Heparin-calibrated anti-factor Xa activity as a surrogate marker for direct oral anticoagulant levels in acute stroke care: a large observational single-center study

Abstract

Direct oral anticoagulants (DOAC) complicate therapeutic decision-making, as rapid assessment of anticoagulant activity is critical in emergency patients, particularly in acute ischemic stroke (AIS). The administration of systemic thrombolysis (IVT) in anti-coagulated patients requires careful risk–benefit evaluation, given the potential for hemorrhagic complications. Substance-specific DOAC plasma level measurements are often unavailable outside tertiary centers. This study evaluated whether widely available heparin-calibrated anti-factor Xa (anti-Xa) activity can reliably estimate clinically relevant DOAC levels and support urgent clinical decision-making. This retrospective single-center study (2015–2023) analyzed 969 patients treated with apixaban, rivaroxaban, or edoxaban. All patients had parallel measurements of anti-Xa activity and specific DOAC levels. Correlations were assessed using Spearman’s coefficient. Receiver operating characteristic (ROC) analyses were performed to derive anti-Xa cut-off values corresponding to DOAC concentrations of 30, 50, and 100 ng/mL. Anti-Xa activity strongly correlated with DOAC plasma levels for all substances (Spearman’s ρ ≈ 0.95–0.96). ROC analyses demonstrated excellent diagnostic accuracy, with areas under the curve ≥ 0.97 for all evaluated thresholds. Similar anti-Xa cut-offs were observed for apixaban and rivaroxaban, whereas these were lower for edoxaban. For all DOACs combined, anti-Xa cut-offs of approximately 0.36U/mL, 0.56U/mL, and 1.19U/mL corresponded to DOAC levels of 30, 50, and 100ng/mL, respectively, with high sensitivity and specificity. Heparin-calibrated anti-Xa activity is a reliable surrogate marker for clinically relevant DOAC levels. Defined anti-Xa cut-offs are highly specific and sensitive and enable rapid and standardized assessment of anticoagulant activity in this cohort.