Objective: Evaluate whether the use of a heparinized flush for an arterial catheter with a closed-loop blood sampling system leads to erroneous coagulation studies.
Design: Prospective study.
Setting: Twenty-two-bed surgical adult intensive care unit in a university hospital.
Patients: Sixty patients, 30 in each phase of the study.
Interventions: Phase 1: coagulation studies on a blood sample from a venous puncture and an arterial blood sample from an arterial catheter with heparinized flush. Phase 2: the same protocol but the arterial catheter was flushed with saline.
Measurements: Activated partial thromboplastin time, fibrinogen, percentage of prothrombin time, and international normalized ratio of prothrombin time on the venous and arterial samples.
Main results: Activated partial thromboplastin time in the arterial blood samples taken from an arterial catheter with heparinized flush was significantly prolonged compared with the venous controls. This was not the case in blood samples from an arterial catheter with saline flush. The magnitude of this difference found by the Bland and Altman analysis was clinically significant. In 23.3% of the patients in phase 1, there was a difference of >50% between the arterial and venous sampled activated partial thromboplastin time, compared with 0% of the patients in phase 2 (p =.018). There was no difference between the venous and the arterial blood samples for the non-heparin-sensitive coagulation studies in both phases of the study. An explanation for these findings could be that there was release of heparin bound to the arterial catheter into the blood sample.
Conclusion: A heparinized flush solution for the arterial catheter, when used together with a closed-loop blood sampling system, leads to erroneous results of heparin-sensitive coagulation studies. Heparin-sensitive coagulation studies, therefore, should not be analyzed on blood samples from such a system if a heparinized flush solution is used.