Anticoagulation management in patients tested for heparin-induced thrombocytopenia in the cardiothoracic intensive care unit diverges from expert guidelines

Abstract

Abstract Heparin-induced thrombocytopenia (HIT) is a rare, life-threatening complication in patients receiving heparin products. Clinical scoring tools, such as the ‘4T’ score, aim to help identify patients at risk for HIT and are recommended by expert guidelines. For example, the American Society of Hematology (ASH) recommends that patients with an intermediate or high-risk 4T score are presumptively treated for HIT by withdrawing heparin and starting an alternative anticoagulant. At our institution, laboratory HIT testing involves an in-house latex immunoassay (LIA) that, if positive, reflexes to a send-out serotonin release assay (SRA) at the Blood Center of Wisconsin. This confirmation may take several days to result, underscoring the importance of presumptive treatment. Anecdotally, we have observed that clinicians have expressed reservations about the use of the 4T score in the cardiothoracic intensive care unit (CTICU). These reservations may lead to increased HIT testing in patients with low 4T scores and delays in the initiation of alternative anticoagulants. We set out to objectively determine if clinical practice in the CTICU is aligned with ASH guidelines. To do so, we used anticoagulant usage as a proxy for medical decisions made at the time of order for HIT laboratory testing. Data was extracted from the electronic medical records of all patients who were admitted to the CTICU between 5/2018 and 7/2022 and administered a heparin product (n = 1,602). Patients that were not tested for HIT (n = 1,449) were excluded. The remaining patients’ (n = 153) medication record timings were normalized such that the time of their first LIA was set to zero to allow for relative comparisons. 25 (16%) of the LIAs were positive, of which 5 (20%) were confirmed by SRA. Of the 153 patients, 136 (89%) received heparin within 24 hours prior to LIA order, while 6 (4%) were already on alternative anticoagulation (bivalirudin). Of the patients on heparin prior to LIA, 18 (13%) were switched to bivalirudin upon LIA order, 66 (49%) had heparin discontinued without an alternative, and 52 (39%) were continued on heparin. Our results suggest that the prevalence of HIT in this cohort is 3.3% by SRA, much lower than the expected pretest probability estimated for intermediate- and high-risk 4T scores in published meta-analyses. In addition, the observed anticoagulation practices varied widely and diverged from ASH guidelines, where only 17% of patients were on alternative anticoagulation when the testing was ordered. Future work will investigate these patterns across LIA positivity and 4T scores while attempting to address the clinical outcomes of these treatment decisions.