| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2007;83:S27-S86
© 2007 The Society of Thoracic Surgeons
a University of Kentucky Chandler Medical Center, Lexington, Kentucky
b University of Florida, Jacksonville, Florida
c University of Pennsylvania Health System, Philadelphia, Pennsylvania
d Harefield Hospital, London, United Kingdom
e Rush Presbyterian St. Lukes’ Medical Center, Chicago, Illinois
f Washington University Medical Center, St. Louis, Missouri
g Center for the Evaluative Clinical Sciences, Dartmouth Medical School, Lebanon, New Hampshire
h Virginia Commonwealth University, Richmond, Virginia
i Montefiore Medical Center, Bronx, New York
j Duke University Medical Center, Durham, North Carolina
k Keenan Research Center in the Li Ka Shing Knowledge Institute of St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada
l Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts
* Address correspondence to Dr Ferraris, Division of Cardiovascular and Thoracic Surgery, University of Kentucky Chandler Medical Center, CTW Bldg, Suite 320, 900 South Limestone, Lexington, KY 40536-0200 (Email: ferraris{at}earthlink.net).
| See Appendix 1 for authors’ financial relationships with industry.
|
 |
Abstract |
|---|
 Top Abstract Introduction1) Methods Used in...2) Risks and Benefits...3) Causes of Blood...4) Indications for Blood...5) Interventions to Limit...6) Prophylaxis--Interventions...7) Transfusion and Blood...8) Summary Treatment Strategy-...9) SummaryAppendix 1Appendix 22. Conduct of the...3. Peer Review and...4. Recommendations of the...FootnotesReferences |
|---|
Methods: We reviewed all available published evidence related to blood conservation during cardiac operations, including randomized controlled trials, published observational information, and case reports. Conventional methods identified the level of evidence available for each of the blood conservation interventions. After considering the level of evidence, recommendations were made regarding each intervention using the American Heart Association/American College of Cardiology classification scheme.
Results: Review of published reports identified a high-risk profile associated with increased postoperative blood transfusion. Six variables stand out as important indicators of risk: (1) advanced age, (2) low preoperative red blood cell volume (preoperative anemia or small body size), (3) preoperative antiplatelet or antithrombotic drugs, (4) reoperative or complex procedures, (5) emergency operations, and (6) noncardiac patient comorbidities. Careful review revealed preoperative and perioperative interventions that are likely to reduce bleeding and postoperative blood transfusion. Preoperative interventions that are likely to reduce blood transfusion include identification of high-risk patients who should receive all available preoperative and perioperative blood conservation interventions and limitation of antithrombotic drugs. Perioperative blood conservation interventions include use of antifibrinolytic drugs, selective use of off-pump coronary artery bypass graft surgery, routine use of a cell-saving device, and implementation of appropriate transfusion indications. An important intervention is application of a multimodality blood conservation program that is institution based, accepted by all health care providers, and that involves well thought out transfusion algorithms to guide transfusion decisions.
Conclusions: Based on available evidence, institution-specific protocols should screen for high-risk patients, as blood conservation interventions are likely to be most productive for this high-risk subset. Available evidence-based blood conservation techniques include (1) drugs that increase preoperative blood volume (eg, erythropoietin) or decrease postoperative bleeding (eg, antifibrinolytics), (2) devices that conserve blood (eg, intraoperative blood salvage and blood sparing interventions), (3) interventions that protect the patient’s own blood from the stress of operation (eg, autologous predonation and normovolemic hemodilution), (4) consensus, institution-specific blood transfusion algorithms supplemented with point-of-care testing, and most importantly, (5) a multimodality approach to blood conservation combining all of the above.
|
|
1) Methods Used in Developing Guidelines |
|---|
|
TopAbstractIntroduction1) Methods Used in...2) Risks and Benefits...3) Causes of Blood...4) Indications for Blood...5) Interventions to Limit...6) Prophylaxis--Interventions...7) Transfusion and Blood...8) Summary Treatment Strategy-...9) SummaryAppendix 1Appendix 22. Conduct of the...3. Peer Review and...4. Recommendations of the...FootnotesReferences |
|---|
|
Evidence-based guidelines are an attempt to reconcile often conflicting lines of evidence, giving greater weight to evidence derived from more methodologically rigorous studies and those for which the overall weight of evidence is most convincing. They must be viewed as guidelines and recommendations, not absolutes. With this in mind, the authors searched numerous sources for available evidence about specific questions relating to the use of blood transfusions and blood conservation before, during, and after cardiac operations.
The committees from The Society of Thoracic Surgeons (STS) and the Society of Cardiovascular Anesthesiologists (SCA) participated in guideline development. The committee members’ financial relationships with industry are listed in Appendix 1. Appendix 2 outlines the steps employed in development of the final guideline document. Within these guidelines, cost analysis was not a primary consideration in formulating the recommendations. Costs change constantly depending upon markets and other forces. The science was the focus, and the risks and benefits from bleeding and blood transfusion should not vary with societal economic forces.
|
|
2) Risks and Benefits of Blood Transfusion—the Dilemma |
|---|
|
TopAbstractIntroduction1) Methods Used in...2) Risks and Benefits...3) Causes of Blood...4) Indications for Blood...5) Interventions to Limit...6) Prophylaxis--Interventions...7) Transfusion and Blood...8) Summary Treatment Strategy-...9) SummaryAppendix 1Appendix 22. Conduct of the...3. Peer Review and...4. Recommendations of the...FootnotesReferences |
|---|
Arguably, the practice of modern transfusion medicine began with the discovery of blood groups by Landsteiner. Even though his work won the Nobel Prize in 1930, the full impact of his discovery lagged; and other developments such as arterial anastomosis (Carrel), blood component therapy, refrigeration of blood components, organization of blood banks, use of anticoagulation, and the urgency of treating war-injured patients were necessary to bring transfusion into the modern era.
As early as 1943, it was recognized that blood transfusion could spread diseases, especially hepatitis [2]. Since that time, other problems such as risks associated with paid donors and concerns about disease transmission, including the current epidemic of human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) and transmission of hepatitis C, raised awareness of the problems associated with blood transfusion (Table 2), almost to the point that the benefits of blood transfusion are overlooked. It is important to realize that the viral and parasitic infectious risks of blood transfusion are dramatically increased in third-world countries or in areas where modern blood banking practices are not available (http://www.cochrane.org/reviews/en/ab002042.html).
|
It is difficult to define the benefits of blood transfusion, as randomized trials to support the use of blood products to treat disease do not exist. Blood transfusion was accepted long before the complications associated with transfusion could be documented. Many traumatic injuries (especially war-related injuries) were almost universally fatal before the advent of blood transfusion. The practice of blood transfusion saved countless lives long before the complications of this therapy were recognized [6].
Enhanced oxygen-carrying capacity [7], improved hemostasis associated with blood component therapy [8, 9], and volume support of cardiac output are three accepted benefits of blood transfusion. Adams and associates, in 1942 (through the pioneering blood banking work of John Lundy) [10, 11], on the basis of clinical observations [10] and animal studies [11], introduced the "10/30" rule of blood transfusion. These authors suggested that the minimal ideal level of oxygen-carrying capacity is maintained by a hematocrit of around 30% and hemoglobin of 10 g/dL. Because of the risks of transfusion with associated costs, and lack of clear evidence regarding the benefit of blood transfusion, the 10/30 arbitrary rule has fallen into disfavor.
There is lack of clear evidence regarding the benefit of blood transfusion. However, clinical reports [12–14] of survival benefit support transfusion in certain clinical situations. A task force of the American Society of Anesthesiologists (ASA) developed a consensus statement based mostly on level B and C evidence that concluded that "red blood cell transfusions should not be dictated by a single hemoglobin ‘transfusion trigger’ but instead should be based on the patient’s risk of developing complications of inadequate oxygenation" [15]. They developed guidelines for transfusion of packed red cells in adults that were accepted by others without much argument and with little high-level evidence to support them. Their guidelines are listed in Table 3. The ASA guidelines do not specifically address the uniqueness of the cardiac surgery patient. Revised ASA guidelines are available on line at: http://www.asahq.org/publicationsAndServices/BCTGuidesFinal.pdf. Because of the lack of randomized trials to define the role of blood transfusion in cardiac surgery and because of concerns about complications of blood transfusion, it is reasonable to review the available evidence supporting transfusion decisions for cardiac operations. The aim of this review is to provide clinically useful guidelines, based on available evidence, to aid cardiothoracic surgeons and anesthesiologists in their decisions about blood conservation and transfusion.
|
|
3) Causes of Blood Transfusion After Cardiac Operations |
|---|
|
TopAbstractIntroduction1) Methods Used in...2) Risks and Benefits...3) Causes of Blood...4) Indications for Blood...5) Interventions to Limit...6) Prophylaxis--Interventions...7) Transfusion and Blood...8) Summary Treatment Strategy-...9) SummaryAppendix 1Appendix 22. Conduct of the...3. Peer Review and...4. Recommendations of the...FootnotesReferences |
|---|
|
Numerous reports identify multivariate predictors of postoperative bleeding and blood transfusion after cardiac operations. Additionally, multiple publications document anecdotal or observational reports of acquired, congenital, or process variables associated with increased bleeding. Table 4 is a summary of these publications with the variables separated by patient-related, procedure-related, and process-related factors. Very few studies listed in Table 4 differentiate between red blood cell transfusion and non–red cell hemostatic factor transfusion. Wherever this distinction was made, it was taken into account in the table. A minority of the studies listed in the table provide odds ratios or p values to support the strength of a particular risk factor as a predictor of blood transfusion. As a consequence, the factors are simply listed without an attempt to add a quantitative measure to each risk factor. Many of the risk factors in Table 4 are listed as a single factor, although it is realized that the risk may be a spectrum spanning low to high intensity. For example, the risk associated with age greater than 75 years is significantly greater than risk for a patient aged 55 years or younger, and it is quite likely that the risk of transfusion associated with age is not a continuous function. Likewise, the risk associated with some antithrombotic drugs (eg, aspirin) is low on the risk spectrum while that of others (eg, clopidogrel) is much higher. No attempt was made to measure the risk spectrum of variables listed in Table 4 as the evidence base for these variables as risk factors is limited and few quantitative descriptors of risk are available.
|
Special clinical situations call for special interventions and should be "red flags" for surgeons faced with these problems. Patients with acquired or congenital coagulopathies, patients scheduled for complex procedures (eg, combined valve/coronary revascularization, aortic dissection with deep hypothermic circulatory arrest), repeat cardiac procedures, sepsis with thrombocytopenia, and Jehovah’s Witnesses fall into this category.
c) Patient-Related Causes of Bleeding
There is evidence that certain patients have an accentuated response to the usual doses of antiplatelet drugs [17]. Certain "hyperresponders" to average doses of aspirin exhibit very prolonged skin bleeding times [18–20]. This accentuated response to aspirin may result in increased perioperative blood loss worsened by preoperative antiplatelet therapy. The mechanisms of these accentuated effects of aspirin and other antiplatelet drugs are undoubtedly related to individual pharmacogenomic variation and involve the antiplatelet, antiinflammatory, anticoagulant, and endothelial-protecting actions of these agents.
Patients with thrombocytopenia from whatever cause (defined as platelet count below 50,000) are at extremely high risk of excessive bleeding after CABG [21–25]. Additionally, patients with preoperative anemia (eg, renal failure, repeated blood drawing during prolonged ICU stay, multiple recent percutaneous procedures, and so forth) exhibit increased perioperative blood transfusion [20, 26–28]. One of the earliest observations about anemia was that bleeding time was prolonged in anemic patients [29, 30]. Anemia-related bleeding abnormalities are likely to be worsened by antiplatelet drugs [17].
Patients with other congenital or acquired qualitative platelet defects are at increased bleeding risk [26–28, 31–37]. Congenital defects include vonWillebrand’s disease, Bernard-Soulier syndrome, Glanzmann’s thrombasthenia, storage-pool disease, and others. Acquired qualitative defects are seen in liver disease, renal disease and drug induced qualitative platelet defects.
d) Physician-Related Causes of Bleeding
There is no doubt that physician practices influence bleeding and blood transfusion. Surgical practices differ widely and can dramatically influence perioperative bleeding and transfusion. Regardless of the patient’s condition at operation or immediately postoperatively, physicians do not consistently apply the same triggers or indications for ordering blood transfusions. Use of transfusions varies widely among centers [38–41]. Stover [2] showed the institution as well as the individual physician are independent multivariate risk factors for allogeneic transfusion. Practice variability in the transfusion of blood products impacts resource utilization [42]. Johnson and associates [43] describe the comparison of outcomes with two different transfusion strategies (conservative versus liberal), and they concluded that no specific transfusion trigger was justified but that the important indicator for transfusion should be patient symptoms and clinical condition. Variation in cardiopulmonary bypass practices, including time on bypass, influences platelet function and perioperative bleeding [44, 45]. Differences in practice patterns relative to recognition, correction, and exploration for excessive postoperative hemorrhage also contribute to wide variability in transfusion practices [46, 47].
e) Procedure-Related Causes of Bleeding
It is difficult to separate the procedure-related variables from those that are the result of physician practices, but certain procedure-related factors stand out. Repeat procedures have higher transfusion rates [48], and the type and urgency of operation are independent predictors for transfusion [49]. Hypothermia related to cardiopulmonary bypass influences platelet function and coagulation [50–52], and the effect can persist into the ICU [53]. Off-pump cardiac surgery is associated with an overall reduction in transfusion requirements [54–56]. Of the coronary artery bypass surgery patients, those who have bilateral internal mammary artery grafts suffer a greater postoperative blood loss than saphenous vein or single mammary artery graft patients [57]. Replacement of the aortic valve with a pulmonary autograft (Ross procedure) is a technically challenging operation, with up to 20% incidence of postoperative bleeding necessitating reexploration [58] and above average mortality rate [59]. Hemostatic dysfunction and concomitant coagulation abnormalities are often seen in candidates for and recipients of ventricular assist devices or artificial hearts [60–62]. These procedure-related variables influence perioperative bleeding and should alert the surgeon and anesthesiologist to additional risk.
f) Drug-Related Causes of Bleeding
Drug therapy has a growing role and influence in the prevention and treatment of cardiovascular disease. Many of the active cardiac drugs provide benefit by inhibiting platelet function, or by causing clot lysis. Thus, any surgical procedure required in the face of such drugs poses a greater than normal risk of perioperative hemorrhage and transfusion requirement. A large body of literature defines the risk/benefit relationship of antiplatelet and anticoagulant drugs, and this subject is covered in detail below.
Stent implantation in coronary arteries and in saphenous vein grafts is accompanied by guideline-driven pharmacologic therapy to maintain patency and prevent thrombo-occlusion [63–66], but with a relatively high incidence of bleeding and vascular complications [67, 68]. Hemorrhagic complications occur at reported rates between 1.9% and 9.8% after placement of stents, with worse outcome among patients receiving transfusions during stent implantation [69–73]. Other studies show even higher rates of bleeding complications, at 21.8% to 27.4%, with varying frequency of blood transfusion ranging between 0% and 9.5% [74, 75].
Preoperative antiplatelet and anticoagulant treatment as prophylaxis for coronary occlusive disease is associated with excessive intraoperative and postoperative bleeding as well as resultant transfusion, in most [20, 76–81], but not all situations [82–86]. Therefore, this aspect of patients’ preoperative medication regimens must be managed for maximum cardioprotective benefit, while minimizing risk of hemorrhagic complication. Guidelines are available to aid in this decision process [87]. Patients with thrombocytopenia or with qualitative platelet defects (eg, renal failure, vonWillebrand’s disease, anemia, and so forth) are particularly sensitive to antiplatelet drugs and represent a high-risk subset that require multiple blood conservation interventions to minimize blood loss and transfusion. Discontinuation of antiplatelet and antithrombotic drugs before CABG in these high-risk patients is prudent.
There is evidence that certain patients have an accentuated response to the usual doses of preoperative aspirin [17]. Certain "hyperresponders" to average doses of aspirin exhibit very prolonged skin bleeding times [18, 19]. This accentuated response to aspirin may result in increased perioperative blood loss [17]. The mechanisms of these effects of aspirin are undoubtedly multifactorial and include the antiplatelet, antiinflammatory, anticoagulant, and endothelial-protecting actions of aspirin. For the vast majority of patients with acute coronary syndromes who require urgent CABG, the best option seems to be to continue aspirin until operation [88]. For the elective patient who requires CABG and does not have an acute coronary syndrome, it may be reasonable to discontinue aspirin for a few days (2 to 3 days) with the expectation that there will be less perioperative bleeding and blood transfusion.
Dipyridamole reduced postoperative blood loss and transfusions in one study, and was thought to preserve platelets during operative intervention [89]. Dipyridamole and aspirin in combination increased bleeding compared to placebo or aspirin alone [90–93], and graft occlusion was reduced, but not eliminated [94]. Breyer and associates [95] demonstrated increased incidence of delayed postoperative cardiac tamponade with routine perioperative administration of aspirin and dipyridamole. According to the Antithrombotic Trialists’ Collaboration, the addition of dipyridamole to aspirin adds no significant benefit over aspirin alone and may increase bleeding risk [96].
Thienopyridines are a class of antiplatelet drugs that reduce adenosine diphosphate (ADP)–mediated platelet activation, with significant improvement of clinical outcomes in many coronary and cardiovascular conditions. Ticlopidine was the first available thienopyridine, but initially, widespread use was limited by frequent side effects, as well as neutropenia and thrombotic thrombocytopenic purpura [97–99]. In contrast, clopidogrel has a much better safety profile, and has become standard therapy after coronary stent implantation [97]. Because clopidogrel is so well tolerated, and pretreatment before stent implantation is advantageous to coronary artery patency, it is a common occurrence that patients undergoing urgent or emergent coronary artery bypass surgery have recent exposure to dual antiplatelet therapy of clopidogrel and aspirin. This dual therapy, although safe and effective during coronary intervention [100–103], results in higher postoperative bleeding, more transfused blood products, and higher rate of reexploration for mediastinal hemorrhage during emergency CABG [104–107]. To address this disturbing finding, Ley [108] described the implementation of a quality improvement initiative around preoperative exposure to clopidogrel. The current ACC/AHA guidelines and the current STS guidelines recommend discontinuing ADP-inhibitors 5 to 7 days before cardiac operations, if possible, recognizing that operations sooner than 5 days in patients on ADP-inhibitors risk increased perioperative bleeding and transfusions and possibly worse long-term outcomes [87, 109]. However, with the new drug-eluting stents sudden withdrawal of platelet inhibition appears to cause thrombotic risk [110, 111]. The drug-eluting stents exhibit delayed endothelialization compared with bare metal stents, and once platelet inhibition is removed, the foreign surface inside the vessel wall creates a nidus for platelet and white cell activation as well as a potential for hypersensitivity reaction [112, 113]. Minimal evidence is available to guide therapy in this situation. Consensus favors some element of platelet inhibition in the patient who requires CABG but has a coated stent in place. Discussion with all members of the cardiovascular team including cardiologists, hematologists, and the operating team is essential. Possible alternatives include shifting to a short half-life glycoprotein IIb/IIIa inhibitor (GPI) or to a direct thrombin inhibitor while waiting for the effects of thienopyridines to disperse, but more evidence is required to determine the best option.
Platelet GPIs, introduced a decade ago, play a profound role in abolition of platelet aggregation and platelet thrombus formation, and reduce the risk of acute ischemic complications after coronary occlusion and coronary angioplasty. Coller and colleagues [114] reported an approximately twofold increased risk of major bleeding in patients treated with abciximab. Short-acting GPIs, tirofiban and eptifibatide, in combination with aspirin and heparin, are used during coronary intervention without a significant increase in serious catheterization-related bleeding events [115–117]. The longer acting GPI, abciximab, improves the clinical outcome of percutaneous coronary intervention, but causes thrombocytopenia and access site bleeding [118, 119]. Tirofiban and abciximab have comparable efficacy and bleeding complications [120]. Limited evidence is available to guide the use of GPIs in patients with acute myocardial infarction or with unstable coronary syndromes who also need cardiac operations [87, 109]. These patients should be considered at high risk for bleeding and blood transfusion, and clinical judgment regarding timing of intervention and transfusion needs is likely the most important determinant of outcomes. Again, guidelines based on consensus agreement are available to assist in this decision analysis [87].
Heparin is an integral component of therapy for acute coronary events. Systematic review of six randomized controlled trials looking at heparin in patients with acute myocardial infarction treated with thrombolytic therapy showed severe bleeding to be similar between those receiving and not receiving heparin [121]. Low-molecular-weight heparins (LMWH) have an acceptable risk of bleeding in management of unstable angina [122], and short-term unfractionated heparin or LMWH is used in acute coronary syndromes. Low-molecular-weight heparin after coronary artery stenting had a 10.5% incidence of hematomas or false aneurysm, and a 3.25% incidence of blood transfusion or surgical repair [123]. One study showed hemorrhage and reexploration for bleeding after CABG to be significantly higher in patients receiving enoxaparin versus unfractionated heparin—but the timing of preoperative dosing is not described in this report [124]. Patients receiving LMWH within 12 hours of cardiac surgery have significantly greater blood loss and increased blood transfusion compared with patients receiving intravenous heparin or a dose of LMWH more than 12 hours before operation [125]. Medalion and associates [126] showed that enoxaparin administered more than 8 hours before coronary artery bypass surgery is not associated with increased postoperative bleeding or transfusion requirement.
Despite the almost universal use of unfractionated heparin, especially during cardiopulmonary bypass, there remains concerns with this drug including heparin rebound, heparin resistance, protamine reaction, heparin-induced thrombocytopenia, and heparin-induced platelet dysfunction [127, 128]. Investigation of alternatives to unfractionated heparin resulted in the availability of some direct-acting thrombin inhibitors (eg, hirudin, bivalrudin, and argatroban) to limit thromboses in patients with cardiovascular disease. Potential advantages of direct thrombin inhibitors include a more predictable anticoagulant response than heparin. Based on randomized trials direct thrombin inhibitors are superior to heparin in preventing arterial thromboses. For example, hirudin is superior to heparin in patients with acute myocardial infarction [129], but safety concerns about increased bleeding and inability to monitor anticoagulation limit use of this agent. The shortest acting direct thrombin inhibitor, bivalirudin, has a half-life of 25 minutes, while that of argatroban is 45 minutes, hirudin 60 minutes, and ximelagatran 3 hours. All of the direct thrombin inhibitors are plagued by the lack of reversibility, difficulties in monitoring the level of anticoagulation, and a prolonged half-life compared with heparin. Argatroban is cleared mainly by the liver, and its use is limited in patients with hepatic dysfunction, but may be preferred for treatment of heparin-induced thrombocytopenia in patients with renal dysfunction.
Experience with direct thrombin inhibitors is limited in patients having cardiac procedures after coronary interventions [134]; however, a recent controlled randomized trial using bivalirudin during off-pump coronary artery bypass surgery (OPCAB) suggests that bivalirudin is safe and effective as an anticoagulant during OPCAB [135]. Likewise, argatroban was used as an anticoagulant during OPCAB with satisfactory results [136]. Lack of a ready reversibility for direct thrombin inhibitors is a major limitation, especially with longer acting agents, and can lead to catastrophic bleeding when used in patients requiring CPB [137]. Bivalirudin is at least as effective as heparin in ischemic heart disease, but with significant reduction in major hemorrhage [130–133]. The Randomized Evaluation of PCI Linking Angiomax to Reduced Clinical Events (REPLACE)-1 trial compared bivalirudin and heparin in patients undergoing percutaneous coronary interventions, and compiled a large prospective dataset of bivalirudin administered concomitantly with planned GPIs. Occurrence of major bleeding was comparable between bivalrudin and heparin in nonoperated patients, but experience with this agent is limited in patients having cardiac procedures [134]. The Acute Catheterization and Urgent Intervention Triage strategY (ACUITY) trial reported improved outcome with decreased bleeding for percutaneous coronary intervention patients who received bivalirudin compared with unfractionated heparin. Trials of bivalirudin for off pump and on pump cardiac surgery showed equivalent 30 day success rates compared with unfractionated heparin [138–140]. Bleeding was not dramatically different, although there appeared to be a tendency for early reexploration of the patients with bivalirudin, as this was an unblinded series of studies. One study from New Zealand showed that graft flow in off-pump CABG patients who received bivalirudin was better than in patients receiving unfractionated heparin for on-pump CABG [135]. Bivalirudin may prove most helpful in patients with heparin induced thrombocytopenia who require coronary revascularization or other cardiac procedures [141].
|
4) Indications for Blood Transfusion—"Transfusion Trigger" |
|---|
|
TopAbstractIntroduction1) Methods Used in...2) Risks and Benefits...3) Causes of Blood...4) Indications for Blood...5) Interventions to Limit...6) Prophylaxis--Interventions...7) Transfusion and Blood...8) Summary Treatment Strategy-...9) SummaryAppendix 1Appendix 22. Conduct of the...3. Peer Review and...4. Recommendations of the...FootnotesReferences |
|---|
Objective evidence to support transfusion decisions is incompletely developed. The hemoglobin, hematocrit, or platelet count at which red cell or platelet transfusion is administered is often referred to as the "transfusion trigger." While surgeons, anesthesiologists, and critical care specialists developed a long-standing tradition of transfusion when the hemoglobin falls below 10 g/dL or hematocrit below 30%, this practice no longer appears appropriate since these recommendations were not evidence based. There is little scientific support for relying upon a specific hemoglobin or hematocrit level as a transfusion trigger. Likewise, estimates of blood loss and intravascular blood volume are often inadequate to assess the needs for transfusion. More advanced measurements such as whole body oxygen-carrying capacity, oxygen consumption, oxygen extraction ratios, and oxygen delivery provide more accurate means to estimate the need for red blood cell transfusions [142].
During cardiac operations, cardiopulmonary bypass provides oxygenation and systemic perfusion in a nonphysiologic manner. Utilizing systemic heparinization, hemodilution and nonpulsatile blood flow in addition to hypothermia, creates a number of physiologic consequences that affect systemic physiology and blood formed elements [143]. With respect to blood transfusion requirements, the use of balanced physiologic solutions as a pump prime, leads to normovolemic hemodilution during cardiopulmonary bypass. The putative advantages of hemodilution during bypass include improvements in tissue perfusion by reducing blood viscosity. In spite of reductions in hemoglobin-bound oxygen transport to tissues, adequate oxygen delivery to tissues is well maintained with hematocrits well below baseline physiologic levels. Hemodilution may contribute to perioperative coagulopathy by diluting the concentration of coagulation factors and platelets. Anemia at the conclusion of cardiopulmonary bypass may theoretically be detrimental to the recovering myocardium in the context of coronary revascularization [144]. However, one study performed in patients immediately after CPB showed no change in myocardial lactate extraction or production even with elective hemodilution to hemoglobin of 5.0 g/dL [145]. Studies done in the early 1980s suggest that hemodilution to hemoglobin of 5.0 g/dL provides adequate oxygen delivery during CPB, but these studies need confirmation in patients with compromised ventricular function and using modern perfusion techniques [146, 147].
The physiologic threshold at which oxygen consumption starts to decrease because of insufficient oxygen delivery in humans is a hemoglobin level of 3 to 4 g/dL, an oxygen extraction ratio of 0.44, a mixed venous oxygen of 34 mm Hg, and a mixed venous oxygen saturation of 56% [148]. Of interest, the lowest limit of critical oxygen delivery of 333 mL O2/min/m2, below which oxygen delivery suffers, is the same for all mammalian species [149]. In well-controlled euvolemic hemodilution studies, whole body shift to anaerobic metabolism (ie, falling below critical oxygen delivery threshold) might be the absolute indicator for transfusion. In the context of coronary artery disease, moderate hemodilution during clinical cardiopulmonary bypass is well tolerated by cardiac patients [146]. However, the coronary artery blood flow reserve and oxygen extraction may be exceeded by extreme hemodilution to less than 6 g/dL [150]. Weisel and colleagues [148] found that patients at risk for perioperative ischemia, presumably most patients having cardiac procedures, showed a delay in myocardial metabolic recovery with substantial hemodilution when postoperative hemoglobin levels fell below 6.0 g/dL. However, transfusion of banked blood was not controlled for in this study. Increases in cardiac output, oxygen extraction ratio, and red cell transit time in capillaries are the primary compensatory mechanisms for reduced oxygen carrying capacity. Patients with compromised ventricular function may not be able to increase cardiac output. There are few data for cardiac surgery patients to address critical end-organ oxygen delivery and utilization in the diseased state, although there are studies that report that tissue oxygenation is maintained in healthy, normovolemic persons with a hemoglobin of 6 to 7 g/dL and hematocrits of 15% to 20% [151, 152]. Fang and colleagues [153] could not find an increased mortality rate among cardiac surgical patients until the perioperative hemoglobin was 5.0 or less. Observational studies provide general guidelines to guide practice, although there are a number of clinical series that provide conflicting data, particularly in specialized patient populations such as the Jehovah’s Witness patients.
The original report by Ott and Cooley [154] of cardiac surgery in Jehovah’s Witness children demonstrated that such high-risk surgery was possible with a mortality rate of 9.4%. This report helped to define the standards for open heart surgery in this group of patients. In reviews of published series between 1983 and 1990 involving 1,404 operations on Jehovah’s Witnesses, anemia and lack of transfusion influenced operative mortality for only 20 patients (1.4%) [155]. A synopsis of 61 reports of 4,722 Jehovah’s Witnesses identified 23 deaths due to anemia, all but 2 of which were patients with hemoglobin concentrations less than 5 g/dL [156]. Based upon these reports, major cardiovascular procedures including coronary bypass surgery, are performed safely in Jehovah’s Witness patients; and these reports provide level B and C evidence that acute and chronic anemia can be tolerated in certain clinical situations.
Because of the lack of high-level evidence to guide transfusion decisions, recommendations rely on expert opinions. In 1988, the National Institutes of Health convened a consensus conference and published recommendations for blood transfusion, platelet therapy, and administration of fresh frozen plasma [157]. Through these activities and a number of other consensus conferences by the American Association of Blood Banks and the American College of Physicians in the early 1990s, recommendations regarding blood utilization and transfusion practices provided significant focus and attention on this issue [158, 159]. In 1990, the Transfusion Practices Committee of the American Association of Blood Banks issued guidelines for transfusion of patients undergoing coronary artery bypass surgery [160]. In 1994, the ASA convened a task force on blood component therapy to develop evidence-based guidelines on the proper indication for perioperative and peripartum administration of red blood cells, platelets, fresh frozen plasma, and cryoprecipitate [161]. This task force reviewed a total of 1,417 articles, of which 160 articles were considered relevant. Published evidence was considered relevant if it addressed the perioperative or peripartum use of the blood components and measured effectiveness in terms of clinical outcomes. Identifying the primary objective of red blood cell transfusion as the improvement of inadequate oxygen delivery, the task force identified two assumptions as the principles upon which decisions to transfuse patients should be based: (1) surgical patients experience adverse outcomes as a result of diminished oxygen carrying capacity, and (2) red blood cell transfusions, by enhancing oxygen-carrying capacity, can prevent adverse outcomes.
After separating the effects of hypovolemia and anemia related to blood loss in this literature, it appears that the absolute lowest threshold for hematocrit or anemia in patients can not be established; however, hemoglobin concentrations as low as 7 g/dL were found to be adequate to provide tissue oxygenation in normal, healthy, euvolemic persons. The NIH Consensus Conference concluded that the evidence did not support the use of a single criterion for transfusion such as a transfusion trigger of 10 g/dL [157]. The summation of consensus recommendations is shown in Table 3.
c) Indications for Transfusion on Cardiopulmonary Bypass
Much has been written regarding the need for improved transfusion indicators in cardiac surgery, particularly during cardiopulmonary bypass [41]. Many reports have tied the decision to transfuse blood components (or the lowest hematocrit on bypass) to postoperative outcome as a means to either confirm the benefits of blood conservation or identify predictors of poor outcome [162, 163]. Among these reports, a number of prospective single-center and multicenter reports offer insight to the management of hemodilutional anemia on cardiopulmonary bypass. In 1997, Fang and colleagues [153] reported the results of a single institution in 2,738 coronary artery bypass graft operations, concluding that after accounting for differences in patient and disease characteristics, a lowest hematocrit value of less than or equal to 14% for low-risk patients and a value of less than or equal to 17% for high-risk patients was an independent risk factor for mortality. In a comprehensive database study from the Northern New England Cardiovascular Disease Study Group, the authors reported that of 6,980 consecutive patients who underwent coronary artery bypass at six medical centers in Maine, New Hampshire, and Vermont and at the Beth Israel Deaconess Hospital in Boston between 1996 and 1998, those whose hematocrit while on cardiopulmonary bypass was allowed to fall below 19% had higher in hospital mortality rates, higher intraoperative use of balloon support, and more frequent return to cardiopulmonary bypass after initial attempts at separation [164]. Women and patients with small body surface area and patients who were anemic before operation were most likely to be severely anemic on cardiopulmonary bypass. Of note, neither of these studies examined transfusion, the usual physician response to low hematocrit, as a covariate or possible cause of the increased mortality.
Several reports suggest worse outcome associated with anemia during CPB. In a recent report evaluating low hematocrit during cardiopulmonary bypass, Karkouti and coworkers [165] evaluated the relationship between nadir hematocrit during cardiopulmonary bypass and perioperative stroke while adjusting for variables known to have an association with stroke and anemia. In prospectively evaluated patients (10,949 consecutive patients) who underwent coronary artery bypass with extracorporeal circulation from 1999 to 2004, nadir hematocrit during cardiopulmonary bypass was an independent predictor of perioperative stroke. After controlling for confounding variables, each percent decrease in hematocrit was associated with a 10% increase in the odds of suffering perioperative stroke. The authors concluded that there is an independent direct association between degree of hemodilution during cardiopulmonary bypass and a risk of stroke.
In a more recent review of the effects of low hematocrit during cardiopulmonary bypass, Habib and coworkers [166] reported a retrospective analysis of operative results and resource utilization in 5,000 consecutive cardiac operations with cardiopulmonary bypass. Stroke, myocardial infarction, low cardiac output, cardiac arrest, renal failure, prolonged ventilation, pulmonary edema, reoperation due to bleeding, sepsis, and multiorgan failure were all significantly increased as lowest hematocrit value decreased below 22%. Consequently, hospital stays, operative costs, and operative deaths were also significantly greater as a function of hemodilution severity. Long-term survival is improved among patients with higher hematocrits on CPB, suggesting that increased hemodilution severity during CPB is associated with worse perioperative outcomes. Like the reports by Fang and coworkers, Habib and associates did not review the effects of transfusion on outcome in their particular study. Two reports [167, 168] examined large databases to evaluate the effects of anemia on renal function during cardiac procedures. There was an association between lowest hematocrit on CPB and the increased incidence of postoperative renal dysfunction. However, the use of transfusion to treat low hemoglobin or to prevent hemodilutional anemia on bypass multiplied the risks of renal dysfunction twofold to 3.5-fold. Because of these findings, some authors suggest that transfusion of blood products worsens outcomes during CPB [140], but transfusion may just be a marker of disease severity, not a cause of poor outcome. More prospective randomized studies are required to amplify the relationship between blood transfusion and poor outcome after CPB.
Because of the possible association of blood transfusion on CPB with worse outcomes, several investigators evaluated this possibility with observational studies. Surgenor and associates [169] showed that although hemodilutional anemia increases the risk of low-output failure after cardiac procedures, an additional risk-adjusted increase of 27% occurs with the transfusion of 1 to 2 units of packed red blood cells regardless of nadir hematocrit. Intraoperative packed red blood cells transfusion during CABG surgery seems to increase the risk of postoperative low-output heart failure. Engoren and coworkers [170] studied 1,915 patients who underwent first-time isolated coronary artery bypass operations between 1994 and 1997 and found that 649 of the study patients (34%) received a transfusion during their hospitalization. Transfused patients were older, smaller, more likely to be female, and had more comorbidity. The transfused patients also had twice the 5-year mortality (15% versus 7%) of nontransfused patients. After correction for comorbidities and other factors, transfusion was still associated with a 70% increase in mortality. By multivariate analysis, transfusion, peripheral vascular disease, chronic obstructive pulmonary disease, New York Heart Association cardiac functional class IV, and age were significant predictors of long-term mortality. The authors concluded that blood transfusions during or after coronary bypass operations are associated with increased long-term morbidity and mortality. Stroke and death were increased in association with the utilization of platelet transfusions in a study by Spiess and associates [167]. However, a more recent study by Karkouti and coworkers [171] did not find an excess morbidity or mortality in CABG patients who received platelet transfusions. Patients who have lower hematocrits bleed more and otherwise receive more transfusions including platelet transfusions. In a review of more than 15,000 patients undergoing CPB procedures at the Cleveland Clinic, a strong association was demonstrated between the use of transfusions and postoperative infections [168].
All the data reviewed above are derived from observational studies and therefore only document associations, not cause and effect. Anemia may well drive a number of physiologic responses as well as physician behaviors (ie, transfusion). It is therefore unclear at this time how important any level of anemia is in creating organ failure or long-term adverse outcomes. Further, it is uncertain what role blood transfusion plays in this complex process. For these reasons, blood transfusions should be administered with caution while adhering to guideline recommendations whenever possible.
d) Postoperative Indications for Transfusion
Decisions about transfusion after operation in ICU patients are complex. It is obvious that the patient’s clinical situation and disease status are important factors in determining the need and indication for transfusion in patients undergoing coronary artery bypass surgery. The patient’s volume status, pulmonary and cardiac status, cerebrovascular status, the chronicity of anemia, the patient’s symptoms, the potential for blood loss, and the extent of surgery and risk of rebleeding all factor into these decisions. Clinical indicators of hypovolemia and its secondary physiologic effects include tachycardia, hypotension, and oliguria. Physiologic indicators of significant impairments of the oxygen supply and demand ratios include mixed venous oxygen saturation (SVO2) less than 55%, or mixed venous oxygen tension less than 30 mm Hg.
Data to support postoperative transfusion decisions in cardiothoracic practices are sparse. Utilizing clinical judgment in the decision-making process, many clinicians assume that anemia increases the risk of myocardial ischemia after CABG operations, although the previously reviewed experience with Jehovah’s Witness patients and other clinical reports offers a contrary opinion [155, 172]. Clinical indicators for postoperative transfusion in coronary artery bypass graft patients taking into account their clinical and disease status were proposed in 1990, and updated in 1996, to provide guidance with respect to indications for transfusion [173, 174]. In the absence of prospective randomized data, these expert recommendations provide a foundation upon which transfusion practices are based and still appear to be generally applicable to modern cardiothoracic practice (Table 3).
More recent assessments of the impact of transfusion therapies are available. In a retrospective data analysis, Hébert and associates [175] found that patients with cardiac disease admitted to the critical care unit had a significantly higher risk of death with lower hemoglobin values. To further elucidate the potential impact of transfusions on mortality in critically ill patients admitted to a general ICU, the same investigators conducted a prospective, randomized, controlled trial [176] to determine the outcome of a transfusion protocol that maintained hemoglobin level at 7 to 9 g/dL (strict protocol) or at 10 g/dL or more (liberal protocol). They reported that there was no overall difference in mortality between the two groups and that a restrictive protocol was at least as effective and possibly superior to the liberal transfusion protocol. In a subset of patients who had known coronary artery disease, mortality was not different based upon transfusion behavior [177].
In a report from the Multicenter Study of Perioperative Ischemia Research Group, Spiess and coworkers [178] reported that the risk of postoperative myocardial infarction was highest in CABG patients with hematocrit greater than 34% and with more severe left ventricular dysfunction, questioning the rationale for adhering to arbitrary transfusion thresholds in cardiac patients.
Based upon nonrandomized observational studies, descriptive case series, a few prospective randomized clinical trials, and expert panel opinions, consensus suggests that red blood cell transfusion to improve oxygen transport when a hemoglobin level is greater than 10 g/dL is almost never of benefit. The NIH Consensus Conference and the practice guidelines of the ASA concluded that patients with hemoglobin levels greater than 10 g/dL did not require blood, whereas most patients with hemoglobin levels less than 7 g/dL benefit from transfusion [157, 161]. In a recent prospective randomized trial of transfusion of patients after CABG, in-dwelling tissue oxygen probes measured the effect of transfusion with 1 or 2 units of red blood cells [179]. Transfusion did not improve oxygen delivery to tissues, but increasing FiO2 from 50% to 100% did improve tissue oxygen delivery. Other animal work confirmed this finding [180, 181]. Such work in hemorrhagic rat models did show that vital signs (arterial pressure, central venous pressure) and blood gases (systemic arterial and mixed venous) improved by transfusion of banked blood. With standard critical monitoring, transfusion seems to improve physiologic variables (blood pressure, arterial blood gases, mixed venous blood gases), but the tissue oxygen delivery may actually be reduced by typical banked blood [181]. Much more work needs to be done before the physiologic benefit of blood transfusion is determined. Only then can realistic and evidence-based transfusion triggers be established.
|
5) Interventions to Limit Blood Transfusion |
|---|
|
TopAbstractIntroduction1) Methods Used in...2) Risks and Benefits...3) Causes of Blood...4) Indications for Blood...5) Interventions to Limit...6) Prophylaxis--Interventions...7) Transfusion and Blood...8) Summary Treatment Strategy-...9) SummaryAppendix 1Appendix 22. Conduct of the...3. Peer Review and...4. Recommendations of the...FootnotesReferences |
|---|
Three drugs that play a role in perioperative blood conservation and that limit fibrinolysis are in clinical use. Epsilon-aminocaproic acid (EACA [Amicar; Xanodyne Pharmaceuticals, Newport, KY]) inhibits fibrinolysis by the inhibition of plasminogen inhibitors and to a lesser degree through antiplasmin activity. Tranexamic acid (TXA [Cyklokapron; Pharmacia & Upjohn, Somerset County, NJ]) is similar in action to epsilon-aminocaproic acid but it is approximately 10 times more potent. These two drugs are classified as lysine analogues because they inhibit plasminogen by binding to the lysine binding sites on the plasminogen molecule. Aprotinin (Trasylol; Bayer Pharmaceuticals, West Haven, CT) is a bovine protein that inhibits proteases with active serine residues, especially plasmin, resulting in the attenuation of inflammatory responses, fibrinolysis, and thrombin gerenation. Aprotinin and the lysine analogues have very different modes and scope of action but ultimately inhibit fibrinolysis by limiting the action of plasmin.
Since the early 1990s, antifibironlytic therapies have been widely adopted for cardiac surgery. However with wide-spread adoption, there is much concern regarding the saftety and efficacy of these drugs (http://www.fda.gov/c
