Tethered Liquid Perfluorocarbon Coating for 72 Hour Heparin-Free Extracorporeal Life Support

Abstract

Coagulopathic complications during extracorporeal life support (ECLS) result from two parallel processes: 1) foreign surface contact and shear stress during blood circulation and 2) administration of anticoagulant drugs to prevent circuit thrombosis. To address these problems, biocompatible surfaces are developed to prevent foreign surface-induced coagulopathy, reducing or eliminating the need for anticoagulants. Tethered liquid perfluorocarbon (TLP) is a nonadhesive coating that prevents adsorption of plasma proteins and thrombus deposition. We examined application of TLP to complete ECLS circuits (membranes, tubing, pumps, and catheters) during 72 hours of ECLS in healthy swine (n = 5/group). We compared TLP-coated circuits used without systemic anticoagulation to standard of care: heparin-coated circuits with continuous heparin infusion. Coagulopathic complications, device performance, and systemic effects were assessed. We hypothesized that TLP reduces circuit thrombosis and iatrogenic bleeding, without impeding gas exchange performance or causing untoward effects. No difference in bleeding or thrombotic complication rate was observed; however, circuit occlusion occurred in both groups (TLP = 2/5; CTRL = 1/5). TLP required elevated sweep gas rate to maintain normocapnia during ECLS versus CTRL (10–20 vs. 5 L/min; p = 0.047), suggesting impaired gas exchange. Thrombus deposition and protein adhesion on explanted membranes were comparable, and TLP did not preserve platelet or blood cell counts relative to controls. We conclude that neither TLP nor standard of care is an efficacious solution to prevent coagulation disturbances during ECLS. Further testing of promising biomaterials for ECLS utilizing the model outlined here is warranted.

Thrombosis, coagulopathy, and inflammation are major issues associated with the use of blood-contacting medical devices. This challenge is apparent during extracorporeal life support (ECLS) where the following concomitant conditions cause significant damage to blood cells: 1) exposure of blood to foreign materials, 2) shear stress and hemolysis from the blood pump, and 3) turbulent flow caused by varied circuit configurations and catheter dimensions. When blood contacts the circuit, protein adsorption and contact pathway activation rapidly occur. This is accompanied by platelet activation, adhesion, and thrombus formation.¹ Thrombi in the circuit impede flow, increase circuit resistance, and reduce the performance of the device, such as inhibiting gas diffusion across the membrane lung during extracorporeal membrane oxygenation (ECMO).² Simultaneously, platelet consumption, activation of clotting factors, and hemolysis cause risk of consumptive coagulopathy and bleeding.³

The risk of thrombotic complications during ECLS prompts providers to initiate systemic anticoagulation therapy; however, there are no universal guidelines for anticoagulant administration during ECLS, resulting in variable, center-specific protocols.^3,4 Maintaining therapeutic anticoagulation amidst circuit-induced thrombosis, depletion of coagulation factors, and frequent transfusions is extremely challenging. The result is frequent bleeding complications, which can occur at the cannulation site, concomitant wounds, airway or gastrointestinal tract, and can even cause lethal cerebral hemorrhage.⁵

An alternative to systemic anticoagulation is to prevent foreign surface-induced thrombosis using biocompatible surface coatings. Modern ECLS devices utilize biocompatible coatings such as immobilized heparin; however, currently, these coatings alone are insufficient to prevent circuit thrombosis, and thus systemic anticoagulation remains standard practice.⁶ Other investigational coatings have shown promise in the laboratory setting; however, they have been primarily tested when applied to circuit tubing only or to segments of deconstructed membranes and catheters, rather than standard, clinically utilized devices. Additionally, these materials are evaluated using static benchtop tests or in vivo in an arteriovenous shunt model, which does not account for the effects of the blood pump or adequately simulate the range of flows for cardiopulmonary (2–7 L/min) or partial respiratory support (500 ml–2 L/min). Although such studies are useful for initial evaluation and optimization of these coatings, to fully understand the functionality for ECLS, multiday studies that closely simulate the clinical care scenario, such as outlined here, are necessary.

One promising biocompatible coating for ECLS is an “omniphobic” surface called tethered liquid perfluorocarbon (TLP).⁷ Tethered liquid perfluorocarbon is a bilayer coating developed to prevent adhesion of plasma proteins, cells, bacteria or any biologic medium to polymer substrates. The TLP bilayer coating consists of: 1) a tethered perfluorocarbon: an immobilized perfluorinated layer covalently bound to an underlying substrate and 2) liquid perfluorocarbon: a mobile, liquid surface layer retained on the substrate by the tether layer, imparting antiadhesive properties to the polymer substrate. Leslie et al.,⁷ demonstrated that TLP prevents attachment of thrombi and pathogenic microbes to the surface of an array of medical grade materials, including plastics, glasses, and metals. We previously demonstrated that TLP exerts antithrombogenic effects in human blood ex vivo; and that retention of the lubricant on the substrate is necessary for these effects.⁸ Furthermore, TLP applied to polyvinyl chloride (PVC) perfusion tubing showed superior performance to uncoated PVC in an 8 hour arteriovenous shunt model in unheparinized swine; however, the flow rates utilized were significantly lower than for ECLS, and there was no inclusion of a blood pump, catheter or membrane in the simplified tubing loop.⁷

Following these promising results, we performed the first evaluation of TLP coating applied tip-to-tip to standard ECLS circuits (catheter, tubing, blood pump, and membrane).⁹ We compared TLP with heparin-coated ECLS circuits from the manufacturer in vivo in swine during 6 hours of veno-venous (VV) ECLS without administration of anticoagulants. TLP permitted heparin-free circulation without causing untoward effects and reduced thrombus deposition on the membrane compared with controls. These results, however, cannot be extrapolated to a multiday timeline.

In this current study we performed a 72 hour evaluation of TLP applied tip-to-tip to a low-flow (<2 L/min blood flow rate) ECLS device for partial respiratory support. We compared TLP-coated devices used without systemic anticoagulation to the standard of care (heparin-coated circuits with continuous heparin infusion) in healthy, mechanically ventilated swine. We assessed coagulation, thrombus deposition, gas exchange efficiency of the membrane lung, systemic response, and end-organ damage (Figure 1). The objective was to evaluate if TLP is safe and efficacious while permitting heparin-free ECLS more than 72 hours. We hypothesized that TLP reduces thrombotic and bleeding complications relative to standard of care; all without altering gas exchange performance of the membrane or causing untoward systemic effects.

Figure 1.: Schematic of study design for evaluation of TLP coating during 72 hours extracorporeal circulation in healthy swine. A: Depiction of TLP bilayer coating components including immobilized-tether layer and mobile liquid surface layer. B: Timeline of key data collection points is shown. C: Key outcome measures to assess efficacy and biocompatibility of the TLP surface coating which includes coagulation profile, thrombus deposition on circuit components, gas exchange performance of the membrane lung (addition of oxygen [O₂] and removal of carbon dioxide [CO2]), and systemic response including vitals and histology of the lungs (shown), kidney, liver and jejunum. BL, baseline; post-ECLS, after initiation of extracorporeal life support; TLP, tethered liquid perfluorocarbon.

Materials and Methods

This study was carried out in compliance with the Animal Welfare Act, implementing Animal Welfare Regulations, and the principles of the Guide for the Care and Use of Laboratory Animals, National Research Council. An Institutional Animal Care and Use Committee (Bridge PTS, San Antonio, Texas) approved all research conducted (Protocol BRIDGE PTS-17-08). Bridge PTS is fully accredited by Association for the Assessment and Accreditation of Laboratory Animal Care.

Equipment

An ECLS circuit compatible with the CARDIOHELP System (Maquet/Getinge; Rastatt, Germany) with permitted blood flow between 0.2 and 2.8 L/min was utilized. Circuits consisted of a ROTASSIST 2.8 centrifugal blood pump with integrated pressure gauge, a membrane lung (0.98 m² surface area) with polymethylpentene membrane fibers, and ¼″ tube connections. Avalon Elite 19 Fr bicaval dual-lumen catheters (Maquet/Getinge; Rastatt, Germany) were used for VV-ECLS. Five TLP-coated and five heparin-coated devices were available for this study. In the TLP group, ECLS circuits (tubing, pump, and membrane) and catheters as provided by manufacturer were coated tip-to-tip with TLP (see Supplemental Methods, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561). In the control group (CTRL), circuits with BIOLINE coating (immobilized albumin with covalently bound heparin) from the manufacturer were used with standard catheters.

Instrumentation

Female Yorkshire swine (CTRL = 48.6 ± 4.6 kg; TLP = 45.3 ± 2.6 kg; p > 0.05) were anesthetized, intubated, and volume-control ventilated with room air. Arterial lines were placed for monitoring arterial blood pressure and collection of systemic blood samples. Venous lines were placed for administration of fluid and continuous total intravenous anesthesia medications (see Supplemental Methods, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561). Animals received surgical plane anesthesia and analgesia for the entirety of the study. Continuous electrocardiogram, blood pressure and oxyhemoglobin saturation (SpO₂) were monitored using the Dräger Infinity M540 Monitor (Drägerwek AG; Lübeck, Germany).

Extracorporeal Life support

A 2000 U unfractionated heparin bolus was administered before cannulation per manufacturer’s recommendation. The Avalon catheter was placed percutaneously in the right jugular vein and connected to the ECLS system in an air-free fashion. Veno-venous extracorporeal life support was initiated with blood flow rate of 1 L/min, and sweep gas (100% O₂) flow of 5 L/min. After cannulation, continuous heparin infusion was initiated in CTRL group only (40 U/kg/h) and titrated to maintain an ACT of 50% above baseline (BL) value. Ventilator settings were incrementally adjusted to VT of 6 ml/kg (see Supplemental Methods, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561). Data were collected for 72 hours unless the ECLS circuit became occluded inhibiting flow [pump revolutions per minute (RPM) approaching limit of 4,800–5,000 with decline in flow below 0.5 L/min], at which point the study was terminated. All vital signs, point-of-care tests and blood laboratories were taken at BL, immediately post-initiation of ECLS (post-ECLS), and at 3, 6, 12, 24, 48, and 72 hours post-ECLS. If initial ECLS settings were insufficient to support the clinical condition of the animal (arterial partial pressure of CO2 [P_aCO2] 35–45 mm Hg; arterial partial pressure of O₂ [P_aO₂] > 60 mm Hg), sweep gas and blood flow were increased accordingly. Time to onset of required adjustments was reported. If required, vasopressors (epinephrine) were administered for blood pressure support and a loop diuretic (Furosemide, 0.5 mg) was given for oliguria (urine output < 0.5 ml/kg/h). Extracorporeal life support circuit variables including arterial pressure (P_art), venous pressure (P_ven), internal membrane pressure (P_int), pressure drop across the membrane (ΔP), and pump RPM were recorded.

Vital Signs, Respiration, and Chemistry

Vital signs and respiratory measures included heart rate, mean arterial blood pressure, SpO₂, respiratory rate, CO2 production (VCO2), minute ventilation, plateau pressure, compliance, pulmonary resistance and Et CO2. Blood gases [pH, partial pressure CO2 (PCO2) and O₂ (PO2), oxygen saturation (SO2), and lactate] were collected from the systemic circulation, as well as immediately before and after the ECLS membrane for determination of CO2 removal, O₂ transfer, oxygenation index and PaO₂-to-F_iO₂ ratio (PFR) (see Supplemental Methods, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561).

Blood Laboratories and Coagulation

Complete blood count was performed using the ADVIA 2120 (Siemens; Munich, Germany). Coagulation testing included prothrombin time (PT), activated partial thromboplastin time (aPTT), antithrombin III, von Willebrand factor (vWF), and concentrations of fibrinogen and D-dimer measured in platelet poor plasma (PPP) using the STA Compact Max (Stago; Parsippany, New Jersey) as previously reported.^10,11 Activated clotting time (ACT) was measured using the Hemochron Signature Elite Microcoagulation System (Accriva Diagnostics; San Diego, California). Thromboelastography (TEG) was performed using the TEG 5000 Hemostasis Analyzer (Haemonetics Corp.; Boston, Massachusetts) with citrated kaolin assay and heparinase cups. The following variables were recorded: reaction time (R), initial clot formation time (K), clot strength (MA), and lysis (Ly60). Platelet aggregation was measured by whole-blood impedance aggregometer (Multiplate Analyzer, Roche Diagnostics; Basel, Switzerland) using adenosine diphosphate (ADP) (10.75 µM; ADPtest, Roche Diagnostics; Basel, Switzerland) and collagen (3.2 µg/mL; Helena Laboratories; Beaumont, Texas) as activators.¹⁰ Plasma free hemoglobin (PFHb) was measured by direct spectrophotometry.¹² To assess systemic inflammatory response, cytokines (IL-1β, IL-6, IL-8, IL-10, and TNF-α) were analyzed in plasma using a MILLIPLEX Magnetic Bead Panel (Millipore Sigma; Burlington, Massachusetts) and Luminex 200 Analyzer (Luminex Corporation; Austin, Texas). To assess endothelial damage, syndecan-1 expression in plasma was measured using a commercial kit (SEB966PO, Cloud-Clone Corp.; Katy, Texas) (See Supplemental Methods, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561).

End of Study Procedures

Animals underwent deep anesthesia and euthanasia by exsanguination to prevent postmortem clotting caused by routine euthanasia medications. Extracorporeal life support circuits and catheters were explanted and processed for analysis via field emission scanning electron microscopy (FESEM).⁹ Total thrombus area and protein adhesion on tubing and catheters were assessed (see Supplemental Methods, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561). Tissue samples were collected from lungs, kidney, liver, and jejunum for histological assessment (see Supplemental Methods, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561).¹³

Statistics

Statistics were performed using SAS 9.4 (Cary, North Carolina). All tests were two-sided with alpha ≤0.05 for significance. Groups were tested independently using one-way mixed models with repeated measures and Dunnett adjustment to evaluate changes from BL. Between-group differences were examined using a two-way mixed model with repeated measures and Tukey adjustment. A Fisher’s exact test was used to assess complication rates. Survival analysis was performed with log-rank test for significance.

Results

We completed the first multiday evaluation of heparin-free ECLS using TLP applied to standard ECLS circuits. Two TLP-coated circuits occluded before 72 hours (50 ± 18 hours) requiring early study termination; and two TLP animals died before 72 hours (one due to pneumothorax and one after rapid desaturation/destabilization unresponsive to vasopressors) (for TLP at 24 hours, n = 5; at 48 hours, n = 3; at 72 hours n = 1). In the CTRL group where heparin-coated ECLS circuits were used with continuous heparin infusion, all animals survived to end of study with one circuit that occluded at the 72 hour mark. Mean survival time was 57 ± 13 hours in TLP versus 72 ± 0 hours in CTRL (p = 0.026).

Vital signs and respiratory parameters are reported in Table 1. PaCO₂ was reduced versus BL in both groups at post-ECLS, which persisted until 72 hours in CTRL only. Minute ventilation was reduced in both groups following ventilator setting decrements made after the post-ECLS time point. The respiratory rate was reduced versus BL in CTRL from 3 to 72 hours; but was only reduced in TLP from 3 to 12 hours. Control was primarily supported by room air; whereas, TLP required a significant increase in FiO₂ beginning at 12 hours. Within hours of ECLS initiation, plateau pressures were higher in TLP and pulmonary compliance was reduced compared with CTRL. Pulmonary resistance increased in both groups. At 12 hours, oxygenation index was elevated in TLP versus CTRL, and PFR was significantly reduced from BL in TLP. There was no difference between groups in percent reduction of PCO2 across the membrane or oxygen transfer (Figure 2).

Table 1. - Vital Signs and Respiratory Measurements

Variable	Group	Baseline	Post-ECLS	3 Hours	6 Hours	12 Hours	24 Hours	48 Hours	72 Hours
HR (bpm)	CTRL	102 ± 5	103 ± 6	96 ± 5	90 ± 8	77 ± 2Table 1.	73 ± 6Table 1.	76 ± 5Table 1.	78 ± 6Table 1.
	TLP	113 ± 4	118 ± 6	99 ± 4	94 ± 8	82 ± 4Table 1.	94 ± 10	105 ± 13	63
MAP (mm Hg)	CTRL	73 ± 4	73 ± 4	79 ± 3	83 ± 5	83 ± 4	86 ± 3	83 ± 4	82 ± 4
	TLP	80 ± 4	70 ± 3	82 ± 3	83 ± 3	82 ± 3	83 ± 5	78 ± 3	90
Arterial pH	CTRL	7.49 ± 0.02	7.60 ± 0.02Table 1.	7.54 ± 0.03	7.52 ± 0.01	7.53 ± 0.01	7.52 ± 0.01	7.46 ± 0.02	7.47 ± 0.02
	TLP	7.49 ± 0.01	7.51 ± 0.02	7.49 ± 0.02	7.50 ± 0.02	7.49 ± 0.01	7.40 ± 0.07	7.49 ± 0.03	7.53
PaO2 (mm Hg)	CTRL	81 ± 2	92 ± 6	78 ± 3	84 ± 5	80 ± 5	82 ± 2	85 ± 6	83 ± 8
	TLP	87 ± 4	95 ± 5	88 ± 4	92 ± 10	83 ± 2	77 ± 6	76 ± 9	96
PaCO2 (mm Hg)	CTRL	41 ± 2	26 ± 1Table 1.	29 ± 1Table 1.	32 ± 2Table 1.	28 ± 1Table 1.	28 ± 1Table 1.	33 ± 1Table 1.	34 ± 2Table 1.
	TLP	42 ± 1	27 ± 3Table 1.	33 ± 3	36 ± 2	35 ± 2	37 ± 3	38 ± 4	34
MV (L / min)	CTRL	7.6 ± 0.8	7.8 ± 0.8	4.0 ± 0.3Table 1.	4.2 ± 0.5Table 1.	3.9 ± 0.4Table 1.	4.2 ± 0.5Table 1.	4.1 ± 0.5Table 1.	4.4 ± 0.2Table 1.
	TLP	8.6 ± 0.3	8.6 ± 0.3	4.2 ± 0.3Table 1.	4.6 ± 0.3Table 1.	4.7 ± 0.3Table 1.	5.5 ± 0.7Table 1.	5.1 ± 1.0Table 1.	4.2
RR(breaths / min)	CTRL	16 ± 1	16 ± 1	10 ± 1Table 1.	10 ± 1Table 1.	10 ± 1Table 1.	10 ± 1Table 1.	10 ± 1Table 1.	10 ± 2Table 1.
	TLP	19 ± 1	19 ± 1	14 ± 2Table 1.	12 ± 1Table 1.	12 ± 1Table 1.	16 ± 3	19 ± 7	10
FiO2 (%)	CTRL	21 ± 0	21 ± 0	21 ± 0	22 ± 1	22 ± 1	22 ± 1	24 ± 2	30 ± 5
	TLP	21 ± 0	21 ± 0	23 ± 2	25 ± 2	27 ± 4Table 1.	53 ± 19Table 1.	62 ± 22Table 1.	25Table 1.
Et CO2 (mm Hg)	CTRL	38 ± 1	23 ± 1Table 1.	32 ± 1Table 1.	34 ± 2	31 ± 2Table 1.	29 ± 1Table 1.	34 ± 2	32 ± 2Table 1.
	TLP	41 ± 2	28 ± 2Table 1.	33 ± 3	29 ± 4Table 1.	31 ± 2	28 ± 5Table 1.	33 ± 1	35
Plateau pressure (cm H2O)	CTRL	18 ± 1	20 ± 1	16 ± 1Table 1.†	17 ± 1Table 1.Table 1.	18 ± 1†	18 ± 1†	18 ± 1†	22 ± 2
	TLP	18 ± 1	22 ± 1Table 1.	18 ± 1	20 ± 1Table 1.	24 ± 2Table 1.	27 ± 5Table 1.	24 ± 3Table 1.	20
Compliance (mL/cm H2O)	CTRL	36 ± 1	36 ± 2Table 1.	37 ± 1	34 ± 1	31 ± 1Table 1.Table 1.	30 ± 2Table 1.Table 1.	30 ± 1Table 1.Table 1.	24 ± 4Table 1.
	TLP	38 ± 2	29 ± 2Table 1.	32 ± 2	27 ± 3Table 1.	21 ± 2Table 1.	18 ± 4Table 1.	16 ± 5Table 1.	29
Resistance (cm H2O/L/s)	CTRL	6.2 ± 0.1	6.3 ± 0.1	6.9 ± 0.2Table 1.	7.2 ± 0.2Table 1.	7.5 ± 0.4Table 1.	7.8 ± 0.4Table 1.	8.3 ± 0.4Table 1.	10.2 ± 0.7Table 1.
	TLP	6.1 ± 0.2	6.7 ± 0.2	7.2 ± 0.3Table 1.	7.6 ± 0.3Table 1.	9.2 ± 1.0Table 1.	11.4 ± 2.4Table 1.	10.8 ± 0.9Table 1.	8.5
Oxygenation index	CTRL	2.5 ± 0.1	2.3 ± 0.2	2.6 ± 0.1	2.6 ± 0.1†	2.8 ± 0.1Table 1.Table 1.	2.9 ± 0.2	3.1 ± 0.2Table 1.Table 1.	5.0 ± 1.1Table 1.
	TLP	2.3 ± 0.2	2.2 ± 0.2	2.7 ± 0.5	3.4 ± 0.4Table 1.	4.3 ± 0.7Table 1.	17.5 ± 10.5Table 1.	22.0 ± 11.4Table 1.	2.86
PFR	CTRL	387 ± 10	438 ± 27	373 ± 16	384 ± 14	369 ± 20	379 ± 10	360 ± 13	301 ± 47Table 1.
	TLP	413 ± 18	454 ± 22	390 ± 24	376 ± 24	322 ± 35Table 1.	255 ± 79Table 1.	215 ± 84Table 1.	384

Mean ± standard deviation in CTRL versus TLP (n = 5/group).

Significance p < 0.05.

*Significant change from baseline.

†Significant difference between CTRL and TLP groups.

CTRL, control; ECLS, extracorporeal life support; E_t CO₂, end-tidal carbon dioxide; F_iO₂, fraction inspired oxygen; HR, heart rate; MAP, mean arterial pressure; MV, minute ventilation; PaO₂, partial pressure arterial oxygen; PaCO₂, partial pressure arterial carbon dioxide; PFR, P_aO₂-to-F_iO₂ ratio; RR, respiratory rate; TLP, tethered liquid perfluorocarbon.

Figure 2.: Figure represents gas exchange efficiency of membrane lungs with TLP (n = 5) coating versus immobilized-heparin CTRL coating (n = 5). Time points are post-initiation of ECLS (PE) and at 3–72 hours after the start of circulation. A and B: Bars represent means ± standard deviation of partial pressure of carbon dioxide (PCO2) and partial pressure of oxygen (PO2), respectively, for blood gases drawn from the pre-membrane (PRE) venous ECLS circuit line and post-membrane (POST) circuit line located immediately before and after blood passage through the membrane lung. C: Mean ± standard deviation of percent decrease in PCO2 from pre-membrane to post-membrane blood samples. D: Mean ± standard deviation oxygen transfer (ml/min) measured from blood samples directly before and after passage through the membrane lung in the extracorporeal circuit. A two-sided test was performed with p <0.05 accepted for significance. *Significant difference between pre-membrane and post-membrane values in CTRL. **Significant difference between pre-membrane and post-membrane values in TLP. CTRL, control; ECLS, extracorporeal life support; TLP, tethered liquid perfluorocarbon.

Complete blood count and coagulation values are shown in Table 2. Red cell count and hemoglobin concentration dropped significantly in both groups after start of ECLS. Platelet count was similar between groups and decreased during ECLS (at 24–72 hours, 44–56% decrease in CTRL, 50–58% decrease in TLP). aPTT and ACT were elevated in CTRL which received heparin. Plasma-free hemoglobin concentration was significantly elevated in TLP versus CTRL beginning at 12 hours but was below levels indicative of clinical hemolysis (PFHb > 50 g/dl),¹⁴ except in the one TLP animal that survived till the end of study. D-dimer was not different from BL in either group (Figure 3A) but was elevated above a diagnostic threshold for suspected venous thromboembolism (>0.5 µg/mL)¹⁵ in two animals: 1) CTRL animal with circuit occlusion at 72 hours and 2) TLP animal that rapidly destabilized and expired early. Antithrombin III activity was reduced from 3 to 72 hours in CTRL, accompanied by an increased heparin infusion rate (100,000 U/L) from post-ECLS (7.8 ± 2.0 ml/hour) to 72 hours (33.6 ± 1.2 ml/hour) to maintain the target ACT (Figure 3B). von Willebrand factor activity decreased in both groups on day 1 (Figure 3C). Thromboelastography (Figure 4) showed that R was prolonged versus BL in CTRL and was elevated compared to TLP. Potassium increased in both groups. Both groups exhibited a reduction in MA after start of ECLS which returned to BL levels by 24 hours and continued to increase. LY60 was reduced in both groups by 48 hours. With time on ECLS both groups exhibited an increase in platelet aggregation (Figure 5).

Table 2. - Blood Cell Count and Coagulation

Variable	Group	Baseline	Post-ECLS	3 Hours	6 Hours	12 Hours	24 Hours	48 Hours	72 Hours
WBC (103/µL)	CTRL	14.6 ± 1.5	16.1 ± 1.4	15.0 ± 2.1	15.1 ± 1.9	15.0 ± 1.6	14.4 ± 0.9	15.1 ± 0.9	11.1 ± 1.2
	TLP	13.9 ± 1.2	13.9 ± 1.6	11.5 ± 1.4	12.0 ± 1.5	12.4 ± 1.2	12.0 ± 1.1	10.5 ± 1.8	8.74
RBC (106/µL)	CTRL	5.8 ± 0.1	5.6 ± 0.2	5.1 ± 0.1Table 2.	4.7 ± 0.1Table 2.	4.5 ± 0.1Table 2.	4.3 ± 0.2Table 2.	4.2 ± 0.3Table 2.	3.7 ± 0.3Table 2.
	TLP	6.1 ± 0.1	5.7 ± 0.1Table 2.	5.2 ± 0.1Table 2.	5.0 ± 0.1Table 2.	4.8 ± 0.1Table 2.	5.0 ± 0.4Table 2.	4.7 ± 0.4Table 2.	3.3
Hb (g/dL)	CTRL	9.9 ± 0.3	9.6 ± 0.4	8.6 ± 0.2Table 2.	8.1 ± 0.2Table 2.	7.6 ± 0.1Table 2.	7.4 ± 0.3Table 2.	7.2 ± 0.5Table 2.	6.3 ± 0.5Table 2.
	TLP	9.9 ± 0.2	9.3 ± 0.2Table 2.	8.5 ± 0.3Table 2.	8.3 ± 0.3Table 2.	7.9 ± 0.2Table 2.	8.3 ± 0.8	7.7 ± 0.4Table 2.	5.9
Plt (103/μL)	CTRL	274 ± 38	263 ± 35	236 ± 34	218 ± 38	196 ± 30Table 2.	152 ± 16Table 2.	122 ± 11Table 2.	121 ± 7Table 2.
	TLP	308 ± 30	269 ± 27Table 2.	238 ± 21Table 2.	225 ± 20Table 2.	190 ± 24Table 2.	148 ± 23Table 2.	130 ± 22Table 2.	155
PT (s)	CTRL	13.3 ± 0.2	13.7 ± 0.2Table 2.	14.0 ± 0.1Table 2.	14.3 ± 0.1Table 2.	14.5 ± 0.3Table 2.	14.5 ± 0.2Table 2.	14.4 ± 0.2Table 2.	14.2 ± 0.2Table 2.
	TLP	13.4 ± 0.2	13.7 ± 0.2	13.9 ± 0.2Table 2.	14.0 ± 0.2Table 2.	13.8 ± 0.2Table 2.	14.5 ± 0.5Table 2.	14.1 ± 0.3Table 2.	13.9
aPTT (s)	CTRL	21.3 ± 1.6	40.0 ± 3.6	59 ± 30Table 2.	56.7 ± 7.2Table 2.Table 2.	91.6 ± 15.0Table 2.Table 2.	65.0 ± 15.9Table 2.Table 2.	75.2 ± 14.8Table 2.Table 2.	96.2 ± 27.3Table 2.Table 2.
	TLP	19.3 ± 2.3	40.7 ± 6.0Table 2.	18.4 ± 1.8	18.3 ± 1.7	18.0 ± 1.4	17.5 ± 1.3	18.0 ± 1.5	15.4
ACT (s)	CTRL	91 ± 4	129 ± 6Table 2.	130 ± 13Table 2.Table 2.	170 ± 16Table 2.Table 2.	171 ± 20Table 2.Table 2.	137 ± 7Table 2.Table 2.	137 ± 8Table 2.Table 2.	137 ± 6Table 2.Table 2.
	TLP	95 ± 1	126 ± 9	95 ± 3	95 ± 2	94 ± 3	96 ± 3	94 ± 2	94
Fib (mg/dL)	CTRL	188 ± 17	171 ± 14	166 ± 14	165 ± 11	196 ± 14	277 ± 17Table 2.	426 ± 23Table 2.	458 ± 35Table 2.
	TLP	182 ± 20	162 ± 13	159 ± 16	153 ± 12	175 ± 18	208 ± 3	389 ± 61Table 2.	210
PFHb (mg/dL)	CTRL	27 ± 3	10 ± 2Table 2.	10 ± 2Table 2.	11 ± 3Table 2.	12 ± 2Table 2.Table 2.	12 ± 1Table 2.Table 2.	11 ± 1Table 2.Table 2.	24 ± 7
	TLP	25 ± 5	9 ± 2Table 2.	16 ± 3Table 2.	16 ± 4Table 2.	24 ± 3	27 ± 3	28 ± 4Table 2.	68Table 2.
BUN (mg/dL)	CTRL	7.0 ± 1.5	7.2 ± 1.4	8.8 ± 1.5	9.8 ± 1.5	10.8 ± 1.6	13.2 ± 1.4Table 2.	11.4 ± 1.9	10.4 ± 1.2
	TLP	7 ± 1.7	6.8 ± 1.5	7.4 ± 1.3	8.4 ± 1.1	9.6 ± 2.3	13.8 ± 2.0Table 2.	12.3 ± 1.0Table 2.	12
Creatinine(mg/dL)	CTRL	1.4 ± 0.1	1.6 ± 0.2	1.7 ± 0.2	1.7 ± 0.2Table 2.	1.7 ± 0.2Table 2.	1.8 ± 0.2Table 2.	1.4 ± 0.1	1.4 ± 0.1
	TLP	1.4 ± 0.1	1.4 ± 0.1	1.4 ± 0.1	1.6 ± 0.1	1.6 ± 0.1	1.6 ± 0.1	1.4 ± 0.1	1.4
K+ (mmol/L)	CTRL	3.7 ± 0.1	3.6 ± 0.1	3.8 ± 0.1	3.9 ± 0.1	3.7 ± 0.1	3.8 ± 0.1	4.0 ± 0.2	4.1 ± 0.1
	TLP	3.6 ± 0.1	3.5 ± 0.1	3.7 ± 0.1	4.0 ± 0.1Table 2.	4.1 ± 0.1Table 2.	3.9 ± 0.2	4.1 ± 0.2Table 2.	3.9

Mean ± standard deviation. Significance p < 0.05.

*Significant change from baseline.

†Significant difference between CTRL and TLP groups.

ACT, activated clotting time; aPTT, activated partial thromboplastin time; BUN, blood urea nitrogen; CTRL, control; ECLS, extracorporeal life support; Fib, fibrinogen; Hb, hemoglobin; K⁺, potassium; PFHb, plasma free hemoglobin; Plt, platelet; PT, prothrombin time; RBC, red blood cell; TLP, tethered liquid perfluorocarbon; WBC, white blood cell.

Figure 3.: Mean ± standard deviation of (A) D-dimer concentration, (B) ATIII activity, and (C) vWF activity in platelet poor plasma from control animals (CTRL; n = 5) receiving ECLS with heparin-coated circuits and continuous heparin infusion versus TLP animals (n = 5) receiving ECLS using TLP-coated circuits and no systemic anticoagulation for 72 hours circulation. *Significant change from baseline in CTRL. **Significant change from baseline in TLP. All tests were two-sided with significance p <0.05. ATIII, antithrombin III; CTRL, control; ECLS, extracorporeal life support; TLP, tethered liquid perfluorocarbon; vWF, von Willebrand factor.

Figure 4.: Mean ± standard deviation of thromboelastography measurement of (A) clot reaction time (R), (B) initial clot formation time (K) (C) clot strength (MA), and (D) fibrinolysis at 60 minutes (LY60) in whole-blood from control animals (CTRL; n = 5) receiving ECLS with heparin-coated circuits and continuous heparin infusion versus TLP animals (n = 5) receiving ECLS using TLP-coated circuits and no systemic anticoagulation for 72 hours circulation. *Significant change from baseline in CTRL. **Significant change from baseline in TLP. †Significant difference between CTRL and TLP. All tests were two-sided with significance p <0.05. CTRL, control; ECLS, extracorporeal life support; TLP, tethered liquid perfluorocarbon.

Figure 5.: Mean ± standard deviation of (A) normalized platelet count (calculated by [platelet (Plt) × (hemoglobin concentration at specified time point/baseline hemoglobin concentration)], (B) collagen (COL)-stimulated platelet aggregation, and (C) ADP-stimulated platelet aggregation in whole-blood from control animals (CTRL; n = 5) receiving ECLS with heparin-coated circuits and continuous heparin infusion versus TLP animals (n = 5) receiving ECLS using TLP-coated circuits and no systemic anticoagulation for 72 hours circulation. Platelet aggregation was measured via impedance aggregometry and reported as total AUC divided by the platelet count. *Indicates significant change from baseline in CTRL. **Significant change from baseline in TLP. All tests were two-sided with significance p <0.05. ADP, adenosine diphosphate; AUC, area under the curve; CTRL, control; ECLS, extracorporeal life support; TLP, tethered liquid perfluorocarbon.

Circulation-related complications and significant events are reported in Table 3. The only statistically significant difference was the requirement for elevated sweep gas rate in TLP versus CTRL (ECLS settings shown in Table 1, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561). Numerically, bleeding complications were more frequent in CTRL, whereas thrombotic complications were more frequent in TLP. The most common site of bleeding was peripheral cannulas; and the most common site for thrombi was the membrane.

Table 3. - Complications and Significant Events

	CTRL % (Time ± SD)	TLP % (Time ± SD)
Hemorrhage: ECLS cannula site	60% (30 ± 14 h)	N/A
Hemorrhage: peripheral cannula	60% (23 ± 4 h)	20% (47 h)
Hemorrhage: airway	60% (24 ± 20 h)	20% (24 h)
Circuit thrombus: pump	N/A	20% (unknown)
Circuit thrombus: membrane	40% (69 ± 1 h)	60% (48 ± 11 h)
Circuit thrombus: occlusion	20% (72 h)	40% (50 ± 18 h)
Increase sweep gas	N/A	80% (39 ± 22 h)Table 3.
Increase blood flow	N/A	60% (41 ± 21 h)
Vasopressors	20% (32 h)	80% (44 ± 18 h)
Diuretic/furosemide	60% (3 ± 3 h)	60% (13 ± 18 h)

Instance of significant events during extracorporeal circulation including complications, deviations from initial circuit settings necessitated by animal condition and membrane performance, and administration of medications unrelated to anesthesia. The percent instance of each event for CTRL (n = 5) and TLP (n = 5) groups are shown. The mean time to onset of event (hours post-ECLS initiation) ± standard deviation is shown.

*Represents significant difference between CTRL and TLP groups.

CTRL, control; ECLS, extracorporeal life support; N/A, not applicable; TLP, tethered liquid perfluorocarbon.

Postcirculation assessment of circuits did not reveal between-group differences in membrane thrombus deposition area or protein adhesion on catheters and tubing (Figure 6). Scanning electron microscopy imaging indicated that thrombus deposition on TLP-coated materials with heparin-free circulation was similar to heparin-coated CTRL devices with heparin infusion (Figure 7). Developed thrombi and areas of dense fibrin coverage were consistently observed on the membrane inlet and progressively decreased toward the outlet. Thrombus deposition on catheters was similar between groups, with greatest deposition at the inlet and outlet connections (Figure 1, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561). Thrombus deposition on connective tubing was minimal or absent.

Figure 6.: Post-circulation assessment of explanted ECLS catheters, tubing and membrane oxygenators. A: Mean ± standard deviation protein adhesion (total protein concentration/blood-contacting surface area) collected from 19 Fr dual-lumen catheters for veno-venous ECLS follow 72 hours circulation in vivo in swine (n = 5/group). CTRL animals used manufacturer’s standard catheters during ECLS with continuous heparin infusion versus TLP animals that used standard catheters with TLP coating applied during ECLS with no systemic anticoagulation. B: Mean ± standard deviation of total thrombus area determined from digital images of polymethylpentene fiber layers collected from the membrane lung inlet/venous face (INLET), center layer of the membrane (MID), and outlet/arterial face of the membrane (OUTLET). The percent of the total membrane area that was covered in thrombus was scored by three blinded reviewers, and the average of the three scores was taken for each sample. *Significant change versus inlet. All tests were two-sided with significance p <0.05. CTRL, control; ECLS, extracorporeal life support; TLP, tethered liquid perfluorocarbon.

Figure 7.: Field emission scanning electron microscopy images from explanted membrane oxygenators at end of study. Images from membrane lungs with immobilized-heparin CTRL coating (A–F) used with continuous heparin infusion versus membranes with TLP (G–L) coating used without systemic anticoagulation for 72 hours extracorporeal circulation in swine (n = 5/group). Top row images (A and B, G and H) from the inlet/venous face of the membrane lung. Center row images (C and D, I and J) from the center of the membrane lung. Bottom row images (E and F, K and L) from the outlet/arterial face of the membrane lung. Significant thrombus deposition present on all membrane lungs, with no apparent difference between groups. Layers of fibrin mesh cover the gas exchange fibers, potentially inhibiting blood flow and gas transfer. Density and extent of thrombus formation appeared to decrease from the inlet face to the outlet face. CTRL, control; ECLS, extracorporeal life support; TLP, tethered liquid perfluorocarbon.

Systemic cytokine levels (Figure 2A–E, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561) were unremarkable, except an increase in IL-6 in TLP beginning at 48 hours. Additionally, IL-8 was reduced from 12 to 24 hours in both groups. There was no change in syndecan-1 (Figure 2F, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561). After histological analysis, there were no obvious signs of toxic damage or distinctive pathologies in the lungs, kidney, liver, or jejunum for either group. Findings were mostly unremarkable, with the exception that all animals had signs of lower lung consolidation, likely due to positional atelectasis and mechanical ventilation (Figure 3, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A561).

Discussion

To our knowledge, this is the first evaluation of an investigational biocompatible coating applied to standard ECLS circuits evaluated on a multiday time frame in vivo. This study was a logical extension of a previously successful 6 hour testing of TLP-coated circuitry.⁹ The results of this 72 hour assessment stress the importance of evaluating biomaterials at a clinically relevant timeline in a model that closely approximates the clinical scenario before making recommendations to clinical care—as opposed to solely short-term (2–8 hour) ex vivo and small animal testing. We hypothesized that TLP is safe, efficacious and reduces thrombotic and bleeding complications for 72 hours ECLS without heparin relative to CTRL (immobilized-heparin coating + systemic heparin). Although there was no statistical difference in bleeding or thrombotic complication rates between groups, we found that TLP without supplemental anticoagulation is insufficient to prevent circuit thrombosis for 72 hours. Additionally, we found that membrane performance was altered in TLP requiring higher sweep gas flow than CTRLs. Finally, we found that heparin-coated ECLS with continuous heparin infusion also does not prevent occlusive thrombosis for 72 hours, and we document the impact of this therapy on healthy animals.

TLP was identified as a promising coating for ECLS because of ability to resist protein adhesion and thrombus deposition. Protein adsorption is a fundamental step in foreign surface-mediated thrombosis, providing platelet binding sites and stimuli for platelet activation. Although TLP previously reduced protein adsorption and thrombus deposition ex vivo^7,8 and during our 6 hour in vivo ECLS study,⁹ these effects were not substantiated for 72 hours. This may be due to insufficient coating durability, meaning the lubricant layer was not retained on circuit surfaces for the duration and blood flow rates tested. Additionally, this may be a functional limitation of the coating, as the nonadhesive property of the coating does not address the contact pathway of coagulation activation or shear-induced cellular activation that can drive systemic coagulation and inflammation. Likely, both durability and functional limitations contributed to circuit occlusion.

Although bleeding complications were more frequent in CTRL (which received continuous heparin) and thrombotic complications were more frequent in the heparin-free TLP group, specific trends in coagulation were similar between groups. For example, the platelet count, which is widely utilized as a measure of biocompatibility,^16,17 decreased with time on ECLS with no between-group differences. Despite this decrease in platelet count, an increase in platelet aggregation was observed from 24 to 72 hours in both groups. This was consistent with the trend toward elevated TEG clot strength (MA) which is correlated with platelet count and/or aggregation.¹⁸ This proaggregatory response may be attributed to shear stress more than foreign surface-mediated interactions. Shear stress directly activates platelets promoting aggregation; and indirectly induces platelet activation via erythrocyte and platelet lysis, increasing extracellular concentrations of platelet activators such as ADP.^19,20 We also observed a transient decrease in vWF during the first 24 hours on ECLS, which can occur with shear-induced uncoiling and cleavage of high molecular weight multimers of vWF or vWF-mediated platelet adhesion under high shear.²¹ Additionally, red blood cell counts progressively decreased during ECLS in both groups, although PFHb did not indicate clinical hemolysis. Platelets are more susceptible to shear-induced lysis than red blood cells,²² suggesting that shear-mediated platelet disturbances may manifest with subclinical hemolysis. The effects of shear stress observed highlight a functional limitation of biocompatible coatings like TLP and immobilized-heparin, as they do not address shear-induced coagulopathy and blood cell damage.

In addition to functional limitations of nonadhesive coatings like TLP, postcirculation assessment of materials suggested that the lubricant layer was not retained on the TLP surfaces for the study duration, as evidenced by significant membrane thrombus deposition. In our previous 6 hour assessment of TLP versus immobilized-heparin for ECLS where neither group received systemic heparin, thrombus deposition was significantly reduced on TLP- versus heparin-coated membranes.⁹ In the present 72 hour study where the TLP group was unheparinized and CTRL/immobilized-heparin group received systemic heparin, membrane thrombus deposition was similar between groups. Dense fibrin layers with incorporated white and red blood cells and platelets were observed around the polymethylpentene fibers in both TLP and CTRL, and likely impeded gas exchange.² Additionally, protein adhesion on tubing and catheters was not reduced on TLP-coated surfaces as previously reported⁷; however, we utilized higher flow rates and longer study duration. This may suggest that TLP is only suitable for short-term or low-flow applications, or that a more stable formulation of the coating is required. Alternatively, a multifaceted coating that combines the nonadhesive properties of TLP with an inhibitor of thrombin or platelet activation may improve the performance and stability of TLP.

In addition to coagulation, we assessed vital signs and respiratory variables to evaluate safety of TLP and impact on membrane performance. ECLS initially enabled ~50% reduction in mechanical ventilator settings in both groups, as previously reported by our group using VV extracorporeal CO2 removal in swine.²³ In CTRL, reduction in ventilator settings persisted for 72 hours, reflected in the decreased respiratory rate and minute ventilation and sustained reduction in P_aCO2. In TLP, P_aCO2 was only reduced immediately after the start of ECLS. For the remainder of the study, the TLP group required an increase in respiratory rate and F_iO₂; with relative increases in minute ventilation and ECLS blood flow settings [n.s.] to meet management goals. In summary, the TLP group required escalation of ventilator settings compared with CTRL in which reduction in minute ventilation was enabled as is customary during extracorporeal CO2 removal. Cumulatively with the elevation of oxygenation index in TLP, this implies that respiratory performance of TLP-coated membranes was inferior to CTRL. This effect was not observed in our 6 hour evaluation of TLP for ECLS,⁹ and did not manifest until approximately 12 hours—stressing the importance of testing for duration of intended use. Furthermore, in TLP animals PFR gradually decreased with time on ECLS and was indicative of mild acute respiratory distress syndrome (ARDS) from 24 to 48 hours.²⁴ Elevated IL-6 was also observed in TLP, which has been shown to be predictive of poor clinical outcomes during ECLS.²⁵

There was no group difference in O₂ transfer across the membrane lung; and transfer was comparable to previous reports for membranes of similar dimensions.^26,27 Although there was no statistical difference in percent reduction in PCO2 across the membrane, numerically the PCO2 gradient was always lower in TLP versus CTRL (despite elevated sweep gas in TLP to achieve this gradient). Whether this is due to thrombus deposition or damage to the native polymers during coating application altering gas permeability of the membrane itself, remains unclear—although the fact that we observed similar thrombus formation on CTRL membranes without the same alterations in membrane function may suggest the latter.

A limitation of this study is the relatively small sample size because of the number of circuits available for testing; however, the analysis we performed utilizing complete circuits and close replication of clinical timelines and conditions is essential to fully evaluate the efficacy of a biomaterial for ECLS. We also did not have an adequate number of circuits to evaluate use of immobilized-heparin circuits without supplemental heparin; or to assess TLP with heparin anticoagulation—to decipher the effects of the coatings versus anticoagulation regimen. The circuits we evaluated were designed for low-flow extracorporeal CO2 removal, so results may not be directly applicable to other forms of ECLS that utilize different blood flow rates. Additionally, we used healthy animals to first understand the blood-biomaterial interaction without the variability of subject-to-subject coagulopathic response to injury; however, results may be different in injured subjects.

The vascular endothelium incorporates numerous prothrombotic and antithrombotic regulators to maintain blood fluidity. Along these lines, we believe a multifactorial coating that incorporates not only a nonadhesive surface like TLP but also bioactive hemostatic regulators like nitric oxide,^17,28–30 which prevents platelet activation and aggregation, and direct thrombin inhibitors,³¹ may be the ideal solution to prevent coagulopathic complications during ECLS.³² Additionally, just as the endothelium is heterogenous with varied regulatory mechanisms for tissue-specific demands and flow rates, it is possible that a heterogenous, component-specific coating approach may be required.^33,34 Furthermore, the role of circuit design including fluid dynamics, pressure differentials, pump design, and shear stress cannot be overlooked and is vital to improving the hemocompatibility of ECLS, along with the biocompatibility of the materials.

Conclusion

This study outlines a robust system for evaluation of hemocompatible materials for ECLS. We found that TLP did not enable heparin-free ECLS at the blood flow rate and duration tested, and also altered membrane performance. Our results were substantially different compared to our previous work with TLP using 6 hour in vivo testing, stressing the importance of the multiday, clinically relevant study timeline. Going forward, we plan to test other biomaterials utilizing this same format to identify a solution for heparin-free ECLS that is suitable for clinical application.

Acknowledgment

The authors thank the University of South Florida Ph.D. Program in Integrated Biomedical Sciences for support of Dr. Robert’s doctoral training through this study; and Free Flow Medical Devices for application of the immobilized-tether layer to the TLP circuits we provided to them.