Early postoperative bleeding and late thromboembolic complications are major concerns in patients with ventricular assist devices (VADs). The need to balance the risk of bleeding with that of thrombosis complicates treatment in patients with VADs, and the ideal regimen for device-related thrombosis prophylaxis has not been determined. Platelet activation associated with left ventricular assist device (LVAD) support has been reported in a few studies in patients with various assist devices, including the Novacor (Novacor Division, Baxter Health care Corp., Oakland, CA) left ventricular assist system, the Berlin Heart VADs (Berlin Heart, Berlin, Germany), the MicroMed DeBakey VAD (MicroMed Technology, Inc., Houston, TX), and the Thoratec VAD (Thoratec Corp., Pleasanton, CA).1–3 The genetic polymorphism of platelet glycoprotein (GP) IIIa may contribute to the development of complications.4 The interaction between leukocytes and platelets may be an important contributor to hemostasis in cellular models of thrombosis. However, this interaction has not been studied in depth, and key data are missing.
The purpose of this pilot study was to examine the effect of long-term LVAD support on platelet function and platelet-leukocyte interactions by using flow cytometry. We wanted to determine whether cytometric assessments of these interactions can help refine anticoagulation protocols in patients on long-term LVAD support, especially those patients with newer LVADs, such as the Jarvik 2000 (Jarvik Heart, Inc., New York, NY) and the HeartMate II (Thoratec Corp.), in which the platelet biosystem interface has not been studied.
Materials and Methods
This prospective study was approved by our hospital institutional review board. From January 2006 to April 2007, we enrolled in our study 15 patients (11 men; average age 46 ± 13 years) in New York Heart Association functional class IV who signed consent forms and underwent LVAD placement. Patients had to be older than 18 years to be included in the study. Patients were excluded if they had a prestudy platelet count <100 × 109/L, heparin-associated antibodies, or a congenital platelet disorder. We collected the following data: age, body weight, sex, drug regimen [aspirin, heparin, and epsilon aminocaproic acid], cardiopulmonary bypass (CPB) time, insertion and removal of an intra-aortic balloon pump, and clinical hematologic events, including thrombosis, bleeding, or disseminated intravascular coagulation.
Patients were implanted with one of the three different types of LVADs. We used two continuous-flow LVADs: the Jarvik 2000 (n = 6) and the Thoratec HeartMate II (n = 7). The third device used was a pulsatile LVAD—the Thoratec HeartMate XVE (Thoratec Corp.) (n = 2).
Left ventricular assist device implantation was performed according to our institution’s standard protocol.5–8 The device was implanted on-pump using a CPB circuit in 12 patients and off-pump in three patients. Cardiopulmonary bypass was performed on a standard bypass circuit equipped with a roller pump and membrane oxygenator. Anticoagulation was achieved with heparin (3 mg/kg or 300 U/kg body weight). Additional heparin was administered to maintain the kaolin-activated clotting time above 450–500 seconds. Epsilon aminocaproic acid was used as a hemostatic agent. At the end of the CPB period, heparinization was reversed by intravenous administration of protamine sulfate.
Heparin titration to 1.5 mean partial prothromboplastin time was started 12–24 hours after LVAD implantation. After extubation and removal of chest tubes, patients were started on oral warfarin (International normalized ratio [INR], 2–3), aspirin (81–325 mg/d), and dipyridamole (25–75 mg/d); warfarin was not given to the one patient who had the Thoratec XVE LVAD.
Collection of Blood Samples
We used the method of blood collection described by Macey et al.,9 with some modifications. This method allows simultaneous analysis of both platelets and leukocytes. Blood was collected into Cyto-Chex blood collection tubes (BCT) tubes from a radial arterial catheter in patients in intensive care unit or from venous access lines or peripheral venipuncture in patients transferred to the unit or in outpatients. Cyto-Chex BCT contains anticoagulant K3EDTA and a cell preservative that maintains the integrity of the white blood cell cluster of differentiation (CD) markers for immunophenotyping by flow cytometry.10 After mixing with an anticoagulant and a cell preservative in the Cyto-Chex tube, blood was transferred into a citrate-theophyline-adenosine-dipyridamole (CTAD) vacutainer tube (Becton-Dickinson, San Jose, CA), mixed gently, and processed for flow cytometry. CTAD tubes, which contain buffered sodium citrate, theophylline, adenosine, and dipyridamole, are formulated to preserve platelets and prevent in vitro platelet activation.
Blood samples were collected before LVAD implantation and at 3, 7, 14, 21, 30, 60, 90, and 180 days after implantation. We were unable to obtain samples on several postoperative days for the following reasons: two patients underwent heart transplantation, two patients died, one patient was transferred to another center, and venipuncture difficulties were encountered in two samples.
We used flow cytometry to quantify the expression of platelet surface GP IIb (αIIb), GP IIIa (β3), and GP Ibα; platelet activation markers P-selectin, CD63, and thrombospondin (TSP)11,12; circulating levels of monocyte-platelet complexes (MPC), granulocyte-platelet complexes (GPC), and lymphocyte-platelet complexes (LPC); leukocyte expression of P-selectin glycoprotein ligand 1 (PSGL-1); and monocyte expression of the lipopolysaccharide (LPS) receptor, CD14. We also determined platelet GP IIIa PLA polymorphism by flow cytometry, using PLA-specific monoclonal antibodies.
The following monoclonal antibodies were used: anti-CD41 FITC (GP IIb) (Beckman Coulter, Inc., Fullerton, CA), anti-CD61 fluorscein isothiocyanate (FITC) (GP IIIa) (Beckman Coulter), anti-CD42b PE (GP Ibα) (Beckman Coulter), anti-CD62P PE (P-selectin) (Immunotech, Beckman Coulter, Inc.), anti-CD63 PE (platelet GP 53) (Beckman Coulter), anti-thrombospondin PE (TSP) (Immunotech), anti-PSGL-1 (CD 162) PE (BD PharMingen, San Diego, CA), and anti-CD14 PC5 (Beckman Coulter).
For studying platelet markers, an aliquot of well-mixed blood, diluted with filtered fetal bovine serum (FBS) buffer (phosphate-buffered saline with 0.1% sodium azide and 2% FBS), was incubated with antibodies in the dark at room temperature for 15 minutes, and then diluted with 1.5 ml of FBS. Flow cytometric analysis was performed immediately, and flow rate was kept at <1,000 per second. For examining platelet-leukocyte aggregates and monocyte markers, an aliquot of blood was incubated with antibodies in the dark at room temperature for 20 min, and the sample was processed using lysing, stabilizing, and fixing reagents (ImmunoPrep™; Beckman Coulter). The lysed sample was kept in the dark and analyzed within 1 hour. A tube containing the unstained patient sample and an isotype control tube were used to evaluate autofluorescent versus nonspecific binding properties of antibodies. A total of 2,000 monocyte events and 10,000 platelet events were acquired in the list mode file from each sample. Results were expressed either as the percentage of positive events gated or median fluorescence intensity (MFI), or both.
Platelets were identified and distinguished from nonplatelet components (RBCs and leukocytes) by their characteristic forward light scatter (FS) versus side scatter (SS) characteristics (log FS; log SS) and their fluorescence from FL1 (CD61 or CD41), in the log-log dot plot of the FS versus CD61 or CD41. Results for the CD61, CD41, and CD42b were expressed as the MFI, and results for CD62P, CD63, and TSP were expressed as the percentage of positive events gated on CD61 and CD41-positive platelets, respectively. Monocytes were identified by specific staining with CD14 antibody, and MPCs were defined as CD14+ monocytes that were simultaneously positive for the CD41 marker. Lymphocytes and granulocytes were identified by their FS and SS characteristics, and LPCs and GPCs were identified by their staining positive for CD41. Results for cell aggregates were expressed as percentages of positive cells. Results for the CD14 and PSGL-1 were expressed as the MFI.
Platelet PLA1/A2 polymorphisms were differentiated by flow cytometry using appropriate monoclonal antibodies specific for the corresponding alleles. The method depended on the ability of monoclonal antibody to differentiate between the platelet GP IIIa receptors carrying either the PLA1 or PLA2 allele. Results for receptor density were expressed as arbitrary fluorescence intensity units.
Values are presented as the mean ± standard deviation. Repeated measures analysis of variance was used to analyze changes in variables over time. Student’s t test was used to analyze differences between subgroups of patients. Nonparametric Spearman R correlation coefficient was used to analyze the relationship between different markers. A p value of <0.05 was considered significant.
Patients’ preoperative characteristics are presented in Table 1. Before surgery, 11 of 15 patients were on heparin therapy and 6 of 15 were taking aspirin. The average duration of CPB was 112 ± 72 minutes in the 12 patients who underwent on-pump LVAD insertion.
The preoperative average percentages of platelets positive for activation markers were as follows: for CD62P, 27% ± 17%; for CD63, 9.7% ± 8.1%; and for TSP, 9.9% ± 6.8% (Figure 1). Throughout the study, we found only small fluctuations but no significant differences in the percent of activated platelets for these three markers. Of the 15 patients, only three had levels of CD62P-positive platelets <10% before LVAD implantation.
Our flow cytometric analysis of platelet GP IIIa receptor PLA genotype showed that five patients were heterozygous (A1/A2) and 10 were homozygous for the A1 allele. When baseline percentages of CD62P, CD63, TSP, MPC, GPC, and LPCs were analyzed according to the platelet PLA genotype, patients with the A1A2 genotype tended to have higher levels of markers than patients with the A1A1 genotype (for CD62P, 32.8% ± 19.2% vs 21% ± 15.2%; for CD63, 10.6% ± 9% vs 6.3% ± 4%; for TSP, 11.1% ± 8% vs 7.3% ± 4.3%; and for MPC, 13.2% ± 5.5% vs 8.5% ± 1.7%). However, the difference for platelet activation markers either did not reach or had only a borderline statistical significance (p = 0.058 for MPCs; p = 0.050 for GPCs).
The average percentages of the three types of leukocyte platelet complexes, MPC, GPC, and LPC, increased significantly on day 7 when compared with preimplantation levels (p < 0.005), peaked on day 21 (p < 0.004), and remained significantly higher than preoperative levels for up to 60 days after implantation (p < 0.04) (Figure 2). The preoperative baseline level of PSGL-1 surface expression in median fluorescence intensity was higher on monocytes (159.2 ± 20.6) than on granulocytes (90.0 ± 12.6) and lymphocytes (76.9 ± 9.9). After a transient increase on postoperative day 3, PSGL-1 surface expression persistently decreased thereafter (Figure 3).
We found significant inverse correlations between MPC and PSGL-1 expression on monocytes (R = −0.84, p < 0.0001), between LPC and lymphocyte PSGL-1 expression (R = −0.78, p = 0.0005), and between GPC and granulocyte PSGL-1 expression (R = −0.69, p = 0.0004). These findings suggest increased levels of leukocyte and platelet interactions. In addition, the density of CD14 on monocyte membranes increased over time during LVAD support, reaching statistical significance compared with preimplantation levels by day 21 (p < 0.05; Figure 4). Furthermore, a significant positive correlation seen between MPC and CD14 expression (R = 0.60, p = 0.011) indicated ongoing monocyte activation. We found a negative correlation between PSGL-1 density and CD14 (R = −0.46, p = 0.022).
During the study period, cerebrovascular accident (CVA) occurred in only one patient and was associated with extremely high levels of platelet activation markers. The CVA occurred on the last day of the study (180 days after LVAD implantation). The patient was admitted with acute onset of right hemiparesis, low back pain, and blurred vision. Computed tomography revealed left parietal infarction with hemorrhagic conversion. The patient was febrile, and his blood cultures were positive because of a drive line infection. Although the patient was on sodium warfarin and aspirin, his prothrombin time was subtherapeutic at 11.6 seconds, and his INR was 1.1. The platelet count of the patient was 277 × 109/L, and WBC count was 6.84 × 109/L. His percentages of activated platelets were highly increased when compared with the average of the remaining nine patients who were still on LVAD support at 180 days: CD62P, 73.8 vs 20.3% ± 9.0%; TSP, 30.2 vs 7.1% ± 4.9%; and CD63, 53.8 vs 6.7% ± 3.2%. However, his percentages of MPC (14.7%), GPC (14.2%), and LPC (12.8%) were similar to those of other patients. After 9 days, he was transferred to the inpatient rehabilitation unit and was discharged home 2 weeks later.
High levels of LPCs were detected in a 41-year-old patient who was septic and died of multiorgan failure 28 days after LVAD implantation. During the postoperative period after LVAD implantation, his organ function deteriorated, and he required prolonged mechanical ventilation and hemodialysis. He had coagulopathy, thrombocytopenia, leukopenia, and fever. At day 21, he had 55.3% MPCs compared with 22.5% ± 2.8% MPCs in the other 14 patients; day 21 was the peak point for MPC levels in the study population. Another patient died of sudden cardiac death at home 82 days after LVAD implantation; he had low levels of activated cells up to the last study point at day 60.
In this study, we found increased cellular activation and heterotypic cell interactions in patients undergoing long-term LVAD circulatory support. Expression of P-selectin (CD62P) was increased before LVAD implantation and remained so, with oscillations, throughout the study. Our findings of increased platelet activation support results published previously in studies of patients with chronic heart failure and those with LVADs.1–3,13
Other studies have shown increased expression of the platelet activation markers, CD62, CD63, and TSP, in patients with end-stage heart failure before device placement and during the assist period with the Novacor left ventricular assist system (n = 5) and the Berlin Heart (n = 3).1 Bonaros et al.,2 using extensive platelet monitoring, studied the potential thrombogenic properties of a continuous-flow axial pump in 13 patients who received the MicroMed DeBakey VAD as a bridge to transplantation and found that platelet activation markers were upregulated in the postoperative period. Similarly, in a study of 15 patients with a Thoratec pulsatile VAD, the platelet activation profile indicated persistent platelet activation with a consistently high inflammatory state and endothelial activation.3
Platelet PLA polymorphism contributes to the development of coronary thrombotic events,14 partly due to platelet hyperactivity. However, findings from large prospective studies have been controversial.15 In addition, genetic polymorphism of platelet GP IIb/IIIa may contribute to the development of complications in LVAD patients.4 A recent study showed a relationship between the PLA genotype and postoperative complications in 41 patients with pulsatile or axial flow VADs4; patients with the A1A1 genotype developed more bleeding complications (39% vs 0%, p = 0.021), whereas patients with the A1A2 genotype showed a tendency toward more thromboembolic events (13% vs 30%, p = 0.33). We could not confirm genotype-related complications in our patients, most likely because of the small number of patients. However, our finding of higher, but statistically insignificant, levels of platelet markers in A1A2 patients confirms reports from a previous study.
Few reports are available on the role of leukocytes in heart failure. The interaction of circulating monocytes and platelets has recently been shown to be an important mechanism for thrombotic events and inflammation in severe heart failure. In addition, measurement of circulating monocyte-platelet aggregates may be a more sensitive indicator of in vivo platelet activation than circulating activated platelets.16 We found increased monocyte activation and leukocyte-platelet interactions during LVAD implantation. The inflammatory state of heart failure most likely contributes significantly to increased baseline levels of markers of cell activation. Platelets are probably in a state of prolonged activation, possibly related to the inflamed heart and the activation of other cell types. In contrast to platelet activation markers, the percentages of leukocyte-platelet complexes were significantly higher in our patients after LVAD implantation than before LVAD implantation.
The role of circulating monocytes, which are an important source of proinflammatory cytokines, has not been studied extensively in LVAD patients.10 Cardiac surgery and CPB, however, have been associated with monocyte activation.17,18 The multifunctional LPS receptor (CD14) on monocytes serves as a pattern recognition molecule in innate immune responses against microorganisms and other exogenous and endogenous stress factors. Ligation of monocyte CD14 receptor and other coreceptors induces the intracellular signaling pathways that lead to the synthesis and release of proinflammatory cytokines and chemokines. The density of CD14 expression on the surface of monocytes and the serum concentration of soluble CD14 molecules have been increased in patients with acute coronary syndrome.19,20 Our findings of an increasing density of monocyte membrane expression of CD14 and a significant positive correlation between CD14 expression and MPC levels in our patients indicate ongoing monocyte activation during LVAD support.
The formation of leukocyte-platelet complexes depends on the interaction between P-selectin on platelets and its ligand, PSGL-1, on leukocytes. Constitutively expressed on the surface of circulating leukocytes, PSGL-1 contributes to leukocyte-leukocyte, leukocyte-platelet, and leukocyte-endothelial interactions.21 Our results showed decreasing levels of leukocyte PSGL-1 during LVAD support, which is likely due to its engagement in platelet-leukocyte complexes, and a negative correlation between PSGL-1 density and CD14. Surface expression of PSGL-1 significantly decreased on human neutrophils, monocytes, and eosinophils after stimulation with platelet-activating factor and phorbol 12-myristate-13-acetate.22 Platelet-monocyte interactions induce expression of monocyte tissue factor, which contributes to procoagulant responses, thrombogenesis, and the proinflammatory state.23–25
The effect of platelet-monocyte interactions on hemostasis during LVAD support is not well understood. The presence of significantly high numbers of prothrombotic platelet-monocyte aggregates in patients with transient ischemic attack or stroke has been demonstrated in nonsurgical settings.26 Snyder et al.27 demonstrated the presence of both leukocyte-platelet aggregates and monocyte tissue factor in cows with VADs, showing low levels of aggregates at baseline, and a persistent increase after VAD implantation. These results are similar to ours. Loebe et al.28 showed that LVAD implantation was associated with stimulation of the inflammatory system, with an overall increase in inflammatory factors after implantation.
The clinical significance of our findings requires further study. To more fully evaluate the significance of circulating leukocyte-platelet complexes and other markers of cell activation, including cell-derived microparticles, larger groups of patients should be studied and markers correlated with clinical events. Patients with end-stage heart failure undergoing LVAD implantation are terminally ill and are more susceptible to the effects of surgical intervention, CPB, and LVAD initiation. An increase in cell activation from baseline levels should be expected in these patients. It will be advantageous to know at what level adverse event should be expected. In this study, we were able to associate high levels of activation markers with clinical events in only two patients; we saw a marked increase in platelet activation in a patient with a CVA episode and increased MPC in a patient with end-stage multiorgan failure, sepsis, and death. Activated platelets have been associated with increased risk of thrombotic complications.29 Furthermore, increased formation of leukocyte-platelet complexes has been reported in patients with sepsis and multiorgan failure.30 Moderately higher levels seen in other patients may have been in response to surgical injury, CPB, or LVAD initiation and may not be cause for alarm. However, our study group was too small to provide a definitive answer.
Our finding of increased platelet activation in patients before and during circulatory support with LVADs may be a result of their primary disease—severe heart failure—rather than an effect of the support device. Increased monocyte activation and leukocyte-platelet interactions observed during the first month of LVAD support may indicate the development of an inflammatory state, or it could be in response to the surgical procedure and CPB (and LVAD initiation) rather than an effect resulting exclusively from LVAD support.
Supported in part by the MacDonald General Research Fund.
1.Dewald O, Schmitz C, Diem H, et al
: Platelet activation markers in patients with heart assist device. Artif Organs
29: 292–299, 2005.
2.Bonaros N, Mueller MR, Salat A, et al
: Extensive coagulation monitoring in patients after implantation of the MicroMed DeBakey continuous flow axial pump. ASAIO J
50: 424–431, 2004.
3.Houel R, Mazoyer E, Boval B, et al
: Platelet activation and aggregation profile in prolonged external ventricular support. J Thorac Cardiovasc Surg
128: 197–202, 2004.
4.Potapov EV, Ignatenko S, Nasseri BA, et al
: Clinical significance of PlA polymorphism of platelet GP IIb/IIIa receptors during long-term VAD support. Ann Thorac Surg
77: 869–874, 2004.
5.Frazier OH: The development of an implantable, portable, electrically powered left ventricular assist device. Semin Thorac Cardiovasc Surg
6: 181–187, 1994.
6.Frazier OH, Delgado RM III, Kar B, et al
: First clinical use of the redesigned HeartMate II left ventricular assist system in the United States: A case report. Tex Heart Inst J
31: 157–159, 2004.
7.Radovancevic B, Frazier OH, Duncan JM: Implantation technique for the HeartMate left ventricular assist device. J Card Surg
7: 203–207, 1992.
8.Westaby S, Frazier OH, Pigott DW, et al
: Implant technique for the Jarvik 2000 heart. Ann Thorac Surg
73: 1337–1340, 2002.
9.Macey M, McCarthy D, Azam U, et al
: EDTA-CTAD: A novel anticoagulant for the flow cytometric assessment of platelet and neutrophil activation ex vivo in whole blood. Cytometry B Clin Cytom
52B: 30–40, 2003.
10.Warrino DE, DeGennaro LJ, Hanson M, et al
: Stabilization of white blood cells and immunologic markers for extended analysis using flow cytometry. J Immunol Methods
305: 107–119, 2005.
11.Michelson AD, Barnard MR, Krueger LA, et al
: Evaluation of platelet function by flow cytometry. Methods
21: 259–270, 2000.
12.Schmitz G, Rothe G, Ruf A, et al
; for the European Working Group on Clinical Cell Analysis: European working group on clinical cell analysis: Consensus protocol for flow cytometric characterization of platelet function. Thromb Haemost
79: 885–896, 1998.
13.Wilhelm CR, Ristich J, Kormos RL, et al
: Monocyte tissue factor expression and ongoing complement generation in ventricular assist device patients. Ann Thorac Surg
65: 1071–1076, 1998.
14.Weiss AJ, Bray PF, Tayback MT, et al
: The platelet glycoprotein IIIa polymorphism PlA2 and inherited platelet risk factor for coronary thrombotic events. N Engl J Med
334: 1090–1094, 1996.
15.Aleksic N, Juneja H, Folsom AR, et al
: Platelet PlA2 allele and incidence of coronary heart disease: Results from the Atherosclerosis Risk In Communities (ARIC) study. Circulation
102: 1901–1905, 2000.
16.Michelson AD, Barnard MR, Krueger LA, et al
: Circulating monocyte-platelet aggregates are a more sensitive marker of in vivo platelet activation than platelet surface P-selectin: Studies in baboons, human coronary interventions, and human acute myocardial infarction. Circulation
104: 1533–1537, 2001.
17.Fontes ML, Mathew JP, Rinder HM, et al
: Multicenter study of perioperative ischemia (McSPI) research group. Atrial fibrillation after cardiac surgery/cardiopulmonary bypass is associated with monocyte activation. Anesth Analg
101: 17–23, 2005.
18.Rinder CS, Bonan JL, Rinder HM, et al
: Cardiopulmonary bypass induces leukocyte-platelet adhesion. Blood
79: 1201–1205, 1992.
19.Zalai CV, Kolodziejczyk MD, Pilarski L, et al
: Increased circulating monocyte activation in patients with unstable coronary syndromes. J Am Coll Cardiol
38: 1340–1347, 2001.
20.Lee WH, Lee Y, Jeong JO, et al
: Activation of CD14 on circulating monocytes in patients with acute coronary syndrome. Int J Cardiol
80: 135–142, 2001.
21.McEver RP, Cummings RD: Perspectives series: Cell adhesion in vascular biology. Role of PSGL-1 binding to selectins in leukocyte recruitment. J Clin Invest
100: 485–491, 1997.
22.Lorant DE, McEver RP, McIntyre TM, et al
: Activation of polymorphonuclear leukocytes reduces their adhesion to P-selectin and causes redistribution of ligands for P-selectin on their surfaces. J Clin Invest
96: 171–182, 1995.
23.Shibamiya A, Tabuchi N, Chung J, et al
: Formation of tissue factor-bearing leukocytes during and after cardiopulmonary bypass. Thromb Haemost
92: 124–131, 2004.
24.Osterud B: The role of platelets in decrypting monocyte tissue factor. Semin Hematol
38: 2–5, 2001.
25.Bokarewa MI, Morrissey JH, Tarkowski A: Tissue factor as a proinflammatory agent. Arthritis Res
4: 190–195, 2002.
26.Garlichs CD, Kozina S, Fateh-Moghadam S, et al
: Upregulation of CD40-CD40 ligand (CD154) in patients with acute cerebral ischemia. Stroke
34: 1412–1418, 2003. (Erratum in: Stroke
34: 1820, 2003).
27.Snyder TA, Eash HJ, Litwak KN, et al
: Blood biocompatibility assessment of an intravenous gas exchange device. Artif Organs
30: 657–664, 2006.
28.Loebe M, Koster A, Sanger S, et al
: Inflammatory response after implantation of a left ventricular assist device: Comparison between the axial flow MicroMed DeBakey VAD and the pulsatile Novacor device. ASAIO J
47: 272–274, 2001.
29.Gawaz M, Neumann FJ, Ott I, et al
: Role of activation-dependent platelet membrane glycoproteins in development of subacute occlusive coronary stent thrombosis. Coron Artery Dis
8: 121–128, 1997.
Copyright © 2009 by the American Society for Artificial Internal Organs
30.Gawaz M, Fateh-Moghadam S, Pilz G, et al
: Platelet activation and interaction with leucocytes in patients with sepsis or multiple organ failure. Eur J Clin Invest
25: 843–851, 1995.