Thromboelastography Is Associated With Surrogates for Bleeding After Pediatric Cardiac Operations

Perioperative Hyper-coagulation and Thrombosis: Cost Analysis After Congenital Heart Surgery

Thrombosis, a major adverse event of congenital heart surgery, has been associated with poor outcomes. We hypothesized that in CHD patients undergoing cardiac surgery, increased perioperative use of pro-coagulant products may be associated with postoperative thrombosis in the setting of hyperfibrinogenemia, leading to greater hospital and blood product costs. Single-center retrospective study. Data from Boston Children's Hospital's electronic health record database was used in this study. All patients undergoing congenital heart surgery between 2015 and 2018 with postoperative fibrinogen levels above 400 mg/dl were reviewed. Of 334 patients with high plasma fibrinogen levels, 28 (8.4%) developed postoperative thrombosis (median age: one year, 59% male). In our cohort, 25 (7%) demonstrated evidence of baseline hypercoagulability by one or more panel test results. Thrombosis was associated with greater hospital and blood product costs, longer ventilation times, and longer hospital and ICU length of stays. Preoperative hypercoagulable state (odds ratio: 2.58, 95% CI [1.07, 9.99], p = 0.002), postoperative red blood cell transfusion (odds ratio: 1.007, 95% CI [1.000, 1.015], p = 0.04), and single ventricle physiology (univariate odds ratio: 2.94, 95% CI [1.09, 7.89], p = 0.03) were predictors of postoperative thrombosis. Preoperative hypercoagulable state and intraoperative platelet transfusion were predictors of hospital cost. Thrombosis was associated with worse in-hospital outcomes and higher costs. Preoperative hypercoagulable state and postoperative red blood cell transfusion were significant predictors of thrombosis. Risk prediction models that can guide thrombosis prevention are needed to improve outcomes of patients undergoing congenital heart surgery.

Thromboelastography-guided Intraoperative Platelet Transfusion in Pediatric Heart Surgery

BACKGROUND

Postoperative bleeding is associated with significant resource use and is an important contributor to other major adverse events in pediatric patients undergoing complex cardiac surgical procedures. Thromboelastography (TEG; TEG 6S, Haemonetics) can guide perioperative blood product transfusions to reduce the risk of postoperative bleeding. This study validated the use of a previously developed TEG 6S maximum amplitude (TEG-MA)-based platelet transfusion calculator used during cardiac surgical procedures to minimize the risk of postoperative bleeding.

METHODS

In this single-center retrospective study of pediatric patients (aged ≤18 years) who underwent cardiac surgical procedures requiring cardiopulmonary bypass at Boston Children's Hospital (Boston, MA) (N = 1000), the volume of platelet transfusion administered at surgical team discretion was compared with the platelet calculator-recommended platelet transfusion volume by using linear regression analysis. Associations between the adequacy of perioperative platelet transfusion and postoperative bleeding or thrombotic complications within the first 24 hours postoperatively (bleeding) and until hospital discharge (thrombosis) were evaluated by logistic regression analysis.

RESULTS

Lower TEG-MA (≤45 mm) measurements after transfusion were associated with a higher risk for postoperative bleeding (odds ratio, 4.4; 95% CI, 2.6-7.4; P < .01 [significant P value <.05]). The platelet transfusion calculator-recommended platelet transfusion volume (on the basis of TEG-MA measured at the time of rewarming) demonstrated moderate correlation with the measured TEG-MA value after platelet transfusion (Pearson r = 0.7). Intraoperative volumes of platelet transfusion that failed to increase a postoperative TEG-MA of at least 45 mm significantly increased the risk for postoperative bleeding in the first 24 hours after surgical procedures (odds ratio, 3.2; 95% CI, 1.9-5.4; P < .01 [significant P value <.05]). The posttransfusion TEG-MA was not independently associated with thrombosis.

CONCLUSIONS

Customizing perioperative platelet transfusion therapy by using quantitative diagnostic tests can help reduce postoperative bleeding complications.

Thromboelastography in the Perioperative Period: A Literature Review

Assessing coagulation status is essential for prompt intervention to reduce morbidity and mortality related to bleeding and thrombotic complications during the perioperative period. Traditional coagulation tests such as platelet count, activated partial thromboplastin time (aPTT), prothrombin time (PT), international normalized ratio (INR), and activated clotting time (ACT) provide only static evaluation. These tests are not designed for assessment of dynamically changing coagulation conditions during the perioperative time. However, viscoelastic coagulation testing such as thromboelastography (TEG) produces a rapid numerical and graphical representation that helps to detect and direct targeted hemostatic therapy. Searching the literature through PubMed, Medline, Ovid, CINAHL, and ClinicalTrials.gov we retrieved 210 studies, which represent the use of TEG in the perioperative period. The included studies were categorized under various settings such as trauma, obstetrics, orthopedics, intensive care unit (ICU), cardiovascular, transplant, and miscellaneous scenarios. TEG showed promising results in trauma surgeries in predicting mortality, hypercoagulability, and bleeding even when it was compared to conventional methods. TEG was also useful in monitoring anticoagulant therapy in orthopedic and obstetric surgeries; however, its role in predicting thrombotic events, hypercoagulability, or complications was questionable. In ICU patients, it showed promising results, especially in the prediction or improvement of sepsis, coagulopathy, thrombotic events, ICU duration, hospital stay, and ventilator duration. TEG parameters effectively predicted hypercoagulation in transplant surgeries. Regarding cardiovascular surgeries, they were effective in the prediction of the need for blood products, coagulopathy, thrombotic events, and monitoring anticoagulation therapy. More randomized clinical trials comparing TEG parameters with standardized tools are needed to produce robust results to standardize its use in different perioperative settings.

Preoperative thromboelastography in the prediction of post-tonsillectomy hemorrhage by coblation tonsillectomy: a post-hoc analysis

BACKGROUND

Post-tonsillectomy hemorrhage (PTH) affects around 4% of patients after tonsillectomy. We hypothesized that preoperative thromboelastography (TEG) might identify patients at higher risk of PTH.

OBJECTIVE

Investigate whether evaluation of coagulation function by preoperative TEG might help to predict PTH after tonsillectomy by coblation tonsillectomy (TE).

DESIGN

Post-hoc analysis of randomized controlled study.

SETTING

Otolaryngology Department between January 2017 and August 2019.

PATIENTS AND METHODS

This post-hoc analysis included adults who underwent coblation TE for benign tonsillar disorders. Routine blood tests and TEG were performed preoperatively. The TEG parameters evaluated included coagulation reaction time (R) and maximum thrombus amplitude (MA).

MAIN OUTCOME MEASURES

The main outcome was PTH during the 4-week postoperative period.

SAMPLE SIZE AND CHARACTERISTICS

284 RESULTS: The 19 patients (6.7%) that experienced PTH had a higher prevalence of diabetes mellitus, lower use of intraoperative suturing, fewer patients with grade I and II tonsillar enlargement, a higher white blood cell count, lower platelet count, lower fibrinogen level, lower R value, and a lower MA value than patients without PTH (all P<.05). Multivariate logistic regression revealed that diabetes mellitus (P<.053), fibrinogen level ≤2.735 g/L (P<.027), R≤6.55 min (P<.011) and MA≤59.15 mm (P<.012) were independently associated with PTH. A regression model incorporating these four factors predicted PTH with a sensitivity of 73.7% and specificity of 83.8%.

CONCLUSION

Preoperative evaluation of diabetes mellitus history, fibrinogen level, and TEG parameters might help to identify patients at higher risk of PTH after coblation TE.

LIMITATIONS

Single-center study with a small sample size; possibly underpowered statistically. TEG measurements might not accurately reflect coagulation function, and a validation cohort was unavailable.

CONFLICT OF INTEREST

None. CHINESE CLINICAL TRIAL REGISTRY NUMBER OF STUDY USED IN THIS ANALYSIS: ChiCTR2000032171. http://www.chictr.org.cn/showprojen.aspx?proj=52553.

Recent advancements in pediatric cardiopulmonary bypass technology for better outcomes of pediatric cardiac surgery

Background

Pediatric cardiac surgery is in itself very enigmatic and individualized. Presently, there has been a slew of new developments aimed primarily toward pediatric cardiopulmonary bypass for safer, patient-centered pediatric cardiac surgery. Still, lot of technological challenges need to be resolved, and their safer application in pediatric and neonate patients requires further refinement.

Main body of the abstract

Considering various significant yet unresolved issues of pediatric cardiac bypass, an exhaustive literature search was done on various internet databases with standard keywords. There are various new recent improvements; as the first oxygenator explicitly designed for neonatal patients; pediatric oxygenators with low prime volumes and surface areas that allow flows up to 2 L/min; pediatric oxygenators with integrated arterial filters; and miniature ultrafiltration devices that allow for high rates of ultrafiltrate removal. These advancements can significantly reduce cardiopulmonary bypass circuit surface areas and prime volumes. These advancements could reduce or eliminate the requirement for homologous red blood cells during or after surgery with reduction or eliminate bypass-related hemodilution, and inflammation. Because of the immaturity of the neonatal hemostatic system, conventional coagulation tests alone are insufficient to guide neonatal hemostatic therapy. Myocardial preservation techniques, safe temperature with duration are still debatable and yet to be fully explored.

Short conclusion

This review is based on Standards for Quality Improvement Reporting Excellence guidelines to provide a framework for reporting new knowledge to find better management strategy for pediatric cardiac cases.

Thromboelastography for the Prevention of Perioperative Venous Thromboembolism in Orthopedics

We have reviewed a large number of relevant literature to determine the deficiencies of orthopedics in the diagnosis and prevention of venous thromboembolism（VTE）events during the perioperative period, and found that the TEG technology has been widely used after liver transplantation, which may make up for the deficiencies. This review expounds the detection principle and latest thromboelastography (TEG) development, and highlights the advantages of TEG over previous screening methods in diagnosing hypercoagulability. By analyzing the correlation and consistency between TEG and conventional coagulation test, reliable indexes for diagnosing hypercoagulability and important parameters for guiding perioperative anticoagulation treatment were summarized. Furthermore, our work contributes to further studies of TEG in orthopedics. Based on the research results, we believe that TEG may help orthopedists to identify and predict VTE events, use anticoagulants, eventually reduce the occurrence of VTE events.

Plasma and Platelet Transfusions Strategies in Neonates and Children Undergoing Cardiac Surgery With Cardiopulmonary Bypass or Neonates and Children Supported by Extracorporeal Membrane Oxygenation: From the Transfusion and Anemia EXpertise Initiative-Control/Avoidance of Bleeding

OBJECTIVES

To present the recommendations and consensus statements with supporting literature for plasma and platelet transfusions in critically ill neonates and children undergoing cardiac surgery with cardiopulmonary bypass or supported by extracorporeal membrane oxygenation from the Transfusion and Anemia EXpertise Initiative-Control/Avoidance of Bleeding.

DESIGN

Systematic review and consensus conference of international, multidisciplinary experts in platelet and plasma transfusion management of critically ill children.

SETTING

Not applicable.

PATIENTS

Critically ill neonates and children following cardiopulmonary bypass or supported by extracorporeal membrane oxygenation.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

A panel of nine experts developed evidence-based and, when evidence was insufficient, expert-based statements for plasma and platelet transfusions in critically ill neonates and children following cardiopulmonary bypass or supported by extracorporeal membrane oxygenation. These statements were reviewed and ratified by the 29 Transfusion and Anemia EXpertise Initiative-Control/Avoidance of Bleeding experts. A systematic review was conducted using MEDLINE, EMBASE, and Cochrane Library databases, from inception to December 2020. Consensus was obtained using the Research and Development/University of California, Los Angeles Appropriateness Method. Results were summarized using the Grading of Recommendations Assessment, Development, and Evaluation method. We developed one good practice statement, two recommendations, and three expert consensus statements.

CONCLUSIONS

Whereas viscoelastic testing and transfusion algorithms may be considered, in general, evidence informing indications for plasma and platelet transfusions in neonatal and pediatric patients undergoing cardiac surgery with cardiopulmonary bypass or those requiring extracorporeal membrane oxygenation support is lacking.

Coagulation profiles and viscoelastic testing in multisystem inflammatory syndrome in children

OBJECTIVE

To characterize viscoelastic testing profiles of children with multisystem inflammatory syndrome in children (MIS-C).

METHODS

This single-center retrospective review included 30 patients diagnosed with MIS-C from March 1 to September 1, 2020. Thromboelastography (TEG) with platelet mapping was performed in 19 (63%) patients and compared to age- and sex-matched controls prior to cardiac surgery. Relationships between TEG parameters and inflammatory markers were assessed using correlation.

RESULTS

Patients with MIS-C had abnormal TEG results compared to controls, including decreased kinetic (K) time (1.1 vs. 1.7 minutes, p < .01), increased alpha angle (75.0° vs. 65.7°, p < .01), increased maximum amplitude (70.8 vs. 58.3 mm, p < .01), and decreased lysis in 30 minutes (Ly30) (1.1% vs. 3.7%, p = .03); consistent with increased clot formation rate and strength, and reduced fibrinolysis. TEG maximum amplitude was moderately correlated with erythrocyte sedimentation rate (ESR) (r = 0.60, p = .02), initial platelet count (r = 0.67, p < .01), and peak platelet count (r = 0.51, p = .03). TEG alpha angle was moderately correlated with peak platelet count (r = 0.54, p = .02). Seventeen (57%) patients received aspirin (ASA) and anticoagulation, five (17%) received only ASA, and three (10%) received only anticoagulation. No patients had a symptomatic thrombotic event. Six (20%) patients had a bleeding event, none of which was major.

CONCLUSIONS

Patients with MIS-C had evidence of hypercoagulability on TEG. Increased ESR and platelets were associated with higher clot strength. Patients were prophylactically treated with ASA or anticoagulation with no symptomatic thrombosis or major bleeding. Further multicenter study is required to characterize the rate of thrombosis and optimal thromboprophylaxis algorithm in this patient population.

Viscoelastic testing in pediatric patients

A tailored transfusion algorithm based on viscoelastic testing in the perioperative period or in trauma patients is recommended by guidelines for bleeding management. Bleeding management strategies in neonates and children are mostly extrapolated from the adult experience, as published evidence in the youngest age group is scarce. This manuscript is intended to give a structured overview of what has been published on the use of viscoelastic testing to guide bleeding management in neonates and children. Several devices that use either the traditional viscoelastic method or resonance viscoelastography technology are on the market. Reference ranges for children have been evaluated in only some of them. As most of the hemostasis maturation processes can be observed during the first year of life, adult reference ranges for viscoelastic testing could be applied over the age of 1 year. The majority of the published trials in children are based on retrospective analyses describing the correlation between viscoelastic testing and standard laboratory testing or focusing on the prediction of bleeding. Clinically more relevant studies in pediatric patients undergoing cardiac surgery have demonstrated that the implementation of a transfusion algorithm based on viscoelastic testing has significantly reduced transfusion requirements and that this approach has enabled a rapid detection of coagulation disorders in the presence of excessive bleeding. Although further studies are urgently needed, experts have reviewed the use of a transfusion algorithm based on viscoelastic testing in children as a feasible approach, as it has been shown to improve bleeding management and rationalize blood product transfusion.

Thromboelastography During Rewarming for Management of Pediatric Cardiac Surgery Patients

BACKGROUND

Thromboelastography (TEG) predicts bleeding in pediatric patients undergoing cardiac surgery. We hypothesize that TEG parameters at rewarming correlate with post-protamine values and that rewarming TEG is associated with surrogate endpoints for postoperative bleeding in pediatric patients undergoing complex cardiac surgery.

METHODS

In a retrospective study of pediatric (≤18yrs) patients (N=703) undergoing complex cardiac surgery procedures, TEG obtained during rewarming and following protamine administration were compared using linear regression. A composite endpoint of extended blood product transfusion or surgical re-exploration for bleeding was utilized as a surrogate for post-operative bleeding.

RESULTS

By multivariable analysis, longer cardiopulmonary bypass time and lower TEG maximal amplitude (MA) during rewarming were independently associated with risk of composite endpoint in the operating room or intensive care unit (p<0.05). Among patients with MA<45mm during rewarming, those who received platelet transfusion compared to those who did not in the operating room were less likely to reach composite endpoint within the subsequent 24 hours (8%vs.32% respectively; p<0.01). Good correlation was observed between TEG parameters at rewarming vs. after protamine administration (Pearson r≥0.7). The relationship between platelet transfusion volume (ml/kg) and percent change in MA was determined using linear regression and a platelet transfusion calculator was generated.

CONCLUSIONS

Lower MA during rewarming is associated with increased risk of perioperative bleeding. In patients with rewarming MA<45mm, intraoperative platelet transfusion may reduce the risk of subsequent bleeding. Individualized platelet transfusion therapy based on rewarming TEG may reduce the risk of bleeding while minimizing unnecessary platelet transfusion.

Stratification of Bleeding Risk Using Thromboelastography in Children on Extracorporeal Membrane Oxygenation Support

OBJECTIVES

Patients undergoing extracorporeal membrane oxygenation are at high risk for bleeding and thrombotic complications. Current laboratory methods for assessing the coagulation system may be imprecise and complicate clinical decision-making. We hypothesize that thromboelastography may be more strongly associated with bleeding events than traditional methods and can aid extracorporeal membrane oxygenation coagulation management.

DESIGN

In a retrospective study, 40 patients with congenital heart disease requiring extracorporeal membrane oxygenation support yielded a total of 159 patient days of data for thromboelastography analysis.

SETTING

Pediatric cardiac ICU at a single institution.

SUBJECTS

Pediatric patients (≤ 18 yr) with congenital heart disease requiring extracorporeal membrane oxygenation support.

INTERVENTIONS

None.

METHODS

Thromboelastography was performed on whole blood samples collected 6-12 hours following extracorporeal membrane oxygenation initiation and daily for the duration of extracorporeal membrane oxygenation. Bleeding during each 24-hour period was defined as need for re-exploration or need for blood transfusion. Associations between thromboelastography variables and bleeding over each 24-hour period (bleeding vs nonbleeding days) were assessed using mixed effects logistic regression and classification and regression tree analysis.

MEASUREMENTS AND MAIN RESULTS

Bleeding occurred in 25 patients (63%), contributing 87 bleeding days (55% extracorporeal membrane oxygenation days) for analysis. The probability of bleeding within the 24-hour period was not associated with activated partial thromboplastin time (p = 0.6) or anti-Xa levels (p = 0.3) on that day. The strongest correlate of bleeding was a maximum amplitude less than 55.4 mm on thromboelastography (odds ratio, 3.28; 95% CI, 1.63-6.60; p < 0.001). Bleeding occurred on 73% versus 35% of extracorporeal membrane oxygenation days for maximum amplitude less than 55.4 mm versus greater than or equal to 55.4 mm, respectively. Bleeding occurred on all days when a combination of maximum amplitude less than 55.4 mm and a reaction time greater than 12.9 minutes was present. The lowest risk of bleeding (28% of patient days) was associated with maximum amplitude greater than or equal to 55.4 mm and plasma fibrinogen greater than 345 mg/dL.

CONCLUSIONS

Thromboelastography-derived variables maximum amplitude and reaction time, along with plasma fibrinogen levels, can help predict bleeding events in children on extracorporeal membrane oxygenation support. Research based on larger patient samples is needed to confirm the specific thresholds identified for bleeding risk stratification for extracorporeal membrane oxygenation anticoagulation management.

Postoperative hemorrhage following coblation tonsillectomy with and without suture: A randomized study in Chinese adults

PURPOSE

Coblation tonsillectomy (TE) increases gradually in China. Hemorrhage is the main complication after tonsillectomy. The conclusions of the studies about suture after tonsillectomy are conflicting. To compare the post-tonsillectomy hemorrhage (PTH) rates in patients who underwent coblation with vs. without suture.

MATERIALS AND METHODS

This was a randomized controlled study of adult patients who underwent coblation TE at our hospital between 01/2017 and 08/2019. The patients were randomized to TE with or without suture. The primary endpoint was the secondary PTH. The secondary endpoints included the primary PTH, grade of PTH, and incidence of PTH within 4 weeks post-TE.

RESULTS

There were no differences between the two groups regarding sex, age, disease course, and BMI (all P > 0.05). The occurrence of secondary PTH was lower in the suture group compared with the non-suture group (2.8% vs. 7.7%, P = 0.016). Compared with non-suture group, the incidence of PTH within 4 weeks (2.8% vs. 10.6%, P = 0.009) and the PTH degree (P = 0.02) were all significantly lower in the suture group.

CONCLUSION

Intraoperative suture reduces the secondary PTH in adult patients who underwent coblation tonsillectomy. The incidence of PTH within 4 weeks, PTH degree and pain might all improved for intraoperative suture.

Platelet Inhibition With IV Glycoprotein IIb/IIIa Inhibitor to Prevent Thrombosis in Pediatric Patients Undergoing Aortopulmonary Shunting

OBJECTIVES

Shunt thrombosis, a potential complication of aortopulmonary shunting, is associated with high mortality. Commonly used oral antiplatelet drugs such as aspirin demonstrate variable absorption and inconsistent antiplatelet effect in critically ill patients early after surgery. IV glycoprotein IIb/IIIa inhibitors are antiplatelet agents with rapid and reproducible effect that may be beneficial as a bridge to oral therapy.

DESIGN

Retrospective review of pediatric patients undergoing treatment with IV tirofiban. Discarded blood samples were used to determine pharmacokinetic parameters.

SETTING

Pediatric cardiac ICU at a single institution.

PATIENTS

Fifty-two pediatric patients (< 18 yr) undergoing surgical aortopulmonary shunt procedure who received tirofiban infusion as a bridge to oral aspirin.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Primary outcome measures were shunt thrombosis and bleeding events, whereas secondary outcomes included measurement of platelet inhibition by thromboelastography with platelet mapping and pharmacokinetic analysis (performed in a subset of 15 patients). Shunt thrombosis occurred in two of 52 patients (3.9%) after prophylaxis treatment with tirofiban; both thrombosis events occurred after discontinuation of the drug. One patient (1.9%) experienced bleeding complication during the infusion. A tirofiban bolus of 10 µg/kg and infusion of 0.15 µg/kg/min resulted in significantly increased platelet inhibition via adenosine diphosphate pathway (median 66% [43-96] pre-tirofiban compared with 97% [92-99%] at 2 hr; p < 0.05). Half-life of tirofiban in plasma was 142 ± 1.5 minutes, and the average steady-state concentration was 112 ± 62 ng/mL. Age and serum creatinine were significant covariates associated with systemic clearance. Dosing simulations were generated based upon one compartment model.

CONCLUSIONS

IV glycoprotein IIb/IIIa inhibitor as a bridge to oral antiplatelet therapy is safe in pediatric patients after aortopulmonary shunting. Dosing considerations should include both age and renal function. Randomized trials are warranted to establish efficacy compared with current anticoagulation practices.

Hemostatic Testing in Critically Ill Infants and Children

Children with critical illness frequently manifest imbalances in hemostasis with risk of consequent bleeding or pathologic thrombosis. Traditionally, plasma-based tests measuring clot formation by time to fibrin clot generation have been the "gold standard" in hemostasis testing. However, these tests are not sensitive to abnormalities in fibrinolysis or in conditions of enhanced clot formation that may lead to thrombosis. Additionally, they do not measure the critical roles played by platelets and endothelial cells. An added factor in the evaluation of these plasma-based tests is that in infants and young children plasma levels of many procoagulant and anticoagulant proteins are lower than in older children and adults resulting in prolonged clot generation times in spite of maintaining a normal hemostatic "balance." Consequently, newer assays directly measuring thrombin generation in plasma and others assessing the stages hemostasis including clot initiation, propagation, and fibrinolysis in whole blood by viscoelastic methods are now available and may allow for a global measurement of the hemostatic system. In this manuscript, we will review the processes by which clots are formed and by which hemostasis is regulated, and the rationale and limitations for the more commonly utilized tests. We will also discuss selected newer tests available for the assessment of hemostasis, their "pros" and "cons," and how they compare to the traditional tests of coagulation in the assessment and management of critically ill children.

Magnetic Resonance-Based Diagnostics for Bleeding Assessment in Neonatal Cardiac Surgery

PURPOSE

Infants undergoing a cardiac operation are at high risk for postsurgical bleeding. To date, there are no highly predictive models for postsurgical bleeding in this population. This study's objective was to assess the predictive ability of T2 magnetic resonance (T2MR).

DESCRIPTION

T2MR uses magnetic resonance to detect clot formation characteristics on a small blood sample and provides hemostatic indicators that can assess bleeding risk.

EVALUATION

This prospective, single-institution study enrolled 100 patients younger than 12 months old undergoing a cardiac operation from April 27, 2015, to September 21, 2016. The primary end point was postsurgical bleeding within 24 hours after the procedure. T2MR data were modeled with a binary recursive partitioning algorithm with randomized cross-validation. The tight clot metric produced the highest univariate discrimination of bleeding (receiver operator characteristic curve, 0.64; classification accuracy, 72%), and along with the platelet function metric, demonstrated highest relative importance based on Gini index splitting (Salford Systems, San Diego, CA). Multivariate modeling with cross-validation showed mean receiver operator characteristic curve area of 0.74 and classification accuracy of 82%.

CONCLUSIONS

T2MR tight clot and platelet function metrics were associated with bleeding events.