Author + information
- Received April 27, 2016
- Revision received July 1, 2016
- Accepted July 5, 2016
- Published online October 4, 2016.
- Omar Wever-Pinzon, MDa,b,c,
- Stavros G. Drakos, MD, PhDa,b,c,
- Stephen H. McKellar, MD, MSca,c,d,
- Benjamin D. Horne, MPH, PhDe,
- William T. Caine, MDa,e,
- Abdallah G. Kfoury, MDa,e,
- Dean Y. Li, MD, PhDa,b,c,
- James C. Fang, MDa,b,c,
- Josef Stehlik, MD, MPHa,b,c and
- Craig H. Selzman, MDa,c,d,∗ ()
- aUtah Cardiac Recovery Program, Salt Lake City, Utah
- bDivision of Cardiology, University of Utah School of Medicine, Salt Lake City, Utah
- cVeterans Affairs Medical Center, Salt Lake City, Utah
- dDivision of Cardiothoracic Surgery, University of Utah School of Medicine, Salt Lake City, Utah
- eIntermountain Medical Center, Murray, Utah
- ↵∗Reprint requests and correspondence:
Dr. Craig H. Selzman, Division of Cardiothoracic Surgery, University of Utah Health Sciences Center, 30 North 1900 East, SOM 3C 127, Salt Lake City, Utah 84132.
Background The number of centers with left ventricular assist device (LVAD) research programs focused on cardiac recovery is very small. Therefore, this phenomenon has been reported in real-world multi-center registries as a rare event.
Objectives This study evaluated the incidence of cardiac recovery with an a priori LVAD implantation strategy of bridge-to-recovery (BTR) and constructed a recovery predictive model.
Methods The study included LVAD recipients registered in the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS). Cardiac recovery was evaluated in BTR and non-BTR patients. A weighted score was derived and externally validated in patients of the Utah Cardiac Recovery (UCAR) program.
Results Of 15,138 INTERMACS patients, cardiac recovery occurred in 192 (1.3%). The incidence of recovery was 11.2% (n = 14) in BTR compared with 1.2% (n = 178) in non-BTR patients (p < 0.0001). Independent predictors of recovery included: age <50 years, non-ischemic cardiomyopathy, time from cardiac diagnosis <2 years, absence of ICD, creatinine ≤1.2 mg/dl, and LVEDD <6.5 cm (c-index: 0.85; p < 0.0001). A weighted score termed I-CARS, effectively stratified patients based on their probability of recovery. I-CARS was validated in the UCAR cohort (n = 190) with good performance (AUC: 0.94; 95% CI: 0.91 to 0.98). One-year survival after LVAD explantation, available in INTERMACS for 21 (11%) patients, was 86%.
Conclusions The incidence of cardiac recovery is higher in patients implanted with an a priori BTR strategy. We developed a simple tool to help identify patients in whom recovery is feasible. In BTR patients with favorable characteristics, I-CARS suggests a 24% probability of successful LVAD explantation. Large-scale studies to better address post-explantation outcomes are warranted.
Left ventricular assist devices (LVADs) are a standard therapeutic option for patients with advanced heart failure (HF) refractory to medical therapy (1,2), whether used for patients awaiting cardiac transplantation or as definitive therapy for those ineligible for transplantation (3,4). Heart transplantation and chronic mechanical support with LVADs improve survival, symptoms, exercise tolerance, and quality of life in this population (4–9). However, these treatment strategies are complex and associated with adverse effects (4,10–13) that impose a significant burden on patients and negatively affect resource utilization and overall public health (14–16).
Because the usual intent with an LVAD is eventual heart transplantation or lifetime support, research efforts have been focused mainly on improving device safety and performance. Consequently, clinicians and investigators have devoted less time to understanding what happens to the supported heart. Indeed, the mechanically unloaded heart can experience reverse remodeling and significantly improved myocardial function, with the potential for cardiac recovery and device explantation, thereby avoiding transplantation or long-term mechanical circulatory support (17–21). Although retrospective studies suggest a low incidence of myocardial recovery (<1% to 2%) (22,23), the lack of pre-specified protocols to monitor and promote structural and functional changes that lead to cardiac recovery and LVAD explantation calls into question the accuracy of these data (24,25).
LVAD unloading of the failing heart is associated with unique genomic, molecular, cellular, and structural changes that can be linked with functional heart improvement (26–28). To date, few clinical and molecular predictors of myocardial recovery have been identified (25,28). The objective of the current study was to evaluate the incidence of cardiac recovery and patient characteristics according to LVAD implantation strategy (i.e., an a priori bridge-to-recovery [BTR] strategy vs. non-BTR strategy) in a large cohort of patients enrolled in a national registry. In addition, we aimed to construct and validate a predictive model of cardiac recovery leading to LVAD explantation.
Materials and Methods
The design, structure, and goals of the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS), a North American registry of patients receiving mechanical circulatory support device therapy, have been published elsewhere (29). Additional information can be found in the Online Appendix.
De-identified data of patients registered in INTERMACS were provided to the authors. The study analysis included adult (i.e., >18 years of age) LVAD recipients registered in INTERMACS between March 2006 and June 2015. We excluded patients who received a right ventricular assist device without an LVAD, a total artificial heart, or prior heart transplantation. Patients with limited follow-up (<30 days), complex congenital heart disease, and hypertrophic or other forms of restrictive cardiomyopathy were excluded from the competing-risk regression models built to assess the association of patient-related factors with cardiac recovery to eliminate a group of patients with different potential for cardiac recovery. The INTERMACS dataset did not provide detailed information to identify and exclude from the regression model patients with acute forms of HF (e.g., acute myocardial infarction, acute myocarditis). Use of these data was deemed exempt from human subject approval by our institutional review boards.
The Utah Cardiac Recovery (UCAR) program has prospectively enrolled patients who required circulatory support with a continuous-flow LVAD (25). The program prospectively collects longitudinal clinical, hemodynamic, and cardiac imaging data before and after LVAD implantation to better define the structural and functional effects of mechanical unloading on the failing myocardium. We included adult LVAD recipients enrolled from July 2008 to April 2015 and again excluded patients with acute forms of HF, hypertrophic or infiltrative cardiomyopathies, and complex congenital heart disease. Informed consent was obtained from all patients, and the institutional review board of the participating institutions approved the study.
Baseline characteristics were summarized by using standard statistical descriptors and compared by using the Mann-Whitney U test or Pearson chi-square test as appropriate. The primary outcome was the incidence of cardiac recovery, defined as LVAD explantation due to myocardial recovery. Patients who experienced myocardial recovery but had their pump deactivated and left in situ without device removal were also included in the primary outcome. Other outcomes included mortality on LVAD support, transplantation, and LVAD explantation due to nonrecovery reasons. Time to event was the time from LVAD implantation until cardiac recovery, death on LVAD support, transplantation, LVAD explantation due to nonrecovery reasons, or last follow-up in patients who remained on LVAD support.
Outcomes were evaluated in the entire derivation cohort and subgroups of patients on the basis of LVAD implantation strategy: BTR and non-BTR that included bridge-to-transplant, bridge-to-candidacy, destination therapy, and rescue therapy. Survival curves were constructed to illustrate multiple possible outcomes at any time after LVAD implantation: cardiac recovery, death on LVAD support, transplantation, and LVAD explantation due to nonrecovery reasons. Cumulative incidence function curves were estimated by means of the Fine and Gray method and compared by using the Gray test (30,31). Univariable and multivariable proportional hazards models were used to assess predictors for cardiac recovery. The proportional hazards assumption was confirmed by testing covariate interactions with quadratic function of time and checked graphically by using Schoenfeld-type residuals. The Fine and Gray method was used to build the models and generate subhazard ratios and 95% confidence intervals (CIs) to account for death on LVAD support, transplantation, or LVAD explantation due to nonrecovery reasons as competing risk events for cardiac recovery. The models examined the association of factors or characteristics present at time of LVAD implantation with cardiac recovery. A multivariable competing-risk regression model was used to determine the independent association of multiple patient-related factors on the 5-year hazard of cardiac recovery (follow-up was censored after 5 years). Variables presumed to be potentially associated with cardiac recovery based on literature review and clinical judgment, as well as those significant at the p < 0.10 level in unadjusted analyses, were included in the multivariable analysis. Only variables significant at the p < 0.05 level on the basis of the likelihood ratio test were retained in the final model. A 2-tailed p value <0.05 was considered statistically significant. Patients from the UCAR program were excluded from the INTERMACS derivation cohort when building the prognostic scoring model.
A practical prognostic scoring system was developed (using methods described in the Online Appendix), and patients were divided into categories of low, intermediate, or high probability of cardiac recovery. The predictive discrimination of the scoring system was examined by calculating the Harrell’s concordance (C) statistic. The prognostic scoring system was externally validated in an independent cohort of patients from the UCAR program (validation cohort). Analyses were performed by using STATA version 14 (StataCorp LP, College Station, Texas).
There were 15,631 adults enrolled between March 2006 and June 2015. We excluded patients who received a right ventricular assist device without an LVAD (n = 111), a total artificial heart (n = 355), or a previous heart transplant (n = 27). In total, 15,138 patients met the inclusion criteria and comprised the INTERMACS cohort (Online Figure 1).
The distribution and characteristics of the LVAD recipients, stratified according to implantation strategy, are summarized in Table 1. Although between-group comparisons for most baseline characteristics were statistically significant, several elements were particularly relevant. Patients who had an LVAD implanted with an a priori BTR strategy were more likely to be young female subjects with lower body mass index who more often had nonischemic cardiomyopathy and a shorter time since their cardiac diagnosis. This group was less likely to have an implantable cardioverter-defibrillator (ICD) or chronic kidney disease, and had lower left ventricular end-diastolic dimensions (LVEDDs). Finally, these patients had lower INTERMACS profiles (higher acuity), were more likely to require dialysis during the index admission or require a biventricular assist device and less likely to require an intra-aortic balloon pump; they also had lower levels of serum creatinine and B-type natriuretic peptides. Interestingly, they were more likely to receive a pulsatile-flow LVAD; those receiving a continuous-flow device were more likely to receive an axial-flow LVAD.
After a median of 323.3 days (range 176.9 to 656.4 days), 192 (1.3%) of the total INTERMACS patient population experienced cardiac recovery leading to LVAD explantation (n = 172) or device deactivation (n = 20). Nonrecovery reasons for device explantation included device malfunction (n = 528), device infection (n = 116), and device thrombosis (n = 620). A total of 202 patients underwent device explantation or deactivation for other reasons (likely for palliative purposes). The hazard estimates and cumulative incidence function of cardiac recovery in this cohort are shown in Online Figure 2 and Figure 1A, respectively. Cardiac recovery occurred infrequently early after LVAD implantation. Among patients who experienced cardiac recovery (n = 192), only 4 (2.1%) did so in the first month, but recovery frequency increased over time: 28 (14.6%) patients experienced cardiac recovery by 3 months, with 80% of recoveries occurring by 2 years’ post-LVAD implantation (Online Figure 3).
There was no significant variation in the incidence of cardiac recovery among centers with different implant volumes (Online Table 1). LVAD recipients in whom cardiac recovery occurred had evidence of reverse remodeling based on longitudinal improvement in left ventricular ejection fraction (LVEF) and reduction in LVEDD compared with patients with no cardiac recovery (Online Figure 4). Of significance, there were 7,084 patients who had at least 1 follow-up echocardiogram with LVEF assessment after 3 months of LVAD support. In this group, 892 (12.6%) patients achieved an LVEF ≥40%, with a relative increase in LVEF ≥50%. Of these patients who experienced a favorable cardiac response, 97% underwent implantation with the use of a non-BTR strategy.
The incidence of cardiac recovery varied widely depending on the implantation strategy (Figure 1B). Cardiac recovery was observed in 14 (11.2%) patients with an a priori strategy of BTR, corresponding to an incidence rate of 7.3 events per 100 person-years, versus 178 (1.2%) patients with a non-BTR strategy, corresponding to an incidence rate of 0.9 event per 100 person-years (p < 0.0001). Nonrecovery-related explants were also more common in the BTR strategy group compared with the non-BTR strategy group (6.8 vs. 3.6 events per 100 person-years; p = 0.04). Analysis of the different outcomes experienced by LVAD recipients revealed varying rates at 12 and 36 months after implantation (Table 2, Figures 2A to 2C).
To better characterize patients considered to have a good potential for cardiac recovery, we compared the characteristics of a priori BTR strategy patients who experienced cardiac recovery versus those who did not. Successfully explanted patients tended to be younger, were more likely women of Hispanic ethnicity, had lower body mass index and shorter duration of HF, were less likely to have an ICD or previous cardiac surgeries, and were more symptomatic (Online Table 2).
The INTERMACS cardiac recovery score
The assessment of the clinical characteristics of patients who experienced cardiac recovery showed that, with few exceptions, this group was fairly homogeneous, independent of the LVAD implantation strategy (BTR vs. non-BTR strategy) (Table 3). However, there were marked differences in the baseline characteristics of patients with and without cardiac recovery. We performed a competing-risks regression analysis in the INTERMACS cohort that excluded patients with follow-up <30 days (n = 186), complex congenital (n = 81) and restrictive (n = 289) heart disease, and patients from the UCAR program (n = 247). Follow-up was censored at 5 years’ post-LVAD implantation. A total of 14,338 patients (185 cardiac recovery events) were analyzed.
Among the characteristics present at time of LVAD implantation that were included in our analyses, multiple univariable predictors of cardiac recovery were identified (Online Table 3). Of significance, cardiac recovery did not occur in patients with complex congenital and restrictive heart disease. After multivariable adjustment, 6 independent predictors of cardiac recovery were identified: age <50 years, nonischemic cardiomyopathy, time from cardiac diagnosis <2 years, absence of an ICD, serum creatinine level ≤1.2 mg/dl, and LVEDD <6.5 cm (Table 4). The discrimination of the model for cardiac recovery was good, with a C-statistic of 0.85 (95% CI: 0.82 to 0.88; p < 0.0001).
On the basis of these results, a prognostic score we labeled the INTERMACS Cardiac Recovery Score (I-CARS) was derived by assigning each of the 6 prognostic variables a number of points proportional to its regression coefficient (Table 4). A score was determined for each patient by calculating a sum of the points corresponding to his or her risk factors. The score ranged from 0 to 9, and the median was 3 (interquartile range: 1 to 5). Cardiac recovery estimates were used to define 3 groups with significantly different prognoses: a low probability group (0 to 3 points), an intermediate probability group (4 to 6), and a high probability group (7 to 9). The corresponding cardiac recovery rates in the INTERMACS cohort were 0.2%, 1.4%, and 8.9% (Figure 3A). Interestingly, applying I-CARS to the BTR strategy group yielded cardiac recovery rates of 0%, 4.9%, and 24.5% for patients in the low, intermediate, and high probability categories, respectively. Cardiac recovery rates stratified according to I-CARS in the non-BTR strategy group were similar to rates in the main cohort. The performance of I-CARS was good, with an area under the receiver-operating characteristics curve (AUC) of 0.84 (95% CI: 0.81 to 0.87) (Figure 3B). The score distribution of I-CARS stratified according to implantation strategy and the presence of cardiac recovery are shown in Online Figures 5 and 6, respectively.
Cardiac recovery occurred infrequently early after LVAD implantation and increased in frequency over time (Online Figure 3). Thus, including patients who did not have the opportunity to disclose their cardiac recovery potential could affect our results. To address this question, we excluded patients who died (n = 1,396) or underwent transplantation (n = 609) within the first 3 months’ post-LVAD implantation. Interestingly, 5 patients with an a priori BTR implantation strategy underwent transplantation and another 10 patients in this group died within the first 3 months after implantation. Cardiac recovery rates conditional to survival on an LVAD 3 months after implantation were 1.5% in the entire cohort, 1.4% with the non-BTR strategy, and 13.1% in the BTR strategy group. Univariable and multivariable predictors of cardiac recovery conditional to 3-month survival after LVAD implantation did not differ significantly from the original analysis (data not shown).
HF history and cardiac recovery
The potential for cardiac recovery is greater in patients with acute forms of HF (e.g., acute myocarditis or after an acute myocardial infarction); including these patients may therefore result in an overestimation of cardiac recovery and could affect the predictors associated with this outcome. To address this issue, patients who had HF for <1 month and who had normal LVEDD (≤5.8 cm in men and ≤5.2 cm in women) were excluded (32). A total of 228 patients were excluded, including 15 LVAD recipients who experienced cardiac recovery. The cardiac recovery rate in this cohort was 1.5%, and predictors of cardiac recovery did not differ significantly from the original analysis (Online Table 4).
The use of antiremodeling agents was common in LVAD patients (Online Table 5). Patients who experienced cardiac recovery were less frequently taking a beta-blocker, angiotensin-converting enzyme (ACE) inhibitor/angiotensin-receptor blocker (ARB), or aldosterone receptor antagonist before LVAD implantation. After LVAD implantation, compared with those with no recovery, patients with cardiac recovery were more frequently taking a beta-blocker (12 months: 95% vs. 77%; p < 0.01), an ACE inhibitor or ARB (12 months: 80% vs. 53%; p < 0.01), or an aldosterone receptor blocker (12 months: 49% vs. 34%; p = 0.05). The use of beta-blockers, ACE inhibitors or ARBs, and mineralocorticoid receptor antagonists was associated with cardiac recovery on univariable analysis (Online Table 3). Patients with cardiac recovery were less likely to be taking a diuretic compared with patients without recovery.
Within the UCAR program, 206 patients underwent LVAD implantation during the study period. We prospectively excluded subjects with acute forms of HF (n = 13) and hypertrophic or infiltrative cardiomyopathies (n = 3). The remaining 190 subjects with chronic advanced cardiomyopathy formed the UCAR validation cohort. Baseline characteristics (Online Table 6) had a median age of 60 years (range 48 to 68 years), 83% were men, 71% were in New York Heart Association functional class IV, median duration of symptoms was 5 years, and median LVEDD was 6.7 cm (range 6.2 to 7.3 cm). Nearly two-thirds of the patients were inotrope dependent and of higher acuity according to their INTERMACS profile. Nonischemic cardiomyopathy was the most frequent etiology (56%).
In the validation cohort, the prognostic value of I-CARS was investigated in 2 populations: first, patients who had an improvement in LVEF ≥50% relative to baseline and a final LVEF ≥40% (cut-points selected based on our previous research) (25), herein termed “cardiac responders” (n = 27); and second, patients within this group who had their LVAD explanted due to “cardiac recovery” (n = 11). The evaluation protocol for LVAD explantation used is described in the Online Appendix.
I-CARS was used to stratify patients in the validation cohort on the basis of their probability of cardiac response and cardiac recovery. Patients in the high probability category experienced significant improvement in myocardial function, with 50% experiencing cardiac response and 39% experiencing cardiac recovery (Figure 4A). I-CARS demonstrated a good performance in discriminating cardiac response (AUC: 0.72; 95% CI: 0.59 to 0.85) (Figure 4B) and cardiac recovery (AUC: 0.94; 95% CI: 0.91 to 0.98) (Figure 4C) in the UCAR validation cohort.
In addition, I-CARS was able to stratify patients in the INTERMACS cohort (7,084 patients with echocardiographic follow-up) on the basis of their probability of cardiac response, which occurred in 10%, 13%, and 29% of patients in the low, intermediate, and high probability groups, respectively, albeit with a lower discrimination power (AUC: 0.58; 95% CI: 0.56 to 0.60).
In the INTERMACS cohort, 1-year follow-up post-LVAD explantation was available in 21 patients. Of these patients, 18 were alive, 1 died, and 2 underwent transplantation. In the UCAR program, after a median follow-up of 271 days (interquartile range: 182 to 442 days) after LVAD explantation (n = 11), 3 patients had recurrent HF (1 patient died, 1 patient underwent transplantation, and the other patient remains stable on HF medical therapy).
The maladaptive response to volume and pressure overload leading to myocardial remodeling has long been implicated in the progressive myocardial dysfunction observed in HF (33,34). Left ventricular unloading by mechanical support has the potential to disrupt the vicious cycle of volume and pressure overload leading to cardiac remodeling. Several studies have shown reverse remodeling in LVAD-supported patients, in some sufficient to allow for explantation of the device (17,18,22,35–40). These observations have opened the gateway to the field of LVAD-induced cardiac recovery. However, the incidence of cardiac recovery observed in these studies has varied widely, thus promoting confusion and disbelief that threatens the progress of this field.
INTERMACS, which has captured data on >15,000 US Food and Drug Administration–approved LVAD implants in the United States, provides a robust opportunity to examine cardiac recovery and LVAD explantation. In this contemporary analysis, we confirmed, in general terms, that cardiac recovery leading to pump removal is an infrequently reported event. Our study’s main finding, however, is that once a patient is identified as a potential candidate for cardiac recovery by an LVAD team (i.e., a priori designation of BTR), the incidence of cardiac recovery increases 9-fold. This finding begs the question: why is the incidence of cardiac recovery higher in the BTR strategy patients compared with patients bridged to transplant or destination therapy?
The identification of patient characteristics associated with a higher probability of cardiac recovery could partially explain differences in recovery rates observed between patients bridged with an a priori BTR strategy versus a non-BTR strategy. Patient characteristics and the study population’s diversity in its propensity for cardiac recovery have shown a significant impact on the rates of cardiac recovery in previous BTR studies (28). INTERMACS patients implanted with an a priori BTR strategy were younger, and more frequently had nonischemic cardiomyopathy and shorter duration of HF; these characteristics have previously been linked to a higher probability of cardiac recovery (23,25). Additional factors, including post-LVAD management, likely contributed to the observed differences in cardiac recovery between BTR and non-BTR patients. This is suggested by the fact that the incidence of cardiac recovery in BTR patients was higher than in non-BTR patients, even when patients have a similar cardiac recovery profile: 25% versus 8% for I-CARS ≥7 and 5% versus 1% for I-CARS between 4 and 6 (Central Illustration).
In terms of post-implant management, most LVAD centers do not have established clinical or research BTR programs, nor do they have protocols to serially monitor the structural and functional myocardial changes in these patients that ultimately could lead to cardiac recovery. As a result, unless a patient is considered to have the potential for cardiac recovery (an a priori BTR strategy), the occurrence of cardiac recovery will be underappreciated (24). It is also possible that uncertainty regarding outcomes after LVAD explantation and inexperience of LVAD teams in managing these patients could affect the decision to proceed with device removal. For instance, device explantation due to cardiac recovery occurred in 1.3% of patients in the INTERMACS cohort, whereas a favorable “cardiac response” occurred in an additional 12.6% of patients who ultimately did not have their device explanted. Considering that 97% of the “cardiac responders” were not explanted and were implanted by using a non-BTR strategy, a change in strategy would potentially raise the incidence of cardiac recovery to approximately 14%.
Similarly, most of these patients are not subjected to rigorous LVAD weaning protocols, with predefined LVAD explantation criteria and protocols for adjuvant drug therapy that have been shown to be helpful in promoting cardiac recovery, although such protocols are unlikely unless an interest in cardiac recovery exists (17–19,36,41). However, a comprehensive evaluation of such recovery protocols to support this hypothesis was not possible from this registry.
Regarding pharmacotherapy in the INTERMACS cohort, there were no differences in the use of antiremodeling agents after LVAD implantation between BTR and non-BTR strategy patients (data not shown). However, patients experiencing cardiac recovery were more often undergoing adjuvant drug therapy compared with patients who did not recover (Online Table 5). The probability of cardiac recovery was 55% higher with the use of mineralocorticoid receptor antagonists and approximately 250% higher with the use of beta-blockers and ACE inhibitors or ARBs.
Interestingly, patients who experienced cardiac recovery were less frequently receiving beta-blockers and other neurohormonal inhibitors at the time of LVAD placement, compared with patients without recovery. This finding was shown to be a predictor of reverse remodeling and cardiac recovery (42).
Although the identification of several characteristics likely to enhance the probability of cardiac recovery in a given patient may have contributed to the higher incidence of recovery in the BTR patients, better characterization of these patients is needed to more effectively target efforts and maximize results. For instance, BTR patients who failed to recover tended to be older, had a longer duration of HF history (40% with a duration >2 years), and more often had previous cardiac surgeries (Online Table 2).
To further refine the “phenotype” of the ideal candidate for cardiac recovery, we developed a simple clinical predictive model (i.e., the I-CARS) that effectively stratified patients in the derivation cohort on the basis of probability of cardiac recovery. More importantly, in those patients implanted with a BTR strategy, a high probability score suggested a 24% probability of cardiac recovery. We validated I-CARS in an independent cohort, showing good discrimination and a 39% recovery rate in patients with a high probability score. (The Online Appendix presents an I-CARS calculator.) Sustainability of cardiac recovery after LVAD explantation is obviously a major concern. The Berlin group experience showed a 3-year freedom from HF recurrence of 69% after LVAD explantation (43). Similarly encouraging outcomes have also been reported by the Harefield program, in which post-explantation survival was similar to post-transplantation survival (44). Outcomes data after LVAD explantation in the INTERMACS are limited (11% of explanted patients) and, as such, should be viewed with caution. The preliminary experience in our program after device removal was favorable; nevertheless, results from larger scale, prospective studies addressing this question are needed.
This study was based on a retrospective analysis of a national registry with the known limitations of this type of study design, including lack of data on patient selection, monitoring and weaning protocols, and explantation criteria. Importantly, data on outcomes after LVAD explantation are limited. In addition, some patients were excluded from the multivariable analysis and predictive model due to missing data, which could have introduced a selection bias. We also dichotomized variables, a strategy that simplifies creation of a risk score but has the potential of providing less refined information.
LVAD-induced cardiac recovery is a real and underrecognized phenomenon. Its incidence was higher in patients for whom their LVAD was implanted with an a priori intent of cardiac recovery. I-CARS, a simple predictive score system created by using 6 clinical factors, predicted individual probabilities of cardiac recovery and may help in directing patient selection and adjuvant therapy approaches. Further validation of this bridge-to-pump removal strategy has significant patient and fiscal impacts by potentially reducing resource utilization and morbidity associated with long-term mechanical support and transplantation in patients likely to achieve cardiac recovery.
COMPETENCY IN MEDICAL KNOWLEDGE: Sufficient recovery of cardiac function to allow device removal occurs in only a small fraction of patients with LVADs. However, the odds improve substantially for those <50 years of age with a diagnosis of nonischemic cardiomyopathy within 2 years and preserved renal function in whom an automatic defibrillator has not been implanted and the left ventricle has not dilated beyond 6.5 cm.
TRANSLATIONAL OUTLOOK: Studies evaluating biomarkers and molecular and metabolic predictors of cardiac recovery in conjunction with clinical predictors should be undertaken.
For an expanded Methods section as well as tables and figures, and I-CARS calculator, please see the online version of this article.
Dr. Drakos has received research support from Abiomed Inc.; and has served as consultant for HeartWare. Dr. Stehlik has received research support from St. Jude Medical; and Speakers Bureau fees from HeartWare. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- angiotensin-converting enzyme
- angiotensin-receptor blocker
- area under the receiver-operating characteristics curve
- confidence interval
- heart failure
- INTERMACS Cardiac Recovery Score
- implantable cardioverter-defibrillator
- Interagency Registry for Mechanically Assisted Circulatory Support
- left ventricular assist device
- left ventricular end-diastolic dimension
- left ventricular ejection fraction
- Utah Cardiac Recovery
- Received April 27, 2016.
- Revision received July 1, 2016.
- Accepted July 5, 2016.
- American College of Cardiology Foundation
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.
- Jorde U.P.,
- Kushwaha S.S.,
- Tatooles A.J.,
- et al.
- Rogers J.G.,
- Aaronson K.D.,
- Boyle A.J.,
- et al.
- Wever-Pinzon O.,
- Selzman C.H.,
- Drakos S.G.,
- et al.
- Wever-Pinzon O.,
- Drakos S.G.,
- Kfoury A.G.,
- et al.
- Hasin T.,
- Marmor Y.,
- Kremers W.,
- et al.
- Long E.F.,
- Swain G.W.,
- Mangi A.A.
- Smedira N.G.,
- Hoercher K.J.,
- Lima B.,
- et al.
- Birks E.J.,
- George R.S.,
- Hedger M.,
- et al.
- Mann D.L.,
- Barger P.M.,
- Burkhoff D.
- Mancini D.M.,
- Beniaminovitz A.,
- Levin H.,
- et al.
- Adler E.,
- Silva Enciso J.
- Drakos S.G.,
- Wever-Pinzon O.,
- Selzman C.H.,
- et al.
- Terracciano C.M.,
- Hardy J.,
- Birks E.J.,
- Khaghani A.,
- Banner N.R.,
- Yacoub M.H.
- Hall J.L.,
- Birks E.J.,
- Grindle S.,
- et al.
- Drakos S.G.,
- Kfoury A.G.,
- Stehlik J.,
- et al.
- Mann D.L.,
- Bristow M.R.
- Maybaum S.,
- Mancini D.,
- Xydas S.,
- et al.
- Dandel M.,
- Weng Y.,
- Siniawski H.,
- et al.
- Dandel M.,
- Weng Y.,
- Siniawski H.,
- et al.
- Ibrahim M.,
- Yacoub M.H.
- Imamura T.,
- Kinugawa K.,
- Hatano M.,
- et al.