Prospective study investigating the incidence of heparin- induced thrombocytopenia in patients treated with mechanical circulatory support

Abstract

The HILA study was a prospective, two-year, single centered, blinded observational study at the university hospital of Tübingen between 2018 and 2020 investigating all male and female patients ≥ 18 years with heparin application who underwent treatment with mechanical circulatory support systems due to cardiac and/or lung failure. These probands were analyzed for their incidence of HIT development, regarding clinical data as well as serological analyses on day 0, 1, 6 and 10 after MCS start. The primary aim was to investigate the incidence of HIT during MCS; secondary, to evaluate possible risk parameters for a HIT development as well as differences between different devices and their complications. Clinical data included device data, renal replacement therapy, conditions of initial health and at admission, circumstances of MCS implementation, trigger reasons and survival, as well as complications, such as bleeding measured with a newly developed Bleeding Assessment Index (BAI) – considering bleeding subcategories, circumstances and dynamics and as well interventions and transfusion requirements. Further complications evaluated were thrombocytopenia, thromboembolic events with subgroups and infections. 38 of 94 primarily included patients were excluded, mostly because of their untimely death. Overall 56 patients were evaluated in four study groups: vvECMO (group 01), simple vaECMO (group 02) and complex vaECMO (group 03) as well as LVAD (group 04). The complex vaECMO group consisted of patients with a primary need of vaECMO with a new re-implementation of any MCS after an unsuccessful recovery after explantation of their initial vaECMO or a combination of or a change to another MCS device – such as, but not limited to, RVAD, Impella or IABP. Everyone had anticoagulation with heparin and only 12,5 % received alternative anticoagulation at some time point (mostly Argatroban). In 75% a RRT was required. A vast majority had relevant complications such as thrombocytopenia (93%), thromboembolic events (73%, mostly extracorporeal) and bleeding events (95%, mostly cannulation site bleeding). The most common bleeding severity was either minor (BAI °II, 48%) or minimal (BAI °I, 29%), measured with the newly developed Bleeding Assessment Index, which was specifically developed for this study to evaluate and compare bleeding dynamics during and after MCS over time of each patient as well as between different patients and subgroups over their time course on ICU. Probands were approximately 16 days in need of any MCS, remained 21 days on ICU and nearly a month in hospital. 68% were discharged from hospital and the 30 day mortality was 25% for the whole evaluated study cohort, but over 40% when considering patients who were initially excluded due to their untimely death. In regard of the MCS study groups some thromboembolic subcategories were associated with a significantly lower survival, e.g. vvECMO patients with extracorporeal thromboembolism and LVAD patients with ischemic TE. Within our study a HIT diagnosis was not associated with a higher mortality, although the mortality was significantly increased when alternative anticoagulation was applied and when a higher bleeding activity occurred. Over 90% (206/221) of the required blood samples were obtained and analyzed for study purpose. There were many high false positive results for clinical suspicion (nearly 60%; 4T Score) and sero-conversion (around 40%; ELISA) which was clearly time related. The incidence of platelet activation was only slightly more than 5% (3/56), as only three patients showed positive HIPA results – all on PID 6 and/or 10, and all were part of the complex vaECMO group, each representing one of the subgroups – MCS change, combination or re-implementation. As only one of them was suspected outside of our study concept to have a HIT, just one of the three HIT positive patients was changed from UFH to Argatroban. Unfortunately he was also the only HIT positive patient who died. We also found a possible rapid inactivation of beforehand confirmed and activated HIT-antibodies shortly after MCS explantation despite the ongoing heparinization (re-implementation within 48 hours). Furthermore the ongoing application of UFH despite a confirmed HIT seemed not to influence the overall survival and was not proved to be significantly relevant for further complications.

With 5,36% (3/56) is our overall HIT incidence for different MCS in the higher range for HIT among several published trials (0,1 bis 5%). Via a conducted ROC analysis we confirmed a cutoff value of 4 points for the 4T Score (sensitivity 100%; specificity 62%) For the ELISA testing an OD of 0.454 proved to have the best possible cutoff (sensitivity 100%; specificity 88%). In our study, a confirmed HIT was not associated with a clearly higher 4T score or Optical Density than other patients under MCS, especially in the case of clinical suspicion. However, our complex vaECMO group presented a much higher incidence with 21,43% (3/14), so further investigation should be done with it’s focus on special subgroups of MCS and their combinations as assumingly patients with complex vaECMO are especially at risk for HIT development. Complications were rather common but not necessarily associated with any HIT development within the overall group. Only study group 03 – where all HIT cases were found – revealed significantly more thromboembolism than other MCS groups (neglecting LVAD). Prothrombotic effects of artificial surfaces of MCS and RRT could be masking and overlapping the manifestation of HIT relevant thromboembolic events. We recommend the establishment of a standardized HIT screening under MCS. Our suggestion is to screen every patient with more than 6 days under MCS - especially in the case of complex MCS and with relevant thrombocytopenia – via an established clinical and serological test (e.g. 4T Score and ELISA) every 3 days with a clear confirmation via HIPA or another confirmation test, if both screening tests are positive. At least until further research gives valid reason to decide different, for now, we agree with the current cutoff for the 4T Score (≥ 4 points) and the ELISA testing (OD ≥ 0.300). If a HIT is confirmed we suggest another routine HIT testing – including the confirmation test – e.g. one week after MCS explantation to evaluate a possible inactivation of HIT antibodies. We suggest also to await the results of the confirmation test before any anticoagulation is changed. Altogether we do not recommend the change of anticoagulation on clinical suspicion alone as it puts a very high number of patients under unduly risks. An adapted 4T Score for MCS could help to navigate also, e.g. with a time line referenced to MCS start and not to heparin application start only and with less importance for thrombocytopenia with a maximum of one point only for this category, but more points in case of later and ongoing thrombocytopenia longer than 6 days. Considering the questionable impact of a HIT diagnosis for this already severely impaired clientele and the clearly relevant complications in case of an unnecessary change of anticoagulation an overestimation of HIT suspicion should be avoided. To what extent a HIT influences the course and outcome of already severely ill patients in need of MCS cannot sufficiently be answered due to the lack clearly defined parameters to distinguish the clinical relevance of any HIT in this patient clientele. To identify those parameters a rather large cohort of patients with unmistakably confirmed HIT is necessary where different risk levels could be defined and further stratified.