European Journal of Cardio-Thoracic Surgery Advance Access published May 16, 2015
ORIGINAL ARTICLE
European Journal of Cardio-Thoracic Surgery (2015) 1–7 doi:10.1093/ejcts/ezv171
A modified Glenn shunt reduces right ventricular stroke work during left ventricular assist device therapy Petter Schiller, Per Vikholm and Laila Hellgren* Department of Cardiothoracic Surgery, Uppsala University Hospital, Uppsala, Sweden * Corresponding author. Department of Cardiothoracic Surgery, Uppsala University Hospital, S-755 91 Uppsala, Sweden. Tel: +46-186114016; fax: +46-18506143; e-mail: [email protected] (L. Hellgren). Received 25 November 2014; received in revised form 26 March 2015; accepted 2 April 2015
Abstract OBJECTIVES: Right ventricular (RV) failure is a major cause of morbidity and mortality after left ventricular assist device (LVAD) placement and remains hard to predict. We hypothesized that partial surgical exclusion of the RV with a modified Glenn shunt during LVAD treatment would reduce RV stroke work. METHODS: An LVAD was implanted in eight pigs and a modified Glenn shunt was constructed. A conductance pressure–volume catheter was placed in the right ventricle through the apex. Haemodynamic data and pressure–volume loops were obtained at the following time periods: (i) baseline, (ii) open shunt, (iii) LVAD with closed shunt and (iii) LVAD and open shunt. RESULTS: During LVAD therapy, the right atrial (RA) pressure increased from 9 mmHg (9–9) to 15 mmHg (12–15), P = 0.01. RV stroke volume increased from 30 ml (29–40) to 51 ml (42–53), P < 0.01. Also, RV stroke work increased to 708 mmHg ml (654–1193) from 535 mmHg ml (424–717), P = 0.04, compared with baseline. During LVAD therapy in combination with a Glenn shunt, the RA pressure decreased from 15 mmHg (12–15) to 10 mmHg (7–11) when compared with LVAD therapy only, P = 0.01. A decrease in RV stroke work from 708 mmHg ml (654–1193) to 465 mmHg ml (366–711), P = 0.04, was seen when the LVAD was combined with a shunt, not significantly different from the baseline value (535 mmHg ml). The developed pressure in the right ventricle decreased from 29 mmHg (26–32) to 21 mmHg (20–24), P < 0.01. The pressure–volume loops of the RV show a significant reduction of RV stroke work during the use of the shunt with LVAD treatment. CONCLUSIONS: A modified Glenn shunt reduced RV volumes, RV stroke work and RA pressure during LVAD therapy in an experimental model of heart failure in pigs. Keywords: Right heart failure • Glenn shunt • Left ventricular assist device • Right ventricular stroke work
INTRODUCTION Left ventricular assist device (LVAD) treatment has become the standard therapy for end-stage heart failure, most often as a bridge to transplant but also as destination therapy. Despite the clinical and physiological benefits of LVAD therapy, right ventricular (RV) failure of varying degree occurs in 20% of patients, and causes substantial morbidity and mortality [1–3]. In addition, the lack of donor hearts is an increasing issue leading to prolonged LVAD therapy in a number of patients and thereby increased risk for late RV failure. LVAD treatment might also worsen pre-existent or subclinical RV failure, and in many patients, RV failure becomes clinically evident once LVAD support has been initiated. The standard therapy for RV failure includes inotropic support and right ventricular assist device (RVAD). Although results with RVAD therapy have improved over time to even include long-term treatment in selected cases, the fact still remains that patients in
need of RVAD have reduced survival and complications include bleeding, infection and multiple organ failure. Several studies have been conducted with the purpose of establishing risk factors or risk scores for identification of patients at a high risk for RV failure, but many aspects of this problem still remain unsolved. One of the important preoperative risk factors for RV failure on the LVAD is a preoperative low RV stroke work (RVSW), which prevents the RV from adapting to the increased haemodynamic burden on it when LVAD therapy is initiated. Surgical exclusion of RV volume has been proven possible in the staged palliation of congenital heart defects when there is no left ventricular impairment and the pulmonary vascular resistance (PVR) is low [4]. This principle has been used for decades, with a Glenn shunt followed by the Fontan procedure or total cavopulmonary connection [5]. Moreover, some case reports describe successful relief of ischaemic RV failure in adults undergoing heart surgery by volume unloading of the RV with a shunt [6–8]. With regard to LVAD patients with preoperative low or normal PVR,
© The Author 2015. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
TX & MCS
Cite this article as: Schiller P, Vikholm P, Hellgren L. A modified Glenn shunt reduces right ventricular stroke work during left ventricular assist device therapy. Eur J Cardiothorac Surg 2015; doi:10.1093/ejcts/ezv171.
2
P. Schiller et al. / European Journal of Cardio-Thoracic Surgery
volume unloading of the RV with a shunt could be an additional strategy in 40% of the candidates [9, 10] when managing intraoperative RV failure. This study investigates the possibility of using the strategy with a modified Glenn shunt to reduce the increased volume load on the RV induced by LVAD treatment in an experimental model. We hypothesize that partial surgical exclusion of the RV volume with a Glenn shunt during LVAD treatment would reduce RV stroke work.
MATERIALS AND METHODS The study was approved by the Uppsala University Ethical Committee on Laboratory Animal Research. All animals received humane care in compliance with the European Convention on Animal Care and the Principles of Laboratory Animal Care. These principles are formulated by the National Society for Medical Research and the ‘Guide for the Care and Use of Laboratory Animals’ prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH Publication No. 86-23, revised 1996). The study comprised eight pigs of Swedish country breed with a mean weight of 34 kg (31–39).
Anaesthesia The method of anaesthesia has been described in previous studies of ours [11, 12]. The animals were intubated and thereafter ventilated with 40% oxygen through a Siemens Servo-i ventilator (Maquet GmbH & Co., KG, Rastatt, Germany). Ventilation was volume controlled and aimed for an arterial pCO2 between 5.0 and 5.5 kPa. Moreover, a positive end-expiratory pressure of 5 cm was used. General anaesthesia was induced by a subcutaneous injection of xylazine (Rompun® 2.2 mg/kg, Bayer A/S, Lyngby, Denmark), and tiletamine/zolazepam (Zoletil 100® 6.0 mg/kg, Virbac S.A., Carros, France). Anaesthesia was then maintained by infusion of a buffered glucose carrier (Rehydrex® with glucose 25 mg/ml, Fresenius Kabi AB, Uppsala, Sweden), fentanyl 0.04 mg/kg/h (Fentanyl 50 µg/ml, B Braun Medical AB, Danderyd, Sweden), ketamine 30 mg/kg/h (Ketaminol Vet® 100 mg/ml, Intervet AB, Stockholm, Sweden), midazolam 0.1 mg/kg/h (Midazolam Hameln 1 mg/ml, AlgolPharma AB, Kista, Sweden) and pancuronium bromide 0.3 mg/kg/h (Pavulon® 2 mg/ml, Schering-Plough AB, Stockholm, Sweden). Body temperature was stabilized with a heating pad, aiming for a core temperature of 37°C. A silicone catheter was used to catheterize the urinary bladder. The jugular vein was catheterized (BD Careflow 17 G) for blood sampling, monitoring of central venous pressure and administration of drugs. A pulmonary artery (PA) catheter (BD Criticath™ Pulmonary Artery/Thermodilution Catheter) was inserted through the right jugular vein and placed into the pulmonary artery. Continuous monitoring of the arterial blood pressure with a pressure transducer (BD Careflow 20G; Becton-Dickinson AB, Stockholm, Sweden) was undertaken with a catheter in the left femoral artery.
were connected to a monitoring system (Dräger Infinity Delta). A 5-French 12-electrode pressure–volume conductance catheter (Ventri-Cath, Millar Instruments, MI, USA) was introduced in the RV apex and advanced to a level just below the pulmonary valve. The catheter was fixed with a purse-string suture and connected to the Millar Pressure Volume Systems (MPVS) Ultra analysing system. Proper catheter positioning was determined by monitoring individual segmental pressure–volume loops. Extra ventricular segments were excluded from analysis. Animals were heparinized with a bolus of 7500 IU once. The modified Glenn shunt was created by connecting the superior vena cava (SVC) to the main PA with a 12 mm Dacron vascular graft (Gelsoft™, Vascutek Inchinnan, Scotland, UK). A sidebiting clamp was placed first on the SVC and then on the main PA and thereafter the vessels were opened longitudinally and anastomosed with a 4-0 running suture. The clamp was released and the patency of the shunt was evaluated with a 12 mm ultrasonic flow probe (VeriQ Perivascular 12 mm, MediStim ASA, Oslo, Norway). The graft was then closed with a vascular clamp. When the shunt was put in use, the vascular clamp was removed from the shunt, and the SVC was then clamped below the shunt, directing all blood from the SVC to the main PA through the modified Glenn shunt (Fig. 1). The left ventricular apex was cannulated with a 28-Fr venous cannula (Malleable Single Stage Venous Cannula, Edwards Lifesciences) and a 16-Fr arterial cannula (FemFlex, Edwards Lifesciences) was inserted into the ascending aorta. The cannulas were secured with purse-string sutures. The left ventricular apex cannula was connected to the inflow tubing of a preprimed Bio-Pump (Medtronic). The aortic cannula was deaired and connected to the outflow tubing of the Bio-Pump. A Bio-Pump flow probe connector (Medtronic) was connected via the tubing to the aortic cannula.
Measurements Standard lead II electrocardiogram (ECG), heart rate (HR), mean systemic arterial blood pressure (MAP), RA pressure, RV pressure, LA pressure and PA pressure were monitored continuously and recorded. Cardiac output (CO) was measured to 3.3 l/min (range, 2.8–3.4) at baseline using the thermodilution technique. The pressure–volume
Surgical preparation A median sternotomy was performed. The pericardium was opened and pericardial stay sutures were placed. The azygos and hemiazygos veins were ligated. Pressure transducer catheters were surgically placed in the right atrium (RA) and the RV. The catheters
Figure 1: Animal preparation; a modified Glenn shunt was created between the SVC and the main pulmonary artery. A left ventricular assist device was implanted. Pressure volume loops were obtained throughout the experimental model in pigs (n = 8).
P. Schiller et al. / European Journal of Cardio-Thoracic Surgery
Experimental protocol Shunt. The clamp on the vascular graft was released and the SVC was clamped at the junction with the RA to direct all blood flow from SVC to the main PA. Shunt flow was measured with a 12 mm ultrasonic flow probe. The Glenn shunt was opened for 20 min and thereafter measurements were performed. Left ventricular assist device. The Glenn shunt was closed and the clamp on the SVC was released. The LVAD was started and set to a flow corresponding to the baseline CO of each animal measured with the thermodilution technique. This was continued for 20 min and thereafter measurements were performed. Left ventricular assist device combined with a modified Glenn shunt. The clamp on the Glenn shunt was released and the SVC was clamped at the junction with the RA to direct all blood flow from SVC to the main PA. Shunt flow was measured with a 12 mm ultrasonic flow probe. Measurements were recorded after 20 min of stabilization.
Data analysis and statistical methods For pressures (i.e. SVCP, RAP, mPAP and MAP), HR, LVAD flow and Glenn-shunt flow, the value for each pig was calculated as the mean of all measurements for each treatment period, and data are expressed as medians with interquartile ranges in brackets. For blood gas measurements (i.e. pH, PO2, PCO2, SBE, SaO2, SvO2 and Hb) the data are expressed as medians with interquartile ranges in brackets, based on the value obtained for each pig in each treatment period. Data analysis of the pressure–volume loops was performed using Lab Chart™ software (AD Instruments Pty Ltd, Bella Vista, NSW, Australia). As there was no calibration for parallel conductance, the baseline RV ejection fraction was estimated to be 0.6 and the volumes were calibrated according to this figure. With regard to conductance catheter-derived values, the mean value of the loops for each pig in each treatment period was used. The data are expressed as medians with interquartile ranges in brackets. Statistical analysis was conducted using the statistical software R (v. 3.1.1). Non-parametric tests were used due to the small sample size. Comparison was done between (i) baseline versus LVAD and (ii) LVAD versus LVAD and modified Glenn shunt, using the Wilcoxon signed-rank test. P-values ≤0.05 were considered statistically significant.
RESULTS All eight pigs survived the experiment and were included in the study.
Baseline All animals were in a stable haemodynamic state at baseline. The MAP was 86 mmHg (74–99), SvO2 was 58% (46–62) and CO was 2.8 l/min (2.3–3.2), as depicted in Table 1. RA pressure was 9 mmHg (9–9) (Fig. 2). Systolic RV pressure was 31 mmHg (28–33) and diastolic RV pressure was 7 mmHg (6–7) at baseline. RV enddiastolic volume was 50 ml (49–67) and RV stroke volume was 30 ml (29–40), as shown in Figs 3 and 4. Moreover, RV stroke work at baseline was 535 mmHg ml (442–717) (Fig. 5). The pressure development (Pdev) in the RV was 26 mmHg (22–28), as shown in Fig. 6. The representative pressure–volume loops at baseline are shown in Fig. 7.
A modified Glenn shunt without left ventricular assist device The median flow in the shunt was 0.57 l/min (0.48–0.74), Table 1. There were no significant changes in the MAP, SvO2, RA or RV pressure during the phase of Glenn shunting when compared with baseline, Table 1. Moreover, the RV pressure–volume loops did not change significantly during shunting, there was, however, a trend towards decreased RV stroke work, as shown in Figs 5 and 7.
Left ventricular assist device therapy without shunt During LVAD therapy alone (with a closed shunt), the pump flow was 3.3 l/min (2.9–4.0). The MAP increased from 86 mmHg (74–99) to 98 mmHg (88–103) (P = 0.04), and SvO2 increased from 58% (46–52) to 64% (58–70) (P = 0.02) when LVAD therapy alone was compared with baseline values; Table 1. LVAD therapy increased the RA pressure from 9 mmHg (9–9) to 15 mmHg (12–15) (P = 0.01) (Fig. 2), and RV end-diastolic volume from 50 ml (49–67) to 85 ml (55–88) (P = 0.05) (Fig. 3). In addition, there was an increase in the RV stroke volume from 30 ml (29–40) to 51 ml (42–53) (P < 0.01) (Fig. 4). RV stroke work increased from the baseline value at 535 mmHg ml (424–717) to 708 mmHg ml (654–1193) (P = 0.04), as shown in Fig. 5. Finally, there was no increase in the development of pressure in the RV during LVAD therapy from 26 mmHg (22–28) to 29 mmHg (26–32) (P = 0.14) (Fig. 6). The pressure–volume loops for LVAD alone significantly shifted to the right when compared with baseline, as shown in Fig. 7.
Left ventricular assist device therapy combined with modified Glenn shunt The LVAD flow was not altered as the shunt was opened. There was a reduction in the MAP with the LVAD and Glenn shunt from 98 mmHg (88–103) to 84 mmHg (74–89) when compared with LVAD therapy alone (P = 0.02); Table 1. Also, there was a decrease in SvO2 from 64% (58–70) with the LVAD alone to 46% (35–55) with the LVAD and shunt (P < 0.01); Table 1.
TX & MCS
catheter was calibrated for blood resistivity using a rho-cuvette (Millar Instruments, MI, USA) and the alpha-value was calculated using the thermodilution-derived CO for each animal. Pressure– volume loops from the RV were recorded for 25 consecutive heartbeats with the ventilation suspended at end-expiration. Haemodynamic measurements and pressure–volume curves were recorded at the following time periods: (i) at baseline, (ii) at shunting, (iii) with the LVAD and a closed Glenn shunt and (iv) with the LVAD and an open Glenn shunt. The CO values presented in the results are derived from the pressure–volume curves. Blood samples were collected at the same time points for blood gas analysis. Arterial blood gases, haemoglobin concentration (Hb) and mixed venous oxygen saturation (SvO2) were measured with an ABL 500 Radiometer, Medical ApS, Bronshoj, Denmark.
3
P. Schiller et al. / European Journal of Cardio-Thoracic Surgery
4
Table 1: The phases of the experimental study on LVAD in combination with a modified Glenn shunt, n = 8 Parameter
Baseline
Modified Glenn-shunt
LVAD
LVAD + modified Glenn-shunt
Baseline vs LVAD, P-values
LVAD vs LVAD + modified Glenn-shunt, P-values
MAP (mmHg) RVCO (l/min) CO (l/min) SVCP (mmHg) RVEDP (mmHg) RVESP (mmHg) mPAP (mmHg) RVESV (ml) RVEDV (ml) RVSV (ml) RVSW (mmHg ml) HR (beats/min) LVAD (l/min) Modified Glenn-shunt (l/min) pH PO2 (kPa) PCO2 (kPa) SBE (mmol/l) SaO2 (%) SvO2 (%) Hb (g/l)
86 (74–99) 2.8 (2.2–3.2) 2.8 (2.2–3.2) 9 (9–9) 7 (6–7) 31 (28–33) 18 (17–19) 20 (20–27) 50 (49–67) 30 (29–40) 535 (424–717) 84 (76–91) N/A N/A 7.47 (7.47–7.48) 43 (42–44) 5.5 (5.2–5.6) 5.8 (5.0–6.0) 99 (98–99) 58 (46–62) 77 (70–82)
96 (86–101) 2.9 (2.1–3.9) 3.4 (2.7–4.4) 20 (20–24) 6 (6–7) 28 (27–29) N/A 16 (12–23) 51 (39–61) 32 (21–40) 461 (271–718) 93 (90–98) N/A 0.57 (0.48–0.74) 7.46 (7.45–7.48) 41 (40–44) 5.5 (5.2–5.5) 4.8 (3.8–6.0) 100 (99–100) 52 (43–56) 74 (65–77)
98 (88–103) 4.8 (3.7–5.8) 4.8 (3.7–5.8) 14 (13–14) 6 (6–7) 35 (32–39) N/A 24 (11–34) 85 (55–88) 49 (42–53) 708 (654–1193) 90 (89–101) 3.3 (2.9–4.0) N/A 7.46 (7.45–7.49) 42 (39–43) 5.5 (5.2–5.8) 5.2 (4.0–6.4) 99 (99–99) 64 (58–70) 51 (48–57)
84 (74–89) 3.4 (2.8–4.2) 4.0 (3.3–4.7) 18 (18–23) 6 (6–7) 28 (26–29) N/A 22 (14–29) 60 (48–69) 34 (33–38) 465 (366–711) 93 (90–108) 3.3 (2.9–4.0) 0.58 (0.43–0.70) 7.46 (7.44–7.48) 43 (40–43) 5.5 (5.2–5.9) 6.0 (4.0–6.6) 100 (99–100) 46 (35–55) 57 (55–60)
0.04
ORIGINAL ARTICLE
European Journal of Cardio-Thoracic Surgery (2015) 1–7 doi:10.1093/ejcts/ezv171
A modified Glenn shunt reduces right ventricular stroke work during left ventricular assist device therapy Petter Schiller, Per Vikholm and Laila Hellgren* Department of Cardiothoracic Surgery, Uppsala University Hospital, Uppsala, Sweden * Corresponding author. Department of Cardiothoracic Surgery, Uppsala University Hospital, S-755 91 Uppsala, Sweden. Tel: +46-186114016; fax: +46-18506143; e-mail: [email protected] (L. Hellgren). Received 25 November 2014; received in revised form 26 March 2015; accepted 2 April 2015
Abstract OBJECTIVES: Right ventricular (RV) failure is a major cause of morbidity and mortality after left ventricular assist device (LVAD) placement and remains hard to predict. We hypothesized that partial surgical exclusion of the RV with a modified Glenn shunt during LVAD treatment would reduce RV stroke work. METHODS: An LVAD was implanted in eight pigs and a modified Glenn shunt was constructed. A conductance pressure–volume catheter was placed in the right ventricle through the apex. Haemodynamic data and pressure–volume loops were obtained at the following time periods: (i) baseline, (ii) open shunt, (iii) LVAD with closed shunt and (iii) LVAD and open shunt. RESULTS: During LVAD therapy, the right atrial (RA) pressure increased from 9 mmHg (9–9) to 15 mmHg (12–15), P = 0.01. RV stroke volume increased from 30 ml (29–40) to 51 ml (42–53), P < 0.01. Also, RV stroke work increased to 708 mmHg ml (654–1193) from 535 mmHg ml (424–717), P = 0.04, compared with baseline. During LVAD therapy in combination with a Glenn shunt, the RA pressure decreased from 15 mmHg (12–15) to 10 mmHg (7–11) when compared with LVAD therapy only, P = 0.01. A decrease in RV stroke work from 708 mmHg ml (654–1193) to 465 mmHg ml (366–711), P = 0.04, was seen when the LVAD was combined with a shunt, not significantly different from the baseline value (535 mmHg ml). The developed pressure in the right ventricle decreased from 29 mmHg (26–32) to 21 mmHg (20–24), P < 0.01. The pressure–volume loops of the RV show a significant reduction of RV stroke work during the use of the shunt with LVAD treatment. CONCLUSIONS: A modified Glenn shunt reduced RV volumes, RV stroke work and RA pressure during LVAD therapy in an experimental model of heart failure in pigs. Keywords: Right heart failure • Glenn shunt • Left ventricular assist device • Right ventricular stroke work
INTRODUCTION Left ventricular assist device (LVAD) treatment has become the standard therapy for end-stage heart failure, most often as a bridge to transplant but also as destination therapy. Despite the clinical and physiological benefits of LVAD therapy, right ventricular (RV) failure of varying degree occurs in 20% of patients, and causes substantial morbidity and mortality [1–3]. In addition, the lack of donor hearts is an increasing issue leading to prolonged LVAD therapy in a number of patients and thereby increased risk for late RV failure. LVAD treatment might also worsen pre-existent or subclinical RV failure, and in many patients, RV failure becomes clinically evident once LVAD support has been initiated. The standard therapy for RV failure includes inotropic support and right ventricular assist device (RVAD). Although results with RVAD therapy have improved over time to even include long-term treatment in selected cases, the fact still remains that patients in
need of RVAD have reduced survival and complications include bleeding, infection and multiple organ failure. Several studies have been conducted with the purpose of establishing risk factors or risk scores for identification of patients at a high risk for RV failure, but many aspects of this problem still remain unsolved. One of the important preoperative risk factors for RV failure on the LVAD is a preoperative low RV stroke work (RVSW), which prevents the RV from adapting to the increased haemodynamic burden on it when LVAD therapy is initiated. Surgical exclusion of RV volume has been proven possible in the staged palliation of congenital heart defects when there is no left ventricular impairment and the pulmonary vascular resistance (PVR) is low [4]. This principle has been used for decades, with a Glenn shunt followed by the Fontan procedure or total cavopulmonary connection [5]. Moreover, some case reports describe successful relief of ischaemic RV failure in adults undergoing heart surgery by volume unloading of the RV with a shunt [6–8]. With regard to LVAD patients with preoperative low or normal PVR,
© The Author 2015. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
TX & MCS
Cite this article as: Schiller P, Vikholm P, Hellgren L. A modified Glenn shunt reduces right ventricular stroke work during left ventricular assist device therapy. Eur J Cardiothorac Surg 2015; doi:10.1093/ejcts/ezv171.
2
P. Schiller et al. / European Journal of Cardio-Thoracic Surgery
volume unloading of the RV with a shunt could be an additional strategy in 40% of the candidates [9, 10] when managing intraoperative RV failure. This study investigates the possibility of using the strategy with a modified Glenn shunt to reduce the increased volume load on the RV induced by LVAD treatment in an experimental model. We hypothesize that partial surgical exclusion of the RV volume with a Glenn shunt during LVAD treatment would reduce RV stroke work.
MATERIALS AND METHODS The study was approved by the Uppsala University Ethical Committee on Laboratory Animal Research. All animals received humane care in compliance with the European Convention on Animal Care and the Principles of Laboratory Animal Care. These principles are formulated by the National Society for Medical Research and the ‘Guide for the Care and Use of Laboratory Animals’ prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH Publication No. 86-23, revised 1996). The study comprised eight pigs of Swedish country breed with a mean weight of 34 kg (31–39).
Anaesthesia The method of anaesthesia has been described in previous studies of ours [11, 12]. The animals were intubated and thereafter ventilated with 40% oxygen through a Siemens Servo-i ventilator (Maquet GmbH & Co., KG, Rastatt, Germany). Ventilation was volume controlled and aimed for an arterial pCO2 between 5.0 and 5.5 kPa. Moreover, a positive end-expiratory pressure of 5 cm was used. General anaesthesia was induced by a subcutaneous injection of xylazine (Rompun® 2.2 mg/kg, Bayer A/S, Lyngby, Denmark), and tiletamine/zolazepam (Zoletil 100® 6.0 mg/kg, Virbac S.A., Carros, France). Anaesthesia was then maintained by infusion of a buffered glucose carrier (Rehydrex® with glucose 25 mg/ml, Fresenius Kabi AB, Uppsala, Sweden), fentanyl 0.04 mg/kg/h (Fentanyl 50 µg/ml, B Braun Medical AB, Danderyd, Sweden), ketamine 30 mg/kg/h (Ketaminol Vet® 100 mg/ml, Intervet AB, Stockholm, Sweden), midazolam 0.1 mg/kg/h (Midazolam Hameln 1 mg/ml, AlgolPharma AB, Kista, Sweden) and pancuronium bromide 0.3 mg/kg/h (Pavulon® 2 mg/ml, Schering-Plough AB, Stockholm, Sweden). Body temperature was stabilized with a heating pad, aiming for a core temperature of 37°C. A silicone catheter was used to catheterize the urinary bladder. The jugular vein was catheterized (BD Careflow 17 G) for blood sampling, monitoring of central venous pressure and administration of drugs. A pulmonary artery (PA) catheter (BD Criticath™ Pulmonary Artery/Thermodilution Catheter) was inserted through the right jugular vein and placed into the pulmonary artery. Continuous monitoring of the arterial blood pressure with a pressure transducer (BD Careflow 20G; Becton-Dickinson AB, Stockholm, Sweden) was undertaken with a catheter in the left femoral artery.
were connected to a monitoring system (Dräger Infinity Delta). A 5-French 12-electrode pressure–volume conductance catheter (Ventri-Cath, Millar Instruments, MI, USA) was introduced in the RV apex and advanced to a level just below the pulmonary valve. The catheter was fixed with a purse-string suture and connected to the Millar Pressure Volume Systems (MPVS) Ultra analysing system. Proper catheter positioning was determined by monitoring individual segmental pressure–volume loops. Extra ventricular segments were excluded from analysis. Animals were heparinized with a bolus of 7500 IU once. The modified Glenn shunt was created by connecting the superior vena cava (SVC) to the main PA with a 12 mm Dacron vascular graft (Gelsoft™, Vascutek Inchinnan, Scotland, UK). A sidebiting clamp was placed first on the SVC and then on the main PA and thereafter the vessels were opened longitudinally and anastomosed with a 4-0 running suture. The clamp was released and the patency of the shunt was evaluated with a 12 mm ultrasonic flow probe (VeriQ Perivascular 12 mm, MediStim ASA, Oslo, Norway). The graft was then closed with a vascular clamp. When the shunt was put in use, the vascular clamp was removed from the shunt, and the SVC was then clamped below the shunt, directing all blood from the SVC to the main PA through the modified Glenn shunt (Fig. 1). The left ventricular apex was cannulated with a 28-Fr venous cannula (Malleable Single Stage Venous Cannula, Edwards Lifesciences) and a 16-Fr arterial cannula (FemFlex, Edwards Lifesciences) was inserted into the ascending aorta. The cannulas were secured with purse-string sutures. The left ventricular apex cannula was connected to the inflow tubing of a preprimed Bio-Pump (Medtronic). The aortic cannula was deaired and connected to the outflow tubing of the Bio-Pump. A Bio-Pump flow probe connector (Medtronic) was connected via the tubing to the aortic cannula.
Measurements Standard lead II electrocardiogram (ECG), heart rate (HR), mean systemic arterial blood pressure (MAP), RA pressure, RV pressure, LA pressure and PA pressure were monitored continuously and recorded. Cardiac output (CO) was measured to 3.3 l/min (range, 2.8–3.4) at baseline using the thermodilution technique. The pressure–volume
Surgical preparation A median sternotomy was performed. The pericardium was opened and pericardial stay sutures were placed. The azygos and hemiazygos veins were ligated. Pressure transducer catheters were surgically placed in the right atrium (RA) and the RV. The catheters
Figure 1: Animal preparation; a modified Glenn shunt was created between the SVC and the main pulmonary artery. A left ventricular assist device was implanted. Pressure volume loops were obtained throughout the experimental model in pigs (n = 8).
P. Schiller et al. / European Journal of Cardio-Thoracic Surgery
Experimental protocol Shunt. The clamp on the vascular graft was released and the SVC was clamped at the junction with the RA to direct all blood flow from SVC to the main PA. Shunt flow was measured with a 12 mm ultrasonic flow probe. The Glenn shunt was opened for 20 min and thereafter measurements were performed. Left ventricular assist device. The Glenn shunt was closed and the clamp on the SVC was released. The LVAD was started and set to a flow corresponding to the baseline CO of each animal measured with the thermodilution technique. This was continued for 20 min and thereafter measurements were performed. Left ventricular assist device combined with a modified Glenn shunt. The clamp on the Glenn shunt was released and the SVC was clamped at the junction with the RA to direct all blood flow from SVC to the main PA. Shunt flow was measured with a 12 mm ultrasonic flow probe. Measurements were recorded after 20 min of stabilization.
Data analysis and statistical methods For pressures (i.e. SVCP, RAP, mPAP and MAP), HR, LVAD flow and Glenn-shunt flow, the value for each pig was calculated as the mean of all measurements for each treatment period, and data are expressed as medians with interquartile ranges in brackets. For blood gas measurements (i.e. pH, PO2, PCO2, SBE, SaO2, SvO2 and Hb) the data are expressed as medians with interquartile ranges in brackets, based on the value obtained for each pig in each treatment period. Data analysis of the pressure–volume loops was performed using Lab Chart™ software (AD Instruments Pty Ltd, Bella Vista, NSW, Australia). As there was no calibration for parallel conductance, the baseline RV ejection fraction was estimated to be 0.6 and the volumes were calibrated according to this figure. With regard to conductance catheter-derived values, the mean value of the loops for each pig in each treatment period was used. The data are expressed as medians with interquartile ranges in brackets. Statistical analysis was conducted using the statistical software R (v. 3.1.1). Non-parametric tests were used due to the small sample size. Comparison was done between (i) baseline versus LVAD and (ii) LVAD versus LVAD and modified Glenn shunt, using the Wilcoxon signed-rank test. P-values ≤0.05 were considered statistically significant.
RESULTS All eight pigs survived the experiment and were included in the study.
Baseline All animals were in a stable haemodynamic state at baseline. The MAP was 86 mmHg (74–99), SvO2 was 58% (46–62) and CO was 2.8 l/min (2.3–3.2), as depicted in Table 1. RA pressure was 9 mmHg (9–9) (Fig. 2). Systolic RV pressure was 31 mmHg (28–33) and diastolic RV pressure was 7 mmHg (6–7) at baseline. RV enddiastolic volume was 50 ml (49–67) and RV stroke volume was 30 ml (29–40), as shown in Figs 3 and 4. Moreover, RV stroke work at baseline was 535 mmHg ml (442–717) (Fig. 5). The pressure development (Pdev) in the RV was 26 mmHg (22–28), as shown in Fig. 6. The representative pressure–volume loops at baseline are shown in Fig. 7.
A modified Glenn shunt without left ventricular assist device The median flow in the shunt was 0.57 l/min (0.48–0.74), Table 1. There were no significant changes in the MAP, SvO2, RA or RV pressure during the phase of Glenn shunting when compared with baseline, Table 1. Moreover, the RV pressure–volume loops did not change significantly during shunting, there was, however, a trend towards decreased RV stroke work, as shown in Figs 5 and 7.
Left ventricular assist device therapy without shunt During LVAD therapy alone (with a closed shunt), the pump flow was 3.3 l/min (2.9–4.0). The MAP increased from 86 mmHg (74–99) to 98 mmHg (88–103) (P = 0.04), and SvO2 increased from 58% (46–52) to 64% (58–70) (P = 0.02) when LVAD therapy alone was compared with baseline values; Table 1. LVAD therapy increased the RA pressure from 9 mmHg (9–9) to 15 mmHg (12–15) (P = 0.01) (Fig. 2), and RV end-diastolic volume from 50 ml (49–67) to 85 ml (55–88) (P = 0.05) (Fig. 3). In addition, there was an increase in the RV stroke volume from 30 ml (29–40) to 51 ml (42–53) (P < 0.01) (Fig. 4). RV stroke work increased from the baseline value at 535 mmHg ml (424–717) to 708 mmHg ml (654–1193) (P = 0.04), as shown in Fig. 5. Finally, there was no increase in the development of pressure in the RV during LVAD therapy from 26 mmHg (22–28) to 29 mmHg (26–32) (P = 0.14) (Fig. 6). The pressure–volume loops for LVAD alone significantly shifted to the right when compared with baseline, as shown in Fig. 7.
Left ventricular assist device therapy combined with modified Glenn shunt The LVAD flow was not altered as the shunt was opened. There was a reduction in the MAP with the LVAD and Glenn shunt from 98 mmHg (88–103) to 84 mmHg (74–89) when compared with LVAD therapy alone (P = 0.02); Table 1. Also, there was a decrease in SvO2 from 64% (58–70) with the LVAD alone to 46% (35–55) with the LVAD and shunt (P < 0.01); Table 1.
TX & MCS
catheter was calibrated for blood resistivity using a rho-cuvette (Millar Instruments, MI, USA) and the alpha-value was calculated using the thermodilution-derived CO for each animal. Pressure– volume loops from the RV were recorded for 25 consecutive heartbeats with the ventilation suspended at end-expiration. Haemodynamic measurements and pressure–volume curves were recorded at the following time periods: (i) at baseline, (ii) at shunting, (iii) with the LVAD and a closed Glenn shunt and (iv) with the LVAD and an open Glenn shunt. The CO values presented in the results are derived from the pressure–volume curves. Blood samples were collected at the same time points for blood gas analysis. Arterial blood gases, haemoglobin concentration (Hb) and mixed venous oxygen saturation (SvO2) were measured with an ABL 500 Radiometer, Medical ApS, Bronshoj, Denmark.
3
P. Schiller et al. / European Journal of Cardio-Thoracic Surgery
4
Table 1: The phases of the experimental study on LVAD in combination with a modified Glenn shunt, n = 8 Parameter
Baseline
Modified Glenn-shunt
LVAD
LVAD + modified Glenn-shunt
Baseline vs LVAD, P-values
LVAD vs LVAD + modified Glenn-shunt, P-values
MAP (mmHg) RVCO (l/min) CO (l/min) SVCP (mmHg) RVEDP (mmHg) RVESP (mmHg) mPAP (mmHg) RVESV (ml) RVEDV (ml) RVSV (ml) RVSW (mmHg ml) HR (beats/min) LVAD (l/min) Modified Glenn-shunt (l/min) pH PO2 (kPa) PCO2 (kPa) SBE (mmol/l) SaO2 (%) SvO2 (%) Hb (g/l)
86 (74–99) 2.8 (2.2–3.2) 2.8 (2.2–3.2) 9 (9–9) 7 (6–7) 31 (28–33) 18 (17–19) 20 (20–27) 50 (49–67) 30 (29–40) 535 (424–717) 84 (76–91) N/A N/A 7.47 (7.47–7.48) 43 (42–44) 5.5 (5.2–5.6) 5.8 (5.0–6.0) 99 (98–99) 58 (46–62) 77 (70–82)
96 (86–101) 2.9 (2.1–3.9) 3.4 (2.7–4.4) 20 (20–24) 6 (6–7) 28 (27–29) N/A 16 (12–23) 51 (39–61) 32 (21–40) 461 (271–718) 93 (90–98) N/A 0.57 (0.48–0.74) 7.46 (7.45–7.48) 41 (40–44) 5.5 (5.2–5.5) 4.8 (3.8–6.0) 100 (99–100) 52 (43–56) 74 (65–77)
98 (88–103) 4.8 (3.7–5.8) 4.8 (3.7–5.8) 14 (13–14) 6 (6–7) 35 (32–39) N/A 24 (11–34) 85 (55–88) 49 (42–53) 708 (654–1193) 90 (89–101) 3.3 (2.9–4.0) N/A 7.46 (7.45–7.49) 42 (39–43) 5.5 (5.2–5.8) 5.2 (4.0–6.4) 99 (99–99) 64 (58–70) 51 (48–57)
84 (74–89) 3.4 (2.8–4.2) 4.0 (3.3–4.7) 18 (18–23) 6 (6–7) 28 (26–29) N/A 22 (14–29) 60 (48–69) 34 (33–38) 465 (366–711) 93 (90–108) 3.3 (2.9–4.0) 0.58 (0.43–0.70) 7.46 (7.44–7.48) 43 (40–43) 5.5 (5.2–5.9) 6.0 (4.0–6.6) 100 (99–100) 46 (35–55) 57 (55–60)
0.04
