Abstract:
Patients with uncontrolled hemorrhage require massive transfusion therapy and consume a large fraction of blood bank resources. Institutional guidelines have been established for treatment, but early identification and prevention in susceptible patients remains challenging. Uncontrolled hemorrhage was defined as meeting institutional guidelines for recombinant FVIIa administration. Patients who received rFVIIa were compared with patients who did not require the therapy but who were operated on during the same time period. After institutional review board approval, demographic, operative, and transfusion data were analyzed from a prospective database. Patients receiving rFVIIa were more likely to undergo multiple procedures (2.6 ± 0.8 vs. 1.8 ± 0.8; p < .001); aortic surgery (59% vs. 11%; p < .005); have a higher Cleveland Clinic Clinical Severity score (7.8 ± 2.7 vs. 5.5 ± 4.0; p < .005); require longer bypass (265 ± 92 min vs. 159 ± 63 min; p < .001), cross-clamp (182 ± 68 min vs. 112 ± 56 min; p < .001), and circulatory arrest (15 ± 24 minvs.2±7 min; p < .05) times; and require more autotransfusion (2580 ± 1847 mL vs. 690 ± 380 mL; p < .05). Uncontrolled hemorrhage is associated with more complex surgery requiring longer bypass times and more autotransfusion.
Keywords: cardiac surgery, recombinant factor VIIa, hemorrhage
Uncontrolled hemorrhage in cardiac surgery is more frequently encountered than in other surgical specialties, and patients who suffer from hemorrhage are 15-fold more likely to die (1). However, it was Karkouti et al. (2) that first established that major blood loss is an independent risk factor for adverse outcomes, instead of simply a marker for the severity of the illness. The distinction is important, because it mandates that we identify, investigate, and bring into clinical practice therapies to reduce this complication. The logical first step is to identify which patients or operative variables will present with the greatest risk for contributing to major blood loss.
Recently, significant interest in using recombinant FVIIa as a rescue therapy for uncontrolled hemorrhage has been shown by a number of peer-reviewed papers, but most trials have been small case series with differential results (3–8). Despite growing investigative interest, rFVIIa is only approved for use in patients with hemophilia A or B with inhibitors to factor VIII or factor IX or in patients with congenital FVII deficiency (9). Other reported uses are off-label and based on limited clinical data.
Furthermore, patients with circulating tissue factor (TF) or predisposing coagulopathy may be at an increased risk for thromboembolic complications with the use of rFVIIa (9). A recent large review in the Journal of the American Medical Association found that most thromboembolic events occurred with the off-label use of rFVIIa and often resulted in severe morbidity and mortality. The cautious application of rFVIIa is particularly warranted in cardiac surgery, because myocardial ischemia itself is known to contribute to pro-thrombotic states, and cardiopulmonary bypass (CPB) levels of TF are markedly elevated (10). In the presence of large amounts of shed blood (expected with uncontrolled hemorrhage), tissue factor levels are elevated even further (11). Therefore, extreme caution is warranted in the decision to use of rFVIIa to arrest un-controlled hemorrhage.
Despite these cautions, rFVIIa has been used as a com-passionate rescue therapy at our institution in >70 cardiac surgery patients. The initial series of 17 patients led to the definition of specific criteria for the administration of rFVIIa, and a subsequent series found that the use of this criteria resulted in the cessation of the hemorrhagic event in 100% of the patients. In an effort to improve safety, criteria are being established to exclude the potential for giving rFVIIa in the presence of disseminated intravascular coagulation by testing for soluble fibrin monomer complexes and antithrombin III levels before administration.
However, the identification of early markers of uncontrolled hemorrhage and more punctual and accepted interventions may preclude the need for rescue therapy. Such a strategy is more attractive than resorting to an expensive, off-label pharmaceutical of questionable safety. The purpose of this study was to characterize the demographic and operative parameters of the patients who suffered uncontrolled hemorrhage and necessitated rescue therapy in the form of rFVIIa.
MATERIALS AND METHODS
After institutional review board approval, data on 204 patients were prospectively collected. Patients suffering from uncontrolled hemorrhage were defined as those patients who met institutional guidelines for the administration of rFVIIa (rFVIIa, n = 17) and were compared with those who did not meet this criteria (No rFVIIa, n = 187). All patients received similar pre-operative, operative, and post-operative care.
A univariate comparison of the groups was conducted that compared demographic, procedure, operative, volume administration, and transfusion data. All data were loaded onto Microsoft Excel and analyzed with SAS (version 8.0; SAS, Cary, NC). Parametric data were analyzed using two-sample t tests, and non-parametric data were compared using χ2 tests for independence with or without Yates’s correction, and Fisher exact test, as appropriate.
Transfusion decisions were based on an institutional algorithm. Packed red blood cells (RBCs) were transfused for hemocrits <22% for patients ≤65 years old and 24% for patients >65 years old. If non-surgical bleeding occurred, other blood products were used. Fresh frozen plasma (FFP) was transfused when the thromboelastograph (TEG) R-time was >16 minutes or 12–16 minutes with a fibrinogen level <100 mg/dL. Platelets (Plt) were transfused if the TEG maximum amplitude (MA) was <49 mm or if the platelet count was <100,000. Cryoprecipitate (CRYO) was transfused if the R-time was >16 minutes, and the fibrinogen level was <100 mg/dL. Uncontrolled hemorrhage occurred when the criteria for rFVIIa administration was met, which included non-surgical bleeding despite a platelet count >100,000, a fibrinogen level >100 mg/dL, a TEG R-time <12 minutes, a TEG MA >49 mm, and a TEG LY30 of ≥8%.
Anesthesia Management
Patients received 0.5–5.0 mg of midazolam in the preoperative care unit, followed by the placement of two 16-gauge intravenous lines. After arriving in the operative suite, a radial or brachial arterial line was placed, followed by a continuous cardiac output pulmonary artery catheter (Baxter), inserted through an internal jugular introducer. Induction was accomplished by 5.0–10.0 mg midazolam, 0.5–1.0 mg Fentanyl, and 7.0–10.0 mg of pancuronium (cisatracurium or vecuronium were occasionally used instead of Pavulon to maintain a heart rate <100 bpm).
Before CPB, formal anesthesia guidelines were in place to keep fluid administration <1200 mL. Blood pressures were maintained with either Neo-Synephrine and ephedrine or epinephrine infusion through the central line. Fentanyl (as needed), Ativan (2 mg), and Versed (5–10 mg) were given before initiation of CPB at the discretion of the anesthesia staff. In addition,1gof cefazolin was given <60 minutes before skin incision, followed by repeat 1-g doses every 4 hours until closure.
Amicar or aprotinin was used on all cases, with selection based on surgeon preference. If Amicar was used,5gwas infused before CPB. If aprotinin was used, a 1-mL test dose was given, followed by 2,000,000 KIU infused over 30 minutes, which was followed by a 1,000,000-KIU/h drip continued through intensive care unit (ICU) arrival.
CPB Management
The CPB circuit consisted of a an oxygenator with integrated arterial line filter and cardiotomy reservoir (Synthesis; Cobe Cardiovascular, Arvada, CO), SmARTx-coated tubing (Cobe Cardiovascular, Allen, TX), Myocardial Protection System (MPS; Quest), inline blood gas monitoring (CDI 500; Terumo Cardiovascular, Ann Arbor, MI), and centrifugal pump (Revolution; Cobe Cardiovascular). The circuit was primed with 1400 mL of PlasmaLyteA, 35 mEq NaHCO3, 5000 IU heparin, and 25 g mannitol.
Autotransfusion was used with every case (CATS; Terumo Cardiovascular). A ratio of 1 part anticoagulant (30 IU heparin/mL 0.9% NS) to 10 parts collected blood was used and was washed with 0.9% NaCl solution on the “Quality Wash” program.
Autologous priming was attempted on all patients. During CPB, arterial pressure was maintained between 60 and 90 mmHg, cardiac index was maintained >1.8 L/min/m2, and blood gases were managed with alpha-stat physiology (pHa, 7.35–7.45; PaCO2,35–45 mmHg; PaO2, 150–250 mmHg). Patients were cooled to a temperature specified by surgeon preference and surgical procedure, with temperature control according to institutional policy (gradients <6°C, maximum arterial temperature <37°C, rewarming rate <0.5°C/min). Vacuum-assisted venous drainage was only used when needed and never exceeded −40 mmHg. Volume replacement during CPB was accomplished with PlasmaLyteA, with 12.5 g of albumin added per liter crystalloid solution, or 5% albumin added to achieve [Albumin] >3.5 g/dL or colloid oncotic pressure (COP) >14 mmHg.
Induction of cardioplegic arrest was accomplished with 1000–1500 mL of cold (4°C) 4 blood:1 crystalloid. The MPS was set to deliver 20 mEq/L KCl and 15 mEq/L of NaHCO3, with the crystalloid component consisting of 1 L 0.9% saline with 12.5 g mannitol. Subsequent doses were given at 15-to 20-minute intervals and consisted of cold (4°C) blood with 10–16 mEq/L KCl (adjusted to ensure electrical quiescence and avoid hyperkalemia). Antegrade cardioplegia was delivered at a system pressure of 100–150 mmHg, and retrograde cardioplegia was delivered to a coronary sinus pressure of 30–40 mmHg. Antegrade vs. retrograde delivery was based on surgeon preference and surgical procedure. Before removal of the cross-clamp, warm blood (37°C) was delivered for 5 minutes, with additional [KCl] of 8 mEq for the first minute, 4 mEq for the second minute, and 0 mEq for the remaining 3 minutes.
If Amicar was used,5gof Amicar was added to the pump prime, the patient was anticoagulated with a loading dose of 300 IU/kg of heparin, and additional heparin was given during CPB to maintain an activated clotting time (ACT) > 480 seconds (ACTII; Medtronic, Minneapolis, MN) If aprotinin was used, 2,000,000 KIU of aprotinin was added to the pump prime, the patient was anticoagulated with 400 IU/kg of heparin, and additional heparin was given during CPB to maintain ACT > 600 seconds. After termination of CPB and patient stability was assured, heparin was reversed with 0.6 mg of protamine per 100 IU heparin administered.
RESULTS
Patients who had uncontrolled hemorrhage were at a higher pre-operative risk than those that did not according to the Cleveland Clinic Clinical Severity (CCCS) score (5.5 ± 4.0 vs. 7.8 ± 2.7; p = .003), but the groups were otherwise similar in pre-operative demographics (Table 1) (12). Uncontrolled hemorrhage patients were also more likely to have aortic surgery (59% vs. 12%; p < .0001) and undergo multiple procedures (2.6 ± 0.8 vs. 1.8 ± 0.8; p < .0001; Table 1).
Table 1.
Demographic and procedure variables.
| No FVII (n = 187) (Mean ± SD or Percentage) | FVII (n = 17) (Mean ± SD or Percentage) | p | |
|---|---|---|---|
| Demographics | |||
| Age (yr) | 66.5 ± 11.3 | 69.6 ± 8.9 | NS* |
| Weight (kg) | 81.7 ± 18.0 | 87.6 ± 16.5 | NS* |
| Height (cm) | 166.8 ± 12.2 | 172.6 ± 10.2 | NS* |
| BSA (m2) | 1.93 ± 0.25 | 2.03 ± 0.24 | NS* |
| BMI (kg/m2) | 29.5 ± 7.1 | 29.2 ± 3.3 | NS* |
| Percent female | 43% | 29% | NS |
| CCCS score | 5.5 ± 4.0 | 7.8 ± 2.7 | .003* |
| Procedure | |||
| 3 valve | 6% | 35% | .0003† |
| 2 valves | 27% | 47% | NS† |
| 1 valve | 71% | 53% | NS† |
| CAB? | 40% | 41% | NS† |
| CAB only? | 3% | 0% | NS‡ |
| Aortic | 12% | 59% | <.0001† |
| Conversion | 4% | 0% | NS‡ |
| Salvage | 2% | 0% | NS‡ |
| Other | 24% | 18% | NS† |
| No. of procedures | 1.8 ± 0.8 | 2.6 ± 0.8 | <.0001* |
| 3 or more | 16% | 47% | .005† |
| 1 | 40% | 0% | .002† |
| 2 | 44% | 53% | NS† |
| 3 | 13% ± 25 | 29% ± 5 | NS* |
| 4 | 2% | 18% | .002 |
| 5 | 1% | 0% | NS‡ |
| Redo? | 21% | 18% | NS† |
| Minimally invasive? | 19% | 0% | .046† |
Two-tailed t test.
χ2 test.
Fisher exact test.
BSA, body surface area; BMI, body mass index; CAB, coronary artery bypass; NS, not significant.
The groups were similar in operative hematocrit, pre-CPB volume administration and transfusion, and CPB non-blood product volume administration (Tables 2–4). However, uncontrolled hemorrhage patients received more autotransfusion during CPB (339 ± 443 mL vs. 64 ± 170 mL; p = .019) and after CPB (3046 ± 2104 mL vs. 773 ± 490 mL; p < .0001) and received more transfusions during and after CPB (Tables 2–5).
Table 2.
Operative variables.
| Operative Variables | No FVII (n = 187) (Percent or Mean ± SD) | FVII (n = 17) (Percent or Mean ± SD) | p |
|---|---|---|---|
| IABP | |||
| Pre-operative | 5% | 12% | NS* |
| Intra-operative | 4% | 6% | NS† |
| Post-operative | 1% | 6% | NS† |
| Drugs | |||
| Aprotinin? | 67% | 88% | NS* |
| rFVIIa? | 0% | 100% | <.0001* |
| ATIII? | 0% | 29% | <.0001* |
| Hematocrit | |||
| Postinduction | 34.3 ± 4.8 | 34.9 ± 4.7 | NS‡ |
| First CPB | 25.1 ± 4.5 | 26.1 ± 3.8 | NS‡ |
| Last CPB | 25.8 ± 3.8 | 27.1 ± 3.7 | NS‡ |
| Cell saver | |||
| Pre-CPB | 37 ± 116 | 128 ± 196 | NS‡ |
| CPB | 64 ± 170 | 339 ± 433 | .019‡ |
| Post-CPB | 690 ± 380 | 2580 ± 1848 | .001‡ |
| Total CS | 773 ± 490 | 3046 ± 2104 | <.0001‡ |
χ2 test.
Fisher exact test.
Two-tailed t test.
IABP, intra-aortic; CS, cell saver; NS, not significant.
Table 3.
Pre-CPB variables.
| Pre-CPB Variables | No FVII (n = 187) (Percentage or Mean ± SD) | FVII (n = 17) (Percentage or Mean ± SD) | p |
|---|---|---|---|
| Albumin | 237.2 ± 409.7 | 670.6 ± 966.2 | NS* |
| Crystalloid | 1505.5 ± 731.7 | 1511.8 ± 883.8 | NS* |
| RBCs (units) | 0.1 ± 0.5 | 0.7 ± 2.4 | NS* |
| RBCs (Y/N) | 5% | 12% | NS† |
| Urine | 365 ± 296 | 341 ± 182 | NS* |
Two-tailed t test.
χ2 test.
RBC, red blood cells; NS, not significant.
Table 4.
CPB variables.
| CPB Variables | No FVII (n = 187) (Percentage or Mean ± SD) | FVII (n = 17) (Percentage or Mean ± SD) | p |
|---|---|---|---|
| AP Taken | 952 ± 384 | 958 ± 515 | NS* |
| AP Given | 433 ± 418 | 368 ± 388 | NS* |
| Crystalloid | 434 ± 648 | 827 ± 1049 | NS* |
| Albumin | 551 ± 449 | 765 ± 504 | NS* |
| RBCs (PRIME) | 0.1 ± 0.5 | 0.2 ± 0.6 | NS* |
| RBC prime (Y/N) | 7% | 18% | .007† |
| RBCs | 0.6 ± 1.1 | 2.6 ± 2.9 | .012* |
| RBCs (Y/N) | 29% | 65% | .006† |
| FFP | 0.1 ± 0.5 | 1.8 ± 2.6 | .016* |
| FFP (Y/N) | 2% | 41% | <.0001† |
| Other | 251 ± 324 | 270 ± 211 | NS* |
| Urine | 610 ± 444 | 736 ± 861 | NS* |
| Ultrafiltration | 211 ± 632 | 589.7 ± 1086.2 | NS* |
| Ultrafiltration (Y/N) | 16% | 35% | .046† |
| CPB time | 160 ± 63 | 265 ± 92 | .000* |
| XC time | 113 ± 56 | 182 ± 68 | .001* |
| Circ arrest | 1.5 ± 7.0 | 15.2 ± 24.0 | .032* |
| DHCA (Y/N) | 6% | 41% | <.0001† |
Two-tailed t test.
χ2 test.
AP, autologous prime; XC, cross clump; DHCA, deep hypothermic circulatory arrest; NS, not significant.
Table 5.
Post-CPB variables.
| Post-CPB Variables | No FVII (n = 187) (Mean ± SD or Percentage) | FVII (n = 17) (Mean ± SD or Percentage) | p |
|---|---|---|---|
| Crystalloid | 547 ± 549 | 1124 ± 956 | .025* |
| Colloid | 113 ± 278 | 259 ± 344 | NS* |
| RBCs | 0.6 ± 1.2 | 5.4 ± 3.5 | <.0001* |
| RBCs (Y/N) | 28% ± 52 | 100% | <.0001† |
| FFP | 0.6 ± 1.6 | 8.6 ± 5.5 | <.0001* |
| FFP (Y/N) | 12% | 94% | <.0001† |
| CRYO | 0.5 ± 2.2 | 11.4 ± 9.4 | .000* |
| CRYO (Y/N) | 7% | 82% | <.0001† |
| Plt | 0.6 ± 2.2 | 6.5 ± 8.7 | .013* |
| Plt (Y/N) | 14% | 94% | <.0001† |
Two-tailed t test.
χ2 test.
NS, not significant.
DISCUSSION
Dial et al. (13) reported that hemodilution and surgical blood loss are major contributors to transfusion requirements during cardiac surgery. Consistent with these results, patients in this trial required more autotransfusion, beginning during bypass and extending into the post-bypass period. Furthermore, Koster et al. (14) recently reported improved preservation of hemostasis with the avoidance of cardiotomy suction, through passive venting of the heart and minimizing suction return. In this study, shed blood was primarily returned to the CPB circuit, and larger autotransfusion volumes corresponded to much larger cardiotomy suction use. Although proponents of complete avoidance of cardiotomy exist, the technique is limited in application because of the inability to accommodate even moderate levels of blood loss or venting requirements. Central to preserving adequate hemostasis is limiting surgical blood loss. Although autotransfusion and cardiotomy suction provide useful mechanisms for reducing the requirement for allogeneic RBC transfusion, the use of both have the strong potential to reduce hemostatic potential. With autotransfusion, all blood components except RBCs are lost to the effluent, potentially contributing to both dilutional coagulopathy and decreased plasma protein and colloid oncotic pressure. Thus, excessive surgical blood loss, compensated with autotransfusion, may result in higher non-RBC transfusion requirements, excessive extravascular water accumulation, and resultant increases in end organ dysfunction.
The management of shed blood during cardiac surgery will continue to receive a great deal of attention, primarily because none of the perfusionist-driven strategies are attractive. The only strategy that is definitively superior is the early identification and cessation of surgical blood loss at the field. Blood allowed to pool in the pleural and pericardial spaces will accumulate supraphysiologic levels of deleterious biological mediators and necessitates further hemodilution to replace the temporarily “lost” volume. Earlier return is an improvement, but the blood will still be exposed to exposed tissue, additional artificial surfaces, negative pressures, and large air–blood interfaces.
Limiting these exposures should be a primary goal of the entire team, requiring time and attention in all facets of care. From this trial, it seems that it is more difficult (and thus would consume more time) to limit shed surgical blood when the operative course includes more than three procedures or aortic surgery.
Because of the power and observational design of this study, it is not possible to determine whether the differences between non-hemorrhage patients and uncontrolled hemorrhage patients predicted the condition, contributed to the condition, or were a result of the condition. All that can be stated is that patients in the uncontrolled hemorrhage group more frequently required multiple procedures (≥3) and/or aortic surgery; required more autotransfusion, longer bypass times, and more hypothermic circulatory arrest; and required more transfusions. Given these characteristics, efforts to preserve hemostatic potential should be studied further in this patient population.
REFERENCES
- 1.Herwaldt LA, Swartzendruber SK, Edmond MB, et al. The epidemiology of hemorrhage related to cardiothoracic operations. Infect Control Hosp Epidemiol. 1998;19:6–8. [DOI] [PubMed] [Google Scholar]
- 2.Karkouti K, Wijeysundera DN, Yau TM, et al. The independent association of massive blood loss with mortality in cardiac surgery. Transfusion. 2004;44:1453–62. [DOI] [PubMed] [Google Scholar]
- 3.Bishop CV, Renwick WE, Hogan C, et al. Recombinant activated factor VII: treating postoperative hemorrhage in cardiac surgery. Ann Thorac Surg. 2006;81:875–9. [DOI] [PubMed] [Google Scholar]
- 4.von Heymann C, Redlich U, Jain U, et al. Recombinant activated factor VII for refractory bleeding after cardiac surgery—a retrospective analysis of safety and efficacy. Crit Care Med. 2005;33:2241–6. [DOI] [PubMed] [Google Scholar]
- 5.Diprose P, Herbertson MJ, O’shaughnessy D, Gill RS.. Activated recombinant factor VII after cardiopulmonary bypass reduces allogeneic transfusion in complex non-coronary cardiac surgery: randomized double-blind placebo-controlled pilot study. Br J Anaesth. 2005;95:596–602. [DOI] [PubMed] [Google Scholar]
- 6.Raivio P, Suojaranta-Ylinen R, Kuitunen AH.. Recombinant factor VIIa in the treatment of postoperative hemorrhage after cardiac surgery. Ann Thorac Surg. 2005;80:66–71. [DOI] [PubMed] [Google Scholar]
- 7.Gowers CJ, Parr MJ.. Recombinant activated factor VIIa use in massive transfusion and coagulopathy unresponsive to conventional therapy. Anaesth Intensive Care. 2005;33:196–200. [DOI] [PubMed] [Google Scholar]
- 8.Hyllner M, Houltz E, Jeppsson A.. Recombinant activated factor VII in the management of life-threatening bleeding in cardiac surgery. Eur J Cardiothorac Surg. 2005;28:254–8. [DOI] [PubMed] [Google Scholar]
- 9.Novo Nordisk, NovoSeven Coagulation Factor VIIa (Recombinant) Package Insert. Bagsvaerd, Denmark: Novo Nordisk; 1999. [Google Scholar]
- 10.Kalweit G, Bach J, Huwer H, Winning J, Hellstern P.. The impact of cardiac ischemia and reperfusion on markers of activated haemostasis and fibrinolysis during cardiopulmonary bypass: comparison of plasma levels in arterial and coronary venous blood. Thromb Res. 2005;116:33–9. [DOI] [PubMed] [Google Scholar]
- 11.Bull BS, Hay K, Razzouk AJ, Huse WM.. Reinfusion of aspirated pericardial blood during CPB. Part II. Laparotomy sponges are hazardous parts of the CPB circuit? Blood Cells Mol Dis. 2005;34:144–50. [DOI] [PubMed] [Google Scholar]
- 12.Higgins TL, Estafanous FG, Loop FD, et al. Stratification of morbidity and mortality outcome by preoperative risk factors in coronary artery bypass patients. a clinical severity score. JAMA. 1992;268:1860. [PubMed] [Google Scholar]
- 13.Dial S, Delabays E, Albert M, et al. Hemodilution and surgical hemostasis contribute significantly to transfusion requirements in patients undergoing coronary artery bypass. Thorac Cardiovasc Surg. 2005;130:654–61. [DOI] [PubMed] [Google Scholar]
- 14.Koster A, Bottcher W, Merkel F, Hetzer R, Kuppe H.. The more closed the bypass system the better: a pilot study on the effects of reduction of cardiotomy suction and passive venting on hemostatic activation during on-pump coronary artery bypass grafting. Perfusion. 2005;20:285–8. [DOI] [PubMed] [Google Scholar]