How should perfusion teams interpret new real-world evidence comparing unfractionated heparin plus antithrombin concentrate with bivalirudin during adult extracorporeal membrane oxygenation (ECMO)? A retrospective study found associations favoring heparin plus antithrombin for recorded thrombosis, bleeding, and hospital length of stay. However, these findings do not establish treatment superiority, a survival benefit, or significant cost savings.
What the investigators studied
Amanda Arenella and colleagues analyzed two US data sources: TriNetX electronic health records from 2014–2023 and the Premier Healthcare Database from 2018–2021. Adults receiving antithrombin concentrate plus unfractionated heparin (ATc + UFH) or bivalirudin around ECMO initiation were compared using separate 1:1 propensity-matched cohorts.
The matched TriNetX cohort included 552 patients, with 276 per treatment group, and supplied clinical outcomes. The matched Premier cohort included 528 patients, with 264 per group, and supplied hospital length-of-stay and cost outcomes. These were separate analyses, rather than clinical and economic measurements in one shared cohort.
Patients undergoing cardiopulmonary bypass or coronary artery bypass grafting during the ECMO hospitalization, and patients receiving both antithrombin concentrate and bivalirudin, were excluded. The bivalirudin cohort also excluded patients with a documented prior history of heparin-induced thrombocytopenia (HIT). Treatment escalation served as an operational proxy for heparin resistance; a standardized laboratory-confirmed definition was unavailable.
Key findings for perfusion teams
- Recorded thrombosis: 10.1% with ATc + UFH versus 19.6% with bivalirudin in TriNetX. The adjusted odds ratio was 0.41 (95% CI 0.24–0.69). The association was largely driven by deep vein thrombosis.
- Recorded bleeding: 9.1% versus 18.5%, respectively; adjusted odds ratio 0.38 (95% CI 0.21–0.65).
- Mortality and HIT: neither outcome differed significantly between groups. Recorded mortality was unusually low, and death timing was limited to month and year.
- Hospital length of stay: the adjusted incidence rate ratio was 0.85 (95% CI 0.74–0.99), favoring ATc + UFH in the Premier cohort.
- Total hospital costs: mean costs were $271,488 versus $299,947 in 2021 US dollars, but the adjusted comparison was not statistically significant: incidence rate ratio 0.98 (95% CI 0.87–1.12).
The reported odds ratios describe associations in these datasets. They should not be read as equivalent percentage reductions in an individual patient’s absolute risk.
Why the comparator matters
This study compared heparin supplemented with antithrombin against bivalirudin. It did not compare heparin alone with bivalirudin or isolate the incremental effect of antithrombin supplementation. Consequently, it cannot establish whether adding antithrombin improves outcomes compared with otherwise identical heparin treatment.
Heparin’s anticoagulant effect depends on antithrombin, whereas bivalirudin directly inhibits thrombin. That mechanistic difference makes the comparison relevant to discussions of heparin responsiveness. Nevertheless, the study does not establish a supplementation threshold, dosing regimen, monitoring target, or preferred strategy for an individual ECMO patient.
Study limitations and interpretation
Propensity matching can balance measured characteristics but cannot remove unmeasured confounding or differences in why clinicians selected each treatment. Institutional protocols, illness severity, anticoagulation monitoring, and treatment selection may have influenced the observed outcomes. Some regional and hospital characteristics remained imbalanced after matching.
Retrospective electronic records and claims are vulnerable to incomplete capture and coding errors. Circuit changes and several ECMO-specific outcomes were unavailable, intervention timing could not be determined in Premier, and heparin resistance could not be consistently confirmed. The small number of recorded deaths and limited date precision make mortality findings especially difficult to interpret.
The publication also contains reporting inconsistencies. Abstract cohort totals do not match the stated group counts, and bivalirudin length-of-stay values differ between Table 2 and the narrative. This summary uses the matched cohort sizes, clinical percentages, and adjusted effect estimates reported in the detailed results. These discrepancies reinforce the need for cautious interpretation.
Practical relevance
For perfusion teams, this paper provides a hypothesis for further comparative research and a useful topic for multidisciplinary review of local ECMO anticoagulation practice. It does not justify a universal switch from bivalirudin to heparin plus antithrombin. Prospective studies with consistent definitions, detailed monitoring data, and reliable ECMO-specific outcomes are needed to determine whether the observed associations reflect a treatment effect.
Funding and disclosures
The study was supported by Grifols Shared Services North America. Three authors reported employment with Grifols; other authors reported Grifols consulting or advisory relationships. These disclosures are relevant when appraising the findings alongside the study’s methodological limitations.
Source: Arenella A, Zhou J, Oh J, Calloway M, Howden R, Bader SO. Real-world outcomes of unfractionated heparin plus antithrombin concentrate versus bivalirudin in adult patients undergoing extracorporeal membrane oxygenation. Journal of ExtraCorporeal Technology. 2026;58:237–245. doi:10.1051/ject/2026015.
Attribution: © 2026 The Authors. Published by EDP Sciences under CC BY 4.0. This IPA educational summary is independently paraphrased; the downloadable source PDF is unchanged. Cover artwork is a conceptual illustration.
