- Email: [email protected]
End-organ function during chronic nonpulsatile circulation
End-Organ Function During Chronic Nonpulsatile Circulation Satoshi Saito, MD, PhD, Stephen Westaby, PhD, FETCS, David Piggot, FRCA, Sergai Dudnikov, FRCA, Desiree Robson, RN, Pedro A. Catarino, FRCS, Colin Clelland, FRCPath, and Chisato Nojiri, MD, PhD Oxford Heart Centre and Department of Cellular Pathology, Oxford, United Kingdom, and Terumo Cardiovascular System Corp, Ann Arbor, Michigan
Background. Evolving blood pump technology has produced user-friendly continuous flow left ventricular assist devices, but uncertainty exists about the safety of chronic nonpulsatile circulation. We established consistently nonpulsatile blood flow in a sheep model using the Terumo magnetically suspended centrifugal pump. We then compared end-organ function between pulseless and control animals. Methods. Fifteen healthy sheep (65 to 85 kg) were allocated to either left ventricular assist device (n ⴝ 9) or control (n ⴝ 6) groups. We implanted the device through a left thoracotomy and determined the flow rate at which pulse pressure was absent. The flow rate was then adjusted to exceed that rate (4.2 ⴞ 1.5 L/min), and all variables of pump function were continuously monitored by computer. Blood tests were taken serially for hepatic and renal function and plasma renin levels. The sheep were sacrificed electively at 30 (n ⴝ 3), 90 (n ⴝ 4), 180 (n ⴝ 1), and 340 (n ⴝ 1) days. Detailed histologic examination was made of the brain, liver, kidney, myocardium, and major arteries. Results. All animals remained in good condition until
sacrifice. All measures of end-organ function remained within normal limits for both groups. There were no histologic differences between the organs of pulsatile and nonpulsatile animals. Although there was no significant difference in mean blood pressure, plasma renin levels were substantially elevated in pulseless animals (1.4 ⴞ 0.3 pg/mL versus 2.9 ⴞ 0.3 pg/mL; p < 0.05). We also identified thinning of the medial layer of the ascending aorta in nonpulsatile sheep (1.8 ⴞ 0.4 mm in left ventricular assist device animals versus 2.6 ⴞ 0.6 mm in control sheep; p < 0.05). Conclusions. Chronic nonpulsatile circulation was well tolerated, and we found neither functional nor histologic changes in major end organs. The renin-angiotensin system was upregulated, but this did not provide a significant rise in blood pressure. The changes in the aortic wall merit further investigation. As a result of these findings, we consider that nonpulsatile devices can be used safely for long-term circulatory support.
T
user friendly for the patient. However, doubts remain about the safety of continuous flow devices in the long term. The mammalian heart produces pulse by contraction, stroke volume ejection, then relaxation with one-way valves. Given this mechanism, pulse pressure is obligatory, as is a resting phase for the myocyte. However, systolic thrust generated by the heart represents only one third of a cardiac cycle. Pulse is diminished at the capillary and cellular level, and there is little understanding as to whether or not pulse pressure is required for normal end-organ function. The closest scenario is our use of the Jarvik 2000 Heart (up to 22 months), although the native heart provides diminished pulse pressure in these patients [2]. The question as to whether pulse pressure is important in the mammalian circulation can only be answered by experimental findings in an animal model.
he success of mechanical blood pumps as a bridge to transplantation has led to their use as permanent circulatory support systems. The REMATCH (randomized evaluation of mechanical assistance for the treatment of congestive heart failure) trial demonstrated a survival advantage for left ventricular assist devices (LVADs) over continued medical treatment but highlighted the substantial risk of mechanical failure and device-related complications inherent in first-generation pulsatile devices [1]. Displacement blood pumps including the Novacor威 and Thermo Cardio Systems (TCI) LVADs were designed to mimic the native heart in the belief that pulsatility was a physiologic requirement in the human circulation. In contrast, continuous-flow blood pumps can be made smaller, simpler, and more
Presented at the Poster Session of the Thirty-eighth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28 –30, 2002. Address reprint requests to Dr Saito, Oxford Heart Centre, John Radcliffe Hospital, Headley Way, Headington, Oxford OX3 9DU, United Kingdom; e-mail: [email protected].
© 2002 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
(Ann Thorac Surg 2002;74:1080 –5) © 2002 by The Society of Thoracic Surgeons
Dr Nojiri discloses that she has a financial relationship with Terumo Cardiovascular System Corp.
0003-4975/02/$22.00 PII S0003-4975(02)03846-8
Ann Thorac Surg 2002;74:1080 –5
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
1081
We established reliable nonpulsatile circulation in the sheep using a magnetically suspended centrifugal blood pump [3, 4]. The objective of this study was to compare organ function between chronically pulseless and normal control animals.
Material and Methods Terumo DuraHeart Left Ventricular Assist System The Terumo DuraHeart left ventricular assist system (TLVAS [Terumo Cardiovascular System Corp, Ann Arbor]) is a titanium centrifugal pump with a magnetically suspended impeller producing continuous (nonpulsatile) flow up to 10 L/min [3]. The interior surface is heparin coated, and there is no purge system. The bloodcontacting surface and the inflow and outflow cannulas are modified with a heparin immobilization technique. All device variables are continuously transmitted to an online computer system.
Animal Experiments Fifteen healthy Welsh male sheep (65 to 85 kg) between 3 and 4 years of age were allocated to either centrifugal LVAD (n ⫽ 9) or control (n ⫽ 6) groups. All animal data were prospectively entered into a database to record device-related events, non– device-related events, pump function (driving speed and energy consumption), and autopsy findings. Sheep in the LVAD group underwent implantation of the TLVAS, whereas those in the control group did not. Control animals were kept in adjacent pens as companion sheep. The surgical procedures and postoperative care were undertaken humanely by licensed personnel in compliance with United Kingdom Home Office guidelines.
Operation The device was implanted through a left thoracotomy without the use of cardiopulmonary bypass (CPB). Heparin (5,000 U) was given before the pump was connected. The inflow cannula was placed in the left ventricular apex, and a polyethylene terephthalate fiber (Dacron [Terumo Cardiovascular System Corp, Ann Arbor]) outflow graft was anastomosed to the descending thoracic aorta. Air was carefully removed from the system before switching the device on at 1,000 rpm. We then implanted an ultrasonic flow probe (Transonic System, Ithaca, NY) around the outflow graft to determine flow through the device at different rotational speeds. After restoring circulating blood volume and central venous pressure to preoperative levels, we determined the pump speed (1,600 to 1,800 rpm) at which pulse pressure disappeared through capture of all transmitral flow. The process was undertaken at least twice as preload was increased from a central venous pressure of 8 to 12 cm H2O. This established that pulse pressure was abolished even at high preload. At greater speeds (⬎2,000 rpm) there was no pulse pressure in the systemic circulation as confirmed by postoperative invasive monitoring
Fig 1. Representive data of pump flow during the 90 days of left ventricular assist device support. Pump flow measured by the ultrasonic flow probe around the outflow graft was between 4.8 and 7.0 L/min.
for 48 hours, then again before sacrifice. An intercostal drain was inserted, and the wound was then closed in layers. Anesthesia was discontinued, and the sheep were transferred from the operating room to the pen. No anticoagulation was used after the surgical procedure. For reasons of animal welfare the control animals were anesthetized for invasive monitoring before sacrifice but did not undergo a sham operation. They were identical to the LVAD animals for age, body weight, and species. We also did not attempt to insert an arterial catheter into awake animals during the course of the support period; we believed that this risked substantial morbidity such as bleeding and device endocarditis.
Pump Flow Data from the device including motor suspension current, energy requirements, and calculated pump flow were monitored constantly by the online computer. Pump flow was also measured daily by the implanted ultrasound flow probe throughout the study period (Fig 1). Mechanical reliability was determined by recording measures of pump function, by auscultation of the chest, and by inspection of power lines and drive line exit site. The correlation between measured graft flow and calculated pump flow were confirmed.
Recovery The sheep were extubated after 30 to 60 minutes of spontaneous respiration and documentation of satisfactory blood gases and acid-base balance. A vest was placed around the thorax to carry the controller and connect with the electrical power line. Before removal of invasive arterial pressure monitoring at 48 hours, we were able to confirm that the systemic circulation remained nonpulsatile even during vigorous activity during handling. The animals were then free to mobilize, drink, feed, and roam around the sheep pen.
1082
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
Ann Thorac Surg 2002;74:1080 –5
Detailed postmortem examination was performed with macroscopic and histologic examination of the brain, heart, kidneys, liver, and aorta.
Histologic Examination of Major Organs Formalin-fixed, paraffin-embedded tissues from LVAD and control sheep were cut into 3-m-thick sections and stained with hematoxylin and eosin. Differences between LVAD and control specimens were assessed by two independent pathologists who were blind to the origin of the samples. The ascending aorta was cut horizontally into three segments that were embedded in paraffin sections. Sections of 3 m thickness were cut from each block and stained with hematoxylin and eosin. The thickness of the medial layer was determined by computed morphometry (NIH images, Bethesda, MD). Fig 2. Representative computerized arterial pressure traces both in (A) left ventricular assist device (LVAD) and (B) control animals before sacrifice. There was no pulse pressure in left ventricular assist device group animals.
Assessment The sheep were examined for signs of neurologic events, heart failure, systemic thromboembolism, or abnormal behavior. Renal and hepatic function were assessed serially by measurement of blood urea, creatinine, bilirubin, and liver enzymes. Plasma renin activity was measured by standard radioimmunoassay for both groups.
Autopsy Studies Sheep in the LVAD group were electively sacrificed at 30 (n ⫽ 3), 90 (n ⫽ 4), 180 (n ⫽ 1), and 340 (n ⫽ 1) days. Before sacrifice all animals were anesthetized and invasively monitored with an arterial line and Swan-Ganz catheter. The relationship between pump speed and absence of pulse pressure was checked again. Computerized arterial pressure traces before sacrifice are shown in Figure 2. Heparin (5,000 IU/kg) was given intravenously to avoid clot formation on the blood-contacting surfaces of the pump and cannulas. The six control sheep were anesthetized and invasively monitored. They were then sacrificed to compare end-organ morphology.
Statistical Analysis All results for continuous variables were expressed as means ⫾ standard deviations. The paired or unpaired Student’s t test, or Mann-Whitney U test if appropriate, was used to compare continuous variables between two subgroups. Probability values of less than 0.05 were considered to indicate statistical significance.
Results Hemodynamic studies before chest closure showed systemic blood flow to be consistently nonpulsatile at pump rates exceeding 1,600 to 1,800 rpm. All animals in the LVAD group recovered rapidly after operation and survived in excellent condition before elective termination. There was no difference in behavior between LVAD and control animals after recovery from the surgical procedure. The device drive line did not restrict mobility around the pen. There were no neurologic events.
Renal and Hepatic Function Blood tests showed the blood urea and creatinine levels remained stable and within normal limits (Table 1). There were no significant differences between the LVAD and control groups. Indices of hepatic function are shown in Table 2. Hepatic function remained normal for the duration of the study with no difference between the LVAD and the control groups.
Table 1. Comparison of Renal Function Between Left Ventricular Assist Device and Control Groupsa LVAD Group
Variable
Control (n ⫽ 6)
Before Implant (n ⫽ 9)
30 d (n ⫽ 9)
90 d (n ⫽ 6)
180 d (n ⫽ 2)
340 d (n ⫽ 1)
Urea (U/L) Cr (mol/L)
4.2 (1.0) 80 (23)
3.3 (0.5) 76 (21)
3.0 (2.2) 92 (16)
1.6 (1.2) 85 (29)
7.6 (1.1) 85 (7.2)
5.8 84
a
Mean ⫾ SEM.
Cr ⫽ creatinine;
LVAD ⫽ left ventricular assist device.
Ann Thorac Surg 2002;74:1080 –5
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
1083
Table 2. Comparison of Indices of Hepatic Function Between Left Ventricular Assist Device and Control Groupsa LVAD group
Variable
Control Group (n ⫽ 6)
Before Implant (n ⫽ 9)
30 d (n ⫽ 9)
90 d (n ⫽ 6)
180 d (n ⫽ 2)
340 d (n ⫽ 1)
T-Bil (mol/L) GPT (IU/L) GOT (IU/L) ␥-GPT (IU/L)
2.1 (0.7) 38 (23) 92 (25) 56 (11)
2.0 (0.8) 49 (35) 84 (24) 47 (6.0)
2.3 (1.7) 16 (8.1) 104 (24) 63 (7.5)
2.0 (0.2) 15 (2.4) 74 (12) 49 (15)
1.8 (0.2) 14 (1.7) 68 (12) 50 (16)
2.0 24 80 49
a
Mean ⫾ SEM.
GPT ⫽ serum glutamate pyruvate transaminase; GOT ⫽ serum glutamate oxalate transaminase; LVAD ⫽ left ventricular assist device; T-Bil ⫽ total bilirubin.
Plasma Renin Activity and Mean Blood Pressure Mean blood pressure was 94 ⫾ 12 mm Hg in the LVAD animals before implantation and 98 ⫾ 20 mm Hg at the time of sacrifice. Mean blood pressure in the control sheep was 88 ⫾ 36 mm Hg. There was no statistical difference between the groups or before and after LVAD implantation. Plasma renin activity after LVAD implantation was significantly elevated compared with the preoperative values (2.9 ⫾ 0.3 pg/mL versus 1.3 ⫾ 0.4 pg/mL; p ⬍ 0.05). There was also a significant difference between the LVAD animals and the control group (2.9 ⫾ 0.3 pg/mL versus 1.4 ⫾ 0.3 pg/mL; p ⬍ 0.05; Fig 3).
Histology of Major Organs Macroscopic examination of the native heart in LVAD sheep showed minor thrombus formation in the area of stasis between the inflow cannula and ventricular wall. One embolic infarct was found in a single LVAD sheep kidney. Apart from this, detailed histologic comparisons of brain, heart, liver, and kidneys showed no other emboli and no parenchymal differences between the LVAD and the control sheep. The organs remained completely normal in the LVAD sheep for up to 340 days (Fig 4). However, we noted the thickness of the medial
␥-GPT ⫽ gamma guanisine triphosphate;
layer of the ascending aorta in the LVAD group to be significantly diminished (1.8 ⫾ 0.4 mm in the LVAD group versus 2.6 ⫾ 0.6 mm in the control group; p ⬍ 0.05; Figs 5 and 6). Both pathologists agreed on this. Because of the small number of sheep we could not discern a significant difference in aortic wall thickness with time.
Comment The continuous flow pumps are silent and much more compact than the displacement LVADs. They have a smaller and less thrombogenic blood–foreign surface interface, no areas of stasis, and no artificial valves. Device durability may be improved by this approach. Until recently it was inconceivable that a rotary blood pump could function for many months in an animal model without anticoagulation. With careful alignment of the TLVAS inflow cannula we were able to provide consistently nonpulsatile blood flow for virtually 12 months. Measurements of calculated flow correlated well with ultrasonic flowmeter measurements of graft flow, and these levels were constantly above that required to abolish pulse pressure during invasive arterial monitoring and elevated preload. Even when the aortic valve remains continuously closed the left ventricle may genFig 3. Mean blood pressure (BP) and plasma renin activity. (N.S. ⫽ not significant; Pre Op LVAD ⫽ preoperative left ventricular assist device; Post Op LVAD ⫽ postoperative left ventricular assist device.)
1084
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
Ann Thorac Surg 2002;74:1080 –5
Fig 4. Representative photomicrographs for major organs from the left ventricular assist device sheep. (A) Hematoxylin and eosin staining of brain from a sheep on the left ventricular assist device for 90 days. There are no abnormalities. (B) Hematoxylin and eosin staining of liver from a sheep on the left ventricular assist device for 180 days. Sinusoidal structure remained normal. (C) Hematoxylin and eosin staining of myocardium from a sheep on the left ventricular assist device for 340 days. Myocardium and coronary arteries are normal. The parasites found occur frequently in sheep. (D) Hematoxylin and eosin staining of kidney from a sheep on the left ventricular assist device for 340 days. Glomerular and tubular structure are normal.
erate preload to the device, which can be transmitted as pulse pressure to the circulation [2]. Consequently we set the TLVAS at a level at which even this level of pulse pressure could not occur. The most encouraging finding was that long-term circulatory support without pulse pressure can maintain normal end-organ structure and function. Activity, behavior, and neurologic function did not differ from control animals. This outcome gives us confidence that both centrifugal and axial flow blood pumps should provide safe and effective long-term circulatory support in the patient with heart failure. Early clinical experience with axial flow pumps has already shown that diminished pulse pressure and periods without pulse are compatible with recovery from heart failure and improvement in end-organ function [2]. In the absence of a reliable nonpulsatile animal model, most previous investigations of pulseless circulation have centered on CPB [5, 6]. Pulsatile CPB systems were said
Fig 5. Corresponding ascending aortic samples from the left ventricular assist device (LVAD) and the control sheep. There is an easily discernible difference in thickness of the medial layer.
to result in less endothelial damage, reduced systemic vascular resistance, and normal nitric oxide release [6]. Pulsatility plays a role in the movement of lymph in and out of the tissues, perhaps preventing edema and capillary sludging [7]. In the CPB model, cerebral microcirculation was improved by pulsatility, and pulse pressure influenced the activity of the renin-angiotensin system and catecholamine release [8]. In reality these short-term CPB studies fail to predict the effects of long-term pulseless circulation because of the body’s ability to adapt to changes in physiology. In any event pulse pressure is greatly diminished or absent at the capillary level. In nonpulsatile animals upregulation of plasma renin activity could be an adaptive response to lack of pulse pressure. This elevates mean perfusion pressure by increasing the systemic vascular resistance. Systemic vascular resistance is raised during nonpulsatile CPB when compared with pulsatile CPB [9]. Nishimura and colleagues [10] reported that the systemic vascular response
Ann Thorac Surg 2002;74:1080 –5
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
1085
although we were unable to define differences in the sheep. In conclusion, chronic pulseless circulation was well tolerated in the sheep model, and we could not identify changes in major end organs. The renin-angiotensin system was upregulated, although mean blood pressure was not significantly elevated. Although pulseless circulation caused thinning of the aortic wall, our findings still suggest that continuous flow devices can be used safely for long-term circulatory support. This work was supported financially in part by a research grant from Terumo Cardiovascular System Corp. The authors thank the pathologists Drs Derek Roskell and Andrew Graham for their pathologic studies and Drs Maekawa and Yanai of Terumo for their technical support.
References Fig 6. Measured thickness of the medial layer of the aorta from left ventricular assist device (LVAD) and control groups.
to norepinephrine administration decreased markedly in a chronic nonpulsatile left heart bypass model (up to 137 days), although there was no difference in the plasma norepinephrine levels between pulsatile and nonpulsatile animals. This group also reported atrophic changes in the aorta of nonpulsatile goats in comparison with goats with pulsatile left heart bypass [11]. We showed that the medial layer of the ascending aorta in the LVAD group was consistently thinner than in control sheep. This striking and reproducible finding suggests that we succeeded in providing pulseless circulation throughout the duration of the experiment. The arterial changes can be explained by a reduction in cyclic mechanical load, which leads to atrophy of the medial smooth muscle cells, and we are examining this further. The effect of a change in pulse pressure on the arterial wall has been investigated using an in vitro model of cyclic mechanical stretching on cultured vascular endothelial cells and smooth muscle cells [12]. These experiments also showed that a reduction in a pulse pressure resulted in atrophic changes. Assuming that these morphologic changes extend to the entire arterial system, vascular endothelial function such as blood flow regulation and pressure control could also be compromised,
1. Rose EA, Gelijns AC, Moskowitz AJ, et al. Long-term use of a left ventricular assist device for end stage heart failure. N Engl J Med 2001;345:1435– 43. 2. Westaby S, Banning AP, Saito S, et al. Circulatory support for long term treatment of heart failure. Experience with an intraventricular continuous flow pump. Circulation 2002;105: 2588–91. 3. Saito S, Westaby S, Pigott D, et al. Reliable long-term non-pulsatile circulatory support without anticoagulation. Eur J Cardiothorac Surg 2001;19:678– 83. 4. Nojiri C, Kijima T, Maekawa J, et al. Terumo implantable left ventricular assist system: results of long term animal study. ASAIO J 2000;46:117–22. 5. Wilkens H, Regelson W, Hoffmeister FS. The physiologic importance of pulsatile flow. N Engl J Med 267;443– 6. 6. Sumpio BE, Widmann MD. Enhanced production of an endothelium-derived contracting factor by endothelial cells subjected to pulsate stretch. Surgery 1990;108:227– 82. 7. Parsons RJ, McMaster PD. Effects of the pulse upon formation and flow of lymph. J Exper Med 1938;68:353–76. 8. Jett GK. Physiology of non-pulsatile circulation: acute versus chronic support. ASAIO J 1999:119–22. 9. Golding LR, Murakami G, Harasaki H, et al. Chronic nonpulsatile blood flow. ASAIO Trans 1982;28:81–5. 10. Nishimura T, Tatsumi E, Nishinak T, et al. Diminished vasoconstrictive function caused by long-term nonpulsatile left heart bypass. Artif Organ 1999;23:722– 6. 11. Nishimura T, Tatsumi E, Takaichi S, et al. Prolonged nonpulsatile left heart bypass with reduced systemic pulse pressure cause morphological changes in the aortic wall. Artif Organ 1998;22:405–10. 12. Birkov KG, Shrinsky VP, Stepanova OV, et al. Stretch affects phenotype and proliferation of vascular smooth muscle cells. Mol Cell Biochem 1995;144:131–9.
Background. Evolving blood pump technology has produced user-friendly continuous flow left ventricular assist devices, but uncertainty exists about the safety of chronic nonpulsatile circulation. We established consistently nonpulsatile blood flow in a sheep model using the Terumo magnetically suspended centrifugal pump. We then compared end-organ function between pulseless and control animals. Methods. Fifteen healthy sheep (65 to 85 kg) were allocated to either left ventricular assist device (n ⴝ 9) or control (n ⴝ 6) groups. We implanted the device through a left thoracotomy and determined the flow rate at which pulse pressure was absent. The flow rate was then adjusted to exceed that rate (4.2 ⴞ 1.5 L/min), and all variables of pump function were continuously monitored by computer. Blood tests were taken serially for hepatic and renal function and plasma renin levels. The sheep were sacrificed electively at 30 (n ⴝ 3), 90 (n ⴝ 4), 180 (n ⴝ 1), and 340 (n ⴝ 1) days. Detailed histologic examination was made of the brain, liver, kidney, myocardium, and major arteries. Results. All animals remained in good condition until
sacrifice. All measures of end-organ function remained within normal limits for both groups. There were no histologic differences between the organs of pulsatile and nonpulsatile animals. Although there was no significant difference in mean blood pressure, plasma renin levels were substantially elevated in pulseless animals (1.4 ⴞ 0.3 pg/mL versus 2.9 ⴞ 0.3 pg/mL; p < 0.05). We also identified thinning of the medial layer of the ascending aorta in nonpulsatile sheep (1.8 ⴞ 0.4 mm in left ventricular assist device animals versus 2.6 ⴞ 0.6 mm in control sheep; p < 0.05). Conclusions. Chronic nonpulsatile circulation was well tolerated, and we found neither functional nor histologic changes in major end organs. The renin-angiotensin system was upregulated, but this did not provide a significant rise in blood pressure. The changes in the aortic wall merit further investigation. As a result of these findings, we consider that nonpulsatile devices can be used safely for long-term circulatory support.
T
user friendly for the patient. However, doubts remain about the safety of continuous flow devices in the long term. The mammalian heart produces pulse by contraction, stroke volume ejection, then relaxation with one-way valves. Given this mechanism, pulse pressure is obligatory, as is a resting phase for the myocyte. However, systolic thrust generated by the heart represents only one third of a cardiac cycle. Pulse is diminished at the capillary and cellular level, and there is little understanding as to whether or not pulse pressure is required for normal end-organ function. The closest scenario is our use of the Jarvik 2000 Heart (up to 22 months), although the native heart provides diminished pulse pressure in these patients [2]. The question as to whether pulse pressure is important in the mammalian circulation can only be answered by experimental findings in an animal model.
he success of mechanical blood pumps as a bridge to transplantation has led to their use as permanent circulatory support systems. The REMATCH (randomized evaluation of mechanical assistance for the treatment of congestive heart failure) trial demonstrated a survival advantage for left ventricular assist devices (LVADs) over continued medical treatment but highlighted the substantial risk of mechanical failure and device-related complications inherent in first-generation pulsatile devices [1]. Displacement blood pumps including the Novacor威 and Thermo Cardio Systems (TCI) LVADs were designed to mimic the native heart in the belief that pulsatility was a physiologic requirement in the human circulation. In contrast, continuous-flow blood pumps can be made smaller, simpler, and more
Presented at the Poster Session of the Thirty-eighth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28 –30, 2002. Address reprint requests to Dr Saito, Oxford Heart Centre, John Radcliffe Hospital, Headley Way, Headington, Oxford OX3 9DU, United Kingdom; e-mail: [email protected].
© 2002 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
(Ann Thorac Surg 2002;74:1080 –5) © 2002 by The Society of Thoracic Surgeons
Dr Nojiri discloses that she has a financial relationship with Terumo Cardiovascular System Corp.
0003-4975/02/$22.00 PII S0003-4975(02)03846-8
Ann Thorac Surg 2002;74:1080 –5
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
1081
We established reliable nonpulsatile circulation in the sheep using a magnetically suspended centrifugal blood pump [3, 4]. The objective of this study was to compare organ function between chronically pulseless and normal control animals.
Material and Methods Terumo DuraHeart Left Ventricular Assist System The Terumo DuraHeart left ventricular assist system (TLVAS [Terumo Cardiovascular System Corp, Ann Arbor]) is a titanium centrifugal pump with a magnetically suspended impeller producing continuous (nonpulsatile) flow up to 10 L/min [3]. The interior surface is heparin coated, and there is no purge system. The bloodcontacting surface and the inflow and outflow cannulas are modified with a heparin immobilization technique. All device variables are continuously transmitted to an online computer system.
Animal Experiments Fifteen healthy Welsh male sheep (65 to 85 kg) between 3 and 4 years of age were allocated to either centrifugal LVAD (n ⫽ 9) or control (n ⫽ 6) groups. All animal data were prospectively entered into a database to record device-related events, non– device-related events, pump function (driving speed and energy consumption), and autopsy findings. Sheep in the LVAD group underwent implantation of the TLVAS, whereas those in the control group did not. Control animals were kept in adjacent pens as companion sheep. The surgical procedures and postoperative care were undertaken humanely by licensed personnel in compliance with United Kingdom Home Office guidelines.
Operation The device was implanted through a left thoracotomy without the use of cardiopulmonary bypass (CPB). Heparin (5,000 U) was given before the pump was connected. The inflow cannula was placed in the left ventricular apex, and a polyethylene terephthalate fiber (Dacron [Terumo Cardiovascular System Corp, Ann Arbor]) outflow graft was anastomosed to the descending thoracic aorta. Air was carefully removed from the system before switching the device on at 1,000 rpm. We then implanted an ultrasonic flow probe (Transonic System, Ithaca, NY) around the outflow graft to determine flow through the device at different rotational speeds. After restoring circulating blood volume and central venous pressure to preoperative levels, we determined the pump speed (1,600 to 1,800 rpm) at which pulse pressure disappeared through capture of all transmitral flow. The process was undertaken at least twice as preload was increased from a central venous pressure of 8 to 12 cm H2O. This established that pulse pressure was abolished even at high preload. At greater speeds (⬎2,000 rpm) there was no pulse pressure in the systemic circulation as confirmed by postoperative invasive monitoring
Fig 1. Representive data of pump flow during the 90 days of left ventricular assist device support. Pump flow measured by the ultrasonic flow probe around the outflow graft was between 4.8 and 7.0 L/min.
for 48 hours, then again before sacrifice. An intercostal drain was inserted, and the wound was then closed in layers. Anesthesia was discontinued, and the sheep were transferred from the operating room to the pen. No anticoagulation was used after the surgical procedure. For reasons of animal welfare the control animals were anesthetized for invasive monitoring before sacrifice but did not undergo a sham operation. They were identical to the LVAD animals for age, body weight, and species. We also did not attempt to insert an arterial catheter into awake animals during the course of the support period; we believed that this risked substantial morbidity such as bleeding and device endocarditis.
Pump Flow Data from the device including motor suspension current, energy requirements, and calculated pump flow were monitored constantly by the online computer. Pump flow was also measured daily by the implanted ultrasound flow probe throughout the study period (Fig 1). Mechanical reliability was determined by recording measures of pump function, by auscultation of the chest, and by inspection of power lines and drive line exit site. The correlation between measured graft flow and calculated pump flow were confirmed.
Recovery The sheep were extubated after 30 to 60 minutes of spontaneous respiration and documentation of satisfactory blood gases and acid-base balance. A vest was placed around the thorax to carry the controller and connect with the electrical power line. Before removal of invasive arterial pressure monitoring at 48 hours, we were able to confirm that the systemic circulation remained nonpulsatile even during vigorous activity during handling. The animals were then free to mobilize, drink, feed, and roam around the sheep pen.
1082
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
Ann Thorac Surg 2002;74:1080 –5
Detailed postmortem examination was performed with macroscopic and histologic examination of the brain, heart, kidneys, liver, and aorta.
Histologic Examination of Major Organs Formalin-fixed, paraffin-embedded tissues from LVAD and control sheep were cut into 3-m-thick sections and stained with hematoxylin and eosin. Differences between LVAD and control specimens were assessed by two independent pathologists who were blind to the origin of the samples. The ascending aorta was cut horizontally into three segments that were embedded in paraffin sections. Sections of 3 m thickness were cut from each block and stained with hematoxylin and eosin. The thickness of the medial layer was determined by computed morphometry (NIH images, Bethesda, MD). Fig 2. Representative computerized arterial pressure traces both in (A) left ventricular assist device (LVAD) and (B) control animals before sacrifice. There was no pulse pressure in left ventricular assist device group animals.
Assessment The sheep were examined for signs of neurologic events, heart failure, systemic thromboembolism, or abnormal behavior. Renal and hepatic function were assessed serially by measurement of blood urea, creatinine, bilirubin, and liver enzymes. Plasma renin activity was measured by standard radioimmunoassay for both groups.
Autopsy Studies Sheep in the LVAD group were electively sacrificed at 30 (n ⫽ 3), 90 (n ⫽ 4), 180 (n ⫽ 1), and 340 (n ⫽ 1) days. Before sacrifice all animals were anesthetized and invasively monitored with an arterial line and Swan-Ganz catheter. The relationship between pump speed and absence of pulse pressure was checked again. Computerized arterial pressure traces before sacrifice are shown in Figure 2. Heparin (5,000 IU/kg) was given intravenously to avoid clot formation on the blood-contacting surfaces of the pump and cannulas. The six control sheep were anesthetized and invasively monitored. They were then sacrificed to compare end-organ morphology.
Statistical Analysis All results for continuous variables were expressed as means ⫾ standard deviations. The paired or unpaired Student’s t test, or Mann-Whitney U test if appropriate, was used to compare continuous variables between two subgroups. Probability values of less than 0.05 were considered to indicate statistical significance.
Results Hemodynamic studies before chest closure showed systemic blood flow to be consistently nonpulsatile at pump rates exceeding 1,600 to 1,800 rpm. All animals in the LVAD group recovered rapidly after operation and survived in excellent condition before elective termination. There was no difference in behavior between LVAD and control animals after recovery from the surgical procedure. The device drive line did not restrict mobility around the pen. There were no neurologic events.
Renal and Hepatic Function Blood tests showed the blood urea and creatinine levels remained stable and within normal limits (Table 1). There were no significant differences between the LVAD and control groups. Indices of hepatic function are shown in Table 2. Hepatic function remained normal for the duration of the study with no difference between the LVAD and the control groups.
Table 1. Comparison of Renal Function Between Left Ventricular Assist Device and Control Groupsa LVAD Group
Variable
Control (n ⫽ 6)
Before Implant (n ⫽ 9)
30 d (n ⫽ 9)
90 d (n ⫽ 6)
180 d (n ⫽ 2)
340 d (n ⫽ 1)
Urea (U/L) Cr (mol/L)
4.2 (1.0) 80 (23)
3.3 (0.5) 76 (21)
3.0 (2.2) 92 (16)
1.6 (1.2) 85 (29)
7.6 (1.1) 85 (7.2)
5.8 84
a
Mean ⫾ SEM.
Cr ⫽ creatinine;
LVAD ⫽ left ventricular assist device.
Ann Thorac Surg 2002;74:1080 –5
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
1083
Table 2. Comparison of Indices of Hepatic Function Between Left Ventricular Assist Device and Control Groupsa LVAD group
Variable
Control Group (n ⫽ 6)
Before Implant (n ⫽ 9)
30 d (n ⫽ 9)
90 d (n ⫽ 6)
180 d (n ⫽ 2)
340 d (n ⫽ 1)
T-Bil (mol/L) GPT (IU/L) GOT (IU/L) ␥-GPT (IU/L)
2.1 (0.7) 38 (23) 92 (25) 56 (11)
2.0 (0.8) 49 (35) 84 (24) 47 (6.0)
2.3 (1.7) 16 (8.1) 104 (24) 63 (7.5)
2.0 (0.2) 15 (2.4) 74 (12) 49 (15)
1.8 (0.2) 14 (1.7) 68 (12) 50 (16)
2.0 24 80 49
a
Mean ⫾ SEM.
GPT ⫽ serum glutamate pyruvate transaminase; GOT ⫽ serum glutamate oxalate transaminase; LVAD ⫽ left ventricular assist device; T-Bil ⫽ total bilirubin.
Plasma Renin Activity and Mean Blood Pressure Mean blood pressure was 94 ⫾ 12 mm Hg in the LVAD animals before implantation and 98 ⫾ 20 mm Hg at the time of sacrifice. Mean blood pressure in the control sheep was 88 ⫾ 36 mm Hg. There was no statistical difference between the groups or before and after LVAD implantation. Plasma renin activity after LVAD implantation was significantly elevated compared with the preoperative values (2.9 ⫾ 0.3 pg/mL versus 1.3 ⫾ 0.4 pg/mL; p ⬍ 0.05). There was also a significant difference between the LVAD animals and the control group (2.9 ⫾ 0.3 pg/mL versus 1.4 ⫾ 0.3 pg/mL; p ⬍ 0.05; Fig 3).
Histology of Major Organs Macroscopic examination of the native heart in LVAD sheep showed minor thrombus formation in the area of stasis between the inflow cannula and ventricular wall. One embolic infarct was found in a single LVAD sheep kidney. Apart from this, detailed histologic comparisons of brain, heart, liver, and kidneys showed no other emboli and no parenchymal differences between the LVAD and the control sheep. The organs remained completely normal in the LVAD sheep for up to 340 days (Fig 4). However, we noted the thickness of the medial
␥-GPT ⫽ gamma guanisine triphosphate;
layer of the ascending aorta in the LVAD group to be significantly diminished (1.8 ⫾ 0.4 mm in the LVAD group versus 2.6 ⫾ 0.6 mm in the control group; p ⬍ 0.05; Figs 5 and 6). Both pathologists agreed on this. Because of the small number of sheep we could not discern a significant difference in aortic wall thickness with time.
Comment The continuous flow pumps are silent and much more compact than the displacement LVADs. They have a smaller and less thrombogenic blood–foreign surface interface, no areas of stasis, and no artificial valves. Device durability may be improved by this approach. Until recently it was inconceivable that a rotary blood pump could function for many months in an animal model without anticoagulation. With careful alignment of the TLVAS inflow cannula we were able to provide consistently nonpulsatile blood flow for virtually 12 months. Measurements of calculated flow correlated well with ultrasonic flowmeter measurements of graft flow, and these levels were constantly above that required to abolish pulse pressure during invasive arterial monitoring and elevated preload. Even when the aortic valve remains continuously closed the left ventricle may genFig 3. Mean blood pressure (BP) and plasma renin activity. (N.S. ⫽ not significant; Pre Op LVAD ⫽ preoperative left ventricular assist device; Post Op LVAD ⫽ postoperative left ventricular assist device.)
1084
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
Ann Thorac Surg 2002;74:1080 –5
Fig 4. Representative photomicrographs for major organs from the left ventricular assist device sheep. (A) Hematoxylin and eosin staining of brain from a sheep on the left ventricular assist device for 90 days. There are no abnormalities. (B) Hematoxylin and eosin staining of liver from a sheep on the left ventricular assist device for 180 days. Sinusoidal structure remained normal. (C) Hematoxylin and eosin staining of myocardium from a sheep on the left ventricular assist device for 340 days. Myocardium and coronary arteries are normal. The parasites found occur frequently in sheep. (D) Hematoxylin and eosin staining of kidney from a sheep on the left ventricular assist device for 340 days. Glomerular and tubular structure are normal.
erate preload to the device, which can be transmitted as pulse pressure to the circulation [2]. Consequently we set the TLVAS at a level at which even this level of pulse pressure could not occur. The most encouraging finding was that long-term circulatory support without pulse pressure can maintain normal end-organ structure and function. Activity, behavior, and neurologic function did not differ from control animals. This outcome gives us confidence that both centrifugal and axial flow blood pumps should provide safe and effective long-term circulatory support in the patient with heart failure. Early clinical experience with axial flow pumps has already shown that diminished pulse pressure and periods without pulse are compatible with recovery from heart failure and improvement in end-organ function [2]. In the absence of a reliable nonpulsatile animal model, most previous investigations of pulseless circulation have centered on CPB [5, 6]. Pulsatile CPB systems were said
Fig 5. Corresponding ascending aortic samples from the left ventricular assist device (LVAD) and the control sheep. There is an easily discernible difference in thickness of the medial layer.
to result in less endothelial damage, reduced systemic vascular resistance, and normal nitric oxide release [6]. Pulsatility plays a role in the movement of lymph in and out of the tissues, perhaps preventing edema and capillary sludging [7]. In the CPB model, cerebral microcirculation was improved by pulsatility, and pulse pressure influenced the activity of the renin-angiotensin system and catecholamine release [8]. In reality these short-term CPB studies fail to predict the effects of long-term pulseless circulation because of the body’s ability to adapt to changes in physiology. In any event pulse pressure is greatly diminished or absent at the capillary level. In nonpulsatile animals upregulation of plasma renin activity could be an adaptive response to lack of pulse pressure. This elevates mean perfusion pressure by increasing the systemic vascular resistance. Systemic vascular resistance is raised during nonpulsatile CPB when compared with pulsatile CPB [9]. Nishimura and colleagues [10] reported that the systemic vascular response
Ann Thorac Surg 2002;74:1080 –5
SAITO ET AL CHRONIC NONPULSATILE CIRCULATION
1085
although we were unable to define differences in the sheep. In conclusion, chronic pulseless circulation was well tolerated in the sheep model, and we could not identify changes in major end organs. The renin-angiotensin system was upregulated, although mean blood pressure was not significantly elevated. Although pulseless circulation caused thinning of the aortic wall, our findings still suggest that continuous flow devices can be used safely for long-term circulatory support. This work was supported financially in part by a research grant from Terumo Cardiovascular System Corp. The authors thank the pathologists Drs Derek Roskell and Andrew Graham for their pathologic studies and Drs Maekawa and Yanai of Terumo for their technical support.
References Fig 6. Measured thickness of the medial layer of the aorta from left ventricular assist device (LVAD) and control groups.
to norepinephrine administration decreased markedly in a chronic nonpulsatile left heart bypass model (up to 137 days), although there was no difference in the plasma norepinephrine levels between pulsatile and nonpulsatile animals. This group also reported atrophic changes in the aorta of nonpulsatile goats in comparison with goats with pulsatile left heart bypass [11]. We showed that the medial layer of the ascending aorta in the LVAD group was consistently thinner than in control sheep. This striking and reproducible finding suggests that we succeeded in providing pulseless circulation throughout the duration of the experiment. The arterial changes can be explained by a reduction in cyclic mechanical load, which leads to atrophy of the medial smooth muscle cells, and we are examining this further. The effect of a change in pulse pressure on the arterial wall has been investigated using an in vitro model of cyclic mechanical stretching on cultured vascular endothelial cells and smooth muscle cells [12]. These experiments also showed that a reduction in a pulse pressure resulted in atrophic changes. Assuming that these morphologic changes extend to the entire arterial system, vascular endothelial function such as blood flow regulation and pressure control could also be compromised,
1. Rose EA, Gelijns AC, Moskowitz AJ, et al. Long-term use of a left ventricular assist device for end stage heart failure. N Engl J Med 2001;345:1435– 43. 2. Westaby S, Banning AP, Saito S, et al. Circulatory support for long term treatment of heart failure. Experience with an intraventricular continuous flow pump. Circulation 2002;105: 2588–91. 3. Saito S, Westaby S, Pigott D, et al. Reliable long-term non-pulsatile circulatory support without anticoagulation. Eur J Cardiothorac Surg 2001;19:678– 83. 4. Nojiri C, Kijima T, Maekawa J, et al. Terumo implantable left ventricular assist system: results of long term animal study. ASAIO J 2000;46:117–22. 5. Wilkens H, Regelson W, Hoffmeister FS. The physiologic importance of pulsatile flow. N Engl J Med 267;443– 6. 6. Sumpio BE, Widmann MD. Enhanced production of an endothelium-derived contracting factor by endothelial cells subjected to pulsate stretch. Surgery 1990;108:227– 82. 7. Parsons RJ, McMaster PD. Effects of the pulse upon formation and flow of lymph. J Exper Med 1938;68:353–76. 8. Jett GK. Physiology of non-pulsatile circulation: acute versus chronic support. ASAIO J 1999:119–22. 9. Golding LR, Murakami G, Harasaki H, et al. Chronic nonpulsatile blood flow. ASAIO Trans 1982;28:81–5. 10. Nishimura T, Tatsumi E, Nishinak T, et al. Diminished vasoconstrictive function caused by long-term nonpulsatile left heart bypass. Artif Organ 1999;23:722– 6. 11. Nishimura T, Tatsumi E, Takaichi S, et al. Prolonged nonpulsatile left heart bypass with reduced systemic pulse pressure cause morphological changes in the aortic wall. Artif Organ 1998;22:405–10. 12. Birkov KG, Shrinsky VP, Stepanova OV, et al. Stretch affects phenotype and proliferation of vascular smooth muscle cells. Mol Cell Biochem 1995;144:131–9.