Heparin Vs Direct Thrombin Inhibitors in Patients with Previous HIT Undergoing Cardiothoracic Surgery: The Cleveland Clinic Experience

BACKGROUND: Heparin Induced Thrombocytopenia (HIT) is a potentially devastating antibody-mediated complication of heparin therapy in which autoantibodies activate platelets in the presence of heparin. Despite thrombocytopenia, HIT is strongly associated with thrombosis that can affect the arterial or venous systems. Heparin is generally avoided in patients with a history of HIT; however, it is still the anticoagulant of choice for cardiopulmonary bypass surgery (CPB). Experience with alternative anticoagulants such as direct thrombin inhibitors (DTI) during CPB is limited, and the outcomes of cardiac surgery after a diagnosis of HIT are poorly described. We therefore conducted a retrospective cohort study to compare the 30 day mortality rate, rate of new thromboembolic events and rate of bleeding episodes in patients with a history of HIT who were treated with heparin or DTI undergoing cardiothoracic surgery requiring CPB.

METHODS: We identified patients with a history of HIT who underwent cardiothoracic surgery requiring CPB at Cleveland Clinic between 1998 and 2017. We included patients with a history of HIT confirmed by the serotonin release assay. Patients with other prothrombotic disorders were excluded. Demographic and clinical data including details of surgical procedures, anticoagulation, and outcomes (mortality, bleeding, and thrombosis) were collected from the medical record. Bleeding was categorized into minimal (<750 mL), mild (750 ml- 1.5 L), moderate (1.5 L-2 L), and severe (>2 L) bleeding. We evaluated differences in the intraoperative heparin versus DTI groups using the Fisher exact test and T test for categorical and continuous variables, respectively. Associations with study outcomes (mortality, moderate to severe hemorrhage, and thrombotic events) were first evaluated in univariate analyses and then in separate multivariable logistic regression models.

RESULTS: Of 13,998 patients that underwent cardiothoracic at Cleveland Clinic between 1998 and 2017, 72 patients had a confirmed prior diagnosis of HIT. Of these, 59 received heparin and 13 received DTI intraoperatively. The heparin and DTI groups were not different in terms of age (69 versus 65, P=0.102), presence of thrombocytopenia (39.0% vs. 30.8%, P=0.580), time on CPB (P=0.376), heparin-PF4 ELISA optical density (P=0.336), or surgery type (P=0.820).Thirty day mortality was 8.5% and 0% in the heparin and DTI groups, respectively (P=0.277). Thrombotic events occurred in 15 (25.4%) of the patients on heparin and 1 (7.7%) on DTI (P= 0.164). Of the 72 patients, 23 (31.9%) experienced minimal bleeding, 19 (26.4%) experienced mild bleeding, 26 (31.6%) experienced moderate bleeding and 4 (5.6%) experienced severe bleeding. In the heparin group, 16 (27.1%) had minimal bleeding, 14 (23.7%) had mild bleeding, while 25 (42.4%) had moderate bleeding and 4 (6.8%) had severe bleeding. In the DTI group, 7 (53.8%) had minimal bleeding, 5 (38.5%) had mild bleeding, while 1 (7.8%) had moderate bleeding and 0 had severe bleeding. The anticoagulant used intraoperatively was not associated with mortality or thrombosis in multivariable logistic regression models adjusted for gender, and ELISA optical density. To identify associations with bleeding, we categorized bleeding as minimal to mild (class I-II) and moderate to severe (class III -IV). The type of anticoagulant (DTI versus heparin) was significantly associated with moderate to severe hemorrhage (OR 0.080 [95% CI 0.011-0.731], P=0.024) when adjusted for the presence of thrombocytopenia (OR 0.945 [95% CI 0.332-2.690], P=0.915), and time on CPB (2-4 hours, OR 2.74 [95% CI 0.24-31.89], P=0.421), and (>4 hours, OR 3.36 [95% CI 0.32-35.60], P=0.313).

CONCLUSION: For patients with a prior diagnosis of HIT that undergo subsequent cardiothoracic surgery requiring CPB, intraoperative use of DTI versus heparin is associated with a lower rate of moderate to severe hemorrhage (7.7% vs 49.2%, P=0.006) and may be associated with lower rates of 30-day mortality (0% vs 8.5%, P=0.277), and thrombosis (7.7% vs 25.4%, P=0.164). While the differences in mortality and thrombosis rates did not reach statistical significance in our small sample size, these results are concerning and indicate a need for further investigation, ideally in randomized multicenter trials.

View largeDownload slide

View largeDownload slide

Close modal