Discordant High Activated Partial Thromboplastin Time Relative to Anti-Xa Values in Hospitalized Patients is an Independent Risk Factor for Increased 30-day Mortality.

Abstract

Background:  The activated partial thromboplastin time (aPTT) and anti-factor-Xa levels (anti-Xa) are both used to monitor patients on unfractionated heparin. Our previous study demonstrated that patients with discordant high aPTT relative to anti-Xa had higher rates of mortality and bleeding events.

Objective:  To determine if underlying patient characteristics drive both discordance and adverse outcomes or if discordance is an independent risk factor to adverse outcomes.

Methods:  We analyzed all patients hospitalized at the Stanford Hospital between January 2011 and December 2019 who had simultaneous aPTT and anti-Xa levels performed. From the electronic medical record, we extracted and analyzed 51 patient features including baseline coagulation laboratory results, demographics, values of other common laboratories (basic metabolic panel, complete blood count, etc.), diagnostic procedures, medications, and death.

Results:  A total of 17,728 patients had 78,701 paired aPTT and anti-Xa levels. Patients with discordant aPTT and anti-Xa where aPTT (seconds) was elevated beyond the expected therapeutic range had a higher 30-day mortality (odds ratio [OR]: 2.16, 95% confidence interval [CI]: 1.78-2.63, p < 0.001). Sectioning the patients based on the degree of discordance and whether aPTT or anti-Xa were signaling excess anticoagulation, we found those with an elevated aPTT discordant to their anti-Xa level had the highest odds of death (OR: 2.46, 95% CI: 1.99-3.10) compared with the concordant group. This finding was still present after controlling for patient comorbidity and other laboratory results at hospital admission.

Conclusion:  After controlling for patient features strongly associated with increased mortality in heparinized patients, we identified that the discordant pattern of high aPTT to anti-Xa served as an independent predictor of 30-day all-cause mortality, with a higher degree of discordance associated with increased odds of 30-day mortality.