Classification of Heart Failure Events by Severity: Insights From the VICTORIA Trial

Abstract

Background:

Hospitalization due to heart failure (HFH) is a major source of morbidity, consumes significant economic resources and is a key endpoint in HF clinical trials. HFH events vary in severity and implications, but they are typically considered equivalent when analyzing clinical trial outcomes.

Objectives:

We aimed to evaluate the frequency and severity of HF events, assess treatment effects and describe differences in outcomes by type of HF event in VICTORIA (Vericiguat Global Study in Subjects with Heart Failure with Reduced Ejection Fraction).

Methods:

VICTORIA compared vericiguat with placebo in patients with HF with reduced ejection fraction (< 45%) and a recent worsening HF event. All HFHs were prospectively adjudicated by an independent clinical events committee (CEC) whose members were blinded to treatment assignment. We evaluated the frequency and clinical impact of HF events by severity, categorized by highest intensity of HF treatment (urgent outpatient visit or hospitalization treated with oral diuretics, intravenous diuretics, intravenous vasodilators, intravenous inotropes, or mechanical support) and treatment effect by event categories.

Results:

In VICTORIA, 2948 HF events occurred in 5050 enrolled patients. Overall total CEC HF events for vericiguat vs placebo were 43.9 vs 49.1 events/100 patient-years (P = 0.01). Hospitalization for intravenous diuretics was the most common type of HFH event (54%). HF event types differed markedly in their clinical implications for both in-hospital and post-discharge events. We observed no difference in the distribution of HF events between randomized treatment groups (P = 0.78).

Conclusion:

HF events in large global trials vary significantly in severity and clinical implications, which may have implications for more nuanced trial design and interpretation.

Clinical trial registration:

ClinicalTrials.gov (NCT02861534)

Lay Summary

Heart failure (HF) events vary dramatically in severity and prognosis, but this has not traditionally been assessed in clinical trials. We evaluated the frequency and severity of HF events in a large global heart failure trial (VICTORIA). The most common type of HF event was hospitalization for IV diuretics. In our analysis, HF event types differed markedly in their clinical implications for patients, both during the event and after discharge from the hospital. This research may have important implications for how clinical trials are designed and interpreted.

Heart failure hospitalization (HFH) is a major component of the total morbidity rates resulting from chronic HF and is a significant driver of the overall economic burden of HF.¹ HFH also serves as a key endpoint in HF clinical trials and may be the primary measure of clinical benefit that drives regulatory approval of and use of new therapies. HFH is not only a source of morbidity; it often portends a change in the natural history of chronic HF, including increased risk of subsequent hospitalizations or death.² In clinical practice, there is substantial variation in the clinical severity of HFH, from mild signs and symptoms of congestion requiring brief intensification of diuretics to cardiogenic shock requiring intensive care and mechanical support. Contemporary real-world data confirm significant variability in the severity of and treatment patterns for patients hospitalized for HF.^3,4 Urgent HF visits or events with intensification of therapy but not requiring hospitalization are also frequently included in HF composite endpoints in contemporary clinical trials.⁵ In most clinical trials, these widely varied events are grouped together as “HF events” and are treated as being equivalent, potentially obscuring important differences in the overall burden of HF among patient populations or treatment groups. Additionally, effective therapies may not just prevent events but could also attenuate the severity of events that do occur (as seen with other medical interventions, including the COVID-19 vaccination). The objectives of the current analysis in the Vericiguat Global Study in Subjects with Heart Failure with Reduced Ejection Fraction (VICTORIA) study were to: (1) describe the variability in HF event severity in a large global, randomized controlled trial in patients with heart failure with reduced ejection fraction (HFrEF); (2) evaluate the treatment effects of vericiguat on the severity of HF events in the study; and (3) describe any differences in in-hospital and long-term outcomes based on the type of antecedent HF event.

Methods

The methods and primary results of the VICTORIA trial (NCT02861534) have been published previously.^6,7 In VICTORIA, vericiguat reduced the composite primary outcome of HFH or cardiovascular death in patients with chronic HFrEF (ejection fraction < 45%, New York Heart Association class II–IV) and a recent (within the prior 6 months) worsening HF event. All deaths and potential cardiovascular events identified in the conduct of the VICTORIA study were centrally adjudicated by an independent clinical events committee (CEC), whose members were blinding to study treatment, using standardized, previously published definitions.⁸ During the adjudication process of potential HF events, reviewers recorded the specific therapies that were used in the treatment of that event. The trial protocol was approved by ethics committees at participating sites, and all patients provided written informed consent.

For the current analyses, we classified all positively adjudicated HF events (which included HFH and urgent HF visits requiring intravenous [IV] therapy) into 6 mutually exclusive categories based on the greatest intensity of HF treatment required at any time during the event, as identified by the CEC: (1) urgent HF visits (treatment with IV diuretics but without hospitalization); (2) HFH treated with intensification of oral diuretics only; (3) HFH treated with IV diuretics only; (4) HFH treated with IV vasodilators; (5) HFH treated with IV inotropic therapy; and (6) HFH treated with mechanical circulatory support or fluid removal. Events were categorized based on the most intensive therapy the patient received during the HF event. For each of these event types, we describe the outcomes of length of stay, in-hospital mortality rates, 30-day post-discharge outcomes, and long-term mortality rates for the duration of the study’s follow-up. For long-term mortality rates, we further subdivided deaths by adjudicating cause of death into 4 discrete categories: HF death, sudden death, other causes of cardiovascular death, and noncardiovascular death.

Event rates were calculated as the number of events per 100 patient-years of follow-up. Hazard ratios and corresponding P values were determined by using stratified Andersen-Gill models with robust standard errors to account for potential correlation among patients who had recurrent events and stratified 1-sided log-rank tests (α = 0.025), with strata defined by region, race (within North America) and nature of qualifying index event (prior HF hospitalization < 3 months, prior HF within 3–6 months, or outpatient/emergency department visit within 3 months). The treatment effect across intensities of HFH events was evaluated by the Mann-Whitney U test. Comparisons of lengths of stay were performed using the Moode median test, and all other outcomes were compared using ² tests. For the analysis of subsequent mortality after each type of HF event, we limited the analysis to the first HF event experienced by each patient in order to mitigate the confounding effect of intercurrent HF events.

Results

The VICTORIA study enrolled 5050 eligible patients with HFrEF in 42 countries and followed them for a median duration of 10.8 months. The baseline characteristics of the study population have been reported in detail previously.⁹ The median age was 69 years, 76% of patients were male, and 64% were white. By protocol, all patients had had a worsening HF event requiring IV diuretics within the prior 6 months. The primary analysis of the study showed a significant improvement in the primary endpoint for patients randomized to vericiguat compared with placebo (hazard ratio [HR] 0.90, 95% confidence interval [CI] 0.82–0.98; P = 0.02).⁶ Compared with other contemporary HFrEF trials, the VICTORIA study enrolled a population at uniquely high risk for HF events.¹⁰

A total of 2948 adjudicated HF events occurred in 1525 of the 5050 patients enrolled (30% of patients experienced at least 1 HF event). Using the framework for event severity described above, 389 (13.2%) were urgent HF visits not requiring hospitalization, 68 (2.3%) were HFH treated with oral diuretics, 1612 (54.7%) were treated with only IV diuretics, 157 (5.3%) were treated with IV vasodilators, 522 (17.7%) were treated with IV inotropic agents, and 196 (6.6%) were treated with mechanical support. Event rates by event type and associated treatment effects of vericiguat compared with placebo for each type of event are shown in Table 1. Forest plots with 95% confidence intervals for each event type are shown in the Central Illustration. The prevalence of any HF event was significantly lower in patients randomized to vericiguat compared with placebo (43.9 HF events/100 patient-years for vericiguat vs 49.1 HF events/100 patient-years for placebo; P = 0.01). We did not observe a difference in the distribution of HF events between randomized treatment groups (P = 0.78) (Fig. 1).

Table 1.

Incidence of HF hospitalization by severity and treatment effect of vericiguat

Outcome	Vericiguat (n = 2526)		Placebo (n = 2524)		HR (95% CI)*	P Value*
	No. events (%) no. pts	Events/100 pt-yrs	No. events (%) no. pts	Events/100 pt-yrs
Total CEC HF events	1402 (28.9)730	43.94	1546 (31.5) 795	49.06	0.90 (0.83–0.98)	0.010
Urgent HF visit	179 (4.5) 114	5.61	210 (5.0) 126	6.66	0.85 (0.69–1.04)	0.107
Total hospitalizations for HF	1223 (27.4)691	38.33	1336 (29.6) 747	42.40	0.91 (0.84–0.99)	0.025
HFH with oral intensification of diuretics only	38 (1.3) 33	1.19	30 (1.1)29	0.95	1.26 (0.78–2.03)	0.348
HFH with IV diuretics only	766 (19.4) 490	24.01	846 (21.1) 533	26.85	0.90 (0.81–0.99)	0.037
HFH with IV vasodilators	72 (2.5) 63	2.26	85 (2.7) 69	2.70	0.83 (0.61–1.14)	0.257
HFH with IV inotropes	242 (7.0) 177	7.58	280 (8.6)216	8.89	0.86 (0.73–1.03)	0.100
HFH with mechanical circulatory support or fluid removal	102 (3.6)91	3.20	94 (2.8)71	2.98	1.09 (0.82–1.44)	0.552

^*Hazard ratio (95% CI) and P value computed from a stratified Andersen-Gill model.CEC, clinical events committee; CI, confidence interval; HF, heart failure; HFH, heart failure hospitalization; HR, hazard ratio; IV, intravenous; pt, patients; pt-yrs, patient years.

Fig. 1.

(Visual Take Home Graphic). Forest plot of hazard ratios comparing vericiguat with placebo for each type of heart failure event. Rates are per 100 patient-years. Error bars represent 95% confidence intervals.

Data for in-hospital and 30-day post-discharge outcomes are summarized in Table 2. In-hospital outcomes (length of stay and in-hospital mortality) by HF severity subtype are shown in Fig. 2. Both in-hospital mortality rates and lengths of stay were numerically worse in patients with more severe HF subtypes, especially for those requiring inotropic therapy or mechanical cardiac support. Generally, more severe events were associated with higher post-discharge event rates, with the exception of patients requiring IV vasodilators, who had relatively low rates of 30-day cardiovascular mortality (0.6%) and the composite of cardiovascular death or HFH at 30 days (15%).

Table 2.

Clinical outcomes after HF event stratified by severity

Outcome	Overall	Length of Stay (days)				In-Hospital Mortality (%)				30-Day CV Mortality (%)			30-Day Recurrent HFH or CV Mortality (%)
		Vericiguat	Placebo	P Value	Overall	Vericiguat	Placebo	P Value	Overall	Vericiguat	Placebo	P Value	Overall	Vericiguat	Placebo	P Value
Total HF events		-	-			-	-		120 (4.1) 2948	63 (4.5) 1402	57 (3.7) 1546	0.268	627 (21.3) 2948	303 (21.6) 1402	324 (21.0) 1546	0.664
Urgent HF visit	-	-	-		0 (0.0) 389	-	-		14 (3.6) 389	7 (3.9) 179	7 (3.3) 210	0.761	14 (3.6) 389	7 (3.9) 179	7 (3.3) 210	0.761
HF hospitalization	8.5 (5.0, 14.0)	9.0 (5.0, 14.0)	8.0 (5.0, 14.0)	0.527	183 (7.2) 2559	79 (6.5) 1223	104 (7.8) 1336	0.194	106 (4.1) 2559	56 (4.6) 1223	50 (3.7) 1336	0.289	613 (24.0) 2559	296 (24.2) 1223	317 (23.7) 1336	0.778
HFH with oral intensification of diuretics only	7.0(4.0, 11.0)	6.0(4.0, 11.0)	8.5 (6.0, 11.0)	0.134	1 (1.5) 68	1 (2.6) 38	0 (0.0) 30	0.559	0 (0.0) 68	0 (0.0) 38	0 (0.0) 30		11 (16.2) 68	7 (18.4) 38	4 (13.3) 30	0.226
HFH with IV diuretics only	7.0(5.0, 12.0)	7.5 (5.0, 12.0)	7.0 (4.0, 12.0)	0.569	58 (3.6) 1612	24 (3.1) 766	34 (4.0) 846	0.340	66 (4.1) 1612	31 (4.0) 766	35 (4.1) 846	0.927	378 (23.4) 1612	180 (23.5) 766	198 (23.4) 846	0.964
HFH with IV vasodilators	9.0(5.0, 15.0)	9.0 (5.0, 14.5)	8.0 (5.0, 15.0)	0.259	4 (2.5) 157	1 (1.4) 72	3 (3.5) 85	0.292	1 (0.6) 157	1 (1.4) 72	0 (0.0) 85	0.459	23 (14.6) 157	9 (12.5) 72	14 (16.5) 85	0.483
HFH with IV inotropes	11.0 (8.0, 17.0)	11.0 (7.0, 15.0)	11.5 (8.0, 18.5)	0.678	74 (14.2) 522	31 (12.8) 242	43 (15.4) 280	0.405	32 (6.1) 522	20 (8.3) 242	12 (4.3) 280	0.059	145 (27.8) 522	73 (30.2) 242	72 (25.7) 280	0.258
HFH with mechanical circulatory support or mechanical fluid removal	15.5 (8.0, 29.0)	14.5 (8.0, 25.0)	16.0 (9.0,31.0)	0.392	46 (23.5) 196	22 (21.6) 102	24 (25.5) 94	0.513	6 (3.1) 196	3 (2.9) 102	3 (3.2) 94	0.316	54 (27.6) 196	26 (25.5) 102	28 (29.8) 94	0.501

CV, cardiovascular; HF, heart failure; HFH, heart failure hospitalization; IV, intravenous.

Fig. 2.

In-patient outcomes for length of stay and in-patient mortality by HF event severity. Median length of stay (green line) and in-patient mortality (blue bars) for each subtype of heart failure event.

For long-term mortality, the rates of death occurring after each type of first HF event are shown in Fig. 3, stratified into categories by adjudicated cause of death. For all patients experiencing an HF event who subsequently died, HF was the most common cause of death, regardless of the initial HF event type. However, the event rate of HF death varied more than 4-fold based on the severity of the preceding HF event. Among those patients who died after an initial HF event, the disposition of subsequent HF deaths fell broadly into 3 categories. Patients whose HF events were treated with vasodilators had low risks of subsequent HF death (12.6 events per 100 patient-years), whereas HF events treated with diuretics alone had intermediate risk, regardless of diuretic treatment type (HF death rate from 19.7 events per 100 patient-years for HF events treated only with IV diuretics and 21.5 events per 100 patient-years for those treated only with per os [PO] diuretics). HF events suggestive of low output had very high rates of subsequent HF death (35.8 events per 100 patient-years for inotropes to 61.5 events per 100 patient-years for mechanical support). Other causes of death, including sudden death, were generally similar across all HF event types.

Fig. 3.

Mortality rates after each type of first heart failure (HF) event. Long-term mortality after each type of first HF event showing all-cause mortality and grouped by adjudicated cause of death (HF death, sudden death, other causes of cardiovascular death, and noncardiovascular death). Vertical bars represent 95% confidence intervals.

Discussion

Using data from a large global randomized clinical trial in patients with HFrEF, we found that HFH severity varied substantially across types of adjudicated HF events. As anticipated, HF events requiring greater intensity of treatment were generally associated with longer lengths of stay, greater in-hospital mortality rates and higher rates of post-discharge morbidity and mortality. The treatment effect of vericiguat in reducing HF events was generally consistent across all event types. The most common subtype of HFH was treated only with IV diuretics (54% of all HF events), which is similar to prior reports of observational data from the United States.³ Of the various HF severity classifications, patients treated with IV vasodilators were notably associated with relatively lower risk of both in-hospital and post-discharge events. This finding has been shown previously in observational data such as the ALARM-HF (AHF global survey of standard treatment) registry.¹¹ Although it is possible that this is related to the favorable therapeutic effects of vasodilators, it seems more likely that patients selected for vasodilator treatment, who tend to have higher blood pressures at presentation, are a lower risk group, given the known association between higher systolic blood pressure and lower risk in the setting of HFH.¹²

Despite the importance of HFH, relatively little rigorous evaluation of the clinical impact of the severity of different types of HFH have been published. Parikh et al. reported data on the in-hospital and post-discharge trajectories of patients with HF from the American Heart Association Get With The Guidelines registry of presentation severity by using the ADHERE (Acute Decompensated Heart Failure National Registry) risk score, and they demonstrated that patients with higher predicted risk had higher in-patient mortality rates, higher use of cardiovascular procedures and more long-term rehospitalizations and deaths.¹³ The need to escalate treatment during HF hospitalization due to worsening signs and symptoms, usually termed in-patient worsening HF, has been described in multiple datasets and has been used as an endpoint in some clinical trials.¹⁴

Data concerning cause of death have been published from prior clinical trials of HF,^15–17 but they have not been previously described on the basis of severity of a preceding HF event. In our analysis, HF was the predominant cause of death for all patients who died subsequent to an HF event, compared with patients without an HF event, in which death due to HF was rare. Rates of HF death after HF events were broadly similar for diuretic-based HF events, whereas HF events treated with vasodilators were associated with lower risk, and HF events suggestive of low-output states carried the highest risk. Of note, the subsequent rate of HF death was similar for HF events in those treated with IV diuretics as outpatients (ie, urgent HF events), in those treated in-hospital with only PO diuretics, and for those treated in-hospital with IV diuretics, suggesting that these differences in therapy may be more likely to be related to local practice patterns than to the severity of underlying HF.

Limitations

Our study has both limitations and strengths. We examined data from a global randomized clinical trial, and patients enrolled in clinical trials may differ in important ways from the broader HF population. VICTORIA enrolled patients with reduced ejection fraction (< 45%), so these data may not apply to patients with HF and preserved ejection fraction. Although there were a substantial number of HF events in the VICTORIA study, the number of events in some categories was relatively small, limiting the power to make meaningful comparisons. The strengths of our study include the prospective systematic assessment of HF events by blinded adjudicators, more than a 99% complete follow-up, the large number of both HFH and cardiovascular deaths, and the global nature of our study.

Conclusions

HF events in large global trials vary significantly in severity and clinical implications. The treatment benefit of vericiguat compared with placebo was generally similar across various HF event types. These data may have implications for more nuanced interpretation of HF clinical trial data, such as using a hierarchical approach to HF events rather than considering them all to be equivalent when evaluating the efficacy of HF therapeutics.

Bullet Pointswfi.

• Heart failure events may vary dramatically in severity, but this has not typically been captured in clinical trials.

• Differing severities of heart failure events may have profoundly different implications for patients’ outcomes.

• We found that HF events of differing severities had very different outcomes, although the treatment effects of vericiguat were generally similar across event types.

Abbreviations:

CEC

clinical events committee

CI

confidence interval

EF

ejection fraction

HF

heart failure

HFH

heart failure hospitalization

HFrEF

heart failure with reduced ejection fraction

HR

hazard ratio

IV

intravenous

VICTORIA

Vericiguat Global Study in Subjects with Heart Failure with Reduced Ejection Fraction

Biography