Myocardial Recovery Using Ventricular Assist Devices
Prevalence, Clinical Characteristics, and Outcomes
Background— Ventricular assist devices (VADs) are important bridges to cardiac transplantation. VAD support may also function as a bridge to ventricular recovery (BTR); however, clinical predictors of recovery and long-term outcomes remain uncertain. We examined the prevalence, characteristics, and outcomes of BTR subjects in a large single center series.
Methods and Results— We implanted VADs in 154 adults at the University of Pittsburgh from 1996 through 2003. Of these implants, 10 were BTR. This included 2/80 (2.5%) ischemic patients (supported 42 and 61 days, respectively). Both subjects had surgical revascularization, required perioperative left VAD support, and were alive and transplant-free at follow up (232 and 1319 days, respectively). A larger percentage of nonischemic patients underwent BTR (8/74, 11%; age 30±14; 88% female; left ventricular ejection fraction 18±6%; supported 112±76 days). Three had myocarditis, 4 had post-partum cardiomyopathy (PPCM), and 1 had idiopathic cardiomyopathy. Five received biventricular support. After explantation, ventricular function declined in 2 PPCM patients who then required transplantation. Ventricular recovery in the 6 nonischemic patients surviving transplant-free was maintained (left ventricular ejection fraction 54±5%; follow-up 1.5±0.9 years). Overall, 8 of 10 BTR patients are alive and free of transplant (follow-up 1.6±1.1 years).
Conclusions— In a large single center series, BTR was evident in 11% of nonischemic patients, and the need for biventricular support did not preclude recovery. For most BTR subjects presenting with acute inflammatory cardiomyopathy, ventricular recovery was maintained long-term. VAD support as BTR should be considered in the care of acute myocarditis and PPCM.
Ventricular assist devices (VADs) have a long history as essential bridges to transplant (BTT) for patients with end-stage congestive heart failure, and more recently as long-term destination therapy for nontransplant candidates.1,2 For transplant candidates, VAD therapy can improve survival post-transplantation.3,4 Hemodynamic unloading and myocardial rest after VAD placement may lead to recovery of native cardiac function, allowing for removal of the device without cardiac transplantation.5–11 VAD support is also associated with decreases in neurohormonal activation, alterations in myocyte calcium handling, and improvement in the proinflammatory cytokine milieu.12–20 Histological analysis of the explanted heart at the time of transplantation demonstrates decreased fibrosis and myocyte size after VAD placement.16,21,22
Despite these salutatory changes as a result of VAD support, the frequency of bridge to recovery (BTR) in subjects supported long term remains low. Additional data are needed to predict long-term outcomes of BTR patients and the extent to which observed ventricular recovery is sustained. The subjects most likely to benefit from a BTR strategy remain poorly defined. We sought to further define the characteristics of successful BTR patients and their long-term outcomes in a large single center series.
Study Design and Data Collection
This study was approved by the institutional review board of the University of Pittsburgh Medical Center. All subjects who underwent VAD support as intended BTT from 1996 to 2003 were retrospectively reviewed. Pediatric patients and those receiving only devices designed for transient support were excluded. Subjects received 1 of 3 VAD systems: Thoratec VAD (Thoratec Corp), Novacor LVAS (World Heart Corporation), or Heartmate LVAS XVE (Thoratec Corp).
The clinical characteristics of subjects who underwent BTR were compared with the remaining (BTT) subjects. For BTR subjects, clinical outcomes and left ventricular ejection fraction (LVEF) by transthoracic echocardiography were reviewed. Data are expressed as mean±SD. Continuous variables were analyzed using Student’s t test. Categorical variables were analyzed using ANOVA. All tests were 2-tailed. A probability value <0.05 was considered statistically significant. Statistical analysis was performed using SPSS for Windows, version 12 (SPSS, Inc).
All 10 BTR patients and a subset of 36 BTT patients underwent specific echocardiographic weaning studies using automated border detection from mid-ventricular short-axis images and noninvasive arterial pressure, as previously described.5 These echocardiographic data were used as a principal identifier of LV recovery. Subjects were selected for screening if they were on support >30 days and if the clinician felt the patient had a reasonable chance for LV recovery, but no other criteria were prespecified. Initial echocardiographic weaning studies were performed approximately 1 month after transplant and repeated as necessary every 2 to 3 weeks until the time of explantation. Briefly, on-line beat-to-beat responses were recorded during 2- to 5-minute trials of decreased device flow to half. With transient low assist device flow, device removal patients had increases in echo stroke area and fractional area change that did not occur in VAD-dependent BTT patients. Estimates of preload-adjusted maximal power, a relatively load-independent index, were significantly higher in BTR patients compared with BTT patients. An increase in stroke area, >40% increase in fractional area change, or a preload adjusted maximal power >4.0 mW/cm4 with low device flow were associated with successful device removal.5
Additional evaluation of recovery involved exercise physiology with gas exchange analysis and invasive hemodynamics by right heart catheterization, both with the VAD rate lowered. Right heart catheterization was performed using the same VAD flow reduction protocol as for echocardiography. VAD flow was decreased by half; if this was tolerated for 3 minutes, heparin 5000 U was given intravenously, the VAD rate decreased to single strokes every 10 seconds, and hemodynamics reassessed.
Prevalence of BTR and Comparison With BTT
Between 1996 and 2003, a total of 161 VAD implants were performed at the University of Pittsburgh for planned long-term support. Of these, 7 were pediatric patients that were excluded from further analysis. Of the remaining 154 adult VAD implants, 10 (6.5%) were BTR, as recovery of ventricular function allowed successful removal of the device without transplantation (Figure 1). BTR patients were younger, more often female, and less likely to have ischemic cardiomyopathy when compared with the BTT patients; other measures of heart failure (HF) severity were similar between groups (Table). Additionally, BTR patients had higher serum transaminases but lower serum creatinine. Because of their very acute presentation, elevated transaminases in the BTR group were felt to be secondary to hepatic congestion or hypoperfusion. These values normalized after 1 month of ventricular unloading with VAD (aspartate transaminase in the BTR group improved from 568±593 to 59±29 U/L, P=0.02; alanine transaminase improved from 372±436 to 31±23 U/L, P=0.04) and remained normal until the time of device explantation.
Support time was similar between BTR and BTT patients (100±72 days, median 73, versus 125±127 days, median 89, respectively). Similar percentages of patients were supported with biventricular assist devices (BTR 50% versus BTT 40%). All BTR patients had Thoratec extracorporeal devices with the exception of 1 patient who had an implantable Thoratec IVAD, reflecting the majority of female patients with a smaller body surface area precluding larger implantable pumps. The remainder of the BTT patients had either Novacor (29%), Heartmate (14%), or Thoratec (17%) VAD for support of the left ventricle only. The majority of BTR subjects were taking angiotensin-converting enzyme inhibitors (90%) and β-blockers (70%) at the time of explanation.
For BTR subjects overall, recovery of function was generally seen by 1 to 2 months of support. Mean peak oxygen consumption with the VAD rate lowered to 40 bpm was 17±3 mL O2 · kg−1 · min−1. All underwent hemodynamic assessment before explantation. The mean pulmonary capillary wedge pressure was 12±5 mm Hg; mean cardiac output was 4.15±1.71 L/min on minimal support (hand pump once every 10 seconds).
BTR Cohort: Ischemic
Overall, 10 subjects were identified as BTR. This included 2 of 80 ischemic patients (2.5%). Both ischemic BTR subjects had new-onset HF after surgical revascularization for an acute myocardial infarction and required perioperative Left ventricular assist device (LVAD) support, and were felt to be candidates for BTR because of their recent-onset of HF. Patient 1 was a 48-year-old man with an LVEF of 10% after experiencing an acute myocardial infarction and undergoing bypass surgery. Postoperatively, he was transferred to our institution on pressor support and an intra-aortic balloon pump. He was supported for 42 days with a Thoratec VAD. Patient 2 was a 52-year-old man with papillary muscle rupture and severe mitral regurgitation after acute myocardial infarction with an LVEF of 20%. He underwent mitral valve and revascularization surgery and postoperatively required extracorporeal membrane oxygenation. He was supported for 61 days with a Thoratec extracorporeal LVAD.
BTR Cohort: Nonischemic
A larger percentage of nonischemic patients underwent successful BTR (8 of 74, 11%). The 8 nonischemic patients in the BTR group were age 30±14 years and 88% were female, with a mean LVEF 18±6%. Etiology was acute myocarditis in 3 (38% of nonischemic BTR cohort, 33% of myocarditis cases requiring VAD), peripartum cardiomyopathy (PPCM) in 4 (50% of nonischemic BTR cohort, 44% of PPCM cases requiring VAD), and 1 had recent onset idiopathic dilated cardiomyopathy (IDCM, 12% of nonischemic BTR cohort, 3% of IDCM cases requiring VAD). Of note, histology at the time on VAD implantation revealed evidence of myocarditis in only 1 subject and nonspecific inflammation in 1 and was negative in the remaining 6. The subject with recent-onset IDCM was diagnosed with active pulmonary tuberculosis while on VAD support. This was successfully treated, and his ventricle improved and he was subsequently weaned from the device.
Five patients were supported with Thoratec biventricular assist devices and 3 with Thoratec LVAD. Mean duration of support was 112±76 days (median 91 days). The mean duration of heart failure before device implant was significantly shorter in the BTR patients compared with the nonischemic BTT patients (0.9±1.1 months, median 0.4, versus 55.5±40.4 months, median 46.0, P<0.0005; Figure 2). Similar results were seen in comparison with ischemic BTT patients (duration of heart failure 30.2±36.9 months, median 16.5) or all BTT patients (duration of heart failure 44.0±47.2 months, median 31.0). BTR patients also had a smaller left ventricular end-diastolic diameter (56±11 versus 69±12 mm, P=0.04), likely reflective of the shorter duration of heart failure.
Overall, 8 of 10 BTR patients are alive and free of transplant, with a mean length of follow-up 1.6±1.1 years (median 1.5 years, range 3 months to 3.6 years). Both ischemic patients remain alive and transplant-free with follow-up of 8 months and 3.6 years (most recent LVEF 30% and 36%, respectively). Of the 8 nonischemic patients, LV function declined in 2 PPCM patients who required transplantation or repeat VAD support within 1 year. The most recent LVEF was 54±5% in the 6 nonischemic patients surviving transplant-free with follow-up of 1.5±0.9 years (median 1.5 years, range 3 months to 3.0 years).
Complete recovery of LVEF was evident before explantation in the 6 nonischemic patients surviving transplant-free, but not in the 2 patients who subsequently declined (LVEF before explantation 56±3% versus 39±2%, P=0.002). The 6 nonischemic patients surviving transplant-free maintained normal LVEF up to 3 years after explantation. The 2 ischemic patients recovered to LVEFs of 35% and 45%, respectively, before explantation. For these subjects, LVEF subsequently declined in the first month after device removal but then stabilized (Figure 3).
In this study, 6.5% of all adult patients implanted with a VAD for planned long-term support recovered ventricular function sufficient for device removal. The majority of these patients were nonischemic (80% of the cohort, representing 11% of the nonischemic patients), with the etiology in most cases being acute myocarditis or PPCM. This is consistent with the multicenter experience previously reported,6 which suggests BTR is predominantly seen in myocarditis and acute inflammatory cardiomyopathies such as PPCM. The high degree of biventricular failure in this recovery cohort, as evidenced by the large percentage requiring biventricular support, is not apparent in previous series. The present study demonstrates that the need for biventricular support should not preclude consideration of recovery.
The current report also suggests that clinical measures of acute illness, in particular duration of symptoms and LV diameter, are potentially important predictors of recovery. This article also demonstrates the limitations of histology for predicting outcomes. The majority of BTR subjects had no evidence of inflammation on myocardial histology at the time of implant, despite a clinical history suggesting acute inflammatory cardiomyopathy. Though histological evidence of myocarditis may suggest a potentially reversible process, a negative biopsy seems to have little predictive value.
The current report differs from the Berlin experience, which reported successful BTR in subjects with long-standing chronic dilated cardiomyopathy.10 No chronic subject (ie, history of heart failure greater than six months) was successfully weaned in our series. This may reflect a screening bias, as acute patients are may have been examined more closely for BTR than chronic subjects. Many chronic subjects, however, underwent echocardiographic weaning studies but failed to meet recovery criteria, and this cohort of “screened” BTT subjects had a duration of HF similar to that of the remaining BTT patients who were not studied (37±39, median 35, versus 45±48 months, median 30, respectively; P=NS). The finding that BTR subjects all had short durations of symptoms is consistent with the finding in nonischemic cardiomyopathy that dynamic recovery is common in recent-onset disease23 but extremely rare in more chronic disease. In addition to optimal medical therapy for heart failure, future adjunctive treatments, such as stem cells or tissue engineering with gene therapy, will be necessary to assist myocardial recovery in long-standing chronic cardiomyopathy and allow VAD as BTR to become an option for this subset of patients.
Despite the physiological improvements evident with VAD support, the number of BTR patients is low. This may be in part due to inadequate screening. Recovery of function, as defined by improvement in LVEF, was an early phenomenon that occurred within 2 months of VAD support. Routine assessment by transthoracic echo after 1 to 2 months of support would identify all the BTR subjects in the current study and appears to be the most useful clinical tool for more generalized screening for BTR. The 2 subjects who declined after BTR in this series had significant reductions in their LVEF noted within the first month after explantation. This suggests that the recurrence of LV dysfunction after BTR is unfortunately also an early phenomenon that appears soon after device removal restores normal loading conditions.
For subjects with acute onset of inflammatory myocarditis or PPCM, VAD support allows long-term restoration of cardiac function and avoidance of transplantation. Whether VAD support is therapeutic and facilitates recovery in this subset or functions as a bridge to spontaneous innate recovery remains to be determined. As knowledge of the mechanisms of VAD-induced ventricular recovery progresses, the therapeutic role of VAD support in acute inflammatory cardiomyopathy and the appropriate timing of support needs to be more clearly defined.
The results of this retrospective study must be interpreted with caution and considered primarily hypothesis-generating. The predictive value of LV size and the limited value of histology should be revaluated prospectively in a multicenter format. In particular, the absence of BTR in more chronic subjects may reflect screening bias and needs to be better delineated prospectively in a larger series.
In a large single center series, VAD as BTR was seen in 6.5% overall and in 11% of nonischemic patients and was most successful with acute inflammatory cardiomyopathy or PPCM. When used in acute inflammatory cardiomyopathy or PPCM, VAD support appears to allow long-term restoration of cardiac function and the avoidance of transplantation. VAD support as BTR should be an important component of the care of acute inflammatory myocarditis.
Dr McNamara is supported by National Institutes of Health grants R01 HL75038 and K24 HL69912. Dr Simon is supported by National Institutes of Health National Research Service Award 2-T32 HL07820-06. Dr Gorcsan is supported in part by National Institutes of Health award K24 HL04503-01 and American Heart Association National Grant-in-Aid 0050587N.
Dr Kormos is on the medical advisory boards of Thoratec Corporation and World Heart Corporation.
Rose EA, Gelijns AC, Moskowitz AJ, Heitjan DF, Stevenson LW, Dembitsky W, Long JW, Ascheim DD, Tierney AR, Levitan RG, Watson JT, Meier P, Ronan NS, Shapiro PA, Lazar RM, Miller LW, Gupta L, Frazier OH, Desvigne-Nickens P, Oz MC, Poirier VL; Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) Study Group. Long-term mechanical left ventricular assistance for end-stage heart failure. N Engl J Med. 2001; 345: 1435–1443.
Farrar DJ, Holman WR, McBride LR, Kormos RL, Icenogle TB, Hendry PJ, Moore CH, Loisance DY, El-Banayosy A, Frazier H. Long-term follow-up of Thoratec ventricular assist device bridge-to-recovery patients successfully removed from support after recovery of ventricular function. J Heart Lung Transplant. 2002; 21: 516–521.
Khan T, Delgado RM, Radovancevic B, Torre-Amione G, Abrams J, Miller K, Myers T, Okerberg K, Stetson SJ, Gregoric I, Hernandez A, Frazier OH. Dobutamine stress echocardiography predicts myocardial improvement in patients supported by left ventricular assist devices (LVADs): hemodynamic and histologic evidence of improvement before LVAD explantation. J Heart Lung Transplant. 2003; 22: 137–146.
Mancini DM, Beniaminovitz A, Levin H, Catanese K, Flannery M, DiTullio M, Savin S, Cordisco ME, Rose E, Oz M. Low incidence of myocardial recovery after left ventricular assist device implantation in patients with chronic heart failure. Circulation. 1998; 98: 2383–2389.
James KB, McCarthy PM, Thomas JD, Vargo R, Hobbs RE, Sapp S, Bravo E. Effect of the implantable left ventricular assist device on neuroendocrine activation in heart failure. Circulation. 1995; 92: 191–195.
Milting H, El-Banayosy A, Kassner A, Fey O, Sarnowski P, Arusoglu L, Thieleczek R, Brinkmann T, Kleesiek K, Korfer R. The time course of natriuretic hormones as plasma markers of myocardial recovery in heart transplant candidates during ventricular assist device support reveals differences among device types. J Heart Lung Transplant. 2001; 20: 949–955.
Terracciano CM, Hardy J, Birks EJ, Khaghani A, Banner NR, Yacoub MH. Clinical recovery from end-stage heart failure using left-ventricular assist device and pharmacological therapy correlates with increased sarcoplasmic reticulum calcium content but not with regression of cellular hypertrophy. Circulation. 2004; 109: 2263–2265.
Barbone A, Holmes JW, Heerdt PM, The AH, Naka Y, Joshi N, Daines M, Marks AR, Oz MC, Burkhoff D. Comparison of right and left ventricular responses to left ventricular assist device support in patients with severe heart failure: a primary role of mechanical unloading underlying reverse remodeling. Circulation. 2001; 104: 670–675.
Bartling B, Milting H, Schumann H, Darmer D, Arusoglu L, Koerner MM, El-Banayosy A, Koerfer R, Holtz J, Zerkowski HR. Myocardial gene expression of regulators of myocyte apoptosis and myocyte calcium homeostasis during hemodynamic unloading by ventricular assist devices in patients with end-stage heart failure. Circulation. 1999; 100: II216–II223.
Torre-Amione G, Stetson SJ, Youker KA, Durand JB, Radovancevic B, Delgado RM, Frazier OH, Entman ML, Noon GP. Decreased expression of tumor necrosis factor-alpha in failing human myocardium after mechanical circulatory support: a potential mechanism for cardiac recovery. Circulation. 1999; 100: 1189–1193.
Bruckner BA, Stetson SJ, Perez-Verdia A, Youker KA, Radovancevic B, Connelly JH, Koerner MM, Entman ME, Frazier OH, Noon GP, Torre-Amione G. Regression of fibrosis and hypertrophy in failing myocardium following mechanical circulatory support. J Heart Lung Transplant. 2001; 20: 457–464.
Zafeiridis A, Jeevanandam V, Houser SR, Margulies KB. Regression of cellular hypertrophy after left ventricular assist device support. Circulation. 1998; 98: 656–662.
McNamara DM, Holubkov R, Starling RC, Dec GW, Loh E, Torre-Amione G, Gass A, Janosko K, Tokarczyk T, Kessler P, Mann DL, Feldman AM. Controlled trial of intravenous immune globulin in recent-onset dilated cardiomyopathy. Circulation. 2001; 103: 2254–2259.