Secondary Logo

Journal Logo

Biomedical Engineering

Twelve Hours In Vitro Biocompatibility Testing of Membrane Oxygenators

Bleilevens, Christian*; Grottke, Oliver*; Tillmann, Sabine*; Honickel, Markus*; Kopp, RüEdger; Arens, Jutta; Rossaint, Rolf*

Author Information
doi: 10.1097/MAT.0000000000000252
  • Free


Indications for lung support via extracorporeal membrane oxygenation (ECMO) in intensive care units are numerous. Neonatal, pediatric, and adult ECMO systems for patients with cardiopulmonary diseases, diaphragmatic hernia or pertussis, septic shock, acute respiratory distress syndrome, or as bridge to lung transplant for several weeks are available.1–4 Nevertheless, permanent ECMO support for weeks requires frequent exchange of the extracorporeal membrane devices because several adverse events, such as renal failure (52%), bacterial pneumonia (33%), hemorrhage (33%), hemolysis (18%), and thrombosis (10%), were recently described.5 These adverse events may reflect either the insufficient biocompatibility of the membrane lungs or the severe pathology of the patient’s disease.

It remains unclear why these complications occur after a prolonged time of extracorporeal circulation and what might trigger the complications. For systematic investigation of the molecular mechanisms leading to these events, stable extracorporeal circulation test settings are necessary, allowing analysis of possible underlying mechanisms as a result of insufficient biocompatibility of the membrane lungs during long-term use. Available in vitro studies compare the biocompatibility of different ECMO devices or surface coatings; however, these studies were limited to an extracorporeal circulation of 6 hr.6–9 This 6 hr period is sufficient for short time application testing of membrane lungs as it appears in surgical heart and lung procedures but is insufficient for the detection of the complications occurring during long-term use. At present, the long-term use of ECMO systems can only be tested in in vivo models, which is either an animal model or a treated patient, for approximately 30 days.10,11 Therefore, a stable extracorporeal circulation long-term test setting would enable the systematic investigation of the molecular mechanisms and additionally might help replace expensive and complex experimental in vivo animal models.

As a first step to increase the in vitro test time toward a prolonged test setting, this study was designed to prove the feasibility of operating an in vitro test system for membrane oxygenators (mock loop) under physiologic conditions (pH, flow, pressure, etc.) for at least 12 hr. In addition, the hematology, intensity of hemolysis, the platelet (PLT) activation status, the coagulation cascade, cytokine, and microparticle (MP) release were monitored over time as markers for limited biocompatibility.4,12–15

Materials and Methods

Extracorporeal Circuit (Mock Loop)

The mock loop was assembled as shown in Figure 1. The oxygenator resistance was calculated according to the method described by LaFayette et al.8

Figure 1
Figure 1:
An iD-pediatric membrane oxygenator (BIOLINE coating; MAQUET, Rastatt, Germany) (A) was connected via standard extracorporeal membrane oxygenation tubing (15 cm ¾ inch and 15 cm ½ inch) to the deltastream system (B) as a controller for the s.pump device (MEDOS, Stolberg, Germany) (C). The temperature was adjusted via a water bath thermostat (D). Pressures at the pump inlet/outlet (E) and the temperature were monitored (F), and the data were continuously collected on a PowerLab data acquisition system (PowerLab; ADInstruments, Oxford, United Kingdom) (G).

Blood Donation

Approximately 150 ml of whole blood were collected in a standard 450 ml blood bag (FENWALL, Lake Zurich, IL), containing citrate sodium hydrogen phosphate (5.66 mg/ml final concentration in the bag) from healthy male (n = 4) or female (n = 4) domestic pigs from different projects, all approved by the “Landesamt für Natur, Umwelt und Verbraucherschutz NRW.” Blood was withdrawn from arterial catheters. The mock loop was loaded with a total of 119 ± 1 ml citrated whole blood.

Experimental Groups

In the first setting, the mock loop was operated for 12 hr without further modifications (standard group), according to the manufacture’s recommendation for blood (0.6 L/min) and air flow (3 L/min). All blood samples (4.5 ml), withdrawn at defined time points, were replaced with fresh blood (4.5 ml) from the original blood bag to keep the volume constant. In the second setting, room air was replaced by CO2-enhanced gas (Linde AG, Aachen, Germany), and each time after withdrawing blood from the circuit, 4.5 ml of phosphate–adenine–glucose–guanosine–saline–mannitol (PAGGSM) nutrient solution15 was added additionally to the 4.5 ml of fresh blood from the blood bag to the mock loop (4 × 4.5 ml extra volume) (Table 1).

Table 1
Table 1:
Experimental Groups

Blood Sampling and Analysis

The 4.5 ml blood samples (five samples in each experiment: baseline [BL] from the blood bag, rest from the mock loop) were withdrawn at defined time points and analyzed immediately (blood gases, hemogram, thromboelastometry, and fluorescence-activated cell sorting (FACS)) or were centrifuged and stored at −80°C for subsequent analysis (coagulometry, cytokine screening, and enzyme-linked immunosorbent assay (ELISA)) (Table 2).

Table 2
Table 2:
Blood Sampling and Analysis

Thromboelastometry was performed according to a previously established measurement protocol for porcine blood.16

Flow cytometry was used to analyze cell surface expression of P-selectin (CD62P) as an activation marker on porcine CD61-positive PLTs, using fluorescein isothiocyanate (FITC)-conjugated and phycoerythrin (PE)-conjugated CD61 and CD62P or their appropriate isotype controls for 15 min at room temperature as described previously.17 After staining, PLTs were fixed for analysis (Cell Fix; BD Biosciences, Heidelberg, Germany).

Statistical Analysis

Data are presented as the mean ± SEM. Parameters were tested for significant differences between the groups at distinct time points by two-way analysis of variance (ANOVA) with Holm–Sidak correction for multiple comparisons and a confidence interval of 95%, and correlation of activated PLTs with circulating biomarkers was computed by Pearson’s correlation coefficient (r) in a two-tailed test setting with a confidence interval of 95% (Prism 6.04; GraphPad Software, La Jolla, CA). The effects of time or time and group were calculated by multivariate analysis for repeated measurements (IBM SPPS statistics, version 20; IBM Corp, Armonk, NY). The results were regarded as significantly different if the calculated p value was less than 0.05.


Mock Loop Operating Conditions

The initial blood flow was adjusted to 0.6 L/min. In spite of increasing the pump speed in the standard group, the flow decreased after 6 hr constantly to significantly lower values at the end of the experiment (Figure 2A). In contrast, the blood flow was maintained at 0.6 L/min despite a smaller increase of pump speed in the mod group.

Figure 2
Figure 2:
Pump speed (rpm), blood flow (L/min) (A), pressures at the pump inlet/outlet (mm Hg) (B), and the calculated resistance (mm Hg/L/min) (C). Effects of time (*p < 0.05, **p < 0.01) or time and group (§p < 0.05). Differences between the experimental groups (#p < 0.05, ##p < 0.01, ###p < 0.0001). BL, baseline.

The calculated resistance of the oxygenator changed from comparable values at BL to a significantly elevated resistance after 12 hr in the standard group (Figure 2C). The temperature was constant over time in the modified experiments in contrast to the standard group; however, these differences were not significant (data not shown).

Blood Gas Analysis

Within the first hour, the pH value in the standard group increased from pH 7.1 ± 0.1 to pH 8.0 ± 0.2 and, then, it decreased to pH 7.5 ± 0.1 within 12 hr. In the mod group, the pH remained significantly lower until the end of the experiment, but became acidotic after 6 hr (Figure 3A).

Figure 3
Figure 3:
pH value (A), sodium (Na+, mmol/L) and potassium (K+, mmol/L) concentration (B), free hemoglobin (mg/dl) (C), and partial carbon dioxide pressure (pCO2, mm Hg) (D). Effects of time (*p < 0.05, **p < 0.01) or time and group (§p < 0.05). Differences between the experimental groups (#p < 0.05, ##p < 0.01, ###p < 0.0001). BL, baseline.

Hypernatremia and hyperkalemia developed in the standard group in contrast to the mod group (Figure 3B).

Stronger hemolysis was indicated by significantly higher levels of free hemoglobin (Hgb) after 12 hr in the standard group (387.3 ± 141 mg/dl) in contrast to the modified experiments (62.5 ± 14.7 mg/dl) (Figure 3C).

Constantly higher pCO2 from the beginning to the end of the modified experiments was observed in comparison with the standard group (Figure 3D).

The counts for red blood cells (RBCs) and PLTs remained stable over time, whereas the white blood cell counts were significantly lower after 12 hr in the standard group (Table 3) (p < 0.05).

Table 3
Table 3:
RBC, WBC, PLT Counts, and Hgb Concentration Over Time

PLT Activation

After 6 hr, elevated levels of activated PLTs (CD62P-positive PLTs) were observed in the standard group (66.8% ± 4.1%) compared with the mod group (21.2% ± 1.4%) (Figure 4A), correlating with circulating interleukin (IL)-1β (Figure 4B), von Willebrand factor (vWF) (Figure 4C), and procoagulant MP concentration (Figure 4D). In the mod group, only the slightly elevated MP concentration correlated with CD62P-positive cells (Figure 4, D and E).

Figure 4
Figure 4:
Activated CD62P platelets (A). Correlation of CD62P platelets with interleukin (IL)-1β (ng/ml) (B), von Willebrand Factor (vWF, ng/ml) (C), procoagulant microparticles (MPs, nM) (D), and the total concentration of MPs (E). Effects of time (*p < 0.05, ***p < 0.001) or time and group (§§p < 0.01, §§§p < 0.0001). Differences between the experimental groups (#p < 0.05, ###p < 0.0001). BL, baseline.

Coagulation Analysis

The activated partial thromboplastin time (aPTT) increased in both groups during the initial phase of the experiments and peaked at significantly higher levels at 6 hr in the standard group compared with the mod group (Figure 5A).The intrinsic system clotting time (INTEM-CT) values exceeded the cutoff value (450 s) in both experimental groups (Figure 5B), whereas the cutoff for heparinase application clotting time (HEPTEM-CT) was only exceeded in the standard group (Figure 5C). Extrinsic system clotting time (EXTEM-CT) decreased after 1 hr in both experimental groups, followed by stabilization at a reduced level after 7 hr; no differences were observed between the groups after 12 hr (Figure 5D). The thrombin antithrombin (TAT) complex concentration decreased over time, without differences between the experimental groups (Figure 5E).

Figure 5
Figure 5:
Activated partial thromboplastin time (aPTT, s) and prothrombin time (PT, s) (A), clotting time of the intrinsic system (INTEM, s) (B), after heparinase application (HEPTEM, s) (C), and the extrinsic system (EXTEM, s) (D). Concentration of thrombin antithrombin (TAT) complexes (µg/L) (E). Effects of time (*p < 0.05, ***p < 0.001) or time and group (§p < 0.05, §§p < 0.01). Differences between the experimental groups (###p < 0.0001). BL, baseline.

Inflammatory Response

All cytokine levels were significantly elevated after 12 hr in the standard group compared with the mod group (Figure 6, A–C).

Figure 6
Figure 6:
Interleukin (IL)-1α (A), IL receptor antagonist (IL1ra) (B), and IL8 (C). Effects of time (*p < 0.05, ***p < 0.001) or time and group (§p < 0.05). Differences between the experimental groups (##p < 0.01, ###p < 0.0001). BL, baseline.

Parameters in the Blood Bag (Control)

The control values for the analysis performed after 12 hr from the eight blood bags that were used for the experiments are summarized in Table 4, and the values are plotted as the dotted lines in Figures 3–5.

Table 4
Table 4:
Analysis After 12 Hours From the Blood Bags


We aimed to prolong in vitro test settings for extracorporeal gas exchange membranes and could exceed the established 6 hr testing period for extracorporeal membrane gas exchange test circuits (mock loop) using a CO2-enhanced gas and small amounts of a nutrient solution.

Without modifications, the mock loop system could not be stably operated for more than 6 hr. The flow rate could not be kept at a constant level, and increasing the pump speed resulted in foam formation and unsuitable flow conditions and increasing oxygenator resistance over time. An increase in CD62P-positive PLTs, correlating with increasing IL-1β, vWF, and procoagulant MP concentrations, was observed over time, and additionally, the extrinsic coagulation pathway was highly activated, whereas a prolongation of the clotting time of the intrinsic system was observed. Proinflammatory cytokines IL-1α, IL1ra, and IL8 were strongly increased after 12 hr. However, our results in the unmodified setting were comparable with the sparse available data for in vitro tests with porcine blood during the first 6 hr, concerning the hemolysis18 and the oxygenator resistance.8

In contrast, the modified system was stable for 12 hr with constant pump speed, pressures, and flow rate. No initial pH increase, PLT activation, or cytokine release were detected. Furthermore, the results concerning the coagulation system and PLT counts, as well as pCO2 and pH values, were comparable with in vivo settings after 12 hr.19,20 These findings show that conditions comparable with in vivo settings could be achieved in vitro. Nonetheless, further investigations on the influence of the nutrient solution on the rheological properties of the blood (e.g., the dynamic viscosity) and also on the pH values are necessary.

Comparison of Modified Versus Unmodified

The extensive respiratory alkalosis, which is known to be associated with enhanced PLT aggregation and activation,21 was prevented in the modified setting by the application of a CO2-enhanced gas. Consistent with previous studies, we found that increasing levels of vWF correlated with increasing amounts of CD62P-PLTs in the standard group.22 Furthermore, we detected significantly increased levels of IL-1α, IL1ra, and IL8 in the standard group, indicating leukocyte activation and IL-1β, which might be secreted by activated leukocytes or PLTs. Thus, subsequent to the activation of the coagulation system, an inflammatory response was observed that might be triggered by activated PLTs.23

Although the additional volume of 18 ml nutrient solution in the modified experiments suggests a dilatational effect, constant Hgb levels, RBC counts, and a stable filling status of the modified mock loop was observed. In contrast, we measured increasing Hgb levels in the unmodified system, despite the substitution of 4.5 ml fresh blood after each sampling, indicating fluid loss over time. A fluid loss of 0.046 ml/min by evaporation via the oxygenator membrane was recently described to correlate with every liter per minute of gas flow at 37°C.24 Although our primarily intention was the support of circulating blood cells with this stabilizing nutrient solution, we obviously covered the fluid loss over time by adding additional volume, without diluting the blood. Furthermore, PAGGSM provides improved preservation of RBCs, including decreased hemolysis and reduced shedding of MPs from RBCs.25 As we found significantly lower amounts of MPs, as well as significantly decreased hemolysis in our modified system, this might be regarded as an effect of PAGGSM solution.

Another observation was the enormous prolongation of the INTEM-CT values in both experimental groups, indicating a blocked intrinsic cascade, triggered by the release of heparin from the BIOLINE surface coating of the membrane oxygenator (activated factor X [Xa] inhibitor). Unfractionated heparin (UFH) was detected in the circuit, as confirmed by the HEPTEM-CT measurements, and by increasing aPTT as a reliable predictor for circulating UFH.26 These findings generally prove the clinical benefits that MAQUET predicts for their oxygenator: reduced clotting activity (decreased TAT complexes), reduced PLT adhesion, and generated thrombin.

In contrast, the extrinsic system was activated after the beginning of the experiment in both groups, as indicated by decreased EXTEM-CT values. This pathway depends highly on tissue factor (TF), which is predominantly released by endothelial cells, fibroblasts, and smooth muscle cells. As none of these cell types is present in the mock loop, MPs released from PLTs, leukocytes, or erythrocytes might represent the TF source. We detected increasing amounts of MPs in both test settings, with significantly higher concentrations in the unmodified group, possibly because of higher shear stress, leading to the shearing-off of MPs from cell surfaces.27 MPs from these cell types are known to contribute to a procoagulant state, as they carry TF, phosphatidylserine residues, and P-selectin glycoprotein ligands.13,28,29 In addition, the MP release correlates with the amount of CD62P-PLTs. Thus, we have two conditions (PLT activation and/or high shear stress) that might lead to MP release in our system. The lower amount of MPs in the mod group might still be adequate for the observed extrinsic coagulation activation. Microparticles were reported to have a 50 fold to 100 fold higher procoagulant activity than activated PLTs themselves.30 Delayed thrombus formation on oxygenator membranes in clinical ECMO applications could possibly be explained by extenuating heparin activity, provided by the membrane coating, over time. Losing heparin activity means activation of the intrinsic pathway. As this is paralleled by permanent activation of the extrinsic pathway, possibly through MPs, the clotting cascade could be initiated even after days or weeks when the intrinsic pathway is no longer blocked by heparin.

The major limitation we must mention is that the experiments were carried out with porcine blood instead of human blood, as this is a well-accepted model throughout in vivo ECMO research. Consequently, the findings’ translation to clinical relevance is still limited.


We demonstrated that the modified in vitro ECMO testing enables physiologic conditions for more than 6 hr by applying a CO2-enhanced gas as an important buffer and repetitively adding 4.5 ml nutrient solution as extra volume to cover fluid loss resulting from evaporation. Thus, hemolysis, PLT activation, and cytokine release were prevented, and the results of the modified experiments are partly comparable with those from in vivo settings for at least 12 hr of extracorporeal circulation. Independently from the modifications, we detected a blockade of the intrinsic coagulation pathway, whereas the extrinsic coagulation was permanently activated. Further experiments are needed to study the procoagulant properties of MPs in ECMO systems should be studied specifically and to elucidate whether a mock loop system could be operated for more than 12 hr to reduce or replace in vivo studies.


1. Gupta P, McDonald R, Chipman CW, et al. 20-year experience of prolonged extracorporeal membrane oxygenation in critically ill children with cardiac or pulmonary failure. Ann Thorac Surg. 2012;93:1584–1590
2. Rambaud J, Guilbert J, Guellec I, Renolleau S. A pilot study comparing two polymethylpentene extracorporeal membrane oxygenators. Perfusion. 2013;28:14–20
3. Hirshberg E, Miller RR 3rd, Morris AH. Extracorporeal membrane oxygenation in adults with acute respiratory distress syndrome. Curr Opin Crit Care. 2012;1:36–43
4. Nolan H, Wang D, Zwischenberger JB. Artificial lung basics: Fundamental challenges, alternative designs and future innovations. Organogenesis. 2011;7:23–27
5. Zangrillo A, Landoni G, Biondi-Zoccai G, et al. A meta-analysis of complications and mortality of extracorporeal membrane oxygenation. Crit Care Resusc. 2013;15:172–178
6. Mulla H, Lawson G, von Anrep C, et al. In vitro evaluation of sedative drug losses during extracorporeal membrane oxygenation. Perfusion. 2000;15:21–26
7. 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
8. LaFayette NG, Schewe RE, Montoya JP, Cook KE. Performance of a MedArray silicone hollow fiber oxygenator. ASAIO J. 2009;55:382–387
9. Gillogly A, Kilbourn C, Waldvogel J, Martin J, Annich G, Wagner D. In vitro clearance of intravenous acetaminophen in extracorporeal membrane oxygenation. Perfusion. 2013;28:141–145
10. Sato H, Hall CM, Lafayette NG, et al. Thirty-day in-parallel artificial lung testing in sheep. Ann Thorac Surg. 2007;84:1136–43
11. Wu ZJ, Zhang T, Bianchi G, et al. Thirty-day in-vivo performance of a wearable artificial pump-lung for ambulatory respiratory support. Ann Thorac Surg. 2012;93:274–281
12. Li C, Li J, Li Y, et al. Crosstalk between platelets and the immune system: Old systems with new discoveries. Adv Hematol. 2012;2012:384685
13. Owens AP 3rd, Mackman N. Microparticles in hemostasis and thrombosis. Circ Res. 2011;108:1284–1297
14. Ota K. Advances in artificial lungs. J Artif Organs. 2010;13:13–16
15. Veale MF, Healey G, Sparrow RL. Effect of additive solutions on red blood cell (RBC) membrane properties of stored RBCs prepared from whole blood held for 24 hours at room temperature. Transfusion. 2011;51(suppl 1):25S–33S
16. Grottke O, Braunschweig T, Spronk HM, et al. Increasing concentrations of prothrombin complex concentrate induce disseminated intravascular coagulation in a pig model of coagulopathy with blunt liver injury. Blood. 2011;118:1943–1951
17. Shattil SJ, Cunningham M, Hoxie JA. Detection of activated platelets in whole blood using activation-dependent monoclonal antibodies and flow cytometry. Blood. 1987;70:307–315
18. Arens J, Schnoering H, Pfennig M, et al. The Aachen MiniHLM—A miniaturized heart-lung machine for neonates with an integrated rotary blood pump. Artif Organs. 2010;34:707–713
19. Kopp R, Bensberg R, Henzler D, et al. Hemocompatibility of a miniaturized extracorporeal membrane oxygenation and a pumpless interventional lung assist in experimental lung injury. Artif Organs. 2010;34:13–21
20. Jialiang S, Juanhong S, Qiyi C, et al. In-line hemofiltration minimized extracorporeal membrane oxygenation-related inflammation in a porcine model. Perfusion. 2014;6:526–533
21. Baumgarten A, Wilhelmi M, Ganter M, Rohn K, Mischke R. Changes of platelet function and blood coagulation during short-term storage of CPDA-1-stabilised ovine blood. Res Vet Sci. 2011;91:150–158
22. Jenne CN, Urrutia R, Kubes P. Platelets: Bridging hemostasis, inflammation, and immunity. Int J Lab Hematol. 2013;35:254–261
23. Semple JW, Italiano JE Jr, Freedman J. Platelets and the immune continuum. Nat Rev Immunol. 2011;11:264–274
24. Li Li C, Oi Yan T, Ming Chit Arthur K, Hoi Ping S, King Chung Kenny C, Wing Wa Y. Insensible water loss through adult extracorporeal membrane oxygenation circuit: An in vitro study. ASAIO J. 2014;60:508–512
25. Sparrow RL. Time to revisit red blood cell additive solutions and storage conditions: A role for ‘‘omics’’ analyses. Blood Transfus. 2012;10:7–11
26. Maul TM, Wolff EL, Kuch BA, Rosendorff A, Morell VO, Wearden PD. Activated partial thromboplastin time is a better trending tool in pediatric extracorporeal membrane oxygenation. Pediatr Crit Care Med. 2012;13:e363–e371
27. Miyazaki Y, Nomura S, Miyake T, et al. High shear stress can initiate both platelet aggregation and shedding of procoagulant containing microparticles. Blood. 1996;88:3456–3464
28. Italiano JE Jr, Mairuhu AT, Flaumenhaft R. Clinical relevance of microparticles from platelets and megakaryocytes. Curr Opin Hematol. 2010;17:578–584
29. Koshiar RL, Somajo S, Norström E, Dahlbäck B. Erythrocyte-derived microparticles supporting activated protein C-mediated regulation of blood coagulation. PLoS One. 2014;9:e104200
30. Sinauridze EI, Kireev DA, Popenko NY, et al. Platelet microparticle membranes have 50- to 100-fold higher specific procoagulant activity than activated platelets. Thromb Haemost. 2007;97:425–434

extracorporeal membrane oxygenation; in vitro; biocompatibility; 12 hr

Copyright © 2015 by the American Society for Artificial Internal Organs