Author + information
- Received November 11, 2008
- Revision received February 6, 2009
- Accepted February 10, 2009
- Published online June 9, 2009.
- Jolanta Klovaite, MD⁎,
- Finn Gustafsson, MD, PhD†,
- Svend A. Mortensen, MD, DMSc†,
- Kåre Sander, MD, DMSc‡ and
- Lars B. Nielsen, MD, PhD, DMSc⁎,§,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Lars B. Nielsen, Department of Clinical Biochemistry, KB3011, Rigshospitalet, University of Copenhagen, Blegdamsvej 9, DK-2100 Copenhagen, Denmark
Objectives This study investigated the influence of the mechanical blood pump HeartMate II (HMII) (Thoratec Corporation, Pleasanton, California) on blood coagulation and platelet function.
Background HMII is an implantable left ventricular assist device used for the treatment of heart failure. Patients treated with HMII have increased bleeding tendencies.
Methods We measured agonist-induced platelet aggregation in 16 patients on HMII support.
Results The von Willebrand factor (vWF)-dependent ristocetin-induced platelet aggregation was impaired in 11 of the 16 patients, of which 12 had experienced at least 1 minor or major bleeding episode. The impaired ristocetin-induced platelet aggregation was associated both with decreased specific activity of plasma vWF, presumably due to lack of high molecular weight vWF multimers, as well as with attenuated function of the platelets themselves.
Conclusions The results imply that HMII treatment is associated with impaired platelet aggregation, which may contribute to an increased tendency to bleed.
Left ventricular assist devices (LVADs) are mechanical blood pumps used for the treatment of severe heart failure (e.g., as a bridge to heart transplantation, myocardial recovery, or as long-term destination therapy). HeartMate II (HMII) (Thoratec Corporation, Pleasanton, California) is among the most commonly used LVAD systems (1,2).
Exposure of the blood to foreign surfaces may activate the coagulation system. Hence, systemic anticoagulation with vitamin K and platelet inhibitors are recommended in HMII-treated patients (2). However, an increased tendency to bleed above what is normally observed in patients receiving anticoagulation treatment has been observed at both early and late stages after HMII implantation (3).
The HMII contains an axial-flow pump that draws blood from the left ventricle and continuously pumps it into the ascending aorta, exposing the blood to high shear stress (4). High shear stress increases the susceptibility to proteolysis of the largest von Willebrand factor (vWF) multimers (5,6), decreasing the ability of vWF to induce platelet aggregation. Hence, exposure of the blood to high shear stress might impose an acquired form of von Willebrand disease with qualitative vWF defects and increased bleeding tendencies. Such a pathogenic mechanism may explain the tendency to bleed of patients with aortic stenosis (7–9) or defects of cardiac septae (10). vWF-dependent platelet aggregation involves the binding of vWF to the GPIb-IX-V glycoprotein receptor complex on platelets. Exposure of the blood to shear stress (e.g., during hemodialysis) also results in defective GPIb-IX-V receptor function (11,12), which increases bleeding tendency. The aim of this study was to investigate the putative impact of HMII treatment on blood coagulation and specifically to test the hypothesis that HMII may impair vWF-dependent platelet aggregation.
We studied 12 men and 4 women, 19 to 63 years of age, in whom a HMII had been implanted 17 to 459 days (mean 151 days) previously. The indication for HMII treatment was end-stage dilated (n = 10) or ischemic (n = 6) cardiomyopathy. After HMII implantation, low molecular weight heparin and warfarin were initiated when post-operative bleeding was acceptable, typically on day 2 after surgery. In patients with underlying ischemic heart disease, aspirin was added, if tolerated. In case of manifest thromboembolic events, further platelet inhibition with clopidogrel was initiated. The international normalized ratio (INR) was monitored closely, and if it decreased below 1.8, low molecular weight heparin was initiated until the INR was >2. At the time of the present study, all patients received warfarin, 5 received aspirin, and 1 of these patients also received clopidogrel. Five patients underwent heart transplantation (HTx) after 166 to 392 days (mean 295 days) of treatment with HMII.
We also investigated hospitalized patients with congestive heart failure (CHF) (ejection fraction ≤20%; New York Heart Association functional class III to IV) due to dilated (n = 4) or ischemic (n = 5) cardiomyopathy and 6 outpatients who were receiving warfarin therapy because of previous venous thrombosis (INR 2.4 to 4.1). Seven of the patients with CHF received aspirin, with 1 of these patients also receiving clopidogrel and 1 warfarin. Clinical data were obtained from patient medical records. Control samples were obtained from healthy staff members at the Department of Clinical Biochemistry, Rigshospitalet, Copenhagen.
Blood was anticoagulated with 3.8% sodium citrate (Vacuette, Greiner Bio-One GmbH, Kremsmünster, Austria). Platelet-rich plasma (PRP) was prepared by centrifugation at 200 gfor 10 min at 21 to 23°C. The remaining blood was recentrifuged at 2,400 gfor 15 min to obtain platelet-poor plasma (PPP). Plasma for biochemical analysis was analyzed immediately, or stored at −80°C within 2 h after blood collection.
Laboratory analyses were conducted at the Department of Clinical Biochemistry, Rigshospitalet. Plasma von Willebrand factor antigen (vWF:Ag) was measured with sandwich enzyme-linked immunosorbent assay by the use of antibodies from DAKO-Cytomation, SA (Trappes, France). The von Willebrand factor ristocetin-cofactor activity (vWF:RCo) was measured with reagents and the BCT instrument from Dade Behring (Marburg, Germany). Multimeric patterns of vWF were determined by sodium dodecyl sulfate-agarose gel electrophoresis and western blotting at the Department of Clinical Biochemistry, Malmö University Hospital, Malmö, Sweden. Scanned images were analyzed with ImageJ (National Institutes of Health, Bethesda, Maryland). We measured a disintegrin and metalloproteinase with a thrombospondin type 1 motif, 13 (ADAMTS-13) activity with reagents from Pepta Nova (Sandhausen, Germany). Glycocalicin was measured with an enzyme-linked immunosorbent assay (Glycocalicin EIA kit, Takara Biomedical, Ohtsu, Japan).
PRP was mixed with autologous PPP to obtain a final platelet count of ∼250 × 109/l. Platelet aggregation after addition of agonists was measured at 37°C in a Whole Blood Lumi-Aggregometer 560 CA (Chrono-log Corporation, Havertown, Maryland) after the addition of adenosine diphosphate (ADP) (5 μmol/l), collagen (2 μg/ml), arachidonic acid (1 mmol/l), or ristocetin (1.25 mg/ml) (Triolab, Copenhagen, Denmark). We measured the release of adenosine triphosphate with CHRONO-LUME reagent (Triolab). Aggregation was expressed as percent of the maximal aggregation obtained 8 min after the addition of agonists. Patient samples were always analyzed in parallel with platelets from a healthy control subject. When ristocetin-induced platelet aggregation was impaired, patient PRP was mixed with normal PPP (1 + 3), and normal PRP was mixed with patient PPP (1 + 3) before the addition of ristocetin (1.25 mg/ml). Final platelet counts in mixtures were 109 ± 31 × 109/l and 122 ± 22 × 109/l when patient platelets and normal platelets were used, respectively.
The results are mean ± SD unless otherwise indicated. Statistical analyses were conducted with Prism software (version 2.0, Graph Pad, San Diego, California). Two-group comparisons were made with the Student ttest.
Of 16 HMII-treated patients, 12 had major or minor bleeding (Table 1).None had any history of increased bleeding tendency before HMII implantation. Two patients had suspected thromboembolic events (Table 1). Two other patients exhibited temporary bilateral loss of vision and lateralized motor deficits in the post-operative phase, respectively. In both cases, cerebral computed scans were normal and did not reveal cerebral infarction (or bleeding) (Table 1).
Changes in blood coagulation parameters could not explain the increased bleeding tendency (Table 2).However, ristocetin-induced platelet aggregation was severely impaired in patients with a bleeding history (Table 1, Fig. 1A).We found that ADP-induced platelet aggregation was moderately reduced in 7 patients, of whom 2 received aspirin (Fig. 1A). Collagen- and arachidonic acid-induced platelet aggregation (and release of adenosine triphosphate) was only abnormal in 1 patient who received both aspirin and clopidogrel (Fig. 1A).
To differentiate between plasma and platelet defects in patients with impaired ristocetin-induced platelet aggregation, we mixed platelets from the patients treated with HMII with normal plasma (or vice versa) before studying platelet aggregation. Seven of the 11 patients displayed signs of impaired ability of both plasma and platelet factors to mediate ristocetin-induced platelet aggregation (Fig. 1B). Two patients displayed normal ristocetin-induced platelet aggregation when their platelets were mixed with normal plasma, and 2 patients had normal platelet aggregation when their plasma was mixed with normal platelets (Fig. 1B).
Lack of high-molecular weight vWF multimers in HMII-treated patients
The average plasma concentrations of vWF protein (vWF:Ag) and activity (vWF:RCo) were 1.6 ± 0.7 kIU/l and 1.0 ± 0.6 kIU/l, respectively. Hence, although vWF:Ag was elevated, as previously observed in patients with CHF (13), 11 of 16 patients treated with HMII had decreased specific activity of plasma vWF (vWF:RCo/vWF:Ag) (Fig. 2A),and vWF specific-activity was lowest in patients with a history of severe bleeding (Fig. 2A). The patients treated with HMII had fewer large vWF multimers than healthy control patients (Figs. 2B and 2C). This finding was not due to increased plasma ADAMTS-13 levels, which were normal (1.2 ± 0.4 AU/l; reference interval: 0.75 to 1.40 AU/l) in the HMII-treated patients.
Normalized vWF-dependent platelet aggregation after HTx
Ristocetin-induced platelet aggregation (Fig. 3A),plasma vWF-specific activity (Fig. 3B), and vWF multimeric distribution (Figs. 2C and 3C) were all normalized after HTx. Also, patients with CHF and patients receiving warfarin because of a previous venous thrombosis had normal ristocetin-induced platelet aggregation (Fig. 3A) and plasma vWF-specific activity (Fig. 3B). Patients with CHF also had normal vWF multimeric distribution (Fig. 2C).
The present results imply that HMII treatment is associated with impaired platelet aggregation, which may contribute to an increased tendency of bleeding in patients. The most striking finding was that 11 of 16 HMII-treated patients had impaired ristocetin-induced platelet aggregation and that 9 of the 10 patients with severely impaired ristocetin-induced platelet aggregation had a history of minor or major bleedings. The impaired platelet aggregation was most likely due to the HMII treatment because each of the 5 patients examined had normalized platelet aggregation after HTx. We examined whether the defect ristocetin-induced platelet aggregation was due to plasma and/or platelet factors. The results suggest that both platelet and plasma factors and often a combination of the 2 contributed toward the observed abnormality.
The specific activity of vWF was decreased in HMII-treated patients. This finding was probably at least in part due to the lack of HMW vWF multimers. As such, HMII appears to induce a secondary VWD type 2 similar to what has previously been observed in patients with aortic stenosis (7–9) and other cardiac defects (10). This finding is in agreement with those recently reported by Geisen et al. (14). Although the specific activity of plasma vWF was decreased in HMII-treated patients, the vWF activity was still within the normal range. Therefore, the lack of HMW multimers may not fully account for the plasma factor leading to impaired ristocetin-induced platelet aggregation. Hence, it is conceivable that the plasma from HMII-treated patients might contain 1 or more factors that actually inhibit the ristocetin-induced platelet aggregation. One such putative factor could be fragments of vWF generated during the loss of HMW multimers (15,16).
The present findings emphasize a novel aspect of HMII treatment. Thus, it seems important to uncover how impaired platelet function affects the risk of bleeding in HMII-treated patients. The results, albeit based on a rather small number of patients, suggest that the risk of bleeding in HMII-treated patients due to impaired platelet aggregation may be significant, and the authors of future studies should also address whether patients on LVADs with abnormal platelet aggregation would be better off without routine antiplatelet therapy, especially those in whom ristocetin- and/or ADP-induced platelet aggregation is impaired.
The authors thank Marianne Johnson for expert technical assistance.
- Abbreviations and Acronyms
- a disintegrin and metalloproteinase with a thrombospondin type 1 motif, 13
- adenosine diphosphate
- congestive heart failure
- HeartMate II
- heart transplantation
- international normalized ratio
- left ventricular assist device
- platelet-poor plasma
- platelet-rich plasma
- von Willebrand factor
- von Willebrand factor antigen
- von Willebrand factor ristocetin-cofactor activity
- Received November 11, 2008.
- Revision received February 6, 2009.
- Accepted February 10, 2009.
- American College of Cardiology Foundation
- Snyder T.A.
- Tsai H.M.,
- Sussman I.I.,
- Nagel R.L.
- Pareti F.I.,
- Lattuada A.,
- Bressi C.,
- et al.
- Gill J.C.,
- Wilson A.D.,
- Endres-Brooks J.,
- Montgomery R.R.
- Cooke G.E.,
- Eaton G.M.,
- Whitby G.,
- et al.
- Geisen U.,
- Heilmann C.,
- Beyersdorf F.,
- et al.
- Alevriadou B.R.,
- Moake J.L.,
- Turner N.A.,
- et al.