Extracorporeal membrane oxygenation (ECMO) is used to support a critically ill patient who developed acute respiratory or acute cardiac failure.1,2 Despite technical improvements and refinement in anticoagulation management, thrombosis remains one of the most common and feared complication of ECMO support that causes morbidity and mortality in these patients.3,4 The incidence of clots within the ECMO circuit and the membrane oxygenator (MO) was reported to be 3–22% and 21%, respectively.5,6 The underlying pathophysiologic process that occurs at the blood/biomaterial interface has not been fully understood. Activation of coagulation and fibrinolytic pathways as well as a complement-mediated inflammatory response are discussed.6,7
Scanning electron microscopy and multidetector computed tomography were used to visualize and localize the clots within the MO.8,9 Beely et al.9 identified different types of clots within the MOs consisting of platelets and red blood cells that were incorporated in extended fibrin strands. In a preliminary study, our group emphasized the relevance of adherent leukocytes as important copartners during clot formation within the MO for the first time.10,11
The aim of the current study was the quantitative assessment of cellular depositions on the surface of the gas exchange membranes in MOs after use. Furthermore, the existence of membrane-like structures spanning multiple gas capillaries (PM) was analyzed as an indicator of emerging thrombosis within the MO. Therefore, clinical and laboratory parameters of patients with PMs or without PMs within their MOs were analyzed.
MATERIALS AND METHODS
From April 2008 to February 2012, a total of 41 polymethylpentene (PMP) MOs (27 Quadrox PLS, Maquet Cardiopulmonary, Rastatt, Germany; 6 HiliteLT, 4 ILA-activ, and 2 Nova Breath, Xenios, Heilbronn, Germany; 2 ECC.O5, Sorin, Mirandola, Italy) were removed when clinically indicated at the end of therapy (27 were successfully weaned; 5 were removed after death of the patient) or during ECMO therapy due to technical complications (7 with coagulation disorder within the MO, 1 because of risk of infection, and 1 with an acute MO thrombosis).5 The MOs came from 41 randomly selected Caucasian patients (36 men, 5 women) with acute cardiac and respiratory failure that required ECMO support (31 veno-venous and 10 veno-arterial). None of the patients had a known inherent or an acquired coagulation disorder. Standard treatment for ECMO patients was described earlier.5 At the time of MO collection, 30 patients were treated with heparin (5 of 30 with additional aspirin; 2 of 30 with additional clopidogrel), 9 patients got argatroban (1 of 9 with additional aspirin), 1 patient was treated with aspirin alone, and 1 patients did not require any anticoagulants/antiplatelet drug. Central laboratory-based coagulation testing including activated partial thromboplastin time (aspired value: 50–60 seconds), international normalized ratio (aspired value: <1.3), and platelet count was used for coagulation monitoring. Gas capillaries from MOs were analyzed for their cellular coverage by direct counts of 4′,6-diamidino-2-phenylindole (DAPI)-stained nuclei and by image analysis measuring the area covered by DAPI-stained nuclei in microscopic images of MO capillaries. Patient characteristics and laboratory data before ECMO are summarized in Table 1.
Preparation of Oxygenators
After removal from the ECMO circuit, the MO was extensively rinsed with 10 L physiologic saline (1 L/min) and 500 ml fixative (4% paraformaldehyde in 0.1 M phosphate buffer [PB], pH 7.2). Another 500 ml fixative was circulated in closed loop mode through the MO for 30 min at room temperature (RT). Subsequently, the fixative was replaced by rinsing 1.5 L PB through the MO and thoroughly washed out by circulation of another 1.5 L PB in closed loop modus for 30 min. Finally, the MO was rinsed with 500 mL of 10% and 500 mL of 20% saccharose solutions (in PB) followed by 500 mL of 30% saccharose solution in closed loop mode for 30 min as a cryoprotectant. The MO was then frozen at −20°C and stored at −80°C.
The frozen MO was dismantled and the core of gas capillary sheets was segmented in 9 blocks (3 × 3 × 4.5 cm) without thawing. The blocks were wrapped in aluminum foil and stored at −80°C. For the analysis, the frozen blocks of capillary layers were broken up, and small areas of gas capillary sheets (0.5 × 1.3 cm) from the center of the middle block (“sample”) were collected and rehydrated in phosphate-buffered saline (PBS).
For immunohistochemical staining, each sample was washed in PBST (PBS plus 0.1% TritonX100 at RT) and incubated with donkey serum (10% in PBST; 30 min at RT) to prevent nonspecific binding of the secondary antibodies. Specific monoclonal antibodies (mouse antihuman CD45 [1:20] and rabbit antihuman vWF [1:50]; DAKO, Cambridgeshire, United Kingdom) were used to incubate the samples overnight at 4°C. Samples were washed with PBST and stained with secondary antibodies for 90 min at RT (1:1000; donkey antimouse IgG-fluoresceinisothiocyanat [FITC]; donkey antirabbit IgG-TexasRed; DAKO). After washing, samples were counterstained with DAPI (0.75 µg/ml in PBST; Vector Laboratories, Petersborough, United Kingdom) for another 30 min at RT. The samples were carefully placed on a microscope slide, embedded in Fluoromount-G (SouthernBiotech, Birmingham, AL), topped with a coverslip and visualized using a Leica DMRBE fluorescence microscope (Bensheim, Germany). To estimate the extent of cellular deposits, the central capillary of the stained sample was selected and digitalized using a Spot2000 camera (Diagnostic Instruments, Stirling Heights, MI) under software control (Visiview [R]; Visitron Systems GmbH, Puchheim, Germany) resulting in images with 1600 × 1200 pixels. Low-power overviews were obtained by combining 3 slightly overlapping images that were digitized with the 2.5× lens of the microscope (resolution 3.000 µm/pixel). For quantification of cellular coverage, 12 nonoverlapping images of the central capillary were obtained with the DAPI fluorescence channel at higher magnification in the 16-bit mode of the camera using the 20× lense (resolution 0.375 µm/pixel). Details of representative cells and cell accumulations were documented at high magnification using a 40× lens of the fluorescence microscope (resolution 0.187 µm/pixel).
Estimation of Cellular Depositions Using Nuclear Staining
DAPI staining of nuclei was used to visualize the colonization of PMP capillaries by nucleated cells. Accumulations of high amounts of nucleated cells that were incorporated in membrane-like structures spanning multiple gas capillaries (pseudomembrane, PM) were used as a qualitative criterion to define the 2 study groups.10
Two methods were used to obtain quantitative measures of cellular coverage of MO capillaries.
One approach to quantify cellular coverage used image analysis to measure the area covered by DAPI-stained nuclei in images of the capillaries. For this analysis, 12 nonoverlapping images of the central capillary were obtained with the DAPI fluorescence channel in the 16-bit mode of the camera using the 20× lens. The microscope settings and exposure time were determined in pilot experiments and subsequently kept constant for digitizing the specimen. The images were opened using ImageJ (version 1.42q, NIH, http://imagej.nih.gov/ij) and a lower threshold of 1,536 for the gray value was set using the “image,” “adjust,” and “threshold” menue. This threshold value was established in pilot experiments to discriminate between DAPI-stained nuclei and background. Using the “analyze” and “measure” menus in ImageJ, the number of pixels with gray values above threshold was determined for each image and used as a measure of cell coverage. The whole procedure for measuring the number of suprathreshold pixels of an image was implemented as an ImageJ macro allowing fast and standardized analysis.
For the second, more traditional approach to determine the coverage of the capillaries by nucleated cells, the same images of 27 PLS-MOs were used to manually count DAPI-stained nuclei and calculate the density of cells. For regions with medium-to-low coverage counting was easy and straightforward; however, in regions with high coverage and where the cells grew in thicker layers like PM counting became difficult and unreliable.
To evaluate whether measurement of cell coverage by image analysis arrives at similar conclusions as the more traditional and tedious method of counting cell nuclei, the results from both types of measurement were correlated for a subset of 19 PLS-MOs where cell coverage was sufficiently low to obtain manual cell counts (Figure 2A).
Data Collection and Statistical Analysis
Data were collected prospectively. Patient data (including laboratory, respiratory, and physical data) were documented routinely in our Regensburg ECMO database. This study was approved by Ethics Committee of the University of Regensburg (vote # 106-336-104). The device was a waste-product, biological material was fixed with formaldehyde, and therefore, no genetic analysis was intended.
Data for groups with and without PMs are reported as median and IQRs. Nonparametric Mann–Whitney U tests were used to identify significant differences between groups for parameters determined on an analog scale (see Tables 1 and 2). For comparing proportions between the groups with and without PMs (e.g., gender ratio, discharge from hospital), the Fisher exact test was used. The correlation between the 2 measures of cellular coverage of MO capillaries (see Figure 2A) was examined using a Spearman rank order correlation. A p value of ≤0.05 was considered statistically significant. The software package IBM SPSS-Statistics 20.0 (IBM Corp., Armonk, NY) was used for the statistical evaluation.
Classification of PMP-MOs
On PMP membranes cellular deposits identified by DAPI fluorescence were highly variable between different MOs. Capillaries from 18 of 41 MOs were evenly colonized with a low density of single cells (Figure 1A). Another 17 MOs presented a higher cell density especially around the crossing points of adjacent gas capillaries and the warp threads (Figure 1B). The remaining 6 MOs were covered with PM consisting of high amounts of nucleated cells and cell aggregates (Figure 1C). The presence or absence of PMs was used as a qualitative criterion to define the study groups.10
Appraisal of Cellular Accumulations on the Surface of PMP-Gas Capillaries
DAPI fluorescence was used to quantify the extent of cellular deposits. Although cell coverage of 8 of the 27 PLS-MOs was too high for unequivocal counting of DAPI-stained nuclei, samples from the remaining 19 MOs were used to count nuclei and determine the cell density. The cell counts of the 12 images of 1 sample from each of the 19 MOs were converted to cell density (cell count per mm2). Cell density ranged between 60 and 835 cells/mm2 with a median (IQR) of 325 (148–540) cells/mm2.
In addition to cell counts, image analysis was used for the same images to determine the sum of pixels with gray levels above threshold representing the DAPI-stained nuclei. The scatter plot in Figure 2A illustrates the correlation between both measures of cell accumulation on the capillaries (Spearman rank test, rs = 0.890; p < 0.001). Also shown is the best fit linear regression line with a coefficient of determination (r2 = 0.71) demonstrating that the variation of the sum of suprathreshold pixels explains more than 70% of the variation of cell density. The high correlation between cell density and DAPI fluorescence shows that the fast image analysis of DAPI fluorescence provides an adequate measure of cell accumulation that was also used to quantify fluorescence where the high density of cells prevented counting.
Consequently, the extent of cellular deposits on gas capillaries was determined by image analysis of DAPI fluorescence for all 41 MO samples with and without PMs. As shown in Figure 2B, the cell accumulation of samples from PM-free MOs was significantly lower (Mann–Whitney U test, p = 0.013) than in MOs with PM.
Identification of Cells on PMP-Gas Capillaries
The presence of CD45-positive leukocytes on the surface of PMP-gas capillaries and the potential of some of these cells to differentiate to endothelial-like cells under in vitro conditions were shown in a previous study.10 However, there had been no information about the localization of the cells within the gas fibers of the MO. Therefore, immunofluorescence was used in the subset of PLS-MOs to identify CD45-positive leukocytes and endothelial cells (via detection of vWF). In addition, the presence of a DAPI-stained nucleus allowed to discriminate between endothelial cells and platelets.
Only regions with low cell density could be used to unequivocally identify leukocytes and vWF-positive structures. A clear identification of the cellular expression of both markers was not possible in extended areas with cell aggregates and PMs. In addition, at high magnification, because of the shape of the capillaries many cells were not exactly in focus resulting in overlapping and blurred cell images that prevented a quantification of the different cell types. Despite these limitations, areas with single cell coverage were used to estimate the proportion of adherent CD45-positive leukocytes. In one focus level, only 44% (range, 0–78%) of the cells were identified as leukocytes. Cell characterization was possible for 19 of the 27 PLS-MOs. In 4 MOs, single endothelial-like cells were identified by their positive staining for vWF that colocalized with a DAPI-positive nucleus. However, <1% of adherent cells expressed vWF in the cytoplasm of nucleated cells. In 3 of the 27 MOs vWF staining was located directly on the PMP-gas capillaries next to single leukocytes. Samples from another 7 MOs were completely free of vWF staining. Samples from 11 MOs presented vWF-positive filament-like structures. Samples with PMs were associated with a weak Texas Red staining corresponding to a low expression of vWF (4 of the 6 MOs). PMs with vWF-positive filaments were documented in 2 of the 6 MOs. Finally, the deposits around the crossing points of adjacent gas capillaries consisted of large amounts of CD45-positive cells and extended inclusions of vWF (11 of the 27 MOs).
Assignment to Clinical Data
Differences in patient characteristics and ECMO data from groups of patients with and without PMs within their MOs are presented in Tables 1 and 2. MOs that developed PMs derived from significantly younger patients (median [IQR]; age, 36 [31–48] years) compared with those without PMs (61 [51–71] years; p = 0.040). All other patient data and initial clotting parameters were not significantly different between the groups (Table 1). The existence of a PM was associated with an increased white blood cell count before ECMO initiation (median [IQR]; with PMs, 21.4 [16.4–24.4]; without PMs, 15.2 [7.8–18.1] × 109 per L; p = 0.051]. There was no other indication for an increased inflammatory state (no fever, no increase in CRP levels) in the patients with PMs.
The analyzed MOs were removed during ECMO therapy (exchange due to technical complications, no PMs: 8 of the 35; with PMs: 1 of the 6) or at the end of therapy (weaning, no PMs: 22 of the 35, with PMs: 5 of the 6; death, no PMs: 5 of the 35; with PMs: 0 of the 6; Table 2). The MOs were exchanged due to an impending coagulation disorder (no PMs: 6 of the 35; with PMs: 1 of the 6), an acute MO thrombosis (no PMs: 1 of the 35) and suspicion of a MO infection (no PMs: 1 of the 35). Specific laboratory (d-dimers, fibrinogen, platelet count) and technical data (dpMO, blood flow) 2 days before removal of the MO were used to describe the coagulation status within the MO.5 The levels of d-dimers and fibrinogen, the number of platelets (as indicators for imminent coagulation disorder) as well as activated partial thromboplastin time and international normalized ratio were inconspicuous and not different between the 2 study groups (Table 2). Furthermore, there was no indication of significant hemolysis since the levels of free hemoglobin were around 70–80 mg/L in both groups (Table 2). The development of a PM did not affect the pressure drop across the MO (dpMO) and blood flow (Table 2).
The running time of the analyzed MOs tended to be longer for MOs with PMs (median [IQR], 11 [6–19] days) compared with the MOs without PMs (8 [5–11] days). In addition, the total ECMO support time was also higher for patients with PM-containing MOs (median [IQR], 15 [6–25] days) compared with those without PMs (10 [5–17] days; Table 2). However, in view of the high variance of the data, these differences between both groups were not significant. Furthermore, the transfusion supply of blood products (red blood cell, platelet count) had no effect on the extent of cellular depositions within the MOs (Table 2).
Each patient with a diagnosed PM in a MO survived ECMO support and was discharged from hospital. In contrast, only 17 patients (49%) with MOs free of detected PMs were discharged from hospital. The statistical comparison of the presently analyzed samples of MOs with PM and without PM revealed that this difference is significant (Fisher exact test, p = 0.027) and clearly suggests that the development of PMs in the MO may be positive and is not associated with a bad prognosis.
Blood contact with polymeric PMP-gas exchange membranes caused adhesion of nucleated cells with a high variability between the MOs analyzed. Especially younger patients on ECMO developed large-scale and fiber-spanning membrane-like structures within the MO. The PMs had no negative influence on the outcome: All patients with diagnosed PMs in their MO were discharged from the hospital, while in this cohort, this was true only for 49% of the patients with PM-free MOs (17 of the 35).
The safety of clinical application of different ECMO systems was proven by Yu et al.12 The usage of PMP-MOs, centrifugal pumps, and antithrombotic coatings allowed prolongation of support time and a lower rate of procedure-related complications.13 Nevertheless, clot formation within the MO remains a common finding in ECMO despite adequate systemic anticoagulation and heparin coating of the circuit components.9,14 Multidetector computed tomography and scanning electron microscopy was used to identify thrombotic deposits within the same type of MOs mainly in the inlet part of the device.9,14 In these studies, clot volume did not correlate with ECMO support time.14 In the current study, the data show on average a longer ECMO time in the group that developed PMs (15 days) compared with the group without PMs (10 days); however, this difference was not significant because of the large interindividual variability in both groups, suggesting that other parameters may be important for the development of PMs. Different predictive factors indicating an impending system exchange are discussed in the literature.5,14,15 The levels of d-dimers and fibrinogen as well as the platelet count predict device-related coagulation disorder.5 In the current study, these selected laboratory parameters were not different between both study groups (Tables 1 and 2). Furthermore, the development of a PM did not result in an increase of the dpMO or alteration of the blood flow. Obviously, the presence of a PM within a MO was not an indicator for an imminent MO thrombosis. The only acute MO thrombosis was observed in a PM-free MO. It could be speculated that the development of PMs was caused by a reduced blood flow during the weaning process. It has been previously shown that low blood flow caused blood stasis and accelerated clot formation.16 However, the proportion of MOs that were removed at the end of ECMO therapy (weaning and death) was comparable in both groups (no PMs: 77%; with PMs: 83%).
All efforts to elucidate the process of clot formation within the MO are so far not very productive. In a previous study, scanning electron microscopy and fluorescence staining was used to identify cellular deposits within the MO.8 Membrane-like structures consisting of a fibrous network with imbedded platelets, red blood cells, and nucleated cells were identified as shown in a previous study.10 The current study showed in detail that nucleated cells covered gas fibers as single cells and cell clusters and were incorporated into PMs. The heterogeneous arrangement of the cells around the gas fibers aggravated the identification of the different cell types using immunofluorescence. Nevertheless, in areas with single cell coverage about 44% (range 0–78%) of adherent and well-defined cells were identified as CD45-positive leukocytes. In contrast, the detachment of the cells from the surface of the gas fibers with trypsin and flow cytometric analysis revealed almost 100% CD45-positive leukocytes.11 Some of the problems associated with the unequivocal classification of cells in the current study are described in the section “Identification of Cells on PMP-Gas Capillaries” and indicate that our data likely underestimate the proportion of different immunohistochemically stained cell groups like CD45-positive leukocytes. Therefore, based on the flow cytometric data, we suppose that the majority of nucleated cells in the used MOs are leukocytes. The involvement of leukocytes in the clot formation process was already known.17 Contact of blood with polymeric surfaces activates both neutrophils and monocytes and increased cell adhesion. In vivo studies have found activated leukocytes adhering to stents,18 heart valves,19 oxygenators,20,21 and hemodialysis membranes.22 The underlying mechanisms are complex and included inflammatory processes, formation of leukocyte–platelet–aggregates, as well as shear–stress-induced platelet activation.17 Platelet adhesion, activation, and aggregation were mediated under high-flow conditions by the multimeric plasma glycoprotein vWF.23 The current study identified expanded filaments of vWF within the PMs and cellular aggregates in the MOs. It could be speculated that plasma vWF mediates platelet adhesion to surfaces coated with fibrinogen,24 which is adsorbed onto surfaces of many materials used in biomedical instruments, including ECMO.25,26 In a clinical study, it was shown that the mediators of platelet aggregation—the high–molecular weight vWF multimers—disappeared in the blood of patients during ECMO support.27,28 The loss resolved after termination of ECMO. In addition, there was no correlation with bleeding events.27,29 Thus, one hypothesis is that the high–molecular weight vWF multimers aggregated within the PMs and cellular aggregations on the gas fibers. However, up to now, no correlation of PMs and the development of clots within the MO have been shown.
The age of patients with MOs that developed PMs was significantly lower compared with the group that developed no PMs during ECMO. To our knowledge, to date, no study was published showing age-related cellular accumulations during ECMO support. It could be hypothesized that the restricted function of the immune cells in elderly patients prevented extended cellular attachment on PMP surfaces during ECMO therapy.30 Especially, granulocytes and monocytes/macrophages have a central role in the immune system remodeling during age.31,32 Chiricolo et al.33 documented a decrease in monocyte subpopulations bearing the adhesion molecule CD11a/CD18 and an increase in CD44 antigen density on monocytes in elderly. Furthermore, leukocytes from older patients expressed higher levels of CD11b and CD15, mediating an increase in neutrophil adhesion with endothelial cells and platelets.34,35 Also, an increased proportion of CD62L-negative granulocytes in the elderly may lead to an impairment in cell adhesion.36 Remodeling of adhesion molecules on inflammatory cells in elderly people influenced not only leukocyte function but also the interactions with platelets.36 These changes might possibly contribute to the absence of PMs in the MOs from older patients analyzed in the current study. However, there was no indication that MOs from older patients show less clot formation.
Despite the development of large-scale and fiber-spanning membrane-like structures within the MO of some patients, there was no negative effect on the outcome. The extent of cellular colonization of gas fibers and the formation of PMs were not the indicators for increased mortality.
The interpretation of the data from the current study needs to consider the following limitations: 1) small sample size, 2) inclusion of different types of PMP-MOs, 3) usage of veno-venous and veno-arterial ECMO systems, 4) disregard of the reasons for removal of the MO, and 5) data limited to Caucasian patients.
Elderly people on ECMO showed a tendency for reduced leukocyte adhesion on PMP-gas membranes that might reduce the risk of clot formation within the MO. The development of cellular aggregates within MOs did not increase mortality.
The authors gratefully acknowledge the excellent technical assistance of K. Bielenberg. Supported by a research grant from Maquet (Rastatt, Germany). The material presented in this publication is part of the doctoral thesis of J.G. Wilm (http://epub.uni-regensburg.de/33786/).
1. Karagiannidis C, Brodie D, Strassmann S, et al. Extracorporeal membrane oxygenation
: evolving epidemiology and mortality. Intensive Care Med 2016.42: 889–896.
2. Combes A, Bréchot N, Luyt CE, Schmidt M. Extracorporeal membrane oxygenation
: beyond rescue therapy for acute respiratory distress syndrome? Curr Opin Crit Care 2017.23: 60–65.
3. Haneya A, Philipp A, Foltan M, et al. First experience with the new portable extracorporeal membrane oxygenation
system Cardiohelp for severe respiratory failure in adults. Perfusion 2012.27: 150–155.
4. Oliver WC. Anticoagulation and coagulation management for ECMO. Semin Cardiothorac Vasc Anesth 2009.13: 154–175.
5. Lubnow M, Philipp A, Foltan M, et al. Technical complications during veno-venous extracorporeal membrane oxygenation
and their relevance predicting a system-exchange–retrospective analysis of 265 cases. PLoS One 2014.9: e112316.
6. Murphy DA, Hockings LE, Andrews RK, et al. Extracorporeal membrane oxygenation
-hemostatic complications. Transfus Med Rev 2015.29: 90–101.
7. Annich GM. Extracorporeal life support: the precarious balance of hemostasis. J Thromb Haemost 2015.13(suppl 1): S336–S342.
8. Dornia C, Philipp A, Bauer S, et al. Visualization of thrombotic deposits in extracorporeal membrane oxygenation
devices using multidetector computed tomography: a feasibility study. ASAIO J 2013.59: 439–441.
9. Beely BM, Campbell JE, Meyer A, et al. Electron microscopy as a tool for assessment of anticoagulation strategies during extracorporeal life support: the proof is on the membrane. ASAIO J 2016.62: 525–532.
10. Lehle K, Philipp A, Gleich O, et al. Efficiency in extracorporeal membrane oxygenation
-cellular deposits on polymethylpentene membranes increase resistance to blood flow and reduce gas exchange capacity. ASAIO J 2008.54: 612–617.
11. Lehle K, Friedl L, Wilm J, et al. Accumulation of multipotent progenitor cells on polymethylpentene membranes during extracorporeal membrane oxygenation
. Artif Organs 2016.40: 577–585.
12. Yu K, Long C, Hei F, et al. Clinical evaluation of two different extracorporeal membrane oxygenation
systems: a single center report. Artif Organs 2011.35: 733–737.
13. Riley JB, Scott PD, Schears GJ. Update on safety equipment for extracorporeal life support (ECLS) circuits. Semin Cardiothorac Vasc Anesth 2009.13: 138–145.
14. Dornia C, Philipp A, Bauer S, et al. D-dimers are a predictor of clot volume inside membrane oxygenators during extracorporeal membrane oxygenation
. Artif Organs 2015.39: 782–787.
15. Lehle K, Philipp A, Müller T, et al. Flow dynamics of different adult ECMO systems: a clinical evaluation. Artif Organs 2014.38: 391–398.
16. Gartner MJ, Wilhelm CR, Gage KL, Fabrizio MC, Wagner WR. Modeling flow effects on thrombotic deposition in a membrane oxygenator
. Artif Organs 2000.24: 29–36.
17. Gorbet MB, Sefton MV. Biomaterial-associated thrombosis
: roles of coagulation factors, complement, platelets and leukocytes
. Biomaterials 2004.25: 5681–5703.
18. van Beusekom HM, van der Giessen WJ, van Suylen R, Bos E, Bosman FT, Serruys PW. Histology after stenting of human saphenous vein bypass grafts: observations from surgically excised grafts 3 to 320 days after stent implantation. J Am Coll Cardiol 1993.21: 45–54.
19. Chang X, Gorbet M. The effect of shear on in vitro
platelet and leukocyte material-induced activation. J Biomater Appl 2013.28: 407–415.
20. Jobes DR. Safety issues in heparin and protamine administration for extracorporeal circulation. J Cardiothorac Vasc Anesth 1998.12(2 suppl 1): 17–20.
21. Zimmermann AK, Weber N, Aebert H, Ziemer G, Wendel HP. Effect of biopassive and bioactive surface-coatings on the hemocompatibility of membrane oxygenators. J Biomed Mater Res B Appl Biomater 2007.80: 433–439.
22. Hoenich NA. Platelet and leucocyte behaviour during haemodialysis. Contrib Nephrol 1999.125: 120–132.
23. Di Stasio E, De Cristofaro R. The effect of shear stress on protein conformation: Physical forces operating on biochemical systems: the case of von Willebrand factor. Biophys Chem 2010.153: 1–8.
24. Wang Y, Gallant RC, Ni H. Extracellular matrix proteins in the regulation of thrombus formation. Curr Opin Hematol 2016.23: 280–287.
25. Paul A, Straub A, Weber N, Ziemer G, Wendel HP. CD41 Western blotting: a new method to detect platelet adhesion
to artificial surfaces used in extracorporeal circulation procedures. J Mater Sci Mater Med 2009.20: 373–378.
26. Balasubramanian V, Slack SM. The effect of fluid shear and co-adsorbed proteins on the stability of immobilized fibrinogen and subsequent platelet interactions. J Biomater Sci Polym Ed 2002.13: 543–561.
27. Pasala S, Fiser RT, Stine KC, Swearingen CJ, Prodhan P. von Willebrand factor multimers in pediatric extracorporeal membrane oxygenation
support. ASAIO J 2014.60: 419–423.
28. Tauber H, Ott H, Streif W, et al. Extracorporeal membrane oxygenation
induces short-term loss of high-molecular-weight von Willebrand factor multimers. Anesth Analg 2015.120: 730–736.
29. Heilmann C, Geisen U, Beyersdorf F, et al. Acquired von Willebrand syndrome in patients with extracorporeal life support (ECLS). Intensive Care Med 2012.38: 62–68.
30. Linton PJ, Thoman ML. Immunosenescence in monocytes, macrophages, and dendritic cells: lessons learned from the lung and heart. Immunol Lett 2014.162(1 Pt B): 290–297.
31. Wenisch C, Patruta S, Daxböck F, Krause R, Hörl W. Effect of age on human neutrophil function. J Leukoc Biol 2000.67: 40–45.
32. Ginaldi L, De Martinis M, D’Ostilio A, Marini L, Loreto MF, Quaglino D. The immune system in the elderly: III. Innate immunity. Immunol Res 1999.20: 117–126.
33. Chiricolo M, Morini MC, Mancini R, Beltrandi E, Belletti D, Conte R. Cell adhesion
molecules CD11a and CD18 in blood monocytes in old age and the consequences for immunological dysfunction. Preliminary results. Gerontology 1995.41: 227–234.
34. Esparza B, Sanchez H, Ruiz M, Barranquero M, Sabino E, Merino F. Neutrophil function in elderly persons assessed by flow cytometry. Immunol Invest 1996.25: 185–190.
35. Noble JM, Ford GA, Thomas TH. Effect of aging on CD11b and CD69 surface expression by vesicular insertion in human polymorphonuclear leucocytes. Clin Sci (Lond) 1999.97: 323–329.
36. De Martinis M, Modesti M, Ginaldi L. Phenotypic and functional changes of circulating monocytes and polymorphonuclear leucocytes from elderly persons. Immunol Cell Biol 2004.82: 415–420.