The use of cardiopulmonary bypass (CPB) during cardiac surgical procedures causes a significant disruption of the coagulation system that may contribute to a coagulopathy of varying degrees.1 In addition to hemodilution from a crystalloid prime, which reduces levels of clotting factors and platelets, contact of blood with the extracorporeal circuit activates platelets and the extrinsic and intrinsic coagulation systems and triggers fibrinolysis. In fact, systemic heparinization alone causes platelet dysfunction and induces fibrinolysis.2 In addition, cell‐saving devices that are routinely used for red cell salvage eliminate platelets and coagulation factors from the blood.
Off‐pump coronary artery bypass surgery (OPCAB) avoids hemodilution, minimizes platelet activation, and reduces usage of blood products.3 The ability of the antifibrinolytic agents to reduce bleeding during these procedures suggests that low‐grade fibrinolysis is still present.4 Although a coagulopathy after OPCAB is very unusual, it may occur in patients who have sustained substantial blood loss with blood scavenged in and returned from the cell‐saving device. This will result in depletion of coagulation factors and platelets. The occurrence of substantial bleeding after an OPCAB procedure generally indicates a surgical source.
Either 28–32 Fr PVC or silicone malleable chest tubes or 24 Fr silicone fluted (Blake) drains are placed in the mediastinum and opened pleural cavities at the conclusion of surgery. They are connected to a drainage system and placed to −20 cm of H2O suction. They are gently milked or stripped to maintain patency after surgery. Both are equally effective in evacuating blood, although the Blake drains may be more comfortable for the patient.5,6
Some surgeons do not obligatorily place chest tubes into widely opened pleural spaces, especially after off‐pump surgery. However, any bleeding that occurs in the pleural space will tend to accumulate and not be drained by the mediastinal tubes. This can produce a deceptive picture with insidious bleeding that can only be detected by chest x‐ray.
Following minimally invasive surgery, the number and location of tubes may vary. After MIDCABs, only one pleural chest tube is placed, so blood could potentially accumulate around the heart and not be drained through the pericardial opening. Following ministernotomy incisions, one mediastinal tube is placed unless the pleural cavity is entered. With right thoracotomy approaches to the aortic or mitral valve, one mediastinal and one pleural tube are placed. Chest tube positioning is difficult and not ideal after these procedures, so the potential for undetected blood accumulation around the heart or in the pleural spaces is enhanced. Thus, extra vigilance for undrained blood in the unstable patient is imperative.
Postoperative bleeding gradually tapers over the course of several hours in the majority of patients, but about 1–3% of patients will require re‐exploration for persistent mediastinal bleeding. Prompt assessment and aggressive treatment in the intensive care unit (ICU) may frequently arrest “medical bleeding”, but evidence of persistent or increasing amounts of bleeding should prompt early exploration (see section VIII, pages 446–449).
Persistent mediastinal bleeding invariably requires the use of various blood products to maintain normovolemia and adequate hemodynamic parameters, correct anemia to ensure adequate tissue oxygen delivery, and correct a coagulopathy to help arrest the bleeding. Understandably, increased chest tube drainage in the early postoperative period is associated with adverse outcomes. One study showed that drainage of >200 mL/h in one hour or >2 mg/kg for two consecutive hours was associated with an increased risk of stroke, re‐exploration, prolonged ventilation, and mortality.7 The need for transfusions and re‐exploration are both independent risk factors for morbidity and mortality.8
Transfused blood is not benign, contributing to increased postoperative morbidity, including graft occlusion, pulmonary, infectious, neurologic, renal, and gastrointestinal complications, as well as increased short‐term and long‐term mortality.9–13 Although a “restrictive” transfusion strategy (transfusion threshold of a hemoglobin of 7–8 g/dL or a hematocrit [HCT] of 21–24%) might be safe in stable patients,14–16 hemodynamic considerations and potential impairment of tissue oxygen delivery with ongoing bleeding mandate transfusions to maintain a safe HCT, which is probably at least 24%. Blood component therapy ideally should be selected based upon identification of specific coagulation abnormalities by point‐of‐care testing and treatment algorithms, although clinical judgment remains essential in making prompt and appropriate therapeutic decisions.17,18
Mediastinal bleeding can be a highly morbid and lethal problem. Although hypo-volemia can be corrected by volume infusions, the bleeding patient tends to be hemodynamically unstable out of proportion to the degree of bleeding and fluid replacement. Most importantly in the immediate postoperative period is the potential for blood to accumulate around the heart, causing cardiac tamponade. The restriction to cardiac filling may produce severe hemodynamic compromise that can precipitously cause cardiac arrest. Constant attention to the degree of bleeding and to trends in hemodynamic parameters should allow for steps to be taken to avert this problem. If profound hypotension or a cardiac arrest develops, emergency sternotomy in the ICU is indicated.
II. Etiology of Mediastinal Bleeding
Mediastinal bleeding is somewhat arbitrarily categorized as “surgical” or “medical” in nature (Table 9.1). Significant bleeding after uneventful surgery is usually “surgical”, especially when initial coagulation studies are fairly normal. However, persistent bleeding depletes coagulation factors and platelets, causing a coagulopathy that is self‐perpetuating. Bleeding that is noted after complex operations with long durations of CPB is frequently associated with abnormal coagulation studies and is considered “medical”. However, even after correction of coagulation abnormalities, discrete bleeding sites may be present that will not stop without re‐exploration. Most studies suggest that a surgical source is identified in about two‐thirds of patients who are re‐explored.19 Thus, the initial approach to bleeding is to try to identify any contributing factors that might account for the degree of bleeding and then take the appropriate steps to correct them.
A number of risk factors have been identified that increase perioperative bleeding and/or the requirement for transfusions (Table 9.2).14,20–22 One risk stratification model found an excellent correlation of only five factors with the risk of significant bleeding, which included nonelective surgery, more than a CABG or isolated valve operation, presence of aortic valve disease, a body mass index (BMI) >25, and age >75.23 Aside from stopping antiplatelet or anticoagulant medications preoperatively, most risk factors cannot be modified. However, they should alert the healthcare team to the increased risk of a coagulopathy, the necessity of utilizing blood conservation measures, and the importance of early aggressive treatment of bleeding to minimize or prevent hemodynamic compromise and organ system dysfunction.
Surgical bleeding is usually related to:
Anastomotic sites (suture lines)
Side branches of arterial or venous conduits
Substernal soft tissues, sternal suture sites, bone marrow, periosteum
Raw surfaces caused by previous surgery, pericarditis, or radiation therapy
Residual anticoagulant effects of medications taken preoperatively, which include:
Unfractionated heparin that is continued into the operating room
Low‐molecular‐weight heparin (LMWH) given within 18 hours
Nonvitamin K antagonist oral anticoagulants (NOACs), including dabigatran, apixaban, rivaroxaban, and edoxaban when given within 48 hours or even longer in patients with chronic kidney disease
Intravenous direct thrombin inhibitors (bivalirudin)
Indirect factor Xa inhibitors (fondaparinux)
Qualitative platelet defects – a major concern with the liberal use of antiplatelet agents in patients with acute coronary syndromes and recently placed stents.14,24,25
Preoperative platelet dysfunction may result from antiplatelet medications (aspirin and P2Y12 inhibitors [clopidogrel, ticagrelor, prasugrel]), glycoprotein IIb/IIIa inhibitors (tirofiban, eptifibatide, abciximab), herbal medications and vitamins (fish oils, ginkgo products, vitamin E), or uremia.
Exposure of platelets to the CPB circuit with alpha‐granule release and alteration of platelet membrane receptors impairs platelet function. The degree of platelet dysfunction correlates with the duration of CPB and the degree of hypothermia after bypass.
Inadequate heparinization is a potent trigger for thrombin release, which activates platelets.
Quantitative platelet defects
Preoperative thrombocytopenia may result from use of heparin, drug reactions (P2Y12 inhibitors including clopidogrel and prasugrel, but not ticagrelor, antibiotics and IIb/IIIa inhibitors), infection, hypersplenism in patients with liver disease, and other chronic conditions (idiopathic thrombocytopenic purpura [ITP]). If a patient developing thrombocytopenia has recently been given heparin, it is essential to rule out heparin‐induced thrombocytopenia (HIT).
Hemodilution on CPB and consumption in the extracorporeal circuit reduce the platelet count by about 30–50%, and thrombocytopenia will be progressive as the duration of CPB lengthens.
Protamine administration transiently reduces the platelet count by about 30%.
Intraoperative heparin and protamine usage
Residual heparin effect may result from inadequate neutralization with protamine at the conclusion of CPB. Administering fully heparinized “pump” blood towards the end of the protamine infusion will reintroduce unneutralized heparin into the blood. Blood washed in cell‐saving devices is usually given after protamine administration, but it has been shown to contain insignificant amounts of heparin.26
Heparin rebound may occur when heparin reappears from tissue stores after protamine administration. This is more common in patients receiving large amounts of heparin, especially obese patients.
Excessive protamine may cause a coagulopathy.14,27,28
Depletion of coagulation factors
Preoperative hepatic dysfunction, residual warfarin effect with vitamin K‐dependent clotting factor deficiencies, and thrombolytic therapy reduce the level of clotting factors. Increased bleeding may specifically be related to a low fibrinogen level.29
von Willebrand’s disease caused by a deficiency in von Willebrand’s factor, a clotting protein that binds to factor VIII and platelets to form a platelet plug. An acquired form is associated with aortic stenosis.30
Hemodilution on CPB reduces most factors by 50%, including fibrinogen. This is most pronounced in patients with a small blood volume.
Loss of clotting factors results from use of intraoperative cell‐saving devices.
Fibrinolysis results in clotting factor degradation and platelet dysfunction.
Preoperative use of thrombolytic agents causes fibrinolysis.
Use of CPB causes plasminogen activation.
Heparinization itself induces a fibrinolytic state.
Transfuse patients requiring urgent surgery to a HCT >26% preoperatively
Use antifibrinolytic therapy (ε‐aminocaproic acid or tranexamic acid)
Consider off‐pump coronary bypass grafting, if feasible
Autologous blood withdrawal prior to CPB if HCT >30% with plateletpheresis
Use heparin‐coated (biocompatible) circuit, if available
Use miniaturized CPB circuit, if available
Use heparin–protamine titration test to optimize anticoagulation and heparin reversal
Consider retrograde autologous priming of the bypass circuit
Avoid more than mild systemic hypothermia
Avoid use of cardiotomy suction
Salvage pump blood via either hemofiltration or cell saver
Employ meticulous surgical technique with careful inspection of anastomotic sites and all artery and vein side branches before coming off bypass
Complete neutralization of heparin with protamine based on heparin levels to return ACT to baseline
Administer appropriate blood component therapy based upon suspicion of the hemostatic defect (especially platelet dysfunction) or use point‐of‐care testing to direct blood component therapy
Use recombinant factor VIIa for intractable coagulopathic bleeding
Preoperative assessment of the patient’s coagulation system should entail measurement of a prothrombin time, as measured by the international normalized ratio (INR), partial thromboplastin time (PTT), and platelet count. Any abnormality should be investigated and corrected, if possible, prior to surgery. Platelet function testing to assess platelet responsiveness to P2Y12 inhibitors is helpful in determining when the bleeding risk is low enough to proceed with urgent, but not emergent, surgery.31
Considerations in patients with preoperative anemia
Preoperative anemia is associated with increased morbidity and mortality after cardiac surgery, being most likely a surrogate for clinical disease and increasing the requirement for transfusions.32 Preoperative iron supplementation and use of erythropoietin should be considered in anemic patients undergoing elective surgery, especially Jehovah Witness patients, who refuse any blood transfusions (see doses on page 197 in Chapter 3).
Hemodilution during CPB can produce a profound anemia in patients with a low preoperative HCT. This may increase the risk of renal dysfunction if the HCT on CPB is much less than 21%,33,34 and can contribute to neurologic events, such as ischemic optic neuropathy.35,36 Although preoperative transfusion for a HCT <26% will reduce the need for intraoperative transfusions, it is controversial as to whether there is any clinical benefit to this approach. However, patients who require multiple blood transfusions during surgery tend to be more coagulopathic and require additional blood component therapy, providing some justification for a preoperative transfusion strategy.
Heparin‐induced thrombocytopenia (HIT) may develop in patients receiving intravenous heparin for several days before surgery. Thus, it is very important to recheck the platelet count on a daily basis in these patients. If the patient develops thrombocytopenia, with documented heparin antibodies by ELISA testing and a positive functional assay (serotonin release assay or heparin‐induced platelet aggregation test), an alternative means of anticoagulation will be necessary during surgery (see pages 251–253).
Cessation of medications with antiplatelet or anticoagulant effects is essential to allow their effects to dissipate to minimize blood loss.14,22 A more detailed discussion of these medications is presented in Chapter 3 (pages 186–193). Specific recommendations are as follows:
Warfarin should be stopped 4–5 days before surgery to allow for the resynthesis of vitamin‐K‐dependent clotting factors and normalization of the INR. Bridging anticoagulation with either LMWH or unfractionated heparin is indicated in patients at high thromboembolic risk, including those with a mechanical valve, recent pulmonary embolism (<4 weeks), and atrial fibrillation with rheumatic mitral stenosis or a CHA2DS2‐VASc Score >4.14 Otherwise, bridging is not indicated.
If the patient requires urgent surgery, vitamin K should be given to normalize the INR. A slow IV infusion of 5 mg over 30 minutes is effective in promptly correcting the INR, but it is preferable to give 5 mg of oral vitamin K if surgery can be delayed a day or two to avoid the risk of anaphylaxis.
If emergency surgery is indicated, fresh frozen plasma (FFP) or prothrombin complex concentrate (PCC) may be given. PCC is more effective and expeditious in reducing the INR and is the preferred product.37,38 The recommended doses are 25 units/kg for an elevated INR <4, 35 units/kg for an INR of 4–6, and 50 units/kg for an INR >6.
Factor eight inhibitor bypassing activity (FEIBA) is an activated four‐factor PCC. It is an anti‐inhibitor coagulant complex that contains nonactivated factors II, IX, X, activated factor VII and factor VIII inhibitor bypassing activity, and some factor VIII coagulant antigen. It has been used to control warfarin‐related coagulopathic bleeding and thus could be used to achieve rapid reversal of warfarin prior to emergency surgery. It can normalize the INR within an hour, but may be associated with thrombotic events.39
Unfractionated heparin (UFH) is used for patients with acute coronary syndromes, during catheterization, for critical coronary disease, or during use of an intra‐aortic balloon pump (IABP). It can be continued up to the time of surgery without increasing morbidity during line placement or increasing the risk of perioperative bleeding. A common practice is to stop it four hours in advance of surgery for patients at low risk for instability.
Low‐molecular‐weight heparin (LMWH) is given in a dose of 1 mg/kg SC q12h for acute coronary syndromes or as a bridge to surgery once warfarin has been stopped. LMWH has a half‐life of approximately four hours, so the last dose should be given 24 hours prior to surgery to minimize the perioperative bleeding risk, since only 60–80% of LMWH is reversible with protamine.
Aspirin (ASA) should be continued up to the time of surgery in patients undergoing bypass surgery, but it may be stopped 3–5 days prior to noncoronary surgery. Because an 81 mg dose is usually not associated with an increased risk of bleeding and has arguably lowered the risk of infarction and mortality, it can be recommended before all coronary bypass operations.40–42 Antifibrinolytic drugs, which should be routinely used, may reduce bleeding associated with preoperative use of aspirin and possibly P2Y12 inhibitors.43–45
The P2Y12 inhibitors are commonly given to patients presenting with an acute coronary syndrome and are generally recommended for at least a year in patients undergoing percutaneous coronary intervention (PCI) with drug‐eluting stents to avoid stent thrombosis. They exhibit antiplatelet effects that last for the lifespan of the platelet, which is 5–7 days, and an increased risk of bleeding is noted when they are stopped for shorter periods of time prior to surgery.46 It is generally recommended that ticagrelor be stopped 3–5 days, clopidogrel 5 days, and prasugrel 7 days before surgery.14,22,47 Use of PRU testing, which is an abbreviation for both platelet reactivity units and P2Y12 reaction units, may reveal patient resistance to P2Y12 inhibitors, allowing for safer earlier surgery than the recommended five‐day interval.31,48 If surgery is required on an emergent or urgent basis, significant bleeding may be encountered, and exogenously administered platelets may be ineffective if given within several hours of a dose of a P2Y12 inhibitor, because the active metabolite may still be present in the bloodstream. Nonetheless, the urgency of surgery to prevent an ischemic event always takes precedence over the time interval from cessation of a P2Y12 inhibitor. If a patient requires urgent but not emergent surgery after receiving one of these drugs or is taking them for recent stent placement, feasible options are to:
Continue the P2Y12 inhibitor and accept the potential for more bleeding.
Stop it for three days to restore some platelet function while maintaining a lesser degree of platelet inhibition.
Stop it for 5 days and use a short‐acting glycoprotein IIb/IIIa inhibitor for a few days as a bridge to surgery.
Nonvitamin K antagonist (or novel) oral anticoagulants (NOACs), also referred to as direct oral anticoagulants (DOACs), are generally recommended for nonvalvular atrial fibrillation or deep venous thrombosis. However, they are commonly used off‐label for patients with AF and structural heart disease exclusive of rheumatic mitral stenosis, although some studies do suggest an excellent safety profile and clinical benefit even in patients with mitral stenosis.49 NOACs should not be used for patients with mechanical heart valves.50 Since the dissipation of anticoagulant effects generally takes 4–5 half‐lives and NOACs have half‐lives averaging around 12 hours, the last dose of these medications should be taken 48 hours prior to surgery and stopped even earlier if the patient has renal dysfunction.51 If the patient requires more urgent surgery, other products can be used to offset their anticoagulant effects.52
Idarucizumab (Praxbind) 5 g IV can reverse dabigatran
Andexanet alfa (recombinant factor Xa) can be used to reverse the effects of apixaban and rivaroxaban.53 A recommended low‐dose protocol is 400 mg IV given at a target infusion rate of 30 mg/min, to be followed by 4 mg/min for 120 minutes. This is recommended if the patient had been taking apixaban ≤5 mg or rivaroxaban ≤10 mg. The high‐dose protocol is twice that dose at the same infusion rates if the patient had been taking >5 mg of apixaban or >10 mg of rivaroxaban. This product prevents factor Xa inhibition for about one hour after the infusion has been stopped.
Antibody‐based targeted therapy for these medications is on the horizon.54
FEIBA can be used to minimize bleeding associated with NOACs if emergency surgery is required. It is usually given in a dose of 20–30 units/kg.55
In the absence of availability of the above products, PCC can be given in a dose of 25–50 units/kg.
Fondaparinux, a factor Xa inhibitor, has a half‐life of nearly 20 hours and should be stopped at least 60 hours before surgery, although there are reports that 24 hours should suffice.14
Tirofiban (Aggrastat) and eptifibatide (Integrilin) are short‐acting IIb/IIIa inhibitors that allow for recovery of 80% of platelet function within 4–6 hours of being discontinued. They should be stopped about four hours prior to surgery. Some studies have shown that continuing these medications up to the time of surgery may preserve platelet function on pump, leading to increased platelet number and function after CPB with no adverse effects on bleeding.56
Abciximab (ReoPro) is a long‐acting IIb/IIIa inhibitor used for high‐risk PCI that has a half‐life of 12 hours. If surgery needs to be performed on an emergency basis, platelets are effective in producing hemostasis, since there is very little circulating unbound drug. Ideally, surgery should be delayed for at least 12 hours and preferably 24 hours. Although platelet function remains abnormal for up to 48 hours, there is little hemostatic compromise at receptor blockade levels less than 50%.
Intravenous direct thrombin inhibitors are primarily used in patients with HIT, but bivalirudin has been used as an alternative to UFH in patients undergoing PCI. It has a short half‐life of 25 minutes and should not pose a significant issue if emergency surgery is required. Its use as an alternative to heparin during surgery in patients with HIT has been associated with comparable outcomes, although bleeding tends to be more problematic.57
Thrombolytic therapy is an alternative to primary PCI for patients presenting with ST‐elevation myocardial infarctions (STEMIs) in centers without PCI capability. Although most agents have short half‐lives measured in minutes, the systemic hemostatic defects persist much longer. These effects include depletion of fibrinogen, reduction in factors II, V, and VIII, impairment of platelet aggregation, and the appearance of fibrin split products. If surgery is required for persistent ischemia after failed thrombolytic therapy, it should be delayed for at least 12–24 hours. If it is required emergently, blood component therapy with FFP, PCC, and/or cryoprecipitate will probably be necessary to correct the anticipated coagulopathy.
Antifibrinolytic therapy with lysine analogues should be used to reduce intraoperative blood loss in all on‐ and off‐pump surgical cases (see doses on page 249).14,43–45
ε‐aminocaproic acid (Amicar) is an antifibrinolytic agent that preserves platelet function by inhibiting the conversion of plasminogen to plasmin. It is effective in reducing blood loss and the amount of transfusions, although it has not been shown to reduce the rate of re‐exploration for bleeding. Because of its low cost, it is usually the drug of choice for most cardiac surgical procedures.
Tranexamic acid (Cyklokapron) has similar properties and benefits to ε‐aminocaproic acid, with more clinical evidence of benefit in reducing perioperative blood loss in both on‐ and off‐pump surgery. One study found that topical tranexamic acid (2 g/500 mL NS) poured into the pericardial cavity just prior to closure significantly reduced blood loss and the need for transfusions.58
Heparin and protamine dosing
Ideal anticoagulation for CPB should minimize activation of the coagulation cascade, be fully reversible, and minimize perioperative bleeding. The most commonly used drug is heparin, which binds to antithrombin to inhibit thrombin and factor Xa. Empiric dosing of heparin (3–4 mg/kg) to achieve an activated clotting time (ACT) >480 seconds has been recommended, although patients with antithrombin deficiency may be heparin‐resistant and require FFP or antithrombin (Thrombate) to achieve a satisfactory ACT.59 Inadequate heparin dosing increases thrombin generation, which in turn activates platelets and can trigger clotting within the CPB circuit. Lower doses of heparin may be used in biocompatible circuits with an acceptable ACT most likely being 400 seconds, although this has not been well defined. A heparin level of 2 units/mL should be achieved.
Systems that provide heparin–protamine titration tests to measure circulating heparin concentrations and determine dose‐response curves are recommended in order to optimize heparin dosing. Use of these systems may result in more heparin being administered, but usually a lower dose of protamine being necessary to achieve reversal, which is based on heparin levels at the conclusion of CPB. The end result is generally less thrombin and platelet activation, a reduction in fibrinolysis, and a reduction in perioperative bleeding, whether a higher or lower dose of heparin is used than predicted by empiric dosing.14,60
The major advantage of heparin is that its anticoagulant effect can be reversed with protamine. In contrast, other effective anticoagulants that can be used for CPB, such as bivalirudin in HIT patients, are not reversible.
Protamine is usually given in a 1:1 ratio to the dose of heparin. However, using dose‐response curves, lower doses of protamine usually suffice to adequately reverse heparin. This may result in less bleeding, because excessive protamine serves as an anticoagulant that directly impairs platelet function and elevates the ACT.14,60,61
Perfusion considerations to optimize blood conservation include the following (see also Chapter 5):14,22,62
Autologous blood withdrawal before instituting bypass (acute normovolemic hemodilution) protects platelets from the damaging effects of CPB. This blood remains of high quality despite storage during the surgery, preserves red cell mass, and reduces transfusion requirements. However, its efficacy in reducing perioperative bleeding is controversial.63,64 It can be considered when the calculated on‐pump HCT after withdrawal remains satisfactory (>20–22%). This can be calculated using the following equation:
Platelet‐rich plasmapheresis entails the withdrawal of platelet‐rich plasma using a plasma separator at the beginning of the operation with its re‐administration after protamine infusion. This improves hemostasis and reduces blood loss. Although it might be beneficial in reoperations, it is expensive, time‐consuming, and probably of marginal benefit.65,66
The use of biocompatible circuits for CPB (usually heparin‐bonded) may reduce activation of platelets and the coagulation cascade with a subsequent reduction in blood loss.14 These systems usually allow for use of lower doses of heparin to achieve a target ACT of 400 seconds. Despite using lower doses of heparin, which could theoretically increase thrombin generation, there is little evidence that this contributes to a prothrombotic milieu.
Avoidance of cardiotomy suction may reduce perioperative bleeding. Blood aspirated from the pericardial space has been in contact with tissue factor, contains high levels of factor VIIa, procoagulant particles, fat particles, and activated complement proteins, and exhibits fibrinolytic activity.22,67 It may be associated with thrombin generation and cause platelet activation resulting in more bleeding. Blood aspirated with cardiotomy suckers drains into a reservoir and mixes directly with the pump blood that is reinfused through the CPB circuit. Most groups use cardiotomy suction routinely and do not find that it has a significant effect on bleeding.
Miniaturized CPB circuits require low priming volumes (500–800 mL) that limit the degree of hemodilution, thus maintaining a higher HCT on pump. Studies have arguably demonstrated that these systems reduce activation of coagulation and fibrinolysis and minimize blood loss. However, the lack of a cardiotomy reservoir increases the risk of air embolism.68
Retrograde autologous priming of the extracorporeal circuit entails initial withdrawal of crystalloid prime to minimize hemodilution, thus maintaining a higher HCT and colloid oncotic pressure on pump. This also reduces extravascular lung water. In some studies, this has been shown to reduce the rate of transfusion.69–71
Intraoperative autotransfusion of blood that is aspirated from the field into a cell‐saving device is recommended as a routine means of salvaging red blood cells whether cardiotomy suction is used or not. It is most helpful in salvaging red cells from dilute fluids (e.g. after cold saline is poured on the heart during cardioplegic arrest). Cell salvage of pump contents at the conclusion of CPB is routinely performed as well. The cells are centrifuged and washed to remove heparin and cytokines and concentrate red cells, but the washing results in loss of coagulation factors and platelets from the blood. Most studies suggest that routine intraoperative cell salvage reduces transfusion requirements, but large reinfusion volumes (>1000 mL) may impair coagulation.14,72 The routine use of ultrafiltration to remove the pump prime is not recommended, except in patients with large pump primes relative to their blood volume.22
Meticulous surgical technique is the mainstay of hemostasis. Warming the patient to normothermia before terminating bypass improves the function of the coagulation system.
IV. Assessment of Bleeding in the ICU
The appropriate assessment of bleeding in the ICU requires the following steps (Table 9.4):73
Frequent documentation of the amount of blood draining into the collection system and attention to tube patency.
Determination of the color (arterial or venous) and pattern of drainage (sudden dump when turned or continuous drainage).
Monitoring of hemodynamic parameters with ongoing awareness of the possibility of cardiac tamponade.
Identification of potential causative factors by review of coagulation studies.
Suspicion of undrained blood in the mediastinum or pleural spaces by review of a chest x‐ray (looking for a widened mediastinum or haziness in the pleural cavity as blood layers posteriorly), auscultating decreased breath sounds on examination, or noting elevation of peak inspiratory pressures on the ventilator.
Obtaining an echocardiogram if tamponade is suspected based upon the pattern of bleeding, hemodynamic derangements, or abnormalities on chest x‐ray.
Quantitate the amount of chest tube drainage. Make sure that the chest tubes are patent, because the extent of ongoing hemorrhage may be masked when the tubes have clotted or blood has drained into an open pleural space. Note: when patients are turned or moved, they will occasionally drain a significant volume of blood that has been accumulating in the chest for several hours. This may suggest the acute onset of bleeding and the need for surgical exploration. The presence of dark blood and minimal additional drainage are clues that this does not represent active bleeding. Serial chest x‐rays may be helpful in identifying residual blood in the pleural space.
The amount of bleeding that warrants treatment is somewhat arbitrary. Some have defined “active bleeding” as bleeding occurring at a rate exceeding 1.5 mL/kg/h for six consecutive hours (which would correspond to >100–150 mL/h for the average adult).74 Other studies suggest that a rate of bleeding exceeding 2 mg/kg/h for three hours (corresponding to >150–200 mL/h for the average adult) is associated with adverse outcomes.23 Yet another study considered significant bleeding to be a bleeding rate of >3 mL/kg/h (about 250–300 mL/h) for more than two consecutive hours.75 It may be inferred that persistent bleeding that exceeds 200 mL/h for several hours requires specific treatment. It is worth remembering that there may be more bleeding occurring than is evident from the amount of chest tube drainage.
A “universal definition” of perioperative bleeding was proposed in 2014.76 This defined five classes of bleeding, based on chest tube output, blood and blood component therapy given, the need for delayed sternal closure, and the need for re‐exploration. By evaluating the total amount of drainage in 12 hours, moderate bleeding was defined as 801–1000 mL/12h, severe was 1001–2000 mL/12h and massive was >2000 mL/12h. This categorization was retrospective and is not that helpful in making treatment decisions, since it is not predictable whether significant early postoperative bleeding will taper without treatment.
Assess hemodynamics with the Swan‐Ganz catheter. Maintenance of adequate filling pressures and cardiac output is essential and is generally accomplished using crystalloid or colloid solutions. However, in the bleeding patient, these will produce hemodilution and progressive anemia, and may potentiate a coagulopathy. It should be noted that unstable hemodynamics are frequently seen in the bleeding patient even if filling pressures are maintained.
If filling pressures are decreasing and nonheme fluid is administered, one needs to anticipate a decrease in the HCT from hemodilution, but more so with ongoing bleeding. Five percent albumin will have a dilutional effect on clotting factors and the HCT, but it is the preferred colloid for volume expansion (other than blood or blood component therapy). One should avoid use of any hetastarch (HES) compounds in the bleeding patient. The high‐molecular‐weight HES‐based compounds adversely affect fibrin formation and platelet function (perhaps slightly less with HES in balanced electrolyte solution [Hextend] than HES 6% in saline [Hespan]).77,78 Although the low‐molecular‐weight HES compounds, such as pentastarch and tetrastarch, appear to have minimal effect on coagulation, they may still be associated with postoperative bleeding and should probably be avoided.79
The administration of volume in the form of clotting factors and platelets promotes hemostasis but must be accompanied by red cell transfusions to maintain a safe HCT. Anemia not only reduces oxygen‐carrying capacity of the blood but also reduces blood oncotic pressure and viscosity, which contribute to hypotension.
Evidence of rising filling pressures and decreasing cardiac outputs may suggest the development of cardiac tamponade. Equilibration of intracardiac pressures may be noted with postoperative tamponade, but, more commonly, accumulation of clot adjacent to the right or left atrium or either ventricle will produce variable elevation in intracardiac pressures that may also be consistent with right (RV) or left ventricular (LV) failure, respectively.
If hemodynamic measurements suggest borderline cardiac function and tamponade cannot be ruled out, transesophageal echocardiography (TEE) is invaluable in making the correct diagnosis. Tamponade should be suspected when hemodynamic compromise is associated with excessive bleeding, bleeding that has abruptly stopped, or even minimal chest tube drainage caused by clotted tubes or spillage into the pleural space. Although a transthoracic study (TTE) may identify a large effusion, acoustic windows are often not ideal, especially with dressings, ECG leads, chest tubes, and tape on the chest. TEE is more helpful in detecting clot behind the heart and should be considered if the TTE is nondiagnostic.80
Coagulation studies should be obtained after administration of protamine in the OR, and abnormal studies associated with “nonsurgical” coagulopathic bleeding should be promptly treated prior to chest closure. Upon arrival in the ICU, these studies are usually repeated, and serial HCTs should then be obtained if the patient is bleeding. If studies are abnormal, but the patient has an insignificant amount of mediastinal bleeding, use of blood component therapy is not indicated.
If hemostasis was difficult to achieve in the OR or hemorrhage persists in the ICU, lab tests may be helpful in assessing whether a coagulopathy is contributing to mediastinal bleeding. Tests for some of the more common nonsurgical causes of bleeding (residual heparin effect, thrombocytopenia, and clotting factor deficiency [primarily fibrinogen levels]) are readily available, but documentation of platelet dysfunction and fibrinolysis requires additional technology. Although no individual test correlates that well with the amount of bleeding, together they can usually direct interventions in a somewhat scientific manner.
No matter what the results of coagulation testing are, clinical judgment remains paramount in trying to ascertain whether the bleeding is more likely to be of a surgical nature (which tends to persist) or due to a coagulopathy (which might improve). If normal coagulation studies are present upon arrival in the ICU, significant bleeding usually requires surgical re‐exploration (see pages 446–447). If markedly abnormal coagulation studies are present, yet bleeding persists despite their correction, surgical exploration is also indicated.
Prothrombin time (PT) measured as the INR assesses the extrinsic coagulation cascade. The INR may be slightly increased after a standard pump run, but clotting factor levels exceeding 30% of normal should allow for satisfactory hemostasis. An abnormal INR is usually corrected with FFP, but administration of PCC may be more effective.81,82
Partial thromboplastin time (PTT) assesses the intrinsic coagulation cascade and can also detect residual or recurrent heparin effect (“heparin rebound”). When an elevated PTT occurs as an isolated abnormality, or with slight elevation of the INR, protamine may be beneficial in correcting the PTT and controlling bleeding. One study found that heparin rebound documented by elevated Xa levels and an abnormal clotting time was present in virtually all patients after surgery and could be abolished by a continuous infusion of 25 mg/h of protamine for six hours.83 However, another study found little correlation of elevated anti‐Xa levels with elevated PTTs, because the latter may be related to clotting factor deficiency or excessive protamine rather than residual heparin.84 It is sometimes worthwhile obtaining an ACT in the ICU, although it may also be elevated in the absence of residual heparin.
Platelet count. Although CPB reduces the platelet count by about 30–50% and also produces platelet dysfunction, platelet function is usually adequate to produce hemostasis. Platelet transfusions may be justified in the bleeding patient with thrombocytopenia (generally <100,000/μL) or for suspicion of platelet dysfunction (usually for patients on aspirin or P2Y12 inhibitors) even in the absence of thrombocytopenia.
Fibrinogen (factor I) levels should be assessed in patients with moderate–severe bleeding. Low levels preoperatively and upon arrival in the ICU are risk factors for bleeding and re‐exploration.29 Fibrinogen is essential for proper platelet function by promoting platelet–platelet interaction leading to platelet aggregation. It is also a cell adhesion molecule that enhances platelet adhesion to endothelial cells. If the patient has significant bleeding and a fibrinogen level <100 mg/dL, transfusion of cryoprecipitate, which is rich in factors I, VIII, and XIII, is helpful. Alternatively, fibrinogen concentrate can be used to increase clot stability, and in combination with platelets, can improve platelet aggregation.85 One study found that fibrinogen concentrates (about 8 g) were more effective in reducing bleeding than the combination of eight units of FFP and four units of platelets.86
Fibrinolysis is invariably present in all patients having heart surgery, although it may be attenuated by the use of lysine analogues, which exhibit antifibrinolytic properties. Test results consistent with fibrinolysis are nonspecific and include elevations in the INR and PTT, decreased levels of factors I and VIII, and elevated fibrin split products (such as D‐dimer). The best means of identifying fibrinolysis are thromboelastography (TEG) and a Sonoclot analysis.
Platelet function can be assessed by a variety of available technologies, including TEG and those that measure whole blood impedance platelet aggregometry.73,87 Although the correlation of these tests with the occurrence of bleeding is not specific, qualitative platelet abnormalities in the bleeding patient do suggest that platelet transfusions are indicated. One comparative study of TEG and platelet aggregometry found that both were effective in identifying impaired hemostasis, but neither correlated with the amount of blood loss.87 In most centers, suspicion of platelet dysfunction is based upon preoperative use of antiplatelet agents and prompts platelet transfusions without point‐of‐care testing.
Thromboelastography (TEG) and rotational thromboelastometry (ROTEM) give a qualitative measurement of clot strength. These tests evaluate the interaction of platelets with the coagulation cascade from the onset of clot formation through clot lysis and have a distinct contour for a variety of coagulation abnormalities, including fibrinolysis (Figure 9.1). Although these tests are very helpful in guiding therapy in patients with coagulopathic bleeding, they do not necessarily identify patients who are going to bleed, and treatment is not indicated for abnormalities noted in the absence of bleeding.88–91 However, in the bleeding patient, they provide more rapid assessment of hemostatic abnormalities than standard blood coagulation testing, thus allowing for more prompt and effective therapy. Studies have found that use of a TEG‐guided protocol reduces blood loss and transfusions.92 ROTEM has been used to identify platelet dysfunction and also exclude residual heparin as the cause of an elevated ACT.93
Sonoclot analysis (Figure 9.2) is another viscoelastic method of evaluating clot formation and retraction that allows for the assessment of coagulation factors, fibrinogen, and platelet activity. The device measures the changing impedance to movement imposed by the developing clot on a small probe that vibrates at an ultrasonic frequency within a blood sample. Studies have suggested that Sonoclot is more predictive of bleeding than routine coagulation studies.94 This device has seen limited use but can direct appropriate therapy in patients with persistent bleeding.
Repeat a chest x‐ray
Note the overall width of the mediastinum. A widened mediastinum may suggest undrained clotted blood accumulating within the pericardial cavity that could cause cardiac tamponade. Comparison with preoperative films can be misleading because of differences in technique, but any difference between the immediate postoperative supine film and a repeat film is of concern (Figure 9.3). A widened superior mediastinum is noted when there is significant clot accumulation around the great vessels.
Note the distance between the edge of the mediastinal silhouette and the location of the Swan‐Ganz catheter in the right atrium or the right atrial pacing wires (if placed on the right atrial free wall). If this distance widens, suspect clot accumulation adjacent to the right atrium.
Note any accumulation of blood within the pleural space that has not drained through the pleural chest tubes. This can be difficult to assess since fluid will layer out on a supine film, so a discrepancy in the haziness of the two pleural spaces should be sought.
Consider obtaining an echocardiogram if any of the above suggests the presence of cardiac tamponade. If the TTE is inconclusive, obtain a transesophageal study.
Table 9.4 Assessment of Postoperative Mediastinal Bleeding
Obtain immediate postoperative chest x‐ray as baseline evaluation of the mediastinum and pleural spaces
Quantify the degree of bleeding into drainage unit frequently
Optimize hemodynamic status while addressing bleeding issues
Obtain coagulation studies:
PT/INR, PTT, platelet count, fibrinogen level
Thromboelastogram, if available
Platelet function testing
Repeat coagulation studies after blood products are administered if still bleeding
Repeat chest x‐ray if concerned about tamponade or undrained blood
Obtain TEE if concerned about tamponade
V. Management of Mediastinal Bleeding
Although there is no role for prophylactic blood product transfusions in the prevention of bleeding following open‐heart surgery, persistent bleeding must be treated immediately and aggressively based on the degree of bleeding and the suspected etiology of hemorrhage (Table 9.5). It is a truism that the longer a patient bleeds, the worse the coagulopathy becomes. In general, the most benign and least invasive treatments should be considered first. If a patient was “dry” at the time of closure and suddenly starts to bleed, the source is usually surgical in nature and requires re‐exploration. In contrast, the patient with persistent bleeding may have a surgical or medical cause for the bleeding. When significant mediastinal bleeding persists despite prompt correction of a suspected coagulopathy, earlier re‐exploration (within 12 hours) significantly lowers the mortality rate.95 Excessive bleeding prompting re‐exploration increases the mortality rate two‐ to fourfold compared with patients with an insignificant amount of bleeding, which reinforces the importance of careful hemostasis at the conclusion of surgery and aggressive management of coagulopathies.20,21,96,97
Ensure chest tube patency. Ongoing bleeding without drainage leads to cardiac tamponade. Gently milk the tubes to remove clot. Aggressive stripping is not necessary.
Warm the patient to 37 °C. Hypothermia produces a generalized suppression of the coagulation mechanism and also impairs platelet function.98 The use of a heated humidifier in the ventilator circuit and a forced‐air warming blanket are beneficial in promoting warming and will reduce the tendency to shiver. All blood products should be delivered through blood‐warming devices, if possible.
Control hypertension with vasodilators (clevidipine, nitroprusside, nicardipine) or β‐blockers (esmolol for the hyperdynamic heart). Higher doses of propofol or morphine can also be used since extubation should not be contemplated in the bleeding patient.
Control agitation in the awake patient with short‐acting sedatives:
Propofol 25–75 μg/kg/min
Dexmedetomidine 1 mg/kg load over 10 minutes followed by a continuous infusion of 0.2–1.5 μg/kg/h
Midazolam 2.5–5.0 mg IV q1–2h
Morphine 2.5–5 mg IV q1–2h
Control shivering with meperidine 25–50 mg IV or dexmedetomidine
Use increasing levels of positive end‐expiratory pressure (PEEP) to augment mediastinal pressure, which may reduce microvascular bleeding.22 However, prophylactic PEEP at levels of either 5 or 10 cm H2O has not been found to be effective in reducing bleeding or transfusion requirements.99 If it is elected to increase PEEP to control bleeding, careful attention to its effects on hemodynamics is essential.
Use blood components to treat early significant bleeding. This should preferably be based on point‐of‐care testing, but appropriate treatment may initially be based on suspicion of the hemostatic defect. For example, the patient who has received aspirin, a P2Y12 inhibitor, a IIb/IIIa inhibitor, or is uremic is likely to have platelet dysfunction and will benefit primarily from platelet transfusions, even if the platelet count is normal. In some cases, desmopressin may also prove beneficial (see section V.K. on page 439)100,101 In contrast, the patient who has recently been on warfarin or has hepatic dysfunction is more likely to have clotting factor deficiencies and may benefit more from an initial transfusion of FFP, PCC, cryoprecipitate, or fibrinogen concentrate. Use of multiple products may be necessary in the patient who has had a long duration of CPB (>3 h) or who has received multiple packed red blood cells (PRBCs) during surgery. Aggressive treatment with blood components should be provided promptly for significant bleeding from a suspected coagulopathy, because persistent bleeding causes progressive depletion of clotting factors and platelets (“coagulopathy begets coagulopathy”).
Once the results of coagulation studies become available, there is more objective information upon which to base therapy. Point‐of‐care testing in the OR is the most expeditious way of assessing the hemostatic profile (Figure 9.4). Some groups preferentially use the thromboelastogram to identify the exact nature of the hemostatic defect, allowing for more prompt initiation of appropriate therapy. This may result in lower transfusion requirements.92 The results of routine coagulation studies drawn after CPB is terminated are usually available in the OR or soon after arrival in the ICU. If bleeding persists despite corrective measures, clotting studies can be repeated to reassess the status of the coagulation system.
An elevated PT implies the need for clotting factors provided by FFP, PCC, and/or cryoprecipitate.
An elevated PTT or ACT suggests a problem with the intrinsic coagulation cascade or persistent heparin effect. Additional protamine may be given first, with an understanding that an elevated PTT may not be related to heparin, and protamine could exacerbate a coagulopathy. FFP and/or cryoprecipitate may also be indicated.
Fibrinogen levels <100 mg/dL warrant administration of cryoprecipitate or fibrinogen concentrate when bleeding is persistent.
A platelet count below 100,000/μL suggests the need for platelet transfusions. Because CPB induces platelet dysfunction, suspicion of a qualitative defect in the actively bleeding patient should be treated with platelets even if the platelet count is adequate.
Note: abnormal results do not need to be treated if the patient has minimal bleeding. Blood samples are frequently drawn from heparinized lines, so they should be repeated if results are markedly abnormal or inconsistent with the amount of bleeding. Platelet transfusions are not indicated in the nonbleeding patient until the platelet count approaches 20,000–30,000/μL, although most patients in the immediate postoperative period will tend to bleed at a platelet count of less than about 60,000/μL.
Blood transfusions are often neglected in the bleeding patient when anemia may be progressive and exacerbated by hemodilution from the administration of FFP and platelets. Although a low transfusion trigger (HCT of 21%) might be acceptable in a nonbleeding, stable young patient with no comorbidities, this is not safe in the bleeding patient. The patient with ongoing mediastinal bleeding should be transfused to maintain a HCT at a reasonable level (>24%) as a safety margin to maintain satisfactory tissue oxygenation. Furthermore, there are a number of clinical indications for transfusion to a higher HCT, especially those suggestive of an impairment in tissue oxygen supply (see section VI.D on page 441). Notably, platelet function is impaired in the profoundly anemic patient.102 Red cells increase platelet‐to‐platelet interaction and facilitate the interaction of platelets with the subendothelium to improve hemostasis. Nonetheless, bleeding does not seem to be exacerbated by HCTs as low as 24% compared with 30%.103
Protamine may be given in a dose of 25–50 mg (5 mg/min) if the PTT is elevated. Generally, the ACT correlates with the PTT but is usually not drawn once the patient leaves the OR. Although the ACT should return to baseline after protamine administration, reinfusion of cell‐saver blood may reintroduce a small amount of heparin, and release of heparin from tissue stores can introduce residual unneutralized heparin that contributes to bleeding. This may occur because the half‐life of protamine is only about five minutes, with virtual elimination from the bloodstream in about 20–30 minutes.104 Thus, a continuous infusion of low‐dose protamine or a few small additional doses for potential heparin rebound is a feasible approach.83 However, prolonged ACTs (and inferentially PTTs) may be noted in the absence of circulating heparin, so if additional protamine is given and the PTT remains elevated, unneutralized heparin may not be the problem. In fact, excessive use of protamine will elevate the ACT and cause bleeding. Excess protamine causes platelet dysfunction, enhances fibrinolysis, and decreases clot strength, emphasizing that indiscriminate use of excessive protamine should be avoided.14,105
Desmopressin (DDAVP) has no role in the prophylaxis of postoperative bleeding, but might be considered in patients with documented von Willebrand’s disease, aortic stenosis with impaired platelet function associated with acquired von Willebrand’s disease, uremia, and possibly drug‐induced platelet dysfunction.100,101
Bleeding following cardiac surgery is often secondary to an acquired defect in the formation of the platelet plug caused by a deficiency in von Willebrand factor. DDAVP increases the level of factor VIII precursors, von Willebrand factor (by approximately 50%), and tissue‐type plasminogen activator by releasing them from vascular endothelium. These factors are responsible for promoting platelet adhesion to the subendothelium.
Patients with aortic stenosis often have acquired type 2A von Willebrand syndrome.30 This develops due to proteolysis of the largest multimers of von Willebrand factor by shear stress generated on blood as it passes through the stenotic valve. These multimers are important for platelet‐mediated hemostasis, so when reduced, they can cause bleeding. Desmopressin given after the induction of anesthesia in patients with abnormal platelet function associated with this syndrome has been shown to significantly reduce perioperative blood loss.101
DDAVP is given in a dose of 0.3–0.4 μg/kg IV over 20 minutes. A slow infusion may attenuate the peripheral vasodilation and hypotension that often follows DDAVP infusion. Peak effects are seen in 30–60 minutes.
Calcium chloride 1 g IV (10 mL of 10% solution) given over 15 minutes may be administered if the patient has received multiple transfusions of CPD preserved blood during a short period of time (e.g. more than 10 units within 1–2 hours). The citrate used as a preservative in CPD blood binds calcium, but hypocalcemia is unusual because of the rapid metabolism of citrate by the liver. However, calcium administration is not necessary when adenine‐saline (AS‐1) is used as the preservative. If hypocalcemia is present, as it often is following CPB, calcium chloride is preferable to calcium gluconate because it provides three times more ionized calcium.
Table 9.5 Management of Postoperative Mediastinal Bleeding
Explore early for significant ongoing bleeding or tamponade
Ensure that chest tubes are patent
Warm patient to normothermia
Control hypertension, agitation, and shivering
Check results of coagulation studies (INR, PTT, platelet count, fibrinogen level, or TEG)
Correct ionized calcium to >1.1 mmol/L
Protamine 25 mg IV for two doses if elevated PTT or infusion of 25 mg/h × 6 h
Consider use of 10 cm PEEP with caution
Packed cells if hematocrit <24%
Platelets, 1–2 “six packs”
Fresh frozen plasma, 2–4 units
Cryoprecipitate, 6 units
Fibrinogen concentrate 25–75 mg/kg
Prothrombin complex concentrate (PCC) 25 units/kg
Desmopressin (DDAVP) 0.3 mg/kg IV over 20 minutes (if suspect platelet dysfunction from uremia or aspirin, von Willebrand’s factor deficiency)
Recombinant factor VIIa 90 mg/kg if severe coagulopathy
Transesophageal echocardiography if concerned about tamponade
Urgent exploration for significant ongoing bleeding or tamponade
Emergency exploration for exsanguinating hemorrhage or near cardiac arrest from tamponade
VI. Blood Transfusions: Red Cells
Red cell transfusions are indicated primarily to increase the oxygen‐carrying capacity of blood to avoid end‐organ ischemia and dysfunction. Tissue oxygen delivery depends on the cardiac output, the hemoglobin level, and the oxygen extraction ratio in tissues. Because the early postoperative period is associated with delayed myocardial metabolic recovery, reduced cardiac output, and significant anemia, oxygen delivery is commonly reduced by at least 25% postoperatively. Tissue oxygenation may be maintained in healthy patients with a hemoglobin as low as 6–7 g/dL (HCT around 18–21%), and the safe lower limit for the HCT in the stable postoperative patient is probably around 22%.
Nonetheless, the approach to the bleeding patient requires extra vigilance and a margin of safety to ensure adequate tissue oxygenation, minimize myocardial ischemia, and prevent hemodynamic compromise. It is therefore safest to administer blood if the HCT is less than 24% when there is ongoing substantial blood loss with the predictable hemodilution from administration of blood components and platelets. However, there is no indication for transfusing to a HCT greater than 30%.
Blood transfusions are not benign, because they are associated with significant postoperative morbidity and compromised short‐ and long‐term survival after open‐heart surgery.9,106 Blood contains cytokines and proinflammatory mediators that are immunomodulatory. Transfusions are associated with numerous potential complications, including, but not limited to:
An increased risk of graft occlusion, renal, neurologic, gastrointestinal, and pulmonary morbidity after heart surgery.8–13,107
Fever associated with hemolytic or nonhemolytic transfusion reactions. The former is caused by circulating antibodies directed against donor leukocytes or HLA antigens. The latter may be related to cytokines in the donor product.
Allergic reactions ranging from urticarial to anaphylaxis
Viral infections, especially with cytomegalovirus (CMV), which is present in about 50% of donor units. However, HIV, hepatitis B, and hepatitis C are rarely transmitted with effective screening. Bacterial infections are also uncommon, but the risk of pneumonia is increased and may be related to transfusion‐related immunomodulation (TRIM). Leukocyte depletion by the blood bank can reduce the risk of infection and may lower mortality rates as well.14,108
Transfusion‐related acute lung injury (TRALI). This may be caused by donor antibody–recipient leukocyte interactions and possibly neutrophil activation by bioactive substances from transfusions, resulting in lung damage and a capillary leak occurring within six hours of a transfusion. Although most patients recover from TRALI in a few days, the associated mortality rate is noted to be 5–15%.9
Transfusion‐related circulatory overload (TACO) is a syndrome of cardiogenic pulmonary edema manifested by hypoxemia, hypertension, and tachycardia occurring within 12 hours of transfusion. This is related to the aggressive administration of too much volume of blood products too rapidly to a patient with preexisting fluid overload from heart failure, LV dysfunction, or chronic kidney disease.109 Its occurrence can be avoided by giving a diuretic, such as furosemide, at the same time as a blood transfusion. This is most applicable to the profoundly anemic patient who is somewhat fluid overloaded but not actively bleeding.
Transfusion triggers should be determined by clinical criteria, including hemodynamic factors (hypotension, tachycardia, low cardiac output states with low mixed venous oxygen saturation, metabolic acidosis, elevated lactate), or evidence of neurologic impairment, respiratory insufficiency, or renal dysfunction. Despite the fundamental concept that blood will dramatically improve oxygen delivery, it must be recognized that transfusions may provide minimal improvement in oxygen‐carrying capacity immediately after transfusion, may reduce microcirculatory flow, and in fact could prove more detrimental than beneficial. This is because the 2,3‐diphosphoglycerate (2,3‐DPG) level in blood is very low, especially with longer durations of storage, resulting in a leftward shift of the oxyhemoglobin dissociation curve with more avid binding of oxygen to hemoglobin and less release to tissues. Fortunately, 2,3‐DPG levels return to 50% of normal within 24 hours after transfusion.
Use of blood filters is beneficial in removing microaggregates of blood. Blood filters of at least 170 μm pore size must be used for all blood transfusions. Filters of 20–40 μm pore size are more effective in removing microaggregates of fibrin, platelet debris, and leukocytes that accumulate in stored blood. These filters have been shown to decrease the incidence of nonhemolytic febrile transfusion reactions and may reduce the adverse effects of multiple transfusions on pulmonary function. Blood lines should be primed with isotonic solutions (preferably normal saline), avoiding lactated Ringer’s, which contains calcium, and D5W, which is hypotonic and will produce significant red cell hemolysis.
Note: care should be taken to avoid transfusing cold blood products. Blood warmers should generally be used if the patient receives rapid transfusions. If one unit is to be transfused, it should be allowed to sit at ambient room temperature or under a heating hood for several minutes to warm.
Packed red blood cells (RBCs) contain approximately 200 mL of red cells and 70 mL of a preservative, most commonly citrate‐phosphate‐dextrose (CPD) with various additives to extend shelf life. The addition of adenine (CPDA‐1) extends shelf life to 35 days, and solutions with more dextrose and adenine and with mannitol (AS‐1 [Adsol] and AS‐5 [Optisol]) increase shelf life to 42 days. Each unit has an average HCT of 70%, and one unit will raise the HCT of a 70 kg man by 3%. At least 70% of transfused cells survive 24 hours, and these cells have a normal lifespan. Since packed cells contain no clotting factors, administration of FFP should be considered to replace clotting factors if a large number of units (generally more than five) is given over a short period of time.
Despite the extended shelf life with improved preservation solutions, significant changes still occur in the red cells with storage. These include increased levels of cytokines, which produce more systemic inflammation, increased lactate with a reduction in pH, loss of deformability, which increases capillary transit time, depletion of 2,3‐DPG, which reduces oxygen unloading from hemoglobin, and increased cell lysis leading to hyperkalemia.110 Although some studies suggest that the risk of infection is greater with prolonged storage, other studies indicate that the storage time of packed RBCs does not affect outcomes.14,111,112
Leukoreduction of red cells is beneficial in reducing some of the febrile and nonhemolytic transfusion reactions. In some hospitals, this is done routinely.108
Fresh whole blood (less than six hours old) has a HCT of about 35% and contains clotting factors and platelets. One unit has been shown to provide equivalent, if not superior, hemostasis to that of 10 units of platelets.113 It is probably the best replacement product, but is usually not available since most blood banks fractionate blood into components. Whole blood stored over 24 hours has few viable platelets, but levels of other clotting factors are reasonably well maintained for several days.
Cell‐saver blood (shed and washed in the OR) is rinsed with heparinized saline and is devoid of clotting factors and platelets. A small amount of heparin may be present after centrifugation, but this is not considered clinically significant. The survival, function, and hemolysis of washed red blood cells is equivalent to that of nonprocessed blood.114
Hemofiltration blood is obtained by placing a hemofilter in the extracorporeal circuit. This provides concentrated red cells and also preserves platelets and clotting factors. Studies have shown superior blood salvage and hemostasis with use of a hemofilter than with cell‐saving devices, but ultrafiltration of pump blood is generally not recommended.22
Autotransfusion of shed mediastinal blood in the ICU is a controversial means of blood salvage. It has arguably been shown to reduce the need for transfusions, and most systems have been designed for reinfusion through 20–40 μm blood filters without washing. Blood filters do not completely remove lipid particles and blood microaggregates, and the reinfused blood contains low levels of factor I and VIII, a low level of platelets which are dysfunctional, elevated levels of fibrinolytics (fibrin split products), inflammatory cytokines, endotoxin, tissue factor, and free hemoglobin. Washing can remove some of these factors but will also eliminate all clotting factors and platelets. If unwashed blood is returned in moderate amounts (>500 mL), an apparent coagulopathy will be present with an elevation in INR, PTT, and D‐dimers, and a reduction in fibrinogen.115,116 There may also be an increased incidence of wound infections with reinfusion of unwashed shed mediastinal blood.117 Thus, if autotransfusion is to be used as part of a blood conservation program, blood should be washed in a cell‐saving device prior to reinfusion, and the amount of reinfused blood should be limited.22
VII. Blood Components, Factor Concentrates, and Colloids
Platelets should be given to the bleeding patient if the platelet count is less than 100,000/μL. Furthermore, since platelets are dysfunctional in patients receiving antiplatelet medications preoperatively and as a result of CPB, one should not hesitate to administer platelets for ongoing bleeding even if the platelet count exceeds 100,000/μL. Platelets are not indicated in the nonbleeding patient unless the count is perilously low (<20–30,000/μL).
Platelets are provided as a pooled preparation from one or several donors, usually as a six‐unit bag, which is the usual amount given to an average‐sized adult. Each unit contains approximately 8 × 1010 platelets and should increase the platelet count by about 7000–10,000/μL in a 75 kg adult. One unit of platelets contains 70% of the platelets in a unit of fresh blood, but platelets lose some of their functional capacity during storage. Platelets stored at room temperature can be used for up to five days and have a lifespan of eight days. Those stored at 4 °C are useful for only 24 hours (only 50–70% of total platelet activity is present at six hours) and have a lifespan of only 2–3 days.
Platelet function is impaired in patients with hypofibrinogenemia and when the HCT is less than 30%. Thus, administration of cryoprecipitate or fibrinogen concentrates along with platelets will improve platelet aggregation and clot stability.85 Red cell transfusion to raise the HCT to 26% may also improve platelet function.102
Transfused platelets will be less effective when given within 4–6 hours of a dose of a P2Y12 inhibitor, because the active compound may still be present in the bloodstream.
ABO compatibility should be observed for platelets, but it is not essential. For each donor used, there is a similar risk of transmitting hepatitis and HIV as one unit of blood.
Platelets should be administered through a 170 μm filter. Several filters are available (such as the Pall LRF 10 filter) that can be used to remove leukocytes from platelet transfusions. Use of these filters may be beneficial in reducing the risk of allergic reactions caused by red and white cells present in platelet packs. Pretreatment with diphenhydramine (50 mg IV), ranitidine (150 mg IV) (H1 and H2 blockers), and steroids (hydrocortisone 100 mg IV) might also attenuate these reactions, but is usually not necessary.
Despite some claims that platelet transfusions are associated with higher risks of infection, respiratory complications, stroke, and death, it is more likely that the need for platelets is simply a surrogate marker for sicker patients.118 A study of nearly 33,000 patients from the Cleveland Clinic confirmed increased postoperative morbidity in patients receiving platelets, but after risk adjustment, there was no increase in morbidity or mortality from transfused platelets.119
Fresh frozen plasma (FFP) contains all clotting factors at normal concentrations with a slight reduction in factor V (66% of normal) and factor VIII (41% of normal). It is devoid of red cells, white cells, and platelets. When cryoprecipitate is obtained from the same unit of blood, FFP will contain low levels of factors I, VIII, XIII, von Willebrand factor, and fibronectin. Only 30% of the normal level of most clotting factors is essential to provide hemostasis, and the INR generally has to exceed 1.5 before a clinically significant factor deficiency exists. However, due to the hemodilutional effects of CPB and the progressive loss of clotting factors during ongoing bleeding, one should not hesitate to administer FFP to improve hemostasis in the bleeding patient even if the INR is mildly elevated. Although some degree of coagulopathy with a reduction in clotting factors is present after CPB, there is no documented benefit of prophylactically administering FFP.120 Because of the importance of factors I and VIII in promoting platelet aggregation and adhesion to the endothelium, the additional transfusion of cryoprecipitate should be considered if fibrinogen levels are low.
One unit of FFP contains about 250 mL of volume. The amount given is usually 2–4 units for the average adult. Four units will increase the level of clotting factors by 10%, which is considered the amount necessary to improve coagulation status.
FFP should be ABO compatible and given through a 170 μm filter. FFP is not viral inactivated, and since each unit is derived from one unit of whole blood from one donor, it has a similar risk of transmitting hepatitis or HIV as one unit of blood.
FFP may be given to patients with antithrombin (AT) deficiency to achieve adequate anticoagulation for CPB.59 This may only be recognized when significant heparin resistance is noted in the OR. To minimize the amount of volume infused, a concentrated source of AT is commercially available (Thrombate III). The amount required is based on an estimate of the level of AT present (see calculation on page 251). Antithrombin is not recommended to reduce bleeding following CPB.
Note: the administration of FFP and platelets not only provides clotting factors but also raises filling pressures. These blood products will therefore lower the HCT and can precipitate fluid overload. If the HCT is less than 24% or not yet available and the patient is bleeding, anticipate the need for blood if other volume is being administered. Remember that the HCT does not change with acute blood loss until replacement fluids are administered.
Prothrombin complex concentrate (PCC) contains the vitamin‐K‐dependent coagulation factors II, IX, and X (three‐factor PCC [Profilnine]) and may contain variable amounts of factor VII as well (four‐factor PCC [Kcentra]).121
PCC is able to reduce an elevated INR more rapidly than FFP (usually in less than an hour) and with less volume.37 It comes in 500‐ and 1000‐unit vials, reflecting units of factor IX. The dosage is calculated based on the patient’s weight and INR. It can also be used to offset the anticoagulant effects of the NOACs in the absence of the more expensive antidotes noted on page 426.
PCC is very effective in the control of refractory bleeding after cardiac surgery.122 It is more effective than FFP in reducing blood loss and transfusions when used for postsurgical bleeding, albeit with a slightly increased risk of thromboembolic events and acute kidney injury.123 Studies comparing three‐ and four‐factor PCC with recombinant factor VII have found them to be equally efficacious in reducing bleeding, but with less renal impairment.124,125 The preoperative dose used for warfarin reversal is 25 units/kg for an elevated INR <4, 35 units/kg for an INR of 4–6, and 50 units/kg for an INR >6. Postoperatively, the INR is rarely that high, and the dose may be based on thromboelastographic findings. The usual dose is 25–35 units/kg. There is potential benefit in reducing bleeding following surgery in patients taking a NOAC preoperatively.
Cryoprecipitate represents the cold insoluble portion of plasma that precipitates when FFP is thawed at 1–6 °C. It is then refrozen at 20 °C within one hour. Approximately 15 mL is derived from one unit which is then suspended in 15 mL of plasma and pooled into a concentrate of 5–6 units, containing about 200 mL. Viral inactivation is not performed, so there is a risk of pathogen transmission.126
Each unit of 10–20 mL provides concentrated levels of factors I, VIII, and XIII. This amounts to about 150–250 mg of fibrinogen, 80–100 units of factor VIII procoagulant activity (VIII:C), 40–50% of the original plasma content of von Willebrand factor, factor XIII (fibrin‐stabilizing factor), and fibronectin (a tissue integrin involved in wound healing). Factors I and VIII are essential for proper platelet aggregation and platelet adherence to endothelium.
Hypofibrinogenemia can significantly impair hemostasis, and when a patient has significant bleeding and a fibrinogen level <100 mg/dL, cryoprecipitate given with platelets is more effective than FFP in reducing bleeding. Although often utilized empirically for significant bleeding due to slow turnaround times in obtaining fibrinogen levels, TEG may be helpful in determining whether it should be given.
The amount given is usually 1 unit/7–10 kg of body weight (e.g. 7 units to a 70 kg patient). One unit will raise the fibrinogen level of a 70 kg man by 7–10 mg/dL. Cryoprecipitate must be thawed before infusion and should be given through a 170 μm filter within 4–6 hours of thawing. ABO compatibility should be observed. One can also calculate the number of units that will be required from the following equations: