Thromboelastography to monitor anticoagulation in ICU patients receiving a continuous infusion of unfractionated heparin.

In intensive care unit (ICU) patients receiving unfractionated heparin (UFH) infusion, activated partial thromboplastin time (aPTT) and anti-factor Xa (anti-Xa) levels have limited correlation and significant discordance. We aimed to test whether thromboelastographic-derived citrated kaolin reaction time (CKR) and/or the difference (R-difference) and/or ratio (R-ratio) between CKR and heparinase-corrected CKR might help resolve such discordance

METHODS

We obtained simultaneous triple-paired samples for aPTT, anti-Xa and thromboelastography (TEG) in ICU patients receiving UFH infusion. We assessed correlation between UFH dose and R-difference and concordance for therapeutic ranges between aPTT, anti-Xa, and CKR or R-difference or R-ratio. On multivariate analysis, we assessed factors associated with TEG R-based measurements

RESULTS

We studied 136 samples from 24 patients with a median APACHE III score of 50 and median UFH dose of 14.6 [10.5, 18.3] units/kg/h. for each 1 unit/kg increase in UFH dose, the R-difference increased by 2 min (95% C.I. 0.23-3.7, p < 0.05) but the correlation was weak (R2 = 0.1). The Cohen's kappa coefficient for agreement was 0.39 for CKR-aPTT and 0.17 for CKR-anti-Xa. Similar results were seen for R-difference and R-ratio. When aPTT and anti-Xa were discordant for therapeutic ranges, TEG did not resolve discordance. On multivariate analysis, CKR, R-difference and R-ratio were independently positively associated with UFH dose but also with APACHE III score and, negatively, with iCa

CONCLUSIONS

In ICU patients receiving UFH infusion, TEG R-based measurements have limited correlation and concordance with anti-Xa and aPTT for defined therapeutic ranges and do not help resolve their discordance.