Cardiovascular Surgery

Transcription

1 Cardiovascular Surgery Meta-Analysis Comparing the Effectiveness and Adverse Outcomes of Antifibrinolytic Agents in Cardiac Surgery Jeremiah R. Brown, PhD; Nancy J.O. Birkmeyer, PhD; Gerald T. O Connor, PhD, ScD Background Since the 1980s, antifibrinolytic therapies have assisted surgical teams in reducing the amount of blood loss. To date, however, serious questions remain regarding the safety and effectiveness of these agents. Methods and Results We conducted a meta-analysis to compare aprotinin, -aminocaproic acid, and tranexamic acid with placebo and head to head on 8 clinical outcomes from 138 trials. Published randomized controlled trial data were collected from OVID/PubMed. Outcomes included total blood loss, transfusion of packed red blood cells, reexploration, mortality, stroke, myocardial infarction, dialysis-dependent renal failure, and renal dysfunction (0.5-mg/dL increase in creatinine from baseline). All agents were effective in significantly reducing blood loss by 226 to 348 ml and the proportion of patients transfused with packed red blood cells over placebo. Only high-dose aprotinin reduced the rate of reexploration (relative risk, 0.49; 95% CI, 0.33 to 0.73). There were no significant risks or benefits for any agent for mortality, stroke, myocardial infarction, or renal failure. However, high-dose aprotinin significantly increased the risk of renal dysfunction (relative risk, 1.47; 95% CI, 1.12 to 1.94), 12.9% versus 8.4%. Compared head to head, high-dose aprotinin demonstrated significant reduction in total blood loss over -aminocaproic acid ( 184 ml; 95% CI, 256 to 112) and tranexamic acid ( 195 ml; 95% CI, 286 to 105). There were no significant differences among any agent when compared head to head on other outcomes. Conclusions All antifibrinolytic agents were effective in reducing blood loss and transfusion. There were no significant risks or benefits for mortality, stroke, myocardial infarction, or renal failure. However, high-dose aprotinin was associated with a statistically significant increased risk of renal dysfunction. (Circulation. 2007;115: ) Key Words: aminocaproic acids aprotinin meta-analysis surgery tranexamic acid Intraoperative bleeding and postoperative bleeding are frequent and often serious complications of cardiac surgery. Two thirds of patients having coronary artery bypass surgery are transfused with packed red blood cells (prbcs). Cardiac surgery patients receive 20% of all prbcs used in the United States. 1 Blood transfusion benefits patients, but it is not without risk. Chelemer et al 2 showed that each unit of prbcs was associated with an increase in the incidence of bacterial infections. Speiss 3 described the typical unit of prbcs: the low ph, high potassium level, and limited ability to deliver oxygen to tissue. Others suggested that less transfusion may be equally effective 4 and that there may be long-term adverse effects of blood transfusion. 5 Editorial p 2790 Clinical Perspective p 2813 Three drugs used to decrease blood loss and the need for transfusion are in clinical use. -Aminocaproic acid (Amicar, Xanodyne Pharmaceuticals, Inc, Newport, Ky) inhibits fibrinolysis by the inhibition of plasminogen inhibitors and, to a lesser degree, through antiplasmin activity. Tranexamic acid (Cyklokapron, Pharmacia & Upjohn Co, Piscataway, NJ) is similar in action to -aminocaproic acid but is 10 times more potent. Aprotinin is a natural serine protease inhibitor (Tasylol, Bayer Pharmaceuticals Corp, West Haven, Conn) that affects multiple mediators, resulting in the attenuation of inflammatory responses, fibrinolysis, and thrombin generation. These antifibrinolytic agents have been studied in hundreds of randomized trials, yet questions remain. An observational study by Mangano et al 6,7 found that aprotinin was associated with increased risks of adverse cardiovascular, cerebrovascular, and renal outcomes. In a recent letter to the editor of the New England Journal of Medicine, we reported the results of a meta-analysis of the renal outcomes of aprotinin versus placebo. 8 Subsequently, we extended this meta-analysis to include -aminocaproic acid and tranexamic acid. We summarized the effects of these agents compared with placebo Received October 19, 2006; accepted March 26, From the Center for the Evaluative Clinical Sciences (J.R.B., G.T.O.) and Departments of Medicine and of Community and Family Medicine (G.T.O.), Dartmouth Medical School, Hanover, NH, and Michigan Surgical Collaboration for Outcomes Research and Evaluation, University of Michigan, Ann Arbor (N.J.O.B.). The online-only Data Supplement, consisting of a table and references, is available with this article at full/circulationaha /dc1. Correspondence to Jeremiah R. Brown, PhD, Clinical Research Section, Rubin 505, Dartmouth-Hitchcock Medical Center, 1 Medical Center Dr, Lebanon, NH Jeremiah.Brown@Dartmouth.edu 2007 American Heart Association, Inc. Circulation is available at DOI: /CIRCULATIONAHA

2 2802 Circulation June 5, 2007 Figure 1. RCT selection. RCT indicates randomized controlled trial. and in head-to-head comparisons. The metrics included bleeding volume, transfusion rate, and adverse events, including mortality, stroke, myocardial infarction, and renal dysfunction (insufficiency and dialysis-dependent failure). Methods Study Selection We conducted a meta-analysis of randomized controlled trials including adult coronary artery bypass (CABG), isolated valve, or combined CABG/valve surgery. Both primary operations and reoperations were included in this analysis. OVID/Medline was used to identify published randomized controlled trials from 1985 through July Key words used to search included trasylol or aprotinin, aminocaproic acid or amicar, tranexamic acid, cardiac surgery or thoracic surgery, coronary artery bypass, or valve. The search yielded 242 published human randomized controlled trials. The search was further limited to the English language (226) and adults (185). Among the 185 studies remaining, 38 did not report outcomes of interest, and 9 were excluded for pump-only fibrinolytic therapy, leaving 138 trials with outcomes of interest to this meta-analysis (Figure 1) We abstracted data from the trials on 8 outcomes, including bleeding (intraoperative and postoperative blood loss was recorded in milliliters); the proportion of patients transfused with prbcs; the proportion of patients returning to the operating room for reexploration; adverse events, including mortality, stroke, and myocardial infarction; and renal complications stratified by dialysis-dependent renal failure and renal dysfunction (0.5-mg/dL increase in creatinine). We defined renal failure as new onset of dialysis. There was 1 exception to this rule: Dialysis-dependent renal failure was not reported, but a 2.0-mg/dL increase in creatinine was reported, which we defined as a renal failure and occurred in 1 study. Renal dysfunction was defined as a 0.5-mg/dL increase in creatinine. We followed the appropriate methods for conducting a meta-analysis as stipulated in the Consolidated Standards of Reporting Trials (CONSORT) statement. 147 Two independent reviewers selected trials for information outcomes and recorded data on spreadsheets. Jadad criteria were assessed and tabulated with study characteristics (see the Table in the online Data Supplement). 148 Outcomes were stratified by Jadad score and were determined not to influence the results. Funnel plots were generated to evaluate possible publication bias. Patterns of bias were observed for high- and low-dose aprotinin with regard to total blood loss and the proportion of patients transfused with prbcs. Patterns of bias also were observed for aminocaproic acid and tranexamic acid for the proportion of patients transfused with prbcs. None of the other funnel plots by agent and outcome indicated evidence of publication bias. To control for the dosing effect of aprotinin, we stratified our analysis by high- and low-dose aprotinin (as suggested by the manufacturer). High-dose (also referred to as full-dose) aprotinin consisted of a 2 million kallikrein-inhibiting units (KIU) IV loading dose, 2 million KIU pump-priming dose, and 0.5 million KIU IV/h maintenance dose. Low-dose (also referred to as half-dose) aprotinin consisted of a 1 million KIU IV loading dose, 1 million KIU pump-priming dose, and 0.25 million KIU IV/h maintenance dose. Cardiopulmonary bypass pump-only aprotinin was not included in our analysis because there were insufficient data (8 studies) to report outcomes of interest. -Aminocaproic acid dosing varied from trial to trial but was consistent within each trial. We stratified our analysis by dividing the doses into high and low strata. We discovered that there was no significantly different effect between the high- and low-dosing categories for -aminocaproic acid. Therefore, we report the summary effect of all dosing of -aminocaproic acid. Tranexamic acid dosing varied from trial to trial. After stratification analysis into 2 groups (high or low dose), we discovered that there was no significant difference in the effect of high-dose versus low-dose tranexamic acid. We report the summary effects for tranexamic acid for all dosing together. There were 22 published trials reporting on head-to-head comparisons, ie, each agent compared with a competing agent. We compared aprotinin with tranexamic acid (15 studies), aprotinin with -aminocaproic acid (9 studies), and -aminocaproic acid with tranexamic acid (5 studies). All head-to-head comparisons involving aprotinin used high-dose aprotinin. Statistical Analysis All outcome comparisons and treatment effects were calculated with the Cochrane Collaborative software, RevMan Effect modification was assessed by the Mantel-Haenszel test of homogeneity across subcategories of procedure type (CABG, CABG/valve, and valve) and for high and low dosing for -aminocaproic acid and tranexamic acid. In addition, we calculated the I 2 to evaluate the percentage of heterogeneity among all the trials incorporated into the summary estimate. 149 To control for heterogeneity, we used randomeffects modeling. The weighted mean difference (WMD) and 95% CIs were calculated for continuous data (total blood loss in milliliters). For all dichotomous data (proportion of patients transfused with prbcs, return to operating room/reexploration, stroke, mortality, myocardial infarction, renal failure, and renal dysfunction), relative risk (RR) and 95% CIs were calculated for each independent study and for the summary statistic. If any single cell (either in treatment or control) had no events, the software automatically added 0.5 to maintain its contribution in the analysis. Any study reporting no events in both the treatment and control groups was not included in the summary statistic calculation. The analysis was divided into 2 parts. First, each agent was compared with placebo. Second, a head-to-head analysis was done to compare each agent with a competing agent. Methods for the calculation of the above statistics have been reported previously. 150,151 The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written. Results Study Selection There were 138 randomized controlled trials that reported outcomes of interest to this meta-analysis: total blood loss,

3 Brown et al Antifibrinolytic Meta-Analysis 2803 proportion of patients transfused with prbcs, return to operating room/reexploration, stroke, mortality, myocardial infarction, and renal failure or dysfunction complications. For placebo-controlled studies, there were 78 high-dose and 32 low-dose aprotinin studies, 18 -aminocaproic acid studies, and 31 tranexamic acid studies. For head-to-head comparative studies, there were 15 aprotinin versus tranexamic acid, 9 aprotinin versus -aminocaproic acid, and 5 -aminocaproic acid versus tranexamic acid studies. We found no evidence for effect modification by procedure type for all agents or for dosing effects for -aminocaproic acid or tranexamic acid. Placebo Comparisons Bleeding Comparisons of antifibrinolytic agents with placebo by outcomes and effect estimates are reported in Table 1. Patients receiving either high- or low-dose aprotinin, -aminocaproic acid, or tranexamic acid had considerably less total blood loss compared with placebo-controlled patients (Figure 2). Patients receiving high-dose aprotinin on average lost 348 ml (95% CI, 416 to 281; P 0.001) less blood than did their counterparts receiving placebo. Low-dose aprotinin (6 studies, 515 patients) reduced total blood loss by 226 ml on average (95% CI, 277 to 175 ml; P 0.001). -Aminocaproic acid resulted in 240 ml less blood loss (95% CI, 341 to 140 ml; P 0.001), and tranexamic acid resulted in 285 ml less blood loss (95% CI, 394 to 175 ml; P 0.001). All antifibrinolytic agents were significantly effective at reducing the proportion of patients requiring transfusion of prbcs compared with placebo-controlled patients (Figure 3). Patients treated with high-dose aprotinin had a 40% reduced rate of transfusion compared with placebo-controlled patients (RR, 0.60; 95% CI, 0.53 to 0.67; P 0.001). A 24% reduced rate of transfusion (RR, 0.76; 95% CI, 0.66 to 0.86; P 0.001) was observed for patients treated with low-dose aprotinin. -Aminocaproic acid reduced the rate of transfusion by 37% (RR, 0.63; 95% CI, 0.44 to 0.90; P 0.010). Tranexamic acid reduced the rate of transfusion by 25% (RR, 0.75; 95% CI, 0.60 to 0.92; P 0.007). High-dose aprotinin significantly reduced the rate of reexploration by 51% (RR, 0.49; 95% CI, 0.33 to 0.73; P 0.001) compared with placebo. Low-dose aprotinin did not significantly reduce the rate for reexploration compared with placebo-treated patients (RR, 0.69; 95% CI, 0.41 to 1.18; P 0.18). Neither -aminocaproic acid (RR, 0.51; 95% CI, 0.15 to 1.82; P 0.30) nor tranexamic acid (RR, 0.70; 95% CI, 0.44 to 1.11; P 0.12) significantly reduced the rate of reexploration compared with placebo-controlled patients. However, all agents demonstrated a trend toward reduced rates of reexploration with similar summary effect measures. Adverse Outcomes Adverse outcomes, including mortality, stroke, and myocardial infarction, are reported in Figure 4 and Table 1. Fortythree studies reported adverse outcomes for high-dose aprotinin; 16 studies reported adverse outcomes for low-dose aprotinin. There were 8 studies for -aminocaproic acid and 18 studies for tranexamic acid. There were no statistically significant risks or benefits for any of these antifibrinolytic agents with regard to mortality, stroke, or myocardial infarction (Table 1 and Figure 4). High-dose aprotinin did not significantly protect patients from mortality or place patients at risk of death (RR, 0.89; 95% CI, 0.65 to 1.21; P 0.46). There was no mortality risk or benefit for low-dose aprotinin (RR, 1.37; 95% CI, 0.72 to 2.59; P 0.34), -aminocaproic acid (RR, 1.82; 95% CI, 0.55 to 5.98; P ), or tranexamic acid (RR, 0.67; 95% CI, 0.33 to 1.37; P 0.28). There was no significant risk or benefit for stoke as observed for high-dose aprotinin (RR, 0.67; 95% CI, 0.30 to 1.47; P 0.32), low-dose aprotinin (RR, 0.47; 95% CI, 0.09 to 2.36; P 0.36), -aminocaproic acid (RR, 0.60; 95% CI, 0.13 to 2.81; P 0.52), or tranexamic acid (RR, 1.31; 95% CI, 0.59 to 2.93; P 0.51). There was no significant risk or protection against myocardial infarction from any of the antifibrinolytic agents: high-dose aprotinin (RR, 1.10; 95% CI, 0.83 to 1.45; P 0.52), low-dose aprotinin (RR, 0.94; 95% CI, 0.58 to 1.54; P 0.82), -aminocaproic acid (RR, 1.14; 95% CI, 0.50 to 2.60; P 0.76), or tranexamic acid (RR, 0.94; 95% CI, 0.51 to 1.74; P 0.85). Antifibrinolytic therapies do not place patients at increased risk of mortality, stroke, or myocardial infarction, nor do they demonstrate a significant protective effect. Renal Complications No agent demonstrated a significant increased risk of dialysis-dependent renal failure (Figure 5): high-dose aprotinin (RR, 1.09; 95% CI, 0.68 to 1.77; P 0.71), low-dose aprotinin (RR, 1.86; 95% CI, 0.07 to 49.26; P 0.71), and tranexamic acid (RR, 1.43; 95% CI, 0.30 to 6.85; P 0.66). There were no renal outcomes reported for -aminocaproic acid. However, as previously shown, 8 high-dose aprotinin resulted in a 47% increased risk of renal dysfunction (RR, 1.47; 95% CI, 1.12, 1.94; P 0.006), defined as a 0.5- mg/dl increase in creatinine from baseline. No increase in the risk of renal dysfunction was observed for low-dose aprotinin (RR, 1.01; 95% CI, 0.69 to 1.49; P 0.96). There were no trials reporting renal dysfunction with -aminocaproic acid to include in this analysis. There was no statistically significant risk of renal dysfunction observed with tranexamic acid (RR, 2.02; 95% CI, 0.73 to 5.60; P 0.18), although a nonsignificant trend toward renal dysfunction was observed. Stratified Analysis We investigated the possibility that results stratified by the type of procedure (eg, CABG, CABG/valve, valve only) could affect the results. We compared each agent with placebo for the proportion of patients transfused with prbcs. High-dose aprotinin stratified by procedures resulted in similar summary estimates: RR, 0.60, 95% CI, 0.53 to 0.67 (all procedures); RR, 0.62, 95% CI, 0.55 to 0.70 (CABG); RR, 0.44, 95% CI, 0.34 to 0.59 (CABG/ valve); and RR, 0.41, 95% CI, 0.21 to 0.80 (valve). Low-dose aprotinin estimates were similar across procedure type: RR, 0.76, 95% CI, 0.66 to 0.86 (all); RR, 0.75, 95% CI, 0.64 to 0.88 (CABG); RR, 0.87, 95% CI, 0.78 to 0.97 (CABG/valve); and RR, 0.89, 95% CI, 0.58 to 1.37 (valve). -Aminocaproic acid resulted in similar estimates

4 2804 Circulation June 5, 2007 TABLE 1. Comparison of Antifibrinolytic Agents With Placebo Agent Versus Placebo Studies Subjects Mean Difference (Random Effects) RR (Random Effects) 95% CI P Bleeding Total blood loss by agent, ml Aprotinin high dose to Aprotinin low dose to Aminocaproic acid to Tranexamic acid to Patients transfused with prbc Aprotinin high dose to Aprotinin low dose to Aminocaproic acid to Tranexamic acid to Return to operating room Aprotinin high dose to Aprotinin low dose to Aminocaproic acid to Tranexamic acid to Adverse Outcomes Mortality Aprotinin high dose to Aprotinin low dose to Aminocaproic acid to Tranexamic acid to Stroke Aprotinin high dose to Aprotinin low dose to Aminocaproic acid to Tranexamic acid to Myocardial infarction Aprotinin high dose to Aprotinin low dose to Aminocaproic acid to Tranexamic acid to Renal failure Aprotinin high dose to Aprotinin low dose to Aminocaproic acid 0 Tranexamic acid to Renal dysfunction Aprotinin high dose to Aprotinin low dose to Aminocaproic acid 0 Tranexamic acid to across procedures: RR, 0.63, 95% CI, 0.44 to 0.90 (all); RR, 0.60, 95% CI, 0.39 to 0.91 (CABG); and RR, 0.77, 95% CI, 0.35 to 1.70 (CABG/valve), with no estimates for valve surgery. Tranexamic acid stratified by procedure type showed similar results with the exception of valve surgery: RR, 0.75, 95% CI, 0.60 to 0.92 (all); RR, 0.75, 95% CI, 0.57 to 0.98 (CABG); RR, 0.69, 95% CI, 0.48 to 0.98 (CABG/valve); and RR, 1.40, 95% CI, 1.17 to 1.69 (valve). When stratified by the type of procedure, the estimates were similar. Tranexamic acid had a significantly increased risk for the proportion of patients transfused with prbcs among valve-only procedures; however, this was not observed for total blood loss by procedure type.

5 Brown et al Antifibrinolytic Meta-Analysis 2805 Total Blood Loss (ml) by Agent Epsilon-Aminocaproic Acid Weighted Mean Difference (Random Effects) Figure 2. Total blood loss by antifibrinolytic agent compared with placebo. The WMD of total blood loss (ml) by antifibrinolytic agent, each compared with placebo. WMD (diamond) and 95% CIs (horizontal bars) summarize the effect using a random-effects model. All antifibrinolytic effects are to the left of the 0, representing a significant reduction in total blood loss over placebo. Head-to-Head Comparisons Bleeding Head-to-head comparisons of antifibrinolytic agents are reported in Table 2. High-dose aprotinin resulted in 195 ml less total blood loss compared with tranexamic acid (95% CI, 286 to 105 ml; P 0.001). High-dose aprotinin resulted in 184 ml less total blood loss compared with -aminocaproic acid (95% CI, 256 to 112 ml; P 0.001; Figure 6). -Aminocaproic acid and tranexamic acid did not differ significantly in total blood loss (WMD, 64 ml; 95% CI, 214 to 85 ml; P 0.40). High-dose aprotinin did not significantly reduce the rate of transfusion compared with tranexamic acid (RR, 0.81; 95% CI, 0.60 to 1.09; P 0.17) or -aminocaproic acid (RR, 0.93; 95% CI, 0.71 to 1.22; P 0.60). There was no significant difference in the rate of transfusion when -aminocaproic acid was compared with tranexamic acid (RR, 0.81; 95% CI, 0.49 to 1.34; P 0.41). High-dose aprotinin did not significantly reduce the rate of reexploration compared with tranexamic acid (RR, 0.90; 95% CI, 0.50 to 1.63; P 0.73) or -aminocaproic acid (RR, 1.04; 95% CI, 0.40 to 2.72; P 0.94). There was no statistically significant difference in the rate of reexploration when -aminocaproic acid was compared with tranexamic acid (RR, 0.72; 95% CI, 0.20 to 2.56; P 0.62). Adverse Outcomes There were no statistically significant differences for highdose aprotinin over tranexamic acid or -aminocaproic acid in reducing mortality, stroke, or myocardial infarction (Table 2 and Figure 7) when antifibrinolytic agents were compared head to head. High-dose aprotinin compared with tranexamic Patients Transfused with prbc Epsilon-Aminocaproic Acid Return to Operating Room Epsilon-Aminocaproic Acid Relative Risk (Random Effects) acid was not significantly different with regard to mortality (RR, 1.29; 95% CI, 0.63 to 2.65; P 0.49). Likewise, mortality, when high-dose aprotinin is compared with -aminocaproic acid, was not significantly different (RR, 0.97; 95% CI, 0.06 to 15.22; P 0.98). -Aminocaproic acid did not significantly differ from tranexamic acid in mortality (RR, 3.85; 95% CI, 0.43 to 34.28; P 0.23). Compared with tranexamic acid, high-dose aprotinin was not significantly different with regard to stroke (RR, 1.52; 95% CI, 0.58 to 4.03; P 0.39) or myocardial infarction (RR, 0.89; 95% CI, 0.52 to 1.50; P 0.65). Compared with -aminocaproic acid, high-dose aprotinin did not significantly differ in regard to stroke (RR, 3.96; 95% CI, 0.44 to 35.27; P 0.22) or myocardial infarction (RR, 1.20; 95% CI, 0.30 to 4.89; P 0.79). -Aminocaproic acid did not significantly differ from tranexamic acid in terms of stroke (RR, 0.33; 95% CI, 0.03 to 3.10; P 0.33) or myocardial infarction (RR, 1.00; 95% CI, 0.29 to 3.46; P 1.00). Overall, there was no significant difference among antifibrinolytic agents with regard to mortality, stroke, or myocardial infarction. There is no demonstrated risk or benefit of 1 agent over another. Renal Complications Published studies reporting on head-to-head comparisons for renal failure and renal dysfunction outcomes were rare. One study that compared high-dose aprotinin with tranexamic acid reported renal failure outcomes. Two studies of the same comparison reported renal dysfunction outcomes (Table 2). There was no significant difference in renal failure between high-dose aprotinin and tranexamic acid (RR, 1.01; 95% CI, 0.36 to 2.85; P 0.99). 34 Renal dysfunction was not signifi- Figure 3. Bleeding outcomes by antifibrinolytic agent compared with placebo. The RRs of bleeding outcomes (number of patients transfused with prbcs and any return to the operating room or reexploration) by antifibrinolytic agent compared with placebo are plotted. The RR (diamond) and 95% CIs (horizontal bars) summarize the effect using a random-effects model. All antifibrinolytic effects are to the left of the 1.0, representing a reduction in the rate of patients transfused with prbcs and reexploration over placebo. When the horizontal bars cross 1.0, the effect is not significantly different from the placebo group

6 2806 Circulation June 5, 2007 Mortality Relative Risk (Random Effects) Epsilon-Aminocaproic Acid Stroke Epsilon-Aminocaproic Acid Myocardial Infarction Epsilon-Aminocaproic Acid Figure 4. Adverse outcomes by antifibrinolytic agent compared with placebo. The RRs of adverse outcomes (mortality, stroke, and myocardial infarction) by antifibrinolytic agent vs placebo are plotted. The RR (diamond) and 95% CIs (horizontal bars) summarize the effect using a random-effects model. Effects left of 1.0 favor the antifibrinolytic agent over placebo; effects to the right favor placebo over antifibrinolytic agent. When the horizontal bars cross 1.0, the effect is not significantly different from the comparison group; this is the case for all agents for all adverse events (mortality, stroke, myocardial infarction) plotted here cantly different between high-dose aprotinin and tranexamic acid (RR, 1.07; 95% CI, 0.55 to 2.10; P 0.84). 34,152 One head-to-head study compared low-dose aprotinin with -aminocaproic acid in 100 patients and reported no renal dysfunction events. 153 No studies that reported renal outcomes compared -aminocaproic acid with tranexamic acid. Discussion In this meta-analysis of published randomized clinical trials of antifibrinolytic agents, we found that aprotinin, -aminocaproic acid, and tranexamic acid all reduce total blood loss and rates of transfusion in cardiac surgery compared with placebo. Although only high-dose aprotinin significantly reduced reexploration rates, there was a similarly protective effect observed for all of the agents. There were no significant effects of the drugs on mortality, stroke, or myocardial infarction. High-dose aprotinin, however, was associated with a significant increase in rates of renal dysfunction (12.9% versus 8.4%). 8 The present study adds 5 years and nearly doubles the number of studies included in previously reported meta-analyses on this topic. Our analysis included all cardiac surgery procedures, in contrast to the Sedrakyan et al 154 study, and Renal Failure Renal Dysfunction Relative Risk (Random Effects) analyzed the aprotinin trials stratified by dose, in contrast to the Cochrane Collaborative study. 150 Our results do not confirm the findings of Sedrakyan et al of a significant reduction in the risk of stroke among coronary bypass patients receiving aprotinin in either the combined results or those stratified according to procedure. In contrast to the recently published observational study by Mangano et al, 6,7 aprotinin did not significantly increase risks of mortality, stroke, myocardial infarction, or renal failure requiring dialysis in our meta-analysis of published trials. However, as we previously reported, 8 our results confirm the findings of Mangano et al 6 and Karkouti et al 155 of an increased risk of renal dysfunction among patients receiving aprotinin. We believe that these discrepancies between our findings and those of Mangano et al lie in the difference between observational and experimental study designs. The major limitation of observational cohort studies for evaluating treatment efficacy is confounding by indication, also known as treatment selection bias. Although standard statistical methods can be used to control for measured differences in prognosis between the treatment groups, confounding by unmeasured factors cannot be removed. Randomization deals Figure 5. Renal failure and dysfunction by antifibrinolytic agent compared with placebo. The RRs of renal complications (dialysis-dependent renal failure and renal dysfunction defined as 0.5- mg/dl increase in creatinine) by antifibrinolytic agent vs placebo are plotted. The RR (diamond) and 95% CIs (horizontal bars) summarize the effect using a random-effects model. The only statistically significant effect shown here is the increased risk of renal dysfunction with the use of high-dose aprotinin vs placebo

7 Brown et al Antifibrinolytic Meta-Analysis 2807 TABLE 2. Head-to-Head Comparisons of Antifibrinolytics Head-to-Head Comparison Studies Subjects Mean Difference (Random Effects) RR (Random Effects) 95% CI P Bleeding Total blood loss by agent, ml Aprotinin vs tranexamic acid to Aprotinin vs -aminocaproic acid to Aminocaproic acid vs tranexamic acid to Patients transfused with prbc Aprotinin vs tranexamic acid to Aprotinin vs -aminocaproic acid to Aminocaproic acid vs tranexamic acid to Return to operating room Aprotinin vs tranexamic acid to Aprotinin vs -aminocaproic acid to Aminocaproic acid vs tranexamic acid to Adverse outcomes Mortality Aprotinin vs tranexamic acid to Aprotinin vs -aminocaproic acid to Aminocaproic acid vs tranexamic acid to Stroke Aprotinin vs tranexamic acid to Aprotinin vs -aminocaproic acid to Aminocaproic acid vs tranexamic acid to Myocardial Infarction Aprotinin vs tranexamic acid to Aprotinin vs -aminocaproic acid to Aminocaproic acid vs tranexamic acid to Renal failure Aprotinin vs tranexamic acid to Aprotinin vs -aminocaproic acid 0 -Aminocaproic acid vs tranexamic acid 0 Renal dysfunction Aprotinin vs tranexamic acid to Aprotinin vs -aminocaproic acid Aminocaproic acid vs tranexamic acid 0 effectively with both measured and unmeasured confounding; therefore, randomized trials are regarded as the most rigorous design for studying the effectiveness of alternative treatments. The downside of randomized trials is that they are conducted under such highly controlled, ideal circumstances that their results may not be applicable to usual clinical conditions. The Mangano et al study included samples drawn from consecutive cardiac surgery patients and found that those treated with antifibrinolytics were at higher risk for adverse outcomes than the untreated control patients. Patients included in the randomized trials summarized in this meta-analysis consisted of low-risk, elective patients whose antifibrinolytic treatments were randomly allocated. The definition of renal dysfunction as a 0.5-mg/dL elevation in postoperative serum creatinine is widely accepted. It is the definition consistently used and reported in the trials and other cardiac surgery studies. Renal dysfunction defined in this way has been shown to increase the risk of 30-day mortality with cardiac surgery by 18-fold, demonstrating its usefulness as a metric in clinical trials. 156 With recent studies showing an increase in renal dysfunction in patients treated with aprotinin, some have raised the possibility that elevations in serum creatinine among aprotinin-treated patients may be artificial because of the accumulation of the drug in the proximal renal tubules, which does not occur among the lysine analogues. The Food and Drug Administration advisory panel reviewing the new evidence was not swayed by this argument, however, and has recommended labeling changes to alert physicians to the risks of kidney damage with the use of aprotinin. 157

8 2808 Circulation June 5, 2007 Weighted Mean Difference (Random Effects) Total Blood Loss by Agent (ml) Aprotinin vs E-Aminocaproic Acid E-Aminocaproic Acid vs Figure 6. Total blood loss by antifibrinolytic agent: head-to-head comparisons. The WMDs of total blood loss (ml) by antifibrinolytic agent compared head to head are plotted. WMD (diamond) and 95% CIs (horizontal bars) summarize the effect using a random-effects model. Aprotinin refers to high-dose aprotinin. The WMDs for aprotinin vs other antifibrinolytic agents are to the left of the 0, representing a significant reduction in total blood loss over tranexamic acid and -aminocaproic acid. The WMD for -aminocaproic acid vs tranexamic acid is to the left of 0, representing a reduction in total blood loss for -aminocaproic acid over tranexamic acid, but the horizontal line includes 0, denoting no significant difference. Our findings should be interpreted in light of limitations that are common to all meta-analyses. In meta-analysis, heterogeneity is the term used to describe inconsistencies in outcomes between studies. One potential source of heterogeneity in this meta-analysis is the combination of trials that include patients undergoing different cardiac procedures or treated with different antifibrinolytic doses. To address these issues, we performed stratified analyses to assess the extent to which our results varied according to procedure type and aprotinin dose. Finding no important interactions, we present only the combined results but have included the raw data in an appendix (online Data Supplement), allowing interested readers to make a more detailed examination of the results. Another limitation of meta-analysis is variability in the quality of the individual trials. In our study, 36 of the published trials included in this meta-analysis violated the intention-to-treat principle by excluding patients from the analysis after they had been randomized. Twenty-three trials excluded patients from the analysis for the occurrence of reexploration. Eighteen of those trials reported the study arm Relative Risk (Random Effects) Patients Transfused with prbc Aprotinin vs E-Aminocaproic Acid E-Aminocaproic Acid vs Return to Operating Room Aprotinin vs E-Aminocaproic Acid E-Aminocaproic Acid vs Mortality Stroke Aprotinin vs E-Aminocaproic Acid E-Aminocaproic Acid vs Aprotinin vs E-Aminocaproic Acid E-Aminocaproic Acid vs Myocardial Infarction Aprotinin vs E-Aminocaproic Acid E-Aminocaproic Acid vs Renal Failure Renal Dysfunction Figure 7. Adverse outcomes by antifibrinolytic agents: head-to-head comparison. The RRs of adverse outcomes (mortality, stroke, and myocardial infarction) by antifibrinolytic agent compared head to head are plotted. The RR (diamond) and 95% CIs (horizontal bars) summarize the effect using a random-effects model. Effects left of 1.0 favor agent 1 over agent 2, where agent 1 vs agent 2 is listed under the outcomes of interest; effects to the right favor agent 2 over agent 1. When the horizontal bars cross 1.0, the effect is not significantly different from the comparison group; this is the case for all comparisons plotted here.

9 Brown et al Antifibrinolytic Meta-Analysis 2809 from which the excluded patients (for reexploration) were derived. In our analysis, we reconciled these patients back to their originally assigned arms and included them in this meta-analysis on the intention-to-treat basis. However, 5 of 23 trials excluding patients for reexploration did not report which study arm the excluded patients were assigned; therefore, we were unable to reconcile those patients to their original arm. 140, In the case of high-dose aprotinin, 2 of these 5 studies reported reexploration as an outcome. 159,160 Both studies reported reexploration events for the placebo group but none for the aprotinin group despite the fact that some patients were excluded from the analysis for reexploration. In these 2 studies, there was concern about reporting bias. For this reason, we removed these studies from our reexploration analysis. Once removed, the summary effect (RR) changed from 0.47 (95% CI, 0.32 to 0.69; P 0.001) to 0.49 (95% CI, 0.33 to 0.73; P 0.001) and did not change the results. One reason to perform a meta-analysis is that numerous small trials will be performed to evaluate the effectiveness of a treatment. By combining studies, we increase the power to evaluate rare adverse effects. We conducted a post hoc analysis to assess the power of this meta-analysis to evaluate some of the rare adverse outcomes reported and found it lacking. For example, the power of our study ranged from 0.05 for dialysis-dependent renal failure to 0.35 for stroke. This study would have required patients in each treatment group to detect a significant difference in rates of dialysis-dependent renal failure. Renal dysfunction was the only adverse outcome for which we had sufficient power (0.85) to detect a significant difference between treatment groups. In summary, we have conducted an updated meta-analysis, confirming previous findings of meta-analyses that captured only published clinical trials through and We extended these findings by expanding the review of trials through July 2006 and by evaluating the dosing effect of aprotinin on blood loss and adverse outcomes. Our findings demonstrate that both high- and low-dose aprotinin, -aminocaproic acid, and tranexamic acid reduce total perioperative blood loss during cardiac surgery, reducing the need for transfusion. However, high-dose aprotinin results in less total blood loss than either of the competing antifibrinolytic agents and results in lower rates of reexploration. There was no difference among these agents in risks of mortality, stroke, myocardial infarction, or renal failure, but high-dose aprotinin significantly increased the rates of renal dysfunction. None. Disclosures References 1. Stover EP, Siegel LC, Parks R, Levin J, Body SC, Maddi R, D Ambra MN, Mangano DT, Spiess BD. Variability in transfusion practice for coronary artery bypass surgery persists despite national consensus guidelines: a 24-institution study. Anesthesiology. 1998;88: Chelemer SB, Prato BS, Cox PM, O Connor GT, Morton JR. Association of bacterial infection and red blood cell transfusion after coronary artery bypass surgery. Ann Thorac Surg. 2002;73: Speiss BD. Transfusion and outcome in heart surgery. Ann Thorac Surg. 2002;74: Johnson RG, Thurer RL, Kruskall MS, Sirois C, Gervino EV, Critchlow J, Weintraub RM. Comparison of 2 transfusion strategies after elective operations for myocardial revascularization. J Thorac Cardiovasc Surg. 1992;104: Busch ORC, Hop WCJ, Vanpapendrecht MAWH, Marquet RL, Jeekel J. Blood-transfusions and prognosis in colorectal-cancer. N Engl J Med. 1993;328: Mangano DT, Tudor IC, Dietzel C; for the Multicenter Study of Perioperative Ischemia Research Group, Ischemia Research and Education Foundation. The risk associated with aprotinin in cardiac surgery. N Engl J Med. 2006;354: Mangano DT, Miao Y, Vuylsteke A, Tudor IC, Juneja R, Filipescu D, Hoeft A, Fontes ML, Hillel Z, Ott E, Titov T, Dietzel C, Levin J. Mortality associated with aprotinin during 5 years following coronary artery bypass graft surgery. JAMA. 2007;297: Brown JR, Birkmeyer NJO, O Connor GT. Aprotinin in cardiac surgery. N Engl J Med. 2006;354: Alajmo F, Calamai G, Perna AM, Melissano G, Pretelli P, Palmarini MF, Carbonetto F, Noferi D, Boddi V, Palminiello A, Vaccari M. High-dose aprotinin: hemostatic effects in open-heart operations. Ann Thorac Surg. 1989;48: Alderman EL, Levy JH, Rich JB, Nili M, Vidne B, Schaff H, Uretzky G, Pettersson G, Thiis JJ, Hantler CB, Chaitman B, Nadel A. Analyses of coronary graft patency after aprotinin use: results from the International Multicenter Aprotinin Graft Patency Experience (IMAGE) trial. J Thorac Cardiovasc Surg. 1998;116: Alvarez JM, Jackson LR, Chatwin C, Smolich JJ. Low-dose postoperative aprotinin reduces mediastinal drainage and blood product use in patients undergoing primary coronary artery bypass grafting who are taking aspirin: a prospective, randomized, double-blind, placebocontrolled trial. J Thorac Cardiovasc Surg. 2001;122: Alvarez JM, Quiney NF, McMillan D, Joscelyne K, Connelly T, Brady P, Deal C, Wilson R. The use of ultra-low-dose aprotinin to reduce blood loss in cardiac surgery. J Cardiothorac Vasc Anesth. 1995;9: Andreasen JJ, Nielsen C. Prophylactic tranexamic acid in elective, primary coronary artery bypass surgery using cardiopulmonary bypass. Eur J Cardiothorac Surg. 2004;26: Armellin G, Casella S, Guzzinati S, Pasini L, Marcassa A, Giron G. Tranexamic acid in aortic valve replacement. J Cardiothorac Vasc Anesth. 2001;15: Asimakopoulos G, Kohn A, Stefanou DC, Haskard DO, Landis RC, Taylor KM. Leukocyte integrin expression in patients undergoing cardiopulmonary bypass. Ann Thorac Surg. 2000;69: Baele PL, Ruiz-Gomez J, Londot C, Sauvage M, Van Dyck MJ, Robert A. Systematic use of aprotinin in cardiac surgery: influence on total homologous exposure and hospital cost. Acta Anaesthesiol Belg. 1992; 43: Bailey CR, Kelleher AA, Wielogorski AK. Randomized placebocontrolled double-blind study of three aprotinin regimens in primary cardiac surgery. Br J Surg. 1994;81: Bailey CR, Wielogorski AK. Randomised placebo controlled double blind study of two low dose aprotinin regimens in cardiac surgery. Br Heart J. 1994;71: Basora M, Gomar C, Escolar G, Pacheco M, Fita G, Rodriguez E, Ordinas A. Platelet function during cardiac surgery and cardiopulmonary bypass with low-dose aprotinin. J Cardiothorac Vasc Anesth. 1999;13: Bennett-Guerrero E, Sorohan JG, Gurevich ML, Kazanjian PE, Levy RR, Barbera AV, White WD, Slaughter TF, Sladen RN, Smith PK, Newman MF. Cost-benefit and efficacy of aprotinin compared with epsilon-aminocaproic acid in patients having repeated cardiac operations: a randomized, blinded clinical trial. Anesthesiology. 1997;87: Bernet F, Carrel T, Marbet G, Skarvan K, Stulz P. Reduction of blood loss and transfusion requirements after coronary artery bypass grafting: similar efficacy of tranexamic acid and aprotinin in aspirin-treated patients. J Card Surg. 1999;14: Bidstrup B. Aprotinin for bleeding after cardiopulmonary bypass. Lancet. 1990;335: Bidstrup BP, Hunt BJ, Sheikh S, Parratt RN, Bidstrup JM, Sapsford RN. Amelioration of the bleeding tendency of preoperative aspirin after aortocoronary bypass grafting. Ann Thorac Surg. 2000;69:

10 2810 Circulation June 5, Bidstrup BP, Royston D, Taylor KM, Sapsford RN. Reduction in blood-loss and blood use after cardiopulmonary bypass with high-dose aprotinin (Trasylol). J Thorac Cardiovasc Surg. 1989;97: Bidstrup BP, Underwood SR, Sapsford RN, Streets EM. Effect of aprotinin (Trasylol) on aorta-coronary bypass graft patency. J Thorac Cardiovasc Surg. 1993;105: Blauhut B, Gross C, Necek S, Doran JE, Spath P, Lundsgaardhansen P. Effects of high-dose aprotinin on blood-loss, platelet-function, fibrinolysis, complement, and renal-function after cardiopulmonary bypass. J Thorac Cardiovasc Surg. 1991;101: Blauhut B, Harringer W, Bettelheim P, Doran JE, Spath P, Lundsgaardhansen P. Comparison of the effects of aprotinin and tranexamic acid on blood-loss and related variables after cardiopulmonary bypass. J Thorac Cardiovasc Surg. 1994;108: Boldt J, Schindler E, Knothe C, Hammermann H, Stertmann WA, Hempelmann G. Does aprotinin influence endothelial-associated coagulation in cardiac surgery? J Cardiothorac Vasc Anesth. 1994;8: Brown RS, Thwaites BK, Mongan PD. Tranexamic acid is effective in decreasing postoperative bleeding and transfusions in primary coronary artery bypass operations: a double-blind, randomized, placebocontrolled trial. Anesth Analg. 1997;85: Casas JI, Zuazujausoro I, Mateo J, Oliver A, Litvan H, Munizdiaz E, Aris A, Caralps JM, Fontcuberta J. Aprotinin versus desmopressin for patients undergoing operations with cardiopulmonary bypass: a double-blind placebo-controlled study. J Thorac Cardiovasc Surg. 1995; 110: Casati V, Bellotti F, Gerli C, Franco A, Oppizzi M, Cossolini M, Calori G, Benussi S, Alfieri O, Torri G. Tranexamic acid administration after cardiac surgery: a prospective, randomized, double-blind, placebocontrolled study. Anesthesiology. 2001;94: Casati V, Della Valle P, Benussi S, Franco A, Gerli C, Baili P, Alfieri O, D Angelo A. Effects of tranexamic acid on postoperative bleeding and related hematochemical variables in coronary surgery: comparison between on-pump and off-pump techniques. J Thorac Cardiovasc Surg. 2004;128: Casati V, Gerli C, Franco A, Torri G, D Angelo A, Benussi S, Alfieri O. Tranexamic acid in off-pump coronary surgery: a preliminary, randomized, double-blind, placebo-controlled study. Ann Thorac Surg. 2001;72: Casati V, Guzzon D, Oppizzi M, Bellotti F, Franco A, Gerli C, Cossolini M, Torri G, Calori G, Benussi S, Alfieri O. Tranexamic acid compared with high-dose aprotinin in primary elective heart operations: effects on perioperative bleeding and allogeneic transfusions. J Thorac Cardiovasc Surg. 2000;120: Casati V, Guzzon D, Oppizzi M, Cossolini M, Torri G, Calori G, Alfieri O. Hemostatic effects of aprotinin, tranexamic acid and epsilonaminocaproic acid in primary cardiac surgery. Ann Thorac Surg. 1999; 68: Cicek S, Demirkilic U, Kuralay E, Ozal E, Tatar H. Postoperative aprotinin: effect on blood loss and transfusion requirements in cardiac operations. Ann Thorac Surg. 1996;61: Cicek S, Demirkilic U, Ozal E, Kuralay E, Bingol H, Tatar H, Ozturk OY. Postoperative use of aprotinin in cardiac operations: an alternative to its prophylactic use. J Thorac Cardiovasc Surg. 1996;112: Coffey A, Pittmam J, Halbrook H, Fehrenbacher J, Beckman D, Hormuth D. The use of tranexamic acid to reduce postoperative bleeding following cardiac surgery: a double-blind randomized trial. Am Surg. 1995;61: Cosgrove DM, Heric B, Lytle BW, Taylor PC, Novoa R, Golding LAR, Stewart RW, Mccarthy PM, Loop FD. Aprotinin therapy for reoperative myocardial revascularization: a placebo-controlled study. Ann Thorac Surg. 1992;54: D Ambra MN, Akins CW, Blackstone EH, Bonney SL, Cohn LH, Cosgrove DM, Levy JH, Lynch KE, Maddi R. Aprotinin in primary valve replacement and reconstruction: a multicenter, double-blind, placebo-controlled trial. J Thorac Cardiovasc Surg. 1996;112: Daily PO, Lamphere JA, Dembitsky WP, Adamson RM, Dans NF. Effect of prophylactic epsilon-aminocaproic acid on blood loss and transfusion requirements in patients undergoing first-time coronary artery bypass grafting: a randomized, prospective, double-blind study. J Thorac Cardiovasc Surg. 1994;108: De Bonis M, Cavaliere F, Alessandrini F, Lapenna E, Santarelli F, Moscato U, Schiavello R, Possati GF. Topical use of tranexamic acid in coronary artery bypass operations: a double-blind, prospective, randomized, placebo-controlled study. J Thorac Cardiovasc Surg. 2000; 119: Defraigne JO, Pinc J, Larbuisson R, Blaffart F, Limet R. Cytokine release and neutrophil activation are not prevented by heparin-coated circuits and aprotinin administration. Ann Thorac Surg. 2000;69: Delrossi AJ, Cernaianu AC, Botros S, Lemole GM, Moore R. Prophylactic treatment of postperfusion bleeding using EACA. Chest. 1989;96: Dernek S, Sevin B, Aslan R, Kaklikkaya I, Atalay C, Kural T. Effects of aprotinin on plasma levels of neutrophil elastase and postoperative blood loss in cardiac surgery. Cardiovasc Surg. 1996;4: Dietrich W, Barankay A, Hahnel C, Richter JA. High-dose aprotinin in cardiac surgery: three years experience in 1,784 patients. J Cardiothorac Vasc Anesth. 1992;6: Dietrich W, Dilthey G, Spannagl M, Jochum M, Braun SL, Richter JA. Influence of high-dose aprotinin on anticoagulation, heparin requirement, and celite- and kaolin-activated clotting time in heparinpretreated patients undergoing open-heart surgery: a double-blind, placebo-controlled study. Anesthesiology. 1995;83: Dietrich W, Spannagl M, Jochum M, Wendt P, Schramm W, Barankay A, Sebening F, Richter JA. Influence of high-dose aprotinin treatment on blood loss and coagulation patterns in patients undergoing myocardial revascularization. Anesthesiology. 1990;73: Diprose P, Herbertson MJ, O Shaughnessy D, Deakin CD, Gill RS. Reducing allogeneic transfusion in cardiac surgery: a randomized double-blind placebo-controlled trial of antifibrinolytic therapies used in addition to intra-operative cell salvage. Br J Anaesth. 2005;94: Dryden PJ, O Connor JP, Jamieson WR, Reid I, Ansley D, Sadeghi H, Burr LH, Munro AI, Merrick PM. Tranexamic acid reduces blood loss and transfusion in reoperative cardiac surgery. Can J Anaesth. 1997;44: Englberger L, Kipfer B, Berdat PA, Nydegger UE, Carrel TP. Aprotinin in coronary operation with cardiopulmonary bypass: does low-dose aprotinin inhibit the inflammatory response? Ann Thorac Surg. 2002; 73: Englberger L, Markart P, Eckstein FS, Immer FF, Berdat PA, Carrel TP. Aprotinin reduces blood loss in off-pump coronary artery bypass (OPCAB) surgery. Eur J Cardiothorac Surg. 2002;22: Fauli A, Gomar C, Campistol JM, Alvarez L, Manig AM, Matute P. Kidney-specific proteins in patients receiving aprotinin at high- and low-dose regimens during coronary artery bypass graft with cardiopulmonary bypass. Eur J Anaesthesiol. 2005;22: Feindt P, Seyfert U, Volkmer I, Huwer H, Kalweit G, Gams E. Is there a phase of hypercoagulability when aprotinin is used in cardiac surgery? Eur J Cardiothorac Surg. 1994;8: Feindt PR, Walcher S, Volkmer I, Keller HE, Straub U, Huwer H, Seyfert UT, Petzold T, Gams E. Effects of high-dose aprotinin on renal function in aortocoronary bypass grafting. Ann Thorac Surg. 1995;60: Forestier F, Belisle S, Robitaille D, Martineau R, Perrault LP, Hardy JF. Low-dose aprotinin is ineffective to treat excessive bleeding after cardiopulmonary bypass. Ann Thorac Surg. 2000;69: Fraedrich G, Weber C, Bernard C, Hettwer A, Schlosser V. Reduction of blood transfusion requirement in open heart surgery by administration of high doses of aprotinin: preliminary results. Thorac Cardiovasc Surg. 1989;37: Golanski R, Golanski J, Chizynski K, Iwaszkiewicz A, Zaslonka J, Pietrucha T, Chrul S, Watala C. Low doses of aprotinin in aortocoronary bypass surgery: advantages and disadvantages. Med Sci Monit. 2000;6: Gott JP, Cooper WA, Schmidt FE Jr, Brown WM 3rd, Wright CE, Merlino JD, Fortenberry JD, Clark WS, Guyton RA. Modifying risk for extracorporeal circulation: trial of four antiinflammatory strategies. Ann Thorac Surg. 1998;66: Green D, Sanders J, Eiken M, Wong CA, Frederiksen J, Joob A, Palmer A, Trowbridge A, Woodruff B, Moerch M. Recombinant aprotinin in coronary artery bypass graft operations. J Thorac Cardiovasc Surg. 1995;110: Greilich PE, Brouse CF, Rinder CS, Smith BR, Sandoval BA, Rinder HM, Eberhart RC, Jessen ME. Effects of epsilon-aminocaproic acid and aprotinin on leukocyte-platelet adhesion in patients undergoing cardiac surgery. Anesthesiology. 2004;100: