Concordance of Activated Partial Thromboplastin Time and Anti‒Factor Xa Levels in Monitoring Heparin Therapy Among Cardiac Surgery Patients in Pre- and Postoperative Settings.

OBJECTIVES

To quantify aPTT‒anti-Xa assay concordance in patients on heparin therapy undergoing cardiac surgery and identify factors influencing discordance in the perioperative setting.

DESIGN

Retrospective single-center cohort study.

SETTING

A quaternary care hospital.

PARTICIPANTS

Adults (≥18 years) undergoing cardiac surgery with cardiopulmonary bypass on perioperative heparin infusion.

INTERVENTIONS: None.

MEASUREMENTS AND MAIN

RESULTS

We collected 1,482 paired aPTT and anti-Xa measurements from 155 adult cardiac surgery patients. After the exclusion of concordantly subtherapeutic samples, the effective sample size was reduced to 589 paired samples from 112 unique patients (median age 68 years, 29% female). Surgical procedures most commonly included coronary artery bypass grafting (45%, 50/112), combined valve coronary artery bypass grafting (31%, 35/112), and ventricular assist device implantation (9.8%, 11/112). Median cardiopulmonary bypass time was 147 minutes (IQR: 94-186). Among the effective sample size, 377 paired measurements were taken from the preoperative period and 212 samples were from the postoperative period. In the preoperative cohort, 48.3% (182/377) were discordant compared to 77.8% (165/212) in the postoperative cohort (p < 0.0001). Unadjusted mixed-effects logistic regression models demonstrated that diabetes was the only significant predictor of assay discordance in the postoperative cohort, with patients with diabetes having 68% lower odds of having discordant measurements (OR = 0.32, 95% CI: 0.11-0.89, p = 0.031).

CONCLUSION

Significant discordance exists in aPTT‒anti-Xa assay results, particularly in the postoperative period. Diabetes was the only significant predictor of discordance, suggesting metabolic status may influence assay agreement. These findings highlight the importance of further investigation into optimal perioperative monitoring strategies.