Effect of Sacubitril/Valsartan vs Standard Medical Therapies on Plasma NT-proBNP Concentration and Submaximal Exercise Capacity in Patients With Heart Failure and Preserved Ejection Fraction: The PARALLAX Randomized Clinical Trial

Key Points

Question  What is the effect of sacubitril/valsartan compared with standard medical therapy on plasma N-terminal pro–brain natriuretic peptide (NT-proBNP) concentration and submaximal exercise capacity in patients with heart failure and preserved ejection fraction?

Findings  This randomized clinical trial included 2572 participants with heart failure and left ventricular ejection fraction higher than 40%. Treatment with sacubitril/valsartan vs standard treatment with enalapril, valsartan, or placebo resulted in a statistically significant reduction in plasma NT-proBNP levels at 12 weeks (adjusted geometric mean ratio, 0.84) but did not significantly change the 6-minute walk distance at 24 weeks (9.7 m vs 12.2 m).

Meaning  Compared with standard medical therapies, sacubitril/valsartan resulted in a significantly greater decrease in plasma N-terminal pro–brain natriuretic peptide at 12 weeks but did not improve submaximal exercise capacity at 24 weeks.

Importance  There is limited evidence on the benefits of sacubitril/valsartan vs broader renin angiotensin system inhibitor background therapy on surrogate outcome markers, 6-minute walk distance, and quality of life in patients with heart failure and mildly reduced or preserved left ventricular ejection fraction (LVEF >40%).

Objective  To evaluate the effect of sacubitril/valsartan on N-terminal pro–brain natriuretic peptide (NT-proBNP) levels, 6-minute walk distance, and quality of life vs background medication–based individualized comparators in patients with chronic heart failure and LVEF of more than 40%.

Design, Setting, and Participants  A 24-week, randomized, double-blind, parallel group clinical trial (August 2017-October 2019). Of 4632 patients screened at 396 centers in 32 countries, 2572 patients with heart failure, LVEF of more than 40%, elevated NT-proBNP levels, structural heart disease, and reduced quality of life were enrolled (last follow-up, October 28, 2019).

Interventions  Patients were randomized 1:1 either to sacubitril/valsartan (n = 1286) or to background medication–based individualized comparator (n = 1286), ie, enalapril, valsartan, or placebo stratified by prior use of a renin angiotensin system inhibitor.

Main Outcomes and Measures  Primary end points were change from baseline in plasma NT-proBNP level at week 12 and in the 6-minute walk distance at week 24. Secondary end points were change from baseline in quality of life measures and New York Heart Association (NYHA) class at 24 weeks.

Results  Among 2572 randomized patients (mean age, 72.6 years [SD, 8.5 years]; 1301 women [50.7%]), 2240 (87.1%) completed the trial. At baseline, the median NT-proBNP levels were 786 pg/mL in the sacubitril/valsartan group and 760 pg/mL in the comparator group. After 12 weeks, patients in the sacubitril/valsartan group (adjusted geometric mean ratio to baseline, 0.82 pg/mL) had a significantly greater reduction in NT-proBNP levels than did those in the comparator group (adjusted geometric mean ratio to baseline, 0.98 pg/mL) with an adjusted geometric mean ratio of 0.84 (95% CI, 0.80 to 0.88; P < .001). At week 24, there was no significant between-group difference in median change from baseline in the 6-minute walk distance with an increase of 9.7 m vs 12.2 m (adjusted mean difference, −2.5 m; 95% CI, −8.5 to 3.5; P = .42). There was no significant between-group difference in the mean change in the Kansas City Cardiomyopathy Questionnaire clinical summary score (12.3 vs 11.8; mean difference, 0.52; 95% CI, −0.93 to 1.97) or improvement in NYHA class (23.6% vs 24.0% of patients; adjusted odds ratio, 0.98; 95% CI, 0.81 to 1.18). The most frequent adverse events in the sacubitril/valsartan group vs the comparator group were hypotension (14.1% vs 5.5%), albuminuria (12.3% vs 7.6%), and hyperkalemia (11.6% vs 10.9%).

Conclusions and Relevance  Among patients with heart failure and left ventricular ejection factor of higher than 40%, sacubitril/valsartan treatment compared with standard renin angiotensin system inhibitor treatment or placebo resulted in a significantly greater decrease in plasma N-terminal pro–brain natriuretic peptide levels at 12 weeks but did not significantly improve 6-minute walk distance at 24 weeks. Further research is warranted to evaluate potential clinical benefits of sacubitril/valsartan in these patients.

Trial Registration  ClinicalTrials.gov Identifier: NCT03066804

More than half of all patients with heart failure have preserved or mildly reduced left ventricular ejection fraction (LVEF >40%),¹ which is associated with substantial morbidity and mortality.² Clinical trials of renin-angiotensin-aldosterone system antagonists have not demonstrated similar benefits among patients with heart failure with reduced ejection fraction (HFrEF) that was 40% or lower.^3-6 However, post hoc analyses indicated that patients whose LVEF levels that ranged from 40% to 50% while being treated with either candesartan⁷ or spironolactone⁸ were admitted to the hospital less often.

In the PARAGON-HF (Prospective Comparison of ARNI [angiotensin receptor–neprilysin inhibitor] with ARB [angiotensin-receptor blockers] Global Outcomes in HFpEF) trial⁸ involving patients with heart failure and LVEF levels of 45% or higher, sacubitril/valsartan compared with valsartan resulted in a rate ratio (RR) of 0.87 (95% CI, 0.75-1.01; P = .06)⁹ for heart failure hospitalization and cardiovascular death. However, there is no evidence-based recommended therapy for renin-angiotensin-aldosterone system blockade in heart failure with LVEF higher than 40%,¹⁰ with a wide variation in background therapy in routine practice.^11,12

Quiz Ref IDIn the phase 2 PARAMOUNT (Prospective Comparison of ARNI With ARB on Management of Heart Failure With Preserved Ejection Fraction) trial¹³ involving patients with heart failure and an LVEF level of 45% or higher, sacubitril/valsartan compared with valsartan reduced N-terminal pro–brain natriuretic peptide (NT-proBNP), a prognostic biomarker in heart failure, and left atrial volume. In PARAGON-HF, reduction in NT-proBNP was associated with lower subsequent risk of heart failure hospitalizations and cardiovascular death.¹⁴

Quiz Ref IDExercise capacity, quality of life, and routine daily activities are impaired in heart failure and are regarded as important treatment targets.¹⁵ Effects of pharmacotherapies on these outcomes in heart failure with preserved ejection fraction (HFpEF) have been neutral,¹⁵ and the effects of sacubitril/valsartan on exercise capacity and quality of life in patients with heart failure and LVEF higher than 40% are incompletely understood.¹⁶

This trial was designed to test the hypothesis that sacubitril/valsartan is more effective than background medication–based individualized comparators in reducing NT-proBNP levels and improving submaximal exercise capacity and quality of life in patients with chronic heart failure and an LVEF higher than 40%.

The PARALLAX (Prospective Comparison of ARNI vs Comorbidity-Associated Conventional Therapy on Quality of Life and Exercise Capacity) trial was a 24-week, randomized, double-blind, parallel group, active-controlled (for patients taking angiotensin-converting enzyme [ACE] inhibitors or angiotensin II receptor blockers [ARBs] prior to recruitment) or placebo-controlled (for patients who were naive to a renin angiotensin system [RAS] inhibitors prior to recruitment) trial. Details of the rationale and trial design have been published¹⁷ and are available in the trial protocol (Supplement 1). This study was conducted between August 2017 and October 2019. Ethics committee approval was provided at each trial center. An independent data and safety monitoring committee monitored trial conduct and patient safety. All patients provided written informed consent.

Quiz Ref IDPatients 45 years or older, with symptomatic heart failure requiring the use of diuretics, New York Heart Association (NYHA) functional class II through IV, elevated plasma NT-proBNP levels (>220 pg/mL for patients in sinus rhythm, and >600 pg/mL for patients with atrial fibrillation or atrial flutter); with evidence of structural heart disease (either left atrial enlargement or left ventricular hypertrophy) as demonstrated by echocardiography at screening or within 6 months of randomization; with an LVEF of 40% or higher (initially defined as LVEF ≥45%, but we changed it to >40% on September 12, 2018, Amendment 2 in the trial protocol); and with an impaired health–related quality of life (Kansas City Cardiomyopathy Questionnaire Clinical Summary Score [KCCQ-CSS] <75 [range, 0-100; higher scores indicate a better quality of life]) were included. Patients taking an ACE inhibitor or ARB (background therapy) were required to have a history of hypertension.¹⁷ Detailed inclusion and exclusion criteria are listed in eTable 1 in Supplement 2. To assess potential differences in demographic data and treatment effects, patients self-reported race based on an open-ended question that was included in the data collection.

Eligible patients were stratified via interactive response technology into 1 of 3 strata according to RAS inhibitor treatment modality as prescribed by their treating physicians prior to trial recruitment: ACE inhibitor (n = 1066), ARB (n = 1174), or neither (n = 326). Each patient was assigned a randomization number using a validated system that automates the random assignment. Patients in each stratum were randomized in a 1:1 ratio to receive either sacubitril/valsartan or the background medication–based individualized comparator (Figure 1). Patients, investigator staff, persons performing the assessments, and data analysts were blinded to the identity of the treatment.

Patients in each stratum received twice daily either sacubitril/valsartan at a target dose of 97 mg/103 mg; background medication–based individualized comparator: enalapril at a target dose of 10 mg (ACE inhibitors stratum), valsartan at a target dose of 160 mg (ARB stratum); or placebo (no RAS inhibitor stratum). Sites were instructed to uptitrate study medication within 4 weeks to a maximally tolerated dose. Study visits were conducted every 2 to 4 weeks until week 24. Details of the study procedures are provided in Supplement 1 and the statistical analysis plan in Supplement 3. Both have been published.¹⁷ Training was provided to sites to ensure standardization of the 6-minute walk distance test.

The 2 primary end points for this study were change in plasma NT-proBNP concentrations from baseline to week 12 and change in the 6-minute walk distance from baseline to week 24 (in the subgroup of patients with a baseline ability of walk between 100 m and 450 m). We chose these cutoffs because walk less than 450 m during the 6-minute walk distance test could indicate impaired exercise capacity^18-20 and less than 100 m could indicate nonmodifiable causes of exercise limitation. We changed the 6-minute walk distance measure from a secondary end point to a coprimary end point on September 12, 2018 (Amendment 2, Supplement 1) reflecting the importance of functional improvement as a therapeutic target in patients with HFpEF. An increase by 30 m was considered as a minimal clinically important difference.²¹

Secondary end points included the mean change from baseline in KCCQ-CSS at week 24, proportion of patients with a 5-point score change in deterioration or improvement at week 24, and change in NYHA functional class. A 5-point increase or decrease in the KCCQ-CSS score is considered a minimal clinically important difference.²² NYHA class was graded based on functional capacity; dyspnea with mild exertion, class II; with moderate exertion, class III; and at rest, class IV.

Exploratory end points included the change in NT-proBNP levels at week 24 and the rate of change in the estimated glomerular filtration rate (eGFR) from baseline.

Safety was based on adverse events reported during the study. Suspected cases of angioedema were adjudicated by a blinded and independent angioedema adjudication committee.

A sample size of 2500 patients was estimated to provide adequate power to evaluate primary and secondary end points.¹⁷ The significance level of a 1-sided α of .025 (2-sided .05) was split between the 2 primary end points for treatment comparisons: 90% for NT-proBNP and 10% for the 6-minute walk distance. The 1-sided α was significant at .025 and the 2-sided α, at .05. With a 1-sided α of .0225, based on results of the PARAMOUNT study, the power for the NT-proBNP end point was at least 92% to detect a relative reduction from baseline of at least 11% at week 12, assuming an SD of 0.81 for change from baseline in the log-transformed NT-proBNP and an overall dropout rate of 10%.¹³ With a 1-sided α of .0025, the power for the 6-minute walk distance was at least 90% to detect a mean difference of 22 m or more in change from baseline to week 24, assuming an SD of 120 m, an overall dropout rate of 10%, and an overall proportion of 88% for patients with a baseline walk distance of between 100 m and 450 m.^17,23

Efficacy variables were analyzed based on the full analysis set. The full analysis set included all validly randomized patients except for those who were misrandomized or misstratified and did not receive any study drug. Patients were analyzed according to the treatment to which they were assigned at randomization. The safety set, which included all randomized patients receiving at least 1 dose of study drug, was used for the analyses of adverse effects.

To control the family-wise type I error rate at the 1-sided .025 significance level, a sequential rejective multiple testing procedure was used.^17,24At first, between the 2 primary end points, the overall α level was split in a 9:1 ratio. When either of them was rejected, the corresponding assigned α was propagated and accumulated to test the KCCQ-CSS. If the 6-minute walk distance was not rejected but the KCCQ-CSS was rejected at the α level inherited from the rejection of NT-proBNP change, the corresponding α level was propagated to 6-minute walk distance, together with the original assigned α, to test the 6-minute walk distance again. If both the KCCQ-CSS and 6-minute walk distance measures were rejected, the NYHA class was tested at the full level of α. Otherwise, the testing procedure was stopped. Based on this sequential rejective strategy, a significant treatment effect on either of the 2 primary end points was sufficient to consider the study successful.

The primary end point NT-proBNP was analyzed using a mixed model for repeated measures,¹⁷ from which the estimates and the 2-sided 95% CIs were provided for the adjusted geometric mean ratios for plasma NT-proBNP concentration at week 12 to the baseline in the 2 treatment groups and for sacubitril/valsartan in comparison to background medication–based individualized comparators. Similarly, the 6-minute walk distance was analyzed using the mixed model for repeated measures in the subgroup of patients with baseline walk distance ranging from 100 m to 450 m.¹⁷ The secondary end point, change in the KCCQ-CSS score at 24 weeks from baseline, was analyzed using a similar mixed model for repeated measures. The NYHA class change (improved, unchanged, and worsened) at week 24 from baseline was analyzed using a proportional cumulative odds model.¹⁷ Patients experiencing a 5-point or more improvement or deterioration in KCCQ-CSS scores were summarized by treatment group and visit, using the number of patients showing more than a 5-point change and the proportion of patients with a 5-point or more change. and analyzed using binary logistic regression models.

In the analyses of NT-proBNP concentration, missing data were handled using the likelihood method, with the assumption that they were missing at random. In the analyses of the 6-minute walk distance, KCCQ-CSS, and NYHA class change, death was considered the worst possible outcome and scheduled visits after death were imputed to the worst value accordingly. Other missing data were assumed to be missing at random.¹⁷

Statistical analyses were performed using SAS version 9.4 or higher (SAS Institute Inc).

A sensitivity analysis was carried out to assess cases for which data were missing not at random; in particular, the sensitivity analysis assumed that patients treated with sacubitril/valsartan who had discontinued due to adverse events, death, or lack of efficacy would not have adhered to therapy if they had stayed in the study.

To assess consistency of treatment effects across key clinically relevant subgroups, subgroup analyses were performed using the corresponding analysis model used for the primary and secondary end points, after adding factors of subgroup and subgroup-treatment interaction to the model. However, to allow for an unbiased estimation of within-subgroup treatment effects and an assessment of homogeneity of treatment effects at the specific target visit, post hoc subgroup analyses were instead conducted by adding fixed-effect factors of subgroup, subgroup × treatment interaction, subgroup × visit interaction, and subgroup × treatment × visit interaction into the original primary analysis model. The estimated treatment effect, 2-sided 95% CIs, and interaction P values are reported for each of the subgroups.

Between August 22, 2017, and May 7, 2019, 4632 patients were screened at 396 centers in 32 countries, and 2572 patients were randomized. The study ended as planned after the last patient completed the 24-week follow-up visit on October 28, 2019. Five patients who were misrandomized and 1 patient who was mistakenly assigned to a wrong stratum were withdrawn from the trial without having received study medication and excluded from primary analysis. The primary intention-to-treat population consisted of 2566 patients (1281 randomized to the sacubitril/valsartan group and 1285 randomized to background medication–based individualized comparator group; Figure 1). Before randomization, 1066 patients (41.5%) were taking ACE inhibitors (ACE inhibitor stratum), 1174 patients (45.8%) were taking ARBs (ARB stratum), and 326 patients (12.7%) were not taking any RAS inhibitors. Baseline demographic and clinical characteristics were balanced between the treatment groups (Table 1 and Table 2). A comparison of clinical characteristics among the 3 strata has been described previously,¹⁷ with fewer patients who were not taking a RAS inhibitors reporting a history of hypertension, hyperlipidemia, or coronary artery disease and who had slightly lower blood pressure at entry into the study than those in the other 2 strata.

At week 12, NT-proBNP levels were significantly reduced in the sacubitril/valsartan group vs the comparator group. The adjusted geometric mean ratio was 0.84 (95% CI, 0.80-0.88; P < .001; Figure 2). Reductions in NT-proBNP levels in the sacubitril/valsartan group were evident as early as week 4 and remained lower than levels in the background medication–based individualized comparator group at week 24 (adjusted geometric mean ratio to baseline, 0.82 vs 0.98 pg/mL; Figure 2). Furthermore, reductions in NT-proBNP levels in the sacubitril/valsartan group was not affected by prior treatment, with no significant interactions among strata (Figure 3). The reduction in NT-proBNP levels at week 12 was consistent across all prespecified subgroups with no significant subgroup by treatment interaction (Figure 3).

At week 24, no significant difference existed between study groups in the 6-minute walk distance from baseline (a mean increase, 9.7 m for the sacubitril/valsartan group vs 13.2 m for the comparator group). The adjusted mean difference was −2.5 m (95% CI, −8.5 to 3.5 m; P = .42; Figure 2). This finding was consistent across subgroups and prestudy treatment strata, except for a modest effect modification by geographical region and sex: among patients in the sacubitril/valsartan group compared with the background medication–based individualized comparator group, walking distance improved among women but decreased among men and decreased among those living in Latin America; however, the effect sizes remained small (eFigure in Supplement 2). A similar proportion of patients improved their walk distance by 30 m or more: 36.0% in the in the sacubitril/valsartan group and 35.4% in the comparator group. A prespecified sensitivity analysis that included all randomized patients (irrespective of their baseline 6-minute walk distance) confirmed the results of the primary analysis (adjusted mean difference, −3.0 m; 95% CI, −8.8 to 2.8 m; P = .31).

There was no significant difference between groups in the KCCQ-CSS score at week 24 compared with baseline (12.3 points in the sacubitril/valsartan group and 11.8 points in the comparator group, P = .24; Table 3). The proportion of patients with an improvement or decrease by 5 or more points were not significantly different in the treatment groups. Changes in NYHA class were also not significantly different between treatment groups. Change from baseline in KCCQ-CSS and NYHA class at week 24 by strata are presented in eTable 2 in Supplement 2.

At week 24, The sacubitril/valsartan group had a significantly lower rate of monthly change from baseline (slope) in eGFR than did the background medication–based individualized comparator group (−0.25 vs −0.43 mL/min/1.73 m², respectively, P = .02). The annualized rate of eGFR decline was −2.95 mL/min/1.73 m² in the sacubitril/valsartan group and −5.14 mL/min/1.73 m² in the background medication–based individualized comparator group, corresponding to an annualized adjusted mean treatment difference of 2.19 mL/min/1.73 m² (95% CI, 0.41 to 3.97; P = .01).

Permanent study drug discontinuation was reported in 175 patients (13.6%) in the sacubitril/valsartan group and 147 patients (11.6%) in the comparator group. Adverse events were the main reason for permanent study drug discontinuation: 116 patients (9.0%) in the sacubitril/valsartan group and 87 (6.8%) in the comparator group.

A total of 1087 patients (84.9%) in the sacubitril/valsartan group and 1030 patients (80.2%) in the background medication–based individualized comparator group had at least 1 adverse event (P = .002). Overall, a higher proportion of patients in the sacubitril/valsartan group reported study drug–related adverse events vs patients in the background medication–based individualized comparator group (30.5% vs 22.6%; P < .001; eTable 3 in Supplement 2). Hypotension (14.1% vs 5.5%), albuminuria (12.3% vs 7.6%), and hyperkalemia (11.6% vs 10.9%) were the most commonly reported adverse events in the sacubitril/valsartan group, especially in the no RAS inhibitor stratum (eTable 4 and eTable 5 in Supplement 2). A higher proportion of patients in the sacubitril/valsartan group had a systolic blood pressure less than 90 mm Hg and higher than a 20 mm Hg decrease from baseline, notably in the no RAS inhibitor stratum (eTable 5 in Supplement 2).

Seventy patients (5.5%) in the sacubitril/valsartan group had cardiac disorders reported as serious adverse events compared with 94 (7.3%) in the comparator group.

Quiz Ref IDIn this trial involving patients with heart failure and LVEF higher than 40%, sacubitril/valsartan was more effective in reducing plasma NT-proBNP levels but did not improve 6-minute walk distance, quality of life (as assessed by KCCQ-CSS), or NYHA functional class compared with background medication–based individualized comparators.

The reduction in NT-proBNP observed with sacubitril/valsartan was consistent across 9 prespecified subgroups. This observation corroborates previous findings from the PARAMOUNT¹³ and PARAGON-HF⁹ trials involving patients with an LVEF of 45% or higher. Although there is robust evidence showing NT-proBNP reduction with sacubitril/valsartan in patients with HFrEF and HFpEF (LVEF ≥45%), the results of this study extend these findings to patients in the LVEF range higher than 40%, encompassing both heart failure with mildly reduced ejection fraction (HFmrEF) and HFpEF, without a trial run-in period and compared with different RAS inhibitor blockers or placebo. The reduction of NT-proBNP levels with sacubitril/valsartan in this study may primarily be attributed to neprilysin inhibition because the ARB irbesartan had a negligible effect on NT-proBNP levels in the Irbesartan in Heart Failure With Preserved Systolic Function (I-Preserve) trial,⁴ and the reduction was seen across all 3 strata in the current study.

The second primary end point, change from baseline in 6-minute walk distance at week 24, was not significantly different between groups. The observed sex interaction for the 6-minute walk distance with women deriving more benefit than men may need further consideration because women benefited more from angiotensin receptor neprilysin inhibitors than did men in the PARAGON-HF trial.⁹

Prior randomized trials of heart failure pharmacotherapies have failed to show benefits for exercise capacity in patients with HFpEF, with 6-minute walk distance improving in both the control and active intervention groups in several trials.¹⁵ Specifically, there was no improvement in 6-minute walk distance with vericiguat,²⁵ praliciguat,²⁶ sildenafil,²⁷ and neladonoson,²⁸ and submaximal exercise capacity decreased slightly with spironolactone²⁹ compared with placebo. One explanation may be that the 6-minute walk distance measure is not sufficiently sensitive to detect differences with pharmacological interventions in heart failure compared with implantable electronic devices³⁰ or a ventricular assist device.³¹ Additionally, therapies might reduce heart failure hospitalizations without improving exercise capacity. In HFrEF, changes in 6-minute walk distance over time bear prognostic information.³² The sodium-glucose co-transporter-2 inhibitor empagliflozin reduced heart failure hospitalizations³³ but had no effect on exercise capacity in HFrEF or HFpEF.³⁴ Similarly, sacubitril/valsartan reduced heart failure hospitalizations and cardiovascular mortality in HFrEF³⁵ but did not improve 6-minute walk distance.³⁶

Quiz Ref IDAlternatively, the lack of improvement of 6-minute walk distance with sacubitril/valsartan may be explained by the heterogeneity of heart failure, with distinct pathophysiological pathways underlying reduced exercise capacity. Better phenotyping³⁷ might help to tailor therapy to specific subsets of patients with heart failure that benefit from a specific intervention. In addition, patients in this trial were older (>80% >65 years), with considerable comorbidity, and the heart failure syndrome may not have been the sole limiting factor for submaximal exercise capacity. Future research should test the combination of therapies with beneficial effects on cardiac remodeling (eg, sacubitril/valsartan⁹ or spironolactone²⁹) with therapies beneficial for quality of life (eg, exercise training).³⁸

The adverse event profile of sacubitril/valsartan was consistent with previous trials.^9,13,35 Hypotension was more common with sacubitril/valsartan, particularly in the no RAS inhibitor stratum, in which patients treated with sacubitril/valsartan more commonly had hyperkalemia and kidney impairment than did patients treated with placebo. Hence, initiation of sacubitril/valsartan at low dosages and a longer, gradual uptitration may result in improved tolerability in RAS inhibitor–naive patients.³⁹

Unlike many other clinical trials, this study recruited only patients with reduced quality of life at screening. This approach was chosen because patients with lower quality of life may have more room for improvement. In line with this, large improvements from baseline in mean KCCQ scores were observed at week 24 in both treatment groups (with no significant differences between groups). However, because low KCCQ scores in patients with heart failure are often multifactorial, it may be that (similarly to 6-minute walk distance) further improvements in KCCQ are affected by persistent comorbidities.

This study has several limitations. First, the study did not have a run-in phase as used in previous trials. The advantage of this design is that the data are more likely to reflect clinical practice; the disadvantage is that differences between groups are smaller because patients who are intolerant of target drug dosages dilute the difference between groups.

Second, the comparator therapy was stratified based on the use of RAS inhibitor therapy prior to randomization. This approach was justified by the lack of data supporting routine use of RAS inhibitor therapy in patients with heart failure and LVEF higher than 40%. In this trial, nearly 90% of patients were taking either an ACE inhibitor or ARB prior to randomization, reflecting routine practice. These data complement the PARAMOUNT¹³ and PARAGON-HF⁹ trials, which showed that NT-proBNP plasma levels are reduced by sacubitril/valsartan in comparison with an ARB.

Third, although plasma natriuretic peptides have been shown to be associated with heart failure outcomes, this trial was of relatively short duration (24 weeks). Whether a longer treatment duration would result in significant differences between treatment groups in outcomes such as 6-minute walk distance, KCCQ scores, or both remains to be elucidated.

Fourth, obesity is a frequent comorbidity in patients with HFpEF, and obese patients have lower NT-proBNP levels. This study included patients with heart failure and a body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) of 40 or lower and elevated NT-proBNP levels. Although patients with BMIs higher than 40 were excluded from this study because of its possible effect on the walk distance measure, the median BMI of the patients enrolled was 30.4 (IQR, 27.0-34.1) demonstrating that the majority of the patients were overweight or obese.

Among patients with heart failure and left ventricular ejection fraction higher than 40%, sacubitril/valsartan treatment compared with standard renin angiotensin system inhibitor treatment or placebo resulted in a significantly greater decrease in plasma N-terminal pro–brain natriuretic peptide levels at 12 weeks but did not significantly improve 6-minute walk distance at 24 weeks. Further research is warranted to evaluate potential clinical benefits of sacubitril/valsartan in these patients.

Corresponding Author: Burkert Pieske, MD, Department of Internal Medicine and Cardiology, Charité University Medicine, Campus Virchow Klinikum, Augustenburger Platz 1, Berlin 13353, Germany (burkert.pieske@charite.de).

Accepted for Publication: September 28, 2021.

Author Contributions: Drs Shah and Pieske had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Pieske, Wachter, Shah, Szecsödy, Ibram, Shi, Cowie.

Acquisition, analysis, or interpretation of data: Pieske, Wachter, Shah, Baldridge, Szecsödy, Ibram, Shi, Zhao, Cowie.

Drafting of the manuscript: Pieske, Wachter, Ibram, Cowie.

Critical revision of the manuscript for important intellectual content: Pieske, Wachter, Shah, Baldridge, Szecsödy, Shi, Zhao, Cowie.

Statistical analysis: Baldridge, Zhao.

Obtained funding: Ibram.

Administrative, technical, or material support: Shah, Szecsödy.

Supervision: Pieske, Wachter, Shah, Szecsödy, Ibram, Cowie.

Conflict of Interest Disclosures: Dr Pieske reported receiving personal fees from Bayer, Bristol Myers Squib, Medscape, MSD, Novartis, and Servier and grants and personal fees from Astra-Zeneca. Dr Wachter reported receiving personal fees from Daiichi Sankyo, Gilead, Novartis, Pfizer, Pharmacosmos, and Servier; grants and personal fees from Boehringer Ingelheim, CVRx, and Medtronic; grants from Bundesministerium für Bildung und Forschung, the European Union, and Deutsche Forschungsgemeinschaft. Dr Shah reported receiving grants and personal fees from Actelion, AstraZeneca, Novartis, and Pfizer; grants from Corvia; and personal fees from Abbott, Amgen, Aria CV, Axon, Bayer, Boehringer-Ingelheim, Boston Scientific, Bristol Myers Squib, Cardiora, CVRx, Cyclerion, Cytokinetics, Edwards Lifesciences, Eisai, Imara, Ionis, Ironwood, Janssen, Lilly Medical, Merck, MyoKardia, Novo Nordisk, Prothena, Regeneron, Sanofi, Shifamed, Tenax, and United Therapeutics. Drs Szecsödy, Shi, and Zhao and Ms Ibram, reported being employees of study sponsor (Novartis). Dr Cowie reported receiving personal fees from AstraZeneca, Bayer, Boehringer Ingelheim, Medtronic, Novartis, Novo Nordisk, Roche, and Servier and grants and personal fees from Abbott and Boston Scientific. No other disclosures were reported.

Funding/Support: This study was funded by Novartis Pharma AG.

Role of the Funder/Sponsor: Novartis participated in the design and conduct of the study; collection, management, analysis, and interpretation of the data; review of the manuscript; and decision to submit the manuscript for publication. All statistical analyses were verified by an independent academic statistician (Ms Baldridge). The sponsor did not have the right to veto publication and did not have control regarding to which journal the manuscript was submitted. Some of the individuals employed by the sponsor.

Steering/Executive Committee: Burkert Pieske (chair), Rolf Wachter, Martin R. Cowie, Sanjiv J. Shah.

Data Monitoring Committee: Henry Dargie (chair), Western Infirmary, Glasgow, Scotland, United Kingdom; Michel Komajda, Goupe Hospitalier Pitiesalpetriere Institut De Cardiologie, Paris, France; Robert Foley, Chronic Disease Research Group Minneapolis, Minnesota; Gary Francis, University of Minnesota, Minneapolis; Stuart Pocock, London School of Hygiene and Tropical Medicine, London, United Kingdom.

Angioedema Adjudication Committee: Allen P. Kaplan (chair), Medical University of South Carolina, Charleston; Nancy Brown, Vanderbilt School of Medicine, Nashville, Tennessee; Bruce Zuraw, University of California, San Diego; Paula J. Busse, Division of Clinical Immunology, Mount Sinai School of Medicine, New York, New York.

Group Information: The PARALLAX study committees and investigators are listed in Supplement 4.

Data Sharing Statement: See Supplement 5.

Additional Contributions: We thank Nagabhushana Ananthamurthy, PhD, Viswanatha Matada, MPharm, and Shalini Verma, PhD (associates of Novartis) for editorial support to the authors.