Development status of Terumo implantable left ventricular assist system.

We have been developing an implantable left ventricular assist system (T-ILVAS) featuring a magnetically suspended centrifugal pump (MSCP) since 1995. In vitro and in vivo studies using a prototype MSCP composed of a polycarbonate housing and impeller (196 ml) have demonstrated long-term durability and excellent blood compatibility for up to 864 days, and excellent stability of the magnetic bearing of the MSCP. These preliminary results strongly suggested that the magnetic bearing of the MSCP is reliable and is a most feasible mechanism for a long-term circulatory assist device. We have recently devised a clinical version pump made of titanium (180 ml) with a new position sensor mechanism and a wearable controller with batteries. Cadaver fit study confirmed that the Type IV pump could be implanted in a small patient with a body surface area as small as 1.3. The in vitro performance tests of the Type IV pump demonstrated excellent hydrodynamic performances with an acceptable hemolysis rate. New position sensors for the titanium housing showed more uniform sensor outputs of a magnetic bearing than in the prototype polycarbonate pump. The Type IV pump then was evaluated in vivo in 6 sheep at the Oxford Heart Centre. Four sheep were electively sacrificed at 3 months and were allowed to survive for more than 6 months for long-term evaluation. In this particular series of experiments, no anticoagulant/antiplatelet regimen was utilized except for a bolus dose of heparin during surgery. There was a left ventricular mural thrombi around the inflow cannula in 1 sheep. Otherwise, there was no mechanical failure nor sign of thromboembolism throughout the study.

Reliable long-term non-pulsatile circulatory support without anticoagulation.

OBJECTIVE: The Terumo implantable left ventricular assist system (T-ILVAS) consists of a titanium centrifugal pump with a unique magnetically suspended impeller producing continuous (non-pulsatile) flow up to 10 l/min. The interior surface is heparin-coated and there is no purge system. We implanted the device into six sheep to ascertain in-vivo haemodynamic function, mechanical reliability and biocompatibility. METHODS: The T-ILVAS was implanted via left thoracotomy without cardiopulmonary bypass. The inflow cannula was placed in the left ventricular apex and a Dacron outflow graft anastomosed to the descending aorta. All animals recovered well. No anticoagulation (heparin or warfarin) was given after the surgery. Suspension position, motor current, impeller speed and pump flow were continuously monitored and stored by on-line computer. Serial blood samples were collected to determine haematological and biochemical indices of renal function, liver function and haemolysis. All animals were electively euthanized between 3 and 7 months postoperatively. The explanted pumps were examined for mechanical reliability and thrombus formation. Major organs were examined macroscopically and histologically for thromboembolism. RESULTS: All animals appeared completely normal for up to 210 days. At speeds between 1500 and 2000 rev./min the device pumped up to 8 l/min capturing all mitral flow. There were no major complications (pump failure, thromboembolism, haemorrhage, or driveline infection). Indices of haemolysis, liver and renal function remained within normal limits. All pumps were mechanically sound and free from thrombus. One embolus was found in a sectioned kidney. CONCLUSION: The T-ILVAS successfully supported the systemic circulation without anticoagulation for up to 210 days. Mechanical reliability and biocompatibility were demonstrated. Organ function remained within normal limits during continuous non-pulsatile flow.

Compact, reliable ventricular assist device as a bridge to recovery or for semipermanent use.

OBJECTIVES: The magnetically suspended centrifugal pump has gained attention as an implantable ventricular assist device for long-term use. We report recovery-oriented pump operation and results of a chronic animal experiment. METHODS: In an acute experiment in 8 sheep having microspheres injected to induce heart failure, left ventricular assist was implemented by 2 inflow cannulas, 1 each in the left atrium and left ventricle. The pressure-volume loop of the left ventricle and myocardial oxygen consumption were measured varying the assist rate. The chronic animal experiment used 10 sheep whose native heart was kept intact. Continuous hemodynamic monitoring and periodic blood sampling were conducted. RESULTS: In the acute study, myocardial oxygen consumption decreased proportionally with increasing assist rate in left atrial drainage, but was significantly less at a 100% assist rate in left ventricular drainage. External work in left ventricular drainage did not decrease until a 75% assist rate, suggesting that the left ventricle shape and size were maintained despite decreased myocardial oxygen consumption. In the chronic experiment, the pumping duration was 14 to 248 days. No thrombi or emboli in the pump or any major organ were found in sacrificed sheep after living more than 100 days. Hepatic and renal function were within an almost normal range throughout the experiment. CONCLUSION: Left ventricular blood drainage effectively reduced oxygen consumption, maintaining the shape of the left ventricle. The magnetically suspended centrifugal pump is suitable for recovery of a failing heart or semipermanent use.

Physical model-based indirect measurements of blood pressure and flow using a centrifugal pump.

This article describes a technique offering indirect measurements of pump pressure differential and flow with certain accuracy independent of changes in blood viscosity. This technique is based on noninvasive measurements of the motor current and rotation speed using the physical model equations of the centrifugal pump system. Blood viscosity included in the coefficients of the dynamic equations is first estimated, and then substitution of the estimated viscosity into the steady equations of the model provides pump flow and pressure differential. In vitro tests using a Capiox pump showed a sufficient linear correlation between actual values and their estimates for pressure differential and pump flow. An in vivo test using a 45 kg sheep showed that the proposed algorithm needs robustness for the convergence of estimates of viscosity. An overall evaluation, however, of the developed algorithm/model showed indications of success in terms of efficient computation and modeling.

Engineering analysis of diamond-like carbon coated polymeric materials for biomedical applications.

Diamond-like carbon (DLC) films have received much attention recently owing to their properties, which are similar to diamond: hardness, thermal conductivity, corrosion resistance against chemicals, abrasion resistance, good biocompatibility, and uniform flat surface. Furthermore, DLC films can be deposited easily on many substrates for wide area coat at room temperature. DLC films were developed for applications as biomedical materials in blood contacting-devices (e.g., rotary blood pump) and showed good biocompatibility for these applications. In this study, we investigated the surface roughness by Atomic Force Microscopy (AFM) and Hi-vision camera, SEM for surface imaging. The DLC films were produced by radio frequency glow discharge plasma decomposed of hydrocarbon gas at room temperature and low pressure (53 Pa) on several kinds of polycarbonate substrates. For the evaluation of the relation between deposition rate and platelet adhesion that we investigated in a previous study, DLC films were deposited at the same methane pressure for several deposition times, and film thickness was investigated. In addition, the deposition rate of DLC films on polymeric substrates is similar to the deposition rate of those deposited on Si substrates. There were no significant differences in substrates' surface roughness that were coated by DLC films in different deposition rates (16-40 nm). The surface energy and the contact angle of the DLC films were investigated. The chemical bond of DLC films also was evaluated. The evaluation of surface properties by many methods and measurements and the relationship between the platelet adhesion and film thickness is discussed. Finally, the presented DLC films appear to be promising candidates for biomedical applications and merit investigation.

Improved blood compatibility of DLC coated polymeric material.

There is currently an increasing interest in the use of DLC (diamond like carbon) films in biomedical applications. These investigations making use of DLC in the biomedical area indicate its attractive properties. In this study, we succeeded in depositing DLC on polymer substrates and found the best conditions and method for this application. We evaluated the blood compatibility of polycarbonate substrates coated by DLC (PC-DLC) under different conditions by using epifluorescent video microscopy (EVM) combined with a parallel plate flow chamber. Segmented polyurethane (SPU), which has been used to fabricate medical devices including an artificial heart, and proven to have acceptable blood compatibility, was compared with polycarbonate substrates coated with DLC film. The EVM system measured platelet adhesion on the surface of the DLC, by using whole human blood containing Mepacrine labeled platelets perfuse at a wall shear rate of 100 s(-1) at 1 min intervals for a period of 20 min. PC-DLC demonstrated that Tecoflex showed higher complement activation than PC-DLC. There were significant differences between the PC-DLC substrates. On the basis of these results, it is recommended for use as a coating material in implantable blood contacting devices such as artificial hearts, pacemakers, and other devices. This DLC seems to be a promising candidate for biomaterials applications and merits further investigation.

Measurement of blood hematocrit inside the magnetically suspended centrifugal pump using an optical technique: application to assessment of pump flow.

To measure blood hematocrit inside the magnetically suspended centrifugal pump, we have performed both forward and backward light scattering measurements using a specially designed optical cell. In the forward scattering measurement, an optical fiber was used to guide the near infrared light at 780 nm into a 250 microns gap region, and the light that forward scattered toward a detector fiber was measured using a phototransistor. The light intensity decreased exponentially with an increase in the hematocrit to around 20%. The forward scattering method suffered from sensitivity at the hematocrit levels around 25-45% due to the diffusion effect. By making the optical path length larger than several millimeters, the sensitivity of the forward scattering method in terms of hematocrit change can be improved. In the back scattering method, however, better sensitivity in terms of hematocrit change from 0-50% was obtained. By making the optical fiber separation distance less than 1 mm, the system will measure the first order back scattering from the shallow layer while, by making the fiber separation distance larger than several millimeters, the system will primarily measure the diffuse reflectance from the deeper layer. Both approaches will yield sensitive optical intensity change in terms of the physiological hematocrit range.

Comparative study of biocompatibility between the open circuit and closed circuit in cardiopulmonary bypass.

The theoretical benefit of a centrifugal pump or heparin coating demonstrated through in vitro or in vivo studies is not recognizable in cardiopulmonary bypass (CPB) during chemical open heart surgery. The objective of this study was to investigate the influence of the interface of air and blood in current CPB with an open circuit system and its relative significance in relationship to the heparin dose and heparin coating. Using the same oxygenator and circuit, an open circuit and closed circuit CPB with the same priming volume were prepared for a 4 h perfusion experiment using diluted and heparinized (3.6 U/ml) fresh human blood. In these experiments, both heparin-coated and noncoated circuits were examined. Blood was sampled before and 2, 30, 60, 120, and 240 min after the start of perfusion, and the platelet and white blood cell counts and beta-thromboglobulin (beta-TG) and C3a levels were measured. The amount of adsorbed protein in the hollow fibers was also measured after retrieval. Although the results demonstrated significantly better biocompatibility of the heparin-coated circuit than the noncoated circuit, the difference between the open and closed circuits was unexpectedly small and insignificant with either the heparin-coated circuit or noncoated circuit. In contrast, the C3a level was higher in the closed circuit than the open circuit. However, the amount of adsorbed protein was markedly lower in the closed circuit (0.7 microgram/cm2) than in the open circuit (11.1 micrograms/cm2). An immunoblot of the adsorbed protein showed a higher density of fibrinogen bands and conversion to fibrin in the open circuit. We speculate that the lower blood C3a level in the open circuit suggests that C3a was taken in by the adsorbed protein. In conclusion, analysis of the adsorbed protein indicates the lower biocompatibility of the open circuit. Similar experiments with less heparin use and more severe conditions will be necessary to elucidate the essential benefit of making a CPB closed circuit.

Recent progress in the development of Terumo implantable left ventricular assist system.

The research group of the Terumo Corporation, the NTN Corporation, and Setsunan University (T. Akamatsu) has been developing an implantable left ventricular assist system (ILVAS) featuring a centrifugal blood pump with a magnetically suspended impeller (MSCP). The impeller of the MSCP is suspended by a magnetic bearing, providing contact-free rotation of the impeller inside the pump housing. Thus the MSCP is expected to provide years of long-term durability. Ex vivo chronic sheep experiments using the extracorporeal model (Model I) demonstrated long-term durability, nonthrombogenicity, and a low hemolysis rate (plasma free Hb <6 mg/dl) for more than 2 years. The prototype implantable model (Model II; 196 ml, 400 g) was evaluated ex vivo in 2 sheep and intrathoracically implanted in a small sheep (45 kg). These experiments were terminated at 70, 79, and 17 days, respectively, because of blood leakage through the connector system within the housing of Model II. There was no thrombus formation on the retrieved pump surfaces. A new connector system was introduced to the Model II pump (modified Model II), and the pump was intrathoracically implanted in a sheep. Pump flow rate was maintained at 3-7 L/min at 1700-1800 rpm. The temperature elevation on the surfaces of the motor and the electromagnet inside the pump casing was kept less than 6 degrees C. The temperature of the tissue adjacent to the pump casing became normal 10 days postoperatively. The sheep survived for more than 5 months without any sign of mechanical failure or thromboembolic complication. In vitro real-time endurance tests of motor bearings made of stainless steel and silicone nitride have been conducted for more than 1 year without any sign of bearing wear. The next prototype system (Model III), with an implantable controller and a new MSCP with reduced input power, has been developed with a view toward a totally implantable LVAS.

Results of chronic animal experiments with a new version of a magnetically suspended centrifugal pump.

We have developed a magnetically suspended centrifugal pump (MSCP) for long-term ventricular support. This study reports results of chronic animal experiments using a new version of the MSCP. Three sheep weighing 50-70 kg were used in this study. A left heart assist system was established with cannulas into the descending aorta and the left ventricular apex. In two sheep the MSCP was positioned outside the body and in one sheep implanted on the chest wall. The pumping flow was estimated by the motor current and motor speed. The temperature of the pump and the muscle near the pump was recorded for 10 days after operation. The duration of continuous pumping was 60, 140, and 230 days+ (ongoing), respectively. The cause of termination was infection associated with thrombus formation in the first, and failure of magnetic suspension in the second sheep. No thrombus or embolus was observed after sacrifice of the second sheep. The third sheep has been going well despite skin necrosis around the pump pocket. The estimation of pumping flow was reliable even at 140 days. Temperature of the pump surface was 42 degrees C immediately after the operation and gradually reduced to 41 degrees C. The MSCP is a reliable pump for long-term circulatory assist.

Comparative blood compatibility of polyether vs polycarbonate urethanes by epifluorescent video microscopy.

The segmented polyether urethanes (PEUs) have been used in implantable medical devices due to excellent mechanical properties, acceptable blood compatibility, and good biostability. However, recent studies demonstrate that the polyether soft segment of PEU is susceptible to oxidative degradation in vivo due to scission of the polyether group. Recently, polycarbonate urethanes (PCUs) having no ether linkage in the soft segment have been developed, and show improved stability against oxidative degradation over PEUs. The current study evaluates blood compatibility of these PCUs in comparison with PEUs using epifluorescent video microscopy (EVM) combined with a parallel plate flow cell. The authors selected two PCUs, Corethane 80A (Corvita Corporation, Miami, FL) and PCU(1560), and two PEUs, Pellethene 2363-80AE (Dow Chemical Japan, Tokyo, Japan) and Tecoflex EG80A (Thermedics, Inc., Woburn, MA), all of which have similar hard segment compositions (MDI or HMDI:1,4-butanediol(BD)) and the same hardness of 80A. The EVM measured the amount of platelet coverage on the surfaces using human whole blood perfused at a wall shear rate of 100/sec for 20 min. Complement activation (C3a) also was measured. Both PEUs, especially Pellethane, showed significantly higher platelet adhesion than the PCUs (p < 0.05). There were no significant differences in platelet adhesion between the two PCUs. As for C3a measurements, Tecoflex showed higher complement activation than the others. Based on these results, it is recommended that PEUs should be replaced by ether free PCUs for use in implantable blood contacting devices such as artificial hearts and pacemaker lead insulators.

More than 1 year continuous operation of a centrifugal pump with a magnetically suspended impeller.

The authors have been developing a centrifugal pump with a magnetically suspended impeller (MSCP) designed for total artificial heart and long-term ventricular support. The MSCP consists of a magnetic bearing, an impeller and housing, and a driving motor. The impeller is suspended by a magnetic bearing, therefore providing contact free rotation of the impeller inside the pump. This study was designed to evaluate long-term durability and nonthrombogenicity of the MSCP in a chronic sheep model. The blood contacting surfaces of the pump and conduits were completely modified by a heparin immobilization technique (Hepaface). The MSCP was placed paracorporeally as a left heart bypass between left ventricle and descending aorta in three adult sheep. Coumadin was given orally to maintain prothrombin time at 15-20 sec. The coagulation and hematologic parameters, including plasma free hemoglobin, were periodically monitored throughout the experiment. Under daily movement in the cage, the pump could produce average flow rates of 3-6 L/min (50-100 ml/kg) at 1,700-2,000 rpm. Although the arterial pulse contour decreased, there was no physiologic deterioration. The axial impeller excursion monitored by a position sensor was < 25 microns. Plasma free hemoglobin level remained at < 5 mg/dl throughout the experiment. There was no increase in the motor current, which indicates no massive thrombus formation around the impeller. One experiment was terminated at 70 days due to Hall sensor dysfunction of the motor. The retrieved pump was entirely free from thrombus formation. There was no detectable thrombus formation inside the pump or the inflow and outflow conduits. Hematologic, renal, and hepatic parameters remained within the normal range throughout the experiment. The other two sheep have survived for more than 395 and 41 days without major complication. These studies demonstrated that the MSCP has significant potential for long-term use.

Photochemical grafting of alpha-propylsulphate-poly(ethylene oxide) on polyurethane surfaces and enhanced antithrombogenic potential.

The aim of this study is the grafting of photoreactive alpha-propylsulphate-poly(ethylene oxide) (PEO-SO3), one end of which is capped with an azidophenyl group, on polyurethane (PU) surfaces via a photochemical technique. The anti-Factor Xa activity and the platelet adhesion characteristics of the modified PU surface were evaluated by a chromogenic assay method and by a flow-controlled chamber method, respectively. X-ray photoelectron spectroscopy analysis showed that PEO-SO3 was covalently grafted on the PU surface. The grafted surface showed anti-Factor Xa activity in the presence of antithrombin III, and significantly reduced platelet adhesion characteristics as compared with those of the unmodified PU surface. These results suggest that the grafting of PEO-SO3 improves the antithrombogenicity of PU surfaces.

Can heparin immobilized surfaces maintain nonthrombogenic activity during in vivo long-term implantation?

The authors previously demonstrated that heparin immobilized surfaces showed excellent nonthrombogenic properties for extracorporeal membrane oxygenation experiments as long as 168 hr. The characteristics of the heparin immobilized surfaces include high heparin bioactivity and prevention of platelet adhesion and complement activation. However, it is not known whether the heparin immobilized surfaces would be effective for in vivo long-term implantation. Heparin bioactivity may be lost because of complete degradation or blocking of binding sites on heparin by adsorbed proteins. This study attempted to elucidate the in vivo long-term fate of heparin immobilized surfaces. The blood contacting surfaces of the ventricular assist device (VAD) made from polyurethane was modified with heparin immobilization and evaluated in a long-term sheep left VAD (LVAD) model for as long as 3 months. After removal of the VAD, heparin bioactivity was measured by Factor Xa assay. The blood contacting surfaces were analyzed with a scanning electron microscope, and the adsorbed proteins on the surfaces of the diaphragm were analyzed by SDS-PAGE and Western blotting. The thickness of adsorbed proteins on the surfaces also was measured by a confocal laser microscope. For the control ventricular assist devices, thrombus formation was observed within 1 month, whereas heparin immobilized VADs were able to operate thrombus free for periods as long as 3 months. The control surfaces demonstrated a thick adsorbed protein layer on thin surfaces, whereas heparin immobilized surfaces maintained thinner adsorbed proteins on thin surfaces. Anti Factor Xa activity of the heparinized surfaces disappeared after 15 days, but the surfaces remained nonthrombogenic even after heparin bioactivity was completely lost. The protein composition analyzed by SDS-PAGE showed an albumin dominant pattern on the heparinized surfaces. The band of 110 kD corresponding to C3b was detected only on the control surfaces, which possibly activated complement, and subsequently activated platelets and coagulation. Immunoblot showed degradation products of fibronectin and vitronectin on the control surfaces, which probably were promoted by surface generated protease, whereas the heparinized surfaces showed minimal degradation throughout the experimental periods. These results suggest that the heparin moiety has an ability to control adsorbed proteins, thereby inhibiting thrombus formation during in vivo long-term implantation.

In vitro studies of immobilized heparin and sulfonated polyurethane using epifluorescent video microscopy.

In situ surface modification techniques to improve the blood compatibility of blood contacting surfaces of medical devices have been developed by the authors. The techniques include heparin immobilization and sulfonated polymer grafting onto a polyurethane (PU) surface by using either ozone oxidation or photo reaction. These modified PUs were evaluated using an epifluorescent video microscope combined with a parallel plate flow cell. The epifluorescent video microscope system measured the amount of platelet coverage on the PU surfaces using whole human blood containing mepacrine labeled platelets perfused at a wall shear rate of 100 sec-1 for 20 min. Platelet activation and complement activation were also measured. Both immobilized heparin and sulfonated PUs showed significantly lower levels of platelet adhesion than the control PU. The platelet activation levels of these modified PUs also correspond to the results of the platelet adhesion. As for complement activation, heparin the immobilized surface showed the least complement activation, while sulfonated PU and the control PU showed higher levels of complement activation. In situ surface modification techniques, which use either ozone oxidation or photo reaction, are useful in a variety of medical devices even of a complex design, such as membrane oxygenators or artificial hearts.

Evaluation of a new heparin bonding process in prolonged extracorporeal membrane oxygenation.

A novel heparin bonding method has been developed for in situ surface modification using ozone oxidation, and evaluated in vitro and in vivo during prolonged extracorporeal membrane oxygenation (ECMO) experiments. The ECMO system consisted of a Capiox hollow fiber membrane oxygenator (MO; Terumo Corp., Tokyo, Japan) with an integral heat exchanger, a Capiox centrifugal pump (CP), and an extracorporeal circuit. The blood contacting surfaces of the system were completely modified using the heparin bonding process, and evaluated in a chronic sheep model for extended period of time, ranging from 96 to 168 hr, under minimal systemic heparinization. The heparin bonded surface was able to maintain high levels of heparin bioactivity, and showed improved blood compatibility in in vitro epifluorescent video microscopy experiments by suppressing platelet adhesion/activation and complement activation. For the ECMO experiments, extracorporeal blood flow was maintained at 3 L/min and the activated clotting time was maintained at 150 sec. There was no significant change in the gas transfer capability or mean pressure drop across the MO over experimental times of up to 168 hr. Platelet count and other coagulation parameters remained stable within the physiologic range throughout the experiment. There were no detectable thrombi in the CP, tubing, or connectors in the test circuit. Local thrombus formation was noted in a stagnant area of the MO, although this did not interfere with its function or lead to any significant embolization. Based on these results, the heparin bonded ECMO system appears to be a safe and effective device for prolonged extracorporeal circulation under minimal systemic heparinization.

Nonthrombogenic polymer vascular prosthesis.

Although many synthetic vascular grafts have been developed and evaluated experimentally or clinically, none of them have met long-term patency when applied as a small diameter vascular substitute. We have recently developed a small caliber vascular graft (3 mm i.d.) using a nonthrombogenic polymer coating. The graft consists of three layered structures: Dacron for the outer layer, polyurethane in the middle layer, and a HEMA/styrene block copolymer (HEMA-st) coating for the inner layer. HEMA-st is an amphiphilic block copolymer composed of 2-hydroxyethyl methacrylate and styrene which has demonstrated improved blood compatibility over existing biomedical polymers in both in vitro and ex vivo experiments. Ten grafts were evaluated in a dog bilateral carotid replacement model. The grafts were electively retrieved at 7, 14, 30, 92, and 372 days after implantation. All grafts were patent without detectable thrombi along the graft length including anastomotic sites. Scanning electron micrographs of retrieved graft lumen showed fairly clean surfaces covered with a homogenous protein-like layer without microthrombi or endothelial cell lining. The thickness of the surface protein layer measured by a transmission electron microscopy was what can be described as monolayer protein adsorption regardless of implantation periods of as much as 372 days. A stable monolayer adsorbed protein layer formed on HEMA-st surfaces demonstrated nonthrombogenic activities in vivo and secure long-term patency of small caliber vascular grafts with the absence of an endothelial cell lining.

Tricuspid valve replacement with the bileaflet St. Jude Medical valve prosthesis.

Case histories of 39 patients who underwent tricuspid valve replacement with the St. Jude Medical prosthesis between June 1979 and August 1992 were reviewed in March 1993. The average patient age at the time of the operation was 46 +/- 11 years (range from 17 to 68 years). Concomitant mitral and/or aortic valve replacements were performed in 30 patients. All patients were given warfarin to maintain thrombotest between 10% to 25%. This number was between 2.8 to 1.6 times the control value in the International Normalized Ratio of prothrombin time. Three operative deaths occurred (7.7%). Among six late deaths, two patients died suddenly of unknown causes, and the remaining patient deaths were not valve-related. The actuarial survival rate at 14 years was 54.7%. Valve thrombosis occurred in one patient and was successfully treated with intravenous urokinase. This was the only valve-related complication (0.67%/patient-year). No reoperations were necessary in the tricuspid position. In conclusion, the St. Jude Medical valve is our choice of prosthesis for tricuspid valve replacement in adult patients who can receive proper anticoagulation therapy.