Effects of Sacubitril/Valsartan vs Valsartan in De Novo vs Acute on Chronic HFpEF and HFmrEF

Background Decompensated heart failure (HF) can be categorized as de novo or worsening of chronic HF. In PARAGLIDE-HF (Prospective comparison of ARNI with ARB Given following stabiLization In DEcompensated HFpEF), among patients with an ejection fraction >40% that stabilized after worsening HF, sacubitril/valsartan led to a significantly greater reduction in N-terminal pro-B-type natriuretic peptide (NT-proBNP) and was associated with clinical benefit compared to valsartan. Objectives This prespecified analysis characterized patients with de novo vs worsening chronic HF in PARAGLIDE-HF and assessed the interaction between HF chronicity and the effect of sacubitril/valsartan. Methods Patients were classified as de novo (first diagnosis of HF) or chronic (known HF prior to the index event). Time-averaged proportional change in NT-proBNP from baseline to weeks 4 and 8 was analyzed using an analysis of covariance model. A win ratio consisting of time to cardiovascular death, number and times of HF hospitalizations during follow-up, number and times of urgent HF visits during follow-up, and time-averaged proportional change in NT-proBNP was assessed for each group. Results Of the 466 participants, 153 (33%) had de novo HF and 313 (67%) had chronic HF. De novo patients had lower rates of atrial fibrillation/flutter and lower creatinine. There was a nonsignificant reduction in NT-proBNP with sacubitril/valsartan vs valsartan for de novo (0.82; 95% CI: 0.62-1.07) and chronic HF (0.88; 95% CI: 0.73-1.07), interaction P = 0.66. The win ratio was nominally in favor of sacubitril/valsartan for both de novo (1.12; 95% CI: 0.70-1.58) and chronic HF (1.24; 95% CI: 0.89-1.71). Conclusions There is no interaction between HF chronicity and the effect of sacubitril-valsartan.

A cute decompensated heart failure (HF) can be subclassified into 2 categories: worsening chronic HF and de novo HF. 1 The distinction between these 2 entities has important implications for prognosis, responses to medical therapy, and the risk of future decompensation. 2In a previous analysis of patients with acute decompensated HF and an ejection fraction (EF) <40%, patients with de novo HF were younger, had fewer comorbidities, and experienced a lower cumulative incidence of cardiovascular (CV) death or HF rehospitalization, but still gained a similar benefit from sacubitril/valsartan as compared to enalapril with regard to reduction in N-terminal pro-B-type natriuretic peptide (NT-proBNP) at weeks 4 and 8. 3 Sacubitril/valsartan is safe, well tolerated, and may lead to clinical benefit in patients with chronic stable HFpEF, particularly in patients with left ventricular ejection fraction (LVEF) on the lower end of the spectrum. 4The recently published PARAGLIDE-HF (Prospective comparison of ARNI with ARB Given following stabiLization In DEcompensated HFpEF) study extended these findings to patients with an EF >40% and a recent worsening HF event. 5The trial showed a greater time-averaged reduction in NT-proBNP with sacubitril/valsartan compared to valsartan, with a lower risk of worsening renal function in the sacubitril/valsartan arm but an increased incidence of symptomatic hypotension.
In the primary analysis of PARAGLIDE-HF, stratification by prior history of HF showed no significant interaction for the primary outcome of reduction in NT pro-BNP.However, a comparison of baseline demographics and an analysis of key secondary endpoints, including safety outcomes and clinical events, have not been performed.Furthermore, as the primary results suggested greater efficacy of sacubitril/ valsartan among patients with an EF below normal (#60%), subsequent analysis would benefit from highlighting this particular subgroup.Thus, given the importance of further characterizing patients with either de novo or worsening chronic HF, we undertook the current prespecified analysis to determine the relative effect of sacubitril/valsartan compared to valsartan in these important patient populations.

METHODS
TRIAL DESIGN.The trial design and rationale for PARAGLIDE-HF have previously been reported. 6In brief, PARAGLIDE-HF was a multicenter, randomized,  Study drug dose levels (ie, dose levels 1, 2, and 3, no treatment, and off treatment) were summarized by treatment arm in participants with de novo vs chronic HF.Data were presented as counts (percentages), and differences between treatment arms were assessed using the chi-squared test or Fisher's exact test, as appropriate.The full statistical analysis plan is available upon request.

RESULTS
Of the 466 trial participants, 153 (33%) had de novo HF, and 313 (67%) had chronic HF.Baseline demographics were overall similar between the groups, with de novo patients having lower rates of atrial fibrillation or flutter and lower serum creatinine (Table 1).
When comparing all patients regardless of treatment group, those with chronic HF showed a higher baseline NT-proBNP, while those with de novo HF saw a larger time-averaged proportional change in NT-proBNP at weeks 4 and 8 (Figure 1A).The rate of composite clinical events of CV death, HF hospitalizations, and urgent HF visits was lower among those with de novo compared to chronic HF (HR 0.34 [95% CI: 0.21-0.55],P < 0.001) (Figure 1B).
The time-averaged proportional change in NT-proBNP was similar for sacubitril/valsartan in both de novo and chronic HF (Central Illustration).There was no interaction by HF chronicity (P interaction ¼ 0.66) (Figure 2A).Among the group of patients with LVEF #60%, a similar effect was seen for both de novo and chronic HF (P interaction ¼ 0.56) (Figure 2B).
Secondary outcomes are presented in Table 2.The composite hierarchical outcome by win ratio favored sacubitril/valsartan among both de novo and chronic HF, although neither group met statistical significance.Among patients with EF #60%, the win ratio again showed a more pronounced effect of sacubitril/ valsartan in both groups.
There was no significant difference in the effects of sacubitril/valsartan in de novo or chronic HF on the secondary composite outcome of CV death, HF hospitalizations, or urgent HF visits.However, those with chronic HF and EF #60% saw a significantly lower rate of HF hospitalizations with sacubitril/valsartan (RR: 0.55; 95% CI 0.32-0.95;P ¼ 0.03).This Murray et al Sacubitril/Valsartan in De novo vs Chronic HFpEF/HFmrEF effect was not seen in de novo HF patients with EF #60% (RR: 1.22; 95% CI: 0.33-4.57;P ¼ 0.77).The interaction term was not significant for this effect (P interaction ¼ 0.28).
Time-to-first recurrent event (CV death, HF hospitalization, or urgent HF visit) for sacubitril/valsartan vs valsartan did not significantly differ between de novo and chronic HF (Figure 3A).Among the EF #60% subgroup, the hazard ratio favored sacubitril/valsartan in the chronic group (HR: 0.55; 95% CI: 0.34-0.89;P ¼ 0.02) but not the de novo group; however, the interaction term was not significant (P ¼ 0.56) (Figure 3B).
There were no significant differences in the effect of sacubitril/valsartan on the composite renal endpoint of renal death, End Stage Renal Disease, or >50% decline in eGFR.Rates of symptomatic hypotension were significantly higher with sacubitril/ valsartan in the chronic HF group, while the de novo group saw nominally higher rates of hypotension that did not meet statistical significance.For the endpoint of worsening renal function (increase in creatinine >0.5 and worsening of eGFR by >25%), the point estimate showed lower rates with sacubitril/valsartan in all groups, but only a significant reduction with sacubitril/valsartan in those with chronic HF and EF #60%.There were no differences in drug dose level achieved among either group (Supplemental Tables 1 and 2).

DISCUSSION
Among patients with EF >40% stabilized after a recent WHF event, both those with de novo and worsening chronic HF saw nominally larger reductions in NT-proBNP with sacubitril/valsartan compared to valsartan, without interaction by HF chronicity.These data suggest that sacubitril/valsartan may provide a similar effect in both de novo and worsening chronic HF patients in reducing natriuretic peptides.The effect of sacubitril/valsartan seen in the overall trial population 5 appeared similar among both de novo and chronic HF patients.Although the larger reduction of NT-proBNP with sacubitril/valsartan did not meet statistical significance in either de novo or chronic HF individual subgroups, this can likely be explained by the loss of statistical power when accounting for the smaller sample size.Importantly, there was no significant interaction by HF chronicity, indicating that sacubitril/valsartan's effect on reduction in natriuretic peptides did not appear to be modified by de novo vs chronic HF group.
Several previous studies have described differences in de novo and worsening chronic HF patients. 1,8,9De novo patients are typically younger with a lower burden of comorbidities than their counterparts with chronic HF.In this study, we also found a higher rate of prior atrial fibrillation or flutter and a higher serum creatinine in the chronic HF group, despite similar age, vital signs, and medication use at baseline.This higher burden of comorbidities likely played a role in the worse clinical outcomes among this group.
The risk of cardiovascular events has been consistently demonstrated in the literature to be higher among chronic patients compared to de novo patients. 10,11This was again shown in the data from PARAGLIDE-HF presented in this study, where chronic HF patients saw significantly higher rates of clinical events and higher NT-proBNP in follow-up compared to de novo patients.This lower event rate among de novo patients, combined with the underpowering of this study for clinical endpoints, may explain the lack of statistically significant clinical benefit with sacubitril/valsartan seen by the de novo group.Furthermore, when looking specifically at the subset of patients with chronic HF and an EF #60%, which likely represent the highest-risk subgroup, the win ratio significantly favored sacubitril/valsartan, and there was a significantly lower rate of HF hospitalizations and clinical events.The lack of significant statistical interaction for these endpoints by HF chronicity, however, shows that these differential effects may be driven by higher event rates in this group rather than a difference in efficacy of sacubitril/ valsartan among the de novo vs chronic phenotypes.
Safety endpoints including changes in renal function and rates of symptomatic hypotension were also important considerations in this trial.In the overall trial results, sacubitril/valsartan had higher rates of symptomatic hypotension compared to valsartan.
This effect was seen among chronic HF patients but not in de novo patients.There were also lower rates of worsening renal function with sacubitril/valsartan in the overall trial results, an effect that persisted in the chronic HF group with LVEF #60% in this analysis.These findings suggest that chronic HF patients are at high risk of medication side effects and should be monitored closely for adverse events; however, they still stand to benefit from the drug, especially with regard to preventing renal dysfunction.
This study is the first to look at the effect of sacubitril/valsartan among de novo vs chronic HF patients in patients with a mildly reduced or preserved EF (>40%).A previous analysis of patients from PIONEER-HF similarly showed that among patients with an EF #40%, sacubitril/valsartan was safe, well-tolerated, and led to a greater reduction in NT-proBNP compared to enalapril, regardless of prior history of HF. 3 This study extends those findings to patients across the range of EFs, again suggesting similar outcomes of sacubitril/valsartan regardless of chronicity of HF with respect to lowering natriuretic peptide levels.
BELOW-NORMAL EF.This study included an additional analysis of patients with a "below normal" (#60%) EF.The rationale for including this is because this appears to be the population for which sacubitril/ valsartan has the largest benefit with respect to reducing clinical events 12 and specific United States  sacubitril/valsartan for patients in this population. 13kewise, in the primary analysis of PARAGLIDE-HF, the win ratio for clinical events significantly favored sacubitril/valsartan only in the EF #60%

For the primary endpoint
of change in NT pro-BNP, the time-averaged proportional change from baseline on a natural logarithmic scale was analyzed using an analysis of covariance model using data from Weeks 4 and 8 with treatment arm, HF group (de novo vs chronic), in-hospital/out-of-hospital randomization status, gender, and baseline LVEF (# median, > median) as fixed effect factors, age and the logarithmic baseline NT-proBNP as covariates, and the interaction of treatment by HF group.The values from weeks 4 and 8 were averaged, and the change from baseline in log transformed NT-proBNP was calculated as follows: ln (average post dose value)ln (baseline A B B R E V I A T I O N S A N D A C R O N Y M S ACEI = angiotensin converting enzyme inhibitor ARB = angiotensin receptor blocker HFmrEF = heart failure with mid-range ejection fraction HFpEF = heart failure with preserved ejection fraction NT-proBNP = amino terminal pro-B type natriuretic peptide WHF = worsening heart failure value).The original study was powered only for the primary endpoint (change in NT pro-BNP) and not for secondary endpoints, which are considered exploratory in the present work.The secondary endpoint of the composite hierarchical outcome consisting of (in order of priority): 1) time to CV death; 2) number and times of HF hospitalizations during follow-up; 3) number and times of urgent HF visits during follow-up; and 4) timeaveraged proportional change in NT-proBNP (from baseline to weeks 4 and 8) was analyzed, estimating the unmatched win ratio by comparing every participant in the sacubitril/valsartan arm to every participant in the valsartan arm to determine a winner (unmatched pairing method).A component was only used as a tiebreaker in the pairwise comparison between 2 subjects if the comparison of components with a higher priority resulted in a tie.The estimated win ratio (the total number of wins in the sacubitril/ valsartan arm divided by the total number of wins in the valsartan arm) was calculated.A win ratio >1.0 was in favor of the sacubitril/valsartan arm.The analysis was performed separately within the de novo and chronic HF groups.The cumulative number of composite events (CV death, HF hospitalizations, urgent HF visits) was calculated.The time to these recurrent events was analyzed using the semiparametric proportional rate model. 7A rate ratio (RR) <1 indicates an effect in favor of sacubitril/valsartan arm.The RR was estimated from the aforementioned proportional rates model with the HF group, treatment arm, and inhospital/out-of-hospital randomization as fixed factors and the HF-by-treatment arm interaction.Additionally, the time-to-first composite event was analyzed using a Cox proportional hazards model with the HF group, treatment arm, the HF-bytreatment interaction, and in-hospital/out-ofhospital randomization as factors.Incidence of a composite renal endpoint of: renal death, reaching end-stage renal disease (sustained estimated glomerular filtration rate [eGFR] <15 ml/min/m 2 , chronic dialysis, or renal transplant), or $50% decline in eGFR relative to baseline was reported using a negative binomial regression model with the count data as the dependent variable.HF group, treatment arm, and in-hospital/out-of-hospital randomization were included as fixed-effect factors, as well as the HF-bytreatment interaction and log (follow-up duration) as the off-set.The incidence of adverse events of special interests: symptomatic hypotension, hyperkalemia (potassium > 5.5 mEq/L), and worsening renal function, defined as an increase in serum creatinine of $0.5 mg/dL and worsening of the eGFR by at least 25%, was analyzed using a logistic regression model with HF group, treatment arm, in-hospital/out-ofhospital randomization as fixed factors, as well as the HF-by-treatment arm interaction.

FIGURE 1
FIGURE 1 Differences in NT-proBNP and Composite Events for De Novo vs Chronic Heart Failure

Food
and Drug Administration labeling for use of CENTRAL ILLUSTRATION Effects of Sacubitril/Valsartan vs Valsartan in De Novo vs Acute on Chronic HFpEF and HFmrEF Murray EM, et al.JACC Adv.2024;3(6):100984.Time averaged proportional change in NT-proBNP as ratio of geometric means from baseline to weeks 4 and 8 for sacubitril/valsartan vs valsartan in de novo and chronic heart failure.P interaction ¼ 0.66.HF ¼ heart failure; HFpEF ¼ heart failure with preserved ejection fraction; HFmrEF ¼ heart failure with mid-range ejection fraction; LVEF ¼ left ventricular ejection fraction; NT-proBNP ¼ N-terminal pro-brain natriuretic peptide.

FIGURE 2
FIGURE 2 Change in NT-proBNP According to Treatment Arm by Baseline Heart Failure Group (De Novo vs. Chronic)

FIGURE 3
FIGURE 3 Time to First Composite Event According to Treatment Arm by Baseline Heart Failure Group (De Novo vs. Chronic)

TABLE 1
Baseline Characteristics of Randomized Participants According to Heart Failure Group

TABLE 2
adjusted rate per 100 subject-years.b Calculated by a negative binomial (NB) regression model, adjusted for baseline heart failure group, treatment arm, inhospital/out-of-hospital randomized, and the interaction of baseline heart failure group by treatment arm.Log (follow-up duration, study exposure) is the offset variable.Events that occurred in the randomized double-blind treatment period are included in the analysis.c Defined as potassium >5.5 mEq/L.d Increase in serum creatinine of >0.5 mg/dl AND worsening of eGFR by at least 25%.CV ¼ cardiovascular; eGFR ¼ estimated glomerular filtration rate; ESRD ¼ End Stage Renal Disease; HF ¼ heart failure; LVEF ¼ left ventricular ejection fraction; NT-proBNP ¼ N-terminal pro-brain natriuretic peptide.
a EAR: exposure