- Email: [email protected]
Thromboelastography Is Associated With Surrogates for Bleeding After Pediatric Cardiac Operations
Accepted Manuscript Thromboelastography is Associated with Surrogates for Bleeding after Pediatric Cardiac Surgery Sirisha Emani, PhD, Lynn A. Sleeper, ScD, David Faraoni, MD, PhD, Michelle Mulone, BS, Fatoumata Diallo, BS, James A. DiNardo, MD, Juan Ibla, MD, Sitaram M. Emani, MD PII:
S0003-4975(18)30649-0
DOI:
10.1016/j.athoracsur.2018.04.023
Reference:
ATS 31586
To appear in:
The Annals of Thoracic Surgery
Received Date: 29 October 2017 Revised Date:
6 April 2018
Accepted Date: 11 April 2018
Please cite this article as: Emani S, Sleeper LA, Faraoni D, Mulone M, Diallo F, DiNardo JA, Ibla J, Emani SM, Thromboelastography is Associated with Surrogates for Bleeding after Pediatric Cardiac Surgery, The Annals of Thoracic Surgery (2018), doi: 10.1016/j.athoracsur.2018.04.023. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Thromboelastography is Associated with Surrogates for Bleeding after Pediatric Cardiac Surgery
RI PT
Running Head: TEG is associated with bleeding
1
Sirisha Emani, PhD, 1Lynn A. Sleeper, ScD, 3David Faraoni, MD, PhD, 1Michelle Mulone, BS, Fatoumata Diallo, BS, 2James A. DiNardo, MD, 2Juan Ibla, MD, 1Sitaram M. Emani, MD
SC
1
M AN U
1
Cardiovascular Program, Boston Children’s Hospital, Massachusetts, USA
2
Department of Anesthesia and Pain Medicine, Boston Children’s Hospital, Massachusetts, USA
3
Correspondence:
EP
Word count: 4498
TE D
Department of Anesthesia and Pain Medicine, Hospital for Sick Children, Toronto, Canada
AC C
Sitaram Emani, MD
300 Longwood Ave. Boston, MA 02115
Email: [email protected]
1
ACCEPTED MANUSCRIPT
Abstract Background: Perioperative bleeding is a common complication in pediatric patients undergoing cardiac surgery. Although thromboelastography (TEG) has been utilized in patients undergoing
RI PT
adult cardiac surgery, limited data are available in pediatric patients. We hypothesize that TEG parameters may be associated with surrogate endpoints for postoperative bleeding in pediatric patients undergoing complex cardiac surgery.
SC
Methods: In a retrospective study, TEG was obtained after protamine administration and upon intensive care unit (ICU) admission in pediatric patients (≤ 18yrs) undergoing cardiac surgery
M AN U
requiring cardiopulmonary bypass. A composite endpoint of extended blood product transfusion or surgical re-exploration for bleeding was utilized as a surrogate for perioperative bleeding. TEG parameters were compared between patients who did or did not reach the composite endpoint.
TE D
Results: The study included 511 pediatric patients undergoing complex cardiac surgery. Composite endpoint was reached in 52% of patients with MA< 45 mm compared to 31% with MA ≥ 45 mm (P<0.001). Using multivariable regression analysis, MA < 45 mm was
EP
independently associated with composite endpoint (P<0.001). Patients with MA < 45 mm who received platelet transfusion in the OR were less likely to reach composite endpoint within the
AC C
subsequent 24 hours (8%) compared to those who did not receive intraoperative platelet transfusion (24%) (P=0.02).
Conclusions: Intraoperative TEG MA < 45 mm is associated with a surrogate endpoint for intraoperative bleeding in pediatric patients undergoing complex cardiac surgery. In patients with MA < 45 mm, prophylactic platelet transfusion in the OR may be associated with reduction in bleeding endpoints in the ICU.
2
ACCEPTED MANUSCRIPT
Coagulopathy after cardiopulmonary bypass (CPB) is known to be complex and multifactorial, and has been shown to be associated with increased transfusion requirements and hemodynamic instability, increasing both morbidity and mortality (1). The pediatric population is
RI PT
at increased risk of coagulopathy, perioperative bleeding, and transfusion requirements compared to adults (2). Standard coagulation assays have been used for decades to guide anticoagulation in adults and children (3, 4). These tests were not designed to monitor
SC
perioperative coagulopathy or guide the administration of hemostatic agents in bleeding patients (5).
M AN U
Recent experience in patients undergoing cardiac surgery has suggested a potential benefit of thromboelastography (TEG) for risk stratification and blood product management following adult and pediatric cardiac surgery (6, 7). The use of viscoelastic tests is now recommended in both American and European guidelines regarding perioperative bleeding management (8, 9). To date, the predictive accuracy of TEG has only been studied in a small number of retrospective
TE D
studies that included infants and children undergoing both low and high complexity congenital cardiac surgical procedures, leading to heterogeneous and conflicting results (10-12). The objective of this study was to determine whether abnormal TEG parameters are associated
EP
with surrogate endpoints of early postoperative bleeding in a group of pediatric patients undergoing complex cardiac surgical procedures and considered at high risk for perioperative
AC C
bleeding.
Patients and Methods Study Design
3
ACCEPTED MANUSCRIPT
Consecutive pediatric patients (age ≤ 18yrs) undergoing cardiac surgery at Boston Children's Hospital (BCH) between April 2015 and March 2017 were retrospectively studied after approval by Institutional Review Board. The target population was patients at high-risk for bleeding
RI PT
following a cardiac surgical procedure requiring cardiopulmonary bypass (CPB). A patient was eligible if any one of the following criteria were met: 1. Neonatal age at surgery (age ≤ 30 days), 2. Single ventricle physiology, 3. Re-operative sternotomy with greater than
SC
two previous operations, 4. Complex cardiac reconstruction with prolonged duration of CPB. Patients were excluded if testing for TEG was not performed at all time points or if patients were
M AN U
transfused prior to testing for TEG in the OR.
For this cohort of patients, CPB circuit was primed with reconstituted whole blood (1 unit of packed red blood cells (RBC) and 1 unit of fresh frozen plasma (FFP)) to obtain target hematocrit of >30% in all patients. RBC obtained by cell salvage of CPB circuit blood was routinely administered in all patients. Anticoagulation with heparin was performed to achieve
TE D
ACT goal of >480, and at least moderate hypothermia (< 32oC) was utilized. Blood gas management included pH stat strategy for patients cooled below 28oC.
EP
Data collection
Patient characteristics data were collected from the medical record. Transfusions of RBC,
AC C
platelets, cryoprecipitate, FFP were recorded along with administration of activated recombinant Factor VII (rFVIIa) and albumin. Procedural records were reviewed to determine any episodes of surgical re-exploration for bleeding following temporary or permanent chest closure. Study Endpoints
A composite endpoint was established as a surrogate for perioperative bleeding that included transfusion requirements or need for surgical re-exploration for bleeding. In our institution, transfusion of RBC and platelets are considered as first line therapy following administration of 4
ACCEPTED MANUSCRIPT
protamine in patients with subjective bleeding. Although RBC and platelet transfusion may also be administered in the absence of bleeding (for anemia or thrombocytopenia), treatment with cryoprecipitate, FFP, or rFVIIa is limited by institutional protocol to patients with subjective
RI PT
bleeding refractory to first line therapy. Thus, extended transfusion was defined as the administration of one or more of these blood products in addition to RBC and platelets. The composite endpoint included either extended transfusion or surgical re-exploration for bleeding.
SC
The composite endpoint was further classified as occurring in the OR or within 24 hours of ICU admission depending upon location of the patient at the time of event. Remote bleeding
Thromboelastography (TEG)
M AN U
episodes beyond the first 24 hours were not included in analysis.
TEG was performed in a CLIA certified laboratory after protamine administration in the OR and upon ICU admission. Activation with Kaolin was achieved in both standard and heparinase
TE D
cups. TEG parameters obtained included Reaction time: time from initiation of clot formation to the clot strength of 2 mm (R-time), kinetic clot time: time to achieve a clot strength of 20 mm (Ktime), rate of clot formation (Angle), maximum amplitude (MA), overall clot strength (G parameter- a calculated value which is the log derivation of MA), and percentage of clot lysis 30
EP
min after the MA is reached (Ly30). TEG values were not utilized to guide transfusion
AC C
management in this cohort.
In preliminary analysis, each TEG parameter was tested based on the risk characterization as high risk (MA< 45 mm), intermediate risk (MA 45-50 mm), and low risk (MA > 50 mm) groups obtained by Cart analysis. As there was no significant difference between intermediate and low risk groups in relation to bleeding outcomes, we reanalyzed using only two groups (MA<45 and >45 mm). Statistical Methods 5
ACCEPTED MANUSCRIPT
Categorical data were described as number with frequency and continuous data as median with interquartile range (IQR) or mean with standard deviation. Association between TEG parameters and composite endpoint was assessed by logistic regression. Association between
RI PT
TEG parameters in OR and composite endpoint in ICU was determined by using Fisher exact test. Generalized additive modeling (GAM) was used to estimate the shape of relationship
between TEG parameters and the probability of reaching composite endpoint. If a significant
SC
non-linear relationship was found, the continuous predictor was categorized, or modeled using a piecewise linear fit to assess association with probability of composite endpoint. Multivariable
M AN U
logistic stepwise regression was conducted to identify independent predictors of reaching composite endpoint using candidate variables found to be significant at the 0.05 level in univariate analysis. Statistical analysis was performed using SAS version 9.4 (SAS Institute,
Results
TE D
Cary, NC) and R version 3.2.1. P-values less than 0.05 were considered statistically significant.
Patient Characteristics, post-bypass TEG parameters and blood products administration
EP
Five hundred and eleven pediatric patients who underwent high-risk cardiac surgical procedures requiring CPB and had TEG testing were retrospectively studied. Patient characteristics
AC C
including age at surgery, body weight, CPB, cross-clamp time (CCT), and procedure type are shown in Table 1. The mean age at surgery was 30 months and the median (IQR) was 11 months (2, 36); 97 (19%) were categorized as neonates (≤ 30 days of age at surgery). Nonautologous blood product transfusion excluding pump priming products (RBC and FFP) was necessary in 360/511 (71%) of patients in the cohort, including RBC (non-salvaged) in 125/511 (25%), platelets in 323/511 (63%), cryoprecipitate in 181/511 (35%), and FFP in 18/511 (4%) of patients. rFVIIa was administered in 28/511 (6%) of patients. The composite endpoint was 6
ACCEPTED MANUSCRIPT
reached in the OR and within 24 hours of ICU admission in 181/511 (35%) and 47/511 (9%) of patients, respectively (Table 1). Re-exploration for bleeding occurred in 6 of 181 patients either in the OR following initial chest closure (2 patients) or ICU (6 patients), with 2 patients requiring
Association of age at surgery with composite endpoint and TEG
RI PT
re-exploration in both locations.
A higher percentage of neonates compared to non-neonates reached composite endpoint in the
SC
OR (70% vs. 27%, P<0.001) but not in the ICU (11% vs. 9%, P= 0.44). Mean post-bypass MA (47 ± 9 mm vs. 49 ± 8 mm, P<0.05) and clot strength G (4.8 ± 1.6 vs. 5.1 ± 1.4 dynes/sec,
M AN U
P<0.05) were significantly lower in neonates compared to non-neonates. However, age at surgery was not correlated with TEG parameters. The absolute Pearson correlation coefficient r for reaction time is 0.05, K- time is 0.04, rate of clot formation is 0.02, maximum amplitude 0.05, clot strength is 0.05 and clot lysis is 0.07 (data not shown).
TE D
TEG parameters and association with composite endpoint in OR
Median values of each TEG parameter are displayed in Table 2 and were compared between patients who did or did not reach the composite endpoint in the OR. Maximal Amplitude, K-
EP
time, and G parameter, but not R-time, Angle, or Ly30 were significantly different between patients who did or did not reach composite endpoint. Figure 1 displays the estimated
AC C
probability of reaching composite endpoint in the OR as a function of the continuous postprotamine TEG parameters. Longer K-time, lower MA and G parameter were significantly associated with probability of composite endpoint. There was a non-linear relationship between clot strength G parameter and probability of composite endpoint. Weaker clot strength ≤ 5.8 dynes/sec was linearly associated with the probability of composite endpoint, but no association was observed at levels > 5.8 dynes/sec. R-time, Angle and Ly30 were not associated with probability of composite endpoint. 7
ACCEPTED MANUSCRIPT
The multivariable regression model for probability of composite endpoint in the OR (Table 3) included only parameters that were significantly associated with composite endpoint in the univariate analysis - neonate vs. non-neonate, continuous age at surgery, and the post-
RI PT
protamine MA. Neonatal age (odds ratio 6.07), and lower MA (odds ratio 0.95 per mm) were independently associated with higher odds of composite endpoint in the OR. The area under the curve (AUC) of 0.71 suggests moderate predictive power of this model to predict composite
SC
endpoint (Table 3). Volume of total blood products (excluding pump prime products)
administered in the OR post-protamine (normalized to body weight) was significantly higher in
M AN U
patients with TEG MA < 45 compared to those with MA > 45 (median 32.3 ml/kg vs. 22.2 ml/kg respectively, P=0.002).
Association between platelet transfusion in OR and composite endpoint in the ICU A subgroup analysis was performed to assess whether there is a differential association
TE D
between platelet transfusion in the OR and reaching composite endpoint within 24 hours of ICU admission depending on MA level measured in the OR. Among patients with post-protamine MA < 45 mm, the percentage reaching composite endpoint in the ICU was lower for the 93 patients who received platelet transfusion in the OR compared to the 33 patients who did not
EP
receive platelet transfusion in the OR (8% vs. 24%, P=0.02). Among patients with normal MA (≥ 45 mm) in the OR, there was no significant difference in the rate of composite endpoint in ICU
AC C
between patients who received or did not receive platelet transfusions in the OR (8% vs. 9% respectively, P=0.71). (Table 4). By logistic regression modeling, intraoperative platelet transfusion in patients with post-protamine MA < 45 mm was associated with significantly lower odds of composite endpoint in the ICU (OR=0.25, P=0.02). In patients with MA ≥ 45 mm, there was no association between intraoperative platelet transfusion and composite endpoint in the ICU (OR=0.86, P=0.67) (Table 5). However, there was insufficient evidence statistically to state
8
ACCEPTED MANUSCRIPT
that the effect of platelets on composite endpoint differs by MA group, with a logistic regression interaction p-value of 0.07.
RI PT
Association between Platelet transfusion and TEG parameters TEG parameters obtained in the ICU were compared to parameters measured in the OR prior to blood product transfusion. For patients with MA < 45 mm in the OR who did not receive platelet transfusion, there was no significant difference in the median (IQR) MA values between the OR
SC
and ICU (41.7 (39.7, 43.4) mm vs. 40.6 (35.8, 46.1) mm, P=0.6). On the other hand, patients with MA < 45 mm in OR who received platelet transfusion (median of 15 cc/kg of apheresis
M AN U
platelets) demonstrated significant increase in median (IQR) MA from 40.8 (37.5, 43) mm in the OR to 63.4 (56.2, 69.7) mm in the ICU (P<0.001) (Figure 2). Patients who received both cryoprecipitate and platelet transfusion in the OR had significant increase in median (IQR) Angle and G parameter between OR and ICU (58 (51, 63.2) degrees and 4.5 (3.6, 5.6) dynes/sec in
TE D
OR vs. 69 (63.8, 73.5) degrees and 9.9 (7.8, 12.6) dynes/sec in the ICU, P<0.001 and P<0.001, respectively). The median (IQR) overall clot strength G parameter measured in the ICU was significantly higher in patients receiving both cryoprecipitate and platelet transfusion compared to patients receiving platelet transfusion alone (9.9 (7.8, 12.6) vs. 7.5 (6.3, 8.8) dynes/sec,
EP
P<0.001). Patients who received platelet transfusion in the OR also demonstrated significant increase in median (IQR) Angle and G parameter between OR and ICU 59.3 (54.5, 63.9)
AC C
degrees and 5.2 (4.3, 6.1) dynes/sec in OR vs. 64.5 (60.6, 69) degrees and 7.5 (6.3, 8.8) dynes/sec in the ICU, P<0.001 and P<0.001, respectively) (Figure 3).
Comment This study of pediatric patients undergoing high-risk cardiac surgery demonstrates an association between TEG parameters obtained immediately following discontinuation of CPB 9
ACCEPTED MANUSCRIPT
and composite endpoint of extended transfusion or re-exploration for bleeding in the OR. Neonatal age (≤ 30 days) at the time of surgery and lower MA were independently associated with increased risk of reaching composite endpoint. Platelet transfusion was associated with
RI PT
significant increase in the MA between OR and ICU. Patients with MA < 45 mm in the OR who did not receive platelet transfusion demonstrated higher rate of composite endpoint in ICU compared to those who did receive platelet transfusion.
SC
Previous studies have not demonstrated a consistent correlation between post-protamine TEG and postoperative bleeding in cohorts of pediatric patients undergoing cardiac surgery (13, 14).
M AN U
In contrast to these previous studies, the current study investigated only patients with a wellaccepted high-risk for bleeding following cardiac surgery. Our cohort of patients was specifically selected since this is the population among whom coagulation testing may be beneficial to mitigate the risk of postoperative bleeding. In this population of high-risk patients, significant association between TEG parameters and the composite endpoint as a surrogate for bleeding
TE D
was observed.
The data in this study is consistent with previous reports that suggest that deficiencies in platelets and fibrinogen, rather than coagulation factors, are associated with postoperative
EP
bleeding following cardiac surgery (15-17). Maximum amplitude measures clot strength, a composite effect of platelet and fibrinogen activity. In this study, the MA frequently normalized
AC C
with platelet transfusion alone, and suggests that cryoprecipitate may be unnecessary in the management of post-bypass coagulopathy in this patient population. Elevated Angle and clot strength G parameter in patients receiving cryoprecipitate and platelet transfusion compared to platelet transfusion alone may increase risk of thrombosis. Previous studies using ROTEM suggested that cryoprecipitate transfusion may result in increased thrombotic risk (18). The TEG functional fibrinogen assay may be useful in determining need for cryoprecipitate transfusion in this setting. 10
ACCEPTED MANUSCRIPT
This study utilized a surrogate endpoint for bleeding based upon blood transfusion practice and need for surgical re-exploration. Although transfusion is a surrogate for bleeding, there are no universally accepted definitions for bleeding in postoperative cardiac surgical patients upon
RI PT
which the study could be based. Bleeding occurs on a continuum, and for ease of analysis a composite endpoint that included threshold level of caregiver response was selected. Given our institutional practice pattern of multiple blood product transfusion and surgical re-exploration
SC
only in patients with significant bleeding, these surrogate measures were selected to define our threshold for bleeding. Chest tube blood loss is an unreliable measure of bleeding in the
M AN U
operating room, although chest tube output in the ICU has been used in several previous studies. Survey-based measures of bleeding are also highly subjective, and can only be performed in a prospective study. Nevertheless, the endpoints chosen for this study carry inherent subjectivity, and represent a limitation of this study.
Patients with abnormally low MA (< 45 mm) who received platelet transfusion in the OR
TE D
appeared to have significantly lower risk of reaching composite endpoint in the ICU compared to patients who did not receive a platelet transfusion. Although this study was unable to further distinguish the nature of this relationship, one possible explanation for this finding is that
EP
patients with persistent coagulopathy continue to demonstrate increase risk of bleeding subsequently in the ICU. The significantly lower risk of bleeding in patients receiving platelet
AC C
transfusion in the OR suggests that correcting the abnormally low MA may prevent subsequent bleeding.
Analysis of patients with normal MA in the OR suggested that the risk of surrogate endpoint of bleeding in the ICU was not significantly different among patients who did or did not receive platelet transfusion. One implication of this finding is that platelet transfusion may not reduce the risk of subsequent bleeding in patients with near-normal MA. The use of TEG MA to guide platelet transfusion warrants further investigation in a prospective trial. 11
ACCEPTED MANUSCRIPT
A major limitation of this study is its retrospective design without a control group. The lack of prospective monitoring of sample collection in OR may have resulted in blood product transfusions prior to collection of the post-protamine TEG sample. TEG reference ranges differ
RI PT
depending upon age (19). Since this study was performed in a specific pediatric population and from a single center, results may not be generalizable to other pediatric populations. Other limitations include lack of fibrinogen quantification in this cohort, lack of quantification of
SC
bleeding, and lack of standardized transfusion protocol. Although some significant associations were found, the model accuracy estimates are modest, suggesting that other important factors
M AN U
exist that influence bleeding risk.
In conclusion, low MA measured by thromboelastography may be associated with increased surrogate endpoints of bleeding in pediatric patients undergoing high-risk cardiac surgery. Platelet transfusion is associated with increase in TEG MA parameter. Prophylactic platelet transfusions may prevent subsequent bleeding in patients with MA < 45 mm, but not in patients
TE D
with MA ≥ 45 mm. Randomized, multicenter clinical trials would be necessary to determine the
AC C
EP
utility of thromboelastography for blood product management in this patient population.
12
ACCEPTED MANUSCRIPT
References 1.
Miller BE, Mochizuki T, Levy JH et al. Predicting and treating coagulopathies after
cardiopulmonary bypass in children. Anesth Analg 1997;85(6):1196-1202. Guzzetta NA, Allen NN, Wilson EC, Foster GS, Ehrlich AC, Miller BE. Excessive
RI PT
2.
postoperative bleeding and outcomes in neonates undergoing cardiopulmonary bypass. Anesth Analg 2015;120(2):405-410.
Hayashi T, Sakurai Y, Fukuda K et al. Correlations between global clotting function tests,
SC
3.
duration of operation, and postoperative chest tube drainage in pediatric cardiac surgery.
4.
M AN U
Paediatr Anaesth 2011;21(8):865-871.
Segal JB, Dzik WH, Transfusion Medicine/Hemostasis Clinical Trials N. Paucity of
studies to support that abnormal coagulation test results predict bleeding in the setting of invasive procedures: An evidence-based review. Transfusion 2005;45(9):1413-1425. 5.
Kitchens CS. To bleed or not to bleed? Is that the question for the ptt? J Thromb
6.
Bolliger D, Tanaka KA. Point-of-care coagulation testing in cardiac surgery. Semin
Thromb Hemost 2017.
Kane LC, Woodward CS, Husain SA, Frei-Jones MJ. Thromboelastography--does it
EP
7.
TE D
Haemost 2005;3(12):2607-2611.
impact blood component transfusion in pediatric heart surgery? J Surg Res 2016;200(1):21-27. American Society of Anesthesiologists Task Force on Perioperative Blood M. Practice
AC C
8.
guidelines for perioperative blood management: An updated report by the american society of anesthesiologists task force on perioperative blood management*. Anesthesiology 2015;122(2):241-275. 9.
Kozek-Langenecker SA, Ahmed AB, Afshari A et al. Management of severe
perioperative bleeding: Guidelines from the european society of anaesthesiology: First update 2016. Eur J Anaesthesiol 2017;34(6):332-395.
13
ACCEPTED MANUSCRIPT
10.
Rafiq S, Johansson PI, Kofoed KF, Olsen PS, Steinbruchel DA. Preoperative hemostatic
testing and the risk of postoperative bleeding in coronary artery bypass surgery patients. J Card Surg 2016;31(9):565-571. Rizza A, Ricci Z, Pezzella C et al. Kaolin-activated thromboelastography and standard
RI PT
11.
coagulation assays in cyanotic and acyanotic infants undergoing complex cardiac surgery: A prospective cohort study. Paediatr Anaesth 2017;27(2):170-180.
Welsh KJ, Padilla A, Dasgupta A, Nguyen AN, Wahed A. Thromboelastography is a
SC
12.
suboptimal test for determination of the underlying cause of bleeding associated with
M AN U
cardiopulmonary bypass and may not predict a hypercoagulable state. Am J Clin Pathol;2014. p. 492-497. 13.
Pekelharing J, Furck A, Banya W, Macrae D, Davidson SJ. Comparison between
thromboelastography and conventional coagulation tests after cardiopulmonary bypass surgery in the paediatric intensive care unit. Int J Lab Hematol 2014;36(4):465-471. Niebler RA, Gill JC, Brabant CP et al. Thromboelastography in the assessment of
TE D
14.
bleeding following surgery for congenital heart disease. World J Pediatr Congenit Heart Surg 2012;3(4):433-438.
Ferraris VA, Bolanos MD. Use of antiplatelet drugs after cardiac operations. Semin
EP
15.
Thorac Cardiovasc Surg 2014;26(3):223-230. Hornykewycz S, Odegard KC, Castro RA, Zurakowski D, Pigula F, DiNardo JA.
AC C
16.
Hemostatic consequences of a non-fresh or reconstituted whole blood small volume cardiopulmonary bypass prime in neonates and infants. Paediatr Anaesth 2009;19(9):854-861. 17.
Woodman RC, Harker LA. Bleeding complications associated with cardiopulmonary
bypass. Blood 1990;76(9):1680-1697. 18.
Faraoni D, Emani S, Halpin E et al. Relationship between transfusion of blood products
and the incidence of thrombotic complications in neonates and infants undergoing cardiac surgery. J Cardiothorac Vasc Anesth 2017. 14
ACCEPTED MANUSCRIPT
19.
Chan KL, Summerhayes RG, Ignjatovic V, Horton SB, Monagle PT. Reference values
for kaolin-activated thromboelastography in healthy children. Anesth Analg 2007;105(6):1610-
AC C
EP
TE D
M AN U
SC
RI PT
1613, table of contents.
15
ACCEPTED MANUSCRIPT
Table 1. Patient characteristics and surgical details (N= 511) Median (IQR) or N (%)
Age at surgery (months)
11 (2,36)
Neonates
97 (19%)
Weight (kg)
7.8 (4.1, 13.5)
Body surface area
0.4 (0.3, 0.6)
Male
302 (59%)
CPB time, time (min)
151 (112, 194)
Cross-Clamp time, time (min)
86 (43, 121)
Cardiac Procedure Type
Number of Patients
SC
M AN U
Valve procedures
RI PT
Patient Characteristic
101
Complex bi-ventricular repairs Single ventricle palliation
73
Right ventricular outflow tract repair
HLHS Stage 1 CAVC repair
EP
Pulmonary vein surgery
70
44
TE D
Aortic reconstruction repair BT Shunt
90
35 32 19 17 14
Truncus repair
6
AC C
Total anomalous pulmonary venous return repair
Left ventricular outflow obstruction repair
5
Ventricular assist device placement
4
Heart transplant
1
Number (%) reaching composite endpoint in OR
181 (35%)
Number (%) reaching composite endpoint in ICU
47 (9%)
Median (IQR) or number (%) are represented in the table. IQR: Inter Quartile Range; CPB: Cardiopulmonary bypass; BT shunt: Blalock–Thomas–Taussig shunt; HLHS: Hypoplastic Left Heart Syndrome; CAVC: Complete Atrioventricular Canal 16
ACCEPTED MANUSCRIPT
Table 2. Post-bypass TEG parameter profile in the OR (n=511)
Median (IQR) Overall
Patients
Patients who
P-value
RI PT
TEG Parameters
reaching
did not reach
composite
composite
endpoint
endpoint
6.9 (5.7, 8.1)
0.43
6.9 (5.8, 8)
6.9 (6.0, 8.0)
K-Time minutes
2.6 (2.0, 3.4)
2.8 ((2.0, 3.8)
2.5 (2.0, 3.2)
0.03
Angle degrees
58.2 (51.7, 63.3)
57.9 (49.5, 63.3)
58.5 (53.1, 63.3)
0.20
MA mm
50.6 (45.0, 54.4)
47.7 (41.9, 52.9)
51.3 (46.8, 54.6)
<0.001
G parameter dynes/sec
2.1 (4.1, 5.9)
4.6 (3.6, 5.6)
5.3 (4.4, 6.0)
<0.001
Ly30 %
0.0 (0.0, 0.3)
0.0 (0.0, 0.1)
0.0 (0.0, 0.4)
0.3
TE D
M AN U
SC
R-Time minutes
AC C
EP
TEG: Thromboelastography; MA: Maximum amplitude; Ly30: clot lysis % after 30 minutes
17
ACCEPTED MANUSCRIPT
Table 3. Multivariable logistic regression model for the probability of reaching composite
Odds ratio
95% CI
P-value
Neonate
6.07
3.70, 9.96
<0.001
Maximum amplitude, mm
0.95
0.92, 0.97
<0.001
AC C
EP
TE D
M AN U
SC
Models
RI PT
endpoint in the OR N=181, AUC=0.71)
18
ACCEPTED MANUSCRIPT
Table 4. Association of reaching composite endpoint in ICU with platelet transfusion in OR Overall (N=511)
Composite endpoint
0.15
YES (N=323)
25 (7.7%)
NO (N=188)
22 (11.7%)
Abnormal MA (< 45 mm)
RI PT
Platelet transfusion in the OR
P-value
0.02
SC
Platelet transfusion in the OR 7 (7.5%)
NO (n=33)
8 (24.2%)
Normal MA (≥ 45 mm) Platelet transfusion in the OR YES (n=230)
M AN U
YES (n=93)
0.71
18 (7.8%)
14 (9.0%)
AC C
EP
TE D
NO (n=155)
19
RI PT
ACCEPTED MANUSCRIPT
SC
Table 5. Logistic regression model results for the composite endpoint in the ICU (N = 47)
Odds Ratio
M AN U
Post-bypass MA group
95% CI
P-Value 0.10
Administration of platelet transfusion in OR
0.02
Administration of platelet transfusion in OR & MA
0.07
MA < 45 mm: Platelet transfusion vs. No Platelet
0.25
0.08, 0.77
0.02
TE D
interaction
0.86
0.41, 1.78
0.67
transfusion in OR
MA ≥ 45 mm: Platelet transfusion vs. No Platelet
AC C
EP
transfusion in OR
20
ACCEPTED MANUSCRIPT
Figure Legends Figure 1: Probability of reaching the composite endpoint in the OR as a function of each of the continuous post-bypass TEG parameters. The solid line indicates the estimated probability
RI PT
based on a linear or piecewise linear fit from logistic regression. The dashed line indicates the estimated probability based on a flexible fit from a generalized additive model. Only clot strength G parameter had a significant nonlinear relationship with composite endpoint.
SC
Figure 2: Maximum amplitude by TEG in the operating room (OR) and intensive care unit (ICU) in patients with MA < 45 mm who did or did not receive platelet transfusion. The error bars
M AN U
indicate median (IQR). The dashed line indicates abnormally low MA cut-off (45 mm). In patients who did not receive platelet transfusion in the OR there was no significant difference in the MA values between the OR and ICU P=0.6. Patients who received platelet transfusion in the OR demonstrated a significant increase in MA between OR and ICU (P<0.001). P-values
TE D
less than 0.05 were considered statistically significant.
Figure 3: Association between blood product transfusions and TEG parameters measured in the OR and ICU. The error bars indicate median (IQR). (A) Angle significantly increased from
EP
OR to ICU among patients who received either platelet transfusion alone (P<0.001) or platelet transfusion and cryoprecipitate (P<0.001) in the OR. (B) G parameter significantly increased
AC C
from OR to ICU among patients who received either platelet transfusion alone (P<0.001) or platelets and cryoprecipitate (P<0.001) in the OR. G parameter was significantly higher in patients who received platelet transfusion with cryoprecipitate compared with patients who received platelet transfusion alone in the OR (P<0.001). P-values less than 0.05 were considered statistically significant.
21
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0003-4975(18)30649-0
DOI:
10.1016/j.athoracsur.2018.04.023
Reference:
ATS 31586
To appear in:
The Annals of Thoracic Surgery
Received Date: 29 October 2017 Revised Date:
6 April 2018
Accepted Date: 11 April 2018
Please cite this article as: Emani S, Sleeper LA, Faraoni D, Mulone M, Diallo F, DiNardo JA, Ibla J, Emani SM, Thromboelastography is Associated with Surrogates for Bleeding after Pediatric Cardiac Surgery, The Annals of Thoracic Surgery (2018), doi: 10.1016/j.athoracsur.2018.04.023. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Thromboelastography is Associated with Surrogates for Bleeding after Pediatric Cardiac Surgery
RI PT
Running Head: TEG is associated with bleeding
1
Sirisha Emani, PhD, 1Lynn A. Sleeper, ScD, 3David Faraoni, MD, PhD, 1Michelle Mulone, BS, Fatoumata Diallo, BS, 2James A. DiNardo, MD, 2Juan Ibla, MD, 1Sitaram M. Emani, MD
SC
1
M AN U
1
Cardiovascular Program, Boston Children’s Hospital, Massachusetts, USA
2
Department of Anesthesia and Pain Medicine, Boston Children’s Hospital, Massachusetts, USA
3
Correspondence:
EP
Word count: 4498
TE D
Department of Anesthesia and Pain Medicine, Hospital for Sick Children, Toronto, Canada
AC C
Sitaram Emani, MD
300 Longwood Ave. Boston, MA 02115
Email: [email protected]
1
ACCEPTED MANUSCRIPT
Abstract Background: Perioperative bleeding is a common complication in pediatric patients undergoing cardiac surgery. Although thromboelastography (TEG) has been utilized in patients undergoing
RI PT
adult cardiac surgery, limited data are available in pediatric patients. We hypothesize that TEG parameters may be associated with surrogate endpoints for postoperative bleeding in pediatric patients undergoing complex cardiac surgery.
SC
Methods: In a retrospective study, TEG was obtained after protamine administration and upon intensive care unit (ICU) admission in pediatric patients (≤ 18yrs) undergoing cardiac surgery
M AN U
requiring cardiopulmonary bypass. A composite endpoint of extended blood product transfusion or surgical re-exploration for bleeding was utilized as a surrogate for perioperative bleeding. TEG parameters were compared between patients who did or did not reach the composite endpoint.
TE D
Results: The study included 511 pediatric patients undergoing complex cardiac surgery. Composite endpoint was reached in 52% of patients with MA< 45 mm compared to 31% with MA ≥ 45 mm (P<0.001). Using multivariable regression analysis, MA < 45 mm was
EP
independently associated with composite endpoint (P<0.001). Patients with MA < 45 mm who received platelet transfusion in the OR were less likely to reach composite endpoint within the
AC C
subsequent 24 hours (8%) compared to those who did not receive intraoperative platelet transfusion (24%) (P=0.02).
Conclusions: Intraoperative TEG MA < 45 mm is associated with a surrogate endpoint for intraoperative bleeding in pediatric patients undergoing complex cardiac surgery. In patients with MA < 45 mm, prophylactic platelet transfusion in the OR may be associated with reduction in bleeding endpoints in the ICU.
2
ACCEPTED MANUSCRIPT
Coagulopathy after cardiopulmonary bypass (CPB) is known to be complex and multifactorial, and has been shown to be associated with increased transfusion requirements and hemodynamic instability, increasing both morbidity and mortality (1). The pediatric population is
RI PT
at increased risk of coagulopathy, perioperative bleeding, and transfusion requirements compared to adults (2). Standard coagulation assays have been used for decades to guide anticoagulation in adults and children (3, 4). These tests were not designed to monitor
SC
perioperative coagulopathy or guide the administration of hemostatic agents in bleeding patients (5).
M AN U
Recent experience in patients undergoing cardiac surgery has suggested a potential benefit of thromboelastography (TEG) for risk stratification and blood product management following adult and pediatric cardiac surgery (6, 7). The use of viscoelastic tests is now recommended in both American and European guidelines regarding perioperative bleeding management (8, 9). To date, the predictive accuracy of TEG has only been studied in a small number of retrospective
TE D
studies that included infants and children undergoing both low and high complexity congenital cardiac surgical procedures, leading to heterogeneous and conflicting results (10-12). The objective of this study was to determine whether abnormal TEG parameters are associated
EP
with surrogate endpoints of early postoperative bleeding in a group of pediatric patients undergoing complex cardiac surgical procedures and considered at high risk for perioperative
AC C
bleeding.
Patients and Methods Study Design
3
ACCEPTED MANUSCRIPT
Consecutive pediatric patients (age ≤ 18yrs) undergoing cardiac surgery at Boston Children's Hospital (BCH) between April 2015 and March 2017 were retrospectively studied after approval by Institutional Review Board. The target population was patients at high-risk for bleeding
RI PT
following a cardiac surgical procedure requiring cardiopulmonary bypass (CPB). A patient was eligible if any one of the following criteria were met: 1. Neonatal age at surgery (age ≤ 30 days), 2. Single ventricle physiology, 3. Re-operative sternotomy with greater than
SC
two previous operations, 4. Complex cardiac reconstruction with prolonged duration of CPB. Patients were excluded if testing for TEG was not performed at all time points or if patients were
M AN U
transfused prior to testing for TEG in the OR.
For this cohort of patients, CPB circuit was primed with reconstituted whole blood (1 unit of packed red blood cells (RBC) and 1 unit of fresh frozen plasma (FFP)) to obtain target hematocrit of >30% in all patients. RBC obtained by cell salvage of CPB circuit blood was routinely administered in all patients. Anticoagulation with heparin was performed to achieve
TE D
ACT goal of >480, and at least moderate hypothermia (< 32oC) was utilized. Blood gas management included pH stat strategy for patients cooled below 28oC.
EP
Data collection
Patient characteristics data were collected from the medical record. Transfusions of RBC,
AC C
platelets, cryoprecipitate, FFP were recorded along with administration of activated recombinant Factor VII (rFVIIa) and albumin. Procedural records were reviewed to determine any episodes of surgical re-exploration for bleeding following temporary or permanent chest closure. Study Endpoints
A composite endpoint was established as a surrogate for perioperative bleeding that included transfusion requirements or need for surgical re-exploration for bleeding. In our institution, transfusion of RBC and platelets are considered as first line therapy following administration of 4
ACCEPTED MANUSCRIPT
protamine in patients with subjective bleeding. Although RBC and platelet transfusion may also be administered in the absence of bleeding (for anemia or thrombocytopenia), treatment with cryoprecipitate, FFP, or rFVIIa is limited by institutional protocol to patients with subjective
RI PT
bleeding refractory to first line therapy. Thus, extended transfusion was defined as the administration of one or more of these blood products in addition to RBC and platelets. The composite endpoint included either extended transfusion or surgical re-exploration for bleeding.
SC
The composite endpoint was further classified as occurring in the OR or within 24 hours of ICU admission depending upon location of the patient at the time of event. Remote bleeding
Thromboelastography (TEG)
M AN U
episodes beyond the first 24 hours were not included in analysis.
TEG was performed in a CLIA certified laboratory after protamine administration in the OR and upon ICU admission. Activation with Kaolin was achieved in both standard and heparinase
TE D
cups. TEG parameters obtained included Reaction time: time from initiation of clot formation to the clot strength of 2 mm (R-time), kinetic clot time: time to achieve a clot strength of 20 mm (Ktime), rate of clot formation (Angle), maximum amplitude (MA), overall clot strength (G parameter- a calculated value which is the log derivation of MA), and percentage of clot lysis 30
EP
min after the MA is reached (Ly30). TEG values were not utilized to guide transfusion
AC C
management in this cohort.
In preliminary analysis, each TEG parameter was tested based on the risk characterization as high risk (MA< 45 mm), intermediate risk (MA 45-50 mm), and low risk (MA > 50 mm) groups obtained by Cart analysis. As there was no significant difference between intermediate and low risk groups in relation to bleeding outcomes, we reanalyzed using only two groups (MA<45 and >45 mm). Statistical Methods 5
ACCEPTED MANUSCRIPT
Categorical data were described as number with frequency and continuous data as median with interquartile range (IQR) or mean with standard deviation. Association between TEG parameters and composite endpoint was assessed by logistic regression. Association between
RI PT
TEG parameters in OR and composite endpoint in ICU was determined by using Fisher exact test. Generalized additive modeling (GAM) was used to estimate the shape of relationship
between TEG parameters and the probability of reaching composite endpoint. If a significant
SC
non-linear relationship was found, the continuous predictor was categorized, or modeled using a piecewise linear fit to assess association with probability of composite endpoint. Multivariable
M AN U
logistic stepwise regression was conducted to identify independent predictors of reaching composite endpoint using candidate variables found to be significant at the 0.05 level in univariate analysis. Statistical analysis was performed using SAS version 9.4 (SAS Institute,
Results
TE D
Cary, NC) and R version 3.2.1. P-values less than 0.05 were considered statistically significant.
Patient Characteristics, post-bypass TEG parameters and blood products administration
EP
Five hundred and eleven pediatric patients who underwent high-risk cardiac surgical procedures requiring CPB and had TEG testing were retrospectively studied. Patient characteristics
AC C
including age at surgery, body weight, CPB, cross-clamp time (CCT), and procedure type are shown in Table 1. The mean age at surgery was 30 months and the median (IQR) was 11 months (2, 36); 97 (19%) were categorized as neonates (≤ 30 days of age at surgery). Nonautologous blood product transfusion excluding pump priming products (RBC and FFP) was necessary in 360/511 (71%) of patients in the cohort, including RBC (non-salvaged) in 125/511 (25%), platelets in 323/511 (63%), cryoprecipitate in 181/511 (35%), and FFP in 18/511 (4%) of patients. rFVIIa was administered in 28/511 (6%) of patients. The composite endpoint was 6
ACCEPTED MANUSCRIPT
reached in the OR and within 24 hours of ICU admission in 181/511 (35%) and 47/511 (9%) of patients, respectively (Table 1). Re-exploration for bleeding occurred in 6 of 181 patients either in the OR following initial chest closure (2 patients) or ICU (6 patients), with 2 patients requiring
Association of age at surgery with composite endpoint and TEG
RI PT
re-exploration in both locations.
A higher percentage of neonates compared to non-neonates reached composite endpoint in the
SC
OR (70% vs. 27%, P<0.001) but not in the ICU (11% vs. 9%, P= 0.44). Mean post-bypass MA (47 ± 9 mm vs. 49 ± 8 mm, P<0.05) and clot strength G (4.8 ± 1.6 vs. 5.1 ± 1.4 dynes/sec,
M AN U
P<0.05) were significantly lower in neonates compared to non-neonates. However, age at surgery was not correlated with TEG parameters. The absolute Pearson correlation coefficient r for reaction time is 0.05, K- time is 0.04, rate of clot formation is 0.02, maximum amplitude 0.05, clot strength is 0.05 and clot lysis is 0.07 (data not shown).
TE D
TEG parameters and association with composite endpoint in OR
Median values of each TEG parameter are displayed in Table 2 and were compared between patients who did or did not reach the composite endpoint in the OR. Maximal Amplitude, K-
EP
time, and G parameter, but not R-time, Angle, or Ly30 were significantly different between patients who did or did not reach composite endpoint. Figure 1 displays the estimated
AC C
probability of reaching composite endpoint in the OR as a function of the continuous postprotamine TEG parameters. Longer K-time, lower MA and G parameter were significantly associated with probability of composite endpoint. There was a non-linear relationship between clot strength G parameter and probability of composite endpoint. Weaker clot strength ≤ 5.8 dynes/sec was linearly associated with the probability of composite endpoint, but no association was observed at levels > 5.8 dynes/sec. R-time, Angle and Ly30 were not associated with probability of composite endpoint. 7
ACCEPTED MANUSCRIPT
The multivariable regression model for probability of composite endpoint in the OR (Table 3) included only parameters that were significantly associated with composite endpoint in the univariate analysis - neonate vs. non-neonate, continuous age at surgery, and the post-
RI PT
protamine MA. Neonatal age (odds ratio 6.07), and lower MA (odds ratio 0.95 per mm) were independently associated with higher odds of composite endpoint in the OR. The area under the curve (AUC) of 0.71 suggests moderate predictive power of this model to predict composite
SC
endpoint (Table 3). Volume of total blood products (excluding pump prime products)
administered in the OR post-protamine (normalized to body weight) was significantly higher in
M AN U
patients with TEG MA < 45 compared to those with MA > 45 (median 32.3 ml/kg vs. 22.2 ml/kg respectively, P=0.002).
Association between platelet transfusion in OR and composite endpoint in the ICU A subgroup analysis was performed to assess whether there is a differential association
TE D
between platelet transfusion in the OR and reaching composite endpoint within 24 hours of ICU admission depending on MA level measured in the OR. Among patients with post-protamine MA < 45 mm, the percentage reaching composite endpoint in the ICU was lower for the 93 patients who received platelet transfusion in the OR compared to the 33 patients who did not
EP
receive platelet transfusion in the OR (8% vs. 24%, P=0.02). Among patients with normal MA (≥ 45 mm) in the OR, there was no significant difference in the rate of composite endpoint in ICU
AC C
between patients who received or did not receive platelet transfusions in the OR (8% vs. 9% respectively, P=0.71). (Table 4). By logistic regression modeling, intraoperative platelet transfusion in patients with post-protamine MA < 45 mm was associated with significantly lower odds of composite endpoint in the ICU (OR=0.25, P=0.02). In patients with MA ≥ 45 mm, there was no association between intraoperative platelet transfusion and composite endpoint in the ICU (OR=0.86, P=0.67) (Table 5). However, there was insufficient evidence statistically to state
8
ACCEPTED MANUSCRIPT
that the effect of platelets on composite endpoint differs by MA group, with a logistic regression interaction p-value of 0.07.
RI PT
Association between Platelet transfusion and TEG parameters TEG parameters obtained in the ICU were compared to parameters measured in the OR prior to blood product transfusion. For patients with MA < 45 mm in the OR who did not receive platelet transfusion, there was no significant difference in the median (IQR) MA values between the OR
SC
and ICU (41.7 (39.7, 43.4) mm vs. 40.6 (35.8, 46.1) mm, P=0.6). On the other hand, patients with MA < 45 mm in OR who received platelet transfusion (median of 15 cc/kg of apheresis
M AN U
platelets) demonstrated significant increase in median (IQR) MA from 40.8 (37.5, 43) mm in the OR to 63.4 (56.2, 69.7) mm in the ICU (P<0.001) (Figure 2). Patients who received both cryoprecipitate and platelet transfusion in the OR had significant increase in median (IQR) Angle and G parameter between OR and ICU (58 (51, 63.2) degrees and 4.5 (3.6, 5.6) dynes/sec in
TE D
OR vs. 69 (63.8, 73.5) degrees and 9.9 (7.8, 12.6) dynes/sec in the ICU, P<0.001 and P<0.001, respectively). The median (IQR) overall clot strength G parameter measured in the ICU was significantly higher in patients receiving both cryoprecipitate and platelet transfusion compared to patients receiving platelet transfusion alone (9.9 (7.8, 12.6) vs. 7.5 (6.3, 8.8) dynes/sec,
EP
P<0.001). Patients who received platelet transfusion in the OR also demonstrated significant increase in median (IQR) Angle and G parameter between OR and ICU 59.3 (54.5, 63.9)
AC C
degrees and 5.2 (4.3, 6.1) dynes/sec in OR vs. 64.5 (60.6, 69) degrees and 7.5 (6.3, 8.8) dynes/sec in the ICU, P<0.001 and P<0.001, respectively) (Figure 3).
Comment This study of pediatric patients undergoing high-risk cardiac surgery demonstrates an association between TEG parameters obtained immediately following discontinuation of CPB 9
ACCEPTED MANUSCRIPT
and composite endpoint of extended transfusion or re-exploration for bleeding in the OR. Neonatal age (≤ 30 days) at the time of surgery and lower MA were independently associated with increased risk of reaching composite endpoint. Platelet transfusion was associated with
RI PT
significant increase in the MA between OR and ICU. Patients with MA < 45 mm in the OR who did not receive platelet transfusion demonstrated higher rate of composite endpoint in ICU compared to those who did receive platelet transfusion.
SC
Previous studies have not demonstrated a consistent correlation between post-protamine TEG and postoperative bleeding in cohorts of pediatric patients undergoing cardiac surgery (13, 14).
M AN U
In contrast to these previous studies, the current study investigated only patients with a wellaccepted high-risk for bleeding following cardiac surgery. Our cohort of patients was specifically selected since this is the population among whom coagulation testing may be beneficial to mitigate the risk of postoperative bleeding. In this population of high-risk patients, significant association between TEG parameters and the composite endpoint as a surrogate for bleeding
TE D
was observed.
The data in this study is consistent with previous reports that suggest that deficiencies in platelets and fibrinogen, rather than coagulation factors, are associated with postoperative
EP
bleeding following cardiac surgery (15-17). Maximum amplitude measures clot strength, a composite effect of platelet and fibrinogen activity. In this study, the MA frequently normalized
AC C
with platelet transfusion alone, and suggests that cryoprecipitate may be unnecessary in the management of post-bypass coagulopathy in this patient population. Elevated Angle and clot strength G parameter in patients receiving cryoprecipitate and platelet transfusion compared to platelet transfusion alone may increase risk of thrombosis. Previous studies using ROTEM suggested that cryoprecipitate transfusion may result in increased thrombotic risk (18). The TEG functional fibrinogen assay may be useful in determining need for cryoprecipitate transfusion in this setting. 10
ACCEPTED MANUSCRIPT
This study utilized a surrogate endpoint for bleeding based upon blood transfusion practice and need for surgical re-exploration. Although transfusion is a surrogate for bleeding, there are no universally accepted definitions for bleeding in postoperative cardiac surgical patients upon
RI PT
which the study could be based. Bleeding occurs on a continuum, and for ease of analysis a composite endpoint that included threshold level of caregiver response was selected. Given our institutional practice pattern of multiple blood product transfusion and surgical re-exploration
SC
only in patients with significant bleeding, these surrogate measures were selected to define our threshold for bleeding. Chest tube blood loss is an unreliable measure of bleeding in the
M AN U
operating room, although chest tube output in the ICU has been used in several previous studies. Survey-based measures of bleeding are also highly subjective, and can only be performed in a prospective study. Nevertheless, the endpoints chosen for this study carry inherent subjectivity, and represent a limitation of this study.
Patients with abnormally low MA (< 45 mm) who received platelet transfusion in the OR
TE D
appeared to have significantly lower risk of reaching composite endpoint in the ICU compared to patients who did not receive a platelet transfusion. Although this study was unable to further distinguish the nature of this relationship, one possible explanation for this finding is that
EP
patients with persistent coagulopathy continue to demonstrate increase risk of bleeding subsequently in the ICU. The significantly lower risk of bleeding in patients receiving platelet
AC C
transfusion in the OR suggests that correcting the abnormally low MA may prevent subsequent bleeding.
Analysis of patients with normal MA in the OR suggested that the risk of surrogate endpoint of bleeding in the ICU was not significantly different among patients who did or did not receive platelet transfusion. One implication of this finding is that platelet transfusion may not reduce the risk of subsequent bleeding in patients with near-normal MA. The use of TEG MA to guide platelet transfusion warrants further investigation in a prospective trial. 11
ACCEPTED MANUSCRIPT
A major limitation of this study is its retrospective design without a control group. The lack of prospective monitoring of sample collection in OR may have resulted in blood product transfusions prior to collection of the post-protamine TEG sample. TEG reference ranges differ
RI PT
depending upon age (19). Since this study was performed in a specific pediatric population and from a single center, results may not be generalizable to other pediatric populations. Other limitations include lack of fibrinogen quantification in this cohort, lack of quantification of
SC
bleeding, and lack of standardized transfusion protocol. Although some significant associations were found, the model accuracy estimates are modest, suggesting that other important factors
M AN U
exist that influence bleeding risk.
In conclusion, low MA measured by thromboelastography may be associated with increased surrogate endpoints of bleeding in pediatric patients undergoing high-risk cardiac surgery. Platelet transfusion is associated with increase in TEG MA parameter. Prophylactic platelet transfusions may prevent subsequent bleeding in patients with MA < 45 mm, but not in patients
TE D
with MA ≥ 45 mm. Randomized, multicenter clinical trials would be necessary to determine the
AC C
EP
utility of thromboelastography for blood product management in this patient population.
12
ACCEPTED MANUSCRIPT
References 1.
Miller BE, Mochizuki T, Levy JH et al. Predicting and treating coagulopathies after
cardiopulmonary bypass in children. Anesth Analg 1997;85(6):1196-1202. Guzzetta NA, Allen NN, Wilson EC, Foster GS, Ehrlich AC, Miller BE. Excessive
RI PT
2.
postoperative bleeding and outcomes in neonates undergoing cardiopulmonary bypass. Anesth Analg 2015;120(2):405-410.
Hayashi T, Sakurai Y, Fukuda K et al. Correlations between global clotting function tests,
SC
3.
duration of operation, and postoperative chest tube drainage in pediatric cardiac surgery.
4.
M AN U
Paediatr Anaesth 2011;21(8):865-871.
Segal JB, Dzik WH, Transfusion Medicine/Hemostasis Clinical Trials N. Paucity of
studies to support that abnormal coagulation test results predict bleeding in the setting of invasive procedures: An evidence-based review. Transfusion 2005;45(9):1413-1425. 5.
Kitchens CS. To bleed or not to bleed? Is that the question for the ptt? J Thromb
6.
Bolliger D, Tanaka KA. Point-of-care coagulation testing in cardiac surgery. Semin
Thromb Hemost 2017.
Kane LC, Woodward CS, Husain SA, Frei-Jones MJ. Thromboelastography--does it
EP
7.
TE D
Haemost 2005;3(12):2607-2611.
impact blood component transfusion in pediatric heart surgery? J Surg Res 2016;200(1):21-27. American Society of Anesthesiologists Task Force on Perioperative Blood M. Practice
AC C
8.
guidelines for perioperative blood management: An updated report by the american society of anesthesiologists task force on perioperative blood management*. Anesthesiology 2015;122(2):241-275. 9.
Kozek-Langenecker SA, Ahmed AB, Afshari A et al. Management of severe
perioperative bleeding: Guidelines from the european society of anaesthesiology: First update 2016. Eur J Anaesthesiol 2017;34(6):332-395.
13
ACCEPTED MANUSCRIPT
10.
Rafiq S, Johansson PI, Kofoed KF, Olsen PS, Steinbruchel DA. Preoperative hemostatic
testing and the risk of postoperative bleeding in coronary artery bypass surgery patients. J Card Surg 2016;31(9):565-571. Rizza A, Ricci Z, Pezzella C et al. Kaolin-activated thromboelastography and standard
RI PT
11.
coagulation assays in cyanotic and acyanotic infants undergoing complex cardiac surgery: A prospective cohort study. Paediatr Anaesth 2017;27(2):170-180.
Welsh KJ, Padilla A, Dasgupta A, Nguyen AN, Wahed A. Thromboelastography is a
SC
12.
suboptimal test for determination of the underlying cause of bleeding associated with
M AN U
cardiopulmonary bypass and may not predict a hypercoagulable state. Am J Clin Pathol;2014. p. 492-497. 13.
Pekelharing J, Furck A, Banya W, Macrae D, Davidson SJ. Comparison between
thromboelastography and conventional coagulation tests after cardiopulmonary bypass surgery in the paediatric intensive care unit. Int J Lab Hematol 2014;36(4):465-471. Niebler RA, Gill JC, Brabant CP et al. Thromboelastography in the assessment of
TE D
14.
bleeding following surgery for congenital heart disease. World J Pediatr Congenit Heart Surg 2012;3(4):433-438.
Ferraris VA, Bolanos MD. Use of antiplatelet drugs after cardiac operations. Semin
EP
15.
Thorac Cardiovasc Surg 2014;26(3):223-230. Hornykewycz S, Odegard KC, Castro RA, Zurakowski D, Pigula F, DiNardo JA.
AC C
16.
Hemostatic consequences of a non-fresh or reconstituted whole blood small volume cardiopulmonary bypass prime in neonates and infants. Paediatr Anaesth 2009;19(9):854-861. 17.
Woodman RC, Harker LA. Bleeding complications associated with cardiopulmonary
bypass. Blood 1990;76(9):1680-1697. 18.
Faraoni D, Emani S, Halpin E et al. Relationship between transfusion of blood products
and the incidence of thrombotic complications in neonates and infants undergoing cardiac surgery. J Cardiothorac Vasc Anesth 2017. 14
ACCEPTED MANUSCRIPT
19.
Chan KL, Summerhayes RG, Ignjatovic V, Horton SB, Monagle PT. Reference values
for kaolin-activated thromboelastography in healthy children. Anesth Analg 2007;105(6):1610-
AC C
EP
TE D
M AN U
SC
RI PT
1613, table of contents.
15
ACCEPTED MANUSCRIPT
Table 1. Patient characteristics and surgical details (N= 511) Median (IQR) or N (%)
Age at surgery (months)
11 (2,36)
Neonates
97 (19%)
Weight (kg)
7.8 (4.1, 13.5)
Body surface area
0.4 (0.3, 0.6)
Male
302 (59%)
CPB time, time (min)
151 (112, 194)
Cross-Clamp time, time (min)
86 (43, 121)
Cardiac Procedure Type
Number of Patients
SC
M AN U
Valve procedures
RI PT
Patient Characteristic
101
Complex bi-ventricular repairs Single ventricle palliation
73
Right ventricular outflow tract repair
HLHS Stage 1 CAVC repair
EP
Pulmonary vein surgery
70
44
TE D
Aortic reconstruction repair BT Shunt
90
35 32 19 17 14
Truncus repair
6
AC C
Total anomalous pulmonary venous return repair
Left ventricular outflow obstruction repair
5
Ventricular assist device placement
4
Heart transplant
1
Number (%) reaching composite endpoint in OR
181 (35%)
Number (%) reaching composite endpoint in ICU
47 (9%)
Median (IQR) or number (%) are represented in the table. IQR: Inter Quartile Range; CPB: Cardiopulmonary bypass; BT shunt: Blalock–Thomas–Taussig shunt; HLHS: Hypoplastic Left Heart Syndrome; CAVC: Complete Atrioventricular Canal 16
ACCEPTED MANUSCRIPT
Table 2. Post-bypass TEG parameter profile in the OR (n=511)
Median (IQR) Overall
Patients
Patients who
P-value
RI PT
TEG Parameters
reaching
did not reach
composite
composite
endpoint
endpoint
6.9 (5.7, 8.1)
0.43
6.9 (5.8, 8)
6.9 (6.0, 8.0)
K-Time minutes
2.6 (2.0, 3.4)
2.8 ((2.0, 3.8)
2.5 (2.0, 3.2)
0.03
Angle degrees
58.2 (51.7, 63.3)
57.9 (49.5, 63.3)
58.5 (53.1, 63.3)
0.20
MA mm
50.6 (45.0, 54.4)
47.7 (41.9, 52.9)
51.3 (46.8, 54.6)
<0.001
G parameter dynes/sec
2.1 (4.1, 5.9)
4.6 (3.6, 5.6)
5.3 (4.4, 6.0)
<0.001
Ly30 %
0.0 (0.0, 0.3)
0.0 (0.0, 0.1)
0.0 (0.0, 0.4)
0.3
TE D
M AN U
SC
R-Time minutes
AC C
EP
TEG: Thromboelastography; MA: Maximum amplitude; Ly30: clot lysis % after 30 minutes
17
ACCEPTED MANUSCRIPT
Table 3. Multivariable logistic regression model for the probability of reaching composite
Odds ratio
95% CI
P-value
Neonate
6.07
3.70, 9.96
<0.001
Maximum amplitude, mm
0.95
0.92, 0.97
<0.001
AC C
EP
TE D
M AN U
SC
Models
RI PT
endpoint in the OR N=181, AUC=0.71)
18
ACCEPTED MANUSCRIPT
Table 4. Association of reaching composite endpoint in ICU with platelet transfusion in OR Overall (N=511)
Composite endpoint
0.15
YES (N=323)
25 (7.7%)
NO (N=188)
22 (11.7%)
Abnormal MA (< 45 mm)
RI PT
Platelet transfusion in the OR
P-value
0.02
SC
Platelet transfusion in the OR 7 (7.5%)
NO (n=33)
8 (24.2%)
Normal MA (≥ 45 mm) Platelet transfusion in the OR YES (n=230)
M AN U
YES (n=93)
0.71
18 (7.8%)
14 (9.0%)
AC C
EP
TE D
NO (n=155)
19
RI PT
ACCEPTED MANUSCRIPT
SC
Table 5. Logistic regression model results for the composite endpoint in the ICU (N = 47)
Odds Ratio
M AN U
Post-bypass MA group
95% CI
P-Value 0.10
Administration of platelet transfusion in OR
0.02
Administration of platelet transfusion in OR & MA
0.07
MA < 45 mm: Platelet transfusion vs. No Platelet
0.25
0.08, 0.77
0.02
TE D
interaction
0.86
0.41, 1.78
0.67
transfusion in OR
MA ≥ 45 mm: Platelet transfusion vs. No Platelet
AC C
EP
transfusion in OR
20
ACCEPTED MANUSCRIPT
Figure Legends Figure 1: Probability of reaching the composite endpoint in the OR as a function of each of the continuous post-bypass TEG parameters. The solid line indicates the estimated probability
RI PT
based on a linear or piecewise linear fit from logistic regression. The dashed line indicates the estimated probability based on a flexible fit from a generalized additive model. Only clot strength G parameter had a significant nonlinear relationship with composite endpoint.
SC
Figure 2: Maximum amplitude by TEG in the operating room (OR) and intensive care unit (ICU) in patients with MA < 45 mm who did or did not receive platelet transfusion. The error bars
M AN U
indicate median (IQR). The dashed line indicates abnormally low MA cut-off (45 mm). In patients who did not receive platelet transfusion in the OR there was no significant difference in the MA values between the OR and ICU P=0.6. Patients who received platelet transfusion in the OR demonstrated a significant increase in MA between OR and ICU (P<0.001). P-values
TE D
less than 0.05 were considered statistically significant.
Figure 3: Association between blood product transfusions and TEG parameters measured in the OR and ICU. The error bars indicate median (IQR). (A) Angle significantly increased from
EP
OR to ICU among patients who received either platelet transfusion alone (P<0.001) or platelet transfusion and cryoprecipitate (P<0.001) in the OR. (B) G parameter significantly increased
AC C
from OR to ICU among patients who received either platelet transfusion alone (P<0.001) or platelets and cryoprecipitate (P<0.001) in the OR. G parameter was significantly higher in patients who received platelet transfusion with cryoprecipitate compared with patients who received platelet transfusion alone in the OR (P<0.001). P-values less than 0.05 were considered statistically significant.
21
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT