Allogeneic Blood Transfusion and Postoperative Outcome Before and After Implementing a SONOCLOT Based Coagulation Management Algorithm in Cardiac Surgery: A Quality Control Study
Allogeneic Blood Transfusion and Postoperative Outcome Before and After Implementing a SONOCLOT Based Coagulation Management Algorithm in Cardiac Surgery: A Quality Control Study
Implementing coagulation management algorithms based on Point-of-care (POC) testing devices such as thromboelastography (TEG) or rotational thromboelastometry (ROTEM) have been shown to reduce transfusion rates and the occurrence of adverse patient outcomes. Apart from the TEG and ROTEM, another viscoelastic POC measurement technique, the SONOCLOT, can be easily used at the bedsite. However Information regarding a SONOCLOT based coagulation management algorithm however is lacking.
Transfusion rates of allogeneic blood products (Red blood cells = RBC, Fresh frozen plasma = FFP, Platelet concentrates = PLT) and patient adverse outcomes in two cohorts before and after the implementation of a SONOCLOT based coagulation management algorithm were evaluated on an observational basis in 1754 patients undergoing different types of cardiac surgery. The effect of the cohort and 19 predictors on transfusion was assessed by relative R2-values obtained by bootstrapping and a hurdle regression model, comprising a binomial and a Poisson count component. Association of predictors with post-operative outcomes was assessed with a logistic regression model.
Patient data and all relevant procedure related data of cardiac surgery procedures at the Triemli City Hospital, Switzerland are recorded in a dedicated database with the approval of the institutional research ethics board and patient informed consent. Data analysis focusing on coagulation management using a dedicated algorithm in the present quality control study is done for 1754 out of 2199 consecutive patients undergoing cardiac surgery comprising the initial cohort from 2009 to 2011 (before implementation of the algorithm) and the subsequent cohort from 2013 to 2015 (after implementation of the algorithm).
In the time period analyzed for this study anesthesia and intensive care were performed according to standardized protocols. Fluid management was done with lactated Ringer's solution and synthetic colloids. Different blood conservation strategies were used and included consistently cell salvage, standardized application of tranexamic acid and retrograde autologous cardiopulmonary bypass (CPB) priming. Anticoagulant medication was suspended at least 5 days prior to surgery for all elective procedures. Hematocrit measurements were performed using a standard POC blood gas analyzer available in the OR. Indication for red blood cells (RBC) transfusion were a hematocrit (HCT) level below 21% during CPB, below 25% for patients with preserved left ventricular function and 28% for patients with a left ventricular function ≤ 30% and patients undergoing emergency procedures. Standard laboratory tests (blood count, international normalized ratio (INR) / prothrombin time (PT), activated partial thromboplastin time (aPTT) fibrinogen concentration (FIB) were routinely available for guiding transfusion of coagulation blood products. Coagulation management before the implementation of the algorithm was performed based on institutional guidelines in accordance with published recommendations. Kaolin activated clotting time (ACT) measurements were used to guide heparin and protamine management. After heparin reversal and ongoing bleeding prothrombin complex concentrate was applied at a PT ≤ 70% / INR threshold of 1.5 and in selected cases with ongoing major bleeding fresh frozen plasma (FFP) was used to normalize PT/INR. Moreover, fibrinogen concentrate was given below a fibrinogen level of 1 g/l. Platelet concentrates (PLT) were transfused when platelet count was ≤ 100 x 109/lit. PLT were also given in suspected platelet dysfunction. When severe bleeding occurred at the end of the surgical procedure coagulation management was started on an empirical basis by the anesthesiologist in charge.
The SONOCLOT coagulation management algorithm is based on existing literature on this subject and the experience of the authors with the specific POC technique: It considers standard coagulation lab results and platelet function tests and includes the use of the MULTIPLATE analyzer. SONOCLOT measurements consist basically of three dedicated tests:
First coagulation measurements in each patient undergoing cardiac surgery are executed in the pre-intervention period. The established baseline values trigger the preparation of coagulation blood products that will be applied in the post-interventional period. At the end of the surgical procedure, heparin reversal is performed according to the Bulls protocol. After initial antagonisation with protamine kaolin ACT and glass-bead ACT as well as a heparinase glass-bead ACT are done to primarily detect any residual heparin. For any registered serum heparin the treatment and measurement loop is repeated. In case of an ongoing bleeding after heparin reversal a surgery-related cause has to be excluded and then another treatment and management loop based on glass-bead ACT is initiated and repeated until coagulation has been restored. As soon as results of the standard lab test are available in the postoperative period these are included in the decision making process.
Primary outcome parameters were transfusion rates during the operation and the treatment in the ICU. Secondary outcome parameters included morbidity: an mortality in the postoperative period during the hospital stay.
Statistical analysis will be performed with the software package R. To compare patient-related demographic and perioperative data prior and post algorithm implementation and univariable analysis are performed. Based on the data format and distribution, either Fisher's exact test, t-tests or Wilcoxon rank sum test were applied. For the assessment of the difference of transfusion in the two cohorts and controlling potential influence of other factors two approaches have been chosen. First, after a continuity-corrected log-transformation of the units of RBC given, a sequence of preoperative and procedure related predictors were included in the analysis. To check for multi-colinearity, variance inflation factors are assessed with the R package. Furthermore, to assess the association between potential risk factors and the incidence of postoperative infections a generalized linear model with a binomial link and logit function was chosen and performed with the R package.
Inclusion
Exclusion