PB0082 - Investigating inter-assay variability between DOAC calibrated anti-factor Xa assays: A substudy of the PAUSE trial.

Background: DOAC calibrated anti-Xa assays can assess residual anticoagulant levels in patients requiring procedures or surgery. Previous studies have shown variability between anti-Xa levels determined by different reagent-instrument combinations, which may be related to their use of lyophilized plasma, DOAC-spiked plasma, or assaying at various laboratories. We determined anti-Xa levels in previously consented patients from the PAUSE trial which examined patients on anticoagulation who had interrupted factor-Xa inhibitor therapy before elective surgery.

Aims: 1.Describe inter-assay variability in apixaban and rivaroxaban anti-Xa levels using three coagulation reagent-instrument platforms.
2.Determine if these differences in anti-Xa levels are clinically relevant.

Methods: We analyzed apixaban (n=76) and rivaroxaban (n=61) anti-Xa levels on three different reagent-instrument combinations using manufacturer specific reagents, calibrators and quality controls - STA® reagents on the STA CompactMAX (Diagnostica Stago), HemosIL on the ACLTOP 300 (Werfen) and Biophen (Hyphen Biomed) on the BCS XP (Siemens). We analyzed inter-assay correlations at 2 levels ( < 30 ng/mL and ≥ 30 ng/mL) and compared median anti-Xa levels using analysis of variance.

Results: Anti-Xa levels determined using the 3 reagent-instrument combinations showed good-to-excellent correlation coefficients (r) for both apixaban (r=0.74–0.95) and rivaroxaban (r=0.64–0.92). For anti-Xa activity levels < 30ng/mL, there was a significant difference between Biophen/BCS XP and HemosIL/ACLTOP for apixaban, and for rivaroxaban when comparing HemosIL/ACLTOP to both Biophen/BCS XP and Stago/STA CompactMAX (Figure 1). No difference was found in the ≥30ng/mL group. In 7.3% (10/136) of samples, the reagent-instrument combinations discrepantly classified anti-Xa levels across the 30ng/mL threshold.

Conclusion(s): We demonstrated that anti-Xa levels determined by three common reagent-instrument combinations correlate well with each other. Although there are statistically significant differences between median anti-Xa levels in the < 30ng/mL group these differences are not clinically significant and result in discrepant classification across the 30ng/mL threshold in only 7.3% of samples.