The fibrinogen prothrombin time‐derived method is not useful in patients anticoagulated with low molecular weight heparins or rivaroxaban

Essentials

• Fibrinogen prothrombin time‐derived (FIBPT‐d) behavior in anticoagulated patients is under studied.
• FIBPT‐d method overestimates fibrinogen in rivaroxaban and low molecular weight heparin samples.
• Unfractionated heparin and dabigatran samples showed similar bias to the control group.
• Rabbit brain and human recombinant thromboplastin behavior was different in rivaroxaban samples.

Summary

Background

The fibrinogen prothrombin time‐derived (FIBPT‐d) method with photo‐optical coagulometers is easy and economical. However, there are few reports on the behavior of this test on samples from patients anticoagulated with direct oral anticoagulants or low molecular weight heparin (LMWH).

Objective

To compare fibrinogen results obtained with the Clauss (FIB C) method and the FIBPT‐d method with two thromboplastins in anticoagulated patients.

Population

The study population comprised 295 consecutive anticoagulated patients: 99 treated with vitamin K antagonists (VKAs), 49 treated with unfractionated heparin (UFH), 47 treated with LMWH, 50 treated with rivaroxaban, 50 treated with dabigatran, and 100 normal controls (NCs).

Methods

Dabigatran samples were analyzed by the use of FIB C with HemosIL Fibrinogen C or 100 NHI thrombin units mL^?1 reagents; rabbit brain and human recombinant thromboplastins with HemosIL PTFibrinogen HS plus (HS) and Recombiplastin 2G (RP) were used for FIBPT‐d method. Heparin and rivaroxaban levels were assessed with HemosIL Liq antiXa with specific calibrators; dabigatran levels were determined with the HemosIL Direct Thrombin Inhibitor Assay. All assays were performed on the ACL TOP platform in two laboratories. Percentage biases for the FIBPT‐d method versus the FIB C method were calculated by the use of Bland–Altman plots.

Results

Positive biases of the FIBPT‐d method versus the FIB C method with both thromboplastins were seen in NC samples (13.7% and 18.9% for HS and RP, respectively), but biases with HS in rivaroxaban and VKA patient samples were higher than that in NC samples, at 31.9% and 34.0%, respectively. LMWH patient samples showed higher bias than NC samples: 26.5% and 29.3.0% with HS and RP, respectively. UFH and dabigatran patient samples showed similar bias as NC samples.

Conclusion

The FIBPT‐d method should not be used in anticoagulated patients, because the FIBPT‐d mathematical algorithm has been validated only in normal subjects, so overestimation could occur in these patients.