Blood Products and Recombinant Factor VIIa

Changing from Aprotinin to Tranexamic Acid Results in Increased Use of Blood Products and Recombinant Factor VIIa for Aortic Surgery Requiring Hypothermic Arrest Roman M. Sniecinski, MD,* Edward P. Chen, MD,† Sunal S. Makadia, BS,* Mutsuhito Kikura, MD,‡ Daniel Bolliger, MD,* and Kenichi A. Tanaka, MD, MSc* Objective: Aprotinin, once used to reduce allogeneic blood product transfusion during cardiac surgery, was withdrawn from the market in late 2007 over concerns of causing increased mortality. This study was undertaken to determine what, if any, the impact of changing antifibrinolytic agents (from aprotinin to tranexamic acid) for deep hypothermic circulatory arrest cases would have on blood bank resource utilization.

Design: This a retrospective review.

Setting: All cases were performed at a single university hospital.

Participants: All patients underwent cardiac surgical procedures requiring deep hypothermic circulatory arrest performed by a single cardiac surgeon between January 2006 and November 2008.

Intervention: All patients prior to November 15, 2007 received aprotinin as antifibrinolytic therapy, while those after that date received tranexamic acid for antifibrinolytic therapy.

D

EEP HYPOTHERMIC CIRCULATORY ARREST (DHCA) commonly is used for operations involving the thoracic aorta that require interruption of cerebral blood flow.1,2 While the technique provides vital neuroprotection and allows for a bloodless surgical field,3 it places considerable stress on the patient’s coagulation system.4 Extensive surgical dissection, drastic temperature changes, periods of blood stasis, and severe hemodilution due to prolonged use of cardiopulmonary bypass (CPB) result in severe postoperative coagulopathy. Antifibrinolytic agents, mainly aprotinin and the lysine analogs, ␧-aminocaproic acid and tranexamic acid (TXA), frequently are used to reduce systemic activation of fibrinolytic enzymes, and a premature breakdown of hemostatic clot.5,6 When adequate anticoagulation with heparin was ensured, aprotinin had been shown to be safe and effective at reducing the need for allogeneic blood products following cardiac surgery requiring DHCA.4,7 Prior to November 15, 2007, the use of aprotinin was part of the standard institutional practice in high-risk cardiac surgery, including all DHCA cases. Aprotinin was voluntarily withdrawn from the market following the release of clinical data from the Canadian BART (Blood conservation using Antifibrinolytics in a Randomized Trial) study that demonstrated a trend of increased 30-day morbidity and mortality with aprotinin relative to the lysine analogs.8 As a result, the institution began using TXA as antifibrinolytic therapy for all cardiac surgery involving CPB. Several months following the change, it was believed that the use of blood products for cardiac surgery requiring DHCA had significantly increased. The authors hypothesized that TXA would be associated with increased hemostatic product usage and transfusion cost; thus, the authors conducted a retrospective review of DHCA cases receiving aprotinin or TXA. The goal was also to help quantify the impact, if any, of changing antifibrinolytic therapies for DHCA cases.

Measurements and Main Results: Blood transfusion data and recombinant factor VIIa use during the pre- and immediate postoperative period was collected for all patients during the study time period. There were no significant differences between the aprotinin (n ⴝ 82) and tranexamic acid (n ⴝ 78) groups with regard to baseline coagulation status or operative characteristics. Patients treated with tranexamic acid required more fresh frozen plasma (2.5 units, p < 0.001), platelets (0.5 units, p < 0.01), and cryoprecipitate (25 units, p < 0.001), and had a higher incidence of recombinant factor VIIa use (34.6% v 12.2%, p < 0.01) compared with patients in the aprotinin group.

Conclusions: Patients treated with tranexamic acid required more clotting factors than the control group receiving aprotinin. © 2010 Elsevier Inc. All rights reserved.

KEY WORDS: aortic surgery, blood conservation, hypothermia/circulatory arrest, antifibrinolytics METHODS Following IRB approval, the authors conducted a retrospective chart review for all DHCA cases performed by a single cardiac surgeon (E.P.C.) from January 2006 through November 2008. All reviewed charts involved aortic surgery (ascending, descending, and/or arch) both with and without combined coronary or valve procedures. Collected data were the patient demographics (age, sex, weight), operative characteristics (reoperation, emergency, CPB time, and DHCA time), pre- and postoperative (1st lab values drawn in the ICU) hematology test results: hematocrit (%), platelet count (⫻103/␮L), fibrinogen level (mg/dL), international normalized ratio (INR), and activated partial thromboplastin time (aPTT [s]). Additionally, use of blood products, packed red blood cells (RBCs), platelets, fresh frozen plasma (FFP), cryoprecipitate, and recombinant factor VIIa (mg) in the OR and during the first 24 hours in the ICU were recorded. Clinically important events such as the need for re-exploration, acute (new-onset) renal failure requiring dialysis, stroke, seizure, and death also were noted.

The institutional CPB anticoagulation protocol (400 units/kg of heparin initial dose, with additional doses to maintain kaolin ACT ⬎ 480 s) was used in all cases and did not change during the study period.

For antifibrinolytic therapy, patients prior to November 15, 2007 were administered full-dose aprotinin: 50,000 kallikrein inhibitory units (KIU) test dose after sternotomy (all were negative), 2 ⫻ 106 KIU bolus, followed by 0.5⫻ 106 KIU/h infusion until the end of surgery.

Following the withdrawal of aprotinin, tranexamic acid was adminis-

From the *Departments of Anesthesiology and †Surgery (cardiothoracic), Emory University School of Medicine, Atlanta, GA and ‡Hamamatsu Medical Center, Hamamatsu, Shizuoka, Japan.

Supported in part by Emory University Department of Anesthesiology, and the Foundation for Anesthesia Education and Research (S.S.M.).

Address reprint requests to Roman Sniecinski, MD, Department of Anesthesiology, Emory University Hospital, 1364 Clifton Road, NE, Atlanta, GA 30322. E-mail: roman.sniecinski@emory.edu © 2010 Elsevier Inc. All rights reserved.

1053-0770/2406-0009$36.00/0 doi:10.1053/j.jvca.2010.02.018

Journal of Cardiothoracic and Vascular Anesthesia, Vol 24, No 6 (December), 2010: pp 959-963

959

960

tered at the dose of a 2-g bolus followed by a 0.5-g/h infusion until the end of surgery.

The surgical conduct of DHCA for all cases included right axillary arterial cannulation and either atrial or femoral cannulation for venous return. Following institution of full CPB, patients were aggressively cooled for 25 minutes and reached nadir temperatures of 18-20°C (nasopharyngeal). Sodium pentothal (250-500 mg) was administered prior to circulatory arrest to obtain an isoelectric EEG pattern. Antegrade cerebral perfusion was used during the DHCA time period at a pressure of 60-70 mmHg. Intraoperative red-cell salvage (“cell saver”) use was standard in all cases. RBCs were transfused to maintain hematocrit above 21% while in the OR. Crystalloid salt solutions and albumin, but not hetastarch derivatives, also were used for volume expansion when necessary.

At the conclusion of CPB, 250 mg of protamine were administered for heparin neutralization. If the ACT remained above the baseline value or if heparinized blood from the CPB machine was administered, an additional 25-50 mg of protamine were given. If bleeding continued at 200 mL/h without any obvious surgical cause, hemostatic blood products (FFP, platelets, and cryoprecipitate) were administered at the discretion of the attending anesthesiologist. Following consultation with an attending hematologist, the use of recombinant activated factor VII (NovoSeven, Bagsbaerd, NovoNordisk) was permitted at the dose of 2.4 mg-4.8 mg when microvascular bleeding continued despite transfusion of 4 units of FFP, 2 plateletpheresis units, and 20 units of cryoprecipitate. Postoperative care was directed by a separate intensive care staff who administered blood products according to abnormal laboratory results (platelet count ⬍100 ⫻ 103/␮L, INR ⬎1.5, and/or fibrinogen ⬍150 mg/dL) in the presence of clinical bleeding (⬎200 mL/h blood loss). Transfusion of RBCs in the ICU was at the discretion of the attending intensivist and tailored to the clinical situation.

The primary endpoint was the usage of blood products including RBC, FFP, platelets, and cryoprecipitate. Secondary endpoints were clinical events including rFVIIa use, re-exploration rate, renal failure, stroke, seizure, in-hospital death, and total hospital stay in days. Based on the usage of approximately 2 units of FFP (520 ⫾ 360 mL) in the operating room in the aprotinin-treated DHCA cases, group sample sizes of 82 (aprotinin) and 78 (TXA) would achieve 80% power to detect a difference of ⫺156 (mL) between the null hypothesis that both group means are 520 (mL) and the alternative hypothesis that the mean of group 2 is 676 (mL) with estimated group standard deviations of 360 (mL) and 360 (mL) and with a significance level (␣) of 0.05 using a 2-sided Mann-Whitney test assuming that the actual distribution is logistic.

Potential predictors of bleeding in the entire cohort (n ⫽ 160) were examined using multiple linear regression analysis. The influence of age, gender, body weight, type of antifibrinolytic, CPB time, DHCA time, preoperative hematocrit, INR, and platelet count were evaluated on the 24-hour transfusion (mL) of RBC, FFP, platelets, and cryoprecipitate. The results of the final reduced model using best fit were determined by the R2 value.

Statistical analysis comparing aprotinin-treated and TXA-treated patients initially was carried out using t-test for continuous variables, and using Fisher’s exact test for categoric valuables. Effects of different antifibrinolytic therapies on potential increases in blood product usage were assessed using a multiple logistic regression. The increased blood product usage was defined as follows: RBC ⬎ 5 units, FFP ⬎ 5 units, plateletpheresis ⬎ 3 units, cryoprecipitate ⬎ 15 units, or rFVIIa administration. All analyses were controlled for baseline covariates (age, sex, weight, hematocrit, platelet count, INR) and differences in operative procedure (cell-saver usage, CPB time, circulatory arrest time) among patients. Using a forward stepwise procedure, weight, platelet count, cell saver (in units), and CPB time were included in the final model as significant covariates, and the odds ratio was calculated for

SNIECINSKI ET AL

Table 1. Demographic and Surgical Data

Male sex Age Weight Laboratory data Hematocrit (%) Platelet (⫻103/␮L) INR PTT (s) ACT (s) Serum creatinine Surgical site Aortic root Arch involvement DTAA Redo surgery Emergency CPB time (min) DHCA time (min)

Aprotinin (n ⫽ 82)

TXA (n ⫽ 78)

p Value

55 (67.1%) 57.9 ⫾ 14.1 87.9 ⫾ 22.0

50 (64.1%) 56.9 ⫾ 14.4 85.8 ⫾ 22.2

0.74 0.67 0.54

38.4 ⫾ 5.2 227 ⫾ 69.9 1.1 ⫾ 0.17 39.2 ⫾ 26.8 130 ⫾ 19.1 1.13 ⫾ 0.28

37.3 ⫾ 5.3 225 ⫾ 62.1 1.0 ⫾ 0.13 32.4 ⫾ 3.8 136 ⫾ 18.6 1.11 ⫾ 0.30

0.21 0.85 ⬍0.05 0.07 0.06 0.74 0.16

26 (31.7%) 50 (61.0%) 6 (7.3%) 21 (25.6%) 7 (8.5%) 198 ⫾ 62.7 28.4 ⫾ 13.9

20 (25.6%) 48 (61.5%) 10 (12.8%) 21 (26.9%) 10 (12.8%) 195 ⫾ 59.1 35.9 ⫾ 15.9

0.86 0.45 0.72 ⬍0.01

NOTE. Surgical site analysis done for different categories as a whole.

Abbreviations: DTAA, descending thoracoabdominal aneurysm (ie, descending aorta only); TXA, tranexamic acid; INR, international normalized ratio; PTT, partial thromboplastin time; ACT, activated coagulation time; CPB, cardiopulmonary bypass; DHCA, deep hypothermic circulatory arrest.

each product if applicable. All statistical analyses were performed using SPSS 11.5 (SPSS Inc., Chicago, IL); a p value of ⬍0.05 was considered significant.

RESULTS

The demographic and clinical data from 160 consecutive patients were included (Table 1). There was no statistically significant difference between aprotinin and TXA groups in the demographic data. Preoperative laboratory and surgical data indicated statistically significant differences in preoperative PT and DHCA time. The difference in PT seemed clinically insignificant, but DHCA time was approximately 7 minutes longer in the TXA group. These 2 parameters and other potential covariates were controlled in a multivariate logistic regression model as described below.

There was no significant influence of age, sex, body weight, CPB/circulatory arrest durations, hematocrit, and INR on the amount of transfusion in the entire cohort (n ⫽ 160). The final model of multiple linear regression presented as the best-fit model demonstrated that only the type of antifibrinolytic and preoperative platelet count were significantly associated with the amount of transfused RBC, FFP, platelets, and cryoprecipitate (Table 2).

Blood-product usage was significantly increased for RBC, FFP, plateletpheresis, cryoprecipitate, and rFVIIa (Table 3) in the TXA group. In particular, the use of FFP and cryoprecipitate was increased for both intraoperative and postoperative periods. For RBC and plateletpheresis, the intraoperative usage was not different between the 2 groups, but the postoperative usage was significantly higher in the TXA group. The use of rFVIIa was significantly increased in the TXA group compared

CHANGING FROM APROTININ TO TRANEXAMIC ACID

961

Table 2. Best-Fit Model for Potential Predictors of Blood Product Use for Entire Cohort

R2

Final model, Weight CPB time DHCA time Preop INR Antifibrinolytic Preop platelet count

RBC

FFP

Platelets

Cryoprecipitate

0.194 0.541 0.054 0.693 0.344 0.0001 0.040

0.182 0.132 0.261 0.216 0.323 0.003 0.004

0.119 0.869 0.323 0.717 0.944 0.022 0.006

0.202 0.450 0.649 0.271 0.335 0.0001 0.016

with the aprotinin group (34.6% v 12.2%, p ⬍ 0.003). The amount of blood processed by the cell salvage was not statistically different between the groups (8.6 v 9.6 units, p ⫽ 0.23).

The hematologic data drawn immediately postoperatively in the ICU (Table 4) were similar between the 2 groups, although platelet count was statistically lower in the TXA group. The rate of re-exploration was significantly higher in the TXA group (Table 4). There was also a trend for more seizures in the TXA group, but it did not reach statistical significance. Other clinically pertinent parameters including renal insufficiency, stroke, in-hospital death, and length of hospital stay were the same between the 2 groups.

In the multivariate logistic regression analysis to evaluate increased blood-product usage (Table 5), weight, platelet count, cell saver (in units), and CPB time were included as significant covariates as described in the methods. Relative to aprotinin treatment, the use of TXA did not increase the odds ratio for RBC transfusion. However, TXA was associated with 4-fold increases in FFP and plateletpheresis transfusion, and 50-fold increase in cryoprecipitate transfusion. The use of rFVIIa also

Table 3. Transfusion Requirements–Total and Breakdown Between OR and ICU

RBC (mL) Total Intraoperative Postoperative FFP (mL) Total Intraoperative Postoperative Platelet (mL) Total Intraoperative Postoperative Cryo (mL) Total Intraoperative Postoperative Cell saver (U) Use of rFVIIa

Aprotinin (n ⫽ 82)

TXA (n ⫽ 78)

p Value

2651 (2153-3034) 1410 (1157-1662) 1144 (799-1490)

3558 (3056-4059) 1217 (976-1458) 2340 (1986-2695)

⬍0.05 0.21 ⬍0.001

790 (620-959) 521 (442-601) 268 (139-397)

1424 (1251-1599) 819 (723-915) 605 (474-738)

⬍0.001 ⬍0.001 ⬍0.001

498 (426-570) 368 (327-410) 134 (86.2-183)

693 (609-778) 392 (351-434) 301 (212-390)

⬍0.01 0.14 ⬍0.01

336 (283-389) 251 (202-299) 86.3 (23.1-150) 8.6 (7.6-9.7) 10 (12.2%)

748 (637-860) 446 (396-495) 303 (215-390) 9.6 (8.5-10.6) 27 (34.6%)

⬍0.001 ⬍0.001 ⬍0.001 0.23 ⬍0.01

NOTE. pRBC⬃350 mL/unit, FFP⬃250 mL/unit, plateletpheresis⬃350 mL/unit, cryoprecipitate⬃15 mL/unit Abbreviations: RBC, red blood cells; FFP, fresh frozen plasma; Cryo, cryoprecipitate; rFVIIa, recombinant factor VIIa.

Table 4. Immediate Postoperative Laboratory Data and Postoperative Events Aprotinin (n ⫽ 82)

Laboratory data ACT after protamine (s) Hematocrit (%) Platelet (⫻103/␮L) INR aPTT (s) Fibrinogen (mg/dL) Postoperative events Re-exploration Renal failure Stroke Seizure In-hospital death Hospital stay (days)

134 31.2 103 1.52 64.7 253

⫾ ⫾ ⫾ ⫾ ⫾ ⫾

24.5 3.9 30.9 0.31 22.2 80

2 (2.4%) 7 (8.5%) 2 (2.4%) 0 (0%) 6 (7.3%) 13.7 ⫾ 10.7

TXA(n ⫽ 78)

p Value

⫾ ⫾ ⫾ ⫾ ⫾ ⫾

0.80 0.74 ⬍0.05 0.09 ⬍0.05 0.40

133 31.0 90.5 1.40 54.5 271

18.5 4.2 29.9 0.40 22.8 107

9 (11.5%) 6 (7.7%) 2 (2.6%) 5 (6.4%) 4 (5.1%) 13.1 ⫾ 8.0

⬍0.05 1.00 1.00 0.02 0.51 0.75

was increased by 10-fold in TXA therapy compared with aprotinin therapy.

The drug cost of TXA was about $1,000 less than that of aprotinin (Table 6). However, the increased use of blood products costs approximately $3,000 more in TXA treatment than in aprotinin treatment. This increase does not include blood bank technicians’ time for preparations, and the additional use of rFVIIa. DISCUSSION

The use of antifibrinolytics to reduce postoperative bleeding following cardiac surgery has become routine at most highvolume centers.5 For high-risk cases (combined valve/CABG operations, redo sternotomies, and aortic procedures requiring DHCA) at the institution, aprotinin was used almost exclusively. This practice abruptly ended on November 15, 2007, when the drug was withdrawn from the U.S. market following concerns over increased morbidity/mortality compared with the lysine analogs.8 The institution then switched to tranexamic acid as the sole antifibrinolytic agent in high-risk cases. The authors were particularly interested in the impact this change would have on aortic cases requiring DHCA, which commonly result in severe coagulopathy. This study was undertaken to determine what effect, if any, the change in agents would have on the resource utilization. The results of this study suggested that switching from aprotinin to TXA for DHCA cases has contributed to an increased need for blood-product transfusion.

The authors have seen an increase of approximately 3 units of RBCs (915 mL), 2.5 units of FFP (599 mL), 0.5 plateletpheresis (179 mL), and 25 units (399 mL) of cryoprecipitate on a per-case basis. The higher transfusion requirements occurred both intraoperatively (for FFP and cryoprecipitate), and post-

Table 5. Increased Transfusion with TXA Relative to Aprotinin

FFP Plateletpheresis Cryoprecipitate rFVIIa

Odds Ratio (95% CI)

p Value

4.4 (1.9-10.6) 3.9 (1.9-10.6) 52.8 (9.8-1000) 10.5 (3.5-39.0)

⬍0.01 ⬍0.01 ⬍0.001 ⬍0.001

962

SNIECINSKI ET AL

Table 6. Average Cost of Antifibrinolytic Therapy and Transfusion Aprotinin (n ⫽ 82)

Drug cost Transfusion RBC FFP Plateletpheresis Cryoprecipitate

TXA (n ⫽78)

p Value

$1,268 ⫾ 90.8

$293 ⫾ 32.9

⬍0.001

$2,651 ⫾ 1710 $281 ⫾ 234 $1,181 ⫾ 794 $1,868 ⫾ 1342

$3,558 ⫾ 2676 $506 ⫾ 315 $1,662 ⫾ 1019 $4,130 ⫾ 2728

0.012 ⬍0.001 ⬍0.001 ⬍0.001

operatively (for all hemostatic blood products). There were no differences in length of hospital stay, renal failure, stroke, or death between the 2 groups.

The finding that aprotinin reduced blood transfusion compared with TXA is consistent with other studies comparing the 2 antifibrinolytics. The investigators for the BART study reported decreased postoperative bleeding and massive transfusion in agreement with this data, but at the cost of increased 30-day mortality. Indeed, the re-exploration rates for bleeding were 5.5% for aprotinin, 8.1% for TXA, and 8.2% for ␧-aminocaproic acid.8 It is difficult to directly compare the results with the BART study due to different endpoints, available data, and the likely variability in transfusion practices among many different institutions.9 Unlike this study’s primary focus on the use of FFP, platelets, and cryoprecipitate, BART investigators used postoperative blood loss of more than 1.5 liters as the primary endpoint, and they defined massive transfusion as the use of more than 10 units of RBCs.8 Dietrich et al recently reported a lower transfusion rate of RBCs with aprotinin use compared with TXA in a prospective, randomized, double-blind study involving 220 patients.10 Hemostatic blood components were not examined, but since only primary CABG and primary AVR operations were included, the need for these components would be expected to be low.

Martin et al also compared aprotinin with TXA in a cohort study of 1188 patients and found a higher transfusion rate for RBCs and FFP in the TXA group.11 The use of platelets was not different between the 2 groups, but only a small number of the patients (⬍4%) underwent procedures involving DHCA. The fact that this study dealt exclusively with patients undergoing DHCA provides evidence that the clinical impact (eg, efficacy, side effects) of aprotinin is likely different among target surgical populations. For example, Martin et al suggested that aprotinin may adversely increase the incidence of myocardial infarction in CABG patients.11 Such a finding does not likely apply to DHCA patients, who experience a much greater dilution of procoagulant factors, as well as an increase in fibrinolytic activity.12 In agreement with this data, NicolauRaducu et al recently reported the retrospective analysis of 48 aprotinin-treated and 36 TXA-treated patients undergoing DHCA, demonstrating that the blood-product usage tends to be more with TXA relative to aprotinin.13 The use of aprotinin was associated with postoperative renal dysfunction, but there were no differences between TXA and aprotinin in cardiac, neurologic, respiratory, or survival outcomes after DHCA. Lack of increases in stroke or myocardial infarction with aprotinin is also supported by Ehrlich et al who failed to show any difference in those events comparing aprotinin with placebo in DHCA cases.14

In an attempt to further quantify the impact changing to tranexamic acid had at the institution, the authors took an economic perspective. Although the drug cost of aprotinin was higher than tranexamic acid ($1,268 v $293), overall the institution is now spending more to treat coagulopathy in DHCA patients. The tranexamic acid group’s blood product transfusion costs were more than 60% higher. This is consistent with a prior study by Smith et al that demonstrated an overall perioperative cost-savings benefit from aprotinin when factors other than just drug cost were taken into account.15 In this study, it is also important to note that the tranexamic acid group had a higher re-exploration rate compared with the aprotinin group. Of the 11 re-explored patients in the study cohort, only 3 had identifiable surgical bleeding, suggesting most bleeding was the result of medical coagulopathy rather than lack of surgical hemostasis. Returning to the operating room clearly represents an additional, although difficult to quantify, financial cost.

Recombinant activated factor VII (rFVIIa) has been used successfully to treat postoperative coagulopathy in cardiac surgical patients.16,17 It has been used at the institution since 2002 to treat CPB-induced coagulopathy unresponsive to hemostatic blood products or in those patients unwilling to receive blood transfusions.18 Factor VIIa use in the present study functioned as a rescue treatment, when microvascular bleeding persisted despite multiple rounds of hemostatic blood products. A significantly higher percentage of patients in the tranexamic acid group (34.6% v 12.2%, p ⬍ 0.01) required rFVIIa rescue therapy. Economic implications aside, there may be some concern about tripling the use of rFVIIa. A recent multicenter trial found an increased, although not statistically significant, incidence of thrombotic complications in cardiac surgical patients receiving the drug as rescue therapy for bleeding.19 Given this precaution, it may be speculated that the major effect of aprotinin’s withdrawal is merely the trading of one drug’s potential complications for another drug’s potential complications.

An unexpected result of this analysis was the increased incidence of seizures in the TXA group (5 patients v none in the aprotinin group). This finding also was reported recently by Martin et al.11 The proconvulsant effect of TXA is presumably mediated by gamma-aminobutyric acid-receptor antagonism 20 in a dose-dependent manner.21 The authors speculate that DHCA patients are more susceptible to this complication because of prolonged TXA infusions during CPB (⬎3 hours, Table 2). Of course, the length of circulatory arrest also is associated with seizure risk,22 and the average DHCA time was longer in the TXA group (36 min v 28 min, p ⬍ 0.01).

However, the DHCA times for the patients who experienced seizures were 15-35 (median 32) minutes, all less than the TXA group average, and under the 40-minute limit for increased seizure risk according to Gaynor et al.22 As a result of this study, the clinical protocol changed the TXA infusion rate to a weight-based formula consisting of a loading dose of 15 mg/kg and an infusion rate of 7.5 mg/kg/h.

Limitations There are obvious limitations to this study, including that it was retrospective, nonrandomized, and relatively small. How-

CHANGING FROM APROTININ TO TRANEXAMIC ACID

963

ever, the advantage of this design is that patients underwent similar procedures by a single surgeon. Therefore, surgical technique, which can have a significant effect on transfusion requirements and outcome, would be expected to have limited variability. Indeed, the fact that there were no differences in CPB time or intraoperative cell-saver use would suggest the actual operative characteristics between the 2 groups were nearly identical. The anesthesia staff did not prophylactically transfuse hemostatic products until the absence of visible clots in the surgical field was subjectively confirmed. Transfusion of products in the ICU was handled separately by intensivists, and was laboratory data-driven. The similarly increased bloodproduct usage in the OR and ICU suggests that bleeding tendency related to TXA was not simply an intraoperative event.

In addition, a closer look at the pattern of rFVIIa administration showed that it was given only in the OR in 4/10 (40.0%) patients in the aprotinin group. This was similar to the 14/27 (51.8%) TXA-treated patients receiving rFVIIa only in the OR, suggesting that the OR anesthesiologists were not biased towards administering rFVIIa to TXA patients.

In conclusion, the authors experienced the increased use of blood bank resources for DHCA cases following the switch from aprotinin to TXA. Although 30-day mortality was not affected by increased transfusion associated with TXA, the authors cannot exclude the possibility that long-term outcomes of patients may be affected adversely by the exposure to excess blood products, re-exploration procedures, or the occurrence of seizures. Other institutions performing cardiac operations at high risk for postoperative coagulopathy should be aware of the potential need for these additional resources. Aprotinin is no longer available, but a newer potent antifibrinolytic agent is being considered as its replacement.23 It is thus possible that an adequately powered prospective randomized study can be performed to evaluate the safety and efficacy of TXA relative to such new interventions.

ACKNOWLEDGMENT The authors acknowledge helpful discussions with Dr. Jerrold Levy.

The authors are also grateful to Katherine Egan, Brady Rumph, Kyle Mavros, and Matthew Klopman, Department of Anesthesiology, Emory University for their help with data collections.

REFERENCES 1. Crawford ES, Saleh SA: Transverse aortic arch aneurysm: Improved results of treatment employing new modifications of aortic reconstruction and hypothermic cerebral circulatory arrest. Ann Surg 194:180-188, 1981 2. Ehrlich M, Grabenwoger M, Luckner D, et al: The use of profound hypothermia and circulatory arrest in operations on the thoracic aorta. Eur J Cardiothorac Surg 11:176-181, 1997 3. Gega A, Rizzo JA, Johnson MH,T et al: Straight deep hypothermic arrest: Experience in 394 patients supports its effectiveness as a sole means of brain preservation. Ann Thorac Surg 84:759-766; discussion 766-767, 2007 4. Green JA, Spiess BD: Current status of antifibrinolytics in cardiopulmonary bypass and elective deep hypothermic circulatory arrest.

Anesthesiol Clin North America 21:527-551 viii, 2003 5. Henry DA, Carless PA, Moxey AJ, et al: Anti-fibrinolytic use for minimising perioperative allogeneic blood transfusion. Cochrane Database Syst Rev 2007:CD001886 6. Mannucci PM, Levi M: Prevention and treatment of major blood loss. N Engl J Med 356:2301-2311, 2007 7. Mora Mangano CT, Neville MJ, Hsu PH, et al: Aprotinin, blood loss, and renal dysfunction in deep hypothermic circulatory arrest.

Circulation 104:1276-1281, 2001 8. Fergusson DA, Hebert PC, Mazer CD, et al: A comparison of aprotinin and lysine analogues in high-risk cardiac surgery. N Engl J Med 358:2319-2331, 2008 9. Ide M, Bolliger D, Taketomi T, et al: Lessons from the aprotinin saga: Current perspective on antifibrinolytic therapy in cardiac surgery.

J Anesthesia 24:96-106, 2010 10. Dietrich W, Spannagl M, Boehm J, et al: Tranexamic acid and aprotinin in primary cardiac operations: An analysis of 220 cardiac surgical patients treated with tranexamic acid or aprotinin. Anesth Analg 107:1469-1478, 2008 11. Martin K, Wiesner G, Breuer T, et al: The risks of aprotinin and tranexamic acid in cardiac surgery: A one-year follow-up of 1188 consecutive patients. Anesth Analg 107:1783-1790, 2008 12. Wilde JT: Hematological consequences of profound hypothermic circulatory arrest and aortic dissection. J Card Surg 12:201-206, 1997

13. Nicolau-Raducu R, Subramaniam K, Marquez J, et al: Safety and efficacy of tranexamic acid compared with aprotinin in thoracic aortic surgery with deep hypothermic circulatory arrest. J Cardiothorac Vasc Anesth 24:73-79, 2009 14. Ehrlich M, Grabenwoger M, Cartes-Zumelzu F, et al: Operations on the thoracic aorta and hypothermic circulatory arrest: Is aprotinin safe? J Thorac Cardiovasc Surg 115:220-225, 1998 15. Smith PK, Datta SK, Muhlbaier LH, et al: Cost analysis of aprotinin for coronary artery bypass patients: Analysis of the randomized trials. Ann Thorac Surg 77 635-642; discussion 642-643, 2004 16. Dunkley S, Phillips L, McCall P, et al: Recombinant activated factor VII in cardiac surgery: Experience from the Australian and New Zealand Haemostasis Registry. Ann Thorac Surg 85:836-844, 2008 17. Romagnoli S, Bevilacqua S, Gelsomino S, et al: Small-dose recombinant activated factor VII (NovoSeven) in cardiac surgery.

Anesth Analg 102:1320-1326, 2006 18. Tanaka KA, Waly AA, Cooper WA, et al: Treatment of excessive bleeding in Jehovahâ&#x20AC;&#x2122;s Witness patients after cardiac surgery with recombinant factor VIIa (NovoSeven). Anesthesiology 98:1513-1515, 2003 19. Gill R, Herbertson M, Vuylsteke A, et al: Safety and efficacy of recombinant activated factor VII: A randomized placebo-controlled trial in the setting of bleeding after cardiac surgery. Circulation 120: 21-27, 2009 20. Furtmuller R, Schlag MG, Berger M, et al: Tranexamic acid, a widely used antifibrinolytic agent, causes convulsions by a gammaaminobutyric acid(A) receptor antagonistic effect. J Pharmacol Exp Ther 301:168-173, 2002 21. Yeh HM, Lau HP, Lin PL, et al: Convulsions and refractory ventricular fibrillation after intrathecal injection of a massive dose of tranexamic acid. Anesthesiology 98:270-272, 2003 22. Gaynor JW, Nicolson SC, Jarvik GP, et al: Increasing duration of deep hypothermic circulatory arrest is associated with an increased incidence of postoperative electroencephalographic seizures. J Thorac Cardiovasc Surg 130:1278-1286, 2005 23. Dietrich W, Nicklisch S, Koster A, et al: CU-2010 â&#x20AC;&#x201C;A novel small molecule protease inhibitor with antifibrinolytic and anticoagulant properties. Anesthesiology 110:123-130, 2009