Mechanical circulatory support using an implantable ventricular assist device (VAD) is applied routinely in patients needing temporary circulatory assistance while awaiting cardiac transplantation or to provide permanent support for patients with end-stage heart failure who are not candidates for heart transplantation.1 Although pulsatile flow conditions are considered the most physiologic, continuous-flow VADs (CFVADs) are also being developed to reduce complications associated with pulsatile technology. Some common complications being addressed are infection, thromboembolic events, device failure, and the needs of patients who are too small to be fitted with current pulsatile VADs.2 Left ventricular assist device (LVAD) recipients had 1 and 2 year survivals that were two to three times greater than patients treated with medical therapy alone.3 Additional studies have demonstrated similar efficacy for LVAD use as a bridge-to-transplant in patients awaiting a donor heart.4 Early follow-up of CFVAD recipients demonstrated improved functional status and quality of life 3 months after implantation, with a 6 months survival of 75%.5 Long-term follow-up of patients with CFVADs demonstrated not only preservation of, but also improvement in, hepatic and renal function.6 Some organs, however, may be very sensitive to nonpulsatile (or reduced pulsatile) blood flow. Gastrointestinal (GI) bleeding occurs in 15–30% of patients implanted with a CFVAD and is 10 times more frequent in patients supported with CFVADs than in patients with pulsatile pumps.7,8 This bleeding appears to be directly associated with reduced pulsatile blood flow, as similar bleeding is not seen in patients with mechanical valves and comparable levels of anticoagulation.9 Unfortunately, the mechanisms underlying GI bleeding are not well outlined and need to be understood to improve the treatment of patients with congestive heart failure (CHF) needing mechanical circulatory support. Several explanations for the association between GI bleeding and CFVAD interaction have been proposed and include: 1) acquired von Willebrand’s disease because of shear stress in CFVAD-implanted patients, 2) increased intraluminal pressure coupled with muscular contraction which may result in dilated mucosal veins and the development of arteriovenous communications, 3) neurovascular etiology in which increased sympathetic tone results in smooth muscle relaxation and a subsequent propensity for angiodysplasia or hypoperfusion of the intestines, and 4) lowered pulse pressure (PP) which leads to intestinal hypoxia, vascular dilation, and angiodysplasia.10 Whether the cause of GI bleeding is related to one or a combination of these factors, GI perfusion assessment is crucial to understand the complex interaction between GI bleeding and CFVAD physiology, and to improve device design to prevent or reduce this serious complication.
The primary purpose of our study was to investigate renal and intestinal perfusion changes in response to continuous and intermittent reduced speed (IRS) modes using microsphere and positron emission tomography/computed tomography (PET/CT) techniques in an acute ovine model implanted with a CFVAD.
The HeartWare HVAD Pump (Miami Lakes, FL) is a continuous-flow, centrifugal VAD. The design of the pump has been described elsewhere.11 For this study, an IRS algorithm was implemented using a custom software application developed on a Labview platform (National Instruments, Austin, TX). The IRS algorithm reduced the rotational speed by 20% of the set speed for 5 seconds with a 10 seconds offset in between each cycle. In this study, we tested two continuous modes (low and high flow) and two IRS modes (low and high flow). Low continuous (LC) and low IRS (LIRS) modes targeted partial left ventricular unloading with a flow ranging between 2.4 and 3.5 L/minute. High continuous (HC) and high IRS (HIRS) modes aimed to fully unload the left ventricle which was possible with flows higher than 3.5 L/minute depending on the size of the animal. During high-flow modes, the aortic valve was closed and there was some degree of ventricular suction without any premature ventricular contractions confirmed through pressure, flow, and electrocardiography signals.
Experiments were conducted in healthy Suffolk Cross sheep (n = 4), each weighing 68 ± 15 kg. Animals received humane care in compliance with the “Principles of Laboratory Animal Care,”12 formulated by the National Society for Medical Research, and the National Institute of Health’s “Guide for the Care and Use of Laboratory Animals.”13 Our institute performed the study in accordance with the Animal and Plant Health Inspection Service, United States Department of Agriculture, Animal Welfare Act, 9 Code of Federal Regulations, Parts 1, 2, and 3 as applicable; and according to an Animal Use Protocol (AUP) approved by the University’s Institutional Animal Care and Use Committee (IACUC).
Anesthesia and Surgical Preparation
After the endotracheal intubation, general anesthesia was maintained with isoflurane (0.5–4.0%) in oxygen (100%). Lactated Ringer’s solution for fluid support was given intravenously (IV) at a rate of 5–20 ml/kg/h via the venous catheter.
The HeartWare HVAD pump was implanted through a left thoracotomy without the use of cardiopulmonary bypass (CPB). The left carotid artery and jugular vein were exposed for monitoring of arterial and venous pressures. Once the heparin (250 IU/kg) was administered, the HeartWare HVAD pump outflow graft was anastomosed to the descending thoracic aorta. A titanium/polyester sewing cuff was sewn to the ventricular apex. Then the HeartWare HVAD pump was inserted into the ventricular apex and secured to the sewing ring. After de-airing the graft, a 10 mm ultrasonic flow probe (Transonics Inc., Ithaca, NY) was attached to the outflow graft for flow measurement. The graft was clamped to avoid retrograde flow into the left ventricle until data collection began. A fluid-filled catheter was inserted into the left ventricle for left ventricular pressure monitoring.
Via a left flank incision to the abdomen, 14 gauge catheter was inserted into the cranial mesenteric vein to selectively capture intestinal venous drainage and for serial venous blood gas measurements to calculate intestinal metabolism. The catheter end was tunneled through the skin. The chest and abdomen were closed in a single layer with a #1 nylon suture to avoid exposure to room air. The animal was then transported to the PET/CT suite.
After pump implantation, hemodynamics were allowed to stabilize for 60 minutes. Hemodynamics were assessed at the baseline and with the pump operating at the four experimental settings described previously (LC, HC, LIRS, and HIRS modes). After each change to a new pump setting, hemodynamics were allowed to stabilize for 30–45 minutes before collecting hemodynamic data. Microsphere injections were performed concomitantly, immediately followed by Cu62 injections. Heart rate (HR), systolic aortic pressure (AoPs), diastolic aortic pressure (AoPd), mean aortic pressure (AoPm), PP, central venous pressure (CVP), left ventricular systolic pressure (LVP), and pump flow (PF) were measured at each setting.
Pressure and flow data were continuously recorded with a computer data acquisition system (ADInstruments Power Lab ML880; LabChart Pro, Colorado Springs, CO).
Intestinal and Renal Blood Perfusion Assessment with Microspheres
Regional blood flow (RBF) distribution to the proximal jejunum, distal ileum, and right renal cortex was evaluated by microsphere analysis using 15 μm nonradioactive microspheres (BioPAL, Worcester, MA), as described in detail elsewhere.14 In brief, the labeled microspheres (gold, samarium, ytterbium, europium, and terbium) were injected into the left atrium at the baseline and at each experimental pump setting (LC, HC, LIRS, and HIRS modes). Arterial reference blood samples were taken from the left common carotid artery and used for normalizing tissue sample readings. At study termination, target abdominal organs were harvested and 2 g of tissue were taken from each for microsphere analysis. A neutron activation technique was used to detect microspheres in tissue samples.
Intestinal and Renal Blood Perfusion Assessment with PET/CT
The animal was transported to the PET/CT suite (Siemens 128 slice Biograph mCT PET/CT; Siemens Molecular Imaging, Knoxville, TN) and secured to the CT table in right lateral recumbency. A baseline CT study was obtained to provide anatomic reference and attenuation correction. Omnipaque 350 (1 ± 0.1 ml/kg IV) was infused at a rate of 4 ml/second during CT scans to highlight the blood pool. Abdominal PET perfusion studies were performed for baseline as well as for each pump setting. 62Cu-labeled pyruvaldehyde bis (N-4-methylthiosemicarbazone) copper (62Cu-PTSM), a freely diffusible, metabolically trapped perfusion tracer, was infused IV at a dose of 5–10 mCi/animal/study, during each PET study. All studies were compensated for body mass and the actual amount of tracer injected. Scans were started 10 seconds before infusion to obtain baseline activity information, and data were acquired listmode for 600 seconds total duration. Image data were reframed at 3 seconds/frame for the first 20 frames to accurately define the input bolus, followed by 1 minute/frame to obtain the final disposition. All intestinal and renal perfusion images were reconstructed iteratively with Ordered Subsets Expectation Maximization (OSEM), scatter and attenuation correction applied. Approximately 30–45 minutes of interval was given for 62Cu decay (9.67 minutes half-life) and hemodynamic adjustment to the new speed change. The arterial input function (AIF) was extracted from the time sequence images using automated methods (U.S. Patent Application 61/736,242) and applied to the remainder of the regions-of-interest (ROIs) to generate parametric images of tissue perfusion. This automated approach allowed the selection of an AIF from the entire three-dimensional volume at one time, which increased the quantitative measurement efficiency and interoperator consistency.
Intestinal Arteriovenous Oxygen Difference Assessment
Immediately before each microsphere injection, arterial and venous blood samples were concomitantly withdrawn from the carotid artery and cranial mesenteric vein catheters. This occurred at baseline and at each subsequent speed setting to calculate intestinal arteriovenous oxygen difference and CO2 production (TCO2) changes which are indirect indicators of arteriovenous shunting (AVS) or low oxygen utilization (LOU) and cell metabolism. The differentiation between AVS and LOU is based on the formula proposed by Rozin et al.15: (Venous TCO2-Arterial TCO2). Increased or unchanged CO2 production along with venous hyperoxia compared with baseline is considered a result of AVS, whereas decreased CO2 production is considered sign of LOU.
All statistical tests were performed using Microsoft Office Excel software (Microsoft Corp., Redmond, WA). Analysis of variance (ANOVA) was used to compare continuous variables. p values of less than 0.05 were considered significant.
All sheep (n = 4) underwent successful pump implantation, experienced no surgical complications or mechanical device failures, and survived till study termination.
Hemodynamic data and LVAD-related parameters are summarized in Table 1. The interactions between AoPs, AoPm, AoPd, PP, and CVP at baseline and the different pump settings are shown in Figure 1. Changes in LVP at different pump settings are shown in Figure 2.
HR, AoPs, and AoPm were not significantly affected compared with baseline values at different pump settings. AoPd increased significantly at HC and LIRS settings (p < 0.05 versus baseline), whereas no significant changes were observed in LC and HIRS modes despite higher AoPd values compared with baseline settings. PP was maintained close to baseline values at CL setting, whereas significant decreases were observed at CH, LIRS, and HIRS settings. LIRS and HIRS settings had higher PP compared with CH, but the difference was not statistically significant. CVP remained near baseline values at all speed changes. At LC and LIRS settings, LVP remained close to baseline values without significant statistical changes. LVP decreased significantly at HC and HIRS settings compared with baseline, LC, and LIRS levels, respectively (p < 0.05).
Intestinal and Renal Blood Perfusion Assessment with Microspheres and PET/CT
The results of microsphere and PET/CT analyses of renal and GI perfusion are summarized in Table 2.
As shown in Figure 3, A and B, kidney, proximal jejunum, and distal ileum regional perfusions measured with microspheres increased in all pump settings except the CL mode compared with baseline levels where the increase was not statistically significant. PET/CT data revealed similar trends of increased perfusion in the kidney and proximal jejunum in CL, LIRS, and HIRS modes, whereas distal ileum perfusion remained close to the baseline values and showed nonsignificant perfusion increases at HIRS mode only.
Intestinal Arteriovenous Oxygen Difference Assessment
Serial aortic and cranial mesenteric vein blood gas analysis revealed significant increases in venous partial oxygen pressures at the intestinal level compared with baseline values in all PF modes (Figure 4). Intestinal arteriovenous CO2 production (TCO2) significantly increased in all pump settings compared with baseline levels (p < 0.05) except LIRS mode (p = 0.07; Figure 5). There was no statistical significance between various pump settings.
By using an LVAD (Heartware HVAD pump) capable of producing both continuous and IRS flow modes in a healthy ovine model, we have shown that renal, proximal jejunal, and distal ileal RBFs measured with microsphere and PET/CT techniques did not produce any significant over- or underperfusion changes compared with baseline values in either continuous or reduced pulsatile modes. Although there were fundamental differences between the ROI measurements provided with 62Cu-PTSM/PET/CT and microspheres techniques, the trends closely matched. The microsphere technique used a very small amount of tissue sample (2–3 g) to measure the RBF with a high degree of accuracy and relies on capillary size to trap the material, whereas PTSM is a fully diffusible small molecule that is metabolically bound in place. Nearly all microspheres are lodged in the first pass, whereas it takes several passes for PTSM to bind. PET/CT provides uniform sampling throughout the field of view, enabling measurement over a large volume of tissue as opposed to microspheres. For example, perfusion over the entire right kidney and GI tract was measured with PET/CT versus only 2 g of tissue samples collected from the right kidney and the first and last 10 cm of small intestine for microspheres. Therefore, because only a small sample was taken for microsphere analysis, small regional differences that can bias a microsphere measurement are evened out statistically by selecting a larger ROI with PET. Although the microsphere technique is widely accepted as a “gold standard” in perfusion assessment studies, our study showed that the PET/CT technique with 62Cu-PTSM is equally reliable in renal and GI regional perfusion measurements.
The second important observation of our study was a significant increase in cranial mesenteric vein partial oxygen pressures in both continuous and reduced speed flow modes. Selective cannulation of this vein allowed us to measure changes in partial oxygen pressures specific to the small and large intestines. Considering that this is a healthy heart model, under partial (CL and LIRS modes) or full (HC and HIRS modes) ventricular unloading simulations, the total cardiac output and blood flow to the intestines theoretically should not change, nor should intestinal tissue perfusion, which was proven with microsphere and PET/CT measurements. Therefore, the only variable that may cause immediate venous hyperoxia is the effect of the continuous-flow physiology. Two conditions, AVS and LOU, may exist simultaneously as result of venous hyperoxia in the tissues; thus, there is a need to differentiate one condition from the other. It is well known that under normal tissue perfusion conditions, one O2 molecule is used as one CO2 molecule is produced; however, LOU is associated with low CO2 production. Therefore, AVS may be associated with relatively high O2 use from the capillary vessels and increased CO2 production. Our studies revealed a significant increase in TCO2 production at the intestinal region in both continuous and reduced speed modes as calculated by the formula proposed by Rozin et al.15 We believe that this is strong evidence of immediate intestinal AVS opening under continuous-flow physiology without intestinal tissue perfusion defect as proved by microsphere and PET/CT measurement. Although one may speculate that the immediate occurrence of AVS at the GI level may be linked to the development of intestinal arteriovenous malformations and intestinal bleeding complication, we refrain from making such predictions without chronic in vivo studies.
Favorable outcomes of the continuous-flow technology are mostly attributed to device-related benefits such as smaller pump size, ease of implantation, reduced infection, and postoperative bleeding rates, rather than the flow pattern itself. Therefore, there is a growing interest to provide more induced pulsatility to the circulatory system of the patients supported by continuous-flow pumps to provide better unloading without left ventricular suction, aortic valve opening, pulse transmission, and end-organ perfusion.16,17 In our study, the perfusion trends in target organs (renal and GI system) were similar in both continuous and IRS modes which are similar to one of our earlier studies comparing the effects of continuous flow and induced pulse modes on renal perfusion.18 Nevertheless, it must be considered that in our current experiments, only one asynchronous IRS mode under low and high PF conditions was used. Additional studies are needed to investigate the effects of the electrocardiography triggered or other artificially induced pulse modes on various end-organs. To our knowledge, there are no previous studies investigating the acute or chronic effects of pure continuous or induced pulse flow physiology on GI perfusion and AVS formation provided by a continuous-flow VAD. Our acute study is the first and only in vivo study focused on this topic.
Our study did possess some limitations. First, this is a healthy and anesthetized animal model which does not directly simulate the real chronic heart failure and awake patient clinical scenario. Second, our study used relatively small animal numbers, but we believe that the consistency of our preliminary results in these four animals are important and may shed some light toward future studies focused on this topic. Third, the sheep is an herbivore with four stomachs that have nearly a 20–30 L capacity. Their small intestine is seven times longer than that of humans and does not precisely simulate human omnivore histology and physiology. Nevertheless, considering the relatively minimal limitations of using animal models for assist device testing, we believe that our results provide some insight into unknown complex device–intestine interaction.
Results of the current study suggest that renal and intestinal tissue perfusion is not adversely and differentially affected by a cardiac assist device loaded with continuous and IRS modes. Immediate cranial mesenteric venous hyperoxia and increased CO2 production occurred during all pump modes and may be evidence of intestinal AVS. Longer-term studies are warranted to test the validity of these findings in chronic heart failure models.
1. Frazier OH, Myers TJ, Westaby S, Gregoric ID. Use of the Jarvik 2000 left ventricular assist system as a bridge to heart transplantation or as destination therapy for patients with chronic heart failure. Ann Surg. 2003;237:631–636
2. John R, Kamdar F, Liao K, Colvin-Adams M, Boyle A, Joyce L. Improved survival and decreasing incidence of adverse events with the HeartMate II left ventricular assist device as bridge-to-transplant therapy. Ann Thorac Surg. 2008;86:1227–1234
3. Rose EA, Gelijns AC, Moskowitz AJ, et al.Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) Study Group. Long-term use of a left ventricular assist device for end-stage heart failure. N Engl J Med. 2001;345:1435–1443
4. Slaughter MS, Rogers JG, Milano CA, et al.HeartMate II Investigators. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med. 2009;361:2241–2251
5. Miller LW, Pagani FD, Russell SD, et al.HeartMate II Clinical Investigators. Use of a continuous-flow device in patients awaiting heart transplantation. N Engl J Med. 2007;357:885–896
6. Russell SD, Rogers JG, Milano CA, et al.HeartMate II Clinical Investigators. Renal and hepatic function improve in advanced heart failure patients during continuous-flow support with the HeartMate II left ventricular assist device. Circulation. 2009;120:2352–2357
7. Letsou GV, Shah N, Gregoric ID, Myers TJ, Delgado R, Frazier OH. Gastrointestinal bleeding from arteriovenous malformations in patients supported by the Jarvik 2000 axial-flow left ventricular assist device. J Heart Lung Transplant. 2005;24:105–109
8. Crow S, John R, Boyle A, et al. Gastrointestinal bleeding rates in recipients of nonpulsatile and pulsatile left ventricular assist devices. J Thorac Cardiovasc Surg. 2009;137:208–215
9. Slaughter MS. Hematologic effects of continuous flow left ventricular assist devices. J Cardiovasc Transl Res. 2010;6:618–624
10. Suarez J, Patel CB, Felker GM, Becker R, Hernandez AF, Rogers JG. Mechanisms of bleeding and approach to patients with axial-flow left ventricular assist devices. Circ Heart Fail. 2011;4:779–784
11. Larose JA, Tamez D, Ashenuga M, Reyes C. Design concepts and principle of operation of the HeartWare ventricular assist system. ASAIO J. 2010;56:285–289
12. . National Society for Medical Research. Principles of laboratory animal care. Bio-Medical Purview. 1961;1
13. . National Institutes of Health. Guide for the Care and Use of Laboratory Animals. 1963 Bethesda, MD National Institutes of Heath
14. Reinhardt CP, Dalhberg S, Tries MA, Marcel R, Leppo JA. Stable labeled microspheres to measure perfusion: Validation of a neutron activation assay technique. Am J Physiol Heart Circ Physiol. 2001;280:H108–H116
15. Rozin AP, Attias J, Presser D, Rosenberg H, Moscovitz M, Bentur Y. Alcohol poisoning and venous hyperoxia. Toxicol Mech Methods. 2008;18:745–750
16. Ando M, Nishimura T, Takewa Y, et al. Electrocardiogram-synchronized rotational speed change mode in rotary pumps could improve pulsatility. Artif Organs. 2011;35:941–947
17. Pirbodaghi T, Axiak S, Weber A, Gempp T, Vandenberghe S. Pulsatile control of rotary blood pumps: Does the modulation waveform matter? J Thorac Cardiovasc Surg. 2012;144:970–977
18. Eya K, Tuzun E, Conger J, et al. Effect of pump flow mode of novel left ventricular assist device upon end organ perfusion in dogs with doxorubicin induced heart failure. ASAIO J. 2005;51:41–49
continuous flow; cardiac assist device; gastrointestinal perfusion; gastrointestinal bleeding