart failure and reduced ejection fraction HF, heart failure; LVEF, left ventricular ejection fraction; N/R, not reported; NT-proBNP, N-terminal pro-B-type natriuretic peptide. aAll-cause death. bWorsening HF event. Main findings from glucagon-like peptide-1 receptor antagonist trials including patients with heart failure and reduced ejection fraction HF, heart failure; LVEF, left ventricular ejection fraction; N/R, not reported; NT-proBNP, N-terminal pro-B-type natriuretic peptide. aAll-cause death. bWorsening HF event. Turning to clinical outcomes, the recent subgroup analysis of SELECT accounts for the largest number of patients with HFrEF, greatly exceeding the combined patient exposure in the two dedicated HF trials with liraglutide (FIGHT and LIVE), which first raised potential concerns.9–13 SELECT enrolled 17 604 participants with a BMI ≥27 kg/m² and established cardiovascular disease. Overall, 4286 participants had an investigator-reported history of HF, of which 2273 (53%) were categorized as having HFpEF (LVEF ≥50%), 1347 (31%) with HFrEF (LVEF <50%), and 666 (15%) with ‘unclassified’ HF.12 Among patients with HFrEF, 76 (11%) in the placebo group and 46 (7%) in the semaglutide group died from cardiovascular causes (HR, 0.63, 0.43–0.91) over a median follow-up of ∼40 months.12 Curiously, only 35 (5%) patients in the placebo group had an episode of worsening HF vs 36 (6%) in the semaglutide group (HR 1.08, 0.68–1.72). In HFrEF trials, the number of patients with worsening HF is usually similar to or exceeds the number of cardiovascular deaths. This unusual pattern in SELECT was also observed in patients assigned to placebo in the subgroup analysis of EXSCEL (Table 2).9–13 The rates of both types of events in SELECT were also low compared with recent HFrEF trials e.g. worsening HF events 1.6 per 100 person-years in the SELECT HFrEF subgroup vs 10.1 per 100 person-years in DAPA-HF. The corresponding rates for cardiovascular death were 3.5 per 100 person-years in the SELECT HF subgroup vs 7.9 per 100 person-years in DAPA-HF.3 The event rates in the HFrEF subgroup of EXSCEL (defined by a lower LVEF of <40%) assigned to placebo were 4.7 per 100 person-years for HF hospitalization and 6.5 per 100 person-years for cardiovascular death.9 Overall, therefore, the patterns and rates of clinical events described in the largest available datasets do not appear to be typical of those observed in dedicated HFrEF trials, and the number of events in any individual trial is too small to draw any meaningful conclusion about the effect of GLP-1 RA on clinical outcomes in HFrEF. The history of HF drug development has several examples of misleading conclusions drawn from small trials with few events. Indeed, it seems implausible that, despite the wealth of data showing a reduction in cardiovascular mortality with GLP-1RA across a range of cardiovascular, kidney, and metabolic disorders, these drugs would specifically increase the risk of cardiovascular death in patients with HFrEF (and the data reviewed here do not suggest this, even the opposite).14 The clear benefit of GLP-1 RA therapy in patients with prior myocardial infarction, many of whom will have some degree of left ventricular systolic dysfunction, also argues against a safety concern related to cardiovascular death. In support of this, a recent observational study from the Swedish Heart Failure Registry (SwedeHF) also found a lower rate of cardiovascular death in HFrEF patients treated with a GLP-1 RA compared with propensity-matched participants not treated in this way.15 In addition, experimentally, semaglutide was recently shown to reduce the late sodium current and diastolic sarcoplasmic reticulum calcium leak, and improve systolic calcium transients and contractility, in cardiomyocytes and multicellular preparations from patients with end-stage HFrEF.16 Finally, patients with HFrEF should also benefit from the favourable effects of GLP-1 RA therapy on kidney function, shared by other recently successful treatments in HFrEF.1,3,13,15 The question of GLP-1 RA safety in HFrEF could also be tested in future genetic studies.17 As discussed, the data on worsening HF are particularly difficult to interpret. Moreover, given the stronger association between obesity and worsening HF than between obesity and mortality, it is important to note that among the trials examined, only SELECT specifically enrolled patients with obesity, and only in SELECT did GLP-1 RA treatment lead to substantial weight loss.1,3 In other words, if the benefit of anti-obesity drugs in HFrEF is mainly related to weight loss, this hypothesis has not been adequately tested in the trials available (Table 2). In summary, potential concerns about the safety of GLP-1 RA in patients with HFrEF are unconvincing and based on data that are incomplete, inconsistent, and even confusing. There is not one prospective randomized trial with a sufficient number of endpoints to draw any firm conclusions about the efficacy or safety of these compounds in HFrEF. Although subgroup analyses are available from other large trials, the patients enrolled are poorly characterized from a HF perspective, with no information about history of prior HF hospitalization, when LVEF was measured, or natriuretic peptide levels. Moreover, the hypothetical risk related to elevated heart rate has not been borne out in other vulnerable and related populations such as survivors of myocardial infarction. Thus, the question of whether GLP-1 RAs would reduce morbidity and mortality in patients with HFrEF remains unanswered, and equipoise surely exists. Since GLP-1 RA have been shown to reduce cardiovascular mortality across a wide range of cardiovascular, metabolic, and kidney conditions—many of which are common in HFrEF patients—a prospective, dedicated HFrEF trial with this class of agents, is both warranted and necessary. Similar considerations apply to dual (e.g. tirzepatide) and triple (e.g. retatrutide) hormone therapies, where the potential benefits and harms of the non-GLP-1 components in HFrEF require further investigation.7,18 J.