Aprotinin in Cardiac Surgery
http://www.100md.com
《新英格兰医药杂志》
To the Editor: Although potent and effective drugs may have undesirable side effects in some patients, we believe that the study by Mangano et al. (Jan. 26 issue)1 overemphasizes the risk of aprotinin for the prevention of bleeding during cardiac surgery without a demonstration of its substantial benefits. This overemphasis may have occurred because of a flawed study design, imprecision in the definition of variables, selection bias during development of the study model, and a lack of proper weighting of evidence from the literature.
The observational study involved institutions in many countries with no reported uniformity in indications for drug administration. The period in which postoperative mortality was assessed was not stated. The sampling method that was used to identify study patients was peculiar, with more weight given to programs with a low volume of patients. The exclusion criteria that were used need further explanation, since the death rate among patients who were excluded from the study was three times that in the experimental group.1,2 Furthermore, the rate of renal failure in a previous article from the same database was 7.8 percent,3 not too different from the rate of 8 percent shown for aprotinin in Figure 2 of the article and substantially different from the control rate of 3 percent.1
Many important determinants of outcome were not described. These data included the use of antithrombotic drugs before surgery, the duration of cardiopulmonary bypass or other operative details, the amounts of blood transfused, and important postoperative details, such as use of inotropic drugs and the occurrence of hypotensive episodes, all of which can affect outcome. Moreover, the article minimizes the risk associated with blood transfusion and recommends the use of antifibrinolytic drugs (tranexamic acid and aminocaproic acid) that are either not approved for use in Australia, Japan, and Europe or not approved for this indication.
During the development of guidelines on blood-conservation strategies, the Workforce on Evidence-Based Surgery of the Society of Thoracic Surgeons found several dozen randomized trials, including a meta-analysis, and a Cochrane Collaboration summary that indicated no significant change in mortality, myocardial infarction, or the risk of renal failure associated with the use of aprotinin. Aprotinin was associated with a reduced risk of stroke and a trend toward a reduced rate of postoperative atrial fibrillation.4,5
The best evidence suggests that aprotinin has an acceptable risk–benefit profile and is indicated for blood conservation in patients at increased risk for bleeding (e.g., patients undergoing complex or repeated cardiac procedures and patients with either congenital or acquired abnormalities of the hemostatic system). We trust that the study by Mangano et al. will not deter surgeons from the use of aprotinin to reduce bleeding and the need for blood transfusion in such patients. Although routine use of any drug with potentially substantial side effects should be avoided, a reasoned selective use with consideration of the risk and benefit is always appropriate.
Victor A. Ferraris, M.D., Ph.D.
University of Kentucky Chandler Medical Center
Lexington, KY 40536
ferraris@earthlink.net
Charles R. Bridges, M.D., Sc.D.
University of Pennsylvania Health System
Philadelphia, PA 19106
Richard P. Anderson, M.D.
Virginia Mason Clinic
Seattle, WA 98102
for the Blood Conservation Guideline Taskforce
Dr. Ferraris reports having received research grant support from Aventis, Bayer (for a phase 4 trial), Guilford, and the Medicines Company and having received lecture fees from Aventis, Bayer, and the Medicines Company.
References
Mangano DT, Tudor IC, Dietzel C. The risk associated with aprotinin in cardiac surgery. N Engl J Med 2006;354:353-365.
Mangano DT. Aspirin and mortality from coronary bypass surgery. N Engl J Med 2002;347:1309-1317.
Mangano CM, Diamondstone LS, Ramsay JG, Aggarwal A, Herskowitz A, Mangano DT. Renal dysfunction after myocardial revascularization: risk factors, adverse outcomes, and hospital resource utilization. Ann Intern Med 1998;128:194-203.
Sedrakyan A, Treasure T, Elefteriades JA. Effect of aprotinin on clinical outcomes in coronary artery bypass graft surgery: a systematic review and meta-analysis of randomized clinical trials. J Thorac Cardiovasc Surg 2004;128:442-448.
Henry DA, Moxey AJ, Carless PA, et al. Anti-fibrinolytic use for minimising perioperative allogeneic blood transfusion. Cochrane Database Syst Rev 2001;1:CD001886-CD001886.
To the Editor: In 1999, we conducted a meta-analysis of randomized trials to compare the effectiveness and risks of both aminocaproic acid and aprotinin with those of placebo.1 We found that the two drugs were equally effective in preventing postoperative bleeding, with an insignificant trend toward increased renal dysfunction with full-dose aprotinin (odds ratio, 1.46; 95 percent confidence interval, 0.92 to 2.33). We have updated our analyses to incorporate new trials (Figure 1).
Figure 1. Meta-Analysis of the Relative Risk of Renal Failure and Renal Dysfunction Associated with Full-Dose Aprotinin, as Compared with Placebo.
Randomized, controlled trials are listed chronologically by the year of publication. Renal failure was defined as the need for dialysis or an increase of at least 2.0 mg per deciliter in the creatinine level, as compared with baseline. Renal dysfunction was defined as an increase of at least 0.5 mg per deciliter in the creatinine level, as compared with baseline. Trials with no events in both the aprotinin group and the placebo group could not be used to estimate relative risk. The size of each square denotes the weight of each trial's relative risk in the calculation of the combined relative risk. Diamonds represent the combined relative risk at the center; the opposing points of the diamonds represent the 95 percent confidence interval (CI). I2 denotes the percentage of total variation among studies that is due to heterogeneity, rather than chance. A complete list of references is in the Supplementary Appendix, available with the full text of this letter at www.nejm.org.
Randomized trials comparing full-dose aprotinin (at a dose of 2.0 million kallikrein-inhibitor units ) with placebo were included. The combined relative risk was 1.09 (95 percent confidence interval, 0.68 to 1.76) for renal failure and 1.47 (95 percent confidence interval, 1.12 to 1.94) for renal dysfunction. (Renal dysfunction occurred in 8.4 percent of patients receiving placebo and 12.9 percent of those receiving aprotinin.) There were too few trials to estimate reliably the renal outcome of patients receiving half-dose aprotinin.
Our analyses of the combined evidence from clinical trials support the finding of Mangano et al. that patients receiving aprotinin have an increased rate of renal dysfunction but not of renal failure.
Jeremiah R. Brown, M.S.
Dartmouth Medical School
Lebanon, NH 03756
jeremiah.r.brown@dartmouth.edu
Nancy J.O. Birkmeyer, Ph.D.
Michigan Surgical Collaboration for Outcomes Research
and Evaluation
Ann Arbor, MI 48109
Gerald T. O'Connor, Ph.D., D.Sc.
Dartmouth Medical School
Lebanon, NH 03756
References
Munoz JJ, Birkmeyer NJ, Birkmeyer JD, O'Connor GT, Dacey LJ. Is epsilon-aminocaproic acid as effective as aprotinin in reducing bleeding with cardiac surgery? A meta-analysis. Circulation 1999;99:81-89.
To the Editor: Mangano et al. report on an observational, nonrandomized database that was subjected to extensive statistical analyses. The underlying question related to any analysis involving nonrandomized patients is why patients receive particular therapies.1 In observational studies, clinicians control the treatment that is assigned. Various groups that are evaluated may have large differences in their observed covariates, differences that can lead to biased estimates of treatment effects.2 Propensity scoring reduces bias but does not eliminate it; thus, sicker patients receive different treatments.2 Factors that influence outcomes — such as the duration of cardiopulmonary bypass, the exact dose of aprotinin, and the use of platelet transfusion and anticoagulation — are not reported.
Mangano et al. identified an association between patients in whom aprotinin was used and the patients who had complications, but such identification does not establish a cause-and-effect relationship any more than the study by Connors et al. established a cause-and-effect relationship between the use or nonuse of Swan–Ganz catheters in patients receiving intensive care and mortality.3 We also find it unusual that none of the investigators who collected the data are authors of the manuscript, despite the presence of a well-defined structure in the Multicenter Study of Perioperative Ischemia Research Group that is designed specifically for this purpose.
Jerrold H. Levy, M.D.
James G. Ramsay, M.D.
Robert A. Guyton, M.D.
Emory University School of Medicine
Atlanta, GA 30322
jerrold.levy@emoryhealthcare.org
Dr. Levy reports having received consulting fees from Bayer Pharmaceutical.
References
D'Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17:2265-2281.
Byar DP. Problems with using observational databases to compare treatments. Stat Med 1991;10:663-666.
Connors AF Jr, Speroff T, Dawson NV, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 1996;276:889-897.
To the Editor: In the study by Mangano et al. of outcome data from 4374 patients undergoing coronary revascularization, significant dose-dependent renal dysfunction or dialysis was associated with the administration of aprotinin. In a study cited in the article,1 my colleagues and I prominently described a dose-dependent, statistically significant relationship between renal dysfunction or dialysis and diabetes mellitus in patients receiving aprotinin during cardiac surgery. Unfortunately, the analysis by Mangano et al. did not include an inquiry into this potentially causal relationship. Since our study did not have sufficient power to detect an adverse outcome of renal failure, it would have been logical for Mangano et al. to have performed an analysis of the correlation or covariance of preexisting diabetes and aprotinin therapy with renal adverse outcomes.
Since 1996, we have advocated the avoidance of aprotinin therapy among patients with diabetes mellitus, and it would be gratifying to know whether the data reported by Mangano et al. support this stance. Such an analysis would most certainly augment our understanding of the risk profile of this potent drug. Since the presence of diabetes can be easily assessed preoperatively, it is essential for the authors to report this information to us.
Michael N. D'Ambra, M.D.
Brigham and Women's Hospital
Boston, MA 02115
mdambra@partners.org
Dr. D'Ambra reports having received grant support and an honorarium from Bayer.
References
D'Ambra MN, Akins CW, Blackstone EH, et al. Aprotinin in primary valve replacement and reconstruction: a multicenter, double-blind, placebo-controlled trial. J Thorac Cardiovasc Surg 1996;112:1081-1089.
The author replies: Hunter's incisive Perspective article1 in the same issue of the Journal as our article heralds a fundamental paradigm shift in the methods that are used to assess drug safety. I refer Drs. Ferraris and Levy et al. to that article regarding their concern about the design of our study. Dr. Ferraris and colleagues have also erred in their citation of a previous study from the database of the Multicenter Study of Perioperative Ischemia Epidemiology II; careful review will reveal that our findings are entirely consistent. The oft-quoted but erroneous claim that aprotinin has beneficial effects on stroke and atrial fibrillation is not based on rigorous science but only on post hoc exploratory analyses. Finally, given the steadfast positions that have emerged, we welcome the review by Dr. Ferraris and others of our underlying raw data, which indicate the serious safety problems of aprotinin.
The findings of Dr. Brown and colleagues are consistent with ours, as are those of a recently published study by Karkouti et al.2 showing a significant 41 percent increase in renal dysfunction or renal failure associated with aprotinin. Finally, added to these results are the 1993 findings of the Food and Drug Administration (FDA) that "kidney toxicity also was a problem in some patients in the trials."3
Dr. Levy and colleagues suggest that the unilateral approach to drug safety is the randomized, controlled trial. However, this suggestion raises an associated troubling issue. The largest sponsor-supported, randomized, controlled trial of aprotinin — the 1998 International Multicenter Aprotinin Graft Patency Experience (IMAGE) trial, which involved 870 patients4 — showed that aprotinin was associated with a statistically significant 41 percent increase in the incidence of acute coronary vein–graft closure. However, even then this serious safety finding was effectively dismissed with the use of post hoc adjustment arguments — a fact that negates the "gold standard" arguments implied by Dr. Levy and colleagues. Regarding causality, we agree, but we note that causal inference is supported by the magnitude of the association between the use of aprotinin and renal failure reported in our study and by the particularly powerful dose–response relationship.
Finally, we are also concerned about the authorship question. Two members of the Multicenter Study of Perioperative Ischemia Research Group who were entitled to authorship declined to have their names published in the list of authors after learning that the results of the trial were negative. Because of that fact, and because both investigators had a financial conflict of interest, neither one reviewed the manuscript.
Our multivariable analyses formally addressed Dr. D'Ambra's concern about the association of the use of aprotinin with diabetes. The findings prevailed for multiple characterizations of diabetes (Table 1).
Table 1. Incidence of Renal Events among 4374 Study Patients, According to History of Diabetes Mellitus.
In summary, we believe that the 1993 proclamation of the FDA regarding renal toxic effects is both telling and disturbing, since during the subsequent decade, no sufficiently large randomized, controlled trial has been performed to address the FDA's concern. Now, 13 years later, several independent studies have validated that concern and have shown that safe, effective, and far less expensive alternatives exist. Therefore, shouldn't we now err on the side of protecting the patient ("First, do no harm") instead of protecting the drug?
Dennis T. Mangano, Ph.D., M.D.
Ischemia Research and Education Foundation
San Bruno, CA 94066
dtb@iref.org
References
Hunter D. First, gather the data. N Engl J Med 2006;354:329-331.
Karkouti K, Beattie WS, Dattilo KM, et al. A propensity score case-control comparison of aprotinin and tranexamic acid in high-transfusion-risk cardiac surgery. Transfusion 2006;46:327-338.
Approval of aprotinin: FDA press release, December 30, 1993. (Accessed April 13, 2006, at http://www.fda.gov/bbs/topics/NEWS/NEW00453.html.)
Alderman EL, Levy JH, Rich JB, et al. Analyses of coronary graft patency after aprotinin use: results from the International Multicenter Aprotinin Graft Patency Experience (IMAGE) trial. J Thorac Cardiovasc Surg 1998;116:716-730.
The observational study involved institutions in many countries with no reported uniformity in indications for drug administration. The period in which postoperative mortality was assessed was not stated. The sampling method that was used to identify study patients was peculiar, with more weight given to programs with a low volume of patients. The exclusion criteria that were used need further explanation, since the death rate among patients who were excluded from the study was three times that in the experimental group.1,2 Furthermore, the rate of renal failure in a previous article from the same database was 7.8 percent,3 not too different from the rate of 8 percent shown for aprotinin in Figure 2 of the article and substantially different from the control rate of 3 percent.1
Many important determinants of outcome were not described. These data included the use of antithrombotic drugs before surgery, the duration of cardiopulmonary bypass or other operative details, the amounts of blood transfused, and important postoperative details, such as use of inotropic drugs and the occurrence of hypotensive episodes, all of which can affect outcome. Moreover, the article minimizes the risk associated with blood transfusion and recommends the use of antifibrinolytic drugs (tranexamic acid and aminocaproic acid) that are either not approved for use in Australia, Japan, and Europe or not approved for this indication.
During the development of guidelines on blood-conservation strategies, the Workforce on Evidence-Based Surgery of the Society of Thoracic Surgeons found several dozen randomized trials, including a meta-analysis, and a Cochrane Collaboration summary that indicated no significant change in mortality, myocardial infarction, or the risk of renal failure associated with the use of aprotinin. Aprotinin was associated with a reduced risk of stroke and a trend toward a reduced rate of postoperative atrial fibrillation.4,5
The best evidence suggests that aprotinin has an acceptable risk–benefit profile and is indicated for blood conservation in patients at increased risk for bleeding (e.g., patients undergoing complex or repeated cardiac procedures and patients with either congenital or acquired abnormalities of the hemostatic system). We trust that the study by Mangano et al. will not deter surgeons from the use of aprotinin to reduce bleeding and the need for blood transfusion in such patients. Although routine use of any drug with potentially substantial side effects should be avoided, a reasoned selective use with consideration of the risk and benefit is always appropriate.
Victor A. Ferraris, M.D., Ph.D.
University of Kentucky Chandler Medical Center
Lexington, KY 40536
ferraris@earthlink.net
Charles R. Bridges, M.D., Sc.D.
University of Pennsylvania Health System
Philadelphia, PA 19106
Richard P. Anderson, M.D.
Virginia Mason Clinic
Seattle, WA 98102
for the Blood Conservation Guideline Taskforce
Dr. Ferraris reports having received research grant support from Aventis, Bayer (for a phase 4 trial), Guilford, and the Medicines Company and having received lecture fees from Aventis, Bayer, and the Medicines Company.
References
Mangano DT, Tudor IC, Dietzel C. The risk associated with aprotinin in cardiac surgery. N Engl J Med 2006;354:353-365.
Mangano DT. Aspirin and mortality from coronary bypass surgery. N Engl J Med 2002;347:1309-1317.
Mangano CM, Diamondstone LS, Ramsay JG, Aggarwal A, Herskowitz A, Mangano DT. Renal dysfunction after myocardial revascularization: risk factors, adverse outcomes, and hospital resource utilization. Ann Intern Med 1998;128:194-203.
Sedrakyan A, Treasure T, Elefteriades JA. Effect of aprotinin on clinical outcomes in coronary artery bypass graft surgery: a systematic review and meta-analysis of randomized clinical trials. J Thorac Cardiovasc Surg 2004;128:442-448.
Henry DA, Moxey AJ, Carless PA, et al. Anti-fibrinolytic use for minimising perioperative allogeneic blood transfusion. Cochrane Database Syst Rev 2001;1:CD001886-CD001886.
To the Editor: In 1999, we conducted a meta-analysis of randomized trials to compare the effectiveness and risks of both aminocaproic acid and aprotinin with those of placebo.1 We found that the two drugs were equally effective in preventing postoperative bleeding, with an insignificant trend toward increased renal dysfunction with full-dose aprotinin (odds ratio, 1.46; 95 percent confidence interval, 0.92 to 2.33). We have updated our analyses to incorporate new trials (Figure 1).
Figure 1. Meta-Analysis of the Relative Risk of Renal Failure and Renal Dysfunction Associated with Full-Dose Aprotinin, as Compared with Placebo.
Randomized, controlled trials are listed chronologically by the year of publication. Renal failure was defined as the need for dialysis or an increase of at least 2.0 mg per deciliter in the creatinine level, as compared with baseline. Renal dysfunction was defined as an increase of at least 0.5 mg per deciliter in the creatinine level, as compared with baseline. Trials with no events in both the aprotinin group and the placebo group could not be used to estimate relative risk. The size of each square denotes the weight of each trial's relative risk in the calculation of the combined relative risk. Diamonds represent the combined relative risk at the center; the opposing points of the diamonds represent the 95 percent confidence interval (CI). I2 denotes the percentage of total variation among studies that is due to heterogeneity, rather than chance. A complete list of references is in the Supplementary Appendix, available with the full text of this letter at www.nejm.org.
Randomized trials comparing full-dose aprotinin (at a dose of 2.0 million kallikrein-inhibitor units ) with placebo were included. The combined relative risk was 1.09 (95 percent confidence interval, 0.68 to 1.76) for renal failure and 1.47 (95 percent confidence interval, 1.12 to 1.94) for renal dysfunction. (Renal dysfunction occurred in 8.4 percent of patients receiving placebo and 12.9 percent of those receiving aprotinin.) There were too few trials to estimate reliably the renal outcome of patients receiving half-dose aprotinin.
Our analyses of the combined evidence from clinical trials support the finding of Mangano et al. that patients receiving aprotinin have an increased rate of renal dysfunction but not of renal failure.
Jeremiah R. Brown, M.S.
Dartmouth Medical School
Lebanon, NH 03756
jeremiah.r.brown@dartmouth.edu
Nancy J.O. Birkmeyer, Ph.D.
Michigan Surgical Collaboration for Outcomes Research
and Evaluation
Ann Arbor, MI 48109
Gerald T. O'Connor, Ph.D., D.Sc.
Dartmouth Medical School
Lebanon, NH 03756
References
Munoz JJ, Birkmeyer NJ, Birkmeyer JD, O'Connor GT, Dacey LJ. Is epsilon-aminocaproic acid as effective as aprotinin in reducing bleeding with cardiac surgery? A meta-analysis. Circulation 1999;99:81-89.
To the Editor: Mangano et al. report on an observational, nonrandomized database that was subjected to extensive statistical analyses. The underlying question related to any analysis involving nonrandomized patients is why patients receive particular therapies.1 In observational studies, clinicians control the treatment that is assigned. Various groups that are evaluated may have large differences in their observed covariates, differences that can lead to biased estimates of treatment effects.2 Propensity scoring reduces bias but does not eliminate it; thus, sicker patients receive different treatments.2 Factors that influence outcomes — such as the duration of cardiopulmonary bypass, the exact dose of aprotinin, and the use of platelet transfusion and anticoagulation — are not reported.
Mangano et al. identified an association between patients in whom aprotinin was used and the patients who had complications, but such identification does not establish a cause-and-effect relationship any more than the study by Connors et al. established a cause-and-effect relationship between the use or nonuse of Swan–Ganz catheters in patients receiving intensive care and mortality.3 We also find it unusual that none of the investigators who collected the data are authors of the manuscript, despite the presence of a well-defined structure in the Multicenter Study of Perioperative Ischemia Research Group that is designed specifically for this purpose.
Jerrold H. Levy, M.D.
James G. Ramsay, M.D.
Robert A. Guyton, M.D.
Emory University School of Medicine
Atlanta, GA 30322
jerrold.levy@emoryhealthcare.org
Dr. Levy reports having received consulting fees from Bayer Pharmaceutical.
References
D'Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17:2265-2281.
Byar DP. Problems with using observational databases to compare treatments. Stat Med 1991;10:663-666.
Connors AF Jr, Speroff T, Dawson NV, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 1996;276:889-897.
To the Editor: In the study by Mangano et al. of outcome data from 4374 patients undergoing coronary revascularization, significant dose-dependent renal dysfunction or dialysis was associated with the administration of aprotinin. In a study cited in the article,1 my colleagues and I prominently described a dose-dependent, statistically significant relationship between renal dysfunction or dialysis and diabetes mellitus in patients receiving aprotinin during cardiac surgery. Unfortunately, the analysis by Mangano et al. did not include an inquiry into this potentially causal relationship. Since our study did not have sufficient power to detect an adverse outcome of renal failure, it would have been logical for Mangano et al. to have performed an analysis of the correlation or covariance of preexisting diabetes and aprotinin therapy with renal adverse outcomes.
Since 1996, we have advocated the avoidance of aprotinin therapy among patients with diabetes mellitus, and it would be gratifying to know whether the data reported by Mangano et al. support this stance. Such an analysis would most certainly augment our understanding of the risk profile of this potent drug. Since the presence of diabetes can be easily assessed preoperatively, it is essential for the authors to report this information to us.
Michael N. D'Ambra, M.D.
Brigham and Women's Hospital
Boston, MA 02115
mdambra@partners.org
Dr. D'Ambra reports having received grant support and an honorarium from Bayer.
References
D'Ambra MN, Akins CW, Blackstone EH, et al. Aprotinin in primary valve replacement and reconstruction: a multicenter, double-blind, placebo-controlled trial. J Thorac Cardiovasc Surg 1996;112:1081-1089.
The author replies: Hunter's incisive Perspective article1 in the same issue of the Journal as our article heralds a fundamental paradigm shift in the methods that are used to assess drug safety. I refer Drs. Ferraris and Levy et al. to that article regarding their concern about the design of our study. Dr. Ferraris and colleagues have also erred in their citation of a previous study from the database of the Multicenter Study of Perioperative Ischemia Epidemiology II; careful review will reveal that our findings are entirely consistent. The oft-quoted but erroneous claim that aprotinin has beneficial effects on stroke and atrial fibrillation is not based on rigorous science but only on post hoc exploratory analyses. Finally, given the steadfast positions that have emerged, we welcome the review by Dr. Ferraris and others of our underlying raw data, which indicate the serious safety problems of aprotinin.
The findings of Dr. Brown and colleagues are consistent with ours, as are those of a recently published study by Karkouti et al.2 showing a significant 41 percent increase in renal dysfunction or renal failure associated with aprotinin. Finally, added to these results are the 1993 findings of the Food and Drug Administration (FDA) that "kidney toxicity also was a problem in some patients in the trials."3
Dr. Levy and colleagues suggest that the unilateral approach to drug safety is the randomized, controlled trial. However, this suggestion raises an associated troubling issue. The largest sponsor-supported, randomized, controlled trial of aprotinin — the 1998 International Multicenter Aprotinin Graft Patency Experience (IMAGE) trial, which involved 870 patients4 — showed that aprotinin was associated with a statistically significant 41 percent increase in the incidence of acute coronary vein–graft closure. However, even then this serious safety finding was effectively dismissed with the use of post hoc adjustment arguments — a fact that negates the "gold standard" arguments implied by Dr. Levy and colleagues. Regarding causality, we agree, but we note that causal inference is supported by the magnitude of the association between the use of aprotinin and renal failure reported in our study and by the particularly powerful dose–response relationship.
Finally, we are also concerned about the authorship question. Two members of the Multicenter Study of Perioperative Ischemia Research Group who were entitled to authorship declined to have their names published in the list of authors after learning that the results of the trial were negative. Because of that fact, and because both investigators had a financial conflict of interest, neither one reviewed the manuscript.
Our multivariable analyses formally addressed Dr. D'Ambra's concern about the association of the use of aprotinin with diabetes. The findings prevailed for multiple characterizations of diabetes (Table 1).
Table 1. Incidence of Renal Events among 4374 Study Patients, According to History of Diabetes Mellitus.
In summary, we believe that the 1993 proclamation of the FDA regarding renal toxic effects is both telling and disturbing, since during the subsequent decade, no sufficiently large randomized, controlled trial has been performed to address the FDA's concern. Now, 13 years later, several independent studies have validated that concern and have shown that safe, effective, and far less expensive alternatives exist. Therefore, shouldn't we now err on the side of protecting the patient ("First, do no harm") instead of protecting the drug?
Dennis T. Mangano, Ph.D., M.D.
Ischemia Research and Education Foundation
San Bruno, CA 94066
dtb@iref.org
References
Hunter D. First, gather the data. N Engl J Med 2006;354:329-331.
Karkouti K, Beattie WS, Dattilo KM, et al. A propensity score case-control comparison of aprotinin and tranexamic acid in high-transfusion-risk cardiac surgery. Transfusion 2006;46:327-338.
Approval of aprotinin: FDA press release, December 30, 1993. (Accessed April 13, 2006, at http://www.fda.gov/bbs/topics/NEWS/NEW00453.html.)
Alderman EL, Levy JH, Rich JB, et al. Analyses of coronary graft patency after aprotinin use: results from the International Multicenter Aprotinin Graft Patency Experience (IMAGE) trial. J Thorac Cardiovasc Surg 1998;116:716-730.