Frequency of Anti-PF4/Heparin Antibodies in Patients Treated with Extra-Corporeal Membrane Oxygenation and Detection By Sensitized Serotonin Release Assay of Pathogenic Heparin-Induced Thrombocytopenia Antibodies

Introduction: Extra-Corporeal Membrane Oxygenation (ECMO) provides circulatory support in case of severe cardiac failure and is associated, as cardiopulmonary bypass, with strong activation of platelets, which likely leads to the release in patient blood of large amounts of platelet factor 4 (PF4). Therefore, systematic use of unfractionated heparin (UFH) for maintaining patency of circuits and to prevent thrombosis, exposes patients to heparin-induced thrombocytopenia (HIT). However, the risk of HIT in this particular clinical setting remains poorly documented. Objectives: In this monocentric study, the frequency of development of antibodies to PF4 modified by heparin (anti-PF4/H Abs) was prospectively evaluated in patients who had underwent ECMO, and the impact of these Abs on platelet count (PC) and clinical course was assessed. Patients and Methods: From February 2014 to January 2018, we enrolled in the University Hospital of Tours all consecutive adult patients with ECMO and treated with UFH for at least 5 days. Plasma samples were collected daily, and PF4-specific Abs were systematically detected using HAT45 ® (GTI), an ELISA assaying IgG, IgA and IgM isotypes, and another kit specific for IgG to modified PF4 (HAT45G ® ; GTI). Serotonin release assay (SRA) was also performed to detect platelet-activating Abs (and therefore potentially pathogenic) in plasma samples containing significant levels of anti-PF4/H IgG. Noticeably, SRA was also done after adding 10 μg/ml of PF4 in the reaction mixture (PF4-SRA), since we recently showed this modification might improve the detection of pathogenic HIT antibodies (Vayne et al, Brit J Haematol, 179(5):811-819, 2017). Demographics, PC, and clinical events were collected in all patients until the end of ECMO. Results: Fifty-seven patients with a median age of 57 years (range: 24-76) and who were mainly men (sex ratio W/M: 0.3) were enrolled. ECMO was performed due to medical issues (dilated cardiomyopathy, myocardial infarction, acute myocarditis) in 70 % of cases and in surgical settings in the others. Significant titres of anti-PF4/H IgG/A/M Abs were detected before ECMO in 6 of 57 patients (10.5%), with anti-PF4/H IgG present in only 3 cases (5%). After ECMO was initiated, 28 patients (49%) developed IgG/A/M anti-PF4/H Abs, with anti-PF4/H IgG present in 17 of them (30%). Half of these patients were immunized within 5 days after ECMO initiation and 95% of them within 10 days, with a median delay of 8 days before Abs detection (range: 4-26 days). Only 3 patients (5%) exhibited pathogenic anti-PF4/H IgG with positive SRA and the level of platelet release was always stronger when the test was performed with exogenous PF4 (SRA-PF4). Moreover, in one patient, pathogenic HIT Abs were detected earlier by PF4-SRA, i.e. 2 days before conventional SRA was positive. An abnormal PC pattern was evidenced in 2 of the 3 patients with positive SRA, with a dramatic PC fall after day 5 (of 49% and 31.5 % between days 9-10 and 5-6 respectively). The diagnosis of HIT was confirmed in these 2 cases, which also had suggestive clinical manifestations (arterial thrombosis and cutaneous necrosis). In the third patient, HIT was not suspected, because of low PC before ECMO with persistent thrombocytopenia post-operatively and early death on day 9. In the patients with negative SRA, the development of anti-PF4/H IgG did not appear to impact the biological and clinical evolution, with similar rates of thrombosis and death than in non-immunized patients. Therefore, patients with non-pathogenic anti-PF4/H IgG displayed the same PC pattern than non-immunized patients, with decreasing PC from ECMO initiation until day 5, followed by a slow PC recovery. Conclusion : The development of anti-PF4/H Abs is frequent during ECMO but these antibodies are rarely pathogenic, and not associated with a less favourable prognosis. PC evolution analysis during ECMO appears reliable for suspecting HIT, and a dramatic PC fall or a non-recovery of PC after day 5 should prompt to look for platelet-activating anti-PF4/H Abs. Therefore, IgG specific immunoassays should always be combined with platelet activation tests to improve the specificity of HIT diagnosis in this particular clinical condition. Interestingly, the addition of exogenous PF4 when performing SRA could allow an earlier detection of HIT pathogenic antibodies in patients undergoing ECMO. Disclosures No relevant conflicts of interest to declare.