Author + information
- Received September 26, 2016
- Revision received January 30, 2017
- Accepted February 21, 2017
- Published online June 26, 2017.
- Randall C. Starling, MD, MPHa,∗ (, )
- Jerry D. Estep, MDb,
- Douglas A. Horstmanshof, MDc,
- Carmelo A. Milano, MDd,
- Josef Stehlik, MD, MPHe,
- Keyur B. Shah, MDf,
- Brian A. Bruckner, MDb,
- Sangjin Lee, MS, MDg,
- James W. Long, MD, PhDc,
- Craig H. Selzman, MDe,
- Vigneshwar Kasirajan, MDf,
- Donald C. Haas, MDh,
- Andrew J. Boyle, MDi,
- Joyce Chuang, PhDj,
- David J. Farrar, PhDj,
- Joseph G. Rogers, MDd,
- ROADMAP Study Investigators
- aCleveland Clinic, Cleveland, Ohio
- bHouston Methodist Hospital, Houston, Texas
- cIntegris Baptist Medical Center, Oklahoma City, Oklahoma
- dDuke University, Durham, North Carolina
- eUniversity of Utah, Salt Lake City, Utah
- fVirginia Commonwealth University, Richmond, Virginia
- gSpectrum Health, Grand Rapids, Michigan
- hAbington Memorial Hospital, Abington, Pennsylvania
- iThomas Jefferson University, Philadelphia, Pennsylvania
- jAbbott, Pleasanton, California
- ↵∗Address for correspondence:
Dr. Randall C. Starling, Kaufman Center for Heart Failure, Heart and Vascular Institute, Cleveland Clinic, 9500 Euclid Avenue, Desk J3-4, Cleveland, Ohio 44195.
Objectives The authors sought to provide the pre-specified primary endpoint of the ROADMAP (Risk Assessment and Comparative Effectiveness of Left Ventricular Assist Device and Medical Management in Ambulatory Heart Failure Patients) trial at 2 years.
Background The ROADMAP trial was a prospective nonrandomized observational study of 200 patients (97 with a left ventricular assist device [LVAD], 103 on optimal medical management [OMM]) that showed that survival with improved functional status at 1 year was better with LVADs compared with OMM in a patient population of ambulatory New York Heart Association functional class IIIb/IV patients.
Methods The primary composite endpoint was survival on original therapy with improvement in 6-min walk distance ≥75 m.
Results Patients receiving LVAD versus OMM had lower baseline health-related quality of life, reduced Seattle Heart Failure Model 1-year survival (78% vs. 84%; p = 0.012), and were predominantly INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) profile 4 (65% vs. 34%; p < 0.001) versus profiles 5 to 7. More LVAD patients met the primary endpoint at 2 years: 30% LVAD versus 12% OMM (odds ratio: 3.2 [95% confidence interval: 1.3 to 7.7]; p = 0.012). Survival as treated on original therapy at 2 years was greater for LVAD versus OMM (70 ± 5% vs. 41 ± 5%; p < 0.001), but there was no difference in intent-to-treat survival (70 ± 5% vs. 63 ± 5%; p = 0.307). In the OMM arm, 23 of 103 (22%) received delayed LVADs (18 within 12 months; 5 from 12 to 24 months). LVAD adverse events declined after year 1 for bleeding (primarily gastrointestinal) and arrhythmias.
Conclusions Survival on original therapy with improvement in 6-min walk distance was superior with LVAD compared with OMM at 2 years. Reduction in key adverse events beyond 1 year was observed in the LVAD group. The ROADMAP trial provides risk-benefit information to guide patient- and physician-shared decision making for elective LVAD therapy as a treatment for heart failure. (Risk Assessment and Comparative Effectiveness of Left Ventricular Assist Device and Medical Management in Ambulatory Heart Failure Patients [ROADMAP]; NCT01452802)
The ROADMAP (Risk Assessment and Comparative Effectiveness of Left Ventricular Assist Device and Medical Management in Ambulatory Heart Failure Patients) study’s 1-year results showed that survival with improved functional status at 1 year was better with left ventricular assist devices (LVADs) compared with optimal medical management (OMM) in a patient population of ambulatory New York Heart Association (NYHA) functional class IIIB/IV patients who were not dependent on intravenous inotropic support (1). Survival was similar in both groups in the intention-to-treat analysis. However, as-treated event-free actuarial survival over a 1-year period was significantly better with LVAD than OMM (80 ± 4% vs. 63 ± 5%). Differences in the primary endpoint between LVAD and OMM were primarily due to the use of delayed LVADs in the OMM group. Factors beyond survival seem paramount to decision making surrounding LVAD implantation in this ambulatory patient population. At 1-year follow-up, patients in the OMM group avoided LVAD surgery and LVAD-associated adverse events (AEs); however, patients observed on OMM did not achieve the primary benefits of functional improvements and patient-reported health-related quality of life (HRQoL) with LVAD support. To better understand the long-term benefits and risks of LVAD compared with OMM, we now provide the pre-specified primary endpoint and other important study results after 2 years of follow-up.
The ROADMAP study was a prospective, multicenter (N = 41), nonrandomized, controlled, observational study to evaluate the effectiveness of LVAD versus OMM (2). Enrollment began in October 2011 with patients being followed for up to 2 years. The primary composite endpoint was survival with improvement in 6-min walk distance (6MWD) ≥75 m at 1 year, which has been previously published (1). This report focuses on the primary endpoint at 2 years and secondary study endpoints, which include actuarial survival, HRQoL, depression, functional status, and AEs, after 2 years of follow-up.
In addition to meeting U.S. Food and Drug Administration–approved indications for HeartMate II LVAD (Thoratec [now Abbott], Pleasanton, California) destination therapy, entrance criteria included NYHA functional class IIIB/IV, at least 1 hospitalization for heart failure (HF) (or 2 unscheduled emergency department/infusion clinic visits) in the last 1 year, and 6MWD <300 m. Subjects were excluded if there was inotrope use within 30 days before enrollment. Of the 200 patients enrolled, 97 selected LVAD therapy and 103 remained on medical management in the OMM arm. Patients in the OMM cohort could receive a delayed LVAD at any point during the study period.
Baseline and follow-up assessments
Baseline assessments included demographic characteristics, medical history, NYHA functional class, INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) profile, 6MWD, serum chemistry, hematologic data, and medications. Patients also completed a HRQoL survey, the EuroQol 5 dimensions, 5-level questionnaire (EQ-5D-5L) including the visual analog scale (VAS), and a depression screening questionnaire, the Patient Health Questionnaire (PHQ-9). A summary of baseline patient characteristics has been previously published (1). Clinical follow-up took place every 6 months for up to 2 years and included assessment of HRQol, depression, functional status, and laboratory parameters. Prevalence, incidence, causes of rehospitalizations, and causes of death were documented by study sites, but not adjudicated. AEs were captured using standardized INTERMACS definitions as reported by investigators.
Continuous variables are reported as mean ± SD or SE, or median and quartiles, and categorical data are reported as percentages. Differences between groups were analyzed by the Fisher exact test. Differences between groups of independent, normally distributed, continuous variables were evaluated using the 2-sample Student t test. Variables that were not normally distributed were compared between treatments using the Wilcoxon rank sum test. A 2-sided p value <0.05 is significant.
The primary endpoint evaluated at 2 years was a composite of survival and improvement in 6MWD of ≥75 m. Urgent heart transplantation (HTx) and explant after LVAD complications in LVAD patients were considered treatment failures. Receipt of delayed LVADs and urgent HTx in OMM patients were also considered treatment failures. Patients who withdrew from the study were excluded from the 2-year primary endpoint analysis.
Actuarial survival as treated on original therapy was performed with the Kaplan-Meier method for LVAD patients free of urgent HTx or explant, and for OMM patients free of LVAD implantation or urgent HTx. Intent-to-treat survival was also determined through 2 years post-enrollment. Differences between groups were determined with the log-rank test.
The prevalence of patients with AEs within 2 years and the incidence rate of events per patient-year (eppy) were determined for both groups. A composite event rate was calculated as the sum of eppy for bleeding, driveline infection, pump thrombus, stroke, arrhythmias, and worsening HF. Risk ratio evaluation and comparison of AE rates were performed using Cochran-Mantel-Haenszel statistics.
Paired changes in 6MWD, PHQ-9, and EQ-5D VAS from baseline to 2 years were compared between patients surviving to 2 years on original therapy using mixed-effects modeling on ranks with Tukey adjustment for pairwise comparisons. Post hoc composite endpoints of favorable outcomes at 2 years were determined, defined as patients alive on original therapy at 2 years with improvement in NYHA functional class, increase in VAS of more than 20 points in subjects with impaired HRQoL at baseline as defined by baseline VAS <68 (top quartile of values), and depression score improvement of at least 5 points in patients at baseline with at least mild depression (PHQ-9 ≥5). A biostatistician independent of the sponsor and the investigators provided an independent validation of study results.
A chart of all patients and main outcomes over the 2-year study period is shown in Figure 1. Of the 103 OMM patients, 18 died, 18 received delayed LVADs, and 9 withdrew during the first year. During the second year, an additional 13 OMM patients died, 5 received delayed LVADs, 1 received a HTx, and 4 withdrew from the study. Of the 97 LVAD patients, 17 died, 3 received HTx, and 3 withdrew during the first year of the study. Of the 74 patients remaining on LVAD support at 1 year, an additional 8 died, 4 received HTx, 1 was explanted, and 1 withdrew from the study during the second year. At the end of 2 years, 35 (34%) OMM patients and 60 (62%) LVAD patients were alive on their original therapy.
At 2 years, more LVAD patients (30%) met the primary endpoint of alive on original therapy with improvement in 6MWD of at least 75 m than OMM patients (12%) (odds ratio: 3.2 [95% confidence interval: 1.3 to 7.7]; p = 0.012). The main reason for fewer OMM patients meeting the primary endpoint compared with the LVAD group was the use of delayed LVADs. In the second year, a larger percentage of deaths and additional delayed LVAD implants in OMM patients contributed to fewer OMM patients meeting the primary endpoint at 2 years (Table 1). A sensitivity analysis for the effect of missing data on the primary endpoint is shown in Online Table 1. Significantly more LVAD patients met the primary endpoint compared with OMM patients even when those withdrawn from the study or with missing 6-min walk test were counted as either treatment successes or failures.
Twenty-four-month survival as treated on original therapy (event-free survival) was greater for LVAD versus OMM (70 ± 5% vs. 41 ± 5%; p < 0.001) (Figure 2). There was no difference in intent-to-treat survival for LVAD versus OMM (70 ± 5% vs. 63 ± 5%; p = 0.307) (Figure 3). A competing outcomes analysis for all LVAD and OMM patients is also shown in Online Figure 1.
Functional status and quality of life
At 2 years, in patients still on original therapy, LVAD patients experience greater improvements in functional status and quality of life (QoL) compared with OMM patients. At study enrollment, no patients were in NYHA functional class I or II. By 2 years, 69% of LVAD patients had improved to class I or II, which was significantly >37% of the remaining OMM patients (p < 0.01) (Figure 4). LVAD patients exhibited a significant increase in 6MWD from baseline by an average of 74 m (p < 0.05) after 2 years of support compared with no significant change in OMM patients (Table 2).
Baseline health-related QoL measures are shown in Table 2. The EQ-5D VAS increased significantly more for LVAD versus OMM at 2 years with an average improvement of 27 points for LVAD patients compared with 8 points for OMM patients (p < 0.001) (Table 2). At 2 years, the PHQ-9 depression score for LVAD patients improved significantly from enrollment by an average of 4.6 points, with no significant change in OMM patients (Table 2).
The primary and secondary composite endpoints showing the percentage of patients alive at 2 years on original therapy with improvements in NYHA functional class, EQ-5D VAS, and depression all showed significant improvements in LVAD versus OMM patients (Figure 5).
The 22 patients with delayed VADs (not including the 1 with a total artificial heart) received implants after a median time from enrollment of 4.9 (min to max: 1.2 to 19.7) months (Table 3). The median INTERMACS profile at enrollment was 5 (min to max: 4 to 6) and dropped to 4 (min to max: 2 to 6) at the time of implantation. Fifty-five percent of OMM patients receiving delayed VADs were inotrope dependent when they received implants, but 45% remained INTERMACS 4 to 6. Median Seattle Heart Failure Model–predicted survival dropped from 89% (quartile 1 to 3: 81% to 92%) at enrollment to 71% (quartile 1 to 3: 40% to 86%) before delayed implant, and median 6MWD decreased from 219 (quartile 1 to 3: 192 to 272) m to 90 (quartile 1 to 3: 0 to 221) m before implantation. There were no deaths within 30 days after delayed LVAD, and survival at 1 year was 90 ± 6%, which was not different from the original LVAD cohort (p = 0.33 log-rank). Additional comparisons of baseline characteristics of LVAD patients implanted at enrollment and before implantation in the delayed group are shown in Online Table 2. Survival curves of both groups are shown in Online Figure 2.
Table 4 lists the prevalence and cumulative AE rates for the entire 2-year follow-up period. Over the 2-year study, the composite event rate was significantly higher for LVAD versus OMM (1.74. vs 0.98 eppy; p < 0.001). Gastrointestinal (GI) bleeding is the most common AE for LVAD patients, and 5 patients contributed more than one-half of all GI bleeding events in the LVAD group. Worsening HF is the most frequent event for OMM patients. Table 5 shows the AE occurrence in the first and the second year of follow-up. The AE profile for OMM did not change significantly over time. In LVAD patients, there was no change in pump thrombus or driveline infection rates from the first to second year. However, LVAD event rates for GI bleeding and ventricular arrhythmias decreased significantly from the first to second year. As a result, the composite event rate was no longer significantly different between LVAD and OMM in the second year of follow-up. Right HF also showed a trend for reduction in the second year. LVAD patients had more hospitalizations than OMM patients throughout the study. Freedom from re-hospitalization at 2 years (Online Figure 3) was 16% for OMM patients and 8% for LVAD patients (p = 0.005). The main reason for readmission of OMM patients was worsening HF. Hospitalizations for LVAD patients were largely due to AEs (predominantly bleeding). In the second year, there was a decrease in the number of LVAD patient readmissions related to AEs (Table 6).
Of the 25 LVAD patients who expired, the leading cause of death was infection (n = 6), followed by stroke (n = 3), device thrombosis (n = 3), multiorgan failure (n = 3), HF (n = 3), and 1 each for bleeding, driveline disconnect, ventricular fibrillation, pleural effusion, respiratory failure, car accident, and unknown. The most common cause of death among the 31 OMM patients who expired was worsening HF (n = 21), followed by cancer (n = 2), 1 each for sudden cardiac death, chronic obstructive pulmonary disease, sepsis, multiorgan failure, and bleeding, and 3 were unknown.
Figure 6 summarizes the benefits and risks of LVAD therapy versus OMM for patients in the ROADMAP study. In the second year of the ROADMAP study, LVAD patients continue to be more likely to reach the primary composite endpoint and their survival as treated on original therapy is significantly better. LVAD patients also continue to be more likely to show improvements in NYHA functional class, HRQoL, and depression at 2 years compared with OMM patients. The composite AE rate during the second year between LVAD and OMM patients is not significantly different.
The principal findings after 2 years of follow-up in the ROADMAP study include: 1) higher survival with improved functional status, improved QoL, and reduced depression in the LVAD group; 2) no increased mortality with delaying LVAD implant while being monitored closely; 3) more hospitalizations in the LVAD than the OMM group throughout the study; and 4) greater rate of major AEs in LVAD than OMM subjects in year 1 but with a reduction in LVAD AEs in year 2. The second-year ROADMAP study results demonstrate the benefit of the HeartMate II LVAD in select functionally limited non–inotrope-dependent HF patients.
Advanced HF is a lethal condition with few options to extend survival and QoL; due to donor shortage, heart transplantation is limited to a fraction of the patients in need. The advent of continuous-flow LVAD therapy heralded the hope to save many lives, and initial encouraging reports sparked a proliferation of LVAD implantations (3–6). Enthusiasm grew with new devices, and subsequently, a National Heart, Lung, and Blood Institute workshop was convened to design a clinical trial to determine whether LVAD use could be safely expanded to “less ill” patients (7). The REVIVE-IT (Registry Evaluation of Vital Information for VADS in Ambulatory Life; NCT01369407) trial intended to compare the effectiveness of LVAD and medical therapy in non–inotrope-dependent patients through a design randomizing patients to the 2 respective treatment approaches. The REVIVE-IT trial faced numerous challenges and was never completed, and has evolved into a registry. The ROADMAP study was conceived at the same time and designed as a nonrandomized trial in which patients and their physicians could determine the choice of LVAD versus OMM therapy, and outcomes of these choices would be carefully assessed. Our goals included determining functional capacity, and the primary endpoint included a clinically relevant incremental walk distance of 75 m on original therapy. Detailed efforts were made to collect data to understand why both physicians and patients chose a therapy and to capture patient-reported outcomes.
The observation that LVAD patients more often met the primary endpoint was anticipated and not surprising based on the 1-year results (1). This finding was also sustained at 2 years with more LVAD patients on original therapy alive and able to walk with an improvement of more than 75 m. For this non–inotrope-dependent population, LVAD mortality in the initial 30 days was a remarkably low 1%; however, the observed AE rates remain similar to what has been reported with contemporary devices (3,4,8–12). What is provocative is that despite the LVAD cohort having more AEs, frequent hospitalizations, and worse pre-LVAD self-reported QoL, LVAD therapy resulted in greater improvement in depression and in QoL at both 1 and 2 years compared with OMM, although potential bias in these types of assessments must be recognized. The AEs at 2 years in patients choosing LVAD included 11% stroke, 12% LVAD thrombosis, 22% arrhythmias, 33% GI bleeding, and 86% required hospitalization, whereas of those choosing to stay on OMM, 50% experienced worsening HF, 13% arrhythmias, 4% stroke, and 76% also required hospitalization. AEs/complications remain a challenge for LVAD therapy (13). However, we are encouraged that in the ROADMAP study, there was a reduction in AEs in the second year of follow-up; this observation was also noted in a single-center report with a reduction in AEs after 6 months (8).
The results show there was no increase in mortality in the delayed LVAD group compared with the original LVAD group post-implant. The delayed LVAD patients were enrolled in the OMM arm with a median INTERMACS profile of 5. At the time of delayed LVAD implantation (a median of 4.9 months after enrollment), the median INTERMACS profile had changed incrementally to 4, which is the same as the original LVAD cohort at implantation. Thus, one would expect similar survival rates post-LVAD. Patients implanted with LVADs later in the ROADMAP study are not delayed compared with standard clinical practice, where most patients today are inotrope dependent in profiles 2 or 3. Hence, results cannot be generalized to patients with greater clinical deterioration before implantation.
The ROADMAP study has demonstrated that patients may make choices contrary to what might be intuitive to the physician. Patients who are living with HF may be willing to accept a burden of AEs to improve their functional capacity and QoL. Earlier referral enables better education and shared decision making with the patient, which is important for weighing benefits and risks of LVAD therapy. Patients may be willing to accept a “trade off”: improvement in QoL and functional capacity versus potential AEs. Bruce et al. (14) described the importance of a patient-centered perspective and discussions that include 3 domains for patient education and decision making regarding LVADs: survival, functional capacity/QoL, and AEs. The importance of developing decision aids for patients with advanced HF considering the extremely complex process and lifestyle of LVAD therapy has been highlighted recently (15).
The study was nonrandomized; hence, a potential exists for bias regarding patient selection and patient-reported outcomes including HRQoL and depression, which were nonblinded. There was a high withdrawal rate from the study, with more withdrawing in the OMM arm than in the LVAD arm, thus introducing a potential bias. However, if the primary endpoint is reanalyzed counting all withdrawn patients as treatment failures, the study findings are maintained, with more LVAD than OMM patients still meeting the endpoint (28% vs. 10%; odds ratio: 3.5 [95% confidence interval: 1.5 to 8.4]; p = 0.004) (Online Table 1). Data were also missing for some endpoints, including the 6-min walk test, as noted in Table 2, which could have an effect on the results. The event free survival analysis did not include pump replacement as an event. However, if pump replacement is included, event-free survival for LVAD patients remains significantly higher than for OMM patients (Online Figure 4). AEs were reported by treating physicians and not adjudicated by a clinical events committee. Patient management was per center practices, thus variations in medical management, anticoagulation, bleeding, and other events may not have been attributed solely to patient or device characteristics. Nonetheless, the ROADMAP study provides important information for clinicians prescribing LVAD therapy. Some patients and clinicians were not at equipoise and either believed LVAD therapy was superior or that OMM therapy was superior, and they made decisions based on this belief. Many HF specialists feel that current LVAD technology is not ready to be utilized in randomized trials of less-sick patients, but the severity of illness of the target patient population is a critical factor. On the basis of our observed 24-month intent-to-treat survival in LVAD (70%) and OMM (63%) patients, a trial would be difficult to enroll, and would need to randomize approximately 700 patients in each arm to have 80% power to demonstrate a significant difference in 2-year survival. However, if restricted to INTERMACS profile 4 patients, where our observed intent-to-treat survival was 66% in LVAD and 49% in OMM patients, a trial would only need to randomize approximately 130 patients in each arm to show such a difference.
The ROADMAP study has demonstrated the role of LVAD therapy in non–inotrope-dependent patients with advanced HF to improve functional capacity and QoL, albeit with concomitant AEs. Delaying the decision for LVAD therapy when monitored closely by a HF specialist with access to LVAD therapy and transplantation does not impair survival, though it does delay QoL and functional improvements. The need for a randomized clinical trial in expanded patient populations that are not inotrope dependent remains. Hopefully, such trials can be realized with the emerging generation of LVADs, which new technology shows promise for reduced AEs (16–18).
COMPETENCY IN MEDICAL KNOWLEDGE: The ROADMAP trial provides the first evidence that non–inotrope-dependent patients electing LVAD therapy will gain functional capacity and QoL compared with similar patients electing medical therapy. Although 22% of patients treated medically received delayed LVADs after clinical deterioration, their survival rates were equivalent. AEs are prevalent with both strategies, and readmissions are not reduced after LVAD. Despite the limitations, LVAD therapy will be preferred by some patients with advanced HF who are not yet inotrope dependent. The ROADMAP trial provides information to assist caregivers when counselling HF patients making important treatment decisions.
TRANSLATIONAL OUTLOOK: LVAD therapy improves functional capacity and QoL, but not survival, compared with medical therapy in less ill patients who are non-inotrope dependent. The observation that QoL and functional capacity improve despite frequent AEs and frequent need for hospitalization is counterintuitive. Patients, in shared decisions with their physicians, should decide what treatment option is preferred. New technology is needed to reduce LVAD AEs. A randomized clinical trial in “less ill” patients with new LVAD technology should be forthcoming with endpoints related to functional capacity and QoL.
For supplemental tables and figures, and acknowledgment of the study investigators and participating institutions, please see the online version of this article.
The ROADMAP trial was sponsored and conducted by Thoratec Corporation (now Abbott). All author relationships with Thoratec and St. Jude Medical are now with Abbott. Dr. Starling has served as a member of the steering committee for Thoratec; has received research support from Thoratec, Medtronic, and St. Jude Medical. Dr. Estep has served as a consultant for Maquet, St. Jude Medical, and Thoratec; has received research support from Thoratec; has received speakers honoraria from St. Jude Medical; and has been a medical advisor for Medtronic. Dr. Horstmanshof has served as a consultant and has received research support from Thoratec. Dr. Milano has received research support from Thoratec; and has been an instructor for St. Jude Medical and Medtronic. Dr. Stehlik has received research support from St. Jude Medical and Thoratec. Dr. Shah has received research support from and been a consultant for St. Jude Medical and HeartWare. Dr. Long has received research support from and provided educational program support for Thoratec. Dr. Selzman has received research support from Thoratec. Dr. Kasirajan has received institutional research support from Thoratec, Atricure, Abiomed, and Suncardia; has served as a consultant for Atricure and Carmat SA; and served on the Boston Scientific Speakers Bureau. Dr. Haas has received research support and speakers honoraria from St. Jude Medical and Thoratec. Dr. Boyle have served as a consultant for Thoratec. Dr. Chuang is an employee of Abbott. Dr. Farrar is an employee of and shareholder in Abbott. Dr. Rogers has received research support from Thoratec. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
John R. Teerlink, MD, served as Guest Editor for this article.
- Abbreviations and Acronyms
- 6-min walk distance
- adverse event
- events per patient-year
- EuroQol 5 dimensions questionnaire
- heart failure
- health-related quality of life
- heart transplantation
- left ventricular assist device
- New York Heart Association
- optimal medical management
- Patient Health Questionnaire
- quality of life
- visual analog scale
- Received September 26, 2016.
- Revision received January 30, 2017.
- Accepted February 21, 2017.
- 2017 The Authors
- Estep J.D.,
- Starling R.C.,
- Horstmanshof D.A.,
- et al.
- Park S.J.,
- Milano C.A.,
- Tatooles A.J.,
- et al.
- Hasin T.,
- Marmor Y.,
- Kremers W.,
- et al.
- Smedira N.G.,
- Hoercher K.J.,
- Lima B.,
- et al.
- Smedira N.G.,
- Blackstone E.H.,
- Ehrlinger J.,
- et al.
- Teuteberg J.J.,
- Slaughter M.S.,
- Rogers J.G.,
- et al.,
- ADVANCE Trial Investigators
- Rogers J.G.
- Bruce C.R.,
- Blumenthal-Barby J.S.,
- Meyers D.
- Thompson J.S.,
- Matlock D.D.,
- McIlvennan C.K.,
- Jenkins A.R.,
- Allen L.A.
- Mehra M.,
- Naka Y.,
- Uriel N.,
- et al.
- Netuka I.,
- Sood P.,
- Pya Y.,
- et al.
- Uriel N.,
- Adatya S.,
- Mehra M.R.