Registry Evaluation of Vital Information for VADs in Ambulatory Life (REVIVAL): Rationale, design, baseline characteristics, and inclusion criteria performance

Abstract

INTRODUCTION:

Improved understanding of the clinical course of ambulatory advanced chronic systolic heart failure may improve the provision of appropriate care and is central to the design of clinical trials in this population.

METHODS:

Twenty-one implanting ventricular assist device (VAD) centers enrolled 400 subjects in the Registry Evaluation of Vital Information for VADs in Ambulatory Life (REVIVAL), a prospective, observational study in ambulatory, chronic, advanced systolic heart failure, designed to identify a cohort with an approximately 25% 1-year risk of the primary composite outcome of death, urgent transplant, or durable mechanical circulatory support. Inclusion criteria utilized only information collected during routine clinical care. Exclusion criteria identified patients with contraindications to VAD. Study inclusion required at least 1 of 10 high-risk criteria derived from established hospitalization and non-hospitalization markers of increased mortality risk. We evaluated the test performance characteristics of the high-risk criteria.

RESULTS:

Data on 373 subjects evaluable for the primary composite outcome at the 1-year visit are presented. Baseline data were consistent with a less advanced cohort than Medical Arm for Mechanically Assisted Circulatory Support or Risk Assessment (MedaMACS) and Comparative Effectiveness of Left Ventricular Assist Device and Medical Management in Ambulatory Heart Failure Patients (ROADMAP). Freedom from the primary composite outcome was 75.9%. Non-hospitalization inclusion criteria identified 89% of patients with events.

CONCLUSIONS:

Using routinely obtained clinical information for enrollment, REVIVAL successfully recruited an ambulatory chronic systolic heart failure cohort with an approximately 25% annual risk of the primary composite outcome. Information from this registry will be relevant to the planning of future trials of earlier VAD use and of other interventions in this population.

Despite widespread use of evidence-based medical therapies, including neurohormonal blockers and biventricular pacing, mortality and morbidity from systolic heart failure (HF) remain high.¹ Breakthroughs in mechanical circulatory support (MCS) technology have extended survival and improved quality of life in patients with the most advanced HF awaiting cardiac transplantation and in inotrope-dependent patients who are ineligible for transplant.² Left ventricular assist device (LVAD) therapy offers the promise of relieving HF symptoms in patients with less advanced stages of HF but comes with its own significant adverse effect burden.³

Analyses of ventricular assist device (VAD) clinical trial databases and of the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) have greatly enhanced our knowledge of those who are most likely to have a favorable outcome with VAD therapy.³ However, present prognostic models have been limited by both the quality of the prognostic data and the narrowness of the outcome data (e.g., survival outcome only rather than more comprehensive measures like quality-adjusted life years). The prognostic value of registry data is often limited by the restricted scope of the prognostic variables collected, poor measurement quality, and missing data. Additionally, a better understanding of the factors that might drive a patient’s choice to receive an early VAD or continue with standard medical therapy is currently lacking, yet necessary to inform shared-decision making.

The Registry Evaluation of Vital Information for VADs in Ambulatory Life (REVIVAL) addresses these concerns by establishing a prospective, observational, multicenter patient cohort of ambulatory patients with chronic, advanced, systolic HF that will provide a greater understanding of their clinical trajectory (rates of hospitalizations, stroke, transplantation, MCS use, death, temporal course of their functional status, frailty measures, and quality of life) and of how baseline clinical risk measures are related to prognosis. Through inclusion of the patient perspective, REVIVAL will investigate patients’ self-estimates of prognosis and preferences for VAD therapy, as well as the impact of advanced HF care on caregiver quality of life and caregiver burden, both of which are critical to wellinformed shared decision making.

REVIVAL was designed to achieve the following specific aims: (1) To characterize clinical outcomes, health-related quality of life (HRQOL), and functional impairment over 2 years in a population of ambulatory patients on evidence-based therapy with advanced chronic systolic HF who may benefit from VAD therapy; (2) To evaluate the relationship between the HF subject’s modeled prognosis, self-assessed prognosis, preferences for end of life care, and thresholds for considering VAD implant; (3) To evaluate caregiver burden associated with the HF subject’s measures of HF severity, HRQOL, functional limitations, preferences for care, and thresholds for considering device implant; (4) To determine health-associated costs for HF subjects in the registry; and (5) To link the REVIVAL data-set to INTERMACS to perform comparative analyses of outcomes of patients treated with medical vs VAD therapy.

Here we report the baseline clinical characteristics of the REVIVAL cohort and the performance of the study’s inclusion criteria for identifying a high-risk ambulatory HF cohort consistent with REVIVAL’s aims.

Methods

Study population

Twenty-one implanting VAD centers screened 927 patients and enrolled 400 subjects with ambulatory advanced chronic systolic HF in a prospective, observational study between July 2015 and June 2016. (See Supplementary Figure S1 at www.jhltonline.org for CONSORT diagram). Entry criteria utilized only information collected during routine clinical care. Inclusion criteria were chosen with the goal of identifying a high risk but ambulatory, advanced, chronic systolic HF cohort receiving guideline-directed medical therapy, as tolerated (Table 1). The specific goal was a 25% risk for the primary composite outcome of death, durable MCS implantation, or urgent transplantation. Exclusion criteria sought to identify patients with contraindications to LVAD, although the risk for right HF was not considered (Table 2).

Table 1.

REVIVAL Inclusion Criteria

Eligible Subjects Must Meet Each of the Following Criteria

Ambulatory. Chronic systolic heart failure ≥ 12 months. NYHA II – IV for at least 45 of the last 60 days. Last documented left ventricular ejection fraction ≤ 35% by any imaging modality. Age 18 – 80 years. Currently under the care of a cardiologist at study site. On appropriate evidenced -based heart failure medications - ACE inhibitor, ARB or sacubitril-valsartan; beta blocker; aldosterone antagonist; hydralazine/long-acting nitrate [required of African-American subjects only] for ≥ 3 months absent contraindications or intolerances. Has ICD or CRT-D. If CRT-D, present for ≥ 3 months. Demonstrated advanced heart failure, including any one of the following*: Serum sodium ≤ 135 mEq/L (obtained as an outpatient)** Serum BNP ≥ 750 pg/mL or NT-proBNP ≥ 3000 pg/mL** (obtained as an outpatient) Seattle Heart Failure Model (SHFM) one year predicted survival ≤ 85%** Heart Failure Survival Score (HFSS) ≤ 7.19** Peak VO2 ≤ 55% of predicted for age by Wasserman equation or ≤ 14 ml/kg/min, with RER ≥ 1.05 *** VE/VC02 slope ≥ 40*** 6 minute walk test (6MWT) distance ≤ 350 m without significant non-cardiac limitation** Currently listed as UNOS Heart Transplant Status 2 due to heart failure limitation History of one (1) hospitalization (≥ 24 hours) for acute or acute on chronic heart failure in the past year with either serum BNP ≥ 500 pg/mL or NT-proBNP ≥ 2000 pg/mL** (obtained as an outpatient) History of two (2) hospitalizations (≥ 24 hours) for acute or acute on chronic heart failure in the past year. Willingness to continue to receive heart failure care from the enrolling advanced heart failure clinic over the next two (2) years and to come for all scheduled study visits. Written informed consent given.

^*Qualifying measure must be the most recent of that type of measure obtained (i.e., a BNP ≥ 1000 obtained 2 months prior would not qualify the heart failure subject if a more recent BNP was < 1000)

^**Using values obtained within the prior 90 days, except for peak VO₂ within 365 days

^***Obtained within the prior 365 days

Table 2.

REVIVAL Exclusion Criteria

Serial no	Subjects are Ineligible if Any of the Following Criteria are Met
1.	Known serious medical problem other than heart failure that would be expected to limit 2-year survival (≥50% mortality within 2 years from non-heart failure diagnosis)
2.	Patient is not likely to be compliant with the protocol, in the opinion of the Investigator
3.	Currently hospitalized
4.	Current use of an intravenous inotrope
5.	Primary functional limitation from non-cardiac diagnosis even if not likely to limit survival
6.	Chronic hemodialysis or peritoneal dialysis or serum creatinine value of ≥3 mg/dl at time of enrollment
7.	Cardiac amyloidosis, cardiac sarcoidosis, constrictive pericardial disease, active myocarditis, or congenital heart disease with significant structural abnormality
8.	Hypertrophic cardiomyopathy unless dilated LV and no outflow gradient
9.	Cardiac conditions that are amenable to surgical or percutaneous procedures (other than VAD or transplant) that would substantially improve prognosis and for which this subject is a reasonable candidate, regardless of whether the procedure will or will not be performed
10.	Uncorrected hyperthyroidism or hypothyroidism
11.	Pregnancy

Abbreviations: LV, left ventricle; REVIVAL, Registry Evaluation of Vital Information for VADs in Ambulatory Life; VAD, ventricular assist device.

An independent Observational Study Monitoring Board over saw the conduct of REVIVAL. Institutional Review Boards overseeing the Data Coordinating Center at the University of Michigan and each participating center approved the protocol. Written informed consent was obtained for all study participants.

Data collection

All enrolled subjects were followed and assessed for events for 2 years post-enrollment or until earlier occurrence of a study end point. Study visits occurred at enrollment (Baseline A); 2 months (Baseline B); and at 6, 12, 18, and 24 months. Visit windows extended ±30 days at 2 months and ±45 days for the 6, 12, 18, and 24 month visits. Events included hospitalizations, stroke, implantation of a temporary or durable MCS, heart transplant, and death. Site investigators classified hospitalizations as related to HF or not and death as cardiovascular related or not. United Network of Organ Sharing (UNOS) waitlist status (pre–October 2018 revision levels 1A, 1B, or 2) was recorded at Baseline A, at the time of durable MCS implant, and at the time of transplant. INTERMACS Patient Profile was noted at each study visit and at the time of durable MCS implant. Death, UNOS 1A or 1B heart transplant, and implantation of durable MCS comprise the primary composite outcome, and subjects were not followed further after experiencing one of the constituent events.

Baseline and serial data collection included demographics; medical history (including medication use and cardiac procedures); routine labs; subjective (New York Heart Association [NYHA] class and INTERMACS Patient Profile) and objective (6-minute walk) functional assessments; frailty measures (15-foot walk time and handgrip strength); events (hospitalizations, stroke, cardiac procedures, transplantation, and VAD); and instruments for self-assessment of HRQOL (EuroQoL-EQ-5D-3L and Kansas City Cardiomyopathy Questionnaire), depression (Personal Health Questionnaire-8 item), anxiety (State Trait Anxiety Inventory-Form Y), patient-estimated remaining life-expectancy, and VAD treatment preference.^4–7 In addition, at the Baseline B visit 2 months after enrollment, there was a one-time collection for dedicated core lab analyses of transthoracic echocardiograms; maximal cardiopulmonary treadmill exercise tests with gas measurements; and blood banked for serum biomarkers, DNA, and RNA. Also beginning at Baseline B, identified caregivers were asked to provide information on their health history, HRQOL, and caregiver burden. Additional details are provided in the Supplementary Appendix online.

REVIVAL database

Information about the REVIVAL database, data management, and data security are presented in the Supplementary Appendix online.

Statistical analysis

Baseline characteristics are displayed as mean ( ± standard deviation) or median (25th, 75th percentiles) for normally and non-normally distributed continuous data, respectively. Categorical data are displayed as percentages. The Heart Failure Survival Score (HFSS) risk groups and Seattle Heart Failure Model (SHFM) predicted 1-year survival were determined as previously published.^8,9

The Kaplan–Meier method was used to estimate the time dependent risk of the primary combined end point. The competing risk method was also used to estimate the cumulative incidence function of a patient moving into 1 of the 3 mutually exclusive events of death, implantable durable MCS, or urgent transplantation.

Study inclusion criteria 9i–9viii (non-hospitalization criteria), 9ix (≥1 HF hospitalization in the past year and elevated natriuretic peptide level), and 9x (≥2 HF hospitalizations), as specified in Table 1, were derived from a melding of Randomized Evaluation of VAD InterVEntion before Inotropic Therapy (REVIVE-IT), Medical Arm for Mechanically Assisted Circulatory Support (MedaMACS), and Risk Assessment and Comparative Effectiveness of Left Ventricular Assist Device and Medical Management in Ambulatory Heart Failure Patients (ROADMAP) criteria,^10–13 with some thresholds relaxed to enrich the sample with advanced American College of Cardiology/American Heart Association Stage C (rather than Stage D) patients. The test performance characteristics of these study inclusion criteria, along with other selected clinical data, were assessed for prediction of the primary composite outcome (death, implantation of durable MCS, or urgent UNOS Status 1A or 1B transplant) at 1-year of follow-up. Sensitivity (percent test positive among those who had an event at or before the 1-year follow-up visit), specificity (percent test negative among those who did not have an event at or before the 1-year follow-up visit), positive predictive value (percent of those with a positive test result who had an outcome event) and negative predictive value (percent of those with a negative test result who did not have an outcome event) were calculated. Univariable logistic regression was used to calculate c-statistics. All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).

Results

Analysis cohort

Of 400 subjects enrolled in the study, 2 provided no follow-up information beyond enrollment, leaving 398 evaluable subjects. Data on 25 evaluable subjects whose follow-up did not extend through 1-year and who did not experience a primary outcome event were excluded from subsequent analyses. This included 4 who were lost to follow-up (median follow-up, 234 [186, 303] days), 15 who withdrew from the study (median follow-up, 165 [77, 238] days), and 6 who received UNOS status 2 transplants (median, 111 [32, 152] days). The remaining 373 patients constituted the analysis cohort.

Baseline characteristics

Baseline characteristics for the analysis cohort are shown in Table 3 (Baseline characteristics for the entire 400 patient cohort are included in the Supplementary Appendix online).The cohort averaged age 60 years, 26% were women, and 26% self-identified as African American. Of the cohort, 70% were classified as NYHA class III, IIIB, or IV, but 72% were classified as INTERMACS Profiles 6 or 7. Guideline-directed medical therapy was strongly evidenced, and only 1% of subjects did not have an implanted cardiac resynchronization therapy defibrillator or implantable cardioverter-defibrillator. Renal dysfunction (estimated glomerular filtration rate by 4-variable Modification of Diet in Renal Disease equation < 60 ml/min/1.73 m²) was present in 60% of the cohort. Functional testing showed substantially impaired 6-minute walk test performance (median, 336 [279, 397] meters).

Table 3.

Baseline Visit Data for Subjects Evaluable for the Primary Combined Outcome at 1-Year Follow-Up (n = 373 Unless Otherwise Noted)

Demographic	Value
Age, years	60.3 ± 11.2
Female	26
Race
African American/black	25.5
Am Indian/Alaskan native	0.5
Asian	1.3
White	68.9
Other/unknown	2.4
More than one race	1.3
Ethnicity
Hispanic	8.0
Unknown/undisclosed	2.4
Clinical characteristics
Ischemic etiology	45
HR, bpm (n = 368)	74 ± 12
SBP, mm Hg (n = 371)	108 ± 16
NYHA I/II/III/IIIB/IV	2/29/59/8/3
INTERMACS Profile 4/5/6/7	8/19/39/33
Active on transplant list	15
Time followed by heart failure specialists at study site
0–3 months	11
>3 months–1 year	31
>1 year–2 years	17
>2 years	42
LVEF for study entry, %	21 ± 7
CRT-D/ICD/neither	49/50/1
LBBB ≥ 150 msec w/o CRT, %	1
Serum sodium, mg/dl	138 ± 4
Serum creatinine, mg/dl	1.4 ± 0.7
eGFR < 60/ml/1.73 m2	60
ALT, IU/liter (n = 369)	28 ± 22
Uric acid, mg/dl (n = 348)	8.1 ± 2.3
Total cholesterol, mg/dl (n = 356)	160 ± 46
ACEI/ARB/ARNI	43/26/15
Enalapril equivalents,a mg/day (n = 162)	9 (5, 10)
Valsartan equivalents,b mg/day (n = 152)	81 (40, 160)
Beta blocker	95
Carvedilol equivalents,c mg/day	25 (12.5, 50)
Loop diuretic	92
Furosemide equivalents,d mg/day	80 (40, 160)
Mineralocorticoid antagonist	74
SHFM predicted 1-year survival, % (n = 370)	93 (88, 97)
Selected functional and patient reported outcomes and self-assessed prognosis
6-minute walk test, meters (n = 344)	336 (279, 397)
15-foot walk time, seconds (n = 343)	4.1 (3.5, 5.0)
EQ-5D-Visual Analogue Scale score (n = 343)	65 (50, 75)
Kansas City Cardiomyopathy Questionnaire Overall Summary Score (n = 349)	64 (48, 78)
Patient self-assessed remaining life, years (n = 268)	6.5 (5.0. 10.0)

Abbreviations: ACEI, angiotensin-converting enzyme inhibitor; ALT, alanine aminotransferase; Am Indian, American Indian; ARB, angiotensin receptor blocker; ARNI, angiotensin receptor-neprilysin inhibitor; CRT, cardiac resynchronization therapy; CRT-D, cardiac resynchronization therapy defibrillator; eGFR, estimated glomerular filtration rate from the 4-variable Modification of Diet in Renal Disease equation; HR, heart rate; ICD, implantable cardioverter-defibrillator; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; LBBB, left bundle branch block; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; SBP, systolic blood pressure; SHFM, Seattle Heart Failure Model; w/o, without.

Continuous variables expressed as mean ± SD or median (25th, 75th percentiles), as appropriate. Categorical variables expressed as percentage.

^aEnalapril equivalents: enalapril 40 mg = lisinopril 40 mg = fosinopril 40 mg = quinapril 40 mg = captopril 150 mg = ramipril 10 mg = trandolapril 4 mg = benazepril 40 mg.

^bValsartan equivalents: valsartan 40 mg = candesartan 4 mg = losartan 25 mg = irbesartan 75 mg = olmesartan 10 mg = telmisartan 20 mg. Valsartan from sacubitril/valsartan was multiplied by 1.583.

^cCarvedilol equivalents: carvedilol (short acting) 50 mg = Coreg CR (carvedilol sustained release) 80 mg = metoprolol tartrate 200 mg = Toprol XL (metoprolol succinate) 200 mg = bisoprolol 10 mg = sotalol 320 mg.

^dFurosemide equivalents: furosemide 40 mg per os = torsemide 20 mg per os = bumetanide 1 mg per os.

Outcomes

At 410 days (365 + 45-day visit window), 373 subjects experienced 90 combined primary outcome events. Of these, 46 subjects underwent implantation of a durable implantable LVAD, 14 received a UNOS 1A or 1B transplant, and 30 died. The time courses of these events are shown as competing outcome curves in Figure 1. Kaplan-Meier–estimated event risks at the 1-year visit were 24.1% for the primary composite outcome, 12.3% for durable implantable LVAD, 3.8% for a UNOS 1A or 1B transplant, and 8.0% for death. For the entire 400 patient cohort, freedom from the primary composite outcome was 23.5%. Figure 2 displays the distribution of outcome events over the first year overall and as identified by each inclusion criterion and selected additional criteria.

Figure 1.

Kaplan-Meier curve. Curve shows the estimate for the primary combined outcome and competing outcome curves displaying the cumulative incidence function estimates for death, UNOS 1A or 1B transplant, or durable MCS implantation. MCS, mechanical circulatory support; UNOS, United Network of Organ Sharing.

Figure 2.

Timing of primary combined outcome events that were identified (true positives) by each study inclusion criteria and selected additional criteria through the 12-month study visit. 6MW, 6-minute walk; BNP, B-type natriuretic peptide; HFSS, Heart Failure Survival Score; Hosp, hospitalization; NT-proBNP, N-terminal prohormone of B-type natriuretic peptide; SHFM, Seattle Heart Failure Model; UNOS, United Network of Organ Sharing; VCO₂, carbon dioxide production; VE, ventilation; VO₂, oxygen consumption.

Table 4 displays the high-risk inclusion criteria (criteria 9i–9viii, 10, and 11 in Table 1) and selected additional characteristics. For each, the number of positive and negative tests results, sensitivity, specificity, positive predictive value, negative predictive value, and c-statistic are displayed. Serum sodium, available in 91% of patients, and the natriuretic peptide entry criterion (B-type natriuretic peptide ≥ 750 pg/ml or N-terminal prohormone of B-type natriuretic peptide ≥ 3,000 pg/ml), available in 78% of patients, had similarly high specificities (75% and 78%, respectively), but serum sodium was much less sensitive (28% vs. 46%). Routinely collected clinical information required to calculate 2 predictive risk scores (SHFM and HFSS) was not available in most patients. Both displayed perfect specificity, but threshold values were rarely (SHFM) or never (HFSS) met. Cardiopulmonary exercise test data were available for less than half of the subjects. A peak oxygen consumption ≤ 14 ml/min/kg or ≤55% of that predicted by the Wasserman equation had moderate sensitivity and specificity (58% and 52%, respectively); ventilatory equivalents for carbon dioxide ≥ 40 was much less sensitive (21%) but much more specific (76%). UNOS Status 2 (elective) transplant waitlist priority was very insensitive (17%) but highly specific (87%). If any one (or more) of the non-hospitalization criteria listed previously was present, the sensitivity rose to 89%, at the cost of a decline in specificity to 17%. At least 1 HF hospitalization in the prior year had only slightly lower sensitivity (57%) and specificity (47%) than the peak oxygen consumption criterion. Note, however, that a single HF hospitalization in the prior year was not a study inclusion criterion; 111 of these patients met inclusion criteria by having had ≥ 2 HF hospitalizations in the prior year, and 91 met one of the non-hospitalization high risk entry criteria. Two or more heart failure hospitalizations in the prior year had high specificity (71%) at the cost of lower sensitivity (31%). In comparison, requiring one of the non-hospitalization high-risk features in addition to at least 1 HF hospitalization increased specificity less (63%) and reduced sensitivity less (46%). The SHFM (0.679) and natriuretic peptide (0.622) criteria were the most discriminative individual predictors, with c-statistics of 0.679 and 0.622, respectively.

Table 4.

Test Performance Characteristics of the REVIVAL High-Risk Inclusion Criteria for the Combined Primary Outcome (Death, ble MCS, or UNOS 1A or 1B Transplant) at 1-Year Follow-Up

Characteristic	N	Test POS	Test NEG	Sensitivity	Specificity	PPV	NPV	C-statistic
Serum sodium ≤ 135 mg/dl	339	86	253	28%	75%	27%	76%	0.516
BNP ≥ 750 pg/ml or NT-proBNP ≥ 3,000 pg/ml	292	81	211	46%	78%	42%	81%	0.622
SHFM ≤85% predicted 1-year survival	59	5	54	36%	100%	100%	83%	0.679
HFSS high risk	49	0	49	0%	100%	—	86%	—
Peak VO2 ≤ 14 or ≤55% predicted w/ RER ≥ 1.05	162	81	81	58%	52%	26%	81%	0.554
VE/VCO2≥ 40	156	36	120	21%	76%	19%	78%	0.512
6-minute walk ≤ 350 meters w/o significant non-cardiac limitation	178	91	87	58%	50%	21%	84%	0.540
UNOS 2 listed	358	51	307	17%	87%	29%	76%	0.519
Any non-hospitalization high risk	373	316	57	89%	17%	25%	82%	0.528
≥2 HF hospitalizations in past year	373	111	262	31%	71%	25%	76%	0.509
≥1 HF hospitalizations in past yeara	373	202	171	57%	47%	25%	77%	0.517
≥1 HF hospitalization in past year + any non-hospitalization high riska	373	147	226	46%	63%	28%	78%	0.541

Abbreviations: BNP, B-type natriuretic peptide; HF, heart failure; HFSS, Heart Failure Survival Score; MCS, mechanical circulatory support; NEG, negative; NPV, negative predictive value; NT-proBNP, N-terminal prohormone of B-type natriuretic peptide; POS, positive; PPV, positive predictive value; RER, respiratory exchange ratio; REVIVAL, Registry Evaluation of Vital Information for VADs in Ambulatory Life; SHFM, Seattle Heart Failure Model; UNOS, United Network of Organ Sharing; VAD, ventricular assist device; VCO₂, carbon dioxide production; VE, ventilation; VO₂, oxygen consumption.

^aNot a REVIVAL inclusion Criterion.

Discussion

REVIVAL was designed as a prospective registry of ambulatory patients with advanced HF that would better define the clinical trajectory and relationship between baseline characteristics and outcomes of patients who might be considered for future studies of earlier VAD use. As such, a patient cohort generally considered not ill enough for LVAD or transplant, yet demonstrated here to be at substantial risk for death, urgent transplant, or durable MCS, was enrolled from advanced HF programs at sites offering these therapies. REVIVAL successfully enrolled 400 ambulatory outpatients with chronic systolic HF and achieved the targeted annualized primary combined event rate of ~ 25%.

REVIVAL utilized routinely obtained clinical information to identify a high-risk cohort, as the goal was to simulate recruitment for a clinical trial and the usual information available in that setting. These individual criteria all have been studied extensively and the thresholds chosen are all well-established in the medical literature. However, use of only a single criterion to define HF risk would have resulted in insufficient sensitivity for successful study enrollment. Allowing inclusion if any one of these many inclusion criteria were met allowed identification of a cohort that was very close to the desired 25% annual event risk.

REVIVAL aimed to enroll a less advanced HF sample than those of ROADMAP or MedaMACS (Table 5) and that had been earmarked for enrollment in REVIVE-IT.^3,10,11 ROADMAP targeted patients who were deemed candidates for LVAD therapy by their advanced HF team. MedaMACS employed similar high-risk inclusion criteria to those used in REVIVAL, but set more strict thresholds (e.g., 6-minute walk distance < 300 meters and serum sodium < 135 mg/dl on 2 separate occasions at least 7 days apart) and required that there also be at least 1 HF hospitalization in the prior 12 months. Of the MedaMACS subjects, 80% met enrollment criteria based on having at least 2 HF hospitalizations in the prior 12 months, whereas only 30% of REVIVAL patients had been hospitalized this frequently before study enrollment. As hoped, baseline characteristics in REVIVAL are consistent with a less advanced HF cohort than those of either ROADMAP or MedaMACS. REVIVAL subjects had much greater functional capacity (mean 6-minute walk distance, 336 vs. 196 meters), greater overall generic HRQOL (EQ visual analogue scale, 65 vs. 56), and better HF-related QOL (Kansas City Cardiomyopathy Questionnaire overall summary score, 64 vs. 51) than subjects in the MedaMACS Pilot.¹² Freedom from death, heart transplant (UNOS 1A or 1B for REVIVAL, any transplant for both MedaMACS studies and the optimal medical management arm of ROADMAP), or durable MCS at 1-year for each cohort are consistent with these differences in enrollment criteria and baseline characteristics: 76% for REVIVAL, ~ 63% for the MedaMACS Pilot, 63% for the optimal medical management arm of ROADMAP, and 47% for the MedaMACS Screening Pilot. Thus, REVIVAL represents a less advanced chronic systolic HF cohort that is centered beneath the current threshold at which the advanced HF community believes surgical intervention with a transplant or durable MCS is generally indicated. As a result, REVIVAL is uniquely positioned to map the trajectories of patients between Stage C and Stage D heart failure.

Table 5.

Overview of Clinical Studies of Ambulatory Patients with Advanced Heart Failure Receiving Optimal Medical Management

Characteristic	MedaMACS Screening pilot11	ROADMAP10	MedaMACS Pilot Study12	REVIVAL
Number of patients Enrollment period Design	166 2010–11 Prospective, Observational 10 Centers Single Arm	103 2011–13 Prospective, Observational 41 Centers Comparator Arm with HeartMate II	161 2013–15 Prospective, Observational 11 Centers Single Arm	400 2015–16 Prospective, Observational 21 Centers Single Arm
Inclusion criteria Patient status at enrollment NYHA class Ejection fraction Selected high risk features	Inpatient or Outpatient III–IV ≤30% 2 HF hosp in 12 months, or 1 HF hosp + 1 high risk feature: Peak VO2 < 16 ml/kg/min for men Peak VO2 < 14 ml/kg/min for women w/RER > 1.08 or <55% predicted, or 6MW < 300 m, or BNP >800 pg/ml, or Serum Na+ < 135 mEq/dl	Inpatient or Outpatient IIIB/IV ≤25% ≥1 HF hosp in 12 months, or 2 unscheduled EW visits for HF, and 6MW < 300 m	Inpatient or Outpatient III–IV ≤35% 1 HF hosp in 12 months and one high risk feature: Additional HF hosp in 12 months, or Peak VO2 < 16 ml/kg/min for men Peak VO2 < 14 ml/kg/min for women 6MW < 300 m, or Outpatient NT-BNP > 4,000 pg/ml, or SHFM 1-year predicted survival ≤83%	Outpatient II–IV ≤35% 2 HF hosp in 12 months, or 1 HF hosp in 12 months with outpatient BNP ≥500 pg/ml or NT-proBNP ≥ 2,000 pg/ml, or If no HF hosp, ≥1 high risk feature: Peak VO2 ≤ 16 ml/kg/min or ≤55% predicted with RER ≥1.05, or 6MW ≤ 350 m, or BNP ≥750 pg/ml or NT-proBNP ≥ 3,000 pg/ml as outpatient (w/o HF hosp) SHFM 1-year predicted survival ≤85%
Exclusion criteria	Intravenous inotropes Listed for transplant	Intravenous inotropes Listed for transplant DT VAD ineligible	Intravenous inotropes Listed for transplant Congenital heart defect Amyloidosis	Intravenous inotropes Listed UNOS 1A/B for transplant Congenital heart defect Amyloidosis Dialysis or serum Cr >3 mg/dl
INTERMACS Profiles Enrolled	4%–22% 5%–32% 6%–34% 7%–12%	4%–34% 5%–28% 6%–34% 7%–2%	4%–12% 5%–32% 6%–49% 7%–7%	4%–8% 5%–21% 6%–39% 7%–32%
Kaplan–Meier outcomes at 1 yeara	47% Alive on medical therapy 23% Death w/o VAD/transplant 16% Transplant 15% VAD	63% Alive on medical therapy 19% Death w/o VAD/transplant 0% Transplant 18% VAD	~63% Alive on medical therapy ~19% Death w/o VAD/transplant ~9% Transplant ~9% VAD	76% Alive on medical therapy 8% Death without VAD/urgent transplant 4% Urgent transplant 12% VAD

Abbreviations: 6MW, 6-minute walk; BNP, B-type natriuretic peptide; Cr, creatinine; DT, destination therapy; EW, emergency ward; HF, heart failure; hosp, hospitalization; HMII, HeartMate II assist device; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; MedaMACS, Medical Arm for Mechanically Assisted Circulatory Support; Na⁺, sodium; NT-proBNP, N-terminal prohormone of B-type natriuretic peptide; NYHA, New York Heart Association; RER, respiratory exchange ratio; REVIVAL, Registry Evaluation of Vital Information for VADs in Ambulatory Life; ROADMAP, Risk Assessment and Comparative Effectiveness of Left Ventricular Assist Device and Medical Management in Ambulatory Heart Failure Patients; SHFM, Seattle Heart Failure Model; UNOS, United Network of Organ Sharing; VAD, ventricular assist device; VO₂, oxygen consumption; w/o, without.

^aKaplan–Meier outcomes for the MedaMACS Pilot were estimated from Figure 1A, 1B, and 1C from the published study.¹² Urgent transplants were those categorized as UNOS 1A or 1B at the time of transplant.

Obtaining consensus within the advanced HF community as to where the threshold for durable MCS resides remains critically important. REVIVAL was created following the early termination of the REVIVE-IT, a pilot trial that was to compare a strategy of earlier elective use of a durable LVAD to one of continued optimal medical therapy alone in less advanced ambulatory chronic systolic HF.¹³ In the latter group, LVAD was reserved for those subjects who progressed to more severely advanced HF, for whom LVAD therapy was widely accepted as indicated. REVIVE-IT ceased enrollment after just 1 subject was enrolled, when it became clear that the study pump had an unacceptably high risk of pump thrombosis.¹⁴ However, had this not happened, enrollment in REVIVE-IT would still have been very challenging, perhaps fatally so, as site principal investigators (both HF cardiologists and cardiac surgeons) had widely different perceptions of the level of HF severity at which they would be at clinical equipoise for randomization to LVAD vs medical therapy. This absence of a community consensus would also have been expected to result in differing thresholds for VAD therapy to be used as patients in the initial medical therapy alone arm developed greater HF severity and limitations. Because heart transplantation and LVAD therapy are available as standard of care options for appropriate patients, knowing their rates of use for different profiles of patients with advanced HF is essential information for sample size determination in clinicals trial of earlier durable LVAD use or of any novel therapies in this population. These issues brought into focus the need for REVIVAL.

Although only modest in size at 400 patients, the richness of the data collected in REVIVAL will facilitate deep phenotypic profiling. Patient risk for HF events will be classified by INTERMACS Patient Profile, SHFM and HFSS, serum biomarkers, microRNAs, and DNA polymorphisms. Information from all 5 domains of the Fried frailty phenotype will allow categorization in this important domain. Echocardiographic parameters will define cardiac systolic, diastolic, and valvular function, and maximal cardiopulmonary exercise testing with respiratory gas collection will provide information about functional capacity and ventilatory function, cardiac work, and efficiency. Serial assessments of hospitalizations, medication changes, functional capacity, quality of life, health utility, depression, anxiety, willingness to receive VAD and other end-stage disease life-sustaining therapies, self-assessment of remaining years of life, caregiver quality of life, and caregiver burden will provide a more nuanced understanding of the patient experience in ambulatory advanced HF.

Napoleon Hill, a mid-20th century author of self-help books, wrote that “Opportunity often comes disguised in the form of misfortune or temporary defeat.”¹⁵ Following the lost possibilities of REVIVE-IT, REVIVAL will deepen our understanding of the clinical course and patient and caregiver experience of advanced ambulatory HF, inform the selection of currently available therapies, and provide critical information for the design of future studies in this important patient group.