Free
Perioperative Medicine  |   March 2015
Evaluation of a Novel Transfusion Algorithm Employing Point-of-care Coagulation Assays in Cardiac Surgery: A Retrospective Cohort Study with Interrupted Time–Series Analysis
Author Notes
  • From the Department of Anesthesia and Pain Management, and Toronto General Research Institute, Toronto General Hospital, University Health Network, University of Toronto and Institute of Health Policy, Management, and Evaluation, University of Toronto, Toronto, Ontario, Canada (K.K.); Department of Anesthesia (S.A.M.), Department of Laboratory Medicine (R.S.), and Division of Cardiac Surgery, Department of Surgery (V.R.), Toronto General Hospital, University Health Network, University of Toronto, Toronto, Ontario, Canada; Department of Laboratory Medicine and Pathobiology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada (J.C.); Department of Medicine, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada (J.F.); Department of Anesthesia, Université Paris-Diderot, Paris, France (T.T.); and Department of Anesthesia, University of Toronto, Toronto, Ontario, Canada (D.R.).
  • Submitted for publication June 11, 2014. Accepted for publication November 10, 2014.
    Submitted for publication June 11, 2014. Accepted for publication November 10, 2014.×
  • Address correspondence to Dr. Karkouti: Department of Anesthesia, Toronto General Hospital, 200 Elizabeth Street, 3EN, Toronto, Ontario, Canada M5G 2C4. keyvan.karkouti@uhn.ca. Information on purchasing reprints may be found at www.anesthesiology.org or on the masthead page at the beginning of this issue. Anesthesiology’s articles are made freely accessible to all readers, for personal use only, 6 months from the cover date of the issue.
Article Information
Perioperative Medicine / Clinical Science / Cardiovascular Anesthesia / Coagulation and Transfusion / Hematologic System
Perioperative Medicine   |   March 2015
Evaluation of a Novel Transfusion Algorithm Employing Point-of-care Coagulation Assays in Cardiac Surgery: A Retrospective Cohort Study with Interrupted Time–Series Analysis
Anesthesiology 03 2015, Vol.122, 560-570. doi:10.1097/ALN.0000000000000556
Anesthesiology 03 2015, Vol.122, 560-570. doi:10.1097/ALN.0000000000000556
Abstract

Background:: Cardiac surgery requiring the use of cardiopulmonary bypass is frequently complicated by coagulopathic bleeding that, largely due to the shortcomings of conventional coagulation tests, is difficult to manage. This study evaluated a novel transfusion algorithm that uses point-of-care coagulation testing.

Methods:: Consecutive patients who underwent cardiac surgery with bypass at one hospital before (January 1, 2012 to January 6, 2013) and after (January 7, 2013 to December 13, 2013) institution of an algorithm that used the results of point-of-care testing (ROTEM®; Tem International GmBH, Munich, Germany; Plateletworks®; Helena Laboratories, Beaumont, TX) during bypass to guide management of coagulopathy were included. Pre- and postalgorithm outcomes were compared using interrupted time-series analysis to control for secular time trends and other confounders.

Results:: Pre- and postalgorithm groups included 1,311 and 1,170 patients, respectively. Transfusion rates for all blood products (except for cryoprecipitate, which did not change) were decreased after algorithm institution. After controlling for secular pre- and postalgorithm time trends and potential confounders, the posttransfusion odds ratios (95% CIs) for erythrocytes, platelets, and plasma were 0.50 (0.32 to 0.77), 0.22 (0.13 to 0.37), and 0.20 (0.12 to 0.34), respectively. There were no indications that the algorithm worsened any of the measured processes of care or outcomes.

Conclusions:: Institution of a transfusion algorithm based on point-of-care testing was associated with reduced transfusions. This suggests that the algorithm could improve the management of the many patients who develop coagulopathic bleeding after cardiac surgery. The generalizability of the findings needs to be confirmed.

Abstract

Institution of a practical point-of-care-based transfusion algorithm in cardiac surgery with cardiopulmonary bypass was associated with a statistically and clinically significant reduction in blood component transfusion.

What We Already Know about This Topic
  • The coagulopathy of cardiac surgery is thought to be due to a combination of hyperfibrinolysis, platelet dysfunction, and reduced clotting factor levels caused primarily by the contact of blood with the cardiopulmonary bypass circuit and synthetic grafts, cardiopulmonary bypass–related hemodilution, hypothermia, and surgical trauma.

  • Several point-of-care tests are now available that can rapidly identify the specific causes of coagulopathic bleeding. This before-and-after single-center observational study assessed the safety and effectiveness of a point-of-care–based transfusion algorithm.

What This Article Tells Us That Is New
  • Institution of a practical point-of-care–based transfusion algorithm in cardiac surgery with cardiopulmonary bypass was associated with a statistically and clinically significant reduction in blood component transfusion.

CARDIAC surgery requiring the use of cardiopulmonary bypass (CPB) is frequently complicated by coagulopathic bleeding that, largely due to the shortcomings of conventional coagulation tests, is difficult to manage. As a result, the majority of patients receive one or more blood component transfusions. A recent study at 798 U.S. sites found that in low-risk cardiac surgery, the average transfusion rates for erythrocyte, platelets, and plasma were 56, 25, and 19%, respectively.1  These rates are substantially higher in complex surgeries,*01  and up to 20% of patients develop severe coagulopathy that necessitates the transfusion of large amounts of blood components.2 
Blood is an expensive, limited resource and carries serious risks.2–4  Each erythrocyte unit costs U.S. $500 to $1,200 to deliver from the donor to the patient,5  and forecasting models predict that the demand may outstrip supply in the near future.6  Transfusions can lead to life-threatening complications such as infections, acute hemolytic reactions, acute lung injury, and volume overload.2,4,7,8  Moreover, a proportional relation between blood transfusion and mortality has been noted9,10  but not consistently.11 
The coagulopathy of cardiac surgery is thought to be due to a combination of hyperfibrinolysis, platelet dysfunction, and reduced clotting factor levels caused primarily by the contact of blood with the CPB circuit and synthetic grafts, CPB-related hemodilution, hypothermia, and surgical trauma.9,12  Available therapies for this coagulopathy include blood component transfusions and administration of hemostatic agents such as antifibrinolytic drugs and coagulation factor concentrates. Unless the specific causes of coagulopathy are rapidly diagnosed, however, appropriate therapy cannot be instituted. To identify the causes of coagulopathy, clinicians currently rely on conventional laboratory tests that have long turnaround times and cannot detect important coagulation defects such as excessive fibrinolysis, platelet dysfunction, or specific coagulation factor deficiencies.12–14  As a result, clinicians frequently have to delay therapy until the test results become available or resort to empiric therapy based on their clinical judgment. Both these strategies are suboptimal and potentially harmful as they can lead to underuse of blood components in some patients, resulting in excessive blood loss and possibly re-exploration, and to overuse of blood components in others, exposing them to unnecessary risks.9,15 
Several point-of-care (POC) tests are now available that can rapidly identify the specific causes of coagulopathic bleeding.2,16–18  Incorporation of these tests in integrated transfusion algorithms has been shown to reduce transfusions and adverse outcomes.2,16–18  The objective of this before-and-after observational study was to assess the safety and effectiveness of a POC-based transfusion algorithm that was instituted at our hospital in 2013.
Materials and Methods
Study Setting and Population
This is a retrospective analysis of data collected on consecutive patients who underwent cardiac surgery from January 1, 2012 to December 13, 2013 at the Toronto General Hospital. A full range of adult cardiac surgical procedures are performed at this teaching hospital. After approval from institutional research ethics board (University Health Network, Toronto, Ontario, Canada), data were obtained from institutional databases and medical records for all patients who underwent cardiac surgery with CPB during the study period. Full-time research personnel blinded to the details of this study adjudicated all patient outcomes. For patients who were readmitted for additional operations requiring CPB during the study period, only data from their first admission were used.
Transfusion Algorithm
The transfusion algorithm, which was based on existing literature,2,16–28  incorporated POC tests that were conducted during CPB (upon rewarming) to ensure that the results were made available to the clinicians at the end of CPB in a timely manner. The algorithm included an objective measure of ongoing bleeding (weighing of sponges packed in the mediastinum for 5-min after reversal of heparin) to better focus the transfusion of blood products to the coagulopathic patient27  and used a simple stepwise decision tree to encourage standardized component therapy (fig. 1).
Fig. 1.
Transfusion algorithm. Point-of-care (POC) tests consist of ROTEM® (Rotation Thromboelastometry; Tem International GmBH, Munich, Germany) and Plateletworks® (Helena Laboratories, Beaumont, TX) systems. A10-EXTEM (clot amplitude at 10 min in EXTEM assay; at <35 mm implies impaired clot formation due to low fibrinogen levels, low platelet count, or impaired platelet function), A10-FIBTEM (clot firmness amplitude at 10 min in FIBTEM assay; ≤7 mm confirms low fibrinogen levels), and CT-EXTEM (clotting time in EXTEM assay; ≥100 s implied poor clot initiation possibly due to reduced coagulation factor levels or reduced thrombin generation) are ROTEM® measures. Functioning platelet count is obtained by the Plateletworks® assay. During the last month of the study, fibrinogen concentrate was used in place of cryoprecipitate. ACT = activated clotting time; CPB = cardiopulmonary bypass; PCC = prothrombin complex concentrate; RBC = red blood cell; RV = right ventricle.
Transfusion algorithm. Point-of-care (POC) tests consist of ROTEM® (Rotation Thromboelastometry; Tem International GmBH, Munich, Germany) and Plateletworks® (Helena Laboratories, Beaumont, TX) systems. A10-EXTEM (clot amplitude at 10 min in EXTEM assay; at <35 mm implies impaired clot formation due to low fibrinogen levels, low platelet count, or impaired platelet function), A10-FIBTEM (clot firmness amplitude at 10 min in FIBTEM assay; ≤7 mm confirms low fibrinogen levels), and CT-EXTEM (clotting time in EXTEM assay; ≥100 s implied poor clot initiation possibly due to reduced coagulation factor levels or reduced thrombin generation) are ROTEM® measures. Functioning platelet count is obtained by the Plateletworks® assay. During the last month of the study, fibrinogen concentrate was used in place of cryoprecipitate. ACT = activated clotting time; CPB = cardiopulmonary bypass; PCC = prothrombin complex concentrate; RBC = red blood cell; RV = right ventricle.
Fig. 1.
Transfusion algorithm. Point-of-care (POC) tests consist of ROTEM® (Rotation Thromboelastometry; Tem International GmBH, Munich, Germany) and Plateletworks® (Helena Laboratories, Beaumont, TX) systems. A10-EXTEM (clot amplitude at 10 min in EXTEM assay; at <35 mm implies impaired clot formation due to low fibrinogen levels, low platelet count, or impaired platelet function), A10-FIBTEM (clot firmness amplitude at 10 min in FIBTEM assay; ≤7 mm confirms low fibrinogen levels), and CT-EXTEM (clotting time in EXTEM assay; ≥100 s implied poor clot initiation possibly due to reduced coagulation factor levels or reduced thrombin generation) are ROTEM® measures. Functioning platelet count is obtained by the Plateletworks® assay. During the last month of the study, fibrinogen concentrate was used in place of cryoprecipitate. ACT = activated clotting time; CPB = cardiopulmonary bypass; PCC = prothrombin complex concentrate; RBC = red blood cell; RV = right ventricle.
×
The POC tests used in the algorithm were the ROTEM® (Rotation Thromboelastometry; Tem International GmBH, Munich, Germany) and Plateletworks® (Helena Laboratories, Beaumont, TX) systems. Together, the two tests provide a comprehensive assessment of coagulation defects in cardiac surgery, identifying platelet dysfunction, factor deficiencies, and fibrinolysis.16,29–31 
The algorithm was instituted into clinical care for all CPB patients on January 7, 2013. Before its institution, clinicians were made aware of the algorithm and its objectives during educational rounds starting approximately 2 months before institution of the algorithm. Upon institution, two of the investigators (K.K. and S.A.M.) provided on-site support to clinicians for the interpretation of POC tests and algorithm. In November 2013, the algorithm was changed to include the use of fibrinogen concentrate (4 g) in place of cryoprecipitate. POC testing was conducted during CPB by a dedicated technologist at a satellite area close to the operating rooms. Decisions regarding the use of components were driven by the cardiac anesthesiologist assigned to the case. Transfusions were audited and feedback to staff was provided for algorithm deviations.
Clinical Practice
Blood conservation strategies that were used but not standardized throughout the study period (i.e., were not modified by the algorithm) included cell salvage, retrograde autologous priming of the CPB circuit, universal use of the antifibrinolytic drug tranexamic acid, and rescue use of recombinant activated factor VII for refractory bleeding (indications for its use were consistent with Canadian consensus recommendations).32  Antiplatelet and anticoagulant drugs (except for acetylsalicylic acid) were stopped at least 5 days before surgery, when possible, throughout the study period. Data regarding the use of antiplatelet and anticoagulant therapy in the patients were not available and not part of the analysis.
Erythrocyte transfusion triggers, which also were not modified by the algorithm, were a hemoglobin concentration of approximately 7 g/dl during CPB, approximately 8 g/dl in post-CPB, and approximately 9 g/dl in unstable or bleeding patients. Hemoglobin measures throughout the study period were obtained via POC blood gas analyzers that were available in every operating room. Prealgorithm indications for nonerythrocyte components were in accordance with current guidelines2,33  and were guided by standard laboratory tests conducted after protamine administration (complete blood count, international normalized ratio of prothrombin time, activated partial thromboplastin time, and fibrinogen concentration). Indications for platelet transfusion included a platelet count of less than 50 × 109/l or ongoing bleeding after reversal of heparin with a platelet count of less than 80 × 109/l. In bleeding patients with suspected platelet dysfunction (e.g., recent use of antiplatelet drugs or prolonged CPB), platelets were transfused irrespective of platelet counts. In cases of bleeding after full reversal of heparin, plasma (2 to 4 units) was transfused if the international normalized ratio was greater than 1.5 and cryoprecipitate (8 to 10 units) was transfused if the fibrinogen concentration was less than 1.0 g/l. Prothrombin complex concentrate (Octaplex®; Octapharma, Toronto, Ontario, Canada) was administered instead of plasma in select cases of right ventricular failure or severe volume overload. Given the long turnaround time for laboratory tests, plasma and platelets (but not cryoprecipitate) were administered empirically in patients with severe bleeding. There were no major changes to the conduct of CPB practice during the study period.34 
Outcomes
The primary outcomes were the blood component (erythrocyte, platelet, plasma, cryoprecipitate or fibrinogen concentrate, and prothrombin complex concentrate) transfusion rates up to postoperative day 7. We measured transfusions for 7 days (rather than the immediate perioperative period) to determine whether the use of the algorithm was associated with reduced transfusions rather than delayed transfusions. Secondary outcomes were the amount of transfusions, large-volume erythrocyte transfusions (≥4 units) on the day of surgery, rescue use of recombinant activated factor VII, and surgical reexploration. Safety outcomes were acute kidney injury (greater than twofold increase in creatinine concentration or renal replacement therapy up to postoperative day 7) stroke (documented persistent neurological deficit while in hospital), sepsis (based on positive blood culture during hospitalization), sternal infection (superficial signs of infection or deep infection requiring surgical debridement during hospitalization), and death (during hospitalization). Process of care measures included nadir CPB hemoglobin, 24-h chest tube drainage, discharge hemoglobin, and postoperative ventilation time, intensive care unit stay, and hospital stay.
Statistical Analysis
SAS™ version 9.3 (SAS Institute, Inc., Cary, NC) was used for the analyses; P values less than 0.05 were considered statistically significant. Categorical variables were summarized as frequencies with percentages and continuous variables as medians with interquartile ranges or means with SD as appropriate. Patients operated on from January 1, 2012 to January 6, 2013 were included in the prealgorithm group, and those from January 7, 2013 to December 13, 2013 were included in the postalgorithm group. The overall bleeding risk score of the patients was calculated based on a previously developed prediction rule for large-volume transfusion (using age, body surface area, preoperative shock, preoperative platelet count, preoperative hemoglobin concentration, complexity of procedure, urgency, redo, circulatory arrest time, CPB duration, nadir CPB hemoglobin, and preoperative renal dysfunction).36  Fisher exact chi-square and Student t tests (assuming unequal variance between groups) were used to compare patient characteristics, risk status, and outcomes between the two groups.
The primary analysis was performed using multivariable logistic regression modeling with an interrupted time-series autoregressive covariance structure (which assumes that the correlation between two operations is inversely proportional to the time between those two operations). Variables for which the two groups were not balanced (P < 0.30) were included in the analysis. Case-wise deletion from multivariable regression models was used for patients with missing variables. Interrupted time series, which uses segmented regression modeling to determine the effect of the intervention after controlling for pre- and postintervention time trends, was used to account for secular time trends.37,38  As part of this analysis, outcome rates were calculated for each 4-week interval, and time-series plots were constructed for each outcome.
We conducted several sensitivity analyses. First, as the choice of colloids changed during 2013 from 6% hydroxyethyl starch (HES) 130/0.4 (Fresenius Kabi, Richmond Hill, Canada) to 5% human albumin (supplied by the Canadian Blood Services, Toronto, Ontario, Canada) as a result of a Health Canada Advisory released on June 2013,†02  we repeated the multivariable analysis including only patients from the 6 months before and 6 months after the intervention (before HES use was curtailed). Second, because anemic patients frequently require erythrocyte transfusions irrespective of coagulopathic bleeding, we analyzed anemic (hemoglobin <12 g/dl in women and <13 g/dl in men) and nonanemic patients separately. Third, because an increase in cell-salvage and desmopressin use was noted during the study period, we analyzed patients according to the use of cell salvage and desmopressin (separately). Finally, we compared pre- and postalgorithm erythrocyte transfusion rates in patients who underwent off-pump coronary artery bypass grafting. Because the algorithm was not applied to this group, we expected no change in transfusion rates.
Results
During the pre- and postalgorithm periods, 1,311 and 1,170 patients, respectively, underwent cardiac surgery with CPB. Although the groups were similar in many respects, there were important differences, as the postalgorithm group had higher rates of preoperative renal dysfunction, more frequent use of hypothermic circulatory arrest, and higher use of cell salvage and desmopressin, but lower doses of tranexamic acid (tables 1 and 2).
Table 1.
Patient Characteristics and Clinical Status
Patient Characteristics and Clinical Status×
Patient Characteristics and Clinical Status
Table 1.
Patient Characteristics and Clinical Status
Patient Characteristics and Clinical Status×
×
Table 2.
Surgical Parameters
Surgical Parameters×
Surgical Parameters
Table 2.
Surgical Parameters
Surgical Parameters×
×
In unadjusted analyses, erythrocyte, platelet, and plasma transfusion rates were lower in the postalgorithm group, but fibrinogen replacement therapy (primarily with cryoprecipitate and in five patients with fibrinogen concentrate) rates were similar (table 3). Rates of large-volume transfusion, rescue recombinant activated factor VII therapy, and reexploration were also lower in the postalgorithm group (table 3). Adverse events and process of care measures were similar between the two groups (table 3). The results of the multivariable analysis for (that controlled for preoperative platelet count, preoperative stroke or transient ischemic attack, preoperative renal dysfunction, preoperative bleeding risk score, intraoperative intraaortic balloon pump use, circulatory arrest during CPB, cell salvage, desmopressin use, and tranexamic acid dose) were consistent with the unadjusted results (transfusion data shown in figs. 2–5), with the exception of lower rates of acute kidney injury postalgorithm after adjustment for these potential confounders (odds ratio, 0.39; 95% CI, 0.16 to 0.94; P = 0.04). Time-series plots showed that blood component transfusions rates decreased after the algorithm was instituted but gradually increased thereafter (figs. 2–5). When the segmented regression analysis is stratified by patients with high bleeding risk (preoperative bleeding score >2), we can appreciate that for patients at high bleeding risk, the algorithm resulted in lower rates of transfusions and that the effect was sustained in time, with only plasma transfusions showing a significant increase after the algorithm. For patients with low bleeding risk, however, the initial decrease in rates of blood component transfusion after the algorithm and the gradual increase thereafter remained.
Table 3.
Outcomes for the Entire Pre- and Postalgorithm Groups
Outcomes for the Entire Pre- and Postalgorithm Groups×
Outcomes for the Entire Pre- and Postalgorithm Groups
Table 3.
Outcomes for the Entire Pre- and Postalgorithm Groups
Outcomes for the Entire Pre- and Postalgorithm Groups×
×
Fig. 2.
Time-series plots. Erythrocyte transfusion rates up to postoperative day 7. The algorithm was instituted at the beginning of the 13th interval. All odds ratio (OR), 95% CIs, and P values calculated using segmented regression modeling adjusted as described in the Statistical Analysis section. ORs before and after intervention refer to respective changes in outcome per 4-week period. OR for the intervention refers to change in outcome associated with the institution of the algorithm. (A) Entire sample: before intervention: OR, 0.96 (0.92–1.01), P = 0.08; intervention effect: OR, 0.50 (0.32–0.77), P = 0.002; and after intervention: OR, 1.04 (0.97–1.11), P = 0.23. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.90 (0.86–0.94), P < 0.001; intervention effect: OR, 0.51 (0.30–0.86), P = 0.01; and after intervention: OR, 1.15 (1.07–1.14), P < 0.001. High bleeding risk: before intervention: OR, 1.08 (1.03–1.15), P = 0.005; intervention effect: OR, 0.46 (0.23–0.95), P = 0.04; and after intervention: OR, 0.89 (0.81–0.98), P = 0.02.
Time-series plots. Erythrocyte transfusion rates up to postoperative day 7. The algorithm was instituted at the beginning of the 13th interval. All odds ratio (OR), 95% CIs, and P values calculated using segmented regression modeling adjusted as described in the Statistical Analysis section. ORs before and after intervention refer to respective changes in outcome per 4-week period. OR for the intervention refers to change in outcome associated with the institution of the algorithm. (A) Entire sample: before intervention: OR, 0.96 (0.92–1.01), P = 0.08; intervention effect: OR, 0.50 (0.32–0.77), P = 0.002; and after intervention: OR, 1.04 (0.97–1.11), P = 0.23. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.90 (0.86–0.94), P < 0.001; intervention effect: OR, 0.51 (0.30–0.86), P = 0.01; and after intervention: OR, 1.15 (1.07–1.14), P < 0.001. High bleeding risk: before intervention: OR, 1.08 (1.03–1.15), P = 0.005; intervention effect: OR, 0.46 (0.23–0.95), P = 0.04; and after intervention: OR, 0.89 (0.81–0.98), P = 0.02.
Fig. 2.
Time-series plots. Erythrocyte transfusion rates up to postoperative day 7. The algorithm was instituted at the beginning of the 13th interval. All odds ratio (OR), 95% CIs, and P values calculated using segmented regression modeling adjusted as described in the Statistical Analysis section. ORs before and after intervention refer to respective changes in outcome per 4-week period. OR for the intervention refers to change in outcome associated with the institution of the algorithm. (A) Entire sample: before intervention: OR, 0.96 (0.92–1.01), P = 0.08; intervention effect: OR, 0.50 (0.32–0.77), P = 0.002; and after intervention: OR, 1.04 (0.97–1.11), P = 0.23. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.90 (0.86–0.94), P < 0.001; intervention effect: OR, 0.51 (0.30–0.86), P = 0.01; and after intervention: OR, 1.15 (1.07–1.14), P < 0.001. High bleeding risk: before intervention: OR, 1.08 (1.03–1.15), P = 0.005; intervention effect: OR, 0.46 (0.23–0.95), P = 0.04; and after intervention: OR, 0.89 (0.81–0.98), P = 0.02.
×
Fig. 3.
Time-series plots. Platelet transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.95 (0.91–0.99), P = 0.02; intervention effect: OR, 0.22 (0.13–0.37), P < 0.001; and after intervention: OR, 1.16 (1.08–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.88 (0.84–0.92), P < 0.001; intervention effect: OR, 0.30 (0.15–0.60), P = 0.01; and after intervention: OR, 1.27 (1.16–1.40), P < 0.001. High bleeding risk: before intervention: OR, 1.04 (0.99–1.09), P = 0.16; intervention effect: OR, 0.17 (0.08–0.34), P < 0.001; and after intervention: OR, 1.06 (0.96–1.16), P = 0.27.
Time-series plots. Platelet transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.95 (0.91–0.99), P = 0.02; intervention effect: OR, 0.22 (0.13–0.37), P < 0.001; and after intervention: OR, 1.16 (1.08–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.88 (0.84–0.92), P < 0.001; intervention effect: OR, 0.30 (0.15–0.60), P = 0.01; and after intervention: OR, 1.27 (1.16–1.40), P < 0.001. High bleeding risk: before intervention: OR, 1.04 (0.99–1.09), P = 0.16; intervention effect: OR, 0.17 (0.08–0.34), P < 0.001; and after intervention: OR, 1.06 (0.96–1.16), P = 0.27.
Fig. 3.
Time-series plots. Platelet transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.95 (0.91–0.99), P = 0.02; intervention effect: OR, 0.22 (0.13–0.37), P < 0.001; and after intervention: OR, 1.16 (1.08–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.88 (0.84–0.92), P < 0.001; intervention effect: OR, 0.30 (0.15–0.60), P = 0.01; and after intervention: OR, 1.27 (1.16–1.40), P < 0.001. High bleeding risk: before intervention: OR, 1.04 (0.99–1.09), P = 0.16; intervention effect: OR, 0.17 (0.08–0.34), P < 0.001; and after intervention: OR, 1.06 (0.96–1.16), P = 0.27.
×
Fig. 4.
Time-series plots. Plasma transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.91 (0.88–0.94), P < 0.001; intervention effect: OR, 0.20 (0.12–0.34), P < 0.001; and after intervention: OR, 1.15 (1.07–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.85 (0.81–0.89), P < 0.001; intervention effect: OR, 0.34 (0.18–0.78), P = 0.01; and after intervention: OR, 1.22 (1.08–1.37), P = 0.001. High bleeding risk: before intervention: OR, 0.99 (0.94–1.03), P = 0.49; intervention effect: OR, 0.14 (0.08–0.27), P < 0.001; and after intervention: OR, 1.09 (1.00–1.19), P = 0.05.
Time-series plots. Plasma transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.91 (0.88–0.94), P < 0.001; intervention effect: OR, 0.20 (0.12–0.34), P < 0.001; and after intervention: OR, 1.15 (1.07–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.85 (0.81–0.89), P < 0.001; intervention effect: OR, 0.34 (0.18–0.78), P = 0.01; and after intervention: OR, 1.22 (1.08–1.37), P = 0.001. High bleeding risk: before intervention: OR, 0.99 (0.94–1.03), P = 0.49; intervention effect: OR, 0.14 (0.08–0.27), P < 0.001; and after intervention: OR, 1.09 (1.00–1.19), P = 0.05.
Fig. 4.
Time-series plots. Plasma transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.91 (0.88–0.94), P < 0.001; intervention effect: OR, 0.20 (0.12–0.34), P < 0.001; and after intervention: OR, 1.15 (1.07–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.85 (0.81–0.89), P < 0.001; intervention effect: OR, 0.34 (0.18–0.78), P = 0.01; and after intervention: OR, 1.22 (1.08–1.37), P = 0.001. High bleeding risk: before intervention: OR, 0.99 (0.94–1.03), P = 0.49; intervention effect: OR, 0.14 (0.08–0.27), P < 0.001; and after intervention: OR, 1.09 (1.00–1.19), P = 0.05.
×
Fig. 5.
Time-series plots, large-volume (≥4 units) erythrocyte transfusion rate on the day of surgery. (A) Entire sample: before intervention: odds ratio (OR), 0.94 (0.89–0.99), P = 0.03; intervention effect: OR, 0.23 (0.11–0.48), P < 0.001; and after intervention: OR, 1.14 (1.01–1.28), P = 0.03. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.77 (0.70–0.84), P < 0.001; intervention effect: OR, 0.25 (0.02–3.04), P = 0.27; and after intervention: OR, 1.62 (1.20–2.20), P = 0.002. High bleeding risk: before intervention: OR, 0.99 (0.94–1.05), P = 0.84; intervention effect: OR, 0.25 (0.12–0.50), P < 0.001; and after intervention: OR, 1.07 (0.96–1.19), P = 0.23.
Time-series plots, large-volume (≥4 units) erythrocyte transfusion rate on the day of surgery. (A) Entire sample: before intervention: odds ratio (OR), 0.94 (0.89–0.99), P = 0.03; intervention effect: OR, 0.23 (0.11–0.48), P < 0.001; and after intervention: OR, 1.14 (1.01–1.28), P = 0.03. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.77 (0.70–0.84), P < 0.001; intervention effect: OR, 0.25 (0.02–3.04), P = 0.27; and after intervention: OR, 1.62 (1.20–2.20), P = 0.002. High bleeding risk: before intervention: OR, 0.99 (0.94–1.05), P = 0.84; intervention effect: OR, 0.25 (0.12–0.50), P < 0.001; and after intervention: OR, 1.07 (0.96–1.19), P = 0.23.
Fig. 5.
Time-series plots, large-volume (≥4 units) erythrocyte transfusion rate on the day of surgery. (A) Entire sample: before intervention: odds ratio (OR), 0.94 (0.89–0.99), P = 0.03; intervention effect: OR, 0.23 (0.11–0.48), P < 0.001; and after intervention: OR, 1.14 (1.01–1.28), P = 0.03. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.77 (0.70–0.84), P < 0.001; intervention effect: OR, 0.25 (0.02–3.04), P = 0.27; and after intervention: OR, 1.62 (1.20–2.20), P = 0.002. High bleeding risk: before intervention: OR, 0.99 (0.94–1.05), P = 0.84; intervention effect: OR, 0.25 (0.12–0.50), P < 0.001; and after intervention: OR, 1.07 (0.96–1.19), P = 0.23.
×
A number of subgroup analyses (anemic patients, cell saver use, desmopressin use, and operations performed within 6 months of algorithm start) were performed and are reported in table 4. Results for all subgroups were consistent. Off-pump surgery was conducted on 65 and 62 patients during the pre- and postalgorithm periods, respectively. The erythrocyte transfusion rate was similar (34% prealgorithm vs. 45% postalgorithm; P = 0.27).
Table 4.
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm×
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm
Table 4.
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm×
×
Discussion
In this before-and-after observational study, we found that use of a practical POC-based transfusion algorithm at a single hospital was associated with a substantial reduction in blood component transfusions. After controlling for between-group differences and secular time trends, we found that institution of the algorithm was associated with a substantial reduction in transfusion rates for erythrocytes (odds ratio, 0.50; 95% CI, 0.32 to 0.77; P = 0.002), platelets (odds ratio, 0.22; 95% CI, 0.13 to 0.37; P = 0.002), and plasma (odds ratio, 0.20; 95% CI, 0.12 to 0.34; P = 0.002). There were no indications that the algorithm worsened any of the measured processes of care or outcomes.
There is conflicting evidence on the benefits of POC-based transfusion algorithms. A meta-analysis of nine randomized trials (with 776 participants) comparing POC-based algorithms with conventional management found that the former modestly reduced blood loss (mean difference 85 ml; 95% CI, 29 to 141 ml) but had no effect on transfusions or other outcomes.39  These findings, however, were inconclusive because they were derived from a small number of mostly low-quality randomized trials that had small sample sizes.
Subsequent studies have had more encouraging results. In a large before-and-after observational study, Görlinger et al.24  assessed the benefits of a POC-based algorithm in cardiac surgery (instituted in 2005) by comparing the outcomes of all patients operated on during 2004 (n = 1,718) with those operated on in 2009 (n = 2,147). They found reduced erythrocyte (50 vs. 40%; P < 0.0001) and plasma transfusions (19 vs. 1%; P < 0.0001), but platelet transfusions increased (10 vs. 13%; P = 0.004). Of note, they used a complex algorithm that used factor concentrates instead of blood components as first-line therapy; thus, their results may not be generalizable. Other limitations are the 5-yr time span between the groups and failure to control for secular time trends.
In a randomized study that included 100 patients, Weber et al.26  found that a POC-based transfusion algorithm, compared with an algorithm that used standard laboratory tests, substantially reduced transfusion rates and adverse outcomes such as acute renal failure (6% in the intervention arm vs. 20% in the control arm; P = 0.07), sepsis (2 vs. 14%; P = 0.06), and mortality (4 vs. 20%; P = 0.01). The limitations of this study were that it was a small single-center, unblinded study that used a complex algorithm.
Thus, emerging evidence suggests that POC-based transfusion algorithms may improve the management of coagulopathic bleeding in cardiac surgery, thereby reducing blood transfusions and possibly improving clinical outcomes. Our results add to this evidence. Compared with previous studies, we used a simpler algorithm that did not include factor concentrates as first-line therapy; thus, it should be more applicable to the standard North American approach to postcardiac surgery bleeding. Moreover, to better rationalize treatment strategy and to avoid unnecessary blood component transfusions, we adopted an objective measure of ongoing bleeding (5-min bleeding mass measured after reversal of heparin) that had previously only been reported in clinical studies.21  Although the hemostatic contribution of the 5-min bleeding mass to the overall results cannot be determined, it is unlikely that this measure can by itself (i.e., without the information provided by the POC tests) have a substantial effect on blood loss or transfusion rates.
It is noteworthy that despite a reduction in prognostically important outcomes such as reexploration, large-volume transfusion, and acute kidney injury, the algorithm was not associated with reduced length of hospitalization or mortality. Because the indications for reexploration did not change during the study period, one explanation for the reduced rates would be improved management of coagulopathy due to the introduction of the algorithm. Regarding acute kidney injury, we have previously postulated that stored erythrocytes may harm the kidneys because of the changes that they undergo during storage that reduces their posttransfusion functionality and viability.34,40  The lack of effect on length of hospitalization or mortality may be due to a lack of power (i.e., based on our results, a study would need to include approximately 15,000 participants to be adequately powered to determine whether reducing large-volume transfusion by 50% reduces mortality) or short duration of follow-up. Another possibility is that once a potentially fatal complication has occurred, better management of the ensuing coagulopathy and avoidance of large-volume transfusion may not be sufficient to prevent death.
This study has several limitations. Because it was a retrospective observational study without a concurrent control group, it cannot prove causality. It is possible that some unmeasured factors or simultaneous cointerventions may have been responsible for the improved outcomes.41  There were several important practice changes during the study period that may have influenced our findings. First, the change in colloid use from HES 130/0.4 to albumin may explain some of our findings because HES has been shown to impair coagulation and renal function (but primarily in sepsis).42,43  This change in practice, however, likely did not affect our results because HES volumes were limited to 1 l or less in our practice, HES 130/0.4 and albumin have been shown to have similar effects on blood loss and renal function in patients undergoing on-pump cardiac surgery,44,45  and our results were consistent when we limited our sample to the period before HES use was discontinued. We could not adjust for the colloid type used because the data were not available. We identified two other practice changes not specified by the algorithm. One was an increase in the use of desmopressin after the algorithm was instituted. Despite not being part of the algorithm, the ability to diagnose platelet dysfunction seems to have led clinicians to increase the use of this drug. Controlling for the confounding effect of desmopressin usage did not influence our results. This is not surprising given that there is a lack of definitive data on the efficacy of desmopressin in cardiac surgery.46  The other change that occurred during the study period was an increase in the use of cell savage, which in cardiac surgery has been shown to reduce erythrocyte transfusions, but not any other blood components.47  Cell saver use was at the discretion of the surgeon, and during the past several years, our use of cell saver has consistently increased as we have gotten more equipment and funding for its use. Controlling for the confounding effect of cell saver usage also did not influence our results. Of note, the algorithm was not associated with reduced nadir hemoglobin concentration during CPB or discharge hemoglobin, suggesting that reduced erythrocyte transfusions were not due to changes in hemoglobin transfusion triggers.
A general limitation of before-and-after studies is that the comparison of overall mean rates before and after the intervention may be biased by secular time trends.37  We used interrupted time-series modeling to control for this potential bias (as well as other potential confounders) and found that the immediate effects of the intervention remained significant. We also conducted stratified analysis based on patients’ underlying bleeding risk and found that the effects of the intervention for nonerythrocyte blood products were more pronounced in this high-risk group. We noted a slight uptrend in blood component transfusion rates during the postintervention period. Longer-term studies are needed to explore the importance of this finding. Finally, generalizability is an issue because this was a single-hospital study, and there is wide variability in transfusion practices and rates in cardiac surgery across hospitals.1,48 
In conclusion, this before-and-after single-center observational study found that the institution of a practical POC-based transfusion algorithm in cardiac surgery with CPB was associated with a statistically and clinically significant reduction in blood component transfusion. These results suggest that the clinical adoption of POC-based algorithms could improve the management of the many patients who develop coagulopathic bleeding after cardiac surgery. The generalizability of our findings, however, needs to be confirmed by future trials.
Acknowledgments
The study could not have been completed without the tireless work of Cielo Bingley R.N., OnTraC (Ontario Nurse Transfusion Coordinators Provincial Blood Conservation) blood conservation program, Toronto General Hospital, University Health Network, Toronto, Ontario, Canada; and Jo Carroll, R.N., Department of Anesthesia, Toronto General Hospital, University Health Network.
This study was funded by an Innovation Grant from the Peter Munk Cardiac Centre at the Toronto General Hospital, University Health Network, Toronto, Ontario, Canada. Funded in part by a Merit Award from the Department of Anesthesia at the University of Toronto, Toronto, Ontario, Canada (to Dr. Karkouti).
Competing Interests
Drs. Karkouti, Callum, and Rao have received research funding from Tem International GmBH (Munich, Germany) and Helena Laboratories (Beaumont, Texas) for an ongoing multicenter randomized trial of a POC-based coagulation algorithm. The other authors declare no competing interests.
References
Bennett-Guerrero, E, Zhao, Y, O’Brien, SM, Ferguson, TBJr, Peterson, ED, Gammie, JS, Song, HK Variation in use of blood transfusion in coronary artery bypass graft surgery.. JAMA. (2010). 304 1568–75 [Article] [PubMed]
Ferraris, VA, Ferraris, SP, Saha, SP, Hessel, EA, Haan, CK, Royston, BD, Bridges, CR, Higgins, RS, Despotis, G, Brown, JR, Spiess, BD, Shore-Lesserson, L, Stafford-Smith, M, Mazer, CD, Bennett-Guerrero, E, Hill, SE, Body, S Perioperative blood transfusion and blood conservation in cardiac surgery: The Society of Thoracic Surgeons and The Society of Cardiovascular Anesthesiologists clinical practice guideline.. Ann Thorac Surg. (2007). 83 S27–86 [Article] [PubMed]
Williamson, LM, Devine, DV Challenges in the management of the blood supply.. Lancet. (2013). 381 1866–75 [Article] [PubMed]
Goodnough, LT Blood management: Transfusion medicine comes of age.. Lancet. (2013). 381 1791–2 [Article] [PubMed]
Shander, A, Hofmann, A, Ozawa, S, Theusinger, OM, Gombotz, H, Spahn, DR Activity-based costs of blood transfusions in surgical patients at four hospitals.. Transfusion. (2010). 50 753–65 [Article] [PubMed]
Drackley, A, Newbold, KB, Paez, A, Heddle, N Forecasting Ontario’s blood supply and demand.. Transfusion. (2012). 52 366–74 [Article] [PubMed]
Rohde, JM, Dimcheff, DE, Blumberg, N, Saint, S, Langa, KM, Kuhn, L, Hickner, A, Rogers, MA Health care-associated infection after red blood cell transfusion: A systematic review and meta-analysis.. JAMA. (2014). 311 1317–26 [Article] [PubMed]
Vamvakas, EC, Blajchman, MA Blood still kills: Six strategies to further reduce allogeneic blood transfusion-related mortality.. Transfus Med Rev. (2010). 24 77–124 [Article] [PubMed]
Despotis, G, Eby, C, Lublin, DM A review of transfusion risks and optimal management of perioperative bleeding with cardiac surgery.. Transfusion. (2008). 481 suppl 2S–30 [Article] [PubMed]
Karkouti, K, Wijeysundera, DN, Yau, TM, Beattie, WS, Abdelnaem, E, McCluskey, SA, Ghannam, M, Yeo, E, Djaiani, G, Karski, J The independent association of massive blood loss with mortality in cardiac surgery.. Transfusion. (2004). 44 1453–62 [Article] [PubMed]
Karkouti, K, Stukel, TA, Beattie, WS, Elsaadany, S, Li, P, Berger, R, Wijeysundera, DN Relationship of erythrocyte transfusion with short- and long-term mortality in a population-based surgical cohort.. Anesthesiology. (2012). 117 1175–83 [Article] [PubMed]
Johansson, PI, Sølbeck, S, Genet, G, Stensballe, J, Ostrowski, SR Coagulopathy and hemostatic monitoring in cardiac surgery: An update.. Scand Cardiovasc J. (2012). 46 194–202 [Article] [PubMed]
Haas, T, Spielmann, N, Mauch, J, Madjdpour, C, Speer, O, Schmugge, M, Weiss, M Comparison of thromboelastometry (ROTEM®) with standard plasmatic coagulation testing in paediatric surgery.. Br J Anaesth. (2012). 108 36–41 [Article] [PubMed]
Weber, CF, Klages, M, Zacharowski, K Perioperative coagulation management during cardiac surgery.. Curr Opin Anaesthesiol. (2013). 26 60–4 [Article] [PubMed]
Avidan, MS, Alcock, EL, Da Fonseca, J, Ponte, J, Desai, JB, Despotis, GJ, Hunt, BJ Comparison of structured use of routine laboratory tests or near-patient assessment with clinical judgement in the management of bleeding after cardiac surgery.. Br J Anaesth. (2004). 92 178–86 [Article] [PubMed]
Görlinger, K, Shore-Lesserson, L, Dirkmann, D, Hanke, AA, Rahe-Meyer, N, Tanaka, KA Management of hemorrhage in cardiothoracic surgery.. J Cardiothorac Vasc Anesth. (2013). 274 suppl S20–34 [Article] [PubMed]
Schöchl, H, Voelckel, W, Grassetto, A, Schlimp, CJ Practical application of point-of-care coagulation testing to guide treatment decisions in trauma.. J Trauma Acute Care Surg. (2013). 74 1587–98 [Article] [PubMed]
Görlinger, K, Dirkmann, D, Hanke, AA Potential value of transfusion protocols in cardiac surgery.. Curr Opin Anaesthesiol. (2013). 26 230–43 [Article] [PubMed]
Solomon, C, Cadamuro, J, Ziegler, B, Schöchl, H, Varvenne, M, Sørensen, B, Hochleitner, G, Rahe-Meyer, N A comparison of fibrinogen measurement methods with fibrin clot elasticity assessed by thromboelastometry, before and after administration of fibrinogen concentrate in cardiac surgery patients.. Transfusion. (2011). 51 1695–706 [Article] [PubMed]
Martini, J, Maisch, S, Pilshofer, L, Streif, W, Martini, W, Fries, D Fibrinogen concentrate in dilutional coagulopathy: A dose study in pigs.. Transfusion. (2014). 54 149–57 [Article] [PubMed]
Rahe-Meyer, N, Hanke, A, Schmidt, DS, Hagl, C, Pichlmaier, M Fibrinogen concentrate reduces intraoperative bleeding when used as first-line hemostatic therapy during major aortic replacement surgery: Results from a randomized, placebo-controlled trial.. J Thorac Cardiovasc Surg. (2013). 1453 suppl S178–85 [Article] [PubMed]
Chandler, WL, Velan, T Estimating the rate of thrombin and fibrin generation in vivo during cardiopulmonary bypass.. Blood. (2003). 101 4355–62 [Article] [PubMed]
Solomon, C, Rahe-Meyer, N, Sørensen, B Fibrin formation is more impaired than thrombin generation and platelets immediately following cardiac surgery.. Thromb Res. (2011). 128 277–82 [Article] [PubMed]
Görlinger, K, Dirkmann, D, Hanke, AA, Kamler, M, Kottenberg, E, Thielmann, M, Jakob, H, Peters, J First-line therapy with coagulation factor concentrates combined with point-of-care coagulation testing is associated with decreased allogeneic blood transfusion in cardiovascular surgery: A retrospective, single-center cohort study.. Anesthesiology. (2011). 115 1179–91 [PubMed]
Görlinger, K, Dirkmann, D, Solomon, C, Hanke, AA Fast interpretation of thromboelastometry in non-cardiac surgery: Reliability in patients with hypo-, normo-, and hypercoagulability.. Br J Anaesth. (2013). 110 222–30 [Article] [PubMed]
Weber, CF, Görlinger, K, Meininger, D, Herrmann, E, Bingold, T, Moritz, A, Cohn, LH, Zacharowski, K Point-of-care testing: A prospective, randomized clinical trial of efficacy in coagulopathic cardiac surgery patients.. Anesthesiology. (2012). 117 531–47 [Article] [PubMed]
Rahe-Meyer, N, Solomon, C, Hanke, A, Schmidt, DS, Knoerzer, D, Hochleitner, G, Sørensen, B, Hagl, C, Pichlmaier, M Effects of fibrinogen concentrate as first-line therapy during major aortic replacement surgery: A randomized, placebo-controlled trial.. Anesthesiology. (2013). 118 40–50 [Article] [PubMed]
Orlov, D, McCluskey, SA, Selby, R, Yip, P, Pendergrast, J, Karkouti, K Platelet dysfunction as measured by a point-of-care monitor is an independent predictor of high blood loss in cardiac surgery.. Anesth Analg. (2014). 118 257–63 [Article] [PubMed]
Breet, NJ, van Werkum, JW, Bouman, HJ, Kelder, JC, Ruven, HJ, Bal, ET, Deneer, VH, Harmsze, AM, van der Heyden, JA, Rensing, BJ, Suttorp, MJ, Hackeng, CM, ten Berg, JM Comparison of platelet function tests in predicting clinical outcome in patients undergoing coronary stent implantation.. JAMA. (2010). 303 754–62 [Article] [PubMed]
Dalén, M, van der Linden, J, Lindvall, G, Ivert, T Correlation between point-of-care platelet function testing and bleeding after coronary artery surgery.. Scand Cardiovasc J. (2012). 46 32–8 [Article] [PubMed]
van Werkum, JW, Kleibeuker, M, Postma, S, Bouman, HJ, Elsenberg, EH, ten Berg, JM, Hackeng, CM A comparison between the Plateletworks-assay and light transmittance aggregometry for monitoring the inhibitory effects of clopidogrel.. Int J Cardiol. (2010). 140 123–6 [Article] [PubMed]
Karkouti, K, Beattie, WS, Crowther, MA, Callum, JL, Chun, R, Fremes, SE, Lemieux, J, McAlister, VC, Muirhead, BD, Murkin, JM, Nathan, HJ, Wong, BI, Yau, TM, Yeo, EL, Hall, RI The role of recombinant factor VIIa in on-pump cardiac surgery: Proceedings of the Canadian Consensus Conference.. Can J Anaesth. (2007). 54 573–82 [Article] [PubMed]
Practice guidelines for perioperative blood transfusion and adjuvant therapies: An updated report by the American Society of Anesthesiologists Task Force on Perioperative Blood Transfusion and Adjuvant Therapies.. Anesthesiology. (2006). 105 198–208 [Article] [PubMed]
Karkouti, K, Wijeysundera, DN, Yau, TM, McCluskey, SA, Chan, CT, Wong, PY, Crowther, MA, Hozhabri, S, Beattie, WS Advance targeted transfusion in anemic cardiac surgical patients for kidney protection: An unblinded randomized pilot clinical trial.. Anesthesiology. (2012). 116 613–21 [Article] [PubMed]
Cockcroft, DW, Gault, MH Prediction of creatinine clearance from serum creatinine.. Nephron. (1976). 16 31–41 [Article] [PubMed]
Karkouti, K, O’Farrell, R, Yau, TM, Beattie, WS Reducing Bleeding in Cardiac Surgery Research Group, Prediction of massive blood transfusion in cardiac surgery.. Can J Anaesth. (2006). 53 781–94 [Article] [PubMed]
Ramsay, CR, Matowe, L, Grilli, R, Grimshaw, JM, Thomas, RE Interrupted time series designs in health technology assessment: Lessons from two systematic reviews of behavior change strategies.. Int J Technol Assess Health Care. (2003). 19 613–23 [Article] [PubMed]
Penfold, RB, Zhang, F Use of interrupted time series analysis in evaluating health care quality improvements.. Acad Pediatr. (2013). 136 suppl S38–44 [Article] [PubMed]
Afshari, A, Wikkelso, A, Brok, J, Moller, AM, Wetterslev, J Thrombelastography (TEG) or thromboelastometry (ROTEM) to monitor haemotherapy versus usual care in patients with massive transfusion.. Cochrane Database Syst Rev. (2011). CD007871
Karkouti, K Transfusion and risk of acute kidney injury in cardiac surgery.. Br J Anaesth. (2012). 109suppl 1 i29–38 [Article] [PubMed]
Wagner, AK, Soumerai, SB, Zhang, F, Ross-Degnan, D Segmented regression analysis of interrupted time series studies in medication use research.. J Clin Pharm Ther. (2002). 27 299–309 [Article] [PubMed]
Zarychanski, R, Abou-Setta, AM, Turgeon, AF, Houston, BL, McIntyre, L, Marshall, JC, Fergusson, DA Association of hydroxyethyl starch administration with mortality and acute kidney injury in critically ill patients requiring volume resuscitation: A systematic review and meta-analysis.. JAMA. (2013). 309 678–88 [Article] [PubMed]
Haase, N, Perner, A, Hennings, LI, Siegemund, M, Lauridsen, B, Wetterslev, M, Wetterslev, J Hydroxyethyl starch 130/0.38-0.45 versus crystalloid or albumin in patients with sepsis: Systematic review with meta-analysis and trial sequential analysis.. BMJ. (2013). 346 f839 [Article] [PubMed]
Skhirtladze, K, Base, EM, Lassnigg, A, Kaider, A, Linke, S, Dworschak, M, Hiesmayr, MJ Comparison of the effects of albumin 5%, hydroxyethyl starch 130/0.4 6%, and Ringer’s lactate on blood loss and coagulation after cardiac surgery.. Br J Anaesth. (2014). 112 255–64 [Article] [PubMed]
Van der Linden, P, De Villé, A, Hofer, A, Heschl, M, Gombotz, H Six percent hydroxyethyl starch 130/0.4 (Voluven®) versus 5% human serum albumin for volume replacement therapy during elective open-heart surgery in pediatric patients.. Anesthesiology. (2013). 119 1296–309 [Article] [PubMed]
Ferraris, VA, Brown, JR, Despotis, GJ, Hammon, JW, Reece, TB, Saha, SP, Song, HK, Clough, ER, Shore-Lesserson, LJ, Goodnough, LT, Mazer, CD, Shander, A, Stafford-Smith, M, Waters, J, Baker, RA, Dickinson, TA, FitzGerald, DJ, Likosky, DS, Shann, KG 2011 update to the Society of Thoracic Surgeons and the Society of Cardiovascular Anesthesiologists blood conservation clinical practice guidelines.. Ann Thorac Surg. (2011). 91 944–82 [Article] [PubMed]
Wang, G, Bainbridge, D, Martin, J, Cheng, D The efficacy of an intraoperative cell saver during cardiac surgery: A meta-analysis of randomized trials.. Anesth Analg. (2009). 109 320–30 [Article] [PubMed]
Karkouti, K, Wijeysundera, DN, Beattie, WS, Callum, JL, Cheng, D, Dupuis, JY, Kent, B, Mazer, D, Rubens, FD, Sawchuk, C, Yau, TM Variability and predictability of large-volume red blood cell transfusion in cardiac surgery: A multicenter study.. Transfusion. (2007). 47 2081–8 [Article] [PubMed]
Fig. 1.
Transfusion algorithm. Point-of-care (POC) tests consist of ROTEM® (Rotation Thromboelastometry; Tem International GmBH, Munich, Germany) and Plateletworks® (Helena Laboratories, Beaumont, TX) systems. A10-EXTEM (clot amplitude at 10 min in EXTEM assay; at <35 mm implies impaired clot formation due to low fibrinogen levels, low platelet count, or impaired platelet function), A10-FIBTEM (clot firmness amplitude at 10 min in FIBTEM assay; ≤7 mm confirms low fibrinogen levels), and CT-EXTEM (clotting time in EXTEM assay; ≥100 s implied poor clot initiation possibly due to reduced coagulation factor levels or reduced thrombin generation) are ROTEM® measures. Functioning platelet count is obtained by the Plateletworks® assay. During the last month of the study, fibrinogen concentrate was used in place of cryoprecipitate. ACT = activated clotting time; CPB = cardiopulmonary bypass; PCC = prothrombin complex concentrate; RBC = red blood cell; RV = right ventricle.
Transfusion algorithm. Point-of-care (POC) tests consist of ROTEM® (Rotation Thromboelastometry; Tem International GmBH, Munich, Germany) and Plateletworks® (Helena Laboratories, Beaumont, TX) systems. A10-EXTEM (clot amplitude at 10 min in EXTEM assay; at <35 mm implies impaired clot formation due to low fibrinogen levels, low platelet count, or impaired platelet function), A10-FIBTEM (clot firmness amplitude at 10 min in FIBTEM assay; ≤7 mm confirms low fibrinogen levels), and CT-EXTEM (clotting time in EXTEM assay; ≥100 s implied poor clot initiation possibly due to reduced coagulation factor levels or reduced thrombin generation) are ROTEM® measures. Functioning platelet count is obtained by the Plateletworks® assay. During the last month of the study, fibrinogen concentrate was used in place of cryoprecipitate. ACT = activated clotting time; CPB = cardiopulmonary bypass; PCC = prothrombin complex concentrate; RBC = red blood cell; RV = right ventricle.
Fig. 1.
Transfusion algorithm. Point-of-care (POC) tests consist of ROTEM® (Rotation Thromboelastometry; Tem International GmBH, Munich, Germany) and Plateletworks® (Helena Laboratories, Beaumont, TX) systems. A10-EXTEM (clot amplitude at 10 min in EXTEM assay; at <35 mm implies impaired clot formation due to low fibrinogen levels, low platelet count, or impaired platelet function), A10-FIBTEM (clot firmness amplitude at 10 min in FIBTEM assay; ≤7 mm confirms low fibrinogen levels), and CT-EXTEM (clotting time in EXTEM assay; ≥100 s implied poor clot initiation possibly due to reduced coagulation factor levels or reduced thrombin generation) are ROTEM® measures. Functioning platelet count is obtained by the Plateletworks® assay. During the last month of the study, fibrinogen concentrate was used in place of cryoprecipitate. ACT = activated clotting time; CPB = cardiopulmonary bypass; PCC = prothrombin complex concentrate; RBC = red blood cell; RV = right ventricle.
×
Fig. 2.
Time-series plots. Erythrocyte transfusion rates up to postoperative day 7. The algorithm was instituted at the beginning of the 13th interval. All odds ratio (OR), 95% CIs, and P values calculated using segmented regression modeling adjusted as described in the Statistical Analysis section. ORs before and after intervention refer to respective changes in outcome per 4-week period. OR for the intervention refers to change in outcome associated with the institution of the algorithm. (A) Entire sample: before intervention: OR, 0.96 (0.92–1.01), P = 0.08; intervention effect: OR, 0.50 (0.32–0.77), P = 0.002; and after intervention: OR, 1.04 (0.97–1.11), P = 0.23. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.90 (0.86–0.94), P < 0.001; intervention effect: OR, 0.51 (0.30–0.86), P = 0.01; and after intervention: OR, 1.15 (1.07–1.14), P < 0.001. High bleeding risk: before intervention: OR, 1.08 (1.03–1.15), P = 0.005; intervention effect: OR, 0.46 (0.23–0.95), P = 0.04; and after intervention: OR, 0.89 (0.81–0.98), P = 0.02.
Time-series plots. Erythrocyte transfusion rates up to postoperative day 7. The algorithm was instituted at the beginning of the 13th interval. All odds ratio (OR), 95% CIs, and P values calculated using segmented regression modeling adjusted as described in the Statistical Analysis section. ORs before and after intervention refer to respective changes in outcome per 4-week period. OR for the intervention refers to change in outcome associated with the institution of the algorithm. (A) Entire sample: before intervention: OR, 0.96 (0.92–1.01), P = 0.08; intervention effect: OR, 0.50 (0.32–0.77), P = 0.002; and after intervention: OR, 1.04 (0.97–1.11), P = 0.23. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.90 (0.86–0.94), P < 0.001; intervention effect: OR, 0.51 (0.30–0.86), P = 0.01; and after intervention: OR, 1.15 (1.07–1.14), P < 0.001. High bleeding risk: before intervention: OR, 1.08 (1.03–1.15), P = 0.005; intervention effect: OR, 0.46 (0.23–0.95), P = 0.04; and after intervention: OR, 0.89 (0.81–0.98), P = 0.02.
Fig. 2.
Time-series plots. Erythrocyte transfusion rates up to postoperative day 7. The algorithm was instituted at the beginning of the 13th interval. All odds ratio (OR), 95% CIs, and P values calculated using segmented regression modeling adjusted as described in the Statistical Analysis section. ORs before and after intervention refer to respective changes in outcome per 4-week period. OR for the intervention refers to change in outcome associated with the institution of the algorithm. (A) Entire sample: before intervention: OR, 0.96 (0.92–1.01), P = 0.08; intervention effect: OR, 0.50 (0.32–0.77), P = 0.002; and after intervention: OR, 1.04 (0.97–1.11), P = 0.23. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.90 (0.86–0.94), P < 0.001; intervention effect: OR, 0.51 (0.30–0.86), P = 0.01; and after intervention: OR, 1.15 (1.07–1.14), P < 0.001. High bleeding risk: before intervention: OR, 1.08 (1.03–1.15), P = 0.005; intervention effect: OR, 0.46 (0.23–0.95), P = 0.04; and after intervention: OR, 0.89 (0.81–0.98), P = 0.02.
×
Fig. 3.
Time-series plots. Platelet transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.95 (0.91–0.99), P = 0.02; intervention effect: OR, 0.22 (0.13–0.37), P < 0.001; and after intervention: OR, 1.16 (1.08–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.88 (0.84–0.92), P < 0.001; intervention effect: OR, 0.30 (0.15–0.60), P = 0.01; and after intervention: OR, 1.27 (1.16–1.40), P < 0.001. High bleeding risk: before intervention: OR, 1.04 (0.99–1.09), P = 0.16; intervention effect: OR, 0.17 (0.08–0.34), P < 0.001; and after intervention: OR, 1.06 (0.96–1.16), P = 0.27.
Time-series plots. Platelet transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.95 (0.91–0.99), P = 0.02; intervention effect: OR, 0.22 (0.13–0.37), P < 0.001; and after intervention: OR, 1.16 (1.08–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.88 (0.84–0.92), P < 0.001; intervention effect: OR, 0.30 (0.15–0.60), P = 0.01; and after intervention: OR, 1.27 (1.16–1.40), P < 0.001. High bleeding risk: before intervention: OR, 1.04 (0.99–1.09), P = 0.16; intervention effect: OR, 0.17 (0.08–0.34), P < 0.001; and after intervention: OR, 1.06 (0.96–1.16), P = 0.27.
Fig. 3.
Time-series plots. Platelet transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.95 (0.91–0.99), P = 0.02; intervention effect: OR, 0.22 (0.13–0.37), P < 0.001; and after intervention: OR, 1.16 (1.08–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.88 (0.84–0.92), P < 0.001; intervention effect: OR, 0.30 (0.15–0.60), P = 0.01; and after intervention: OR, 1.27 (1.16–1.40), P < 0.001. High bleeding risk: before intervention: OR, 1.04 (0.99–1.09), P = 0.16; intervention effect: OR, 0.17 (0.08–0.34), P < 0.001; and after intervention: OR, 1.06 (0.96–1.16), P = 0.27.
×
Fig. 4.
Time-series plots. Plasma transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.91 (0.88–0.94), P < 0.001; intervention effect: OR, 0.20 (0.12–0.34), P < 0.001; and after intervention: OR, 1.15 (1.07–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.85 (0.81–0.89), P < 0.001; intervention effect: OR, 0.34 (0.18–0.78), P = 0.01; and after intervention: OR, 1.22 (1.08–1.37), P = 0.001. High bleeding risk: before intervention: OR, 0.99 (0.94–1.03), P = 0.49; intervention effect: OR, 0.14 (0.08–0.27), P < 0.001; and after intervention: OR, 1.09 (1.00–1.19), P = 0.05.
Time-series plots. Plasma transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.91 (0.88–0.94), P < 0.001; intervention effect: OR, 0.20 (0.12–0.34), P < 0.001; and after intervention: OR, 1.15 (1.07–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.85 (0.81–0.89), P < 0.001; intervention effect: OR, 0.34 (0.18–0.78), P = 0.01; and after intervention: OR, 1.22 (1.08–1.37), P = 0.001. High bleeding risk: before intervention: OR, 0.99 (0.94–1.03), P = 0.49; intervention effect: OR, 0.14 (0.08–0.27), P < 0.001; and after intervention: OR, 1.09 (1.00–1.19), P = 0.05.
Fig. 4.
Time-series plots. Plasma transfusion rate up to postoperative day 7. (A) Entire sample: before intervention: odds ratio (OR), 0.91 (0.88–0.94), P < 0.001; intervention effect: OR, 0.20 (0.12–0.34), P < 0.001; and after intervention: OR, 1.15 (1.07–1.25), P < 0.001. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.85 (0.81–0.89), P < 0.001; intervention effect: OR, 0.34 (0.18–0.78), P = 0.01; and after intervention: OR, 1.22 (1.08–1.37), P = 0.001. High bleeding risk: before intervention: OR, 0.99 (0.94–1.03), P = 0.49; intervention effect: OR, 0.14 (0.08–0.27), P < 0.001; and after intervention: OR, 1.09 (1.00–1.19), P = 0.05.
×
Fig. 5.
Time-series plots, large-volume (≥4 units) erythrocyte transfusion rate on the day of surgery. (A) Entire sample: before intervention: odds ratio (OR), 0.94 (0.89–0.99), P = 0.03; intervention effect: OR, 0.23 (0.11–0.48), P < 0.001; and after intervention: OR, 1.14 (1.01–1.28), P = 0.03. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.77 (0.70–0.84), P < 0.001; intervention effect: OR, 0.25 (0.02–3.04), P = 0.27; and after intervention: OR, 1.62 (1.20–2.20), P = 0.002. High bleeding risk: before intervention: OR, 0.99 (0.94–1.05), P = 0.84; intervention effect: OR, 0.25 (0.12–0.50), P < 0.001; and after intervention: OR, 1.07 (0.96–1.19), P = 0.23.
Time-series plots, large-volume (≥4 units) erythrocyte transfusion rate on the day of surgery. (A) Entire sample: before intervention: odds ratio (OR), 0.94 (0.89–0.99), P = 0.03; intervention effect: OR, 0.23 (0.11–0.48), P < 0.001; and after intervention: OR, 1.14 (1.01–1.28), P = 0.03. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.77 (0.70–0.84), P < 0.001; intervention effect: OR, 0.25 (0.02–3.04), P = 0.27; and after intervention: OR, 1.62 (1.20–2.20), P = 0.002. High bleeding risk: before intervention: OR, 0.99 (0.94–1.05), P = 0.84; intervention effect: OR, 0.25 (0.12–0.50), P < 0.001; and after intervention: OR, 1.07 (0.96–1.19), P = 0.23.
Fig. 5.
Time-series plots, large-volume (≥4 units) erythrocyte transfusion rate on the day of surgery. (A) Entire sample: before intervention: odds ratio (OR), 0.94 (0.89–0.99), P = 0.03; intervention effect: OR, 0.23 (0.11–0.48), P < 0.001; and after intervention: OR, 1.14 (1.01–1.28), P = 0.03. (B) Stratified according to bleeding risk. Low bleeding risk: before intervention: OR, 0.77 (0.70–0.84), P < 0.001; intervention effect: OR, 0.25 (0.02–3.04), P = 0.27; and after intervention: OR, 1.62 (1.20–2.20), P = 0.002. High bleeding risk: before intervention: OR, 0.99 (0.94–1.05), P = 0.84; intervention effect: OR, 0.25 (0.12–0.50), P < 0.001; and after intervention: OR, 1.07 (0.96–1.19), P = 0.23.
×
Table 1.
Patient Characteristics and Clinical Status
Patient Characteristics and Clinical Status×
Patient Characteristics and Clinical Status
Table 1.
Patient Characteristics and Clinical Status
Patient Characteristics and Clinical Status×
×
Table 2.
Surgical Parameters
Surgical Parameters×
Surgical Parameters
Table 2.
Surgical Parameters
Surgical Parameters×
×
Table 3.
Outcomes for the Entire Pre- and Postalgorithm Groups
Outcomes for the Entire Pre- and Postalgorithm Groups×
Outcomes for the Entire Pre- and Postalgorithm Groups
Table 3.
Outcomes for the Entire Pre- and Postalgorithm Groups
Outcomes for the Entire Pre- and Postalgorithm Groups×
×
Table 4.
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm×
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm
Table 4.
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm
Subgroup Analyses for Selected Outcomes; Risk-adjusted OR for Outcomes Associated with the Beginning of the Algorithm×
×