Effect of High- vs Low-Dose Tranexamic Acid Infusion on Need for Red Blood Cell Transfusion and Adverse Events in Patients Undergoing Cardiac Surgery

Key Points

Question

Among patients undergoing cardiac surgery with cardiopulmonary bypass, is there a difference between high-dose and low-dose infusions of tranexamic acid with respect to the need for red blood cell transfusions and adverse events?

Findings

In this randomized clinical trial that included 3031 patients, high-dose compared with low-dose tranexamic acid infusion significantly reduced the proportion of patients who received allogeneic red blood cell transfusion (21.8% vs 26.0%, respectively). The rate of a composite safety end point of 30-day mortality, seizure, kidney dysfunction, and thrombotic events was 17.6% in the high-dose group and 16.8% in the low-dose group; the 97.55% CI for the difference was within the noninferiority margin of 5%.

Meaning

Among patients who underwent cardiac surgery with cardiopulmonary bypass, high-dose compared with low-dose tranexamic acid infusion resulted in a modest, statistically significant reduction in the proportion of patients receiving allogeneic red blood cell transfusion and met criteria for noninferiority with respect to a composite safety end point.

Abstract

Importance

Tranexamic acid is recommended for reducing blood loss and transfusion in cardiac surgery. However, it remains unknown whether a high dose of tranexamic acid provides better blood-sparing effect than a low dose without increasing the risk of thrombotic complications or seizures in cardiac surgery.

Objective

To compare the efficacy and adverse events of high-dose vs low-dose tranexamic acid in patients undergoing cardiac surgery with cardiopulmonary bypass.

Design, Setting, and Participants

Multicenter, double-blind, randomized clinical trial among adult patients undergoing cardiac surgery with cardiopulmonary bypass. The study enrolled 3079 patients at 4 hospitals in China from December 26, 2018, to April 21, 2021; final follow-up was on May 21, 2021.

Interventions

Participants received either a high-dose tranexamic acid regimen comprising a 30-mg/kg bolus, a 16-mg/kg/h maintenance dose, and a 2-mg/kg prime (n = 1525) or a low-dose regimen comprising a 10-mg/kg bolus, a 2-mg/kg/h maintenance dose, and a 1-mg/kg prime (n = 1506).

Main Outcomes and Measures

The primary efficacy end point was the rate of allogeneic red blood cell transfusion after start of operation (superiority hypothesis), and the primary safety end point was a composite of the 30-day postoperative rate of mortality, seizure, kidney dysfunction (stage 2 or 3 Kidney Disease: Improving Global Outcomes [KDIGO] criteria), and thrombotic events (myocardial infarction, ischemic stroke, deep vein thrombosis, and pulmonary embolism) (noninferiority hypothesis with a margin of 5%). There were 15 secondary end points, including the individual components of the primary safety end point.

Results

Among 3079 patients who were randomized to treatment groups (mean age, 52.8 years; 38.1% women), 3031 (98.4%) completed the trial. Allogeneic red blood cell transfusion occurred in 333 of 1525 patients (21.8%) in the high-dose group and 391 of 1506 patients (26.0%) in the low-dose group (risk difference [RD], −4.1% [1-sided 97.55% CI, −∞ to −1.1%]; relative risk, 0.84 [1-sided 97.55% CI, −∞ to 0.96; P = .004]). The composite of postoperative seizure, thrombotic events, kidney dysfunction, and death occurred in 265 patients in the high-dose group (17.6%) and 249 patients in the low-dose group (16.8%) (RD, 0.8%; 1-sided 97.55% CI, −∞ to 3.9%; P = .003 for noninferiority). Fourteen of the 15 prespecified secondary end points were not significantly different between groups, including seizure, which occurred in 15 patients (1.0%) in the high-dose group and 6 patients (0.4%) in the low-dose group (RD, 0.6%; 95% CI, −0.0% to 1.2%; P = .05).

Conclusions and Relevance

Among patients who underwent cardiac surgery with cardiopulmonary bypass, high-dose compared with low-dose tranexamic acid infusion resulted in a modest statistically significant reduction in the proportion of patients who received allogeneic red blood cell transfusion and met criteria for noninferiority with respect to a composite primary safety end point consisting of 30-day mortality, seizure, kidney dysfunction, and thrombotic events.

Trial Registration

ClinicalTrials.gov Identifier: NCT03782350

This randomized clinical trial assesses the rate of red blood cell transfusion after start of operation as well as a composite safety outcome with use of high-dose vs low-dose tranexamic acid infusion in patients undergoing cardiac surgery with cardiopulmonary bypass.

Introduction

Since its introduction in 1962,¹ and especially after aprotinin use ceased in 2007, tranexamic acid has been a mainstay antifibrinolytic agent for mitigating bleeding in patients undergoing cardiac surgery. However, some clinical studies have associated high-dose tranexamic acid with seizures or thrombotic complications in such patients.^2,3,4 In a randomized clinical trial (RCT) of tranexamic acid in 4631 patients who underwent coronary artery bypass graft surgery, tranexamic acid reduced bleeding and blood transfusion without increasing thrombotic complications or death up to 1 year postoperatively,^5,6 but single doses of 100 mg/kg and 50 mg/kg were associated with postoperative seizures and consequent stroke and death.⁵

A previous study showed that 1 dose of tranexamic acid may not be sufficiently effective in patients undergoing prolonged cardiac surgery.⁷ In contrast, continuous tranexamic acid infusion reportedly produces a more stable antifibrinolytic plasma concentration and lower peak plasma levels than single doses,⁷ potentially indicating both better antifibrinolytic efficacy and lower risk of adverse effects. Although the delivery method of tranexamic acid for high- and low-dose regimens during cardiac surgery with cardiopulmonary bypass was established in the 1990s and 2000s by several pharmacokinetics studies,^8,9,10 the optimal tranexamic acid infusion dose remains controversial because of limited RCTs studying the effects on transfusion rate and volume, postoperative blood loss, and risk of thrombotic events and seizures.^{7,11,12,13,14} These studies were underpowered to assess either the efficacy or the safety of high-dose tranexamic acid infusion. Therefore, this multicenter, double-blind RCT with 1 year of follow-up—the Outcome Impact of Different Tranexamic Acid Regimens in Cardiac Surgery With Cardiopulmonary Bypass (OPTIMAL) trial¹⁵—was designed to compare high and low doses of tranexamic acid for continuous infusion in patients undergoing cardiac surgery.

Methods

The Fuwai Hospital Institutional Review Board (IRB)/Independent Ethics Committee (IEC) approved the study. All 4 participating sites in China (Fuwai Hospital, Beijing; First Affiliated Hospital of Wenzhou Medical University, Wenzhou; Fuwai Central China Cardiovascular Hospital, Zhengzhou; and Yunnan Fuwai Cardiovascular Hospital, Kunming) accepted the central ethics approval or obtained approval from their local IRB/IEC. Written informed consent was obtained from all study participants before surgery. The study rationale and design were published previously.¹⁵ The initial and final study protocol and the detailed statistical analysis plan are available in Supplement 1.

Study Design and Objective

This trial was a multicenter, randomized, double-blind trial of 2 tranexamic acid dose regimens in adult patients undergoing cardiac surgery with cardiopulmonary bypass. The objective of the trial was to compare the efficacy and safety of these regimens. The study hypothesis was that compared with the low-dose tranexamic acid regimen, the high-dose tranexamic acid regimen would have superior efficacy and noninferior safety.

Study Population

Patient inclusion criteria were age 18 to 70 years, awaiting elective cardiac surgery with cardiopulmonary bypass, and willingness and ability to give informed consent for participation in the study. Patients could withdraw from the study at any time.

Exclusion criteria were acquired defective chromatic (color) vision, active intravascular coagulation (deep vein thrombosis, pulmonary embolism, arterial thrombosis, or antithrombin III deficiency) or a history of thrombophilia, previous convulsion or seizure, allergy or contraindication to intravenous tranexamic acid or its components, breastfeeding or pregnancy, terminal illness with a life expectancy of less than 3 months, mental or legal disability, and current enrollment in another perioperative interventional study.

Randomization

After stratification by center, a web-based registry and randomization system was used to randomly allocate participants in a 1:1 ratio to a high-dose tranexamic acid group or a low-dose tranexamic acid group in permuted blocks of 6. Participants, medical staff, and investigators were blinded to the treatment allocation. The site hospital pharmacy prepared the study medication with different concentrations of tranexamic acid fulfilling the target dose regimen of the 2 groups in identical 50-mL syringes labeled with “bolus,” “maintenance,” or “prime,” as well as the randomization number.

Study Intervention

The high-dose tranexamic acid group received an intravenous bolus of 30 mg/kg after anesthesia induction, then a maintenance dosage of 16 mg/kg/h throughout the operation with a pump prime dose of 2 mg/kg. The low-dose tranexamic acid group received an intravenous bolus and maintenance regimen of 10 mg/kg and 2 mg/kg/h, with a pump prime dose of 1 mg/kg. For each dosing group, after anesthesia induction, the bolus was infused via a central venous line at 150 mL/h over 20 minutes. Then the maintenance dose was infused at 30 mL/h throughout the operation; the pump prime dose was delivered to the cardiopulmonary bypass circuit. All doses were administered with a 50-mL syringe.

Regarding the selection of the study dosages in this trial, we intended to choose the minimum effective dose and the maximal dose proposed by previous in vitro and in vivo studies.^{7,8,9,16,17,18,19} The maximal effective dose of tranexamic acid (a 30-mg/kg loading dose and a 16-mg/kg/h infusion) in cardiac surgery has been proposed by Dowd et al⁸ and has been used in at least 3 RCTs.^7,16,17 Grassin-Delyle et al⁹ reported that using this dosing regimen produced a plasma concentration of 114 μg/mL to 209 μg/mL throughout cardiac operations; this is well above the 100-μg/mL concentration that is known to fully inhibit fibrinolytic activity in vitro. The low dosing regimen (10-mg/kg loading dose plus 1-mg/kg/h maintenance dose) did not appear to maintain the tranexamic acid plasma concentration of 20 μg/mL or higher throughout operation in all patients.^8,9,18,20 Nuttal et al¹⁸ suggested a modified dosing regimen comprising a 10-mg/kg loading dose and a 2-mg/kg/h infusion for the minimum effective dose of tranexamic acid. In our trial, we therefore chose this dose as the low-dose regimen.

Study End Points

The primary efficacy end point was the proportion of patients who received any allogeneic red blood cell transfusion between the start of operation and discharge. Allogeneic red blood cells were transfused if a patient’s hemoglobin level was less than 7 g/dL during cardiopulmonary bypass or less than 8 g/dL after cardiopulmonary bypass, after reinfusion of all residual blood in the circuit and of washed red blood cells by the cell saver.

The primary safety end point was the 30-day rate of the composite of postoperative seizure, kidney dysfunction (stage 2 or 3 Kidney Disease: Improving Global Outcomes [KDIGO] criteria), thrombotic events (myocardial infarction, ischemic stroke, deep vein thrombosis, and pulmonary embolism), and all-cause mortality. Each event has been previously defined.¹⁵

The secondary end points included the volume of allogeneic red blood cell or non–red blood cell transfusion after start of operation, non–red blood cell transfusion rate and volumes after start of operation, postoperative bleeding volume, incidence of reoperation, mechanical ventilation duration, intensive care unit (ICU) length of stay, hospital length of stay, and each component of the primary safety end point. The 6-month and 1-year follow-up data for the primary safety end point are not reported in this article. Except for each component of the primary safety end point, all secondary end points were assessed from the start of surgery to discharge.

As a tertiary end point, all patients’ serum D-dimer levels were measured at 3 time points: preoperatively on the day of surgery, 6 hours postoperatively, and 1 day postoperatively. Other tertiary end points included duration of chest closure, volume of cell saver, duration of chest drainage, hemoglobin levels (the first test after cardiopulmonary bypass and at ICU entry), KDIGO classification, intra-aortic balloon counterpulsation, and extracorporeal membrane oxygenation. Repeat cardiopulmonary bypass, length of surgery, sudden ventricular fibrillation, and total dose and dosing duration of tranexamic acid were post hoc end points.

Procedures and Quality Control

Study quality control, data collection, and follow-up data-collection methods were previously described¹⁵ and are detailed in the eAppendix in Supplement 2. Treatment of study participants began in December 2018. In an interim analysis in December 2019, the study’s data monitoring committee recommended that the study proceed without modification.

Sample Size and Power

The sample size calculation was based on both primary end points. For the primary efficacy end point, a sample size of 1320 participants (660 for each group) was estimated to provide 90% power to detect a 7.4% absolute risk reduction and a 28.9% relative risk reduction according to the 2018 registry data from Fuwai Hospital,²¹ with a red blood cell transfusion rate of 25.5% using a low-dose tranexamic acid regimen. The 28.9% relative risk reduction was based on the pilot clinical study before this trial at Fuwai Hospital. For the primary safety end point, the projected 30-day incidence was 20% in the low-dose group.^5,22 With an absolute noninferiority margin of 5%, a 1-sided α level of .0245, and a 10% dropout rate, a sample size of 3008 participants (1504 for each group) was estimated to provide 90% power to verify the high-dose regimen’s noninferiority to the low-dose regimen. The α level was determined in accordance with the α spending function by the Lan-DeMets method to keep the nominal significance level less than 0.025 (1-sided) for both end points.

The noninferiority margin of 5% was chosen based on 2 tranexamic acid–related large clinical trials. One trial found that the additive adverse end points in high-dose tranexamic acid (30-mg/kg bolus, 16-mg/kg/h maintenance, 2-mg/kg prime) was 25.3%.¹⁷ The other trial found a nonsignificant difference (5.4%) in a composite adverse event rate between the high-dose tranexamic acid (100 mg/kg) group (20.3%) and low-dose tranexamic acid (50 mg/kg) group (14.9%).⁵ Moreover, a 5% margin for composite adverse events appears to be acceptable clinically in cardiac surgery based on the opinion of the investigators.

Statistical Analysis

Demographic data and baseline characteristics were summarized as means and standard deviations or medians and interquartile ranges for continuous variables and as numbers and percentages for categorical variables. The statistical analysis was based on the full analysis set. For the primary efficacy end point, we used the full analysis set for analysis, which included all consenting patients who were randomized to the high- or low-dose group, underwent cardiac surgery with cardiopulmonary bypass, and completed the tranexamic acid intervention without protocol violation. For the primary safety end point, we used completed cases with 30-day follow-up data for analysis in the full analysis set. Multiple imputation by the Markov chain Monte Carlo method with Jeffreys noninformative prior distribution was performed as a sensitivity analysis for the primary end points for any missing data. The χ² test was used to compare the primary efficacy and safety end points between the 2 groups. The t test (continuous variables) and the χ² test (categorical variables) were used for the secondary end points. After 30-day postoperative data were obtained from approximately half of the participants, an interim analysis was performed to evaluate the efficacy and safety of the tranexamic acid interventions. For both primary end points, the nominal significance level in the final analysis was P < .0245 (1-sided) after considering the α spent in the interim analysis. Post hoc study site adjustments with stratified analyses were conducted as sensitivity analyses for the primary efficacy and safety end points. A 2-sided P < .05 was considered statistically significant for all other analyses. Given the potential for type I error due to multiple comparisons, findings for secondary end points and analyses should be considered exploratory. All tests were performed with SAS software, version 9.4 (SAS Institute Inc).

Post Hoc Subgroup Analysis

To investigate the consistency of the primary efficacy of high-dose vs low-dose tranexamic acid across clinically important subgroups, post hoc subgroup analyses were performed for the primary efficacy end point in subgroups by age, sex, body weight, chronic kidney insufficiency, previous cardiac surgery, surgery type, lowest rectal temperature, baseline hemoglobin level, and baseline high-sensitivity C-reactive protein level. Post hoc subgroup analyses compared the primary safety end point across subgroups by age, sex, diabetes, previous stroke, previous myocardial infarction, acute coronary syndrome, left ventricular ejection fraction, urgent surgery, previous cardiac surgery, and surgery type. We also explored the interference of total or subtotal aortic arch replacement with high-dose tranexamic acid in chest tube output. In addition, the influence of both high- and low-dose tranexamic acid on seizures was studied in non–open-chamber and open-chamber cardiac surgery. A previous study revealed an association of seizures with total tranexamic acid exposure greater than 2 g.²³ The incidence of seizures in patients with a total tranexamic acid exposure greater than 2 g and less than or equal to 2 g in the low-dose tranexamic acid group was therefore examined in the current study.

Subgroup analyses were performed by using interaction terms in regression models, and a 2-sided P < .05 without adjustment for multiple comparisons was used to determine statistical significance.

Results

Study Participants

From December 2018 to April 2021, 9056 patients aged 18 to 70 years who underwent cardiac surgery at 4 large academic cardiac centers in different regions of China were assessed for eligibility. Among 3079 patients who were randomized to a treatment group (mean age, 52.8 years; 38.1% women), 3031 (98.4%) completed the trial; 1525 received the high-dose tranexamic acid regimen and 1506 received the low-dose regimen (Figure 1). Forty-eight patients (23 in the high-dose group and 25 in the low-dose group) were lost to follow-up and were therefore excluded from the primary safety analysis. The 2 groups had no notable differences in baseline characteristics (eg, demographics, clinical history, medications, recent laboratory values) and surgical characteristics (eg, surgery type, urgent case, reoperation, length of surgery, duration of cardiopulmonary bypass, cross-clamp time) (Table 1).

Figure 1. Participant Flow in the Outcome Impact of Different Tranexamic Acid Regimens in Cardiac Surgery With Cardiopulmonary Bypass (OPTIMAL) Trial.

^aProtocol violation: tranexamic acid was not administrated in accordance with the bolus, maintenance, and prime dosing protocol.

Table 1. Demographic and Baseline Characteristics.

Characteristics	High-dose tranexamic acid (n = 1525)	Low-dose tranexamic acid (n = 1506)
Age, mean (SD), y	52.9 (12.3)	52.7 (11.9)
Sex, No. (%)
Female	573 (37.6)	582 (38.6)
Male	952 (62.4)	924 (61.4)
Height, mean (SD), cm	167.0 (8.4)	166.6 (8.2)
Body weight, mean (SD), kg	68.1 (13.1)	67.9 (12.6)
Body mass index, mean (SD)a	24.3 (3.7)	24.4 (3.5)
Clinical history, No. (%)
Hypertension	530 (35.9)	511 (34.9)
Smoking	519 (34.1)	508 (33.7)
Smoking within 1 mo	216 (14.2)	228 (15.1)
Hyperlipidemia	489 (32.7)	517 (34.7)
Diabetes	167 (11.0)	166 (11.0)
Previous cardiac surgery	89 (5.8)	92 (6.1)
Lacunar infarction	54 (3.6)	37 (2.5)
Stroke	25 (1.7)	32 (2.2)
Peripheral vascular disease	16 (1.1)	20 (1.4)
Endocarditis	15 (1.0)	16 (1.1)
Carotid artery stenosis ≥80%	12 (0.8)	14 (1.0)
Carotid artery surgery	7 (0.5)	13 (0.9)
Chronic kidney dysfunction	4 (0.3)	7 (0.5)
Chronic obstructive pulmonary disease	3 (0.2)	5 (0.3)
Medication, No. (%)
β-blockers	628 (41.9)	605 (40.7)
Anticoagulants	194 (12.7)	198 (13.2)
Warfarin within 3 d before surgery	2 (0.1)	3 (0.2)
Statins	141 (9.4)	149 (10.0)
Angiotensin-converting enzyme inhibitors	118 (7.9)	111 (7.4)
Antiplatelet agents	83 (5.4)	90 (6.0)
Aspirin within 3 d before surgery	7 (0.5)	4 (0.3)
Clopidogrel within 5 d before surgery	1 (0.01)	1 (0.01)
New York Heart Association classification, No. (%)b
I (least impaired)	159 (10.5)	171 (11.4)
II	734 (48.3)	705 (47.0)
III	527 (34.7)	523 (34.8)
IV (most impaired)	99 (6.5)	102 (6.8)
Left ventricular ejection fraction, %, No. (%)
≥50	1431 (93.8)	1435 (95.3)
40-49	67 (4.4)	66 (4.4)
30-39	25 (1.6)	15 (1.0)
20-29	1 (0.1)	1 (0.1)
Left ventricular end-diastolic dimension, mean (SD), mm	52.3 (10.1)	52.0 (10.4)
Urgent surgery, No. (%)	28 (1.9)	30 (2.0)
Surgery type, No. (%)
Open-chamber cardiac surgery	1292 (84.7)	1265 (84.0)
Isolated cardiac surgery without major vascular surgery	973 (64.6)	999 (67.1)
Isolated valve surgery	555 (36.8)	575 (38.6)
Isolated CABG surgery	250 (16.6)	256 (17.2)
Isolated repair of congenital heart disease	99 (6.6)	102 (6.9)
Isolated myectomy	43 (2.9)	38 (2.6)
Isolated cardiac tumor resection	26 (1.7)	28 (1.9)
Major vascular surgery	253 (16.8)	201 (13.5)
AAR/Bentall, Wheat, or David procedure	126 (8.4)	98 (6.6)
Hemiarch repair	36 (2.4)	35 (2.4)
Total aortic arch repair	34 (2.3)	27 (1.8)
Combined cardiac surgery without major vascular surgery	282 (18.7)	290 (19.5)
Other concomitant operationsc	164 (10.9)	174 (11.7)
Combined CABG and valve surgery	118 (7.8)	116 (7.8)
Combined major vascular surgery (>1 type) without total aortic arch repair	57 (3.8)	41 (2.8)
Intraoperative variables, median (IQR)
Surgery duration, h	4.2 (3.5-5.3)	4.2 (3.4-5.3)
Cardiopulmonary bypass duration, min	122.0 (91.0-159.0)	120.0 (89.0-160.0)
Cross-clamp time, min	85.1 (61.5-115.0)	83.0 (60.3-117.0)
Duration of chest closure, min	72.0 (60.0-91.2)	73.8 (60.0-94.2)
Lowest rectal temperature in cardiopulmonary bypass, mean (SD), °C	31.7 (2.0)	31.9 (1.8)
Preoperative baseline laboratory values, No. (%)
N-terminal pro–brain natriuretic peptide >300.0 pg/mL	684 (44.9)	641 (42.6)
Hemoglobin <10.0 g/dL	48 (3.1)	54 (3.6)
Platelets <100.0 × 109/L	27 (1.8)	27 (1.8)
Serum creatinine ≥1.5 mg/dL	26 (1.7)	26 (1.7)
Prothrombin time >21.0 s	14 (0.9)	14 (0.9)
D-dimer >0.5 μg/mL	257 (16.9)	242 (16.1)
High-sensitivity C-reactive protein >6.0 ng/mL	234 (15.3)	224 (14.9)

Abbreviations: AAR, ascending aorta replacement; CABG, coronary artery bypass graft.

^a Calculated as weight in kilograms divided by height in meters squared.

^b New York Heart Association classifications: I, no limitation of physical activity; ordinary physical activity does not cause undue breathlessness, fatigue, or palpitations; II, slight limitation of physical activity; comfortable at rest, but ordinary physical activity results in undue breathlessness, fatigue, or palpitations; III, marked limitation of physical activity; comfortable at rest, but ordinary physical activity results in undue breathlessness, fatigue, or palpitations; IV, unable to carry on any physical activity without discomfort; symptoms at rest can be present; if any physical activity is undertaken, discomfort is increased.

^c Other concomitant operations: CABG surgery or valve surgery with repair of congenital heart disease and/or cardiac tumor resection.

Primary Efficacy End Point

Between the start of operation and discharge, 724 patients received at least 1 allogeneic red blood cell transfusion: 333 (21.8%) in the high-dose group and 391 (26.0%) in the low-dose group. Compared with patients in the low-dose group, those in the high-dose group had a lower risk of allogeneic red blood cell transfusion, with a risk difference of −4.1% (1-sided 97.55% CI, −∞ to −1.1%; P = .004) and a relative risk of 0.84 (1-sided 97.55% CI, −∞ to 0.96) (Table 2). This effect remained statistically significant in post hoc stratified analysis (risk difference, −4.0%; 97.55% CI, −∞ to −1.0%; P = .005) (Table 2). The multiple imputation method was not used for the primary efficacy end point because there were no missing data in the full analysis set for the primary efficacy end point.

Table 2. Primary and Secondary Outcomes.

Outcomes	High-dose tranexamic acid	Low-dose tranexamic acid	Estimate of difference (95% CI)	P value
Full analysis set, No.	1525	1506
Primary efficacy end point
Patients with red blood cell transfusion, No. (%)	333 (21.8)	391 (26.0)	−4.1 (−∞ to −1.1)a	.004
Adjusted for study site			−4.0 (−∞ to −1.0)a	.005
Primary safety end point
30-d composite, No./total (%)	265/1502 (17.6)	249/1481 (16.8)	0.8 (−∞ to 3.9)b	.003
Adjusted for study site			0.9 (−∞ to 3.9)b	.004
Safety end-point components, No. (%)
Clinical seizurec	15 (1.0)	6 (0.4)	0.6 (−0.0 to 1.2)	.05
Kidney dysfunctiond	71 (4.7)	71 (4.7)	−0.1 (−1.6 to 1.5)	.94
Myocardial infarctione	172 (11.3)	167 (11.1)	0.2 (−2.1 to 2.5)	.87
Strokef	10 (0.7)	8 (0.5)	0.1 (−0.5 to 0.7)	.66
Pulmonary embolismg	1 (0.1)	0	0.1 (−0.2 to 0.0)	>.99
Deep vein thrombosish	15 (1.0)	12 (0.8)	0.2 (−0.5 to 0.9)	.58
Deathi	9 (0.6)	10 (0.7)	−0.1 (−0.1 to 0.01)	.80
Secondary end points
Postoperative chest tube output, median (IQR), mL	490 (320-760)	530 (340-825)	−25 (−50 to 0)	.06
Operations without total aortic arch repair	480 (310-750)	520 (331-800)	−30 (−50 to −10)	.03
Concomitant operations	450 (290-684)	510 (330-770)	−50 (−80 to −20)	.001
Previous cardiac operations	760 (373-1165)	635 (428-1039)	55 (−90 to 220)	.12
Reoperation, No. (%)	16 (1.0)	21 (1.4)	−0.4 (−1.2 to 0.5)	.39
Volume of transfusion, median (IQR)
Red blood cells, mL	0 (0-0)	0 (0-300)	0 (0 to 0)	.01
Fresh frozen plasma, mL	0 (0-0)	0 (0-0)	0 (0 to 0)	.91
Platelets, U	0 (0-0)	0 (0-0)	0 (0 to 0)	.93
Cryoprecipitate, U	0 (0-0)	0.0 (0-0)	0 (0 to 0)	.73
Patients undergoing transfusion, No. (%)
Fresh frozen plasma	225 (14.8)	228 (15.2)	−0.4 (−2.9 to 2.2)	.75
Platelets	147 (9.6)	145 (9.6)	0.0 (−2.1 to 2.1)	>.99
Cryoprecipitate	33 (2.2)	33 (2.2)	0.0 (−1.1 to 1.0)	.96
Any (including red blood cells)	474 (31.0)	515 (34.2)	−3.1 (−6.5 to 0.2)	.07
Length of mechanical ventilation, median (IQR), h	15.0 (11.2-19.0)	14.5 (11.0-19.0)	0.0 (−4.0 to 12.0)	.19
Length of ICU stay, median (IQR), d	2.0 (1.0-4.0)	2.0 (1.0-4.0)	0.0 (0.0 to 0.0)	.76
Length of hospital stay, median (IQR), d	7.0 (6.5-10.0)	7.5 (7.0-10.0)	0.0 (0.0 to 0.0)	.25
Additional outcomes
Repeat cardiopulmonary bypass, No. (%)	27 (1.8)	33 (2.2)	−0.4 (−1.4 to 0.6)	.41
Volume of cell saver, median (IQR), mL	200.0 (200.0-220.0)	200.0 (200.0-220.0)	0.0 (0.0 to 0.0)	.76
Duration of chest drainage, median (IQR), d	4.0 (3.0-5.0)	4.0 (3.0-5.0)	0.0 (0.0 to 0.0)	.16
Hemoglobin level, mean (95% CI), g/dL
First sampling after cardiopulmonary bypass	10.3 (10.2-10.3)	10.2 (10.1-10.3)	0.5 (−0.6 to 1.6)	.99
First sampling at ICU entry	11.1 (11.0-11.2)	11.0 (10.9-11.1)	0.7 (−0.6 to 2.0)	.74
Tranexamic acid total dose, mean (95% CI), mg/kg	103.1 (101.8-104.4)	19.9 (19.7-20.0)	83.2 (81.9 to 84.6)	<.001
Tranexamic acid total dose, mean (95% CI), g	7.1 (6.9-7.2)	1.4 (1.3-1.4)	5.7 (5.6 to 5.7)	<.001
Tranexamic acid dosing duration, mean (95% CI), h	4.8 (4.7-4.9)	4.8 (4.7-4.9)	0.0 (−0.1 to 0.1)	.13
KDIGO category, No. (%)j				.37
0	1083 (70.9)	1122 (73.6)	−3.7 (−6.8 to −0.5)
1	372 (24.4)	331 (21.7)	2.4 (−0.6 to 0.1)
2	47 (3.1)	48 (3.2)	−0.1 (−1.4 to 1.2)
3	25 (1.6)	23 (1.5)	0.1 (−0.8 to 1.0)
KDIGO category 1-3	444 (29.1)	402 (26.4)	2.4 (−0. 8 to 5.7)	.10
Continuous kidney replacement therapy, No. (%)	17 (1.1)	15 (1.0)	0.1 (−0.6 to 0.9)	.73
Sudden ventricular fibrillation, No. (%)	6 (0.4)	5 (0.3)	0.1 (−0.4 to 0.5)	.78
Intra-aortic balloon counterpulsation, No. (%)	17 (1.3)	17 (1.3)	−0.0 (−0.8 to 0.8)	.98
Extracorporeal membrane oxygenation, No. (%)	4 (0.3)	5 (0.4)	−0.1 (−0.5 to 0.4)	.75

Abbreviations: ICU, intensive care unit; KDIGO, Kidney Disease: Improving Global Outcomes.

^a One-sided 97.55% CI; superiority test.

^b One-sided 97.55% CI; noninferiority test with 5% margin.

^c Clinical seizure defined as generalized tonic-clonic events, focal attacks, or both. The diagnosis was confirmed by neurology consultation.

^d Kidney dysfunction defined as stage 2 or 3 KDIGO criteria for cardiac surgery–associated acute kidney injury.

^e Myocardial infarction diagnosed in accordance with the third universal definition of myocardial infarction for cardiac surgery from the joint ESC/ACCF/AHA/WHF Task Force.²⁴ The diagnosis is made if cardiac troponin values are greater than 10 times the 99th percentile upper reference limit during the first 48 hours, occurring from a normal baseline value (≤99th percentile upper reference limit), when associated with the appearance of (1) new pathological Q waves or new left bundle branch block, (2) angiographically documented new graft or new native coronary artery occlusion, or (3) imaging evidence of new loss of viable myocardium or new regional wall motion abnormality.

^f Stroke defined as a new focal neurologic deficit lasting more than 24 hours, confirmed by cerebral computed tomography and an attending neurology consultant.

^g Pulmonary embolism defined as thrombosis in the pulmonary vasculature indicated by pulmonary computed tomography angiography.

^h Deep venous thrombosis defined as thrombosis in the deep veins of the lower extremities with leg swelling, which was further confirmed by vascular ultrasonography.

ⁱ Mortality included death due to any cause within 30 days.

^j KDIGO category classification, mainly according to perioperative serum creatinine levels: category 1, ≥1.5 to 1.9 times baseline serum creatinine or >0.3-mg/dL (26.5-μmol/L) increase in serum creatinine within 48 hours; category 2, ≥2.0 to 2.9 times baseline serum creatinine; category 3, ≥3.0 times baseline serum creatinine or increase of serum creatinine to ≥4.0 mg/dL (353.6 μmol/L) or kidney replacement therapy in patients younger than 18 years, decrease of estimated glomerular filtration rate to <35 mL/min/1.73 m².

Primary Safety End Point

Complete 30-day follow-up data were obtained from 2983 of 3031 participants (98.4%). The composite primary safety end point (seizure, kidney dysfunction, thrombotic events, and all-cause mortality) occurred in 265 patients in the high-dose group (17.6%) and 249 patients in the low-dose group (16.8%) (Table 2). The high-dose tranexamic acid regimen was noninferior to the low-dose regimen with regard to the composite outcome (risk difference, 0.8%; 1-sided 97.55% CI, −∞ to 3.9%; P = .003 for noninferiority). This effect remained statistically significant in the post hoc stratified analysis (risk difference, 0.9%; 1-sided 97.55% CI, −∞ to 3.9%; P = .004 for noninferiority) (Table 2). In addition, the post hoc sensitivity analysis using the multiple imputation method for missing data showed a statistically significant result similar to that of the initial completed case analysis (eTable 1 in Supplement 2).

Secondary End Points

The median volume of allogeneic erythrocyte transfusion was 0.0 mL (IQR, 0.0-0.0 mL) and 0.0 mL (IQR, 0.0-300.0 mL) in the high- and low-dose groups, respectively (median difference, 0.0 mL; 95% CI, 0.0-0.0 mL; P = .01) (eFigure in Supplement 2). However, the rates and volumes of fresh frozen plasma, platelets, and cryoprecipitate transfusion were not significantly different between the groups (Table 2). The median total volume of chest drainage after surgery was 490.0 mL (IQR, 320.0-760.0 mL) and 530.0 mL (IQR, 340.0-825.0 mL) in the high- and low-dose groups, respectively (median difference, −25.0 mL; 95% CI, −50.0 to 0.0 mL; P = .06) (Table 2). Sixteen patients in the high-dose group (1.0%) and 21 patients in the low-dose group (1.4%) had reoperation for bleeding (risk difference, −0.4% [95% CI, −1.2% to 0.5%; P = .39]; relative risk, 0.75 [95% CI, 0.39-1.43]).

There was no statistically significant difference between groups in the median length of mechanical ventilation, ICU length of stay, or postoperative hospital length of stay (Table 2). The 30-day risk of myocardial infarction and the incidences of kidney dysfunction, ischemic stroke, pulmonary embolism, and deep vein thrombosis were not significantly different between groups (Table 2). Seizures occurred in 15 patients in the high-dose group (1.0%) and 6 in the low-dose group (0.4%). The high dose of tranexamic acid did not significantly increase the incidence of seizure (risk difference, 0.6% [95% CI, −0.0% to 1.2%]; relative risk, 2.47 [95% CI, 0.96-6.35; P = .05]) (Table 2). eTable 2 in Supplement 2 lists all patients with seizures and their characteristics and outcomes.

Tertiary End Points

Postoperative serum D-dimer levels were statistically significantly lower in the high-dose group than in the low-dose group at 6 hours (median, 0.3 [IQR, 0.1-0.4] μg/mL vs 0.5 [IQR, 0.2-0.8] μg/mL; median difference, −0.2 μg/mL; 95% CI, −0.2 to −0.2 μg/mL; P < .001) and at 1 day (median, 0.5 [IQR, 0.3-1.0] μg/mL vs 0.9 [IQR, 0.6-1.4] μg/mL; median difference, −0.3 μg/mL; 95% CI, −0.3 to −0.3 μg/mL; P < .001).

There was no statistically significant difference between groups in duration of chest closure, cell saver volume, duration of chest drainage, hemoglobin level (the first test after cardiopulmonary bypass and at ICU entry), KDIGO classification, intra-aortic balloon counterpulsation use, or extracorporeal membrane oxygenation use (Table 2).

Post Hoc Analyses and Outcomes

The post hoc outcomes of repeat cardiopulmonary bypass, length of surgery, and sudden ventricular fibrillation showed no significant differences between groups. A post hoc analysis showed that when the 63 patients who underwent total or subtotal aortic arch replacement were not included in the chest tube output analysis, median chest tube output was 480.0 mL (IQR, 310.0-750.0 mL) in the high-dose group vs 520.0 mL (IQR, 331.0-800.0 mL) in the low-dose group (median difference, −30.0 mL; 95% CI, −50.0 to −10.0 mL; P = .03; P = .65 for interaction). The total dose of tranexamic acid in the high-dose group (mean, 103.1 mg/kg [95% CI, 101.8-104.4 mg/kg]) was significantly higher than that in the low-dose group (mean, 19.9 mg/kg [95% CI, 19.7-20.0 mg/kg]; mean difference, 83.2 mg/kg [95% CI, 81.9-84.6 mg/kg]; P < .001). No statistically significant difference was observed between groups in the dosing duration of tranexamic acid (mean difference, 0.0 hours; 95% CI, −0.1 to 0.1 hours; P = .13) (Table 2).

In this trial, the high dose of tranexamic acid exceeded 2 g in all patients. Meanwhile, the majority of patients who received the low dose of tranexamic acid had a total tranexamic acid dose of 2 g or lower (mean, 1.3 g; 95% CI, 1.3-1.3 g). In 71 patients who received the low dose and had a tranexamic acid total dose greater than 2 g (mean, 2.3 g; 95% CI, 2.2-2.4 g), no seizures were observed.

Post Hoc Subgroup Analyses

Regarding the primary efficacy end point, subgroup analyses suggested that high-dose tranexamic acid reduced allogeneic red blood cell transfusion consistently across subgroups except body weight (Figure 2). Regarding the primary safety end point, the subgroup analyses showed no statistically significant interactions between tranexamic acid dose and baseline or perioperative characteristics except previous stroke and urgent surgery (Figure 3). However, this finding should be cautiously interpreted because of the small sample size in these subgroup post hoc analyses.

Figure 2. Post Hoc Subgroup Analysis of the Relative Risk of the Primary Efficacy End Point Among Patients Treated With High-Dose vs Low-Dose Tranexamic Acid.

hs-CRP indicates high-sensitivity C-reactive protein.

Figure 3. Post Hoc Subgroup Analysis of the Relative Risk of the Primary Safety End Point Among Patients Treated With High-Dose vs Low-Dose Tranexamic Acid.

Among the patients who underwent open-chamber cardiac surgery, the incidence of seizures was 1.2% (15/1273) in the high-dose group and 0.4% (5/1245) in the low-dose group (risk difference, 0.8%; 95% CI, 0.1%-1.5%; P = .04). In contrast, among patients who underwent non–open-chamber cardiac surgery, no seizures were observed in the high-dose group (0/229) and 1 seizure was observed in the low-dose group (1/236 [0.4%]) (risk difference, −0.4%; 95% CI, −1.8% to 1.0%; P = .33; P = .16 for interaction).

Discussion

In this trial involving patients who underwent cardiac surgery with cardiopulmonary bypass, high-dose compared with low-dose tranexamic acid infusion resulted in a modest but statistically significant reduction in the proportion of patients who received allogeneic red blood cell transfusion and was noninferior with respect to a composite primary safety end point consisting of 30-day mortality, seizure, kidney dysfunction, and thrombotic events.

This trial extends previous findings by showing that in patients undergoing cardiac surgery, high and low doses of continuously infused tranexamic acid were associated with similar adverse event rates, but the high dose was more efficacious than the low dose with regard to reducing need for red blood cell transfusion, and high-dose tranexamic acid had a superior antifibrinolytic effect as indicated by postoperative plasma D-dimer levels.

Regarding each component of the primary composite safety outcome, there were no significant differences between the high-dose and low-dose tranexamic acid groups, although the point estimate for postoperative seizures was slightly higher in the high-dose tranexamic acid group. In a post hoc analysis, in the non–open-chamber cardiac surgery subgroup, only 1 patient of 465 had clinical seizures, and this patient received low-dose tranexamic acid therapy. In addition, the low incidence of seizures in this subgroup is similar to what was observed in the Aspirin and Tranexamic Acid for Coronary Artery Surgery (ATACAS) trial,⁵ in which only 0.3% of patients receiving tranexamic acid therapy during isolated coronary artery bypass graft surgery had seizures. In contrast, in ATACAS trial patients who underwent open-chamber cardiac surgery, the incidence of seizures was higher in the tranexamic acid group than in the placebo group (2.0 vs 0%). However, no dose-dependent risk of seizures was observed when the total dose was lowered from 100 mg/kg to 50 mg/kg. The mechanism of tranexamic acid’s dose-dependent increase in risk of seizures among patients undergoing open-chamber cardiac surgery is unknown. Interaction of air emboli with tranexamic acid use is suspected.^23,25 Consistent with those findings, in this trial, among patients who underwent open-chamber cardiac surgery, the incidence of seizures was significantly higher in the high-dose tranexamic acid group than in the low-dose group. Although the total high tranexamic acid dose in this trial (a mean of 103.1 mg/kg) was similar to the 100-mg/kg dose used in the ATACAS study, the total low dose in this trial was lower (a mean of 19.9 mg/kg vs 50 mg/kg). This may explain why a dose-dependent risk of seizure in open-chamber cardiac surgery was observed in this trial but not in the ATACAS trial. In addition, this trial used a continuous infusion of tranexamic acid, while the ATACAS trial used a single dose, which might have resulted in different peak plasma tranexamic acid concentrations.⁸ Nevertheless, the risk of seizures in patients who received high-dose tranexamic acid during open-chamber cardiac surgery was still relatively low in this trial (1.2%) and in the ATACAS trial (2%). The decision to use high- vs low-dose tranexamic acid may depend on the risk of operative bleeding in open-chamber vs non–open-chamber cardiac surgery.

For the secondary end points, the continuous infusion of high-dose vs low-dose tranexamic acid reduced the perioperative allogeneic red blood cell transfusion volume. However, there was no significant reduction in frozen plasma, platelet, or cryoprecipitate transfusion with high-dose tranexamic acid. Although there was no significant difference between groups in overall postoperative chest tube output, a subgroup analysis showed that high-dose tranexamic acid significantly reduced chest tube output in patients who underwent combined cardiac surgery, or when patients with total or subtotal aortic arch replacement were excluded from the overall analysis. The complexity of total aortic arch surgery may have masked a statistically significant reduction of overall chest tube output by high-dose tranexamic acid in this trial. Furthermore, there was no statistically significant difference between the 2 dose groups in durations of mechanical ventilation, ICU stay, or hospital stay. Therefore, these results support the notion that high-dose tranexamic acid exerts a dose-dependent allogeneic red blood cell transfusion–sparing effect without prolonging ICU or hospital stay.

Although the ATACAS trial⁵ supports the use of a single dose of tranexamic acid in patients undergoing cardiac surgery, the optimal dose and delivery method have been disputed. The continuous infusion of tranexamic acid has been a mainstay of the practice for cardiac surgery in both Asia and North America. The current study is the only large-scale randomized trial, to our knowledge, to examine the dose-dependent clinical and adverse effects of continuous infusion of tranexamic acid in cardiac surgery patients.

Limitations

This study has several limitations. First, it included patients who underwent total aortic arch repair. These patients had the highest risk of bleeding in the study, and the complexity of the surgery itself may have masked or reduced the dose-dependent effects of tranexamic acid. Second, for practical reasons, only a small portion of the participants had intraoperative electroencephalographic monitoring. Therefore, electroencephalographic monitoring data were not included in the analysis. Nevertheless, most of the previously reported tranexamic acid–associated seizures in clinical trials were identified from clinical signs. Third, the trial is limited to the Chinese population. Caution should be used in applying these findings to other ethnic populations. Fourth, the reduction of the red blood cell transfusion rate by high-dose tranexamic acid in this trial was smaller than the originally expected absolute rate reduction of 7.4% (28.9% relative risk).

Conclusions

Among patients undergoing cardiac surgery with cardiopulmonary bypass, high-dose compared with low-dose tranexamic acid infusion resulted in a modest statistically significant reduction in the proportion of patients who received allogeneic red blood cell transfusion and met criteria for noninferiority with respect to a composite primary safety end point consisting of 30-day mortality, seizure, kidney dysfunction, and thrombotic events.

Supplement 1.

Trial Protocol and Statistical Analysis Plan

Supplement 2.

eAppendix. Supplemental Methods: Procedures and Quality Control

eTable 1. Effect of Missing Value Treatment Methods on the Primary Safety End Point

eFigure. Postoperative RBC Transfusion Volume Among 68 Patients Treated With High-Dose TxA Versus Patients Treated With Low-Dose TxA

eTable 2. Characteristics and Outcomes for Patients With 30-Day Seizure

Supplement 3.

Nonauthor Collaborators. OPTIMAL Study Group

Supplement 4.

Data Sharing Statement