Performance Assessment of Anti-Xa Assay–Based Heparin Dosing Protocol in Pediatric Patients on Extracorporeal Membrane Oxygenation

Background The use of extracorporeal membrane oxygenation (ECMO) in the postoperative cardiac critical care setting is evolving. Anticoagulation monitoring is among the most challenging aspects of pediatrics. However, there is no consensus on the optimal dosing and monitoring of unfractionated heparin in this setting. To address this, we developed an anti-Xa assay–based protocol derived from the best available clinical and anecdotal evidence of ECMO use and assessed its effectiveness in achieving the anti-Xa assay therapeutic target. Methods This prospective single-arm study was conducted in the pediatric carcardiac-surgery intensive care unit of a large tertiary hospital. We used two different anti-Xa assay intensity levels based on the patients’ bleeding status. Results The median patient age was 7 (interquartile range [IQR]: 5-11.25) months, and the median weight was 5.7 (IQR: 3.8-13.82) kg. The median ECMO duration was 6 (IQR: 4.5-7.5) days. The bleeding protocol was used for most patients. Seventy percent achieved the anti-Xa assay therapeutic target during the study period (median: 75.5 h, IQR: 60.5-117.5 h). Hemorrhagic complications were reported in 40% of the patients, and thrombotic complications were reported in 25%. The median length of stay was 37 (IQR: 22-43) days, with a survival-to-discharge rate of 75%. Conclusions Despite a failure to achieve the anti-Xa assay target within the first ECMO days, most patients achieved the target by the median ECMO duration. Moreover, using two different anti-Xa assay levels reduced thrombotic complications.


Introduction
Extracorporeal membrane oxygenation (ECMO) is used in various critical situations to support cardiopulmonary circulation.2][3][4] Hence, two ECMO modes are available: venoarterial ECMO (VA-ECMO), which assists with cardiac and respiratory dysfunction, and venovenous ECMO (VV-ECMO), which supports only respiratory dysfunction. 3CMO use in pediatric patients in postoperative cardiac critical care settings is evolving. 3,5,6Extracorporeal Life Support Organization (ELSO) data indicate that approximately 1,500 to 2,000 children of neonatal age and older are supported by ECMO annually. 7xtracorporeal membrane oxygenation has been proven to improve patient outcomes and decrease mortality.However, hemorrhagic and thrombotic complications are still a concern. 3,8,9Antithrombin is the main natural anticoagulant that inhibits activated coagulation proteases, such as factors IIa and Xa and, to a lesser extent, factors XIa and XIIa. 10,11everal studies have revealed considerable variability in developmental hemostasis across pediatric age groups. 12,13mostasis is not fully developed in neonates and infants.Some fibrinolytic system components, such as vitamin K-dependent coagulation factors, protein C, and antithrombin III (AT III), are low in the first few months of life and increase gradually with age. 12In contrast, other factors, are more active in this group. 13,14Therefore, adequate anticoagulation dosing and monitoring in pediatric patients supported by ECMO are fundamental. 3,15nfractionated heparin (UFH) is the most commonly used anticoagulant in this inpatient setting due to its fast action onset, short half-life, antidote availability, and reasonable cost. 14UFH acts mainly by binding to a specific pentasaccharide sequence on antithrombin.This potentiates antithrombin's inhibitory effect on factor Xa activity. 10,11ver the years, ECMO has attracted clinicians' interest and has continuously improved. 16However, there has been no parallel improvement in necessary anticoagulation practices. 17A recent study reported significant variability in the types of anticoagulation tests used by different hospitals.These tests include activated partial thromboplastin time (aPTT), activated clotting time (ACT), and anti-Xa activity. 18Several studies have concluded that anti-Xa activity better correlates with UFH dosing in pediatrics compared to others. 5,8,11,18This test does not measure UFH concentrations; instead, it measures heparin's effect on enhancing antithrombin's inhibitory effect on factor Xa and, thus, its anticoagulation effect. 19,20t present, there is no consensus on a standard protocol for dosing and monitoring UFH in pediatric patients supported by ECMO.To address this gap, we developed an anticoagulation protocol for pediatric patients receiving ECMO, taking into consideration the best available evidence and local experience.The objective of this study was to investigate the implementation of anti-Xa assay-based heparin protocol and evaluate its performance and time to achieve the target range.

Study Design and Participants
A pilot prospective single-arm interventional study was conducted in the Pediatric Cardiac-Surgery Intensive Care Unit of King Faisal Specialty Hospital & Research Centre, a tertiary care hospital in Riyadh, Saudi Arabia.
All patients under the age of 18 years admitted to the pediatric cardiac surgical intensive care unit from January to December 2021 and supported by either VA-ECMO or VV-ECMO were included in the study.The exclusion criteria were ECMO for less than 24 h, the use of an anticoagulant other than UFH, and a disseminated intravascular coagulation diagnosis according to the International Society on Thrombosis and Hemostasis scoring system. 21The hospitalintegrated clinical information system was used to extract patient data.The data were entered into REDCap version 11, a secure web-based application.

Intervention
A protocol was developed (Supplementary file 1) after a comprehensive literature review and a strict review process by a multidisciplinary team led by clinical pharmacists and cardiac surgeons to ensure the protocol's safety and applicability.After it was completed and presented to the ECMO Steering Committee, the protocol went live for piloting.To ensure adherence to the protocol, multiple structured teaching sessions were provided to healthcare professionals, including consultants, assistant consultants, surgeons, and nurses.The protocol was designed to minimize communication within the team and make it a nurse-run protocol for dose adjustment.Moreover, to minimize deviations from the protocol, a regular adherence assessment was incorporated.To facilitate implementation, the protocol was integrated into our electronic health system.It was agreed with the hematology laboratory to optimize synchrony, thereby minimizing turnaround times for the anti-Xa assay.Clinical pharmacists were accessible for any protocolrelated queries.All issues related to the protocol and its implementation were regularly reviewed by the ECMO Steering Committee.
The protocol is divided into two categories: bleeding and nonbleeding.It guides healthcare professionals on how to manage UFH use throughout ECMO support.The protocol also identifies laboratory parameters that should be monitored before and after UFH initiation: initial and maintenance doses of UFH infusion, dose monitoring frequency, UFH dose adjustment, and criteria for switching between bleeding and nonbleeding protocols.The anti-Xa assay therapeutic range is 0.2 to 0.4 U/mL for the bleeding protocol and 0.3 to 0.7 U/mL for the nonbleeding protocol.

Endpoints
The primary endpoint was the protocol's ability to achieve the anti-Xa assay therapeutic target during ECMO.The secondary endpoint was the time required to achieve the target and to determine the time spent within the target range.Other endpoints included the incidence of thrombosis-namely, patient thrombosis or mechanical thrombosis complications (oxygenator failure, pump failure, circuit change, cannula clots, clots in any circuit component, or hemofilter clots)-and the time to replace the ECMO circuit or any component.
Regarding hemorrhagic events, surgical and nonsurgical site bleeding events (peripheral cannulation site bleeding, mediastinal cannulation site bleeding, and gastrointestinal hemorrhage), central nervous system hemorrhage, surgical exploration for bleeding, and time to surgical exploration were reported.The survival-to-discharge rate was also reported.

Data Analysis
Descriptive statistics were used to analyze the patients' demographics.Normally distributed continuous variables were expressed as means and standard deviations, while nonnormally distributed variables were expressed as medians and interquartile ranges (IQRs).Categorical variables were expressed as frequencies and percentages.It was determined that 20 subjects would be a sufficient sample size for this pilot study.This sample size was determined for the study's primary objective based on the binomial proportion associated with achieving the anti-Xa assay therapeutic range.It was specified to calculate a 90% confidence interval for that proportion, such that the width of the confidence interval would be no more than 0.2.With these constraints, a sample size of 17 subjects was required.To account for dropouts, the number was inflated to 20 subjects.All statistical analyses were performed using the Statistical Analysis System software package version 9.4 (SAS Institute Inc).

Ethical Considerations
This study was approved by the Institutional Research Advisory Council of Faisal Specialist Hospital & Research Centre, Riyadh, Saudi Arabia (number 2191293).Verbal informed consent was obtained from the patients' parents.

Results
Twenty-five patients were screened during the study period.
Five were excluded due to an ECMO duration of less than 24 h.All patients were admitted to the pediatric cardiacsurgery intensive care unit unit after cardiac surgery.The most common surgery performed was valve repair or replacement in nine patients (45%), followed by ventricular septal defect transcatheter repair (40%) and patent ductus arteriosus ligation or division in eight patients (40%).None of the screened patients had any coagulopathy disorder or received an anticoagulant other than UFH.Half of the patients were female.Their median age was 7 (IQR: 5-11.25)months, and their median weight was 5.7 (IQR: 3.8-13.82)kg.The mean estimated glomerular filtration rate according to the modified Schwartz equation was 80 ± 34.7 mL/min/1.73m 2 .The patients' characteristics are described in Table 1.
The main indications for ECMO were failure to wean from cardiopulmonary bypass after surgery in nine patients (45%) and bridging to heart transplant in two patients (10%).Eight patients (40%) exhibited a failure to wean because of cardiac arrest.The median ECMO duration was 6 (IQR: 4.5-7.5)days.The baseline anti-Xa assay before cannulation was 0.1 IU/mL, which remained the same 4 to 6 h after ECMO initiation.Another important baseline laboratory value was AT III.It was available for seven patients at baseline, with a median of 28% (IQR: 17.5%-30.5%).Based on a set of criteria listed in the protocol's bleeding section (Supplementary file 1), most patients (90%) underwent the bleeding protocol.A precanulation heparin bolus of 50 to 100 U/kg was administered to 11 patients (55%) who had an ACT of <300 s.The median UFH administration duration was 3.51 (IQR: 1.84-5.82)days.Holding UFH was required for eight patients (40%).The median holding UFH time was 13.5 (IQR: 10.5-27.5)hours.ECMO protocol-related characteristics are shown in Table 2.
Fourteen patients (70%) achieved the anti-Xa assay therapeutic target at different time points of the study period.The median time to achieve the target was 75.5 (IQR: 60.5-117.5)hours.The target was maintained for a median of 19.5 (IQR: 8-31.5)hours.By the sixth day, 91% of the patients had achieved the therapeutic target (Figure 1).The total number of hemorrhagic complications was reported in eight patients (8/20, 40%), and the total thrombotic complications were reported in five patients (5/20, 25%) within the cohort.The most common bleeding event, occurring in four patients (4/20, 20%), was central nervous system bleeding.Mediastinal cannulation and surgical site bleeding were reported in two patients (2/20, 10%).Both peripheral cannulation site bleeding and gastrointestinal bleeding occurred in one patient (1/20, 5%).

Comment
Establishment of an ECMO program is important.However, in pediatric patients with congenital heart disease, ECMO is challenging.One challenging aspect is dosing and monitoring anticoagulants in the absence of a universal protocol.In this study, we used a protocol with low-intensity anticoagulation for patients with bleeding and higher-intensity anticoagulation for nonbleeding patients.Under this protocol, most pediatric patients achieved the anti-Xa assay therapeutic target within a median follow-up period of three days.This demonstrates that the protocol can ensure reasonable anticoagulation upon ECMO initiation.Despite this, maintaining the therapeutic target was challenging.
One of the key factors for the success of the UFH protocol was developing system synchrony with the hematology laboratory to optimize turnaround times for the anti-Xa assay by labeling patients receiving ECMO so that they could be easily identified by the laboratory staff.In implementing this protocol, the main challenge was physicians' hesitancy to apply it.This issue was resolved by discussing the evidence from published studies and continuously monitoring the results.Moreover, it was crucial to ensure competency by providing structured training to the nursing staff applying ECMO to this patient population.All members of the medical team, including consultants, assistant consultants, and surgeons, participated in the protocol's initial development, helping to achieve a consensus and a full understanding of its implementation.Besides ensuring adherence to the protocol, we ensured the applicability of an approach led by nurses that recognizes their important role in ECMO patient care.These findings align with recently published recommendations for supporting evidence-based nursing care for patients receiving ECMO. 22he developed protocol uses only the anti-Xa assay for monitoring UFH, as it shows a higher correlation with UFH dosing and greater reliability than other tests. 5,8,14,18Our study is unique in that it protocolizes UFH use in post-cardiac surgery pediatric patients according to their bleeding status.Based on the bleeding status, we used two different levels of anti-Xa intensity to minimize the risk of bleeding.Although this was extremely important, it was also instrumental to agree on a clear definition of bleeding, as provided in this protocol (Supplementary file 1).The decision on the type of protocol used was made by a team of intensivists, surgeons, clinical pharmacists, and nurses.The protocol includes well-defined criteria for switching between high-and low-intensity anticoagulation based on bleeding status (Supplementary file 1).This marks a divergence from previous studies, which have used unified ranges regardless of bleeding risk. 5,18,23e comprehensively monitored laboratory parameters and blood product transfusion.The protocol includes a clear replacement guide focusing on AT III levels (Supplementary file 1).We assumed that a reasonable AT III level would optimize the effect of UFH. 5,24As we did not have ready-made AT III, such as Thrombate III or Thrombotrol-VF, at our hospital, we used fresh-frozen plasma.Our goal was to maintain the AT III level above 80%. 5,8,18his protocol is simplified in terms of monitoring test used and frequency.Unfractionated heparin adjustment is guided by anti-Xa assay level alone without the need for ACT or aPTT.In addition, our protocol has the advantage of less frequent sampling compared to other studies.A previous study used an anti-Xa assay to monitor 20 pediatric patients for a median ECMO duration of 88 h.The intensive protocol used in that study, which involved hourly anti-Xa assay monitoring until two subsequent readings were within the target range, followed by a gradual decrease to every 6 h (this timing was adopted to minimize blood sampling compared to ACT-based monitoring). 5In our study, the anti-Xa assay was measured every 8 h.If three consecutive readings were within the target, the reading interval was increased to 12 h.Another study compared two anti-Xa assay protocols: in one protocol, anti-Xa was measured every 12 h, while in the other, it was measured daily.The researchers found that daily monitoring resulted in significantly lower anti-Xa assay levels than twice-daily monitoring. 25Niebler et al also used twice daily measurements which resulted in fewer hemorrhagic complications and decreased demand for circuit change. 23alancing thrombosis and hemorrhage in ECMO patients is of fundamental importance.Forty percent of our patients had hemorrhagic complications.This rate was slightly higher than those reported by Bembea et al 9 and O'Meara et al 5 (33% and 27%, respectively).This can be explained by the postsurgical status of our patients and differences in defining hemorrhage.In contrast, thrombotic complications occurred in only 25% of our patients.This percentage was lower than that reported in a previous study, in which thrombotic complications occurred in 55.5% of the patients undergoing an anti-Xa assaybased protocol. 8Moreover, in our study, fewer patients needed a circuit/oxygenator change compared to a previous study (20% vs 35%). 5ur survival-to-discharge rate with the new protocol is in line with 2021 ELSO guidelines (60%-73%) 7 but higher than that in a previous study. 5Compared to our own historical data, anti-Xa-guided generally did better than ACT-guided protocol with regard to clinical and safety outcomes (data not shown).
Besides this study's prospective nature, its strengths include utilizing a nursing-run protocol to minimize care interruption and provide autonomy to the team.Furthermore, the developed protocol considers the patients' bleeding status and provides clear instructions for switching between high-and low-intensity anticoagulation, as indicated by the anti-Xa assay, to minimize the risk of bleeding and thrombosis.
Certain limitations of this study should be noted.One limitation is that it was conducted entirely in a cardiac surgery unit.Another limitation is that all included patients required VA-ECMO.Therefore, results might not be extrapolated to VV-ECMO.Furthermore, we did not use controls, as there is a scarcity of data for this subset of patients.
In conclusion, our findings show that an anti-Xa assaybased protocol that incorporates two different levels of anticoagulation intensity can achieve the target therapeutic range.Moreover, the early achievement of the target suggests a reduction in thromboembolic complications.Thus, this study supports the use of a protocol with two anticoagulation intensity levels as a standard of care.A larger study is warranted to standardize this protocol for pediatric ECMO cases.

Table 1 .
Baseline Characteristics of Study Patients.
a Some patients had more than one surgery.b Antithrombin III was available for seven patients.
a Criteria for selecting the bleeding protocol-Definition of bleeding is one of the following: bleeding resulting in hemodynamic instability AND drop in Hg ≥