Hospitalization Rates in Patients with Heart Failure and Reduced Ejection Fraction Initiating Sacubitril/Valsartan or Angiotensin-Converting Enzyme Inhibitors/Angiotensin Receptor Blockers: A Retrospective Cohort Study

Heart failure (HF) affects approximately 6 million adults in the United States (US) [1], and its prevalence is projected to increase by 15% globally and by 23% in the US by 2030 [2]. The renin–angiotensin–aldosterone system (RAAS) inhibition is the cornerstone of therapy for HF with reduced ejection fraction (HFrEF; left ventricular ejection fraction [LVEF] ≤ 40%), which has traditionally included angiotensin-converting enzyme inhibitors (ACEi) and angiotensin receptor blockers (ARB) [3, 4]. Sacubitril/valsartan (SAC/VAL) is a first-in-class angiotensin receptor neprilysin inhibitor (ARNI), which simultaneously delivers RAAS inhibition through valsartan and neprilysin inhibition through sacubitril. SAC/VAL was first approved by the US Food and Drug Administration in July 2015 for the treatment of chronic HFrEF [5]. The latest update of the American College of Cardiology (ACC) Expert Consensus Decision Pathway (ECDP) on HF treatment recommends ARNI as the preferred RAAS inhibitor (RAASi) for HFrEF in patients who have been previously treated with an ACEi or ARB, as well as in patients naïve to an ACEi or ARB [6]. SAC/VAL has shown superiority over enalapril in reducing all-cause mortality and HF hospitalization in patients with HFrEF and New York Heart Association (NYHA) class II-IV who had been previously treated with an ACEi/ARB in the PARADIGM-HF trial [7], and in patients hospitalized with acute decompensated HF in the PIONEER-HF trial [8‐10]. Moreover, there is increasing evidence of the effectiveness and safety of SAC/VAL in real-world clinical practice [11‐17]. Subgroup analyses of clinical trials, including PIONEER-HF [10] and PROVE-HF [18], have shown a consistent benefit of SAC/VAL in ACEi/ARB-naïve patients as in those with prior exposure. This is complemented by data showing consistent safety and tolerability in ACEi/ARB-naïve patients from PIONEER-HF [10] and TRANSITION-HF [19]. However, there are limited data on this patient population in the real-world setting. Because patients with HF who are naïve to RAAS inhibition may differ from those with prior exposure in terms of baseline characteristics and tolerability to new treatments, it is of interest to evaluate the real-world effectiveness of SAC/VAL specifically in patients naïve to RAAS inhibition. Therefore, we aimed to compare the hospitalization rates and time to hospitalization between RAASi-naïve patients with HFrEF initiating SAC/VAL and those initiating traditional first-line ACEi/ARB therapy, using data from the US Optum^® Electronic Health Records (EHR) database.

The new SAC/VAL users in this study had significantly lower rates of all-cause hospitalizations and the composite of HF hospitalization or ER visits than the new ACEi/ARB users; this was true in the analyses of both total rates and time to first event. These results were similar to those observed in a systematic literature review of real-world studies, wherein studies comparing SAC/VAL with ACEi/ARB in patients with HFrEF (previously treated or naïve to ACEi/ARB) reported a significantly lower risk of all-cause hospitalization with SAC/VAL [20].

Despite significant differences in the rate of all-cause hospitalization, the rates of HF and cardiovascular hospitalizations were similar between new SAC/VAL users and new ACEi/ARB users—an initially unexpected and clinically counterintuitive finding. However, another US-based real-world study by Tan et al. [17] involving patients with systolic HF observed a similar pattern—a significantly lower risk for all-cause hospitalization in the SAC/VAL group than in the ACEi/ARB group (P < 0.001), but not for HF hospitalization (P = 0.26). Tan et al. considered that differences in coding practices could have influenced the analysis of HF-specific hospitalizations. As was done in the Tan et al. study, we also defined HF hospitalization per ICD-9-CM and ICD-10-CM codes for HF listed within the primary diagnoses associated with a hospitalization. Although the rate of HF hospitalizations was similar between treatment groups, a composite endpoint of HF hospitalization or ER visits did show a significant difference between SAC/VAL and ACEi/ARB, indicating that ACEi/ARB-treated patients are more likely to present to the ER for their HF than SAC/VAL-treated patients. Analyzing emergency visits that do not result in a hospitalization in addition to hospital admissions, may provide a complete picture of the burden of HF-related events to the healthcare system rather than focusing on hospitalizations alone. Moreover, all-cause hospitalization is considered more important by payers and patients than cause-specific hospitalization, as it represents the overall disease burden [21].

Frequent hospitalizations pose a major burden on patients with HF and their caregivers. In addition, consistent evidence indicates that the economic burden of HF is dominated by the hospitalization costs [11, 16, 22, 23]. A US-based analysis found that inpatient costs accounted for 80% of the total lifetime costs in HF [24]. Therefore, we expect that the reduction in all-cause hospitalizations observed among the new SAC/VAL users in our study will be meaningful in reducing the overall burden of hospitalizations for patients. Reducing the risk of total hospitalizations due to any cause is of particular importance in the context of the current coronavirus disease 2019 (COVID-19) pandemic, when it is imperative to keep patients out of hospital.

To the best of our knowledge, the present study is the first real-world study to specifically compare SAC/VAL and ACEi/ARB use in patients with HFrEF naïve to prior RAASi therapy. Previous subgroup analyses have shown that the benefit of SAC/VAL is similar in patients naïve to ACEi/ARB and those pretreated with ACEi/ARB [9, 17]. A subgroup analysis of a US administrative claims database study showed a similar reduction in all-cause hospitalization with SAC/VAL in patients who were previously treated with ACEi/ARB and those naïve to ACEi/ARB [17]. In a sub analysis of the PIONEER-HF trial, the superiority of SAC/VAL over enalapril was consistent in patients with or without prior exposure to ACEi/ARB [9, 10]. The time to onset of HF hospitalization for the PIONEER-HF trial and the present real-world study were different (PIONEER-HF: HR 0.61, 95% CI 0.40, 0.93; present study: HR 0.92, 95% CI 0.80, 0.91). However, it is important to note that time frame investigated in the present study is much longer than the 8-week time frame in PIONEER-HF trial. Due to the nature of the data and the varying index dates in the present study, we defined the follow-up period as until end of the study period (March 31, 2020), death or patient transfer out of the database. Therefore, in the present study, patients could potentially be assessed for time to HF hospitalization over a period greater than 3 years.

Our findings of the real-world benefit of SAC/VAL on all-cause hospitalizations in patients naïve to prior RAASi therapy are consistent with the results from these previous subgroup analyses and support the use of SAC/VAL in this patient population. Moreover, the recent update of the ACC ECDP for optimization of HF treatment has now recommended a “direct-to-ARNI approach,” that is to use ARNI as a de novo RAASi therapy in patients naïve to ACEi/ARB therapy [6].

The findings of the present study should be interpreted in light of several limitations. Because of the retrospective, observational nature of the study, the causal relationships between the clinical outcomes and the treatment could not be inferred. Medication use might have been overestimated because the drug prescription was used as a proxy for drug use. Therefore, it was not possible to determine whether the patients used their medication as prescribed. The use of SAC/VAL or ACEi/ARB may be influenced by the socioeconomic status of a patient. However, this parameter could not be adjusted as details on socioeconomic status were not available in Optum EHRs. Despite matching several key patient characteristics between the SAC/VAL and ACEi/ARB incident cohorts, a possibility of imbalances in the disease severity might exist, with a chance that patients with more persistent symptoms may have been prescribed SAC/VAL. Possibility of differences in the disease severity might be due to the non-availability of some of the variables, including NYHA class, blood pressure, and biomarkers (e.g., N-terminal pro-B-type natriuretic peptide and renal function markers). Another limitation was that the “naïve” populations in the study were defined as “being naïve to ACEi/ARB and SAC/VAL for at least 365 days before the initial prescription for either ACEi/ARB or SAC/VAL”. Although it was unlikely that patients were previously treated with SAC/VAL because of the overlapping of the identification period of our study with the date of licensing of SAC/VAL in the US, it could have been possible that an individual might not be “truly” naïve to ACEi/ARB and had merely not been prescribed ACEi/ARB during the 365 days before their index prescription within the database. Lastly, the results of this study may not be generalized to other populations, such as the new SAC/VAL and new ACEi/ARB users who were excluded from the analysis during propensity score matching, those with different ethnicity or from different regions, and those who were uninsured or on fee-for-service healthcare plans.

Use of SAC/VAL in RAASi-naïve patients with HFrEF resulted in lower rates of all-cause hospitalization and the composite of HF hospitalization or ER visits compared with ACEi/ARB treatment. The rates of HF and cardiovascular hospitalizations were similar between the two cohorts. These findings further strengthen the evidence base for SAC/VAL in ACEi/ARB-naïve patients. Initiating SAC/VAL directly in RAASi-naïve patients with HFrEF can reduce total hospitalizations, thereby reducing the clinical and economic burden of HFrEF in these patients.