Black lines represent patients randomized within 3 months of HF hospitalization, orange lines those randomized within 3 to 6 months of HF hospitalization, and blue lines those randomized with 3 months of intravenous (IV) diuretic use for heart failure. Reported hazard ratios are populated from unadjusted models. HRH indicates heart failure hospitalization.
A, Incidence rates (and their 95% confidence intervals) of the primary composite end point are shown according to finer cuts of time from index event. Data on days to index event were missing in 16 patients. B, The treatment effect of vericiguat (hazard ratio [HR] with 95% confidence interval) is shown as a function of time as a continuous variable from index event to randomization among hospitalized patients. Data are presented for 4219 patients randomized within 200 days.
eTable 1. Candidate variable list
eTable 2. Hazard ratios for primary and secondary endpoints by index admission subgroups
eTable 3. Treatment effect by index admission event subgroup
eTable 4. Safety events of interest and serious adverse events by index event group
eFigure 1. Kaplan-Meier analysis
eFigure 2. Treatment effect of vericiguat compared with placebo analyzed using time from intravenous diuretic administration for outpatient worsening as a continuous variable
Nonauthor Collaborators. The VICTORIA Study Group.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Lam CSP, Giczewska A, Sliwa K, et al. Clinical Outcomes and Response to Vericiguat According to Index Heart Failure Event: Insights From the VICTORIA Trial. JAMA Cardiol. 2021;6(6):706–712. doi:10.1001/jamacardio.2020.6455
How do outcomes in the Vericiguat Global Study in Subjects With Heart Failure With Reduced Ejection Fraction trial compare across worsening heart failure (HF) index admission characteristics, and is there a time-dependent risk of events according to index hospitalization subgroups?
In this secondary analysis of a randomized clinical trial, there was a risk gradient for the primary composite end point; highest in patients closest in time from index HF hospitalization and lowest in those with only outpatient worsening. This risk was reduced by vericiguat in all subgroups, without evidence of treatment heterogeneity.
Among patients with recent worsening HF, proximity to HF hospitalization was associated with increased risk of adverse outcomes, whereas the risk reduction with vericiguat did not differ significantly across subgroups.
The period following heart failure hospitalization (HFH) is a vulnerable time with high rates of death or recurrent HFH.
To evaluate clinical characteristics, outcomes, and treatment response to vericiguat according to prespecified index event subgroups and time from index HFH in the Vericiguat Global Study in Subjects With Heart Failure With Reduced Ejection Fraction (VICTORIA) trial.
Design, Setting, and Participants
Analysis of an international, randomized, placebo-controlled trial. All VICTORIA patients had recent (<6 months) worsening HF (ejection fraction <45%). Index event subgroups were less than 3 months after HFH (n = 3378), 3 to 6 months after HFH (n = 871), and those requiring outpatient intravenous diuretic therapy only for worsening HF (without HFH) in the previous 3 months (n = 801). Data were analyzed between May 2, 2020, and May 9, 2020.
Vericiguat titrated to 10 mg daily vs placebo.
Main Outcomes and Measures
The primary outcome was time to a composite of HFH or cardiovascular death; secondary outcomes were time to HFH, cardiovascular death, a composite of all-cause mortality or HFH, all-cause death, and total HFH.
Among 5050 patients in the VICTORIA trial, mean age was 67 years, 24% were women, 64% were White, 22% were Asian, and 5% were Black. Baseline characteristics were balanced between treatment arms within each subgroup. Over a median follow-up of 10.8 months, the primary event rates were 40.9, 29.6, and 23.4 events per 100 patient-years in the HFH at less than 3 months, HFH 3 to 6 months, and outpatient worsening subgroups, respectively. Compared with the outpatient worsening subgroup, the multivariable-adjusted relative risk of the primary outcome was higher in HFH less than 3 months (adjusted hazard ratio, 1.48; 95% CI, 1.27-1.73), with a time-dependent gradient of risk demonstrating that patients closest to their index HFH had the highest risk. Vericiguat was associated with reduced risk of the primary outcome overall and in all subgroups, without evidence of treatment heterogeneity. Similar results were evident for all-cause death and HFH. Addtionally, a continuous association between time from HFH and vericiguat treatment showed a trend toward greater benefit with longer duration since HFH. Safety events (symptomatic hypotension and syncope) were infrequent in all subgroups, with no difference between treatment arms.
Conclusions and Relevance
Among patients with worsening chronic HF, those in closest proximity to their index HFH had the highest risk of cardiovascular death or HFH, irrespective of age or clinical risk factors. The benefit of vericiguat did not differ significantly across the spectrum of risk in worsening HF.
ClinicalTrials.gov Identifier: NCT02861534
Despite available heart failure (HF) therapy, patients with reduced ejection fraction (HFrEF) remain at high risk of hospitalization and death. Previous chronic HFrEF trials showed that prior HF hospitalization (HFH) strongly predicted cardiovascular (CV) death and HFH, with risk declining with time from the index admission event.1,2 The Vericiguat Global Study in Subjects With Heart Failure With Reduced Ejection Fraction (VICTORIA)3 is, to our knowledge, the first positive large trial to exclusively enroll patients with worsening HFrEF in the vulnerable period (<6 months) soon after HFH, thereby offering unique opportunities to study the spectrum of patients recruited during and immediately following index HFH or an outpatient worsening HF event. Whereas the baseline characteristics and main unadjusted primary outcomes according to prespecified index event subgroups have been reported,3,4 we now provide new data informed by treatment assignment and the effect of multivariable-adjusted analyses of primary and secondary outcomes. We also assessed outcomes of patients randomized while in hospital, as well as immediately following hospitalization, to provide finer continuous detail on the association between time from index event and outcomes.
Briefly,5 the VICTORIA trial showed that vericiguat reduced the composite primary outcome of HFH or CV death in patients with chronic HFrEF (EF <45%, New York Heart Association [NYHA] class II-IV) and a recent (<6 months) worsening HF event. Prespecified subgroups were less than 3 months or 3 to 6 months after HFH, and those with worsening HF receiving outpatient intravenous diuretics at less than 3 months (without HFH in the prior 6 months). Secondary end points were time to HFH, CV death, a composite of all-cause mortality or HFH, all-cause death, and total HFH. The trial protocol was approved by ethics committees at participating sites, and all patients provided written informed consent.
To compare outcomes among index event subgroups, we used Cox proportional hazards regression models, where multivariable adjustment variables were independent predictors for each outcome selected from a set of 66 available candidate variables (eTable 1 in Supplement 1).6 Interaction between treatment and index event subgroup was tested for each outcome. Patients with index HFH were further divided into in-hospital randomization, 30 days or less, 31 to 90 days, and greater than 90 days. Incident rates were calculated for every period and treatment group. Hazard ratios (HRs) for treatment effect were calculated for every period, along with interaction P value between treatment and finer time subgroup. Significance was determined at a 2-sided P value of less than .05. A trend of the HR of the primary end point for the treatment effect is presented over time from the date of the index hospitalization to randomization date, where time was a continuous variable.
Of a total of 5050 patients in the VICTORIA trial, 3378 (67%) were randomized less than 3 months from index HFH (11% in-hospital), 871 (17%) within 3 to 6 months of HFH, and 801 (16%) within 3 months of outpatient worsening HF (median 26 days; interquartile range, 14-46 days). The baseline characteristics were well balanced between treatment arms within each index event subgroup (Table).
We observed a gradient of risk across the index event subgroups (eFigure 1 and eTable 2 in Supplement 1), with highest risk of the primary composite end point in HFH less than 3 months (40.9 events per 100 patient-years), followed by HFH 3 to 6 months (29.6 events per 100 patient-years), and lowest risk in the outpatient worsening subgroup (23.4 events per 100 patient-years). Compared with the outpatient worsening subgroup, the adjusted risk of CV death or HFH was higher in the HFH less than 3 months subgroup (adjusted HR, 1.48; 95% CI, 1.27-1.73), but no longer significantly different in the HFH 3 to 6 months subgroup (adjusted P = .25). Similar trends were observed for the secondary outcomes (eTable 2 in Supplement 1). There was no statistical difference in vericiguat’s effect on primary or secondary outcomes across index event subgroups (eTable 3 in Supplement 1; Figure 1).
When patients with an index HFH were divided into those in hospital (n = 571) and in 30-day intervals (≤30 [n = 1461]; >30-60 [n = 827]; >60-90 [n = 516]; >90-120 [n = 371]; >120-150 [n = 234]; and >150 days [n = 253]) following index hospitalization, a wider gradient of risk was observed that was highest in hospitalized patients (event rate 52.0 per 100 patient-years) (Figure 2A). The primary composite event rate decreased from 40.5 events per 100 patient-years in those randomized less than 30 days, to 41.0, 33.0, 30.9, 29.0, and 27.2 events per 100 patient-years in those randomized 31 to 60, 61 to 90, 91 to 120, 121 to 150, and more than 150 days following index hospitalization, respectively. The relative risk reduction with vericiguat vs placebo was not significantly different among these more finely divided subgroups (interaction P = .35) (Figure 2A). The continuous association between time since HFH and treatment effect of vericiguat showed a trend (interaction P = .09) toward greater benefit with longer duration since HFH (Figure 2B). A trend of the HR of the primary endpoint for the treatment effect is presented over time from the date of the index hospitalization to randomization date, where time was a continuous variable. When the continuous association between time since intravenous diuretic for outpatient worsening and the treatment effect of vericiguat was explored over the 60-day window of observation, there was no association to treatment benefit (eFigure 2 in Supplement 1).
Safety events were infrequent in all subgroups (eTable 4 in Supplement 1), without difference between treatment arms (symptomatic hypotension range 6.2%-10.3% and syncope range 2.9%-5.8% for subgroups). Specifically, for patients randomized in-hospital, there were 24 (8.1%) cases of symptomatic hypotension with vericiguat vs 17 (6.2%) for placebo, and 17 (5.8%) syncopal episodes with vericiguat arm vs 11 (4.0%) with placebo. These adverse event rates were similar or lower than observed in the respective treatment arms of the outpatient worsening subgroup.
Among patients with worsening chronic HF in the VICTORIA trial, those randomized within a 3-month window of HFH experienced an approximately 50% higher rate of CV death or HFH than those with outpatient worsening without HFH, even after adjusting for prognostically relevant clinical covariates, background therapy, and laboratory parameters (including N-terminal pro hormone B-type natriuretic peptide). This risk was further accentuated in those within 1 month of HFH (>40 events per 100 patient-years) and again amplified among patients randomized within their index hospitalization (>50 events per 100 patient-years). Notwithstanding this spectrum of risk, there was no statistical evidence for heterogeneity of treatment benefit with vericiguat across the range of times from index event to 6 months.
Our results are aligned with prior chronic HFrEF trials reporting increasing risk with proximity to a prior HFH event,1,2 but extend those observations to a higher risk population studied both closer to a prior HFH or outpatient worsening without HFH. The primary event rate in the HFH less than 3 months group in VICTORIA (40.9 events per 100 patient-years) was approximately 2 times higher than the corresponding rate in the highest risk group of Candesartan in Heart Failure: Assessment of Mortality and Morbidity (CHARM; those randomized 0-2 months from HFH) and Prospective Comparison of ARNI with ACEI to Determine Impact on Global Mortality and Morbidity in Heart Failure (PARADIGM; also randomized <3 months from HFH). Furthermore, the primary event rate in the VICTORIA outpatient worsening subgroup (23.4 events per 100 patient-years) was more than 2 times higher than the corresponding rates in the lowest risk group of CHARM (those without prior HFH), and more similar to the event rates in the highest-risk CHARM group (those randomized 0-2 months following HFH).1 The additional risk in the VICTORIA subgroups likely reflect the additive effects of the higher natriuretic peptide entry criteria and baseline levels as compared with prior trials.7,8 These results emphasize that patients with outpatient worsening HF (especially with elevated natriuretic peptides) are at significant risk of HFH and CV death comparable with that of HFH9,10 and should be strongly considered for additional disease-modifying therapies. The benefit of vericiguat, added to standard guidelines-directed medical therapy, is particularly noteworthy in this setting, where a numerically greatest relative risk reduction with vericiguat was evident.
We observed a trend toward greater treatment benefit with increasing time from HFH. Whether this would extend beyond the vulnerable 6 months after HFH to more “stable” HFrEF outpatients warrants further study. Whereas vericiguat treatment was not associated with more adverse events than placebo among in-hospital patients their very high risk of recurrent HFH and CV death may not be modifiable as suggested in prior trials that enrolled in-hospital patients with worsening HF.11,12
Our study has both strengths and limitations. Strengths include the unique VICTORIA patient population, enriched for risk by the dual requirements of proximity to the HF index event as well as high natriuretic peptides, while also exploring a risk spectrum from in-hospital randomization to outpatient worsening HF. Nonetheless, VICTORIA was not powered for these exploratory post hoc analyses in smaller subgroups. Underrepresentation of women and minority ethnic groups was also evident.
Our results have implications for future HF trial design, suggesting that high-entry natriuretic peptides and proximity to a worsening HF event select for patients particularly prone to recurrent HFH out of proportion to CV death, compared with HF trials in lower risk HF populations. Furthermore, clinical events accrue very quickly over time in these patients, shortening the follow-up duration in an event-driven trial powered for the composite outcome. Finally, our findings highlight the residual risk and need for novel therapies in patients with recent worsening HF.
Corresponding Author: Paul W. Armstrong, MD, 4-120 Katz Group Centre for Pharmacy and Health Research, University of Alberta, Edmonton, AB T6G 2E1, Canada (firstname.lastname@example.org).
Accepted for Publication: October 26, 2020.
Published Online: November 13, 2020. doi:10.1001/jamacardio.2020.6455
Correction: This article was corrected on February 17, 2021, to correct the list of the VICTORIA Study Group members; on March 31, 2021, to resupply the list of nonauthor collaborators in a Supplement; and on October 6, 2021, to correct an author affiliation.
Author Contributions: Drs Lam and Armstrong 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: Lam, Edelmann, Ezekowitz, Hernandez, O'Connor, Roessig, Patel, Pieske, Anstrom, Armstrong.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Lam, Sliwa-Hahnle, Refsgaard, Bocchi, Ezekowitz, Roessig, Armstrong.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Giczewska, Refsgaard, Anstrom.
Obtained funding: Hernandez, Armstrong.
Administrative, technical, or material support: Sliwa-Hahnle, Edelmann, Hernandez, Patel.
Supervision: Sliwa-Hahnle, Refsgaard, O'Connor, Patel, Armstrong.
Conflict of Interest Disclosures: Dr Lam reported personal fees from Merck Steering Committee of VICTORIA and Bayer Steering Committee of VICTORIA during the conduct of the study; grants from National Medical Research Council of Singapore Supported by a Clinician Scientist Award, Boston Scientific Investigator-sponsored research program (Asian Sudden Cardiac Death in Heart Failure Registry), Medical Advisory Board, Bayer investigator-led research (Asian Sudden Cardiac Death in Heart Failure Registry), Roche Diagnostic, grants from Medtronics, Vifor Pharma, AstraZeneca; and personal fees from Novartis, Amgen, Merck, Janssen Research and Development LLC, Menarini, Boehringer Ingelheim, Abbott Diagnostics, Corvia, Stealth BioTherapeutics, Novo Nordisk, JanaCare, Biofourmis, Darma, Applied Therapeutics, MyoKardia Advisory board, Cytokinetics, WebMD, Radcliffe Group Ltd, and Corpus utside the submitted work; in addition, Dr Lam had a patent for PCT/SG2016/050217 and for 16/216,929 pending and is and cofounder and nonexecutive director of Us2.ai. Dr Giczewska reported other support from Merck during the conduct of the study. Dr Sliwa-Hahnle reported grants from Merck during the conduct of the study. Dr Edelmann reported personal fees from Bayer, Merck, Novartis, Vifor, Pharmacosmos, Berlin Chemie, AstraZeneca, Boehringer Ingelheim, and Servier and grants from Thermo Fischer and Servier during the conduct of the study; personal fees from Bayer, Merck, Novartis, Vifor, Pharmacosmos, Berlin Chemie, AstraZeneca, Boehringer Ingelheim, Servier, and Thermo Fischer outside the submitted work; and grants from Servier outside the submitted work. Dr Bocchi reported personal fees from Bayer/Merck during the conduct of the study; personal fees from Servier, AstraZeneca, Novartis, Jansen, and Boehringer Ingelheim; and grants from Baldacci outside the submitted work. Dr Ezekowitz reported personal fees from Bayer and Merck during the conduct of the study and from Amgen, AstraZeneca, Boeringher-Ingelheim, Servier, and Novartis outside the submitted work. Dr Hernandez reported grants from Merck and Bayer during the conduct of the study; personal fees from AstraZeneca, Amgen, Bayer, Cytokinetics, Myokardia, Novartis, and Relypsa; and grants from Verily outside the submitted work. Dr O'Connor reported serving as a consultant for Bayer, Bristol Myers Squibb Foundation, Dey, and LP Merck. Dr Roessig reported working as a full-time employee for Bayer AG during the conduct of the study and outside the submitted work. Dr Patel reported working as full-time employee at the Merck Research Laboratories. Dr Pieske reported personal fees from Merck, Bayer, Novartis, BMS, Servier, and AstraZeneca during the conduct of the study. Dr Anstrom reported grants from Merck during the conduct of the study. Dr Armstrong reported grants and personal fees from Merck and Bayer during the conduct of the study; grants from Sanofi-Aventis, Boehringer Ingelheim, and CSL Limited; and personal fees from AstraZeneca and Novartis outside the submitted work. No other disclosures were reported.
Funding/Support: The VICTORIA trial was funded by Merck Sharp and Dohme Corp, a subsidiary of Merck and Co Inc, Kenilworth, New Jersey, and Bayer AG, Wuppertal, Germany.
Role of the Funder/Sponsor: The sponsor was aware of the design and conduct of the study, but was not involved in the collection, management, analysis, and interpretation of the data. Representatives from the sponsor were not involved in the preparation and approval of the manuscript but did have the opportunity to review. The sponsor played no role in the decision to submit the manuscript for publication.
Group Information: The VICTORIA Study Group members can be found in Supplement 2.
Meeting Presentation: This paper was presented at the American Heart Association Scientific Sessions 2020; November 13, 2020; virtual conference.
Additional Contributions: We thank Elizabeth Cook, BA, of the Duke Clinical Research Institute for her editorial assistance and preparation of the manuscript submission.
Additional Information: The data will be made available as outlined in the VICTORIA data sharing charter that is on the CVC website: https://thecvc.ca/wp-content/uploads/2017/01/VICTORIA-Data-Sharing-Charter-v1.0-17Jan2019-FINAL.pdf.