- Email: [email protected]
Alterations in coagulation after implantation of a pulsatile Novacor LVAD and the axial flow micromed DeBakey LVAD
Alterations in Coagulation After Implantation of a Pulsatile Novacor LVAD and the Axial Flow MicroMed DeBakey LVAD Andreas Koster, MD, Matthias Loebe, MD, PhD, Roland Hansen, MD, Evgenij V. Potapov, MD, George P. Noon, MD, Hermann Kuppe, MD, PhD, and Roland Hetzer, MD, PhD Department of Anesthesia, and Department of Cardiothoracic and Vascular Surgery, Deutsches Herzzentrum Berlin, Berlin, Germany
Background. The MicroMed DeBakey left ventricular assist device (LVAD) is a chamber and valveless axial flow blood pump. We investigated parameters of the coagulation system in patients after implantation of the axial flow LVAD and patients following implantation of a pulsatile Novacor LVAD. Methods. Six consecutive patients of both groups were investigated over a period of 6 weeks after implantation. -Thromboglobulin, platelet factor 4, factor XIIa, thrombin/antithrombin complexes, plasmin/␣2-antiplasmin complexes, and D-Dimer levels were measured. Results. With the exception of the plasmin/ ␣ 2 antiplasmin levels in the Novacor group, all parameters were elevated in both groups. The levels of -thrombo-
globulin, platelet factor 4, factor XIIa, and plasmin/␣2antiplasmin were significantly increased in the axial flow LVAD group. Conclusions. The axial flow LVAD strongly influences the systems of contact activation and fibrinolysis. The elevation of platelet proteins appears to follow platelet damage. Although no thromboembolic events were observed in both groups, elevation of thrombin/ antithrombin complexes provides convincing evidence of an increased activation of the coagulation system and the concomitant risk for the development of thromboembolism. (Ann Thorac Surg 2000;70:533–7) © 2000 by The Society of Thoracic Surgeons
I
pared with the other systems, the MicroMed DeBakey LVAD is smaller in size and lighter. This contributes to easier implantation when compared with the other systems. The device has been developed by the space agency NASA in collaboration with Baylor College of Medicine in Houston, TX. However, the implantation of all mechanical assist devices is followed by changes in the coagulation system [6 – 8]. In patients following implantation of a VAD, an increased bleeding tendency in the early postoperative period and thrombus formation with consequent embolism in the later postoperative period have been described [6, 7]. The aim of the present investigation was to asses the parameters of the coagulation system, including the procoagulant coagulation cascade, the contact activation system, the fibrinolytic system, and platelet function in patients after implantation of the DeBakey LVAD, and to compare this with observations in patients following implantation of the Novacor system.
mplantation of a left ventricular assist device (LVAD) may serve as a bridge to transplantation or for the recovery of the failing heart [1]. Most of the LVADs, such as the Novacor (Baxter, Oakland, CA) or the HeartMate system (Thermo Cardiosystems Inc, Woburn, MA) are based on the principle of pulsatile blood flow [2–5]. This pulsatile blood flow is achieved by magnetic or pneumatic compression of the blood in the chamber of the device. In contrast, the recently introduced MicroMed DeBakey LVAD (MicroMed Technology, The Woodlands, TX) works by the production of an axial blood flow generated by a turbine with approximately 7,500 to 12,500 r/min. This flow is generated by a rotating impeller, which sucks the blood from the left ventricular apex and pumps it into the ascending thoracic aorta. The pulsatile flow of the patient’s heart and the axial flow of the LVAD are in parallel, which leads to a reduction of the arterial pressure amplitude. This amplitude depends on the difference between the pressure of the device and the native left ventricle (contingent on its residual function). Similar to the Novacor or HeartMate system, an external portable control battery pack allows the patients mobility. Com-
Accepted for publication Feb 24, 2000. Address reprint requests to Dr Koster, Deutsches Herzzentrum Berlin, Augustenburgerplatz 1, D-13353 Berlin, Germany; e-mail: [email protected].
© 2000 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
Patients and Methods Patients Twelve patients, three women and nine men, were enrolled in the investigation. The age ranged from 32 to 63 years, with a mean of 47 ⫾ 7.7 SD. In 10 patients, the 0003-4975/00/$20.00 PII S0003-4975(00)01404-1
534
KOSTER ET AL HEMOSTASIS IN PULSATILE AND AXIAL FLOW LVAD
implantation of the device was necessary because of dilated cardiomyopathy and, in 2, because of ischemic cardiomyopathy. All patients received an isolated left ventricular assist device. Six patients received an axial flow MicroMed DeBakey LVAD and the other 6 patients received a pulsatile Novacor LVAD.
Devices The MicroMed DeBakey LVAD achieves axial blood flow by a valveless rotary blood pump with inflow from the left ventricular apex and an outflow graft into the ascending aorta. The pulsatile Novacor LVAD has a polyurethane pumping chamber and an inflow graft from the left ventricular apex and an outflow graft to the ascending aorta. Both grafts are made of Dacron (C.R. Bard, Haverhill, PA) and contain biological heart valves.
Anticoagulation In both groups, the same protocol of anticoagulation was applied. Before implantation of the LVAD, anticoagulation was performed with unfractionated heparin (Liquemin, Roche, Germany) to an activated thromboplastin time (aptt) of 40 to 60 seconds. Anticoagulation during surgery with cardiopulmonary bypass was accomplished with unfractionated heparin according to the Hepcon HMS regimen (Medtronic, Parker, CO). In all patients, during cardiopulmonary bypass, the kallikrein inhibitor aprotinin was given according to the high-dose regimen with a bolus of 2 ⫻ 106 kIU for the patient, 2 ⫻ 106 kIU for the priming solution, and a continuous infusion of 500,000 kIU/h. After termination of cardiopulmonary bypass, unfractionated heparin activity was totally reversed by protamine. The intravenous systemic heparinization was restarted 6 to 12 hours after surgery to achieve a target activated clotting time between 160 and 180 seconds.
Antiplatelet Therapy Antiplatelet agents were simultaneously administered according to the in vitro-induced platelet aggregation test using the method of Born. After removal of all drainage tubes, the anticoagulation was switched to coumarines and aspirin, and dipyridamole depending on the results of the platelet aggregation test. The test were performed with adenosine diphosphate 20 mol/L, epinephrine 100 l/L, collagen 190 g/mL, and arachidonic acid 500 g/mL. The target value for adenosine diphosphate and epinephrine-induced platelet aggregation was 30% to 50% and less than 40% for arachnidonic acid. A decrease of the aggregation to a value of 70% to 90% was regarded as sufficient for collagen-induced platelet aggregation. Antiplatelet therapy was initiated with aspirin beginning with 50 mg/d up to 200 mg/d as a maximum dosage. If patients did not respond to this therapy, dipyridamole (400 to 1,000 mg/d) was given.
Laboratory Investigation Blood samples were collected on the first and third postoperative day and, thereafter, twice a week for a
Ann Thorac Surg 2000;70:533–7
period of 2 months. Platelet factor 4 (PF4) and -thromboglobulin (-TG) were measured in citrate, theophylline, adenosine, dipyramidole (CTAD) plasma. Platelet factor 4 was measured by the use of the ASSERACHROM PF4 ELISA (Boehringer, Mannheim, Germany; range of reference 0 to 0.5 IU/mL). The -TG tests were performed by the use of the ASSERACHROM -TG ELISA (Boehringer; range of reference 10 to 40 IU/mL). Factor XIIa (FXIIa) D-Dimers, plasmin-␣2-antiplasmin complexes (PAP), and thrombin-antithrombin complexes (TAT) were measured in citrated blood samples. The FXIIa testing was performed with the FXIIa ELISA (PROGEN, Heidelberg, Germany; range of reference 18 to 25 g/mL) and the D-Dimers were evaluated by the use of the AUTO Dimertest (Organon Teknika, Eppelheim, Germany; range of reference between 130 and 200 g/L). The measurement of TAT complexes was performed with the ENZYGNOST TAT micro ELISA (Dade Behring, Marburg, Germany; range of reference 1.0 to 4.0 g/L) and the PAP complexes were measured by the ENZYGNOST PAP ELISA (Dade Behring; range of reference 120 to 700 g/L).
Statistical Analysis Statistical analysis was performed by the use of the Pearson’s correlation coefficient and the Student’s t tests. Differences between groups with a p value less than 0.01 were regarded to be significant.
Results Twelve patients were enrolled in the investigation. There was no significant difference in the preimplantation data between the 2 groups. Moreover, there were no significant differences in the postoperative course of the patients between the groups with regard to transfusion requirements, postoperative blood losses, duration of mechanical ventilation, parameters indicating infection (leukocyte count, C-Reactive Protein), or stay in the intensive care unit after implantation of the device. In all patients, anticoagulation followed the same protocol and there was no significant difference in the values of the plasmatic coagulation assays, platelet count, and platelet aggregation assays between the groups. In no patient was clinical evidence of thromboembolism obvious during the period of investigation. The results for each patient are expressed as the mean value for all 12 measurements of a single parameter over the period of 6 weeks. In the DeBakey group, the -TG value ranged from 218.3 IU/mL to 285.4 IU/mL with a mean of 258.2 IU/mL and a SD of ⫾138.7 (Fig 1). In the Novacor group, the -TG value ranged from 96.0 IU/mL to 224 IU/mL with a mean of 137.8 and a SD of ⫾77.9 IU/mL (Fig 1). The p value between the groups was 0.002. The PF4 value in the DeBakey group ranged from 41.9 IU/mL to 113.3 IU/mL with a mean of 74.1 IU/mL and a SD of ⫾79.0 IU/mL (Fig 1). The PF4 value in the Novacor group ranged from 17.0 IU/mL to 59.4 IU/mL with a mean
Ann Thorac Surg 2000;70:533–7
KOSTER ET AL HEMOSTASIS IN PULSATILE AND AXIAL FLOW LVAD
535
of 45.4 IU/mL and a SD of ⫾46.1 IU/mL (Fig 1). The p value between the groups was 0.001. In the DeBakey group, the factor XIIa ranged from 1.14 to 6.28 ng/mL with a mean of 3.76 ng/mL and a SD of ⫾2.26 ng/mL (Fig 2). In the Novacor group, the factor XIIa ranged from 1.81 ng/mL to 5.37 ng/mL with a mean of 2.84 ng/mL and a SD of ⫾1.51 ng/mL (Fig 2). The p value between the groups was 0.006. The TAT complexes in the DeBakey group revealed a value that ranged from 9.16 g/L to 14.8 g/L with a mean of 11.9 g/L and a SD of ⫾10.5 g/L (Fig 3). In the Novacor group, the values ranged from 8.33 g/mL to 16.5 g/mL with a mean of 12.0 g/L and a SD of ⫾4.49 g/L (Fig 3). The p value between the groups was 0.95. The PAP complexes in the DeBakey group ranged from 820.0 g/L to 1,092 g/L with a mean of 1,037.4 g/L and a SD of ⫾497.5 g/L (Fig 3). In the Novacor group, the PAP value ranged from 508.8 g/L to 865.0 g/L with a mean of 674.1 g/L and a SD of ⫾224 g/L (Fig 3). The p value between the groups was 0.001. The D-Dimer levels in the DeBakey group ranged from 439.0 g/mL to 1,552 g/mL with a mean of 11,270 g/mL and a SD of ⫾769.7 g/mL (Fig 3). The level in the
Fig. 2. Plasmin/anti-plasmin (PAP) complexes (A) and factor XIIa (B). Each patient is represented by a symbol. The line represents the median of the group. For PAP, the p value was p ⬍ 0.001 between the groups and, for factor XIIa, the p value was 0.006.
Novacor group ranged from 1,026.1 g/mL to 1,934 g/mL with a mean of 1,320.7 g/mL and a SD of ⫾544 g/mL (Fig 3). The p value between the groups was 0.66.
Comment
Fig. 1. Platelet factor 4 (PF4) (A) and -thromboglobulin (B). Each patient is represented by a symbol. The line represents the median of the group. For both parameters, the p value was p ⬍ 0.001 between the groups.
Changes in the coagulation system in 6 patients after implantation of the axial flow DeBakey LVAD and in 6 patients after implantation of the pulsatile Novacor LVAD were investigated during the first 6 postoperative weeks. The parameters included platelet factor 4 (PF4) and -thromboglobulin (-TG), which are markers for both platelet activation and platelet degradation. Moreover, the fibrinolytic system was observed by monitoring plasmin/␣2-antiplasmin (PAP) complexes and D-Dimers. The influences on the plasmatic coagulation system were monitored by the thrombin generation via measurement of the thrombin/antithrombin (TAT) complexes. Factor XIIa was monitored to investigate the influences on the contact activation system. The mean values for each group during the entire course of the investigation were elevated for all measured parameters except the PAP complexes in the Novacor group. When compared with
536
KOSTER ET AL HEMOSTASIS IN PULSATILE AND AXIAL FLOW LVAD
Fig. 3. D-Dimers (A) and thrombin-antithrombin (TAT) complexes (B). Each patient is represented by a symbol. The line represents the median of the group. For D-Dimers, the p value was 0.66 between the groups and, for TAT, the p value was 0.95.
the Novacor group, there was a significant increase in the levels of -TG, PF4, Factor XIIa, and the PAP complex in the DeBakey LVAD group. No significant differences between the groups was observed for the D-Dimer and TAT levels. Pulsatile LVADs have been successfully implanted over a period of several years as a bridge to transplantation. However, despite the development of biocompatible surfaces, thromboembolism still remains a problem. After the implantation of the Novacor system, a rather high percentage of thromboembolic complications, despite elaborate anticoagulation protocols, have been observed [9]. However, it remains unclear whether thromboembolic complications result from thrombus formation on the artificial surfaces, in particular in the compression chamber of the system itself, or to distinct flow patterns of the device at the site of the valves or to thrombus formation in the native ventricle [10]. Because totally implanted pulsatile devices need a compliance volume shifting chamber, these devices are large and heavy, which restricts the mobility of the patient and limits the option for permanent implantation [11]. Therefore, the development of a nonpulsatile per-
Ann Thorac Surg 2000;70:533–7
manent rotary blood pump was suggested [11]. However, certain features were necessary prerequisites for such a system, namely, small size, atraumatic blood flow, low thrombogenicity, a durable simple design, and a low energy requirement with easy controllability [11]. The MicroMed DeBakey LVAD was the first such pump to become available in November 1998. Because it does not need a compliance chamber, the rotary DeBakey LVAD is small in size and light. Moreover, certain characteristics of the device’s design are thought to contribute to a reduction of thrombogenicity, for example, because there is no compliance chamber, no residual volume remains in the device, the rapid rotation of the impeller and the fast blood flow may additionally inhibit thrombus formation in the device. Because of the axial flow system, it does not need a valved conduit, which should contribute to a reduction of thrombus formation in the device and blood trauma. Changes in platelet morphology have also been described with the use of the Novacor system [12] and reports have appeared that attribute the use of rotary or centrifugal blood pumps as the cause of platelet damage [13]. These findings are supported by our results in that the levels of PF4 and -TG, which are released from the platelet-alpha granula, were significantly elevated when compared with the values for the Novacor device (Fig 1). Because the platelet activation was comparable in both groups and the patients were treated with antiplatelet agents according to the results of aggregometry, these elevations of PF4 and -TG appear to be predominantly caused by cell damage and not by cell activation. Interestingly, in the DeBakey group, both FXIIa and PAP levels were significantly increased when compared with the Novacor group (Fig 2). These findings are of significance as PF4 inhibits FXII [14], and platelets are known to inhibit fibrinolysis [15]. Therefore, the activation of both systems may be higher as these results indicate. Because FXIIa plays a pivotal role in the initiation of fibrinolysis [16] in both the contact activation on foreign surfaces and the endothelium, the elevated FXIIa levels may be responsible for the initiation of fibrinolysis. However, there was no statistically significant difference in the other measured marker of the fibrinolytic system, the D-Dimers, although in both groups this parameter was elevated. Limitations in the assay, in particular in the higher range of the D-Dimer test, may be responsible for these findings. However, the clinical relevance of an increased fibrinolytic activity in patients on a VAD is unclear. Although fibrinolysis in the early postoperative period may contribute to adverse effects, such as bleeding [6], moderate hyperfibrinolysis in the later periods may contribute to a reduction of thromboembolic complications. There was no significant difference in the values for the TAT complexes in both groups. However, in both groups nearly fourfold elevated levels of TAT, as well as increased D-Dimers, indicate an excessive generation of thrombin (and fibrin). Therefore, although during the period of the investigation no thromboembolic events were observed, these elevated parameters provide evi-
537
Ann Thorac Surg 2000;70:533–7
KOSTER ET AL HEMOSTASIS IN PULSATILE AND AXIAL FLOW LVAD
dence of a strong activation of the plasmatic coagulation cascade and a possible concomitant increase of thrombogenic risk in both groups. We conclude that the implantation of the MicroMed DeBakey high-velocity axial flow LVAD, when compared with the established pulsatile Novacor LVAD, is associated with a significant elevation of platelet-granula proteins. This elevation seems to be caused by platelet damage by the device. Furthermore, systems of contact activation and fibrinolysis are significantly accelerated in these patients when compared with the Novacor group. The elevation of TAT complexes and D-Dimers strongly suggests an increased risk of thromboembolic events in both groups. These findings of an increased thrombin generation and fibrinolysis have also been reported in patients with the HeartMate system, which is thought to have improved biocompatibility of the inner artificial surface because of the generation of a pseudo-endothelium layer [17, 18]. However, reduced rates of thromboembolic events using this system have been reported [17]. Therefore, it can be speculated that the balance between the systems of thrombin generation and fibrinolysis is of pivotal importance for the reduction of thromboembolic complications [17]. Although as described previously, the MicroMed DeBakey high-velocity axial flow LVAD has characteristics that are thought to reduce thrombogenicity in the device, such as lack of a compression chamber, valves, and the fast impeller rotation, its potentially increased thrombogenicity appears to be from systemic effects, such as platelet trauma and activation of plasmatic coagulation systems. However, the balance between the coagulation systems, as a part of the inflammatory system, is not only influenced by the device itself but also by the clinical course of the patient. It may be that, during prolonged periods of implantation and ongoing recovery, the balance between the systems may be shifted toward coagulation as well as fibrinolysis. Therefore, further investigations with larger patient groups over longer periods of implantation are necessary to address this subject.
mechanical cardiac support in patients with idiopathic dilated cardiomyopathy. Circulation 1997;15:542–9. McCarthy PM, Smedira NO, Vargo RL, et al. One hundred patients with the HeartMate left ventricular assist device: evolving concepts and technology. J Thorac Cardiovasc Surg 1998;115:904 –12. Arabia FA, Smith RG, Rose DS, Arzouman DA, Sethi GK, Copeland JG. Success rates of long-term assist devices used currently for bridge to heart transplantation. ASAIO J 1996; 42:M542– 6. Koul B, Solem JO, Stehen S, Casimir-Ahn H, Granfeldt H, Lo¨nn UJ. HeartMate left ventricular assist device as bridge to heart transplantation. Ann Thorac Surg 1998;65:1625–30. Vetter HO, Kaulbach HG, Schmitz C, et al. Experience with the Novacor left ventricular assist system as a bridge to cardiac transplantation including the new wearable system. J Thorac Cardiovasc Surg 1995;109:74– 80. Livingston ER, Fisher CA, Bibidakis EJ, et al. Increased activation of the coagulation and fibrinolytic systems lead to hemorrhagic complications during left ventricular assist implantation. Circulation 1996;94(Suppl II):227–34. Himmelreich G, Ullmann H, Riess H, et al. Pathophysiologic role of contact activation in bleeding followed by thromboembolic complications after implantation of a ventricular assist device. ASAIO J 1995;41:M790 – 4. Loebe M, Gorman K, Burger R, Gage JE, Harke C, Hetzer R. Complement activation in patients undergoing mechanical circulatory support. ASAIO J 1998;M340 – 6. Schmid C, Weyand M, Nabavi DG, et al. Cerebral and systemic embolization during left ventricular support with the Novacor N100 device. Ann Thorac Surg 1998;65:1703–10. Bepu S, Tanaka N, Noda H, et al. High incidence of left ventricular thrombosis and systemic embolism in patients with left ventricular assist system. J Cardiol 1992;22:713–20. Nose Y, Kawahito K. Development of a non-pulsatile permanent rotary blood pump. Eur J Cardiothorac Surg 1997; 11(Suppl):32. Dewald O, Fischlein T, Vetter HO, et al. Platelet morphology in patients with mechanical circulatory support. Eur J Cardiothorac Surg 1997;12:634– 41. Kawahito K, Mohara J, Misawa Y, Fuse K. Platelet damage caused by the centrifugal pump: in vitro evaluation by measuring the release of alpha-granule packing proteins. Artif Organs 1997;21:1105–9. Dumenco LL, Everson B, Culp LA, Ratnoff OD. Inhibition of the activation of Hagemann factor (factor XII) by platelet factor 4. J Lab Clin Med 1988;112:394 – 400. Fay WP, Eitzmann DT, Shapiro AD, Madison EL, Ginsburg D. Platelets inhibit fibrinolysis in vitro by both plasminogen activator-1–inhibitor-dependent and –independent mechanisms. Blood 1994;83:351– 6. Lenich C, Pannell R, Gurewich V. Assembly and activation of the intrinsic fibrinolytic pathway on the surface of human endothelial cells in culture. Thromb Haemost 1995;74:698 – 703. Spanier T, Oz M, Levin H, et al. Activation of coagulation and fibrinolytic pathways in patients with left ventricular assist devices. J Thorac Cardiovasc Surg 1996;112:1090–7. Wang IW, Kottke-Merchant K, Vargo RL, McCarthy PM. Hemostatic profiles of HeartMate ventricular assist device recipients. ASAIO J 1195;41:M782–7.
I thank Miss Annette Gaussmann for the professional preparation of the figures and Miss Tonie Derwent for her editorial work on the text. Moreover, I thank Miss Cornelia Harke for the accurate performance of the laboratory work. I am also indebted to Prof Fritz Mertzlufft, Homburg Saar, for the helpful discussions during the preparation of the manuscript.
References 1. Mueller J, Wallukat G, Weng YG, et al. Weaning from
2.
3.
4. 5.
6.
7.
8. 9. 10. 11. 12. 13.
14. 15.
16.
17. 18.
Background. The MicroMed DeBakey left ventricular assist device (LVAD) is a chamber and valveless axial flow blood pump. We investigated parameters of the coagulation system in patients after implantation of the axial flow LVAD and patients following implantation of a pulsatile Novacor LVAD. Methods. Six consecutive patients of both groups were investigated over a period of 6 weeks after implantation. -Thromboglobulin, platelet factor 4, factor XIIa, thrombin/antithrombin complexes, plasmin/␣2-antiplasmin complexes, and D-Dimer levels were measured. Results. With the exception of the plasmin/ ␣ 2 antiplasmin levels in the Novacor group, all parameters were elevated in both groups. The levels of -thrombo-
globulin, platelet factor 4, factor XIIa, and plasmin/␣2antiplasmin were significantly increased in the axial flow LVAD group. Conclusions. The axial flow LVAD strongly influences the systems of contact activation and fibrinolysis. The elevation of platelet proteins appears to follow platelet damage. Although no thromboembolic events were observed in both groups, elevation of thrombin/ antithrombin complexes provides convincing evidence of an increased activation of the coagulation system and the concomitant risk for the development of thromboembolism. (Ann Thorac Surg 2000;70:533–7) © 2000 by The Society of Thoracic Surgeons
I
pared with the other systems, the MicroMed DeBakey LVAD is smaller in size and lighter. This contributes to easier implantation when compared with the other systems. The device has been developed by the space agency NASA in collaboration with Baylor College of Medicine in Houston, TX. However, the implantation of all mechanical assist devices is followed by changes in the coagulation system [6 – 8]. In patients following implantation of a VAD, an increased bleeding tendency in the early postoperative period and thrombus formation with consequent embolism in the later postoperative period have been described [6, 7]. The aim of the present investigation was to asses the parameters of the coagulation system, including the procoagulant coagulation cascade, the contact activation system, the fibrinolytic system, and platelet function in patients after implantation of the DeBakey LVAD, and to compare this with observations in patients following implantation of the Novacor system.
mplantation of a left ventricular assist device (LVAD) may serve as a bridge to transplantation or for the recovery of the failing heart [1]. Most of the LVADs, such as the Novacor (Baxter, Oakland, CA) or the HeartMate system (Thermo Cardiosystems Inc, Woburn, MA) are based on the principle of pulsatile blood flow [2–5]. This pulsatile blood flow is achieved by magnetic or pneumatic compression of the blood in the chamber of the device. In contrast, the recently introduced MicroMed DeBakey LVAD (MicroMed Technology, The Woodlands, TX) works by the production of an axial blood flow generated by a turbine with approximately 7,500 to 12,500 r/min. This flow is generated by a rotating impeller, which sucks the blood from the left ventricular apex and pumps it into the ascending thoracic aorta. The pulsatile flow of the patient’s heart and the axial flow of the LVAD are in parallel, which leads to a reduction of the arterial pressure amplitude. This amplitude depends on the difference between the pressure of the device and the native left ventricle (contingent on its residual function). Similar to the Novacor or HeartMate system, an external portable control battery pack allows the patients mobility. Com-
Accepted for publication Feb 24, 2000. Address reprint requests to Dr Koster, Deutsches Herzzentrum Berlin, Augustenburgerplatz 1, D-13353 Berlin, Germany; e-mail: [email protected].
© 2000 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
Patients and Methods Patients Twelve patients, three women and nine men, were enrolled in the investigation. The age ranged from 32 to 63 years, with a mean of 47 ⫾ 7.7 SD. In 10 patients, the 0003-4975/00/$20.00 PII S0003-4975(00)01404-1
534
KOSTER ET AL HEMOSTASIS IN PULSATILE AND AXIAL FLOW LVAD
implantation of the device was necessary because of dilated cardiomyopathy and, in 2, because of ischemic cardiomyopathy. All patients received an isolated left ventricular assist device. Six patients received an axial flow MicroMed DeBakey LVAD and the other 6 patients received a pulsatile Novacor LVAD.
Devices The MicroMed DeBakey LVAD achieves axial blood flow by a valveless rotary blood pump with inflow from the left ventricular apex and an outflow graft into the ascending aorta. The pulsatile Novacor LVAD has a polyurethane pumping chamber and an inflow graft from the left ventricular apex and an outflow graft to the ascending aorta. Both grafts are made of Dacron (C.R. Bard, Haverhill, PA) and contain biological heart valves.
Anticoagulation In both groups, the same protocol of anticoagulation was applied. Before implantation of the LVAD, anticoagulation was performed with unfractionated heparin (Liquemin, Roche, Germany) to an activated thromboplastin time (aptt) of 40 to 60 seconds. Anticoagulation during surgery with cardiopulmonary bypass was accomplished with unfractionated heparin according to the Hepcon HMS regimen (Medtronic, Parker, CO). In all patients, during cardiopulmonary bypass, the kallikrein inhibitor aprotinin was given according to the high-dose regimen with a bolus of 2 ⫻ 106 kIU for the patient, 2 ⫻ 106 kIU for the priming solution, and a continuous infusion of 500,000 kIU/h. After termination of cardiopulmonary bypass, unfractionated heparin activity was totally reversed by protamine. The intravenous systemic heparinization was restarted 6 to 12 hours after surgery to achieve a target activated clotting time between 160 and 180 seconds.
Antiplatelet Therapy Antiplatelet agents were simultaneously administered according to the in vitro-induced platelet aggregation test using the method of Born. After removal of all drainage tubes, the anticoagulation was switched to coumarines and aspirin, and dipyridamole depending on the results of the platelet aggregation test. The test were performed with adenosine diphosphate 20 mol/L, epinephrine 100 l/L, collagen 190 g/mL, and arachidonic acid 500 g/mL. The target value for adenosine diphosphate and epinephrine-induced platelet aggregation was 30% to 50% and less than 40% for arachnidonic acid. A decrease of the aggregation to a value of 70% to 90% was regarded as sufficient for collagen-induced platelet aggregation. Antiplatelet therapy was initiated with aspirin beginning with 50 mg/d up to 200 mg/d as a maximum dosage. If patients did not respond to this therapy, dipyridamole (400 to 1,000 mg/d) was given.
Laboratory Investigation Blood samples were collected on the first and third postoperative day and, thereafter, twice a week for a
Ann Thorac Surg 2000;70:533–7
period of 2 months. Platelet factor 4 (PF4) and -thromboglobulin (-TG) were measured in citrate, theophylline, adenosine, dipyramidole (CTAD) plasma. Platelet factor 4 was measured by the use of the ASSERACHROM PF4 ELISA (Boehringer, Mannheim, Germany; range of reference 0 to 0.5 IU/mL). The -TG tests were performed by the use of the ASSERACHROM -TG ELISA (Boehringer; range of reference 10 to 40 IU/mL). Factor XIIa (FXIIa) D-Dimers, plasmin-␣2-antiplasmin complexes (PAP), and thrombin-antithrombin complexes (TAT) were measured in citrated blood samples. The FXIIa testing was performed with the FXIIa ELISA (PROGEN, Heidelberg, Germany; range of reference 18 to 25 g/mL) and the D-Dimers were evaluated by the use of the AUTO Dimertest (Organon Teknika, Eppelheim, Germany; range of reference between 130 and 200 g/L). The measurement of TAT complexes was performed with the ENZYGNOST TAT micro ELISA (Dade Behring, Marburg, Germany; range of reference 1.0 to 4.0 g/L) and the PAP complexes were measured by the ENZYGNOST PAP ELISA (Dade Behring; range of reference 120 to 700 g/L).
Statistical Analysis Statistical analysis was performed by the use of the Pearson’s correlation coefficient and the Student’s t tests. Differences between groups with a p value less than 0.01 were regarded to be significant.
Results Twelve patients were enrolled in the investigation. There was no significant difference in the preimplantation data between the 2 groups. Moreover, there were no significant differences in the postoperative course of the patients between the groups with regard to transfusion requirements, postoperative blood losses, duration of mechanical ventilation, parameters indicating infection (leukocyte count, C-Reactive Protein), or stay in the intensive care unit after implantation of the device. In all patients, anticoagulation followed the same protocol and there was no significant difference in the values of the plasmatic coagulation assays, platelet count, and platelet aggregation assays between the groups. In no patient was clinical evidence of thromboembolism obvious during the period of investigation. The results for each patient are expressed as the mean value for all 12 measurements of a single parameter over the period of 6 weeks. In the DeBakey group, the -TG value ranged from 218.3 IU/mL to 285.4 IU/mL with a mean of 258.2 IU/mL and a SD of ⫾138.7 (Fig 1). In the Novacor group, the -TG value ranged from 96.0 IU/mL to 224 IU/mL with a mean of 137.8 and a SD of ⫾77.9 IU/mL (Fig 1). The p value between the groups was 0.002. The PF4 value in the DeBakey group ranged from 41.9 IU/mL to 113.3 IU/mL with a mean of 74.1 IU/mL and a SD of ⫾79.0 IU/mL (Fig 1). The PF4 value in the Novacor group ranged from 17.0 IU/mL to 59.4 IU/mL with a mean
Ann Thorac Surg 2000;70:533–7
KOSTER ET AL HEMOSTASIS IN PULSATILE AND AXIAL FLOW LVAD
535
of 45.4 IU/mL and a SD of ⫾46.1 IU/mL (Fig 1). The p value between the groups was 0.001. In the DeBakey group, the factor XIIa ranged from 1.14 to 6.28 ng/mL with a mean of 3.76 ng/mL and a SD of ⫾2.26 ng/mL (Fig 2). In the Novacor group, the factor XIIa ranged from 1.81 ng/mL to 5.37 ng/mL with a mean of 2.84 ng/mL and a SD of ⫾1.51 ng/mL (Fig 2). The p value between the groups was 0.006. The TAT complexes in the DeBakey group revealed a value that ranged from 9.16 g/L to 14.8 g/L with a mean of 11.9 g/L and a SD of ⫾10.5 g/L (Fig 3). In the Novacor group, the values ranged from 8.33 g/mL to 16.5 g/mL with a mean of 12.0 g/L and a SD of ⫾4.49 g/L (Fig 3). The p value between the groups was 0.95. The PAP complexes in the DeBakey group ranged from 820.0 g/L to 1,092 g/L with a mean of 1,037.4 g/L and a SD of ⫾497.5 g/L (Fig 3). In the Novacor group, the PAP value ranged from 508.8 g/L to 865.0 g/L with a mean of 674.1 g/L and a SD of ⫾224 g/L (Fig 3). The p value between the groups was 0.001. The D-Dimer levels in the DeBakey group ranged from 439.0 g/mL to 1,552 g/mL with a mean of 11,270 g/mL and a SD of ⫾769.7 g/mL (Fig 3). The level in the
Fig. 2. Plasmin/anti-plasmin (PAP) complexes (A) and factor XIIa (B). Each patient is represented by a symbol. The line represents the median of the group. For PAP, the p value was p ⬍ 0.001 between the groups and, for factor XIIa, the p value was 0.006.
Novacor group ranged from 1,026.1 g/mL to 1,934 g/mL with a mean of 1,320.7 g/mL and a SD of ⫾544 g/mL (Fig 3). The p value between the groups was 0.66.
Comment
Fig. 1. Platelet factor 4 (PF4) (A) and -thromboglobulin (B). Each patient is represented by a symbol. The line represents the median of the group. For both parameters, the p value was p ⬍ 0.001 between the groups.
Changes in the coagulation system in 6 patients after implantation of the axial flow DeBakey LVAD and in 6 patients after implantation of the pulsatile Novacor LVAD were investigated during the first 6 postoperative weeks. The parameters included platelet factor 4 (PF4) and -thromboglobulin (-TG), which are markers for both platelet activation and platelet degradation. Moreover, the fibrinolytic system was observed by monitoring plasmin/␣2-antiplasmin (PAP) complexes and D-Dimers. The influences on the plasmatic coagulation system were monitored by the thrombin generation via measurement of the thrombin/antithrombin (TAT) complexes. Factor XIIa was monitored to investigate the influences on the contact activation system. The mean values for each group during the entire course of the investigation were elevated for all measured parameters except the PAP complexes in the Novacor group. When compared with
536
KOSTER ET AL HEMOSTASIS IN PULSATILE AND AXIAL FLOW LVAD
Fig. 3. D-Dimers (A) and thrombin-antithrombin (TAT) complexes (B). Each patient is represented by a symbol. The line represents the median of the group. For D-Dimers, the p value was 0.66 between the groups and, for TAT, the p value was 0.95.
the Novacor group, there was a significant increase in the levels of -TG, PF4, Factor XIIa, and the PAP complex in the DeBakey LVAD group. No significant differences between the groups was observed for the D-Dimer and TAT levels. Pulsatile LVADs have been successfully implanted over a period of several years as a bridge to transplantation. However, despite the development of biocompatible surfaces, thromboembolism still remains a problem. After the implantation of the Novacor system, a rather high percentage of thromboembolic complications, despite elaborate anticoagulation protocols, have been observed [9]. However, it remains unclear whether thromboembolic complications result from thrombus formation on the artificial surfaces, in particular in the compression chamber of the system itself, or to distinct flow patterns of the device at the site of the valves or to thrombus formation in the native ventricle [10]. Because totally implanted pulsatile devices need a compliance volume shifting chamber, these devices are large and heavy, which restricts the mobility of the patient and limits the option for permanent implantation [11]. Therefore, the development of a nonpulsatile per-
Ann Thorac Surg 2000;70:533–7
manent rotary blood pump was suggested [11]. However, certain features were necessary prerequisites for such a system, namely, small size, atraumatic blood flow, low thrombogenicity, a durable simple design, and a low energy requirement with easy controllability [11]. The MicroMed DeBakey LVAD was the first such pump to become available in November 1998. Because it does not need a compliance chamber, the rotary DeBakey LVAD is small in size and light. Moreover, certain characteristics of the device’s design are thought to contribute to a reduction of thrombogenicity, for example, because there is no compliance chamber, no residual volume remains in the device, the rapid rotation of the impeller and the fast blood flow may additionally inhibit thrombus formation in the device. Because of the axial flow system, it does not need a valved conduit, which should contribute to a reduction of thrombus formation in the device and blood trauma. Changes in platelet morphology have also been described with the use of the Novacor system [12] and reports have appeared that attribute the use of rotary or centrifugal blood pumps as the cause of platelet damage [13]. These findings are supported by our results in that the levels of PF4 and -TG, which are released from the platelet-alpha granula, were significantly elevated when compared with the values for the Novacor device (Fig 1). Because the platelet activation was comparable in both groups and the patients were treated with antiplatelet agents according to the results of aggregometry, these elevations of PF4 and -TG appear to be predominantly caused by cell damage and not by cell activation. Interestingly, in the DeBakey group, both FXIIa and PAP levels were significantly increased when compared with the Novacor group (Fig 2). These findings are of significance as PF4 inhibits FXII [14], and platelets are known to inhibit fibrinolysis [15]. Therefore, the activation of both systems may be higher as these results indicate. Because FXIIa plays a pivotal role in the initiation of fibrinolysis [16] in both the contact activation on foreign surfaces and the endothelium, the elevated FXIIa levels may be responsible for the initiation of fibrinolysis. However, there was no statistically significant difference in the other measured marker of the fibrinolytic system, the D-Dimers, although in both groups this parameter was elevated. Limitations in the assay, in particular in the higher range of the D-Dimer test, may be responsible for these findings. However, the clinical relevance of an increased fibrinolytic activity in patients on a VAD is unclear. Although fibrinolysis in the early postoperative period may contribute to adverse effects, such as bleeding [6], moderate hyperfibrinolysis in the later periods may contribute to a reduction of thromboembolic complications. There was no significant difference in the values for the TAT complexes in both groups. However, in both groups nearly fourfold elevated levels of TAT, as well as increased D-Dimers, indicate an excessive generation of thrombin (and fibrin). Therefore, although during the period of the investigation no thromboembolic events were observed, these elevated parameters provide evi-
537
Ann Thorac Surg 2000;70:533–7
KOSTER ET AL HEMOSTASIS IN PULSATILE AND AXIAL FLOW LVAD
dence of a strong activation of the plasmatic coagulation cascade and a possible concomitant increase of thrombogenic risk in both groups. We conclude that the implantation of the MicroMed DeBakey high-velocity axial flow LVAD, when compared with the established pulsatile Novacor LVAD, is associated with a significant elevation of platelet-granula proteins. This elevation seems to be caused by platelet damage by the device. Furthermore, systems of contact activation and fibrinolysis are significantly accelerated in these patients when compared with the Novacor group. The elevation of TAT complexes and D-Dimers strongly suggests an increased risk of thromboembolic events in both groups. These findings of an increased thrombin generation and fibrinolysis have also been reported in patients with the HeartMate system, which is thought to have improved biocompatibility of the inner artificial surface because of the generation of a pseudo-endothelium layer [17, 18]. However, reduced rates of thromboembolic events using this system have been reported [17]. Therefore, it can be speculated that the balance between the systems of thrombin generation and fibrinolysis is of pivotal importance for the reduction of thromboembolic complications [17]. Although as described previously, the MicroMed DeBakey high-velocity axial flow LVAD has characteristics that are thought to reduce thrombogenicity in the device, such as lack of a compression chamber, valves, and the fast impeller rotation, its potentially increased thrombogenicity appears to be from systemic effects, such as platelet trauma and activation of plasmatic coagulation systems. However, the balance between the coagulation systems, as a part of the inflammatory system, is not only influenced by the device itself but also by the clinical course of the patient. It may be that, during prolonged periods of implantation and ongoing recovery, the balance between the systems may be shifted toward coagulation as well as fibrinolysis. Therefore, further investigations with larger patient groups over longer periods of implantation are necessary to address this subject.
mechanical cardiac support in patients with idiopathic dilated cardiomyopathy. Circulation 1997;15:542–9. McCarthy PM, Smedira NO, Vargo RL, et al. One hundred patients with the HeartMate left ventricular assist device: evolving concepts and technology. J Thorac Cardiovasc Surg 1998;115:904 –12. Arabia FA, Smith RG, Rose DS, Arzouman DA, Sethi GK, Copeland JG. Success rates of long-term assist devices used currently for bridge to heart transplantation. ASAIO J 1996; 42:M542– 6. Koul B, Solem JO, Stehen S, Casimir-Ahn H, Granfeldt H, Lo¨nn UJ. HeartMate left ventricular assist device as bridge to heart transplantation. Ann Thorac Surg 1998;65:1625–30. Vetter HO, Kaulbach HG, Schmitz C, et al. Experience with the Novacor left ventricular assist system as a bridge to cardiac transplantation including the new wearable system. J Thorac Cardiovasc Surg 1995;109:74– 80. Livingston ER, Fisher CA, Bibidakis EJ, et al. Increased activation of the coagulation and fibrinolytic systems lead to hemorrhagic complications during left ventricular assist implantation. Circulation 1996;94(Suppl II):227–34. Himmelreich G, Ullmann H, Riess H, et al. Pathophysiologic role of contact activation in bleeding followed by thromboembolic complications after implantation of a ventricular assist device. ASAIO J 1995;41:M790 – 4. Loebe M, Gorman K, Burger R, Gage JE, Harke C, Hetzer R. Complement activation in patients undergoing mechanical circulatory support. ASAIO J 1998;M340 – 6. Schmid C, Weyand M, Nabavi DG, et al. Cerebral and systemic embolization during left ventricular support with the Novacor N100 device. Ann Thorac Surg 1998;65:1703–10. Bepu S, Tanaka N, Noda H, et al. High incidence of left ventricular thrombosis and systemic embolism in patients with left ventricular assist system. J Cardiol 1992;22:713–20. Nose Y, Kawahito K. Development of a non-pulsatile permanent rotary blood pump. Eur J Cardiothorac Surg 1997; 11(Suppl):32. Dewald O, Fischlein T, Vetter HO, et al. Platelet morphology in patients with mechanical circulatory support. Eur J Cardiothorac Surg 1997;12:634– 41. Kawahito K, Mohara J, Misawa Y, Fuse K. Platelet damage caused by the centrifugal pump: in vitro evaluation by measuring the release of alpha-granule packing proteins. Artif Organs 1997;21:1105–9. Dumenco LL, Everson B, Culp LA, Ratnoff OD. Inhibition of the activation of Hagemann factor (factor XII) by platelet factor 4. J Lab Clin Med 1988;112:394 – 400. Fay WP, Eitzmann DT, Shapiro AD, Madison EL, Ginsburg D. Platelets inhibit fibrinolysis in vitro by both plasminogen activator-1–inhibitor-dependent and –independent mechanisms. Blood 1994;83:351– 6. Lenich C, Pannell R, Gurewich V. Assembly and activation of the intrinsic fibrinolytic pathway on the surface of human endothelial cells in culture. Thromb Haemost 1995;74:698 – 703. Spanier T, Oz M, Levin H, et al. Activation of coagulation and fibrinolytic pathways in patients with left ventricular assist devices. J Thorac Cardiovasc Surg 1996;112:1090–7. Wang IW, Kottke-Merchant K, Vargo RL, McCarthy PM. Hemostatic profiles of HeartMate ventricular assist device recipients. ASAIO J 1195;41:M782–7.
I thank Miss Annette Gaussmann for the professional preparation of the figures and Miss Tonie Derwent for her editorial work on the text. Moreover, I thank Miss Cornelia Harke for the accurate performance of the laboratory work. I am also indebted to Prof Fritz Mertzlufft, Homburg Saar, for the helpful discussions during the preparation of the manuscript.
References 1. Mueller J, Wallukat G, Weng YG, et al. Weaning from
2.
3.
4. 5.
6.
7.
8. 9. 10. 11. 12. 13.
14. 15.
16.
17. 18.