| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
* From the Departments of Anesthesia, Health Policy, Management, and
Evaluation,
Cardiac Surgery, and
Hematology, University Health Network, University of Toronto, Toronto, Ontario, Canada.
Addresss correspondence to: Dr. Keyvan Karkouti, University Health Network, Toronto General Hospital, Department of Anesthesia, EN 3-402, 200 Elizabeth Street, Toronto, Ontario M5G 2C4, Canada. Phone: 416-340-5164; Fax: 416-340-3698; E-mail: keyvan.karkouti{at}uhn.on.ca
 |
Abstract |
|---|
 TOP Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Methods: Data were collected prospectively over the period from 2000 to 2005, on patients who underwent surgery with CPB at one hospital. Patients who received 
five units of RBC within one day of surgery were classified as MBT. Logistic regression was used to appropriately select and weigh perioperative variables in the prediction rule, which was developed on the initial 60% of the sample and validated on the remaining 40%.
Results: Of the 10,667 patients included, 925 (8.7%) had MBT. The clinical prediction rule included 12 variables (listed in order of predictive value: CPB duration, preoperative hemoglobin concentration, body surface area, nadir CPB hematocrit, previous sternotomy, preoperative shock, preoperative platelet count, urgency of surgery, age, surgeon, deep hypothermic circulatory arrest, and type of procedure) and was highly discriminative (c-index = 0.88). In the validation set, those classified as low-, moderate-, and high-risk by a simple risk score derived from the prediction rule had a 5%, 27%, and 58% chance of MBT, respectively.
Conclusion: A clinical prediction rule was developed that accurately identified patients at low-risk or high-risk for MBT. Studies are needed to determine the external generalizability and clinical utility of the prediction rule.
|
Introduction |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
A clinical prediction rule that could accurately estimate the probability of massive blood transfusion (MBT) after cardiac surgery would have clinical and research relevance. Such a prediction rule could be used to rationalize clinical management by, for example, limiting the use of expensive blood conservation modalities (such as aprotinin or cell salvage and washing) in patients identified as being at low-risk for MBT. It could also be used to investigate, and, if found to be appropriate, institute preventive measures (such as modifying risk factors or prophylactic administration of hemostatic agents or coagulation factors) to reduce the risk of MBT in those identified as being at high-risk for MBT.
Previous studies have identified several variables that are associated with RBC transfusion or excessive blood loss in cardiac surgery (Table I
).2,5–21 However, no multivariable prediction rules have been developed that can be used by clinicians to accurately estimate the probability of MBT in cardiac surgical patients. Consequently, we sought to develop and validate such a prediction rule in a cohort of patients who underwent cardiac surgery with cardiopulmonary bypass (CPB) at a single institution.
|
|
Methods |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Clinical practice
Hemostatic and CPB management during the study period were as follows. Antifibrinolytics [tranexamic acid (Cyclokapron®, Pharmacia & UpJohn Inc., Mississauga, ON, Canada) 50 to 100 mg·kg–1 or aprotinin (Trasylol®, Bayer AG, Toronto, ON, Canada) 6 x 106 units] were routinely administered to all patients. Institutional guidelines recommended that aprotinin be reserved for patients who had active endocarditis, were undergoing complex procedures requiring prolonged CPB support, or had at least two previous sternotomies. Anticoagulation for CPB was achieved with heparin to maintain an activated clotting time above 480 sec. The CPB circuit (non-heparin coated; membrane, hollow fibre oxygenator; roller pump) was primed with 1.8 L of Ringer’s lactate and 50 mL of 20% mannitol. Albumin (5% or 25%) and synthetic colloids (Pentaspan®; Bristol-Myers Squibb, Montreal, QC, Canada) were used at the discretion of the clinical team. Management of CPB included systemic temperature drift to 34°C, alpha-stat pH management, targeted mean perfusion pressure between 50–70 mmHg, and pump flow rates of 2.0–2.4 L·min–1·m–2. Myocardial protection was achieved with intermittent antegrade and, occasionally, retrograde blood cardioplegia. When necessary, deep hypothermic circulatory arrest (DHCA) was achieved by cooling to 20–28°C, depending on the anticipated duration of circulatory arrest. Neuroprotection was carried out with antegrade or retrograde cerebral perfusion at the discretion of the operating surgeon. When circulatory arrest was carried out for very short periods of time, cerebral perfusion was often omitted. Pericardial blood was salvaged into the cardiotomy suction reservoir and re-infused via the CPB circuit for as long as patients were anticoagulated. After separation from CPB, heparin was neutralized with protamine sulphate to a target activated clotting time within 10% of baseline.
Perioperative blood management during the study period was as follows. To guide RBC transfusions, hematocrit concentrations were measured before surgery, before CPB, every 15 min during CPB, every 30 min after CPB in the operating room, upon arrival and every four hours thereafter in the intensive care unit for the first 24 hr after surgery (and more frequently in bleeding or unstable patients). Leukoreduced RBC concentrates were transfused to maintain the hematocrit concentration at or above 18–20% during CPB, and at or above 24–27% after CPB.
Platelet counts, partial thromboplastin time, pro-thrombin time and international normalized ratio (INR) were routinely measured before surgery and upon arrival to the intensive care unit after surgery. During surgery, these tests were performed at the discretion of the clinical team in patients with prolonged CPB or microvascular hemorrhage following termination of CPB and neutralization of heparin. Indications for platelet transfusion included a platelet count of < 50 x 109·L–1, ongoing hemorrhage after complete reversal of heparin and a platelet count of < 80 x 109·L–1, or ongoing hemorrhage after prolonged CPB (> two hours was considered prolonged by most clinicians) irrespective of platelet counts. In cases where the risk of developing a qualitative platelet disorder after CPB was deemed to be high (e.g., recent use of long-acting platelet glycoprotein IIb/IIIa receptor antagonist therapy, high-dose clopidogrel, or CPB > three hours), platelets were at times transfused pro-phylactically upon termination of CPB. All platelets (random donor or single donor) were leukoreduced by the Canadian Blood Services. Fresh frozen plasma was transfused to bleeding patients if the INR was greater than 1.5 after complete neutralization of heparin with protamine. Owing to 30 min turnaround times for coagulation tests, fresh frozen plasma and platelets were at times transfused empirically in bleeding patients. Cryoprecipitate transfusion was guided by fibrinogen levels, which were measured in bleeding patients. Patients received eight units of cryoprecipitate if their fibrinogen level was less than 1 g·L–1.
Changes in blood management practice during study period
In November 2002, a clinical protocol for the off-label use of activated recombinant factor VII (NiaStase®, Novo Nordisk, Mississauga, ON, Canada) in patients with excessive, refractory hemorrhage after cardiac surgery was introduced at our institution, the details of which have been described previously.23 In January 2003, the processing times for blood gas and hematocrit measures were reduced to two to five minutes from ten to 15 min owing to the introduction of a point-of-care system. Finally, throughout the study period there was a gradual increase in the RBC transfusion trigger during CPB, likely as a result of our observations that hematocrit concentrations below 20% were associated with increased morbidity.22,24
Dependent and predictor variables
The primary definition for MBT was five units of RBC transfused within one day of surgery. This cutoff was based on the distribution of perioperative RBC transfusion in our patient population (corresponding to the 90th percentile) as well as on the independent relationship between the number of units of RBC transfused and adverse postoperative events.4 To explore the predictive value of clinical variables associated with greater blood loss, a second classification for MBT using a cutoff of seven units of RBC within one day of surgery (95th percentile) was also assessed.
Predictor variables included perioperative variables previously shown to be associated with perioperative blood loss (Table I
), as well as several variables related to adverse postoperative outcomes.
Statistical analyses
SASTM version 8.2 (SAS Institute, Inc., Cary, NC, USA) was used for the statistical analyses. Categorical variables were summarized as frequencies and percentages, continuous variables as means and standard deviations if normally distributed, and medians and interquartile ranges if not normally distributed.
Patients operated on from January 2000 to December 2002 were used for model development (training set) and those operated on from January 2003 to May 2005 were used for model validation (validation set). January 2003 was used as the cutoff because it correlated with the changes of practice noted above and was therefore anticipated to result in training and validation sets that would differ from each other, allowing for a more robust test of the validity of the model. Characteristics of the two groups were compared using the t-test or Wilcoxon rank sum test for continuous variables and the Chi-square test for categorical variables.
Multivariable logistic regression was used to appropriately select and weigh the predictor variables for inclusion in the prediction rule.25 Bivariate analysis (using the Chi-square test for categorical variables and the t-test or Wilcoxon rank-sum test for continuous variables) was first carried out to assess the unadjusted relationship between the predictor variables and MBT in the training set. The shape of the relationships between the continuous predictor variables and MBT (logit transformation) were assessed using restricted cubic spline functions.25 Based on this analysis, some variables were either mathematically transformed or categorized for the logistic regression analysis.25
Backward stepwise logistic regression modelling was then carried out to construct the model for the prediction rule. All significant (P < 0.1) predictor variables were considered for inclusion in the model. Variable retention in the model was guided by the variables’ Wald 
2 statistic and co-linearity and interaction issues.25 For comparison, a second model was constructed using the second cutoff of seven units of RBC as the dependent variable, using similar modelling steps. The models’ calibration was assessed by the Hosmer-Lemeshow goodness-of-fit test (the larger the P value, the better the fit), and discriminative ability was assessed by the c-index [which is equivalent to the area under the receiver operating characteristic (ROC) curve].25
Based on the performance of the primary model ( five units of RBC) at various probability cutoffs, optimal cutoffs were identified for classifying patients as being at low, moderate, or high-risk for MBT. Since application of logistic regression models involves complex calculations, a simple risk score was developed by categorizing all continuous variables in the prediction rule based on the shape of their spline function in the logistic regression analysis, and assigning a risk derived from the ß coefficients of the variables in the logistic regression model. The predictive accuracy of the prediction rule and the simplified risk score were assessed in the validation set.
Data and medical records of a sample (identified randomly based on date of surgery) of patients who had MBT, but were misclassified by the prediction rule as low-risk, were reviewed to obtain the reasons for excessive blood loss.
|
Results |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
five and seven RBC units within one day of surgery, respectively (Table II). A total of 6,651 consecutive patients were assigned to the training set and the remaining 4,016 patients were assigned to the validation set. The training and validation sets had important differences including a higher incidence of MBT ( 5 U: 11% s 7%; P < 0.0001) and procedures other than isolated coronary artery bypass grafting or single valve (34% vs 25%; P < 0.0001) in the validation set (Table III).
|
|
five units) in the training set are shown in Table IV. The shapes of the relationships between some of the continuous variables and MBT are shown in Figure 1. Several continuous variables were categorized to conform to the linearity assumption of logistic regression or to simplify the model. The logistic regression model for MBT ( five units) is shown in Table Va. The model was well calibrated (Hosmer-Lemeshow goodness-of-fit P = 0.2) and discriminative (c-index = 0.88) when tested on the patients in the training set. Table Vb shows the model for the secondary MBT definition ( 7 U). The second model was also well calibrated (Hosmer-Lemeshow goodness-of-fit P = 0.3) and discriminative (c-index = 0.90) when tested on the patients in the training set. Moreover, the same variables were retained in the two models with little change in their predictive odds ratios. Notably, variables related to perioperative hemodilution (body surface area, preoperative hemoglobin, and nadir CPB hematocrit) remained in this second model, albeit their predictive value (as measured by their Wald 2) was reduced.
|
|
|
|
lists the sensitivity, specificity, and positive and negative predictive values of the primary model when it was re-applied to the patients in the training set using various probability cutoffs. The same data are provided as a ROC curve in Figure 2. The 0.10 and 0.35 probability levels were selected as the optimal probability cutoffs for classifying patients into low-, moderate-, and high-risk groups.
|
|
shows the risk score that was developed based on the ß coefficients of the variables in the primary model. The risk score is additive and can range from 0 to 14. Risk classification cutoffs were based on the 75th percentile (2.5) and 95th percentile (4.5) scores to correspond with the cutoffs of the prediction rule: 
2.5 denotes low-risk, between 2.5 to 4.5 denotes moderate-risk, and 4.5 denotes high-risk for MBT.
|
shows the performance of the prediction rule and the risk score in the validation set. For comparison, the predictive accuracies of three clinical variables known to be important predictors for excessive blood loss are also shown in Table VIII.
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Definition of MBT
The primary definition of MBT in this study was the transfusion of at least five units of RBC within one day of surgery. We have previously shown that this threshold, which occurs in only about 10% of our patients, is independently associated with significant increase in mortality.4 A second threshold of seven units of RBC within one day of surgery, corresponding to the 95th percentile in our patient population, was also analyzed. That the logistic regression models for the two outcomes were very similar supports the robustness of the prediction rule.
As a surrogate measure for perioperative blood loss, there are limitations with using units of RBC transfused perioperatively. Since RBC transfusions are based on factors other than perioperative blood loss, such as patients’ baseline red blood cell volume and clinicians’ transfusion practice, one cannot assume that every patient who was categorized as MBT had excessive perioperative blood loss. Owing to the high RBC transfusion thresholds used for defining MBT, however, it is reasonable to assume that few patients categorized as having had MBT did not have excessive perioperative blood loss.
Other measures of perioperative blood loss have their own limitations. Clinical estimates of blood loss are inherently unreliable, even under controlled conditions.28 Others have used surgical re-exploration as the criteria for excessive blood loss.2,16,17,21 Many patients with excessive blood loss, however, never require re-exploration, and some patients undergo re-exploration for causes other than bleeding. In cardiac surgery, the amount of blood in mediastinal tubes can also be used to identify excessive blood loss.15 This method, however, does not capture the significant proportion of patients who bleed excessively in the operating room after termination of CPB.
Independent predictors of MBT as identified by the prediction rule and risk modification opportunities
Massive blood transfusion after cardiac surgery can be due to surgical or non-surgical (i.e., coagulopathy) causes. Surgical bleeding is estimated to account for about 50% of cases that require re-exploration.29 In our sample, about one-third of patients with MBT required re-exploration (344/925). Although surgical bleeding is often unpredictable, the prediction rule included several variables that may be associated with increased risk of surgical bleeding owing to increased surgical complexity. These were type of procedure, urgency of procedure, previous sternotomy, CPB duration, and DHCA. Interestingly, when the surgeon was analyzed as a dichotomous variable based on the MBT median rate, it only had a moderate influence on the risk of MBT. Analyzing each surgeon separately increased the overall importance of the variable in the model but did not alter the discriminative ability or predictive performance of the model (results not shown).
The causes of coagulopathy after cardiac surgery with CPB include platelet-related abnormalities owing to thrombocytopenia or platelet dysfunction, reduction in plasma coagulation factors owing to hemodilution, excessive primary or secondary [due to disseminated intravascular coagulation (DIC)] fibrinolysis, and relative protamine deficiency or excess.26,30 Excessive primary fibrinolysis is unlikely to be a factor in subjects in this study due to the routine use of antifibrinolytics in nearly all patients. Since heparin and protamine doses were not captured in the databases, their predictive value could not be assessed. Not surprisingly, variables associated with platelet-related coagulopathy were the most predictive variables in the prediction rule. Of these, the most important was CPB duration, which is known to cause a progressive impairment in platelet function.29,31 Consistent with this, we found a linear relationship between CPB duration and risk of MBT (Figure 1). Deep hypothermic circulatory arrest, which not only prolongs CPB duration but also has detrimental effects on platelet function that are additive to those of CPB,32 was also an important predictor of MBT. Another important predictor was pre-existing thrombocytopenia, with the risk most pronounced when platelet counts were less than 100 x 109·L–1. Low platelet count has been found to be associated with MBT after cardiac surgery in some,33 but not all,30 previous studies. Our findings support the opinion of some experts that cardiac surgical patients at high-risk for MBT may benefit from platelet transfusions if their platelet count is less than 100 x 109·L–1.34 Of note, preoperative acetylsalicylic acid (ASA) use did not remain in the model, suggesting that in patients with MBT, its effect on bleeding may be overshadowed by other factors that cause platelet-related coagulopathy. Since the databases did not contain information on the use of other drugs that may increase perioperative blood loss (such as clopidogrel and glycoprotein IIb/IIIa receptor blockers),35 their effects on MBT could not be assessed. Our limited chart review of misclassified patients indicates that such drugs may account for about one-fifth of the 5% of patients who were misclassified in the low-risk group. Their effect in the other groups, however, was not determined, and their inclusion may have improved the accuracy of the prediction rule, particularly in the moderate-risk group.
As in previous studies (Table I), this study found that variables associated with low blood volume are independently associated with MBT. Preoperative hemoglobin and body surface area were inversely related to the risk of MBT, and age was directly related to the risk of MBT. Although patients with low baseline blood volume are more likely to require RBC transfusions in response to perioperative blood loss, this alone cannot account for the strong relationship between these variables and the transfusion of more than five or seven units of RBC within one day of surgery. Another plausible explanation relates to dilutional coagulopathy that can occur in patients with small blood volume after initiation of CPB. Since the amount of CPB prime is for the most part constant, coagulation factors are more likely to drop below levels required for normal hemostasis in patients with low blood volumes. This explanation is supported by the observed strong inverse relationship between nadir hematocrit during CPB and risk of MBT in both models. Thus, interventions that limit the amount of hemodilution, such as use of smaller CPB circuits or retrograde priming of the circuit, may reduce the risk of MBT in patients with low baseline blood volume.
Another etiology for excessive blood loss in cardiac surgery is DIC.30 Our prediction rule included preoperative shock, which may predispose patients to DIC.36
Predictive ability of the prediction rule or the risk score
Both the prediction rule and the simplified risk score were highly accurate in identifying patients at low-risk and at high-risk for MBT (Table VIII). In the low-risk group, their negative predictive value was 95%, and in the high-risk group, their positive predictive value was around 60%. These two categories represented over 80% of the population. This compares favourably to the predictive ability of known clinical predictors of bleeding risk, such as CPB duration, urgency of procedure, and history of previous sternotomies (Table VIII). In the moderate-risk category, however, the prediction rule and the risk score did not perform well and were no better than the clinical predictors.
Thus, the prediction rule or the risk score can be applied clinically to accurately stratify about 80% of patients undergoing cardiac surgery to low- or high-risk for MBT. This information can then be used to reduce costs in the low-risk group by, for example, limiting the use of expensive blood conservation modalities (such as aprotinin or cell salvage and washing), and to reduce the risk of MBT in the high-risk group by, for example, instituting preventive measures (such as reducing hemodilution or prophylactic administration of hemostatic agents).
Limitations of the prediction rule and future direction
Owing to lack of data, we could not assess the predictive value of potentially important variables such as preoperative drugs [except for heparin, coumadin (through measurement of INR), and ASA] and use of synthetic colloids.37 Future studies are required to delineate the effects of these factors on MBT.
Owing to institutional variations in transfusion practice,6 our prediction rule may not be valid at other institutions. The prediction rule, however, did perform well in the validation set, even though patients in the validation set differed from those in the training set on several important variables such as date of surgery and case-mix. Moreover, the variables included in the prediction rule are not site-specific; the same variables have been found to be associated with blood loss in other studies. Although these factors suggest that the prediction rule should be externally generalizable, it needs to be assessed at other institutions.
Future studies are also needed to assess the clinical utility of the prediction rule. Moreover, owing to the low predictive accuracy of the prediction rule in the moderate-risk group, there is a need to investigate strategies that may improve its predictive ability in this group, such as point-of-care coagulation testing prior to termination of CPB.38 Finally, it is important to determine why some patients with no clinical risk factors for MBT nevertheless bleed excessively after cardiac surgery. A potentially important area of investigation in these patients is the role of genotype in bleeding.39
In summary, we developed and internally validated a prediction rule (and a simple risk score) that can accurately predict the risk of MBT in the majority of cardiac surgical patients. Future studies are required to refine the predictive ability of the prediction rule and to assess its external generalizability and clinical utility.
|
Footnotes |
|---|
Funding for this project was provided by the Canadian Institutes of Health Research and Canadian Blood Services through an operating grant, and Novo Nordisk through an unrestricted research grant.
Accepted for publication February 28, 2006. Revision accepted March 13, 2006.
|
References |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
2 Moulton MJ, Creswell LL, Mackey ME, Cox JL, Rosenbloom M. Reexploration for bleeding is a risk factor for adverse outcomes after cardiac operations. J Thorac Cardiovasc Surg 1996; 111: 1037–46.
3 Crabtree TD, Codd JE, Fraser VJ, Bailey MS, Olsen MA, Damiano RJ Jr. Multivariate analysis of risk factors for deep and superficial sternal infection after coronary artery bypass grafting at a tertiary care medical center. Semin Thorac Cardiovasc Surg 2004; 16: 53–61.[Medline]
4 Karkouti K, Wijeysundera DN, Yau TM, et al. The independent association of massive blood loss with mortality in cardiac surgery. Transfusion 2004; 44: 1453–62.[Medline]
5 Karkouti K, Cohen MM, McCluskey SA, Sher GD. A multivariable model for predicting the need for blood transfusion in patients undergoing first-time elective coronary bypass graft surgery. Transfusion 2001; 41: 1193–203.[Medline]
6 Surgenor DM, Churchill WH, Wallace EL, et al. The specific hospital significantly affects red cell and component transfusion practice in coronary artery bypass graft surgery: a study of five hospitals. Transfusion 1998; 38: 122–34.[Medline]
7 Surgenor DM, Churchill WH, Wallace EL, et al. Determinants of red cell, platelet, plasma, and cryoprecipitate transfusions during coronary artery bypass graft surgery: the Collaborative Hospital Transfusion Study. Transfusion 1996; 36: 521–32.[Medline]
8 Bilfinger TV, Conti VR. Blood conservation in coronary artery bypass surgery: prediction with assistance of a computer model. Thorac Cardiovasc Surg 1989; 37: 365–8.[Medline]
9 Despotis GJ, Filos KS, Zoys TN, Hogue CW Jr, Spitznagel E, Lappas DG. Factors associated with excessive postoperative blood loss and hemostatic transfusion requirements: a multivariate analysis in cardiac surgical patients. Anesth Analg 1996; 82: 13–21.[Abstract]
10 Arora RC, Legare JF, Buth KJ, Sullivan JA, Hirsch GM. Identifying patients at risk of intraoperative and postoperative transfusion in isolated CABG: toward selective conservation strategies. Ann Thorac Surg 2004; 78: 1547–55.
11 Cosgrove DM, Loop FD, Lytle BW, et al. Determinants of blood utilization during myocardial revascularization. Ann Thorac Surg 1985; 40: 380–4.[Abstract]
12 Moskowitz DM, Klein JJ, Shander A, et al. Predictors of transfusion requirements for cardiac surgical procedures at a blood conservation center. Ann Thorac Surg 2004; 77: 626–34.
13 Covin R, O’Brien M, Grunwald G, et al. Factors affecting transfusion of fresh frozen plasma, platelets, and red blood cells during elective coronary artery bypass graft surgery. Arch Pathol Lab Med 2003; 127: 415–23.[Medline]
14 Chu MW, Wilson SR, Novick RJ, Stitt LW, Quantz MA. Does clopidogrel increase blood loss following coronary artery bypass surgery? Ann Thorac Surg 2004; 78: 1536–41.
15 Dial S, Delabays E, Albert M, et al. Hemodilution and surgical hemostasis contribute significantly to transfusion requirements in patients undergoing coronary artery bypass. J Thorac Cardiovasc Surg 2005; 130: 654–61.
16 Karthik S, Grayson AD, McCarron EE, Pullan DM, Desmond MJ. Reexploration for bleeding after coronary artery bypass surgery: risk factors, outcomes, and the effect of time delay. Ann Thorac Surg 2004; 78: 527–34.
17 Dacey LJ, Munoz JJ, Baribeau YR, et al. Reexploration for hemorrhage following coronary artery bypass grafting. Incidence and risk factors. Arch Surg 1998; 133: 442–7.
18 Ferraris VA, Gildengorin V. Predictors of excessive blood use after coronary artery bypass grafting. A multivariate analysis. J Thorac Cardiovasc Surg 1989; 98: 492–7.[Abstract]
19 Parr KG, Patel MA, Dekker R, et al. Multivariate predictors of blood product use in cardiac surgery. J Cardiothorac Vasc Anesth 2003; 17: 176–81.[Medline]
20 Goodnough LT, Soegiarso RW, Geha AS. Blood lost and blood transfused in coronary artery bypass graft operation as implications for blood transfusion and blood conservation strategies. Surg Gynecol Obstet 1993; 177: 345–51.[Medline]
21 Hall TS, Brevetti GR, Skoultchi AJ, Sines JC, Gregory P, Spotnitz AJ. Re-exploration for hemorrhage following open heart surgery differentiation on the causes of bleeding and the impact on patient outcomes. Ann Thorac Cardiovasc Surg 2001; 7: 352–7.[Medline]
22 Karkouti K, Beattie WS, Wijeysundera DN, et al. Hemodilution during cardiopulmonary bypass is an independent risk factor for acute renal failure in adult cardiac surgery. J Thorac Cardiovasc Surg 2005; 129: 391–400.
23 Karkouti K, Beattie WS, Wijeysundera DN, et al. Recombinant factor VIIa (rF-VIIa) for intractable blood loss after cardiac surgery: a propensity score-matched case-control analysis. Transfusion 2005; 45: 26–34.[Medline]
24 Karkouti K, Djaiani G, Borger MA, et al. Low hematocrit during cardiopulmonary bypass is associated with increased risk of perioperative stroke in cardiac surgery. Ann Thorac Surg 2005; 80: 1381–7.
25 Feinstein AR. Multiple logistic regression. In: Feinstein AR (Ed.). Multivariable Analysis: An Introduction. New Haven: Yale University Press; 1996: 297–330.
26 Whitlock R, Crowther MA, Ng HJ. Bleeding in cardiac surgery: its prevention and treatment - an evidencebased review. Crit Care Clin 2005; 21: 589–610.[Medline]
27 Diprose P, Herbertson MJ, O’Shaughnessy D, Gill RS. Activated recombinant factor VII after cardiopulmonary bypass reduces allogeneic transfusion in complex non-coronary cardiac surgery: randomized doubleblind placebo-controlled pilot study. Br J Anaesth 2005; 95: 596–602.
28 Brecher ME, Monk T, Goodnough LT. A standardized method for calculating blood loss. Transfusion 1997; 37: 1070–4.[Medline]
29 Woodman RC, Harker LA. Bleeding complications associated with cardiopulmonary bypass. Blood 1990; 76: 1680–97.
30 Despotis GJ, Goodnough LT. Management approaches to platelet-related microvascular bleeding in cardiothoracic surgery. Ann Thorac Surg 2000; 70: S20–32.
31 Green JA, Spiess BD. Current status of antifibrinolytics in cardiopulmonary bypass and elective deep hypothermic circulatory arrest. Anesthesiol Clin North America 2003; 21: 527–51.[Medline]
32 Mora Mangano CT, Neville MJ, Hsu PH, Mignea I, King J, Miller DC. Aprotinin, blood loss, and renal dysfunction in deep hypothermic circulatory arrest. Circulation 2001; 104(Suppl I): 276–81.
33 Nuttall GA, Oliver WC, Ereth MH, Santrach PJ. Coagulation tests predict bleeding after cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1997; 7: 815–23.
34 Anonymous. Practice guidelines for blood component therapy. A report by the American Societyof Anesthesiologists Task Force on Blood Component Therapy. Anesthesiology 1996; 84: 732–47.[Medline]
35 Merritt JC, Bhatt DL. The efficacy and safety of perioperative antiplatelet therapy. J Thromb Thrombolysis 2004; 17: 21–7.[Medline]
36 Bick RL. Disseminated intravascular coagulation. Current concepts of etiology, pathophysiology, diagnosis, and treatment. Hematol Oncol Clin N Am 2003; 17: 149–76.[Medline]
37 Barron ME, Wilkes MM, Navickis RJ. A systematic review of the comparative safety of colloids. Arch Surg 2004; 139: 552–63.
38 Cammerer U, Dietrich W, Rampf T, Braun SL, Richter JA. The predictive value of modified computerized thromboelastography and platelet function analysis for postoperative blood loss in routine cardiac surgery. Anesth Analg 2003; 96: 51–7.
39 Welsby IJ, Podgoreanu MV, Phillips-Bute B, et al.; Perioperative Genetics and Safety Outcomes Study (PEGASUS) Investigative Team. Genetic factors contribute to bleeding after cardiac surgery. J Thromb Haemost 2005; 3: 1206–12.[Medline]
This article has been cited by other articles:
 |
|
 |
|
  M. de Perrot, K. McRae, M. Anraku, K. Karkouti, T. K. Waddell, A. F. Pierre, G. Darling, S. Keshavjee, and M. R. Johnston Risk Factors for Major Complications After Extrapleural Pneumonectomy for Malignant Pleural Mesothelioma Ann. Thorac. Surg., April 1, 2008; 85(4): 1206 - 1210. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Karkouti, W. S. Beattie, M. A. Crowther, J. L. Callum, R. Chun, S. E. Fremes, J. Lemieux, V. C. McAlister, B. D. Muirhead, J. M. Murkin, et al. The role of recombinant factor VIIa in on-pump cardiac surgery: Proceedings of the Canadian Consensus Conference: [Le role du facteur VIIa recombinant dans la chirurgie cardiaque avec circulation extra-corporelle : Rapport de la Conference canadienne de consensus] Can J Anesth, July 1, 2007; 54(7): 573 - 582. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Rodrigues, P. R. B. Evora, S. Bassetto, P. M. Luciano, L. Alves Jr, A. S. Filho, and W. V. de Andrade Vicente Efficacy and Safety of Aprotinin Use for Reoperative Valvular Surgery Ann. Thorac. Surg., June 1, 2007; 83(6): 2060 - 2065. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
