While trial evidence supports the benefit of angiotensin receptor-neprilysin inhibitor (ARNI) therapy in heart failure with reduced ejection fraction (HFrEF), its effectiveness in routine clinical practice is less explored. This study investigated the relative and absolute effectiveness of ARNI in patients with HFrEF. This nationwide Danish database study included patients with left ventricular ejection fraction (LVEF) ≤40%, 2018 to 2023. Using a prevalent new user design, 2,446 ARNI initiators were matched 1:2 to 4,892 users of angiotensin-converting enzyme inhibitors (ACE-I) or angiotensin receptor blockers (ARB) based on propensity scores, age, LVEF, and NT-proBNP. The primary outcome was all-cause mortality; secondary outcomes were cardiovascular mortality and hospitalization. There were 279 deaths among ARNI initiators (5.6/100 person-years) and 533 among ACE-I/ARB users (6.7/100 person-years), yielding a hazard ratio (HR) of 0.85 (95% CI, 0.74 to 0.98) for all-cause mortality. A significant interaction was observed for recent hospitalization (p = 0.04), with ARNI yielding a lower HR in this group. HRs were otherwise consistent across age, sex, LVEF, NT-proBNP, NYHA class, ischemic heart disease, chronic kidney disease, and type 2 diabetes. The largest absolute mortality reductions were seen in subgroups with recent hospitalization, NYHA class III to IV, and severely elevated NT-proBNP. ARNI was also associated with a lower risk of cardiovascular death (HR, 0.81; 95% CI, 0.65 to 0.99), but not with other secondary outcomes. In this study, ARNI was associated with a 15% reduction in all-cause mortality vs ACE-I/ARB. Patients with advanced or symptomatic heart failure appeared to experience the greatest absolute benefit.
Heart failure is a common and serious condition, with its prevalence increasing due to an aging population. Heart failure with reduced ejection fraction (HFrEF) is associated with persistently high mortality rates, evidenced by a 5-year survival rate of only 25% following hospitalization. In recent years, heart failure treatment has seen rapid development, including the introduction of angiotensin receptor-neprilysin inhibitor (ARNI) class, marking the first major addition to treatment of HFrEF beyond the traditional cornerstone therapy with angiotensin converting enzyme-inhibitors (ACE-I) or angiotensin receptor blockers (ARB), beta-blockers, and mineralocorticoid receptor antagonists (MRA).
In the landmark randomized controlled PARADIGM-HF trial, treatment with ARNI demonstrated a significant 20% reduction in the risk of cardiovascular death or hospitalization for heart failure compared with traditional renin angiotensin system blockade with enalapril and a 16% reduction in the secondary outcome all-cause mortality, prompting its rapid integration into standard HFrEF therapy. Recent major guidelines recommend ARNI either as a potential part of first-line therapy or as a replacement for ACE-I/ARB in patients who remain symptomatic despite optimal treatment with ACE-I/ARB, a beta-blocker, MRA, and a sodium-glucose cotransporter-2 inhibitor. ,
While solid clinical trial evidence supports the efficacy of ARNI in HFrEF (i.e., benefit under controlled conditions), its effectiveness in the broader heart failure populations commonly encountered in routine clinical practice remains less well explored. The PARADIGM-HF trial has been noted for its stringent eligibility criteria, including a run-in period and the requirement for patients to achieve target doses of ARNI, which may limit its generalizability to the broad heart failure populations encountered in routine clinical settings. ,, Additional evidence is needed to assess the effectiveness of ARNI in the diverse populations of routine heart failure care and across subgroups at varying baseline risk, both in terms of relative and absolute benefit. Whereas a small number of previous database studies have investigated the impact of ARNI in HFrEF, their contribution to the overall body of evidence has been limited due to lack of clinical data on heart failure patients, short follow-up periods, selective inclusion, or limitations in design. ,,,
Using a nationwide clinical registry in Denmark, this database study was conducted to investigate the relative and absolute effectiveness of ARNI, compared with ACE-I/ARB, in patients with HFrEF.
Methods
Study design
This was a noninterventional database study including patients with HFrEF who initiated treatment with ARNI (sacubitril-valsartan, the only ARNI available during the study period) or were treated with ACE-I/ARB from January 1, 2018, to June 30, 2023. While ARNI was first marketed in Denmark in 2016, the study period was chosen to start in 2018 due to limited early use and because very early users of a newly marketed drug may be selected and differential, compared with later, more representative users. Patients with heart failure were identified using the Danish Heart Failure Registry (DHR), which holds records of patients with first-time specialist care primary diagnosis of heart failure (ICD-10 codes: I11.0, I13.0, I13.2, I42.0, I42.6, I42.7, I42.9, I50.0, I50.1, I50.9). Reporting to the registry is mandatory for all hospital departments caring for patients with incident heart failure in Denmark. The study population included patients with left ventricular ejection fraction (LVEF) ≤40%, aged ≥45.
For inclusion, patients were, first, required to have a recorded N-terminal prohormone of brain natriuretic peptide (NT-proBNP) measurement within the previous 180 days; given the strong prognostic value of NT-proBNP on mortality in heart failure, this criterion was included to ensure reliable and valid identification of patients at high baseline risk at the time of the index date. Second, patients were required to have had a specialist care contact for heart failure within 30 days prior to the index date to ensure that all patients had a recent opportunity to be prescribed an ARNI for treatment initiation.
Exclusion criteria included having a prescription for ARNI at any time before the index date (to preserve the integrity of the analytic contrast and avoid carryover effects), heart transplant, end-stage renal disease or kidney transplant, end-stage illness, drug misuse, and residence in a nursing home at the time of the index date. All eligibility criteria are shown in Supplementary Table 1 .
The primary outcome was all-cause mortality; the secondary outcomes were the composite of cardiovascular mortality or hospitalization due to cardiovascular causes and its individual components. The outcome of cardiovascular hospitalization was selected over heart failure hospitalization, which was used in the PARADIGM-HF trial, given the different nature of data; specifically, the strictly adjudicated events of heart failure hospitalization in controlled trials may not readily align with the coding practices encountered in routine clinical settings, , making such a narrow definition less pertinent for the analyses of secondary data.
Employing a prevalent new user (PNU) design, the ARNI treatment group comprised patients who initiated ARNI following heart failure diagnosis. The index date (the date of start of follow-up) for the ARNI group was defined as the date of treatment initiation. The comparator group was defined based on all prescriptions for ACE-I/ARB indicated for HFrEF (enalapril, ramipril, trandolapril, candesartan, valsartan, and losartan) following heart failure diagnosis; the date of each ACE-I/ARB prescription was considered a potential comparator index date. The design and choice of comparator was intended to compare the benefit of initiating ARNI versus initiating or continuing standard-of-care ACE-I/ARB treatment in HFrEF.
To form the study population, initiators of ARNI were matched to the comparator group in a 1:2 ratio based on the number of previous ACE-I/ARB prescriptions, age, LVEF, levels of NT-proBNP, and the propensity score. Details of the matching procedure and selection of comparator index dates are provided in the eMethods. The choice of a prevalent new user design allowed the inclusion of all initiators of ARNI, including both patients who initiated treatment with ACE-I/ARB and then switched to ARNI and those who initiated ARNI treatment as their first renin angiotensin system-acting drug after first-time diagnosis of HF. Each patient could contribute twice to the final study population: once to the ARNI group and once to the comparator group.
Data sources
Information on use of ARNI and ACE-I/ARB was obtained from the National Prescription Register, which records all prescriptions filled at all pharmacies in Denmark, including the date of dispensing and Anatomic Therapeutic Chemical (ATC) code. The ATC codes for the study drugs are shown in Supplementary Table 2 .
The primary outcome was based on information on date of death from the Civil Registration System (CRS), which holds information on vital status for all Danish residents, with virtually complete coverage. Information on cardiovascular hospitalization and cardiovascular death was obtained from primary diagnoses at hospital admission in the National Patient Register (NPR) and recorded causes of death in the Danish Register of Causes of Death (CDR), respectively. , The NPR records all hospital contacts, including information on date of contact and discharge codes, classified according to International Classification of Diseases, 10th revision (ICD-10). The CDR is based on death certificates, holding information on the underlying cause of death for all deaths in Denmark, similarly classified according to ICD-10. For the composite outcomes, the event date was defined by the first occurrence of an outcome component. Definitions of all outcomes are provided in Supplementary Table 3 .
Information on covariates was obtained from the DHR (LVEF and NYHA classification), the CRS (demographics), Statistics Denmark (education), the NPR (medical history, procedures, and healthcare utilization), the National Prescription Register (prescription drug use), and the Danish Register of Laboratory Results for Research (RLRR) (estimated glomerular filtration rate [eGFR] and NT-proBNP). Covariate definitions are provided in Supplementary Table 4 . Missing covariate information on education, NYHA-classification, baseline eGFR, mean eGFR in the previous 365 days, alcohol consumption, and smoking status (the only covariates with missing values) were handled using a missing value category.
Data analysis
Cox proportional hazards regression was used to estimate the risk of the primary and secondary outcomes associated with ARNI, compared with ACE-I/ARB. For the main analyses, treatment status was defined according to the treatment status on the index date, analogous to an intention-to-treat design in a clinical trial. Patients were censored on the date of death, emigration, or end of the study period (June 30, 2023), whichever occurred first. The patients in the comparator group were also censored on the date of a prescription of ARNI. The underlying time scale was time since the start of treatment. The absolute rate differences between ARNI and ACE-I/ARB for the primary outcome were estimated using Poisson regression, using the logarithm of time of follow-up as the offset.
Subgroup analyses of the primary outcome within the overall matched study population were conducted according to age, sex, LVEF, levels of NT-proBNP, NYHA-class, hospitalization due to heart failure in the previous 30 days, history of ischemic heart disease, chronic kidney disease, and type 2 diabetes. The subgroup analyses of rate differences were focused on clinical characteristics and severity of heart failure (NT-proBNP, NYHA-class, hospitalization due to heart failure, history of ischemic heart disease, chronic kidney disease, and type 2 diabetes). Definitions of all subgroups are provided in Supplementary Table 5 . Differences between subgroups were assessed by use of the Wald test for homogeneity, with p-values <0.05 considered significant. preplanned sensitivity analyses for the primary outcome included, first, the use of a ‘per-protocol’ analysis, in which each prescription counted as 180 days of treatment: patients were considered to be on treatment for as long as a new prescription was filled within the 180-day window of the preceding prescription. Patients who did not fill a new prescription within 180 days were censored. Second, by design, in the main analysis patients could crossover only from ACE-I/ARB to ARNI and not vice versa: to assess the potential impact of differential or informative censoring due to this design, inverse probability of censoring weighting (IPCW) were applied to balance the population according to treatment and censoring selection factors that could affect the outcome, including all variables in Table 1 .
Table 1
Baseline characteristics after matching
| Characteristic | ARNI | ACE-I/ARB | SMD |
|---|---|---|---|
| N | 2,446 | 4,892 | |
| Male sex | 1,834 (75) | 3,628 (74) | 0.02 |
| Age, Mean (SD) | 68.9 (9.7) | 69.0 (10.0) | |
| Heart failure characteristics | |||
| LVEF, %, Mean (SD) | 26.4 (8.4) | 26.6 (8.5) | |
| LVEF, %, categorized | |||
| 35-40 | 627 (26) | 1,307 (27) | 0.02 |
| 30-34 | 479 (20) | 905 (18) | 0.03 |
| 25-29 | 417 (17) | 869 (18) | 0.02 |
| <25 | 923 (38) | 1,811 (37) | 0.01 |
| NYHA-classification | |||
| I | 300 (12) | 629 (13) | 0.02 |
| II | 1,647 (67) | 3,306 (68) | 0.01 |
| III | 453 (19) | 858 (18) | 0.03 |
| IV | 16 (1) | 38 (1) | 0.01 |
| Missing | 30 (1) | 61 (1) | ≤0.01 |
| Hospitalized at the time of first HF-diagnosis | 1,094 (45) | 2,260 (46) | 0.03 |
| Baseline NT-ProBNP, pg/mL, Median (IQR) | 1398.0 (713.0-2869.0) | 1381.0 (703.5-2862.0) | |
| Days since HF diagnosis, Median (IQR) | 232.0 (122.0-835.0) | 222.0 (120.0-636.5) | |
| Baseline eGFR (mL/min/1.73m²), Mean (SD) | 68.2 (18.6) | 68.2 (19.1) | |
| Medical history | |||
| Acute coronary syndrome | 621 (25) | 1,185 (24) | 0.03 |
| Ischemic heart disease | 1,029 (42) | 2,019 (41) | 0.02 |
| Stroke | 224 (9) | 424 (9) | 0.02 |
| Atrial fibrillation | 858 (35) | 1,747 (36) | 0.01 |
| Health care utilization | |||
| HF hospitalization in the previous 30 d | 334 (14) | 687 (14) | 0.01 |
| HF hospitalization in the previous 365 d | 1,333 (54) | 2,635 (54) | 0.01 |
| Prescription drug use | |||
| Loop diuretic | 1,825 (75) | 3,655 (75) | ≤0.01 |
| Other diuretic | 2,031 (83) | 4,012 (82) | 0.03 |
| ACEi | 1,767 (72) | 3,578 (73) | 0.02 |
| ARB | 856 (35) | 1,640 (34) | 0.03 |
| Beta-blocker | 2,346 (96) | 4,707 (96) | 0.02 |
| MRA | 1,996 (82) | 3,932 (80) | 0.03 |
| SGLT2 inhibitor | 1,128 (46) | 2,047 (42) | 0.09 |
| Metformin | 515 (21) | 909 (19) | 0.06 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
