Journal of Artificial Organs

, Volume 17, Issue 2, pp 135–141

Novel control system to prevent right ventricular failure induced by rotary blood pump


    • Department of Artificial OrgansNational Cerebral and Cardiovascular Center Research Institute
    • Department of Cardiovascular Surgery, Saitama Medical CenterJichi Medical University
    • Department of Cardiac SurgeryTokyo Metropolitan Geriatric Hospital
    • Department of Cardiothoracic SurgeryUniversity of Tokyo
  • Yoshiaki Takewa
    • Department of Artificial OrgansNational Cerebral and Cardiovascular Center Research Institute
  • Akihide Umeki
    • Department of Cardiothoracic SurgeryUniversity of Tokyo
  • Masahiko Ando
    • Department of Cardiothoracic SurgeryUniversity of Tokyo
  • Yuichiro Kishimoto
    • Department of Artificial OrgansNational Cerebral and Cardiovascular Center Research Institute
  • Yutaka Fujii
    • Department of Artificial OrgansNational Cerebral and Cardiovascular Center Research Institute
  • Shunei Kyo
    • Department of Cardiac SurgeryTokyo Metropolitan Geriatric Hospital
  • Hideo Adachi
    • Department of Cardiovascular Surgery, Saitama Medical CenterJichi Medical University
  • Eisuke Tatsumi
    • Department of Artificial OrgansNational Cerebral and Cardiovascular Center Research Institute
Original Article Artificial Heart (Basic)

DOI: 10.1007/s10047-014-0757-1

Cite this article as:
Arakawa, M., Nishimura, T., Takewa, Y. et al. J Artif Organs (2014) 17: 135. doi:10.1007/s10047-014-0757-1


Right ventricular (RV) failure is a potentially fatal complication after treatment with a left ventricular assist device (LVAD). Ventricular septal shift caused by such devices is an important factor in the progress of RV dysfunction. We developed a control system for a rotary blood pump that can change rotational speed (RS) in synchrony with the cardiac cycle. We postulated that decreasing systolic RS using this system would alter ventricular septal movement and thus prevent RV failure. We implanted the EVAHEART ventricular assist device into seven adult goats weighing 54.1 ± 2.1 kg and induced acute bi-ventricular dysfunction by coronary embolization. Left and RV pressure was monitored, and ventricular septal movement was echocardiographically determined. We evaluated circuit-clamp mode as the control condition, as well as continuous and counter-pulse modes, both with full bypass. As a result, a leftward ventricular septal shift occurred in continuous and counter-pulse modes. The septal shift was corrected as a result of decreased RS during the systolic phase in counter-pulse mode. RV fractional area change improved in counter-pulse (59.0 ± 4.6 %) compared with continuous (44.7 ± 4.0 %) mode. In conclusion, decreased RS delivered during the systolic phase using the counter-pulse mode of our new system holds promise for the clinical correction of ventricular septal shift resulting from a LVAD and might confer a benefit upon RV function.


Artificial organsRotary blood pumpLVADRight ventricular failure


Left ventricular assist device (LVAD) therapy is considered effective for treating patients with end-stage heart failure, either as a bridge to transplantation [1] or as destination therapy [2]. Continuous-flow LVADs are now widely applied, owing to a decrease in complications and improved prognoses compared with pulsatile LVADs [3]. However, differences in physiological effects between continuous-flow and pulsatile LVADs remain controversial [46], especially those involving right ventricular (RV) function [7]. The probability of RV failure does not seem to differ between pulsatile and continuous-flow LVADs [7], but differences in the physiological effects between these devices remain unclear.

RV failure is a life-threatening complication for patients treated with an LVAD [8]. Physiologically, an LVAD is thought to have both beneficial and adverse effects on the right ventricle [9, 10]. One reason for the beneficial effect on RV function is that LVAD unloading leads to a decrease in RV afterload through a decrease in pulmonary artery pressure [11]. However, some reports have indicated that a change in left ventricular (LV) geometry, especially that resulting from a ventricular septal (VS) shift, leads to decreased RV contractility [11, 12]. A VS shift occurs during treatment with LVADs, especially under full or excess LVAD support [10]. Patients with an LVAD who develop RV failure often have concurrent multiple organ failure [13], for which they require high blood-flow support. Therefore, we often encounter the dilemma that increased RS causes RV dysfunction and consequently decreases systemic blood flow. Ultimately, we implant an RV assist device if a patient does not respond to medical therapy [8]. Ideally, an LVAD should provide not only maximum flow support but also RV functional support.

We consider that a change during the LVAD support phase can control RV function, especially that of the ventricular septum. Hence, we developed a novel heart beat-synchronous drive mode for the EVAHEART centrifugal LVAD (Sun Medical Technology Research Corporation, Nagano, Japan) [14]. To create this drive mode, we defined the duration and rotational speed (RS) of the systolic and diastolic phases to drive the EVAHEART in synchrony with the native cardiac cycle [1518]. This driving mode, which is synchronous with the heartbeat, significantly changes coronary flow in normal goats [16, 17]. We speculated that a VS shift under continuous-flow LVAD support would occur during the complete cardiac cycle. However, VS function is more essential in the systolic than the diastolic phase and this novel drive mode might confer a benefit on VS and RV function under full bypass conditions by decreasing RS during the systolic phase.

We investigated whether the new drive mode could prevent RV failure in a goat model of bi-ventricular dysfunction and evaluated interaction between the left and right ventricles, including VS movement.

Materials and methods


Seven adult goats weighing 54.1 ± 2.1 kg were maintained in accordance with the guidelines of the Committee on Animal Studies at the National Cerebral and Cardiovascular Center, and the National Cerebral and Cardiovascular Center Animal Investigation Committee approved the study.

Surgical procedures and implanted devices

All animals were sedated with an intramuscular injection of ketamine hydrochloride (8–10 mg/kg) and anesthetized using isoflurane (1–3 vol %/100 mL in oxygen) inhalation. The animals were placed in the right recumbent position, intubated and mechanically ventilated. We approached the pressure lines for aortic and central venous pressure via a left thoracotomy at the fifth intercostal space from the left intrathoracic artery and vein, respectively. The main pulmonary artery and ascending aorta were dissected and taped, and we attached EMF-1000 electromagnetic flow meters (Nihon Kohden, Tokyo, Japan) with a 12–18-mm diameter to them. After heparinization (300 U/kg), the EVAHEART (Sun Medical Technology Research Corporation, Nagano, Japan) was installed without a cardiopulmonary bypass. First, we sutured the outflow conduit of the device to the descending aorta using a partial clamp. Then we made stitches around the LV apex and punched out the apex with a 16-mm puncher. The inflow cannula was immediately inserted using left hand to control bleeding by squeezing as necessary. We used a 16-mm TS420 ultrasonic flow meter (Transonic Systems, Ithaca, NY, USA) to measure LVAD flow (PF). Total flow (TF) was calculated as the sum of the ascending aortic flow and the PF. The bypass rate was calculated by dividing the PF by the TF.

To obtain pressure–volume loops, we inserted a 6-Fr conductance catheter (2S-RH-6DA-116; Taisho Biomed Instruments Co., Ltd., Osaka, Japan) into the RV from the apex towards the pulmonary valve, and 4-Fr micro-tip catheter pressure transducers (Millar Instruments Inc., Houston, TX, USA) were inserted through the RV and LV walls. In addition, pulmonary artery pressure was monitored via a catheter inserted directly into the artery. Hemodynamic data were recorded using LabChart 7 software (ADInstruments, Castle Hill, New South Wales, Australia). The pressure and volume data of the LV were recorded using a Leycom Sigma 5 system (CardioDynamics, Zoetermeer, The Netherlands). We defined LV and RV systolic duration (%) as the sum of the isovolemic contraction phase and the ejection phase of the respective ventricles divided by the RR interval (Fig. 2). Because blood was ejected during the diastolic phase under LVAD support, we defined the end of ejection phase as the first point when RV volume reached diastolic volume and the first point when the ventricular pressure waveform sharply decreased (Fig. 2). The corresponding systolic duration was measured using LabChart software.

Bi-ventricular dysfunction model

We modeled bi-ventricular dysfunction in this study. We impaired RV and LV function by embolizing the coronary artery through the left anterior descending (LAD) and right coronary (RCA) arteries as described [18]. Briefly, a 4-Fr Amplatz catheter (Create Medic Co., Ltd., Yokohama, Japan) was inserted through a 4-Fr long sheath into the left carotid artery toward the LAD and RCA ostia under fluoroscopic guidance. Microspheres (75-μm diameter) were injected into the LAD (60 spheres/kg) and RCA (30 spheres/kg). The general condition of each goat was observed for a further 30 min to stabilize cardiac function before collecting data.

We controlled the infusion volume and depth of anesthesia during the experiment to maintain stable PF and aortic and central venous pressures, thus avoiding suction at the inflow cannula and sustaining an appropriate after load. All animals were administered with lidocaine 2 % (1 mg/kg/h) during the experiment to prevent ventricular arrhythmia. Heart failure was established to reduce and then maintain cardiac output at approximately 60 % of the cardiac output as TF and PAF according to previously described methods [18].

Study protocol and LVAD control

Our new controller allows the RS of the EVAHEART to change for a specific duration after an R wave appears to synchronize it with the cardiac cycle. We defined the systolic and diastolic phases of the cardiac cycle as 33 and 67 % of the RR interval, respectively, (Fig. 1) according to the published protocol [1518]. The pressure and flow parameters described above were evaluated twice in each goat under the following LVAD controller modes (Fig. 1): circuit-clamp (no pump support due to clamping the LVAD circuit) as the control condition, continuous (constant RS without using the new controller) and counter-pulse (RS decreased during systolic phase using the new controller). We set the RS during the systolic phase to 700–1,000 rpm in the counter-pulse mode and increased the RS in the diastolic phase to achieve an appropriate bypass rate, which was calculated by dividing the PF rate by the sum of the flow rates of the PF and the ascending aorta. We defined full bypass as a rate of about 100 % (range 90–110 %).
Fig. 1

Rotary blood-pump control modes. Circuit-clamp mode (control) does not include pump support because left ventricular assist device circuit is clamped. Rotational speed in continuous mode remains constant without a controller and that during systolic phase in counter-pulse mode is decreased by a controller. Systolic and diastolic phases in cardiac cycle are defined as 33 and 67 %, respectively, of RR interval


We evaluated RV and LV dimensions by echocardiography using a Vivid E9 system and M5S-D sector transducer (GE Healthcare, Horten, Norway). The end-diastolic and end-systolic RV areas were measured in the direct cardiac long-axis view. The %fractional area change was calculated as:
$$\frac{{{\text{end - diastolic RV area}} - {\text{end - systolic RV area}}}}{\text{end - diastolic RV area}} \times 100$$

Statistical analysis

Numerical data are shown as mean ± SE. Differences in values between groups were analyzed by repeated measures analysis of variance followed by Tukey’s multiple comparison test. Statistical significance was accepted at a probability value of <0.05. All data were analyzed using SPSS version 19 (SPSS Inc., Chicago, IL, USA).


Bi-ventricular dysfunction models were established by coronary embolism in which TF decreased from 3.3 ± 0.4 to 2.0 ± 0.3 L/min (p < 0.01) in the circuit-clamp mode. The dP/dt, particularly in the RV, fell from 436 ± 40 to 341 ± 48 mmHg/s (p < 0.05) and Emax fell from 1.35 ± 0.16 to 0.87 ± 0.19 (p < 0.05) in circuit-clamp mode.

Figure 2 shows sample waveforms of the ECG, R waves, RV pressure, LV pressure, PF, and RS. Actual RS in the counter-pulse mode closely followed the command RS, since average RS in the systolic and the diastolic phase was close. LV pressure waveforms were decreased in the continuous mode. The amplitude of the LV pressure waveforms was similar between counter-pulse and circuit-clamp modes. The duration of the systolic phase differed among the three groups (Fig. 3). The duration of systole between the right and left ventricles did not significantly differ in circuit-clamp mode (45.8 ± 2.8 vs. 42.9 ± 3.2 %, respectively). The LV systolic phase was significantly shorter than the RV systolic phase in the continuous mode and the LV systolic phase in circuit-clamp mode (LV vs. RV: 20.2 ± 1.8 vs. 45.0 ± 2.5 %, p = 0.01). The LV systolic phase was significantly shorter in the counter-pulse, than in the circuit-clamp mode, and significantly longer than in the continuous mode (RV 45.9 ± 2.7 vs. LV 36.5 ± 2.6 %; p = 0.01) (Fig. 3).
Fig. 2

Sample pressure and flow waveforms and pressure–volume loops for each mode. a Left ventricular systolic duration is decreased in continuous mode compared with that in circuit-clamp mode. Duration of LV systolic phase is longer in counter-pulse than in continuous mode. LV left ventricular pressure, PF pump flow, RS rotational speed, RVP right ventricular pressure. b, c Left and right ventricular systolic duration (%) is the sum of isovolumic contraction and ejection phases of respective ventricles divided by cardiac cycle
Fig. 3

Systolic duration per mode. Left ventricular systolic phase is significantly shorter than RV systolic phase in continuous mode, and shorter than LV systolic phase in circuit-clamp mode. Left ventricular systolic phase is significantly shorter in counter-pulse than in circuit-clamp mode, and longer than in continuous mode. LV left ventricle, RV right ventricle

Table 1 shows the LV and RV data. The LV end-diastolic pressure was significantly decreased in LVAD modes (continuous and counter-pulse modes) compared with circuit-clamp mode. The RV end-diastolic pressure and volume did not significantly differ among the three modes. In addition, pulmonary artery pressure slightly decreased and pulmonary artery flow was slightly higher during the LVAD modes than in circuit-clamp mode, although mean RS in the counter-pulse mode was significantly lower than that of the continuous mode.
Table 1

Hemodynamic data obtained in each drive mode








 LVEDP (mmHg)

8.8 ± 1.7

4.7 ± 1.4

4.4 ± 1.2

 Mean AoP (mmHg)

59.7 ± 3.0

64.0 ± 3.8

66.0 ± 4.6

 Pump flow (L/min)

0.0 ± 0.0

2.4 ± 0.2a

2.4 ± 0.2b

 Total flow (L/min)

2.0 ± 0.3

2.3 ± 0.3

2.3 ± 0.3

 Bypass rate (%)

0.0 ± 0.0

107.5 ± 2.4a

101.8 ± 3.7b


 RVEDP (mmHg)

8.4 ± 1.8

7.9 ± 1.8

8.0 ± 2.0

 RVEDV (ml)

75.1 ± 7.8

75.0 ± 8.1

77.8 ± 7.5

 dP/dt max (mmHg/s)

341 ± 48

361 ± 54

381 ± 57


0.87 ± 0.19

0.93 ± 0.22

1.05 ± 0.20

 PVA (mmHg/mL)

474 ± 50

522 ± 61

519 ± 50

 Mean PAP (mmHg)

27.6 ± 1.9

26.2 ± 0.9

25.4 ± 1.9

 PA flow (L/min)

2.1 ± 0.3

2.3 ± 0.3

2.4 ± 0.3



 Actual RS (command RS)

  Systolic RS (rpm)

0 ± 0

1557 ± 53a

(1588 ± 38a)

700 ± 23b,c

(719 ± 12b,c)

  Diastolic RS (rpm)

0 ± 0

1557 ± 53a

(1588 ± 38a)

1627 ± 65b

(1612 ± 46b,c)

  Mean RS (rpm)

0 ± 0

1557 ± 53a

(1588 ± 38a)

1290 ± 50b,c

(1292 ± 36b,c)

AoP arterial pressure, LVEDP left ventricular end-diastolic pressure, PA pulmonary artery, PAP pulmonary artery pressure, PVA pressure volume area, RS rotational speed, RVEDP right ventricular end-diastolic pressure, RVEDV right ventricular end-diastolic volume

ap < 0.01: circuit-clamp vs. continuous mode

bp < 0.01: circuit-clamp vs. counter-pulse mode

cp < 0.01: continuous vs. counter-pulse mode

RV free wall movement did not significantly change in counter-pulse mode, but VS movement differed among the groups in the direct cardiac short- and long-axis views (Fig. 4). The VS shifted towards the left ventricle in LVAD modes. However, the VS shift was corrected as a result of RS being decreased during the systolic phase in the counter-pulse mode. The end-diastolic RV areas were greater in the LVAD than in circuit-clamp mode. Therefore, the %fractional area change was greater in counter-pulse than in circuit-clamp and continuous modes (59.0 ± 4.6 vs. 56.1 ± 8.8 and 44.7 ± 4.0 %; Fig. 5).
Fig. 4

Sample echocardiograms obtained in each mode. Leftward ventricular septal (VS) shift is evident in continuous and counter-pulse modes (a, b, d, e), but corrected by decreased rotational speed during systolic phase in counter-pulse mode (c, f)
Fig. 5

Bar graph shows right ventricular fractional area change with each mode, **p < 0.01


To our knowledge, this is the first echocardiographic demonstration of a VS shift being corrected towards the left ventricle in the systolic phase using counter-pulse mode applied to an LVAD. The shift itself might exacerbate RV dysfunction. We consider that LVAD unloading was a major cause of the VS shift, since the duration of the systolic phase was remarkably reduced in continuous mode. The decreased systolic RS provided by the counter-pulse mode significantly corrected LV systolic duration and RV %fractional area change compared with continuous mode. We found that the shortened systolic duration in the continuous mode was corrected in the counter-pulse mode as re-synchronization between the RV and the LV. Thus, we presumed that this favorable effect on RV function was caused by re-synchronizing RV and LV contractions.

Others have discussed the question of whether or not RV function is diminished in patients treated with an LVAD [911], especially of the continuous-flow type [7]. We did not find adverse effects on any parameter of RV contractility, such as Emax, dP/dt and pulmonary artery flow. Umeki et al. have demonstrated that the same counter-pulse mode decreases LV end-diastolic volume (LVEDV). They argued the decreased LVEDV and LVEDP directed the unloading effect according to the Staring law. In addition, lower LVEDP decreased RV afterload. However, improvements in RV contractility, such as increased Emax or RV diastolic volume, were not evident in counter-pulse compared with continuous mode, although pulmonary flow was slightly increased. In contrast, we confirmed by echocardiography that the novel counter-pulse mode used with the EVAHEART corrected the VS shift during the systolic phase while maintaining a full bypass without a decrease in total systemic flow or adversely affecting the RV. We considered that the counter-pulse mode did not improve RV function because this experiment proceeded under open-chest conditions. That is to say, this mode did not affect the ventricular free wall, which in the right ventricle is important for RV pressure and volume. Taken together, this mode might confer benefits on the right ventricle through the ventricular septum. Park et al. showed that RV performance was impaired by full assist of a left heart bypass with continuous-flow LVAD [10]. Yoshioka et al. [13] discovered that patients with an LVAD who develop RV failure often had concurrent multiple organ failure, for which they required high blood-flow support. We often encounter the dilemma that full LV support causes RV dysfunction and consequently decreases systemic blood flow. Ultimately, some patients required RV assist devices if they did not respond to medical therapy [8]. Our counter-pulse mode can provide not only maximum flow support but also optimized VS movement by only LVADs. We believe that the counter-pulse mode can resolve these issues.

Electrocardiography-synchronized RS control confers several advantages on hemodynamics as well as native heart function compared with continuous RS. Ando et al. [16] created pulsatility using their pulsatile mode with increased RS. Kishimoto et al. [19] reported that speed modulation fully opens the aortic valve and thus prevents aortic insufficiency. Electrocardiography-synchronized RS control of the rotary blood pump confers several beneficial effects compared with the conventional constant RS drive mode. Here, we demonstrated that the system has potential RV support by VS shift correction in addition to the advantages described above. RS control provides further treatment options for circulatory support therapy with rotary blood pumps.

Our study had several limitations. We used an animal model of bi-ventricular dysfunction that developed decreased contractility including that of the ventricular septum. Although VS movement could be evaluated in this model, we could not assess the septal thinning that is a frequent feature of patients with an enlarged LV treated with an LVAD. The next step will be to evaluate VS shift using the new control system in a model with a thin ventricular septum. The experiments proceeded under open-chest condition under which, RV pressure might be difficult to evaluate, since RV pressure remarkably changed according to pressure in the pericardial space. Therefore, the effect of the novel control system should be assessed in a model of chronic heart failure with a thinned ventricular septum under closed-chest conditions.


The novel counter-pulse mode optimized LV unloading and thus corrected VS movement while maintaining sufficient blood flow. This might prevent the RV failure that frequently complicates patients treated with an LVAD, especially during the unstable acute phase after LVAD implantation.


This study was presented at the annual conference of The American Society for Artificial Internal Organs in 2012, in San Francisco, CA and supported by a JSAO Grant in 2013. The authors are grateful to Dr. Daisuke Ogawa at the Sun Medical Technology Research Corporation for creating the programs.

Conflict of interest

The authors have no conflicts of interest to disclose.

Copyright information

© The Japanese Society for Artificial Organs 2014