Heparin Anti-Xa Activity, a Readily Available Unique Test to Quantify Apixaban, Rivaroxaban, Fondaparinux, and Danaparoid Levels
Elodie Boissier,Thomas Senage,Antoine Babuty,Isabelle Gouin-Thibault,Bertrand Rozec,Jean-Christian Roussel,Marianne Sigaud,Catherine Ternisien,Marc Trossaert,Marc Fouassier,Karim Lakhal
DOI: https://doi.org/10.1213/ANE.0000000000005114
2021-03-01
Abstract:Background: Despite their usefulness in perioperative and acute care settings, factor-Xa inhibitor-specific assays are scarcely available, contrary to heparin anti-Xa assay. We assessed whether the heparin anti-Xa assay can (1) be used as a screening test to rule out apixaban, rivaroxaban, fondaparinux, and danaparoid levels that contraindicate invasive procedures according to current guidelines (>30 ng·mL-1, >30 ng·mL-1, >0.1 µg·mL-1, and >0.1 IU·mL-1, respectively), (2) quantify the anticoagulant level if found significant, that is, if it exceeded the abovementioned threshold. Methods: In the derivation cohort then in the validation cohort, via receiver operating characteristics (ROC) curve analysis, we evaluated the ability of heparin anti-Xa assay to detect levels of factor-Xa inhibitors above or below the abovementioned safety thresholds recommended for an invasive procedure (screening test). Among samples with relevant levels of factor-Xa inhibitor, we determined the conversion factor linking the measured level and heparin anti-Xa activity in a derivation cohort. In a validation cohort, the estimated level of each factor-Xa inhibitor was thus inferred from heparin anti-Xa activity. The agreement between measured and estimated levels of factor-Xa inhibitors was assessed. Results: Among 989 (355 patients) and 756 blood samples (420 patients) in the derivation and validation cohort, there was a strong linear relationship between heparin anti-Xa activities and factor-Xa inhibitors measured level (r = 0.99 [95% confidence interval {CI}, 0.99-0.99]). In the derivation cohort, heparin anti-Xa activity ≤0.2, ≤0.3, <0.1, <0.1 IU·mL-1 reliably ruled out a relevant level of apixaban, rivaroxaban, fondaparinux, and danaparoid, respectively (area under the ROC curve ≥0.99). In the validation cohort, these cutoffs yielded excellent classification accuracy (≥96%). If this screening test indicated relevant level of factor-Xa inhibitor, estimated and measured levels closely agreed (Lin's correlation coefficient close to its maximal value: 95% CI, 0.99-0.99). More than 96% of the estimated levels fell into the predefined range of acceptability (ie, 80%-120% of the measured level). Conclusions: A unique simple test already widely used to assay heparin was also useful for quantifying these 4 other anticoagulants. Both clinical and economic impacts of these findings should be assessed in a specific study.