Relationship between Cardiopulmonary Bypass Time, Platelet Count, Fibrinogen, ROTEM Measurements, Antithrombin Level, and Bleeding Amount during Cardiovascular Surgery and Intensive Care Unit Stay: A Observational Study

Purpose: Hemoglobin levels after a cardiopulmonary bypass (CPB) are easily estimated; however, decreased brinogen and platelet levels before and after CPB are dicult to predict because of the adsorption and consumption. We hypothesized that brinogen levels, platelet count, and measurements by ROTEM TM may decrease in proportion to CPB time; moreover, we compared the perioperative blood loss by conducting a observational study. Methods: A total of 160 patients were enrolled and divided into three groups depending on the CPB time: < 2, 2–3, and >3 h. Blood samples were obtained at four time points, i.e., baseline, CPB start, CPB end, and intensive care, and platelet counts, ROTEM, brinogen, and antithrombin were measured. Results:


Introduction
Whole blood cells and other components are diluted in cardiac surgery with cardiopulmonary bypass (CPB). A dilution hemoglobin (Hb) ratio can be calculated based on the amount of extracorporeal circulation lling and circulating volume [1]. Therefore, Hb level can be easily estimated after initiating CPB based on the dilution ratios of priming volume; however, decreased brinogen (Fib) and platelet levels before and after CPB are di cult to predict because of consideration of adsorption and consumption [2] in the extracorporeal membrane.
Certain studies suggested that point-of-care (POC) tests, such as rotation thromboelastometry (ROTEM™) [3][4][5][6], and a hemostasis management system (Hepcon HMS Plus™) [7] can re ect real coagulopathy and reduce the requirements for hemostatic products along with shortening the hospital stay compared to conventional laboratory coagulation testing. Thus, we not only measured normal coagulation tests, including brinogen concentration, activated partial thromboplastin time, prothrombin time, antithrombin, and complete blood counts (CBC), but also brinogen polymerization and platelet counts represented by ROTEM™. We rst aimed to analyze these data in proportion to the CPB time. The second goal of the study was to compare the allogeneic blood transfusion rates, bleeding volumes, and perioperative outcomes during surgery and within the rst 48 h after intensive care unit admission monitored using conventional laboratory testing and two POC tests: ROTEM™ and Hepcon HMS Plus™.
We hypothesized whether brinogen, platelet count, antithrombin, and parameters measured by ROTEM™ would decrease in proportion to CPB time. We also compared amounts of transfusion and the perioperative bleeding loss by conducting a case-control study.

Trial Design
This observational, single-center study was conducted at Tokyo Women's Medical University, Medical Centre East, in accordance with the Declaration of Helsinki; was approved by the Institutional Review Board of the same institution (approval #150207 on 10/9/2015); and was registered with the University Hospital Medical Information Network Centre (ID: UMIN000017412 on 5/5/2015). All patients provided written informed consent.

Study Population
A total of 160 patients who underwent cardiac surgery with CPB in October 2015-December 2018 were selected in this study. Exclusion criteria included emergency surgery, history of liver dysfunction, aged < 20 years, and pregnancy. Preoperative antiplatelet therapy except aspirin was terminated at least 7 days preoperatively. Warfarin and direct oral anticoagulants were terminated 2-4 days preoperatively, depending on guidelines [8]. Instead, heparin bridging with weight-adapted unfractionated heparin was administered intravenously until 6 h preoperatively.
Patients were divided into four groups depending on the CPB time: <1, 1-2, 2-3, and > 3 h. Blood samples were obtained at four time points: baseline (before anesthesia induction), immediately after starting CPB, while weaning from CPB, and in the intensive care unit (ICU), and CBC, ROTEM™ [3][4][5][6] were obtained in all study groups. Hepcon HMS Plus™ [7] was used to calculate the appropriate unfractionated heparin dose if we set the patients' body weight and target activated clotting time (ACT; i.e., 450 s) according to control blood samples within the target normal range. Hemostatic therapy comprised packed red blood cells (PRBCs), fresh frozen plasma (FFP), and platelet concentrate (PC). In Japan, one unit of these transfusion products is derived from 200 ml blood. For ROTEM™ measurements, samples were collected at the same four time points. ROTEM™ algorithms were based on three tests: the EXTEM (re ecting extrinsic initiation), INTEM (re ecting intrinsic initiation), and FIBTEM (platelet-inhibited extrinsic activation, re ecting the contribution of brin polymerization to clot rmness). The clotting time (CT) in EXTEM and INTEM tests and the amplitude of clot rmness at 10 min (A10) were used to determine the transfusion [9,10], such as PRBCs, FFP, and PC.

The Main Algorism for Transfusion
The target Hb concentration was 9-10 mg dl − 1 , and the FFP dose was estimated according to previous studies [9,10] to achieve a FIBTEM A10 of > 8-11 mm. These studies suggested that approximately 4 ml kg − 1 of FFP increases the FIBTEM A10 by 1 mm. The PC dose was also calculated according to a study to achieve an EXTEM A10 of > 40 mm [9] and/ or platelet count > 5-100,000/ul. If either the CT of INTEM or/and ACT measured using Hepcon™ was prolonged after protamine administration, we administered an additional protamine calculated using Hepcon™ after CPB. We did not administer any other additional blood products, such as brinogen concentrate, cryoprecipitate, and prothrombin complex concentrate.

Page 5/11
A Cell-Saver 5™ (Hemonetics Japan, Tokyo, Japan) was used to salvage washed erythrocytes for transfusion. Surgical re-exploration was performed if the total chest drain-tube blood loss exceeded 200 ml h − 1 or hemopericardium occurred in the ICU as diagnosed with ultrasound and X-ray.
The number of transfused units (PRBCs, FFP, and PC) and blood loss intraoperatively and in the rst 48 h after ICU admission are the primary outcome variables.
Statistical analyses were performed using JMP ver.15J (SAS Institute Japan, Tokyo, Japan). Data were assessed using analysis of variance and Tukey-Kramer's HSD test between groups. Statistical signi cance was de ned as P < 0.05. Preliminary power analysis setting α = 0.05, β = 0.8, effect size; f = 0.25 and number of groups; 3 suggested n = 159 calculated by statistical software, G* power [11].

Results
One patient had < 1 h, 73 had 1-2 h, 63 had 2-3 h, and 23 had > 3 h CPB time. Therefore, as only one patient had < 1 h, he was included in the 1-2 h group and renamed as < 2 h CPB group resulting in 74 cases. Patient characteristics are presented in Table 1. There was no signi cance except that between the EuroSCORE II of the CPB groups < 2 h and > 3 h. Blood loss in the operating room (OR) and ICU was signi cantly higher in proportion to CPB time as shown in Fig. 1, and components of CPB > 3 h for transfusion, i.e., FFP and PC in OR and RBCs and FFP in ICU, were signi cantly higher than the other two groups (Table 2); however, no signi cant differences were observed in platelet counts and brinogen levels between groups at weaning from CPB (Figs. 2). Furthermore, antithrombin, A10 of EXTEM and FIBTEM were lower in the > 3 h group than those in the other two groups at weaning of CPB (Figs. 2 and 3). Two patients with CPB of 2-3 h and one patient with > 3 h required reoperation for bleeding.  PRBCs OR 4 (0-8) 6 (0-8) 6 (0-10)

Discussion
Blood loss in cardiovascular surgery in the OR and ICU was higher in proportion to the CPB time, particularly in the > 3 h group, i.e., 6.2 and 3.1 times, respectively, higher than in the < 2 h group. Differences in platelet counts and brinogen levels among groups were not signi cant, but ROTEM™ measurement use showed a signi cant reduction in EXTEM and FIBTEM A10 at CPB termination, which are said to re ect the platelet counts and brinogen polymerization [9].
Based on the study, we suggest that blood transfusion management using ROTEM™ might be necessary if CPB time is > 3 h, such as aortic surgery, to add EXTEM and FIBTEM parameters as transfusion algorithms. Therefore, if CPB time can be predicted to be > 3 h perioperatively, we would measure not only CBC and brinogen but also parameters by POC, i.e., ROTEM to determine whether the amount of preparation, such as FFP and PCs, would be enough or not preoperatively.
Haensig et al. [12] also reported that POC was effective if CPB time was > 115 min. However, Girdauskas et al. [13] reported that POC was more useful for the surgery with deep hypothermic circulatory arrest (DHCA), with an average CPB time of approximately 200 min. In brief, they suggested that longer CPB time was more valuable for the measurement of POC coagulation test, which reduced the incidence of massive transfusion. However, only few studies investigated the relationship between CPB time, bleeding volume, and platelet function using POC, such as ROTEM.
Based on this study, two possible causes of reduced EXTEM and FIBTEM A10 were identi ed as represented by ROTEM. The rst one was decreased antithrombin level. As shown in Fig. 2, when CPB exceeded 3 h, antithrombin was signi cantly lower than those in the other CPB time groups. The low antithrombin level might reduce the effectiveness of heparin, resulting in platelet activation and consequently reducing platelet function following the CPB termination [14]. The second reason was based on DHCA. As shown in Table 1, most surgery types with > 3 h CPB were aortic surgeries such as total and hemi-arch replacement that required DHCA which could be re ected not by ASA physical status but EuroSCORE II as well. Moreover, the body temperature in DHCA was approximately 25 °C-29 °C in our hospital, requiring longer time for cooling and rewarming. It is generally said that platelet function is markedly impaired by hypothermia because of inactivation of enzyme insides the platelet [15]. Its hemostatic inability is known to be reversible; therefore, su cient rewarming can also recover the platelet function. We regularly terminate CPB when the rectal temperature exceeded 36 °C in DHCA cases; however, recovery of rectal temperature did not always coincide with the recovery of platelet function, showing time discrepancy. Based on these perspectives, the minimum temperature of DHCA tended to be higher in recent years. Therefore, if DHCA cases were excluded from patients in the CPB of > 3 h group, the group that only included mainly re-do or multi-valve replacement surgeries, decreased platelet function might not occur that often. Several studies have already investigated the usefulness of POC tests to reduce perioperative bleeding, transfusions, and cost for cardiac surgery and trauma patients [16]. Moreover, a recent multicenter study investigated a higher number of patients, i.e., 7402 cardiac surgeries with routine POC algorithm, who have been proven to have reduced major bleeding and PRBCs transfusion and PC [17].
By contrast, based on the results of our study, routine protocol for applying the POC test did not seem to be cost effective [18]. Therefore, patients with longer CPB time, with higher risk factors i.e., higher EuroSCORE II, and who used antiplatelet medications and coagulants should be carefully selected for the measurement of POC test because of high cartridge costs. Cardiac surgeries with CPB time of < 3 h do not require routine POC test but require su cient usual laboratory CBC and coagulation tests including Fib.
Limitations of this study include the small number of patients as an observational study. The in uence of DHCA for the > 3 h group could not be excluded. We noted that patients undergoing DHCA were weaned from CPB at a lower temperature, and this might affect platelet counts and function represented by the ROTEM measurements [19]. If patients with DHCA were excluded, platelet dysfunction may have been lower than that of our data. Only elective cardiac cases were included in this case-control study, and those who underwent emergency aortic dissection were excluded. Longer periods of hospital courses, including any complications, i.e., length of arti cial ventilation, kidney injury, and neurological de cits, should also be considered.

Conclusion
Perioperative bleeding loss and amount of transfusion have been increased in proportion to CPB time mainly because of lower antithrombin levels and platelet counts and brinogen polymerization represented by ROTEM™ measurement values, such as EXTEM and FIBTEM A10.