Comparison of Routine Laboratory Measures of Heparin Anticoagulation for Neonates on Extracorporeal Membrane Oxygenation

Abstract

Our objective was to determine the best measure of heparin anticoagulation in neonatal patients on extracorporeal membrane oxygenation. Activated clotting time (ACT), activated partial thromboplastin time (aPTT), and antifactor Xa levels, along with corresponding heparin infusion rates and heparin bolus volumes, were collected from neonates receiving ECMO at our institution from 2008 to 2013. After natural log transformation of antifactor Xa, ACT, and aPTT, overall correlations between antifactor Xa levels and either ACT or aPTT and correlations between these tests and heparin infusion rates were evaluated using linear mixed models that accounted for both within- and between-patient correlations. Twenty-six neonates with an average weight of 3.4 kg (standard deviation .7) had a total of 27 separate ECMO runs during the study period. Within each patient, ACT (r = .40, p < .0001) and aPTT (r = .48, p < .0001) were both directly correlated with antifactor Xa levels. In contrast, between patients, only aPTT maintained a direct correlation with antifactor Xa (r = .61, p = .07), whereas ACT showed a statistically significant inverse correlation with antifactor Xa (r = −.48, p = .04). Compared with ACT, aPTT is more consistently reflective of the anticoagulation status both within each patient on ECMO and between patients treated with ECMO. Future efforts to develop standardized heparin infusion algorithms for patients on ECMO should consider using aPTT levels to monitor anticoagulation.

Coagulations Studies Do Not Correlate With Each Other or With Hematologic Complications During Pediatric Extracorporeal Membrane Oxygenation*

“…Additionally, prior investigations have indicated that there is moderate to negligible correlation among anti-Xa, aPTT, and ACT values (7)(8)(9)(10)(11)(12). Henderson et al (12) showed in a retrospective study that ACT and aPTT had moderate correlation (r = 0.532), aPTT and anti-Xa had weak correlation (r = 0.205), and ACT and anti-Xa had weak correlation (r = 0.169).…”

“…In a single-center retrospective review of 604 patients, Baird et al (9) reported that UFH and ACT had moderate correlation (r = 0.48), although 18% of the study population was supported on ECMO for primary cardiac disease and the study spanned from 1980 to 2001, over 2 decades during which there may have been more practice variation. Liveris et al (7) found that UFH dose correlated better with anti-Xa, but multiple other studies demonstrated poor correlation (6,10). In a single-center retrospective cohort study of 69 neonatal and pediatric patients, McMichael et al (6) found that UFH dose had statistically significant but weak correlation with anti-Xa in neonates, infants, and children older than 1 year (ρ = 0.25, 0.26, and 0.25, respectively) (6).…”

“…Various measures have been explored to try to balance the risk of thrombosis and hemorrhage for children treated with ECMO (4), including monitoring ACT, aPTT, and anti-Xa levels. Previous studies have demonstrated that there is inconsistent and relatively poor correlation between UFH dose and coagulationrelated laboratory studies (2,(6)(7)(8)(9)(10). In a single-center retrospective review of 604 patients, Baird et al (9) reported that UFH and ACT had moderate correlation (r = 0.48), although 18% of the study population was supported on ECMO for primary cardiac disease and the study spanned from 1980 to 2001, over 2 decades during which there may have been more practice variation.…”

“…Antithrombin-III levels are often monitored since antithrombin-III is required for UFH activity (5). However, studies conflict as to whether these various coagulation-related laboratory measures correlate with UFH dose (2,(6)(7)(8)(9), whether they correlate with each other (2,(6)(7)(8)(10)(11)(12), and whether these laboratory values predict hemorrhage or thrombosis outcomes (2,6,8,9,(13)(14)(15)(16). The few studies that have investigated these correlations have been limited by small sample size, statistical models that fail to account for repeated measures on individual patients, and inclusion of mixed cardiac and noncardiac populations.…”

Poor Reliability of Common Measures of Anticoagulation in Pediatric Extracorporeal Membrane Oxygenation

“…However, we have identified two studies in the past decade evaluating PTT in children supported by ECMO that suggest it does not correlate well with heparin dosing. 10,16 Additionally, aXa has previously been studied in both neonatal and pediatric ECMO and shown to better correlate with heparin dosing. 10,17,18 Of importance, despite the statistically significant correlation between bivalirudin dosing and both PTT and R-time, both analyses had high degrees of variance as highlighted by the low R-squared values.…”