Diagnosis of heparin-induced thrombocytopenia (HIT) is challenging in cardiac surgery patients, as nearly all develop postoperative thrombocytopenia, and 10–50% have platelet factor 4 (PF4) IgG antibodies after surgery. Pre-test probability scores for HIT have not been widely used by cardiac surgeons at our institution due to concerns regarding thromboembolic complications in untreated HIT, leading to overuse of direct thrombin inhibitors (DTIs) as empiric therapy for postoperative thrombocytopenia.
DESIGN/METHODS: From August through September 2010, we incorporated a hematologist into work rounds in the Brigham and Women's Hospital (BWH) Cardiac Surgery Intensive Care Unit (CSICU). The hematologist collaborated with cardiac surgeons to create literature-based, multidisciplinary guidelines on hematology topics in CSICU patients. For postoperative thrombocytopenia, the guidelines (1) discouraged empiric use of DTIs on postoperative days (POD) 0–3, except in suspected catastrophic HIT; (2) allowed for empiric DTI use on POD ≥ 4 while awaiting PF4 antibody test results; and (3) suggested discontinuation of DTIs after two negative PF4 antibody tests drawn on POD ≥ 4. We performed a retrospective analysis of projected DTI use in BWH cardiac surgery patients based on the guidelines.
RESULTS AND DISCUSSION: From February 2010 through February 2011, 80 cardiac surgery patients received DTIs for suspicion of HIT. The median age was 72 years (range, 30–90). Forty-two patients were male. The most common procedures were aortic valve replacement (n = 36), mitral valve replacement (n = 32), and coronary artery bypass grafting (n = 31). Six required an intraaortic balloon pump. Nine were reoperations. PF4 antibody tests, performed using polytypic (Diagnostica Stago) or IgG-specific (Gen-Probe GTI Diagnostics) PF4 antibody ELISA assays, were positive in 19/80 cases (23.75%). Among a subset of 20 patients for whom pre-test probability scores were calculated, correlation between 4 T's and Lillo-Le Louet scores was poor (high probability score, 45% and 15%, respectively; concordance, 50%; p = 0.109 (McNemar's test)).
Initiation of DTIs was in disagreement with the guidelines in 17/80 cases (21.25%), of which PF4 antibody tests were negative in 15 and positive in 2. Of the 15 PF4 antibody-negative cases, 12 had no evidence of thrombosis; two developed superficial or deep venous thromboses but tolerated heparin reexposure, and one with multiple negative PF4 antibody tests developed a radial artery thrombosis on POD 9 and was treated with heparin and thrombectomy. Both of the PF4 antibody positive cases were unusual presentations of HIT; in one case, the guidelines would have supported delaying initiation of DTIs by one day, which would not have changed overall clinical course for this patient.
Patients in whom DTIs were started on POD ≥ 4 were more likely to test positive for PF4 antibodies than patients in the POD 0–3 group (50% vs 11.1%, respectively; p = 0.0004 (Fisher's exact test)). Almost all cases in which initiation of DTIs disagreed with the guidelines occurred in the POD 0–3 group rather than the POD ≥ 4 group (16 vs 1, respectively; p = 0.008). Overall, the guidelines would have delayed DTI initiation by approximately 1–2 days in 20% of patients and shortened duration of DTI therapy in roughly 15%, leading to a 163-day reduction in total bivalirudin use, a 7-day reduction in argatroban use, and a projected yearly cost savings of $81,414.29–83,055.82 in DTI usage, with minimal impact on patient outcomes. Since launching of these guidelines and other coordinated hospital-wide efforts, overall DTI use at BWH has decreased markedly and stabilized.
Our analysis suggests a potential role for multidisciplinary clinical care models incorporating hematologists into surgical and critical care units.
Fanikos:Baxter: ran a symposium at American Society of Health-System Pharmacists.
Asterisk with author names denotes non-ASH members.