The Use of Antifactor Xa Assays in a Comprehensive Pediatric Extracorporeal Membrane Oxygenation Anticoagulation Protocol is Associated with Increased Survival and Significant Blood Product Cost-Savings

Abstract

We augmented our standard extracorporeal membrane oxygenation laboratory protocol to include antifactor Xa assays, thromboelastography, and antithrombin measurements. We performed a retrospective chart review to determine outcomes for patients placed on extracorporeal membrane oxygenation (ECMO) prior to and after the initiation of our anticoagulation laboratory protocol. A total of 663 consecutive ECMO runs were evaluated from January 1, 2007 to June 30, 2018. Of these patients, 252 were on ECMO prior to initiation of the anticoagulation laboratory protocol on September 1, 2011, and 411 patients were on ECMO after initiation of the protocol. There were no major changes to our extracorporeal membrane oxygenation circuit or changes to our transfusion threshold during this continuous study period. Transfusion utilization data revealed statistically significant decreases in almost all blood components, and a savings in blood component inflation-adjusted acquisition costs of 31% bringing total blood product cost-savings to $309,905 per year. In addition, there was an increase in survival to hospital discharge from 45 to 56% associated with the initiation of the protocol ( p  = 0.004). Our data indicate that implementation of a standardized ECMO anticoagulation protocol, which titrates unfractionated heparin infusions based on antifactor Xa assays, is associated with reduced blood product utilization, significant blood product cost savings, and increased patient survival. Future prospective evaluation is needed to establish an antifactor Xa assay-driven ECMO anticoagulation strategy as both clinically superior and cost-effective.

Introduction

Extracorporeal membrane oxygenation (ECMO) utilization has quadrupled in the pediatric population in the past two decades, but patients supported with ECMO continue to be at significant risk for a multitude of complications. ¹ Complications associated with ECMO include hemorrhage, stroke, limb ischemia, thrombosis, and infection. ^{2 3} Bleeding and thrombotic complications remain the leading cause of morbidity and mortality in patients on ECMO.

The bleeding and thrombosis during ECMO (BATE) study found that among pediatric patients, over 70% experienced at least one bleeding event necessitating a transfusion or resulting in intracranial hemorrhage, while 38% of patients experienced a thrombotic event. This study also reported a wide variation across the eight participating centers, with center-specific hemorrhagic complication rates ranging from 51% to as high as 90%. ⁴ The mechanisms of bleeding and thrombosis during ECMO are complex due to multiple alterations in hemostatic factors, anticoagulants used and the fact that ECMO necessitates contact between blood and the artificial, nonendothelialized surface of an extracorporeal circuit. ^{5 6 7}

Red blood cell (RBC) transfusion on ECMO, routinely employed for the augmentation of oxygen carrying capacity, circuit priming and treatment of hemorrhagic complications, ⁸ is often guided by institutional policies targeting specific hematocrit thresholds. The Extracorporeal Life Support Organization (ELSO) Anticoagulation Guidelines acknowledge the lack of evidence regarding optimal transfusion support and threshold for transfusion. ⁹ RBC transfusion volume is independently associated with mortality in pediatric patients on ECMO, as well as increased morbidity including central-line associated bloodstream infections, prolonged duration of mechanical ventilation and prolonged duration of hospitalization. ^{8 10}

Maintaining adequate anticoagulation on ECMO is challenging due to the complexities of the hemostatic system that are compounded by age-related developmental hemostatic changes, variable effects of the underlying illness that may contribute to either a hypercoagulable or a hypocoagulable state, and blood-circuit interaction. ^{11 12} Unfractionated heparin is the most widely utilized anticoagulant during ECMO. There are no definitive guidelines for the management of anticoagulation for ECMO patients.

In 2011, our institution changed our anticoagulation laboratory protocol to use antifactor Xa assays without exogenous antithrombin (AT) to guide heparin titration ( Table 1 ). In a previous publication, we reported this anticoagulation laboratory protocol was associated with a decrease in the incidence of cannula and surgical site bleeding, an increase in ECMO circuit life, and a decrease in blood product transfusion. ¹³ For our current study, we sought to determine if we continued to have fewer hemorrhagic complications, reduced blood product usage, and increased circuit life as well as quantify the cost-savings realized for the potential decrease in blood product utilization.

Table 1. Anticoagulation protocol.

Condition	Activated clotting time	Protime/partial thromboplastin time/international normalized ratio/fibrinogen	Hematocrit/platelets	Antithrombin	Anti-Xa level	Thromboelastogram
Stable	Q1	Daily	Q8	Daily	Q6	Weekly PRN
Bleeding or clotting	Q1 and PRN	Treat and redraw within 4 h	Treat and redraw within 1 h	Daily	Q6 and PRN	Daily PRN
Heparin dose decreasing	Q1 and PRN	Treat and redraw within 4 h QAM if normal	Treat and redraw within 1 h Q8 if normal	Daily	Q6 and PRN	Daily PRN
Heparin dose increasing	Q1 and PRN	Treat and redraw within 4 h QAM if normal	Treat and redraw within 1 h Q8 if normal	Daily and PRN	Q6 and PRN	Daily PRN

Abbreviations: PRN, as needed (pro re nata); Q1H, every hour; Q6H, every 6 hour; QAM, every morning.

Materials and Methods

The ECMO program at Monroe Carell Jr. Children's Hospital at Vanderbilt is an ELSO Center of Excellence. In addition to reporting data to ELSO, we maintain a Research Electronic Data Capture (REDCap) database containing demographics, blood product usage, complications, and outcomes for all ECMO patients at our center. The bedside ECMO specialist records clots in the circuit, and clots are categorized by location in the circuit. It does not account for the number of clots in a particular circuit component. Whether a circuit component had a single clot or multiple clots during the ECMO run, it is only marked as present under the clotting complications section. The presence of a clot in a circuit component does not indicate that the circuit component was changed or removed. A charge log database records the number of circuits used during each ECMO run. We obtained permission from the Vanderbilt University Institutional Review Board to review this data and evaluated all ECMO runs at our institution from January 2007 to June 2018.

In September 2011, the Monroe Carell Jr. Children's Hospital at Vanderbilt instituted a revised protocol for laboratory testing to manage anticoagulation for patients on ECMO. The changes with this protocol included the incorporation of antifactor Xa assays, AT levels, and thromboelastography in addition to our standard laboratory draws for activated clotting time (ACT), platelet count, prothrombin time (PT), international normalized ratio (INR), activated partial thromboplastin time (aPTT), and hematocrit.

Our standard initial ACT range remained 200 to 220 seconds. However, after initiation of the protocol, antifactor Xa assays were used to guide heparin titration, and goal ACT ranges for patients on ECMO were adjusted to maintain a therapeutic antifactor Xa assay of 0.3 to 0.7 IU/mL. After protocol initiation, we followed AT activity level every 24 hours. We considered dosing AT concentrate or fresh frozen plasma (FFP) if the level was low for age and the heparin requirement was greater than 60 U/kg/h.

Thromboelastography was obtained as needed for bleeding or clotting complications. Prior to protocol initiation, ECMO heparin infusions were primarily titrated by using ACTs, with some guidance from aPTT results. Antifactor Xa assays at our institution are performed on the STA-R Evolution automated coagulometer (Diagnostica Stago, Parsippany, New Jersey, United States). The patient's plasma serves as the source of unfractionated heparin and AT, and no exogenous AT is added. The two form an unfractionated heparin-antithrombin complex that binds to bovine factor Xa added to the sample. The residual factor Xa binds to a chromogenic substrate, and the signal is inversely proportional to the concentration of unfractionated heparin in the sample. The test is based on absorbance or optical density of monochromatic light passing through a cuvette when a chromogenic reaction takes place.

During the study period, there were no significant changes in the type or brand of ECMO equipment. For patients equal to or greater than 10 kg, we used the Stöckert Centrifugal Pump Console Trolley System with the Revolution centrifugal cone (Sorin Biomedical). For patients weighing less than 10 kg, we used the SIII single roller head pump (Sorin Biomedical, Irvine, California, United States). This circuit contains a straight 6-inch long silicone bladder connected to 0.25 inch Smart tubing that has a coating of polysiloxane-containing copolymers to reduce the inflammatory response (Sorin, Milan, Italy). For the roller-head raceway, we used 3/8 inch S97-E Tygon tubing (Saint-Gobain, Paris, France). We used the Quadrox-iD Pediatric polymethylpentene hollow-fiber oxygenator (Maquet, Hirrlingen, Germany) with the roller head circuit. The Quadrox-iD Pediatric oxygenator has an albumin-heparin BIOLINE coating to enhance biocompatibility. For patients equal to or greater than 10 kg, we used the Stöckert Centrifugal Pump Console Trolley System with the revolution centrifugal cone (Sorin Biomedical). The circuit used with centrifugal pumps consists of 3/8 inch smart tubing that has a coating of polysiloxane-containing copolymers to reduce the inflammatory response (Sorin) and the Quadrox D polymethylpentene hollow-fiber oxygenator (Maquet). The Quadrox D oxygenator has a heparin-free, synthetic polymer SOFTLINE (Maquet, Hirrlingen, Germany) coating to enhance biocompatibility. Until September 2008, we used the minimax and affinity microporous polypropylene hollow-fiber oxygenators (Medtronic, Minneapolis, Minnesota, United States) before switching to the Quadrox oxygenator (Maquet).

Systemic anticoagulation is achieved with unfractionated heparin, and bedside ACTs are measured with the Hemochron Signature Elite (International Technidyne, Edison, New Jersey, United States). There were no changes in our guidelines for blood product transfusion on ECMO during the study period. Packed RBCs were transfused to maintain a hematocrit of 30 to 40%, with neonates and children with cyanotic congenital heart disease typically transfused at the higher end of that range. Platelets were transfused to maintain counts greater than 80,000/μL. Fresh frozen plasma was transfused to keep an INR less than 1.8, and cryoprecipitate was transfused to maintain a fibrinogen level of at least 150 mg/dL.

Total (including direct and indirect), direct, and variable direct hospital costs for each patient were obtained from our local hospital cost accounting database (EPSi, Chesterfield, Missouri, United States). All cost outcomes are adjusted for inflation to 2016 U.S. dollars by using the Consumer Price Index for All Urban Consumers. ¹⁴

Univariate comparisons were used to quantify the characteristics of the patient population. Continuous variables were compared by using the Wilcoxon rank-sum test, whereas categorical variables were compared by using the Chi-square test. Blood product transfusions for the first 504 patients to extrapolate direct cost-savings were calculated by the two-tailed Student's t tests. All univariate statistical tests were two-tailed, with a p -value of 0.05 or less considered significant. Statistical analysis was performed with R version 3.0.2 (R Foundation for Statistical Computing, Vienna, Austria).

Results

A total of 663 ECMO runs were evaluated from January 2007 to June 2018. Of these patients, 252 were on ECMO prior to initiation of the anticoagulation laboratory protocol on September 1, 2011, and 411 patients were on ECMO after initiation of the protocol. Table 2 compares descriptive statistics for ECMO patients before and after initiation of the protocol. ECMO indication was similar in both groups; however, age distribution and duration of ECMO was found to be statistically different.

Table 2. Descriptive statistics before and after protocol initiation.

Descriptor	Prior to protocol n  = 252	After protocol n  = 411	p -Value
Age distribution			0.003 b
< 1 y	190 (75)	278 (68)
1 to <10 y	28 (11)	59 (14)
10 to <17 y	25 (10)	30 (7)
≥ 18 y	9 (4)	44 (11)
Median duration on ECMO (d)	4.48	4.95	0.015 a
ECMO indication			0.589 b
Cardiac	90 (36)	161 (39)
ECPR	63 (25)	91 (22)
Respiratory	99 (39)	159 (39)
Survival to discharge			0.004 b
Overall
No	139 (55)	180 (44)
Yes	113 (45)	231 (56)
< 1 y			p  < 0.001
No	116 (61.0%)	116 (41.7%)
Yes	74 (39.0%)	162 (58.3%)
1 to <10 y			0.902
No	11 (40.7%)	24 (40.2%)
Yes	17 (59.3%)	35 (59.8%)
10 to <17 y			0.076
No	9 (36.0%)	18 (60%)
Yes	16 (64.0%)	12 (40%)
≥ 18 y			0.361
No	3 (33.3%)	22 (50%)
Yes	6 (66.7%)	22 (50%)

Abbreviations: ECMO, extracorporeal membrane oxygenation; ECPR, extracorporeal cardiopulmonary resuscitation.

^aWilcoxon's signed-rank test.

^bPearson's Chi-squared test.

Survival to hospital discharge increased significantly from 45% before the protocol to 56% after the protocol ( p  = 0.004). There were significant decreases in surgical site bleeding (40–31%; p  = 0.016) and pulmonary hemorrhage (20–12%; p  = 0.004 ) after initiation of the protocol but no significant difference in cannula site bleeding ( Table 3 ). Significantly more oxygenator clots were recorded as being observed after protocol initiation (8–15%; p  = 0.018), but there was no significant difference in the frequency of circuit changes.

Table 3. Bleeding and clotting complications before and after protocol initiation.

Complication	Prior to protocol n  = 252	After protocol n  = 411	p -Value
Hemorrhagic complications
Cannula site bleeding	60 (24)	102 (25)	0.738
Surgical site bleeding	101 (40)	127 (31)	0.016 a
Pulmonary hemorrhage	51 (20)	49 (12)	0.004 a
Clotting complications
Oxygenator clot	21 (8)	60 (15)	0.018 a
Hemofilter clot	19 (8)	40 (10)	0.337
Continuous renal replacement therapy clot	2 (5)	19 (5)	0.925
Isolated oxygenator change	7 (3)	11 (3)	0.948

^aPearson's Chi-squared test.

The use of packed RBCs, fresh frozen plasma, and platelets decreased significantly after protocol initiation ( Table 4 ). Median packed RBC usage decreased by more than half after the initiation of the protocol (48.5–18.2 mL/kg/d of ECMO; p  < 0.001). There were significant reductions in usage of FFP (15.9–6.6 mL/kg/d of ECMO; p  < 0.001) and platelets (24.9–16.7 mL/kg/d of ECMO; p  < 0.001) as well after protocol initiation.

Table 4. Blood product use before and after protocol initiation.

Product (mL/kg/d of ECMO)	Median before protocol (25–75th percentile)	Median after protocol (25–75th percentile)	p -Value
Packed RBCs	48.5 (17.5–109.3)	18.2 (8.2–42.0)	<0.001 a
Fresh frozen plasma	15.9 (4.1–42.2)	6.6 (1.8–18.8)	<0.001 a
Platelets	24.9 (11.5–42.8)	16.7 (5.5–28.6)	<0.001 a
Cryoprecipitate	0.4 (0–2.1)	0.3 (0–1.7)	0.181

Abbreviations: ECMO, extracorporeal membrane oxygenation; RBC, red blood cell.

^aWilcoxon's signed-rank test.

Even though ECMO duration was greater after the protocol, there was still a significant increase in circuit life after the initiation of the protocol ( Table 5 ). Overall, circuit life increased from a median of 4.11 to 5.0 days ( p  < 0.001). In order to eliminate patients with a brief duration of ECMO, circuit life was analyzed separately for patients who were on ECMO for greater than 24 hours. For this population, the difference in circuit life was clinically and statistically significant, with a median of 4.61 days before the protocol and a median of 5.34 days after the protocol ( p  = 0.001).

Table 5. Extracorporeal membrane oxygenation circuit life data.

	n	Median circuit life (d) (25–75th percentile)	p -Value
All patients			<0.001 a
Before	252	4.11 (2.24–6.91)
After	411	5 (2.85–8.27)
Total	663	4.76 (2.61–7.92)
Patients on ECMO ≥1 d			0.003 a
Before	226	4.61 (2.76–7.50)
After	377	5.34 (3.36–8.77)
Total	603	5.06 (3.01–8.18)

Abbreviation: ECMO, extracorporeal membrane oxygenation.

^aWilcoxon's signed-rank test.

Table 6 compares the first 252 postprotocol patients with the 252 pre-protocol patients for the purposes of cost analysis. The postprotocol group showed significant reductions in total RBC, FFP, and platelets. This resulted in a direct cost-savings of $84,585 per year in blood product acquisition costs. Increased AT administration and AT activity level laboratory testing after initiation of the protocol led to an increase in $42,336 per year in the postprotocol group.

Table 6. Descriptive statistics and comparison of blood product and antithrombin usage in patients before and after implementation of anticoagulation protocol.

			Pre-ECMO protocol cost analysis		Post-ECMO protocol cost analysis		p -Value
Number of patients			252		252
Total ECMO days			1,476		1,929
Circuit life			4.11 d		4.84 d		p  = 0.006 b
Average weight of patients			13.5 kg		15.3 kg		p  = 0.17 a
Time period			4.75 y		4.58 y
Age distribution							0.09 a
< 1 y			190 (75)		184 (73)
1 to <10 y			28 (11)		30 (12)
10 to <17 y			25 (10)		17 (7)
≥ 18 y			9 (4)		21(8)
ECMO indication							0.845 a
Cardiac			90 (36)		84 (33)
ECPR			63 (25)		104 (41)
Respiratory			99 (39)		64 (25)
Survival to discharge							0.003 a
No			139 (55)		106 (42)
Yes			113 (45)		146 (58)
	Before protocol initiation				After protocol initiation
	Total	Mean ± SD (median [range])	Direct cost per year	Total	Mean ± SD (median [range])	Direct cost per year	p -Value	% Change
Packed red blood cells (units)	1,291	5.12 ± 8.84 (3.00 [0–96])	$61,968	709	2.78 ± 5.25 (1.00 [0–46])	$30,942	<0.001	45.0%
Fresh frozen plasma (units)	935	3.71 ± 5.25 (2.00 [0–49])	$8,426	600	2.37 ± 4.98 (1.00 [0–52])	$4,917	p  = 0.0034	35.8%
Platelets (units)	1,036	4.20 ± 8.06 (2.00 [0–88])	$153,559	774	2.80 ± 5.93 (1.00 [0–59])	$104,307	p  = 0.027	25.3%
Cryoprecipitate (units)	177	0.70 ± 0.733 (1.00 [0–6])	$13,309	183	0.73 ± 1.50 (1.00 [0–17])	$12,511	p  = 0.82	3.3%

Abbreviations: ECMO, extracorporeal membrane oxygenation; ECPR, extracorporeal cardiopulmonary resuscitation; SD, standard deviation.

Note: Antithrombin administration and antithrombin activity level laboratory testing led to a $42,336 per year cost increase in the postprotocol group. All costs adjusted to 2016 US dollars using the Consumer Price Index for all urban consumers.

^aWilcoxon's signed-rank test.

^bPearson's Chi-squared test.

Discussion

Implementation of a comprehensive anticoagulation monitoring protocol in our institution was associated with decreased blood component utilization and costs, but more importantly was associated with improved survival. Our postprotocol ECMO patients experienced significantly less pulmonary hemorrhage, fewer surgical site hemorrhagic complications, and there was an increase in ECMO circuit life. The initiation of the anticoagulation laboratory protocol is the only major change that was made during the 12-year study period. There were no major changes to our ECMO circuitry or changes to our transfusion threshold for patients on ECMO during this continuous study period. The biggest change that came with the initiation of the anticoagulation laboratory protocol was the use of the anti-factor Xa assay to guide heparin titration. It is possible that the use of the anti-factor Xa assay to maintain therapeutic anticoagulation with unfractionated heparin avoided the excessive bleeding and blood product transfusion that can result from supratherapeutic anticoagulation and avoided thromboembolization and decreased circuit life that comes with subtherapeutic anticoagulation.

Anti-Xa assays provide a measure of the functional level of heparin in plasma. ^{12 15} Although ACT is the least expensive and most readily available test used to monitor heparin's effect in ECMO patients, it was originally designed to measure the anticoagulant effect of heparin during cardiopulmonary bypass. ^{16 17 18 19 20} The anti-factor Xa assay has a stronger correlation to heparin dose than ACT and is a more valuable monitor of heparin effect in pediatric patients supported with ECMO. In a cohort of 22 pediatric ECMO patients managed with an anti-Xa guided protocol compared to 10 patients historical controls managed with ACT, patients managed with the anti-Xa protocol had almost 20 fewer blood draws per day, had a stronger correlation to the heparin dose, and less bleeding. ¹⁷ Baird et al performed a retrospective review of 604 pediatric patients on ECMO and showed that ACT values did not correlate well with changes in heparin dose and suggested that adherence to recommended ACT values may result in inadequate anticoagulation in pediatric ECMO patients. ²⁰

The anti-factor Xa assay also has a stronger correlation to heparin dose than aPTT. aPTT values correlate poorly to anti-factor Xa and unfractionated heparin levels in pediatric patients on ECMO. ^{18 20 21 22} aPTT evaluates contact activation in the intrinsic pathway and is influenced by fibrinogen, factor VIII levels, heparin, and AT levels. ^{11 23} A single-center retrospective review of 47 pediatric patients requiring ECMO support demonstrated a significant decrease in the prevalence of bleeding, but increased circuit clotting complications, when using aPTT versus ACT to manage heparin infusions. ²⁴ aPTT can be unreliable in critically ill patients. ^{23 25} C-Reactive Protein behaves like a lupus anticoagulant and can falsely prolong aPTT. When factor VIII is elevated as an acute phase reactant or thrombin is circulating, aPTT is falsely decreased. ²⁵ In a prospective observational cohort of 17 pediatric ECMO patients, anti-factor Xa levels had a stronger correlation to heparin dose than the ACT or aPTT. ¹⁸

We experienced an associated increase in survival after protocol initiation and an associated reduction in blood component transfusion after initiation of our ECMO anticoagulation protocol. The association between blood component transfusion and hospital length of hospital stay and mortality is well established. A single-center retrospective study at our institution reported a 24% increase in adjusted odds of death for each additional 10 mL/kg/day of RBC transfusion volume among 203 infants and children supported with ECMO for non-cardiac indications. ¹⁰ In a 72 patient retrospective cohort study of neonates supported with ECMO for refractory hypoxemic respiratory failure, a protocol change from a target hematocrit of 40% to 35% was significantly associated with decreases in RBC transfusion volume and shorter ECMO durations. ²⁶ In a single-center study of adult VA ECMO patients, implementation of an ECMO transfusion protocol with an RBC transfusion threshold of 8 g/dL was associated with significantly reduced RBC transfusion burden and improved survival compared to retrospective pre-protocol patients. ²⁷ Taken together, these data suggest that greater transfusion volumes during ECMO support are associated with increased morbidity and mortality.

We found thromboelastography particularly useful in our comprehensive anticoagulation protocol when patients were experiencing excessive bleeding or clotting. For example, a normal thromboelastograph in the face of refractory hemorrhage could be related to surgical bleeding rather than an inherent or induced abnormality in coagulation. Young children have decreased physiologic concentrations of AT compared to adults, and hemodilution can further decrease AT levels due to the volume of blood in the ECMO circuit relative to patient blood volume. ^{11 12} We followed AT activity level daily on every patient and considered dosing antithrombin concentrate or FFP if the level was low for age and the heparin requirement was greater than 60 U/kg/hr. We avoided the use of AT for ECMO patients with serious hemorrhagic complications.

Blood transfusions were significantly reduced in our post-ECMO anticoagulation protocol population. As a result, a substantial cost-savings was realized for direct blood product acquisition costs. Because we had a disproportionate number of postprotocol patients per year compared to pre-protocol, to calculate cost-savings of blood product acquisition per year, we compared the 252 pre-protocol patients with the first 252 postprotocol patients. A total of 504 continuous pediatric ECMO patients were included in the analysis over 9 years. Despite greater average weight and total ECMO days in the postprotocol group, blood transfusions were significantly reduced in our post-ECMO anticoagulation protocol population ( Table 6 ). Total red blood cell transfusions decreased by 45% in the postprotocol group ( p  < 0.001). Significant decreases were also detected in plasma usage by 35.8% and platelets by 25.3% ( p  = 0.0034 and 0.027, respectively). A slight, but not statistically significant, increase was seen in cryoprecipitate usage in the postprotocol era. We realized total savings of $84,585 per year from a reduction in direct blood product acquisition costs. In addition to acquisition costs, transfusion-related (e.g., admission, testing, logistics, and administration) and overhead costs—key components of activity-based costs—must be considered. Each packed red blood cell unit transfused costs an estimated $1,400/unit; FFP of $400/unit; platelet dose of $1100/unit; and cryoprecipitate cost of $360/unit bringing total blood product cost-savings to $309,905 per year. ^{27 28 29} Accounting for a protocol-driven increase in laboratory frequency and AT usage, we incurred an increase in $42,336 in laboratory and AT administration costs per year in the postprotocol group. Further analyses are needed to assess the financial impact of prolonged ECMO circuit life as well as cost-savings associated with fewer hemorrhagic complications after protocol implementation.

Although there was an increase in oxygenator clots observed in the postprotocol group, the recording of an oxygenator clot does not indicate that a circuit change was performed or that the function of the circuit was compromised. Despite the increase in oxygenator clots in the postprotocol group, there was no statistically significant difference in isolated oxygenator or circuit changes in both groups. It is uncertain how clinically significant these findings are in the face of increasing circuit life. It is possible that with increasing circuit life there will be an increase in the number of clots observed in the circuit.

Although our results potentially have significant clinical implications, this single institution study has several limitations. Increase in patient survival may be related to differences in severity of illness and/or improved overall ECMO care. As this is a retrospective review, we cannot control for other factors that might have had an effect on the observed outcome.

Conclusion

A delicate balance between the use of systemic anticoagulation to reduce thrombotic events and prevent hemorrhagic complications is required in pediatric ECMO patients. We show that the use of a comprehensive ECMO anticoagulation laboratory protocol that includes antifactor Xa assays for unfractionated heparin titration, thromboelastography, and AT measurement is associated with decreased blood product administration, decreased hemorrhagic complications, increased circuit life, increased survival, and significant blood product cost savings. Future prospective evaluation is needed to establish an antifactor Xa assay-driven ECMO anticoagulation strategy as both clinically superior and cost-effective.