Both devices that are currently available for percutaneous implantation are indicated for short-term use, only. The TandemHeart™ (CardiacAssist Inc., Pittsburgh, PA, USA) percutaneous ventricular assist device (pVAD) is a centrifugal continuous-flow pump which can be adapted to provide RV support [2, 3]. The RA and pulmonary artery (PA) [2] are cannulated (Fig. 21.1d), with percutaneous access obtained via the femoral vein. A recent review [19] included 46 patients in whom the pVAD (percutaneous and surgical approach) was used for isolated RV as well as biventricular support. Cases included acute MI, myocarditis, chronic LV dysfunction, and patients post valve and CABG surgery. The mean duration of support was 4.8 ± 6.1 days for the percutaneous approach, and 6.5 ± 6.2 days for the surgical approach. Mean flow provided was 4.2 ± 1.3 L/min. Overall in-hospital mortality was 57 %, with cause of death being multiorgan failure [19].

An axial continuous-flow pump, the Impella RP can also provide RV support [3] with flows up to 4.8 L/min [2]. It is a small device, with a diameter of 6.4 mm and weight of 17 g, which permits both percutaneous and central approaches for implantation (Fig. 21.1e) [9]. The inlet cannula is placed in the inferior vena cava (IVC) and outflow in the PA [2]. Advantages over the Centrimag and AB5000 include a much smaller surface area exposed to blood [2], but the device relies on mechanical bearings which increase the risk of hemolysis and thrombosis [9]. As a result, it is presently approved for only 10 days of support [9]. In first-in-man trials in Canada and Europe, the device has been used in patients with RV failure after cardiac surgery and after LVAD implantation. The duration of support has ranged from 1 to 7 days, with >60 % of patients being supported for more than 4 days and having the device explanted upon recovery of the RV [2, 20].

The options for paracorporeal support are comprised of pneumatic pulsatile devices. Of these, only the Thoratec PVAD (Thoratec Corp., Pleasanton, CA, USA) has been approved by the FDA for RV support [3]. As a bridge to transplantation and recovery [3, 9], it has been used for univentricular or biventricular support in over 4,000 patients since 2010 (Fig. 21.2a) [4]. In general, biventricular support has worse outcomes than LV support alone [21] and planned biventricular support is associated with better outcomes than LVAD implantation followed by RV failure requiring a second operation for RVAD implantation [10]. Single center experience with biventricular support using the Thoratec PVAD has reported survival rates of 75 % when excluding those supported for postcardiotomy or post-infarct shock which is known to have very poor outcomes [22].

The AB5000 (Abiomed Inc.) requires a sternotomy to cannulate the RA and PA [16], but the device can be exchanged at the bedside without a re-operation [9]. Duration of support can last up to months [16]. It can generate flows of up to 5–6.5 L/min, and has a fixed drive pressure of 300 mmHg (Fig. 21.2b) [4]. Introduced in 1988 [23], the BerlinHeart Excor (Berlin Heart GmbH) device can be used as a bridge to recovery or transplantation [4]. It can provide ambulatory support for up to 10 h [9], and has been demonstrated to provide biventricular support for up to 575 days (Fig. 21.2c) [23].

The pulsatile Thoratec IVAD™ (Thoratec Corp., Pleasanton, CA, USA) [3] can provide intermediate to long-term support and is indicated for biventricular support as a bridge to transplant or recovery (Fig. 21.3a) [9]. In a multicenter clinical trial of 29 patients, 15 received biventricular support using the IVAD. Of the 14 bridge to transplant candidates, eight patients survived: one was weaned off support, and the other seven were transplanted [24].

Continuous-flow devices such as the HVAD^® (HeartWare International Inc., MA, USA) and HeartMate II^® (Thoratec Corp., Pleasanton, CA, USA) have been used to provide biventricular support [9] in cases such as giant cell myocarditis [25] and after cardiac arrest during non-cardiac surgery [26]. With dual controllers, the Heartware system has been increasingly used as an alternative to the total artificial heart. Such dual support can provide successful physiologic levels of support and can alter flows to respond to changes in preload and afterload [26]. The duration of support has ranged from 7 days [26] to 4 months [25]. Complications include suction events causing RA collapse [26]. Although these devices are not approved by the FDA for RV support, they have been used for RV support in Europe and in the United States (via individual appeals to the FDA for Humanitarian Device Exemption), utilizing separate controllers and certain adaptations by some centers. With such biventricular support, RVAD flows have been set lower than systemic output, to avoid overloading the LV [8].

However, some investigators point out that adapting LVADs for RV support—specifically by reducing pump speeds beyond design specifications—increases thrombosis risk [8]. A continuous-flow pump designed specifically for the RV, the Cleveland Clinic’s DexAide RVAD has been successfully implanted in calves and averaged 24 ± 21days of support, generating flows of 5.4 ± 1 L/min [8, 9]. In animal models of biventricular support using continuous-flow devices, these investigators have found that the circulatory loop is most stable when RVAD flows are lower than the LVAD’s [8]. Specifically, in a biventricular support model, RVAD speeds must ideally (1) adjust so that flows match 50–75 % of the LVAD output at any point in time and (2) have a maximum threshold so that, in the event of hypovolemia or LV failure, the system can avoid suction events and overdriving [8].

The Circulite^® Synergy^® micropump has similarly been used for biventricular support in fibrillating sheep hearts [14]. This miniature pump, which weighs 25 g, has a pressure gradient of 70–80 mmHg and can generate flows up to 4.25 L/min. Lower flow rates of 3 L/min can be generated at the lowest speed of 20,000 rpm and a pressure gradient of 30 mmHg, which make it ideal for RV support (Fig. 21.3b) [14]. In the fibrillating heart model, right and left sided flows always equilibrated, with a proportional decrease in left atrial (LA)-aorta pressure gradient if the RA-PA gradient were increased with increasing RVAD speeds [14]. The Circulite system is currently undergoing revision.

There is potentially a great need for RV support in pulmonary arterial hypertension (PAH), as many patients die from RV failure (see Chap. 14). RV dysfunction in the setting of PAH or pulmonary veno-occlusive disease poses a significant challenge for mechanical support, as the RV pump failure is also accompanied by significant LV diastolic dysfunction [27]. Case reports of patients in florid cardiogenic shock have also described significantly elevated pulmonary pressures with and without associated pulmonary hemorrhage after RVAD implantation [13, 28]. Indeed, a computer simulation of the cardiovascular system in PAH and RV dysfunction incorporating a continuous-flow micropump showed that, while left sided filling and cardiac output improved with mechanical support, pulmonary arterial pressures and PCWP rose significantly [27]. However, the increase in pulmonary arterial pressures could be mitigated by setting lower RVAD flow rates with continued improvement of the systemic hemodynamics (Fig. 21.4) [27].

Such a system has been shown to be feasible in animal models. One such device is the OxyRVAD, which generates flows of up to 3 L/min through the pulmonary vascular bed [29]. It includes both an axial flow pump and a low resistance gas exchanger, with the VAD cannula placed in the RA appendage and the outflow graft anastamosed to the PA. The device successfully provided hemodynamic support for 14 days in healthy sheep [29]. The MC3 BioLung is a thoracic artificial lung (TAL) that has been studied in sheep models with chronic PH [30]. The circuit connects the PA and LA, and it does not (yet) incorporate a blood pump. The total impedance of the TAL in parallel with the native pulmonary circulation is less than the native system alone, thus decreasing pulmonary resistance and RV afterload. More blood can be diverted to the TAL when the PA is banded, but this is at the expense of increased overall impedance and afterload. The systemic output drops when >75 % of blood flow is diverted to the TAL [30].

More recently, the successful use of a paracorporeal artificial lung (PAL) has been described in patients with PH and RV failure [31, 32]. The Novalung, which does not incorporate a blood pump, connects the PA and LA and has been demonstrated to generate flows of 3.5 L/min, reduce PA pressures, and improve systemic hemodynamics [31]. In a retrospective review of patients with PAH who were listed for lung transplantation, the incorporation of ECLS strategy with select patients receiving the Novalung, was shown to reduce mortality and time on the waiting list for transplantation [33].

Similar to the initiation of support for the left heart as a bridge to transplant in INTERMACS 1 and/or 2 patients [34], several institutions have utilized the markedly improved technology of ECLS in the PH patient population awaiting lung transplant as a bridge to transplant or less commonly, as a bridge to recovery as the best means of support for the ultimate failing RV [33, 35–37]. In a newer paradigm of RV support, whereas the RVAD may ultimately be considered in the chronically failing patient (Fig. 21.4), extracorporeal membrane oxygenation (ECMO) support is the choice for the “crash and burn” viable, transplantable, “PH INTERMAS 1 and 2” equivalent in the setting of PH and RV failure. Timing is crucial—to intervene in the patient with imminent but not end-organ injury generally characterized by inotrope dependence or resistance, diuretic resistance, systemic hypotension with renal insufficiency and/or abnormal liver function tests. In particular, transfer of patients to centers where cardiac support device therapies have been established in a timely fashion is advisable. ECLS may include traditional femoral [38] vs. upper torso ambulatory “Sport Model” configurations [39] for VA or VAV ECMO circuits, single catheter VV configurations across existing intracardiac shunts [39] for effective VA support, attempts at VV with larger natural or created PFO configurations and PA-LA Novalung configurations [33].