Special DELIVERy: Reducing High-Complexity Hospitalizations in Heart Failure

This article refers to ‘Effects of dapagliflozin on heart failure hospitalizations according to severity of inpatient course: Insights from DELIVER and DAPA-HF’ by S. Chatur et al., published in this issue on pages 1364–1371. Heart failure (HF) is a global public health epidemic affecting more than 64 million people worldwide and is the leading cause of hospitalization in the United States and Europe.1 Hospital admission for HF is considered a sentinel event in the progression of the disease and frequently results in readmissions, decrements in quality of life, and the need for higher intensity medical care. Attempts to decrease HF admissions and its associated social and economic burden have become a public health priority. While the traditional focus of therapeutics has been to reduce HF hospitalizations, recent insights and methodological advances have broadened our perspectives on endpoint selection in HF trials. First was the recognition that therapies, strategies, and other interventions may not only modify the first HF hospitalization during trial follow-up, but may also modify subsequent HF hospitalizations.2 In addition, recent insights have identified worsening HF episodes occurring outside of traditional hospitalization as carrying important prognostic information, and subsequently, these events have been included in composite endpoint assessments in some recent HF clinical trials.3, 4 The sodium–glucose cotransporter 2 inhibitors (SGLT2i) have now been shown to reduce first and recurrent HF hospitalizations in addition to other worsening HF events across the left ventricular ejection fraction (LVEF) spectrum.5-8 In this issue of the Journal, Chatur et al.9 extend our understanding of how this therapy might differentially modify yet another unexplored aspect: HF hospitalizations of varying complexity and length of stay. The authors used a participant-level pooled data of two large, international randomized controlled trials, DAPA-HF and DELIVER, both of which evaluated the effect of dapagliflozin 10 mg once daily relative to placebo in patients with symptomatic HF, elevated natriuretic peptides and LVEF ≤40% and >40%, respectively. These trials found that dapagliflozin reduced the risk of death from cardiovascular causes and HF events across the spectrum of LVEF.5, 8, 10 In this post hoc analysis, the authors examined the effects of dapagliflozin on adjudicated HF hospitalizations of varying complexity and length of stay. Complicated hospitalizations were defined as those requiring intensive care unit (ICU) stay, intravenous vasoactive therapies, invasive/non-invasive ventilation, mechanical fluid removal, and/or mechanical circulatory support. Hospitalizations not requiring these resources were classified as uncomplicated. In DAPA-HF, 43% of the total 799 hospitalization were categorized as complicated. In DELIVER, 29% of the total 1209 HF hospitalizations were complicated. Unsurprisingly, patients experiencing complicated HF hospitalizations were at significantly higher risk for in-hospital mortality compared with those with uncomplicated hospitalizations in both DELIVER (16.7% vs. 2.3%) and DAPA-HF (15.1% vs. 3.8%). Importantly, the authors found that dapagliflozin similarly reduced the risk of total uncomplicated and complicated hospitalizations across both trial populations. The beneficial effects of dapagliflozin were also consistent in those with shorter (<5 days) or longer (>5 days) hospital lengths of stay. The authors should be commended on performing a comprehensive and novel analysis which demonstrates the consistent beneficial effects of dapagliflozin in reducing complex hospitalizations requiring intensive management beyond standard intravenous diuretics and those with prolonged lengths of hospital stay. The SGLT2i story to date has been one of remarkable consistency; this therapy has improved clinical outcomes irrespective of LVEF, age, sex, and care setting. We now understand that this therapy is also not discerning in the types of HF hospitalization reduced. While Chatur and colleagues extend our knowledge on the effects of this particular therapy, they also explore a new paradigm of evaluating HF hospitalization complexity which may have broader implications for the design and conduct of future HF trials. Current constructs frame HF hospitalizations, regardless of severity, as a homogeneous endpoint. In reality, however, clinicians understand that tremendous variability exists in clinical presentations, resource utilization, and in-hospital complexity of hospitalization.11 In this analysis, the authors propose complicated hospitalization as those requiring intensive therapy including ICU stay, intravenous vasoactive therapies (i.e. vasopressor, inotrope, and/or vasodilator), invasive or non-invasive ventilation, mechanical fluid removal, ultrafiltration, and/or mechanical circulatory support. However, no standardized or validated definition exists. Importantly, the inclusion of ICU requirement may also vary greatly based on inpatient facility, care location, the availability of intensive care services and other factors. Establishing consensus in defining complex HF hospitalization may be important for the HF community. Another adjacent and similarly underexplored question is the extent to which HF contributes to hospitalization. There is a growing recognition that a binary assessment of a hospitalization as HF or non-HF may be inadequate or overly simplistic; probabilistic adjudication approaches may help, particularly among trials in mildly reduced and preserved ejection fractions, where non-HF cardiometabolic comorbid burden is expected to be larger.12 While the overall effects of dapagliflozin versus placebo were consistent across HF complexity in these trials in interaction testing, the point estimates did favour greater numerical reductions in those with more uncomplicated hospitalizations, potentially suggesting these may be more ‘modifiable’ by the therapy. Standardization of definitions on HF complexity and probabilistic approaches identifying the relative contributions of HF to a particular hospitalization may allow for a more nuanced understanding of modifiable and non-modifiable treatment effects in future HF trial programmes. The authors found a greater prevalence of complicated hospitalizations in patients with reduced LVEF as compared to those with mildly reduced LVEF. Despite known epidemiology of HF with a mildly reduced ejection fraction (HFmrEF)/HF with a preserved ejection fraction (HFpEF) as including patients who are older and with greater comorbid burden, cardiovascular events rates in those with HFmrEF/HFpEF have generally been observed to be lower than those in HFrEF.13, 14 When hospitalized, patients with HFpEF on aggregate, have lower in-hospital mortality compared to patients with HFrEF, however, in both cases, post-discharge survival remains soberingly poor.14 While these findings are limited by trial inclusion criteria and variation may exist in usual care settings, these data do reaffirm LVEF as a powerful predictive marker of clinical cardiovascular severity and the resultant need for intensive HF support during hospitalization.13 Overall, the DAPA-HF and DELIVER programmes have effectively demonstrated that compared to placebo, dapagliflozin consistently and robustly reduces the risk of death from cardiovascular causes and HF hospitalization across the full range of LVEF.10 As a class, SGLT2i have shown to improve outcomes across LVEF and care settings, including ambulatory, recently hospitalized and acutely hospitalized patients.4, 10, 15 Chatur et al. now demonstrate the consistent benefit of this class on reducing HF hospitalization, regardless of complexity. Patients with end-stage or refractory HF are complex, underserved group most likely to experience complicated HF hospitalization and prolonged hospital courses. Given the demonstrated safety and efficacy of SGLT2i and the insights of reducing both uncomplicated and complicated hospitalizations as evidenced by this paper, whether this class of therapies could be prospectively evaluated in patients with the most severe forms of HF is an intriguing question for further study. In conclusion, Chatur et al. capture a deeper spectrum of hospitalization of HF in the context of pooled data from large HF clinical trials. The authors also establish safety and efficacy of dapagliflozin in yet another care dimension. In doing so, they also make us think broadly about a more detailed capture of HF events in future trials and the insights that future might hold. Conflict of interest: none declared.