Comorbid Conditions and Health-Related Quality of Life in Ambulatory Heart Failure Patients
Introduction
Ambulatory patients with advanced heart failure (HF) treated with oral medical therapy have comparable survival to those undergoing left ventricular assist device (LVAD) implantation.1 Such ambulatory HF patients, classified as Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profiles 4 to 7, comprise ≈16% of those undergoing contemporary LVAD implantation.2 Absent improved survival—the rationale for LVAD therapy in these patients—is improved health-related quality of life (HRQOL).
HF patients have multiple chronic conditions,3 and these comorbidities may impact HRQOL as much as HF does.4 One potential contributor to continued low HRQOL after LVAD could be the persistence of noncardiac comorbidities. Understanding the influence of noncardiac comorbidities on HRQOL in advanced HF is critical to understanding the potential for improvement in HRQOL after LVAD.
We sought to determine the impact of noncardiac comorbid conditions that would be expected to persist beyond LVAD therapy on HRQOL in ambulatory HF patients eligible for advanced therapies and thereby identify those patients most likely to anticipate improvement in HRQOL following LVAD. We hypothesized that ambulatory HF patients with multiple noncardiac comorbidities as measured by an integrated comorbidity index would have reduced HRQOL irrespective of HF severity.
Methods
Patient Population
We performed a cross-sectional study of 373 of the 400 ambulatory patients in REVIVAL (Registry Evaluation of Vital Information for Ventricular Assist Devices in Ambulatory Life) at the baseline visit. REVIVAL is a 2-year prospective, observational cohort study of ambulatory systolic HF patients with INTERMACS 4 to 7 profiles recruited from 21 US LVAD centers between July 2015 and June 2016.5 Of the 400 patients enrolled, 27 did not complete HRQOL surveys and were not included in this analysis. Patients with New York Heart Association class II to IV symptoms despite optimal medical therapy and a recent nonelective HF hospitalization, heart transplant listing, objective functional limitation, or evidence of neurohormonal activation were eligible for enrollment. Patients were excluded if they had a condition other than HF that would result in ≥50% 2-year risk of death or obvious contraindications to LVAD therapy.
The institutional review board at each clinical site and the data coordinating center approved the study. All subjects provided written informed consent before study participation. The data for REVIVAL have been uploaded to National Heart, Lung, and Blood Institute Biologic Specimen and Data Repository Information Coordinating Center and will become available per the National Institutes of Health policy.
Determination of Comorbidities
Baseline data included demographics, HF severity, medical history, physical exam, submaximal exercise capacity (6-minute walk distance), and laboratory results. The comorbid conditions assessed were selected from those collected in the registry and included in the INTERMACS list of contraindications to transplantation,6 the Charlson Comorbidity Index,7 or the Elixhauser Comorbidity Index.8 Recognizing the complex interplay between HF and comorbid conditions, a noncardiac comorbidity was defined as a condition that may not resolve after either LVAD implantation or heart transplantation. The comorbid conditions included peripheral vascular disease, chronic obstructive pulmonary disease, diabetes mellitus (DM), DM with complications (DM and either coronary artery disease, peripheral vascular disease, neurological event, or estimated glomerular filtration rate <60 mL/min per 1.73 m2), connective tissue disease, unintentional weight loss, liver disease, hypertension (blood pressure ≥140/80 mm Hg), obesity (body mass index ≥30 kg/m2), electrolyte disorder (Na ≤135 or ≥145 mEq/L or K <3.5 or >5 mEq/L), anemia (Hgb <13 g/dL for men and <12 g/dL for women), coagulopathy (platelets <150/μL or international normalized ratio >1.2 while not on warfarin), chronic kidney disease (estimated glomerular filtration rate <60 mL/min per 1.73 m2), and depression (Patient Health Questionnaire [PHQ]-8 score ≥10). A comorbidity index of the conditions was created by taking the sum of these noncardiac comorbidities for a total score of up to 14.
Health Status
Patients completed the EQ-5D-3L, which includes the EuroQol Visual Analogue Scale (EQ-VAS) and 5 single-item dimensions of HRQOL for the assessment of generic health status9 and the Kansas City Cardiomyopathy Questionnaire (KCCQ) for disease-specific health status.10 The EQ-VAS and the KCCQ overall summary score (OSS) each range from 0 to 100, and the EQ-5D-3L Index (calculated from the dimension scores) ranges from 0 to 1. Higher scores for each reflect better health status.
Statistical Analysis
Descriptive statistics for categorical and continuous variables are reported. Multivariable general linear models were used to examine the relationship between each HRQOL score and the individual comorbidities and the comorbidity index, with the HRQOL scores as the dependent variables and the individual comorbidities and the comorbidity index (in separate models) as the primary predictors. Assumptions for general linear model were met. Age, sex, race, INTERMACS profile, and 6-minute walk distance were candidate variables for the multivariable models. INTERMACS profiles 4 and 5 were combined due to the small number of profile 4 participants. The final multivariable models were determined by stepwise backward selection using the Akaike information criterion. The contribution of each variable to the overall model (the percentage of the variance explained by the variable) was determined by the semipartial ω2.
Given the well-established relationship between depression and HRQOL among ambulatory HF patients,11 depression was removed from the index, and the analysis was repeated with inclusion of the residual comorbidity index (without depression as a comorbidity) and depression as a separate candidate variable in the multivariable model. To illustrate the association between the comorbidities and the HRQOL measures, the predicted marginal effect of the comorbidities on HRQOL was determined using the least squares mean scores and 95% CIs. The relationship between depression (dependent variable) and INTERMACS levels (as an indicator of HF severity) was assessed using logistic regression. Statistical analyses were performed using SAS, version 9.4 (SAS Institute, Cary, NC).
Results
Demographic characteristics, including the prevalence of comorbidities and HRQOL scores, are depicted in Table 1. Of the 373 ambulatory HF patients enrolled, 32 (9%) were INTERMACS profile 4, 78 (21%) profile 5, 144 (39%) profile 6, and 119 (32%) profile 7. The most common comorbidities were chronic kidney disease (60%), obesity (44%), and DM (38%). Depression was present in 100 (27%) patients. The median (25th to 75th percentile) number of comorbidities in the cohort was 3 (2–4).
| Characteristic | N (%) or Mean±SD or Median (25th to 75th Percentile) | |
|---|---|---|
| Age, y | 60.3±11.3 | |
| Male sex | 279 (74.8) | |
| Race | ||
| Black | 89 (23.9) | |
| White | 262 (70.2) | |
| Other* | 22 (5.9) | |
| Hispanic ethnicity | 26 (7.0) | |
| INTERMACS profile | ||
| 4 | 32 (8.6) | |
| 5 | 78 (20.9) | |
| 6 | 144 (38.6) | |
| 7 | 119 (31.9) | |
| Comorbid conditions | ||
| Peripheral vascular disease | 15 (4.0) | |
| Hypertension | 17 (4.6) | |
| COPD | 46 (12.3) | |
| DM | 140 (37.5) | |
| Complicated DM | 119 (31.9) | |
| Chronic kidney disease | 224 (60.1) | |
| Liver disease | 52 (13.9) | |
| Connective tissue disease | 12 (3.2) | |
| Coagulopathy | 69 (18.5) | |
| Obesity | 164 (44.0) | |
| Unintentional weight loss | 3 (0.8) | |
| Electrolyte disorder | 112 (30.0) | |
| Anemia | 120 (32.2) | |
| Depression symptoms | 100 (26.8) | |
| No. of comorbid conditions (CCI) | 3 (2–4) | |
| 6-min walk distance, m | 341 (280–401) | |
| EQ-VAS (n=367) | 65 (50–75) | |
| EQ-5D Index (n=372) | 0.82 (0.71–0.86) | |
| KCCQ-OSS (n=373) | 64 (48.0–78.0) | |
The results of the multivariable general linear models are depicted in Table 2. The comorbidity index was associated with a reduction in the KCCQ-OSS and the EQ-5D Index. For each additional comorbidity, HRQOL as measured by the EQ-5D Index (scale, 0–1) and KCCQ-OSS (scale, 0–100) was reduced 0.013 units (−0.022 to −0.004; P=0.005) and 2.0 units (−3.1 to −0.8; P=0.001), respectively.
| EQ-VAS | EQ-5D Index | KCCQ-OSS | ||||
|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 1 | Model 2 | Model 1 | Model 2 | |
| Comorbidity index | Not in the final model | … | −0.013* (−0.022 to −0.004) | … | −2.0* (−3.1 to −0.8) | … |
| Residual comorbidity index | … | Not in the final model | … | Not in the final model | … | Not in the final model |
| Depression (vs no depression) | … | −12.5† (−16.7 to −8.2) | … | −0.153† (−0.184 to −0.123) | … | −25.2† (28.9 to −21.6) |
| 6MWD (per 100 m) | 5.2† (3.1 to 7.3) | 4.7† (2.7 to 6.6) | 0.050† (0.034 to 0.066) | 0.044† (0.030 to 0.058) | 5.6† (3.4 to 7.7) | 4.9† (3.2 to 6.7) |
| Age (per 10 y) | 1.6‡ (−0.1 to 3.4) | 0.027† (0.014 to 0.041) | 0.015§ (0.003 to 0.027) | 4.2† (2.4 to 5.9) | 2.3* (0.8 to 3.7) | |
| INTERMACS profile | ||||||
| 4 and 5 | −6.1§ (−11.2 to −1.0) | −11.1† (−16.3 to 6.0) | −7.0* (−11.3 to −2.7) | |||
| 6 | −3.2 (−7.9 to 1.5) | −6.9* (−11.6 to −2.2) | −5.5* (−9.4 to −1.7) | |||
| 7 (ref) | … | … | … | |||
| Black/African American (vs other) | 5.1§ (0.4 to 9.9) | 8.1† (3.4 to 12.8) | 5.6* (1.7 to 9.5) | |||
| Women (vs men) | 4.2‡ (−0.3 to 8.6) | |||||
| Variance explained by the model (ω2) | 0.10† (0.04 to 0.16) | 0.15† (0.09 to 0.22) | 0.15† (0.09 to 0.22) | 0.32† (0.25 to 0.40) | 0.22† (0.15 to 0.30) | 0.48† (0.41 to 0.54) |
After removal of depression from the comorbidity index, the residual index was not associated with reduced HRQOL by any of the HRQOL measures (Table 2). In the final multivariable general linear models, only depression and 6-minute walk distance were associated with worse generic (EQ-VAS and EQ-5D Index) and disease-specific (KCCQ-OSS) HRQOL. HF severity as measured by INTERMACS profile was associated with a reduction in the KCCQ score but not generic HRQOL measures.
The amount of variance in HRQOL explained by the models improved with the addition of depression. Depression was found to have the highest semipartial ω2, the amount of variation in the HRQOL outcome attributable to a variable across all surveys (EQ-VAS, 0.08 [95% CI, 0.03–0.14]; EQ-5D Index, 0.19 [95% CI, 0.12–0.26]; KCCQ-OSS, 0.27 [95% CI, 0.20–0.35]) and contributed greater than half of the variance explained in each model (from Table 2). For example, in the KCCQ-OSS model, depression accounted for 27% (95% CI, 20%–35%) of the 48% (95% CI, 41%–54%) of the variance explained by the multivariable model. Compared with participants without depression, those with depression would be expected to score 12.5 points (95% CI, 8.2–16.7) lower on the EQ-VAS, 0.15 points (95% CI, 0.12–0.18) lower on the EQ-5D Index, and 25.2 points (95% CI, 21.6–28.9) lower on the KCCQ-OSS as determined by the marginal mean effect of depression on each HRQOL survey. Compared with INTERMACS profile 7 participants, the odds of having depression were similar in INTERMACS profile 6 participants (odds ratio, 1.6 [95% CI, 0.9–3.0]; P=0.14) and higher in INTERMACS profile 4 and 5 participants (odds ratio, 3.4; [95% CI, 1.8–6.6]; P<0.001).
Discussion
Noncardiac comorbidities are common in patients with ambulatory advanced HF. While there was an association between the comorbidity index and a reduction in HRQOL, this association was no longer present when depression was removed from the comorbidity index. Depression was highly associated with a reduction in disease-specific and generic HRQOL. Collectively, these findings provide novel insight into our understanding of LVAD candidacy in patients with noncardiac comorbidities. Our study suggests that noncardiac comorbidities (other than depression) that are not severe enough to be a contraindication to LVAD should not impact the potential gains in HRQOL after LVAD implantation. Conversely, HF severity, which would be expected to improve after LVAD implantation, was strongly associated with depression.
Current mechanical circulatory support guidelines support screening for comorbid conditions.12 There is limited guidance about how individual comorbidities, much less multiple comorbidities, should be considered in the evaluation of candidacy for advanced therapies beyond ensuring that a condition is not life-limiting or associated with increased mortality after LVAD implantation. Despite this, the presence of multiple comorbidities has been suggested as a reason why LVAD implantation may be inappropriate in the majority of patients13 and is associated with increased decisional conflict.3 Our study confirms that multiple comorbid conditions are present in ambulatory advanced HF patients and clarifies the relationship between multiple conditions and patient-centered outcomes.
The lack of an association between HRQOL and individual noncardiac comorbidities other than depression, either individually or together, has important implications in ambulatory advanced HF patients who are candidates for LVAD implantation and heart transplantation. Noncardiac comorbid conditions such as chronic obstructive pulmonary disease and DM would be expected to persist after receipt of advanced therapies. Our findings suggest that the presence of the assessed comorbidities, if not severe enough to contraindicate LVAD, does not impact HRQOL. The clinician’s focus should be to prevent the progression of these comorbid conditions to the point that they may negatively impact LVAD candidacy. Notably, the lack of an association between comorbid conditions and HRQOL is likely the result of purposeful selection bias in our cohort, as participants could not have conditions severe enough to limit their functional status or 2-year life expectancy.
Depression, conversely, has consistently been strongly associated with worse HRQOL than other factors in ambulatory HF patients, including New York Heart Association functional class.11 In our study, depression was associated with a profound decrease in both generic and disease-specific HRQOL. The 25-point decrease in the KCCQ-OSS associated with depression in this analysis is 5× greater than the reported minimal clinically important difference in the measure (5 points).14
The current consensus document on psychosocial evaluation of candidates for LVAD recommends a detailed psychosocial screen in all patients being considered for LVAD with a formal consultation with a psychiatrist and treatment as needed.15 The optimal method for prescreening for depression is not certain,15 and there is likely variation in practice. Without identification of depression in patients without a formal diagnosis of depression, patients who may benefit from increased resources and therapies may not receive additional psychosocial support. Given the increasing literature on the association of outcomes and depressive symptoms determined by PHQ surveys, performing a standardized assessment that includes either the 2-item PHQ-2 survey, which if positive leads to the PHQ-8 or 9 survey, or the PHQ-8 or 9 survey upfront followed by referral to a mental health professional if positive should be explored in registries and clinical practice.
Our study has limitations. First, the comorbidity index was unweighted, and it is possible that the impact of different comorbidities on HRQOL would differ, although this method has frequently been used to assess the impact of multiple comorbid conditions on outcomes.16 Second, the presence of depression was determined from the results of the PHQ-8. While the PHQ-8 does not replace a formal psychiatric evaluation and diagnosis of depression, using a screen has been employed in prior HF studies and removes the potential subjectivity of a depression diagnosis.17
Our findings suggest that while noncardiac comorbid conditions are common in ambulatory advanced HF patients, depression is uniquely associated with decreased generic and disease-specific HRQOL. The presence of multiple noncardiac comorbid conditions other than depression, if not severe enough to otherwise contraindicate LVAD, should not impact expected gains in HRQOL after LVAD implantation and should not affect candidacy for advanced therapies. Future research may focus on whether the association of depression and reduced HRQOL is modifiable in this population. Furthermore, a greater understanding of the impact of depression symptoms on postimplant HRQOL may better inform discussions on the potential benefits of LVAD in ambulatory HF patients. In the meantime, improved screening and best evidence-based practice for the management of depression should be emphasized.
| DM |
diabetes mellitus |
| EQ-VAS |
EuroQol Visual Analogue Scale |
| HF |
heart failure |
| HRQOL |
health-related quality of life |
| INTERMACS |
Interagency Registry for Mechanically Assisted Circulatory Support |
| KCCQ |
Kansas City Cardiomyopathy Questionnaire |
| LVAD |
left ventricular assist device |
| OSS |
overall summary score |
| PHQ |
Patient Health Questionnaire |
| REVIVAL |
Registry Evaluation of Vital Information for Ventricular Assist Devices in Ambulatory Life |
Acknowledgments
We thank the support by funding from the National Institutes of Health, the National Heart, Lung, and Blood Institute (contract No. HHSN268201100026C) for REVIVAL (Registry Evaluation of Vital Information for Ventricular Assist Devices in Ambulatory Life), and the National Center for Advancing Translational Sciences (grant No. UL1TR002240) for the Michigan Institute for Clinical and Health Research.
Sources of Funding
This study was supported by funding from the National Institutes of Health, the National Heart, Lung, and Blood Institute (contract No. HHSN268201100026C), and the National Center for Advancing Translational Sciences (grant No. UL1TR002240) for the Michigan Institute for Clinical and Health Research. The views expressed in this article are those of the authors and do not necessarily represent the views of the National Heart, Lung, and Blood Institute; the National Institutes of Health; or the US Department of Health and Human Services.
Disclosures
Dr Stehlik reports personal consulting fees from Medtronic and Abbott. Dr Ewald reports personal consulting fees from Abbott. Dr Horstmanshof reports speaker/research grant for Abbott Medical. Dr Shah reports personal fees from NuPulse CV and Ortho Clinical Diagnostics and grants from American Heart Association (AHA)/Enduring Hearts, Merck & Co, Medtronic, and Haemonetics. Dr Jorde reports serving as a nonpaid consultant for Abbott. Dr Grady reports grant and personal fees from National Institutes of Health (NIH)/National Heart, Lung, and Blood Institute (NHLBI); grants from NIH/NHLBI, NIH/National Institute on Aging, and NIH/National Institute of Nursing Research; and nonfinancial support from International Society for Heart and Lung Transplantation (for membership on International Society for Heart and Lung Transplantation Board of Directors), AHA, and Heart Failure Society of America. Dr Aaronson reports consultant and contracted clinical research from Medtronic; contracted clinical research from Abbott; belongs to the Scientific Advisory Board at Procyrion; and is a consultant for NuPulse CV. The other authors report no conflicts.
Footnotes
References
- 1.
Ambardekar AV, Kittleson MM, Palardy M, Mountis MM, Forde-McLean RC, DeVore AD, Pamboukian SV, Thibodeau JT, Teuteberg JJ, Cadaret L . Outcomes with ambulatory advanced heart failure from the Medical Arm of Mechanically Assisted Circulatory Support (MedaMACS) registry. J Heart Lung Transplant. 2019; 38:408–417. doi: 10.1016/j.healun.2018.09.021CrossrefMedlineGoogle Scholar - 2.
Kormos RL, Cowger J, Pagani FD, Teuteberg JJ, Goldstein DJ, Jacobs JP, Higgins RS, Stevenson LW, Stehlik J, Atluri P . The Society of Thoracic Surgeons Intermacs database annual report: evolving indications, outcomes, and scientific partnerships. J Heart Lung Transplant. 2019; 38:114–126. doi: 10.1016/j.healun.2018.11.013CrossrefMedlineGoogle Scholar - 3.
Warraich HJ, Allen LA, Blue LJ, Chaussee EL, Thompson JS, McIlvennan CK, Flint KM, Matlock DD, Patel CB . Comorbidities and the decision to undergo or forego destination therapy left ventricular assist device implantation: an analysis from the Trial of a Shared Decision Support Intervention for Patients and Their Caregivers Offered Destination Therapy for End-Stage Heart Failure (DECIDE-LVAD) study. Am Heart J. 2019; 213:91–96. doi: 10.1016/j.ahj.2019.04.008MedlineGoogle Scholar - 4.
Joyce E, Chung C, Badloe S, Odutayo K, Desai A, Givertz MM, Nohria A, Lakdawala NK, Stewart GC, Young M . Variable contribution of heart failure to quality of life in ambulatory heart failure with reduced, better, or preserved ejection fraction. JACC Heart Fail. 2016; 4:184–193. doi: 10.1016/j.jchf.2015.12.011CrossrefMedlineGoogle Scholar - 5.
Aaronson KD, Stewart GC, Pagani FD, Stevenson LW, Palardy M, McNamara DM, Mancini DM, Grady K, Gorcsan J, Kormos R ; REVIVAL Investigators. Registry Evaluation of Vital Information for VADs in Ambulatory Life (REVIVAL): rationale, design, baseline characteristics, and inclusion criteria performance. J Heart Lung Transplant. 2020; 39:7–15. doi: 10.1016/j.healun.2019.09.008CrossrefMedlineGoogle Scholar - 6. STS Intermacs. Appendix M. STS Intermacs® USERS’ GUIDE. 2018, June 28; 2019. Available at: https://www.uab.edu/medicine/intermacs/images/Intermacs-Users-Guide-v5.0-2018-06-28.pdf. Accessed July 12, 2019.Google Scholar
- 7.
Charlson ME, Pompei P, Ales KL, MacKenzie CR . A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987; 40:373–383. doi: 10.1016/0021-9681(87)90171-8CrossrefMedlineGoogle Scholar - 8.
Elixhauser A, Steiner C, Harris DR, Coffey RM . Comorbidity measures for use with administrative data. Med Care. 1998; 36:8–27. doi: 10.1097/00005650-199801000-00004CrossrefMedlineGoogle Scholar - 9. EuroQol Group. EuroQol a new facility for the measurement of health-related quality of life. Health Policy. 1990; 16:199–208.CrossrefMedlineGoogle Scholar
- 10.
Green CP, Porter CB, Bresnahan DR, Spertus JA . Development and evaluation of the Kansas City Cardiomyopathy Questionnaire: a new health status measure for heart failure. J Am Coll Cardiol. 2000; 35:1245–1255. doi: 10.1016/s0735-1097(00)00531-3CrossrefMedlineGoogle Scholar - 11.
Baert A, De Smedt D, De Sutter J, De Bacquer D, Puddu PE, Clays E, Pardaens S . Factors associated with health-related quality of life in stable ambulatory congestive heart failure patients: Systematic review. Eur J Prev Cardiol. 2018; 25:472–481. doi: 10.1177/2047487318755795CrossrefMedlineGoogle Scholar - 12.
Feldman D, Pamboukian SV, Teuteberg JJ, Birks E, Lietz K, Moore SA, Morgan JA, Arabia F, Bauman ME, Buchholz HW ; International Society for Heart and Lung Transplantation. The 2013 International Society for heart and lung transplantation guidelines for mechanical circulatory support: executive summary. J Heart Lung Transplant. 2013; 32:157–187. doi: 10.1016/j.healun.2012.09.013CrossrefMedlineGoogle Scholar - 13.
Allen LA, Stevenson LW, Grady KL, Goldstein NE, Matlock DD, Arnold RM, Cook NR, Felker GM, Francis GS, Hauptman PJ ; American Heart Association; Council on Quality of Care and Outcomes Research; Council on Cardiovascular Nursing; Council on Clinical Cardiology; Council on Cardiovascular Radiology and Intervention; Council on Cardiovascular Surgery and Anesthesia. Decision making in advanced heart failure: a scientific statement from the American Heart Association. Circulation. 2012; 125:1928–1952. doi: 10.1161/CIR.0b013e31824f2173LinkGoogle Scholar - 14.
Spertus J, Peterson E, Conard MW, Heidenreich PA, Krumholz HM, Jones P, McCullough PA, Pina I, Tooley J, Weintraub WS ; Cardiovascular Outcomes Research Consortium. Monitoring clinical changes in patients with heart failure: a comparison of methods. Am Heart J. 2005; 150:707–715. doi: 10.1016/j.ahj.2004.12.010CrossrefMedlineGoogle Scholar - 15.
Dew MA, DiMartini AF, Dobbels F, Grady KL, Jowsey-Gregoire SG, Kaan A, Kendall K, Young QR, Abbey SE, Butt Z . The 2018 ISHLT/APM/AST/ICCAC/STSW recommendations for the psychosocial evaluation of adult cardiothoracic transplant candidates and candidates for long-term mechanical circulatory support. J Heart Lung Transplant. 2018; 37:803–823. doi: 10.1016/j.healun.2018.03.005CrossrefMedlineGoogle Scholar - 16.
de Groot V, Beckerman H, Lankhorst GJ, Bouter LM . How to measure comorbidity. A critical review of available methods. J Clin Epidemiol. 2003; 56:221–229. doi: 10.1016/s0895-4356(02)00585-1CrossrefMedlineGoogle Scholar - 17.
Starling RC, Estep JD, Horstmanshof DA, Milano CA, Stehlik J, Shah KB, Bruckner BA, Lee S, Long JW, Selzman CH ; ROADMAP Study Investigators. Risk assessment and comparative effectiveness of left ventricular assist device and medical management in ambulatory heart failure patients: the ROADMAP study 2-year results. JACC Heart Fail. 2017; 5:518–527. doi: 10.1016/j.jchf.2017.02.016CrossrefMedlineGoogle Scholar