Article ID: CR-25-0224
Background: Managing heart failure complicated by severe right heart failure with implantable mechanical circulatory support remains a significant challenge. One therapeutic strategy is biventricular assist device (BiVAD) support, typically involving the off-label use of 2 implantable left ventricular assist devices (LVADs). Because the available data of implantable BiVAD support remain limited, we reviewed the data for 6 patients on implantable BiVAD support.
Methods and Results: Between January 2010 and March 2019, 6 patients underwent BiVAD implantation at Osaka University Hospital. Their mean age was 31±11 years, and 2 (33%) were male. The right ventricular assist devices (RVADs) utilized were Jarvik2000 (Jarvik Heart, NY, USA) in 4 patients (67%), and HVAD (HeartWare, Framingham, MA, USA) in 2 patients (33%). The survival rates at 1 and 3 years after BiVAD implantation were 83% and 67%, respectively. Of the 6 patients, 4 underwent heart transplantation at 553, 709, 791, and 1,245 days, respectively, following RVAD implantation; 2 patients died during follow-up at 280 and 511 days, respectively, after RVAD implantation. Stroke occurred in 3 patients. Hemolysis or pump thrombosis occurred in 3 patients. Heart failure occurred in 3 patients. Device-related infection occurred in 1 patient.
Conclusions: Although implantable BiVAD support provided a feasible bridge to transplantation with favorable survival, the high incidence of complications indicates that significant challenges remain in optimizing patient outcomes and emphasizes the necessity for RV-specific device development.
Cardiovascular medicine for heart failure (HF) has made significant progress,1,2 and for patients in whom HF does not improve with medical therapy, left ventricular assist devices (LVADs) play a crucial role in clinical management. However, managing HF complicated by severe right HF with implantable mechanical circulatory support remains a significant clinical challenge.3–6 One therapeutic strategy for such cases of severe biventricular failure is biventricular assist device (BiVAD) support, typically involving the off-label use of 2 LVADs. Although several studies have reported clinical outcomes associated with implantable BiVAD support, the available data remain limited.7,8 We experienced 6 cases of BiVAD implantation at Osaka University Hospital, and the present study is our retrospective review of those patients, focusing on their preoperative characteristics, clinical outcomes, and adverse events.
Between January 2010 and March 2019, 6 patients underwent BiVAD implantation at Osaka University Hospital. Devices for RVAD was purchased through our research grant and BiVAD implantation was approved by the Osaka University School of Medicine. The mean age of the patients was 31±11 years, and 2 (33%) were male. The underlying etiologies were fulminant myocarditis in 3 (50%), endstage hypertrophic cardiomyopathy in 1 (17%), arrhythmogenic right ventricular cardiomyopathy in 1 (17%), and cardiac sarcoidosis in 1 patient (17%). The interagency registry for mechanically assisted circulatory support (INTERMACS) profiles were Bridge-to-Bridge in 2 patients (33%), Profile 1 in 3 patients (50%), and Profile 2 in 1 patient (17%) (Table).
Preoperative Characteristics and Clinical Outcomes of Patients on BiVAD Support
| Case no. |
Age (years) |
Sex | Diagnosis | INTERMACS profile |
LVAD device |
RVAD device |
Duration from LVAD to RVAD implantation (days) |
Outcome | Duration as outpatient after RVAD implantation (days) |
Complications |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 47 | M | Sarcoidosis | 1 | EVAHEART2 | HVAD | 50 | HTx (553 days) |
113 | Cerebral hemorrhage, cerebral infarction |
| 2 | 35 | F | Fulminant myocarditis |
B | DuraHeart | Jarvik2000 | 0 | HTx (709 days) |
576 | HF |
| 3 | 13 | F | Endstage HCM |
2 | HVAD | HVAD | 35 | HTx (791 days) |
349 | HF due to AR and PR |
| 4 | 27 | F | ARVC | 1 | Jarvik2000 | Jarvik2000 | 0 | HTx (1,245 days) |
968 | Cerebral hemorrhage, pump thrombosis, hemolysis |
| 5 | 31 | F | Fulminant myocarditis |
B | DuraHeart | Jarvik2000 | 14 | Death (280 days) |
0 | Hemolysis, device infection |
| 6 | 32 | M | Fulminant myocarditis |
1 | EVAHEART | Jarvik2000 | 19 | Death (511 days) |
123 | Cerebral hemorrhage, HF due to AR |
AR, aortic regurgitation; ARVC, arrhythmogenic right ventricular cardiomyopathy; HCM, hypertrophic cardiomyopathy; HF, heart failure; HTx, heart transplantation; INTERMACS, interagency registry for mechanically assisted circulatory support; LVAD, left ventricular assist device; PR, pulmonary regurgitation; RVAD, right ventricular assist device.
All patients and their families provided informed consent to participate in related clinical studies prior to LVAD implantation.
DevicesThe types of LVAD implanted were as follows: DuraHeart (Terumo Heart, Ann Arbor, MI, USA) in 2 patients (33%), EVAHEART (SunMedical, Japan) in 1 patient (17%), EVAHEART2 (SunMedical, Japan) in 1 patient (17%), Jarvik2000 (Jarvik Heart, NY, USA) in 1 patient (17%), and HVAD (HeartWare, Framingham, MA, USA) in 1 patient (17%). The implantable LVADs were used off-label as right ventricular assist devices (RVAD) in this study. The RVADs utilized were Jarvik2000 in 4 patients (67%) and HVAD in 2 patients (33%), selected in consideration of their relatively small size and the associated technical feasibility of implantation as RVADs (Table).
Surgical ProceduresFor patients in whom the Heart Team determined that recovery of right heart function was unlikely, an implantable RVAD was placed simultaneously with the LVAD. In patients with a possibility of right heart function recovery following LVAD implantation, extracorporeal RVAD support was initially employed at the time of LVAD implantation. In all cases, the LVAD inflow cannula was implanted into the left ventricle, and the outflow graft was anastomosed to the ascending aorta. The RVAD outflow graft was anastomosed to the main pulmonary trunk. When the Jarvik2000 was used as the RVAD, the pump was implanted into the RV free wall (Figure 1A). To prevent inflow obstruction, the anterior leaflet of the tricuspid valve was either sutured to the anterior wall of the RV or resected. When the HVAD was used as the RVAD, the inflow cannula was placed into the right atrium at the depth of 10 mm, and the pump body was secured to the chest wall (Figure 1B).
Intraoperative drawings of BiVAD implantation in 2 patients: (A) with Jarvik 2000 for LVAD and RVAD; and (B) with EVAHEART for LVAD and HVAD for RVAD. BiVAD, biventricular assist device; LVAD, left ventricular assist device; RVAD, right ventricular assist device.
The RVAD pump speed was set at the minimum setting (Jarvik2000: dial 1, 8,000 rpm; HVAD: 2,400 rpm). Adequate volume resuscitation and careful adjustment of RVAD flow were performed under transesophageal echocardiography monitoring of the LV and RV volumes during weaning from cardiopulmonary bypass. After weaning, intravenous catecholamines or vasopressors were administered during the early postoperative period as needed.
Anticoagulation ManagementAntiplatelet therapy with aspirin (100 mg/day) was initiated on the day following implantation and continued throughout BiVAD support unless major bleeding occurred. In addition, anticoagulation with warfarin was initiated after adequate hemostasis was achieved, targeting an international normalized ratio (INR) of 2.0–3.0. Systemic heparinization was maintained until the target INR was achieved.
Data CollectionThe patient data included baseline characteristics, etiology, device types, duration of BiVAD implantation, postoperative complications and course, electrocardiography and echocardiography. All patient data were collected from electronic medical and operative records.
Statistical AnalysisKaplan-Meier analysis was used to calculate the survival rate after BiVAD implantation.
The survival rates at 1 and 3 years after BiVAD implantation were 83% and 67%, respectively (Figure 2). Of the patients, 4 underwent heart transplantation at 553, 709, 791, and 1245 days, respectively, following RVAD implantation, with outpatient management durations of 113, 576, 349, and 948 days, respectively (Cases 1–4). There were 2 deaths during follow-up (Cases 5 and 6) at 280 and 511 days, respectively, after RVAD implantation; their outpatient management durations were 0 and 123 days, respectively.
Survival rate after biventricular assist device (BiVAD) implantation.
Timing of RVAD Implantation
In 2 patients (Cases 2 and 4), RVADs were implanted simultaneously with LVADs due to poor right heart function. In Case 2, emergency BiVAD implantation using a Toyobo paracorporeal device (Toyobo-Nipro, Osaka, Japan) was performed for cardiac arrest secondary to fulminant myocarditis. Postoperatively, electrocardiography revealed persistent ventricular asystole, and echocardiography demonstrated no contraction of both ventricles with marked RV dilation (Figures 3,4). At 6 months after paracorporeal BiVAD implantation, progressive liver dysfunction secondary to severe pulmonary regurgitation developed, necessitating pulmonary valve closure, at which time paracorporeal BiVAD was converted to an implantable BiVAD.9,10 In Case 4, the patient had a long history of HF due to arrhythmogenic RV cardiomyopathy and was frequently hospitalized. Following rapid clinical deterioration, she was referred for mechanical circulatory support under INTERMACS Profile 2. Echocardiography demonstrated severe RV dilatation with poor function and small LV (Figure 5). The Heart Team determined that isolated LVAD would be insufficient, and simultaneous BiVAD implantation was performed.11–13
Echocardiographic findings in Case 2 before BiVAD implantation: (A) long-axis view, (B) short-axis view. Ao, aorta; BiVAD, biventricular assist device; LA, left atrium; LV, left ventricle; RV, right ventricle.
ECG in Case 2 after biventricular assist device (BiVAD) implantation. Only the P wave is observed and the QRS complex that should have followed is absent.
Echocardiographic findings in Case 4 before BiVAD implantation: (A) long-axis view, (B) short-axis view. Ao, aorta; BiVAD, biventricular assist device; LA, left atrium; LV, left ventricle; RV, right ventricle.
The remaining 4 patients (Cases 1, 3, 5, and 6) initially received extracorporeal RVAD support for 14–50 days (mean, 30 days) following LVAD implantation. In Cases 5 and 6, repeated RVAD-off test under maximal inotropic support resulted in decreased LVAD flow, elevated central venous pressure, and hypotension, indicating an inability to wean from RVAD support. In Cases 1 and 3, RVAD-off tests were not attempted because of complete absence of RV contraction; even a decrease in RVAD flow caused elevated central venous pressure and decreased LVAD flow.
Adverse EventsStroke occurred in 3 patients (Cases 1, 4, 6: 50%). Hemolysis or pump thrombosis occurred in 2 patients (33%). HF occurred in 3 patients (50%). Valve disfunction occurred in 2 patients (33%). Device-related infection occurred in 1 patient (17%). None of the patients developed renal failure or hepatic disorders (0%).
Of the stroke cases, in Case 1, the patient suffered 2 cerebral hemorrhages and 3 cerebral infarctions between 3 and 15 months post-implantation, resulting in mild lower limb paralysis, dysarthria, and cognitive decline. In Case 4, a cerebral hemorrhage occurred 2 months after RVAD implantation. Temporary reversal of anticoagulation led to pump thrombosis.12 In Case 6, the patient died from cerebral hemorrhage at 511 days after RVAD implantation.
Hemolysis or pump thrombosis occurred in Cases 4 and 5. In Case 4, massive hemoglobinuria and elevated serum lactate dehydrogenase (LDH) levels (∼7,000 IU/L) were observed following temporary reversal of anticoagulation therapy. Hemolysis led to renal failure and anemia requiring frequent transfusions. Severe right HF occurred despite maximum RVAD support. Pump thrombosis of the RVAD was suspected, leading to conversion to a temporary extracorporeal RVAD, which resulted in reduction of the serum LDH to ∼1,500 IU/L. Pump thrombosis of the LVAD was also suspected, and both devices were replaced 5 months after initial implantation, successfully resolving the hemolysis.12 In Case 5, massive hemoglobinuria and elevated serum LDH (∼5,000 IU/L) were noted 5 months post-implantation. Echocardiography revealed a small RV. Reducing the RVAD speed (from dial 1 to dial “6K”, specially customized for RVAD) led to a gradual decrease in the serum LDH to ∼1,500 IU/L over 2 weeks. However, the patient later died from device-related infection.
HF occurred in Cases 2, 3, and 6. In Case 2, the patient developed exertional dyspnea 14 months after RVAD implantation; symptoms improved with an increase in RVAD speed (from dial 1 to 2). In Case 3, neither aortic regurgitation (AR) nor pulmonary regurgitation (PR) was present prior to BiVAD implantation; however, HF secondary to AR and PR developed 10 months post-implantation. Surgical closure of both valves was performed at 329 days post-implantation. In Case 6, mild AR was observed at the time of BiVAD implantation, and 1 year later, the patient was hospitalized for HF due to worsening AR and subsequently died from cerebral hemorrhage during hospitalization.
Device-related infection occurred in Case 5. This patient had a preexisting infection at the cannula exit site of the paracorporeal device prior to conversion to an implantable BiVAD. Culture of the cannula exit site and the LV and RV cuffs revealed Staphylococcus epidermidis. Extensive debridement and omentopexy were performed at the time of device conversion, but 5 months after BiVAD conversion, the patient developed fever, with blood cultures positive for methicillin-resistant Staphylococcus aureus, and computed tomography imaging showed a low-density area around the inflow cannula suggestive of device infection. Despite aggressive management, the patient’s condition deteriorated, leading to death from respiratory failure at 280 days after RVAD implantation.
We report the clinical characteristics and outcomes of 6 patients who underwent implantable BiVAD support, using an implantable LVAD off-label as an RVAD. Of then, 4 patients successfully underwent heart transplantation after 553–1,245 days (mean, 825 days) of BiVAD support. However, all patients experienced adverse events such as stroke, pump thrombosis, hemolysis, and HF during the course of device support.
Previous studies have reported 1-year survival rate of 45–60% for patients on BiVAD support, with a bridge-to-heart transplantation success rate of 10–15%.7,8 Compared with those results, the outcomes in our study were favorable, demonstrating acceptable long-term BiVAD support and a high success rate for bridge-to-heart transplantation (66%). Nevertheless, significant challenges remain. The mean outpatient duration after RVAD implantation was only 41±33%, and adverse events such as stroke, pump thrombosis, hemolysis, renal failure, HF, and infections occurred in the majority of patients. Prolonged BiVAD support does not seem to prevent adverse events, and often requires additional surgical interventions, including valve closure surgery for regurgitation, pump exchange procedures for pump thrombosis, and drainage or antimicrobial therapy for device infections. Readmission due to complications is a significant drawback for patients.14,15 The appropriate timing and types of interventions are required to achieve acceptable outcomes of long-time BiVAD support.
Pump thrombosis remains a major concern during implantable BiVAD support. Previous reports have indicated a RVAD thrombosis rate of approximately 14% when HeartMate 3 devices (Abbott, IL, USA) are used for RV support.7,8 In our study, pump thrombosis was suspected or confirmed in 2 of 6 patients (33%), and 1 patient (Case 4) required pump exchange surgery. It is important to note that although no cases of pump thrombosis were reported in HeartMate 3 LVAD patients in the MOMENTUM 3 trial,16 off-label use of LVAD for RV support may present a higher thrombosis risk, attributable to design limitations, as current LVAD devices are not optimized for the low afterload conditions for the RV and the RVAD rotational speed must generally be set lower than that of the LVAD.17 Moreover, the low pulse pressures may cause blood stasis within the pump.18 These factors likely contribute to the increased risk of pump thrombosis we observed. Furthermore, the reversal of anticoagulation therapy in the setting of bleeding events, as seen in Case 4, is a known risk for pump thrombosis. Development of devices specifically designed for RV support may help address these issues.
Neurological complications were also common, with stroke events occurring in half of the patients. The HeartMate 3 LVAD had demonstrated lower stroke rates than HeartMate II.19 However, in the setting of BiVAD support, stroke prevention remains a challenge. In a recent study using HeartMate 3 for BiVAD support, stroke incidence was 7%, but the incidence of sepsis was high (36%), with more than half of sepsis cases resulting in death.7 Bloodstream infections are a known risk factor for hemorrhagic stroke and worsen overall survival in patients with VADs.20–22 Therefore, preventing infection may be critical for improving outcomes in BiVAD-supported patients. Risk factors for VAD infections include bridge-to-bridge procedures and preoperative use of temporary mechanical circulatory support such as venoarterial extracorporeal membrane oxygenation or percutaneous LVAD.23 Whenever possible, BiVAD implantation should be performed under stable hemodynamic conditions, ideally in patients with INTERMACS Profile 3. Furthermore, simultaneous implantation of LVAD and RVAD may reduce infection risk by avoiding prolonged exposure to temporary devices. In our study, a RVAD was not implanted simultaneously with the LVAD in 66% of cases, possibly reflecting hesitation to proceed with BiVAD implantation in critically ill patients due to concerns regarding prognosis.24
Valve dysfunction also represents a significant issue during long-term BiVAD support. In our cohort of 6 patients, 1 developed significant PR, 1 developed AR, and 1 developed AR and PR, necessitating surgical valve closure in 2 cases. The incidence of valve dysfunction seems to be higher than those reported in isolated LVAD patients,25 although the precise mechanisms behind this remain to be investigated. Severe PR can exacerbate RV volume overload and liver dysfunction, as occurred during paracorporeal BiVAD support in Case 2.8,9 Routine echocardiographic monitoring is critical for early detection, and proactive valve interventions should be considered to prevent secondary organ failure.
Study LimitationsThis was a single-center, retrospective study with a small sample size, limiting generalizability. The RVAD devices used (HVAD and Jarvik2000) varied among the patients, and management strategies for anticoagulation and infection were not standardized, potentially influencing outcomes.
Implantable BiVAD support utilizing off-label LVADs for RV support achieved favorable long-term survival and a high rate of successful bridge to transplantation. However, the incidence of serious adverse events remained high. Development of RV-specific devices, standardized perioperative management strategies, and early detection and intervention for complications are urgently required to improve outcomes in patients requiring BiVAD support.
The authors declare no conflicts of interest regarding this report.
The present study was approved by Osaka University School of Medicine (Reference number: 21372(T1)3).
The deidentified participant data will not be shared.
