Derivation of the population of patients potentially newly eligible for sacubitril/valsartan under expanded US Food and Drug Administration (FDA) labeling based on National Health and Nutrition Examination Survey (NHANES) data from 2015 to 2018. eGFR indicates estimated glomerular filtration rate; HFrEF, heart failure with reduced ejection fraction; LVEF, left ventricular ejection fraction.
A, Single ranges of LVEF are offered for sacubitril/valsartan expanded eligibility in the overall population. B, Potential sex-specific LVEF-based eligibility is considered. HF indicates heart failure.
eMethods. NHANES population estimates, GWTG-HF left ventricular ejection fraction distribution, and PARAGON-HF trial treatment effect estimates.
eResults. Sensitivity analysis of conservative estimate of potentially eligible treatment candidates.
eTable. Characteristics of NHANES participants with heart failure without potential clinical intolerance based on blood pressure or kidney function.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Vaduganathan M, Claggett BL, Greene SJ, et al. Potential Implications of Expanded US Food and Drug Administration Labeling for Sacubitril/Valsartan in the US. JAMA Cardiol. 2021;6(12):1415–1423. doi:10.1001/jamacardio.2021.3651
What are the implications of expanded US Food and Drug Administration (FDA) labeling for sacubitril/valsartan for use in individuals with chronic heart failure (HF) with left ventricular ejection fraction (LVEF) lower than normal in the US?
In this decision analytical model study including an estimated 4 682 098 adults, the number of adults projected to potentially be newly eligible varied by the definition of FDA labeling of lower than normal LVEF from 643 161 (LVEF of 41% to 50%) to 1 838 756 (LVEF of 41% to 60%). Comprehensive implementation of sacubitril/valsartan among newly eligible patients was empirically estimated to prevent up to 69 268 worsening HF events (LVEF of 41% to 50%) to 182 592 worsening HF events (LVEF of 41% to 60%) over 3 years of treatment.
The expanded FDA labeling is positioned to substantially increase the potential HF population eligible for sacubitril/valsartan by up to 1.8 million individuals and has the potential to prevent or postpone as many as 180 000 worsening HF events, depending on the definition of normal LVEF.
The US Food and Drug Administration (FDA) expanded labeling for sacubitril/valsartan for use in individuals with chronic heart failure (HF) with left ventricular ejection fraction (LVEF) lower than normal. The population-level implications of implementation of sacubitril/valsartan at higher LVEF ranges is unknown. While the Prospective Comparison of ARNI With ARB Global Outcomes in HF With Preserved Ejection Fraction (PARAGON-HF) trial did not meet its primary end point, the trial may provide useful information in projecting expected clinical events among treated individuals.
To quantify newly eligible treatment candidates for sacubitril/valsartan under the expanded FDA labeling and to apply treatment effects and the number needed to treat (NNT) to prevent 1 worsening HF event derived from subgroups of the PARAGON-HF trial who fall under the revised FDA label.
Design, Setting, and Participants
Newly eligible treatment candidates were estimated by mapping the LVEF distribution from 559 520 adult patients hospitalized between 2014 and 2019 in the Get With The Guidelines–Heart Failure registry to adults self-identifying with HF in the National Health and Nutrition Examination Survey (2015 to 2018). The NNT with 3 years of treatment for 3 end points of interest (total HF hospitalizations, total HF hospitalizations and cardiovascular death, and total HF hospitalizations and urgent HF visits and cardiovascular death) were estimated from the PARAGON-HF trial. Data were analyzed from February to June 2021.
Main Outcomes and Measures
Number of worsening HF events prevented or postponed if eligible patients were treated with sacubitril/valsartan for 3 years.
Of an estimated 4 682 098 adults, the mean (SE) age was 66.3 (0.8) years, 1 995 037 (42.6%) were women, and 748 045 (16.0%) were Black. The potential number of adults projected to be newly eligible varied by the definition of FDA labeling of lower than normal LVEF from 643 161 (95% CI, 534 433-751 888; LVEF of 41% to 50%) to 1 838 756 (95% CI, 1 527 911-2 149 601; LVEF of 41% to 60%). In the PARAGON-HF trial, the NNT to prevent a worsening HF event (range, 7 to 12 patients) was consistent irrespective of specific LVEF range selected. Comprehensive implementation of sacubitril/valsartan among newly eligible patients was empirically estimated to prevent up to 69 268 (95% CI, 57 558-80 978) worsening HF events (LVEF of 41% to 50%) to 182 592 (95% CI, 151 725-213 460) worsening HF events (LVEF of 41% to 60%).
Conclusions and Relevance
The expanded FDA labeling is positioned to substantially increase the potential HF population eligible for sacubitril/valsartan by up to 1.8 million individuals and has the potential to prevent or postpone as many as 180 000 worsening HF events, depending on the definition of normal LVEF.
Despite experiencing heightened risks of worsening heart failure (HF) events, adverse health-related quality of life, and death,1 patients with HF with preserved ejection fraction (HFpEF) have limited evidence-based therapeutic strategies. On February 16, 2021, the US Food and Drug Administration (FDA) expanded regulatory labeling for sacubitril/valsartan to reduce the risk of cardiovascular death and hospitalization for HF in adult patients with chronic HF. These regulatory changes were based on data from the Prospective Comparison of ARNI With ARB Global Outcomes in HF With Preserved Ejection Fraction (PARAGON-HF) trial (which enrolled participants with HF and left ventricular ejection fraction [LVEF] of 45% or greater), contextualized alongside data from the Prospective Comparison of Angiotensin Receptor Neprilysin Inhibitor With Angiotensin-Converting Enzyme Inhibitor to Determine Impact on Global Mortality and Morbidity in Heart Failure (PARADIGM-HF) trial (which enrolled participants with HF and LVEF of 40% or less). While the PARAGON-HF trial did not meet its primary end point, the totality of evidence, inclusive of supportive prespecified secondary and sensitivity analyses, favored sacubitril/valsartan over valsartan and led the FDA to consider a supplemental new drug application for expanded use.2 After a favorable Cardiovascular and Renal Drugs Advisory Committee vote (12 to 1) on December 15, 2020, the FDA formally expanded labeling for sacubitril/valsartan. Notably, in the PARAGON-HF trial, there was a significant interaction between LVEF and treatment efficacy, whereby participants with an LVEF at or below the median value of 57% appeared to garner greater clinical benefit.3 Indeed, the FDA emphasized that benefits are most clearly evident in patients with LVEF lower than normal. However, the specific label language did not specify a cutoff to define abnormal, recognizing imprecision in the estimation of LVEF.4 To our knowledge, no data are available regarding the potential implications of this expanded labeling on US clinical practice. As such, we projected the number of patients with HF who have become newly eligible for sacubitril/valsartan under FDA labeling and estimated HF events that may potentially be prevented or postponed with comprehensive implementation of sacubitril/valsartan in the US. Recognizing clinical uncertainty regarding identifying potentially eligible patients based on LVEF measurements, we have estimated the number needed to treat (NNT) to prevent a worsening HF event or number needed to harm (NNH) to cause a safety event (hypotension) across a range of LVEF cutoffs to define lower than normal.
The prevalence of HF was estimated using self-reported weighted data from the National Health and Nutrition Examination Survey (NHANES) from 2015 to 2018 (n = 19 255) (eMethods in the Supplement). To capture a clinically actionable cohort for sacubitril/valsartan, we narrowed the potentially eligible population by excluding those with a systolic blood pressure (SBP) less than 100 mm Hg and/or with an estimated glomerular filtration rate (eGFR) less than 30 mL/min/1.73 m2.
In sensitivity analyses, we further excluded those receiving hospice or comfort care and those with inotrope use, ventricular assist devices/transplant, and expected clinical intolerance to renin-angiotensin system inhibitors or sacubitril/valsartan (estimated based on run-in failure to either valsartan or sacubitril/valsartan in the PARAGON-HF trial).5,6
We then leveraged data from the American Heart Association Get With The Guidelines–Heart Failure (GWTG-HF) registry7 to identify a contemporary distribution of LVEF among adult patients 18 years and older hospitalized between January 1, 2014, and September 30, 2019, at 529 US sites (eMethods in the Supplement). Patients with an LVEF of 40% or less were considered previously eligible for sacubitril/valsartan based on prior FDA labeling. We estimated potentially newly eligible patients with HF at higher LVEF ranges for sacubitril/valsartan based on 4 LVEF cutoffs to define lower than normal: (1) 41% to 50%, which covers HF with midrange ejection fraction in clinical practice guidelines8,9; (2) 41% to 55%, in which the upper limit of the confidence interval of the treatment effect is entirely confined to the region of benefit in a pooled analysis of the PARAGON-HF and PARADIGM-HF trials3; (3) 41% to 57%, which was the median LVEF value in the PARAGON-HF trial and defined the prespecified subgroup (LVEF of 57% or less) that preferentially benefited in the trial; and (4) 41% to 60%, which is the nearest 5-point integer below the population-wide average estimates for LVEF in men and women.10
The PARAGON-HF trial was a global randomized clinical trial that examined sacubitril/valsartan against an active comparator valsartan among adults 50 years and older with symptomatic HFpEF with LVEF of 45% or more (eMethods in the Supplement).2,11 To capture the total burden of HF events, 3 end points of interest were considered in the present study: (1) total (first and recurrent) HF hospitalizations; (2) total HF hospitalizations and cardiovascular death, the primary end point of the PARAGON-HF trial; and (3) an expanded composite inclusive of urgent HF visits, which was an additional protocol-specified end point.12 We identified risks of hypotension, defined in 3 ways: (1) hypotension as a serious adverse event leading to permanent drug discontinuation (a prespecified adverse event of interest); (2) hypotension reported as any serious adverse event; and (3) SBP less than 100 mm Hg irrespective of symptoms or clinical action. Incidence rates and confidence intervals for both efficacy and safety events were derived from Poisson models with robust variance estimates. Treatment effects of sacubitril/valsartan vs valsartan were summarized for each of the 3 recurrent event end points as rate ratios (RRs) and 95% CIs obtained using the semiparametric proportional rates method of Lin et al,13 stratified by geographic region. Treatment effects for hypotension (SBP less than 100 mm Hg) were summarized as hazard ratios and 95% CIs using Cox proportional hazards models, stratified by geographic region.
Incidence rates (expressed per 100 patient-years) were estimated using Poisson regression models with robust variance estimation (to account for correlation between events). Recurrent event rates were annualized and assumed to be consistent over the treatment period. Traditionally, NNT reflects the number of patients who need to be treated to prevent 1 patient from experiencing an event. In this case, we estimated an event-based NNT as the number of patients needed to be treated to prevent 1 worsening HF event. This event-based NNT was estimated as the inverse of the absolute risk difference between study arms for each of the 3 end points. The recurrent event measure is expressed as NNT per patient-year of treatment exposure rather than per patient treated.14,15
Treatment effects were estimated for the 4 potentially eligible populations defined by LVEF ranges. Patients with LVEF greater than 40% but less than 45% were not covered by eligibility criteria in either the PARAGON-HF trial or PARADIGM-HF trial. We assumed that risk estimates in this LVEF range would be comparable with that observed among patients with adjacent higher LVEF in the PARAGON-HF trial.
To facilitate interpretability, we forecasted the number of events prevented or caused if patients were treated with sacubitril/valsartan for 3 years. We estimated the number of worsening HF events that would be prevented or postponed and the number of hypotension events caused if sacubitril/valsartan were completely implemented for 3 years among all newly eligible patients across the 4 LVEF ranges.
Given population variation in LVEF distribution10 and treatment heterogeneity by sex in the PARAGON-HF trial,16 sex-specific cutoffs may be considered to define abnormal LVEF. Compared with a more conservative LVEF range in men (41% to 55%), we evaluated a wider LVEF range (41% to 60%) among women, as the upper bounds of the 95% CI was confined to a region of benefit when LVEF was analyzed continuously.3 We applied sex-stratified LVEF distributions to published estimates of HF prevalence to identify potential new treatment candidates. We then estimated treatment effects, 3-year NNT, and 3-year NNH from the PARAGON-HF trial separately in men and women in these indicated LVEF ranges. We finally projected the number of worsening HF events that could be averted and number of hypotension events caused with optimal treatment according to the sex-specific expanded eligibility criteria. Statistical analyses were performed using Stata version 16.1 (StataCorp) and R version 4.0.3 (The R Foundation), and a 2-sided P value less than .05 was considered significant.
A flowchart outlining patient identification under the expanded FDA labeling for sacubitril/valsartan is summarized in Figure 1. The prevalence of self-reported HF among US adults is estimated to be 5 604 076 (95% CI, 4 674 944-6 533 208). After excluding those with SBP less than 100 mm Hg or eGFR less than 30 mL/min/1.73 m2, 4 682 098 (95% CI, 3 890 581-5 473 615) adult patients with HF were identified. Demographic and clinical characteristics of these men and women with HF in NHANES are described in the eTable in the Supplement. The mean (SE) age was 66.3 (0.8) years, 1 995 037 (42.6%) were women, and 748 045 (16.0%) were Black.
Among 586 580 patients hospitalized for HF in the GWTG-HF registry, 27 060 (4.6%) did not have available LVEF measurements. In those with available data (n = 559 520), we described relative proportions of patients who would have been previously eligible under prior FDA labeling and those who would be newly eligible for sacubitril/valsartan (Figure 1). Overall, 45.4% of patients with HF (an estimated 2 124 182 individuals) would have already met LVEF criteria for the prior HF with reduced ejection fraction indication. Expanded treatment eligibility under new FDA labeling ranged from an additional 14% (if LVEF was considered up to 50%) to 39% (if LVEF was considered up to 60%) (Figure 2). The corresponding potential number of adults projected to be newly eligible ranged from 643 161 (95% CI, 534 433-751 888; LVEF of 41% to 50%) to 1 838 756 (95% CI, 1 527 911-2 149 601; LVEF of 41% to 60%).
Among the 4796 patients in the PARAGON-HF trial, the 3-year NNT was 20 for total HF hospitalizations, 19 for total HF hospitalizations and CV death, and 17 for total worsening HF events and CV death (Table 1). Compared with those taking valsartan, among those taking sacubitril/valsartan, hypotension defined as a serious adverse event leading to drug discontinuation (59 of 2407 [2.5%] vs 55 of 2389 [2.3%]; P = .73) or as any serious adverse event (116 of 2407 [4.8%] vs 122 of 2389 [5.1%]; P = .65) was not significantly increased, and as such, NNH could not be estimated. When hypotension was defined by hemodynamic criteria alone (SBP less than 100 mm Hg), sacubitril/valsartan increased risks of hypotension compared with control valsartan (380 of 2407 [15.8%] vs 257 of 2389 [10.8%]; P < .001), and the 3-year NNH to cause 1 hemodynamic hypotensive event was 16 in the overall PARAGON-HF trial. Across each of the 4 LVEF ranges (45% to 50%, 45% to 55%, 45% to 57%, and 45% to 60%) potentially encompassed by the expanded FDA labeling, 3-year NNT ranged from 7 to 12 for all 3 end points of interest, and 3-year NNH ranged from 20 to 21 for hypotension (SBP less than 100 mm Hg) (Table 1). If these risk reductions and treatment estimates are projected to the US population of patients with HF, comprehensive implementation of sacubitril/valsartan would be anticipated to prevent or postpone up to 69 268 (95% CI, 57 558-80 978) worsening HF events (LVEF of 41% to 50%) to 182 592 (95% CI, 151 725-213 460) worsening HF events (LVEF of 41% to 60%) over 3 years of treatment. Comprehensive implementation of sacubitril/valsartan would be expected to cause up to 32 029 (95% CI, 26 614-37 443) hemodynamic hypotensive events (LVEF of 41% to 50%) to 86 164 (95% CI, 71 598-100 730) hemodynamic hypotensive events (LVEF of 41% to 60%) over 3 years of treatment (Table 2).
Among 263 805 women in the GWTG-HF registry, 87 960 of 263 805 (33.3%) had an LVEF that fell within prior FDA labeling for sacubitril/valsartan, and an additional 117 821 of 263 805 (44.7%) would meet expanded FDA labeling if sex-specific LVEF criteria were expanded to 60% or less. Among 295 715 men in the GWTG-HF registry, 165 884 of 295 715 (56.1%) had an LVEF that met prior FDA labeling criteria; an additional 70 781 of 295 715 (23.9%) would newly qualify if sex-specific LVEF criteria were expanded to 55% or less (Figure 2B). We estimate the potential number of women and men projected to be newly eligible to be 1 200 099 (95% CI, 984 086-1 416 122) and 477 523 (95% CI, 365 868-589 178), respectively. In the PARAGON-HF trial, the 3-year NNT within the sex-specific LVEF ranges was 5 (for women) and ranged from 27 to 34 (for men) across all worsening HF end points (Table 3). The 3-year NNH for hemodynamic hypotension within the sex-specific LVEF ranges was 18 in women and 27 in men. When these 3-year NNT and NNH estimates were applied across potentially eligible patients in the US, up to 238 340 (95% CI, 195 439-281 240) and 17 477 (95% CI, 13 391-21 564) worsening HF events in women and men, respectively, could be averted, and 66 849 (95% CI, 54 816-78 881) and 17 723 (95% CI, 13 579-21 867) hemodynamic hypotensive episodes would be caused with complete implementation of sacubitril/valsartan for 3 years (Table 2). Sensitivity analysis of a more restrictive eligible population identified modestly lower potentially averted worsening HF events, but this estimate remained more than twice that of potentially resultant hypotensive episodes with full implementation (eResults in the Supplement).
Clinical HF occurring among patients without frankly reduced LVEF (eg, 40% or less) is poised to continue to grow with population aging and increasing cardiometabolic comorbidities. Worsening HF events in this at-risk population are frequent, costly, prognostically meaningful,17 and may accelerate disease progression.18 The updated FDA labeling for sacubitril/valsartan expands its reach to previously ineligible patients with higher LVEF who may stand to benefit in preventing worsening HF events.
This study attempts to summarize the potential implications of broader implementation of sacubitril/valsartan under the guidance of this expanded FDA labeling. We present data from the PARAGON-HF trial specific to the cohort potentially encompassed by the FDA labeling that quantifies substantial and clinically relevant risk reductions for a number of end points should the effectiveness and safety in clinical practice match the efficacy in the PARAGON-HF clinical trial. In many cases, 10 or fewer patients would need to be treated for 3 years to prevent a worsening HF event. These risk reductions and NNT estimates were largely consistent irrespective of the individual LVEF range selected to define lower than normal. Indeed, across every LVEF range encompassed by the expanded label, approximately twice the number of worsening HF events is projected to be averted as the number of resultant hypotensive episodes, even when applying a liberal definition based on low blood pressure alone (irrespective of symptoms or clinical consequences). These data support a favorable risk-benefit margin for most newly eligible treatment candidates. Additionally, even patients who experience hypotension with initial drug exposure may derive subsequent clinical benefit, as has been previously shown in the PARADIGM-HF trial.19
We describe large variation in the potentially eligible population based on specific LVEF ranges selected. For instance, if an LVEF of 41% to 60% was considered lower than normal, a projected additional 1.8 million patients could be considered for sacubitril/valsartan. In evaluating the entire treatment indication for sacubitril/valsartan in individuals with chronic HF, approximately 4 million patients would meet eligibility for initiation or switching to sacubitril/valsartan. However, notably expanding the treatment eligible population based on the most conservative interpretation of lower than normal (ie, an LVEF of 41% to 50%) would result in far fewer potential new treatment candidates (less than 700 000).
It has been posited that given the higher average LVEF in women than men,10 the cutoff to define lower than normal LVEF should similarly be higher in women. Our data indicate that more men are estimated to have prevalent HF in the US, and a greater number were previously eligible for sacubitril/valsartan. In contrast, applying sex-specific cutoffs would greatly expand the number of women who become eligible under revised FDA labeling. In the PARAGON-HF trial, heterogeneity in treatment effects by sex was observed with greater benefits seen in women.16 In addition, women appeared to derive benefits from sacubitril/valsartan at a higher LVEF range than men.3 Given these amplified benefits, only 5 women need to be treated for 3 years to prevent a worsening HF event, while 18 women would need to be exposed to sacubitril/valsartan over 3 years to result in a hypotensive episode (any SBP less than 100 mm Hg). If broadly implemented among newly eligible women with HF, more than 200 000 worsening HF events may be averted over 3 years; these projected figures are 14-fold the number of events potentially averted if sacubitril/valsartan was fully used among newly eligible men. While sex was a prespecified subgroup in the PARAGON-HF trial and although similar 2-way interactions with sex and LVEF have been described with other therapies in individuals with HFpEF,20 these findings require validation, and their implications for clinical practice remain uncertain.
While many factors may limit the effectiveness of sacubitril/valsartan in preventing worsening HF events in clinical practice, we have attempted to apply conservative estimates of the potentially eligible cohort and treatment effects derived from trial data. First, participants in NHANES who had missing blood pressure and/or eGFR data (each less than 10% missingness) were assumed to not meet treatment eligibility under the expanded label. This conservative assumption partially mirrors clinical practice in which clinicians may be reluctant to treat patients who do not have recent clinical data. Second, the PARAGON-HF trial was an active-controlled trial, and these data highlight the incremental benefit of sacubitril/valsartan over valsartan. As current use of renin-angiotensin system inhibitors in real-world US clinical practice is incomplete (54% to 58% among participants with HF with midrange or preserved LVEF in the GWTG-HF registry),1 these projections may underestimate the treatment benefits of implementation of sacubitril/valsartan among renin-angiotensin system inhibitor–naive patients.21 However, renin-angiotensin system inhibitors have been used as background therapy in approximately 80% of participants in more recent global clinical trials of HFpEF,22 and their use in the community may similarly expand; as such, our projections are anticipated to generalize to many contemporary patients with HF. Third, the LVEF range between 41% and 45% was not encompassed in eligibility criteria of either the PARADIGM-HF trial or the PARAGON-HF trial. As a conservative estimate, we assumed patients in this LVEF range would have similar treatment responses as those at the lower end of eligibility in the PARAGON-HF trial. Notably, while eligibility was determined based on screening echocardiograms interpreted by local sites, 11% of patients had LVEF less than 45% when examining echocardiograms processed and interpreted in a centralized core laboratory, suggesting this range was still somewhat represented in the PARAGON-HF trial.23
Taken together, these data provide population-level estimates of eligible patients, anticipated treatment effects, and potential for safety and efficacy events with implementation of sacubitril/valsartan under the updated FDA labeling. These estimates may help guide economic analyses and possibly encourage payer coverage of this effective therapeutic option. If prior implementation barriers to sacubitril/valsartan are swiftly overcome, population-level impact on worsening HF events in this high-risk population is certainly possible with a favorable safety/efficacy margin for most individuals. However, challenges based on access, cost, and therapeutic inertia should not be underestimated. In light of the substantial variation in estimates of newly eligible patients with HF encompassed by the expanded label, the use of sacubitril/valsartan in real-world settings should be carefully monitored. These projections further highlight the complexity in defining the normal LVEF range in HF. LVEF measurements are subject to variation across demographic characteristics, hemodynamics and loading conditions, imaging modality, operator, image processing, and interpretation.
The study is subject to important limitations. First, NHANES relies on self-reported data of HF diagnoses, and corroborating clinician assessment or health care utilization records were not available. However, this approach is consistent with methods by the American Heart Association Heart Disease and Stroke Statistics24 to estimate the US HF population. Self-reported HF may also most closely approximate the intended target population of symptomatic HF. Clinicians are less likely to discuss and patients are less likely to be aware of asymptomatic detection of echocardiographic abnormalities. Second, while certain clinical parameters (such as blood pressure and kidney function) were available in NHANES, other important factors (such as severity of HF) were not, introducing uncertainty around the estimates of the eligible population for sacubitril/valsartan. Third, this study was not intended to apply the more selective eligibility criteria used in the PARAGON-HF trial (such as natriuretic peptides and evidence of structural heart disease); we instead examined the population encompassed by FDA-approved labeling, which do not specify or require these parameters. While some of the identified patients may be found to have other entities (such as infiltrative or genetic forms of cardiomyopathy) if they underwent deep phenotyping, such screening and evaluation are infrequent in current clinical practice. It is also reassuring that our estimates of the potentially eligible HF population align with prior applications of these actionable criteria in other well-described cohorts.25 Fourth, mapping LVEF distribution from a hospitalized setting (GWTG-HF registry) to the US HF population at large may not be appropriate or accurate. However, as is common in clinical practice, the GWTG-HF registry captured the most recent LVEF measurements (whether obtained during hospitalization or that were previously recorded prior to hospitalization).26 Furthermore, similar bimodal LVEF distributions have been reported in other US-based cohorts.27,28 Despite variation in settings of care, demographic and clinical characteristics of those self-identified with HF in NHANES are largely comparable with those of participants enrolled in the GWTG-HF registry.29 Fifth, risk reductions from the PARAGON-HF trial may not readily translate to real-world settings, especially if sacubitril/valsartan is implemented differently or incompletely. When evaluating treatment effects on recurrent events, our NNT estimates assumed a constant annual rate of event accrual, but individual patient risk may dynamically vary over time. Mortality was not reduced with sacubitril/valsartan in the PARAGON-HF trial and thus is not expected with population-level implementation. Sixth, we considered treatment heterogeneity based on LVEF and sex, which were the only 2 (of 12) prespecified subgroups that were identified with significant interaction terms in multivariable modeling in the PARAGON-HF trial. However, we acknowledge that individual treatment responses are complex and may be influenced by other factors that were unaccounted for in our trial projections. Importantly, certain racial and ethnic groups were underrepresented in the PARAGON-HF trial, making more precise treatment estimates in these subgroups challenging. Seventh, we projected clinical benefits with broader use of sacubitril/valsartan but did not consider other aspects of its implementation, including costs and adherence. A more complete decision analysis model integrating these factors is needed.
The expanded FDA labeling is positioned to substantially increase the HF population potentially eligible for sacubitril/valsartan by up to 1.8 million individuals. Comprehensive implementation of sacubitril/valsartan in this large, at-risk population has the potential to prevent or postpone up to 180 000 worsening HF events. However, variable clinical interpretation of the exact revised eligibility criteria with respect to abnormal LVEF cutoffs will substantially alter the potential treatment candidate pool and forecasted events potentially prevented.
Accepted for Publication: July 1, 2021.
Published Online: September 15, 2021. doi:10.1001/jamacardio.2021.3651
Corresponding Author: Scott D. Solomon, MD, Division of Cardiovascular Medicine, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115 (firstname.lastname@example.org).
Author Contributions: Drs Vaduganathan and Solomon 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.
Study concept and design: Vaduganathan, Bhatt, Fonarow, Solomon.
Acquisition, analysis, or interpretation of data: Vaduganathan, Claggett, Greene, Aggarwal, McMurray, Fonarow.
Drafting of the manuscript: Vaduganathan, Bhatt.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Vaduganathan, Claggett, Aggarwal.
Obtained funding: Solomon.
Administrative, technical, or material support: Bhatt, Fonarow.
Study supervision: Fonarow, Solomon.
Conflict of Interest Disclosures: Dr Vaduganathan has received research grants from Amgen, AstraZeneca, and Boehringer Ingelheim; personal fees for serving on advisory boards for American Regent, Amgen, AstraZeneca, Baxter International, Bayer, Boehringer Ingelheim, Cytokinetics, Lexicon Pharmaceuticals, Relypsa, and Roche Diagnostics; and personal fees for speaking for Novartis and Roche Diagnostics; and participates on clinical end point committees for studies sponsored by Galmed and Novartis. Dr Claggett has received consulting fees from Amgen, Boehringer Ingelheim, Gilead, MyoKardia, Novartis, AOBiome, and Corvia. Dr Greene has received research support from Amgen, AstraZeneca, Bristol Myers Squibb, Cytokinetics, Merck, Novartis, and Pfizer as well as personal fees for serving on advisory boards for Amgen, AstraZeneca, Bayer, Cytokinetics, Merck, and Vifor and serves as a consultant for Amgen and Merck. Dr Bhatt has received honoraria from Sanofi Pasteur and Verve Therapeutics and is supported by the National Heart, Lung, and Blood Institute T32 postdoctoral training grant T32HL094301. Dr McMurray has received payments to his institution for work on clinical trials, consulting, and other activities from Alnylam, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol Myers Squibb, Cardurion, Cytokinetics, DalCor, GlaxoSmithKline, KBP Biosciences, Novartis, Pfizer, and Theracos, and personal fees from Abbott, Hikma, Ionis, Sun Pharmaceuticals, and Servier. Dr Fonarow has received research funding from the National Institutes of Health and personal fees for serving as a consultant for Abbott, Amgen, AstraZeneca, Bayer, CHF Solutions, Edwards, Janssen, Medtronic, Merck, and Novartis. Dr Solomon has received research grants from Actelion, Alnylam, Amgen, AstraZeneca, Bellerophon, Bayer, Bristol Myers Squibb, Celladon, Cytokinetics, Eidos, Gilead, GlaxoSmithKline, Ionis, Lilly, Mesoblast, MyoKardia, the National Heart, Lung, and Blood Institute, Neurotronik, Novartis, NovoNordisk, Respicardia, Sanofi-Pasteur, and Theracos and personal fees for consulting from Abbott, Action Pharma, Akros, Alnylam, Amgen, Arena, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol Myers Squibb, Cardior, Cardurion, Corvia, Cytokinetics, Daiichi-Sankyo, GlaxoSmithKline, Ironwood, Lilly, Merck, MyoKardia, Novartis, Roche, Takeda, Theracos, Quantum Genetics, Janssen, Cardiac Dimensions, Tenaya, Sanofi-Pasteur, Dinaqor, Tremeau, CellProThera, Moderna, American Regent, and Sarepta. No other disclosures were reported.
Funding/Support: The PARAGON-HF trial was sponsored by Novartis. The Get With The Guidelines–Heart Failure program is provided by the American Heart Association and is sponsored, in part, by Novartis, Boehringer Ingelheim, Eli Lilly Diabetes Alliance, NovoNordisk, Sanofi, AstraZeneca, and Bayer.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: Dr Fonarow is Associate Editor for Health Care Quality and Guidelines of JAMA Cardiology, but he was not involved in any of the decisions regarding review of the manuscript or its acceptance.
Meeting Presentation: Data from this article were virtually presented at the European Society of Cardiology Heart Failure; June 29, 2021.