Abstract
Voluntary withdrawal of the drug aprotinin from the US market in 2008 has increased the risk of bleeding in babies and young children after reconstructive cardiac surgery. The data on which the Bayer Corporation and FDA based their decision to withdraw the drug were obtained entirely from adults who are at risk of a different profile of complications following cardiac surgery than children. Furthermore, serious questions have been raised regarding the initial widely publicized article which questioned the safety of aprotinin. Some believe that the article was a misapplication of the new statistical method of propensity scoring. The fact that the principal author has not allowed free access to the raw data has raised further questions. Aprotinin is no longer under patent protection and could be evaluated under the pediatric exclusivity provision. It presents an excellent opportunity for alternative suppliers to undertake a clinical trial in a pediatric population of heart surgery patients.
Keywords
Introduction
In January 2006, an influential article appeared in the New England Journal of Medicine (NEJM) regarding aprotinin, a drug used to control bleeding during and after heart surgery in children and adults. 1 The article, which has subsequently been shown to be seriously flawed and included only adults in its analysis, suggested that aprotinin was associated with an increased risk of death, stroke, and kidney complications. Although no longer under patent protection aprotinin had been marketed exclusively as ‘Trasylol’ in the US by the Bayer Pharmaceutical Corporation. In 2007, the US Food and Drug Administration suggested to Bayer that it should suspend marketing of aprotinin. In May 2008, Bayer notified the FDA that in spite of strenuous protest by pediatric heart surgeons, the company was voluntarily removing all remaining stocks of aprotinin from the US market so that it would no longer be available to reduce the risk of bleeding after heart surgery in newborns and infants. Most other countries initially followed the US lead and discontinued the use of aprotinin, although it was available from other suppliers outside of the US. Two and a half years later, cardiac surgeons in many countries have returned to the use of aprotinin and work with alternative suppliers to the Bayer Corporation. However, in the US where the Bayer Corporation has been until now the sole FDA approved supplier, aprotinin remains unavailable for children and adults undergoing complex cardiac surgery.
The aprotinin story highlights a number of important regulatory issues:
Both the FDA and the Bayer Corporation require the same regulatory status for pediatric application of a drug as is stipulated for adult application even when the data on which the decision to withdraw the drug is based entirely on elderly adults. The FDA denies responsibility in preventing access of children undergoing heart surgery to aprotinin, suggesting that the withdrawal from the market by Bayer was a corporate decision. On the other hand, the Bayer Corporation implicates the FDA, suggesting that the FDA made it untenable to continue distribution of aprotinin. The relatively small size of the pediatric cardiac surgical market in contrast to the large adult cardiac surgical market almost certainly influenced the corporate decision regarding the withholding of aprotinin from the pediatric market. Publication of a single article within a prestigious medical journal like the NEJM can have a profound effect on decisions by both the FDA and pharmaceutical corporations related at least in part to the fear of subsequent malpractice suits.
Background
Aprotinin is a naturally occurring broad-spectrum serine protease inhibitor, which is obtained from bovine lung. 2 By inhibiting proteases such as kallikrein, trypsin, and plasmin, it modulates many important functions such as blood clotting, fibrinolysis (the breakdown of blood clots) as well as many inflammatory processes. During open heart surgery, the anticoagulant heparin is administered to avoid blood clots within the plastic tubing and oxygenator of the heart lung machine. At the conclusion of the surgery, heparin is neutralized with protamine. However, coagulation remains abnormal for several hours and bleeding can be problematic. The problem is particularly challenging in newborn children and infants undergoing complex reconstructive surgery for congenital heart problems. Not only are there long suture lines and fragile tissue but in addition, the coagulation system is frequently immature and platelet function is also suboptimal. Bleeding can also be problematic in elderly adults with serious aortic problems such as a dissection of the ascending aorta or aortic arch.
Aprotinin and bleeding after cardiac surgery
Aprotinin was first introduced in 1963 for the treatment of acute pancreatitis. In the 1980s, the potential application of aprotinin during cardiac surgery was investigated in London for its known anti-inflammatory activity. 3 (A whole body inflammatory response can be a complication of exposure of a patient’s blood to the foreign surfaces of the heart lung circuit.) However, the most obvious beneficial effect of aprotinin in these initial studies was a significant reduction in postoperative bleeding. Early reports as well as many subsequent prospective trials consistently demonstrated that aprotinin reduces blood loss and need for blood transfusion by at least 30%. 4 Subsequent investigation of the underlying mechanisms by which aprotinin reduces blood loss found that multiple complex pathways were involved. In addition to inhibition of fibrinolysis through direct inhibition of plasmin activity, aprotinin was also demonstrated to inhibit plasminogen activator. It also has complex effects on platelet function.5–7
Risk of stroke during cardiac surgery
The fact that aprotinin increases blood clotting led to concern that it might result in undesirable clots, including clots in the brain that can result in a stroke. Although one influential early report suggested that aprotinin might be associated with an increased risk of stroke in patients undergoing circulatory arrest (complete cessation of the circulation usually for aortic surgery) as well as cardiopulmonary bypass, 8 subsequent studies suggested that aprotinin was actually associated with a significant reduction in the risk of stroke during cardiac surgery. Sedrakyan 9 from Yale reviewed a total of 35 trials of aprotinin involving 3887 patients and found a 47% reduction in risk. The patients in this meta-analysis were all adult patients. Laboratory studies at Children’s Hospital Boston demonstrated that aprotinin protects the lining of blood vessels, which have been exposed to ischemia. 10 Others also demonstrated that aprotinin inhibits interaction between white cells and the endothelial lining of blood vessels, thereby reducing the inflammation resulting from the heart lung machine. 11 Aprotinin also has the potential to inhibit neuronal cell death in the brain mediated by natural tissue plasminogen activators. 12
Aprotinin and the NEJM
In January 2006, Dr Dennis Mangano, an anesthesiologist and math PhD who heads a private research institute (‘The Ischemia Research and Education Foundation’) based in California, published a highly influential paper in the NEJM, which emphasized the disadvantages of aprotinin for adult cardiac surgery. 1 The Mangano article suggested that aprotinin was associated with a number of dangerous side effects including strokes, cardiac and renal effects, and increased mortality. The conclusion was not based on a randomized clinical trial but rather on statistical analysis of a large international database. The study included multiple patients from Brazil and Germany, where the mortality rates were substantially higher than in most centers in the US. The analysis involved propensity scoring, a method of statistical analysis which was developed in the mid-1980s to improve the validity of data extracted from observational databases. Propensity scoring, however, requires considerable subjective input by the data analyzers. It is essential that all confounding variables are included in the analysis. Harvard statistician Donald Rubin, who developed propensity scoring has gone on record to state: ‘The analyses by Mangano et al are a misapplication of propensity score methods and should not be the basis of any conclusions regarding the use of aprotinin. I am surprised the study was considered publishable by any journal. My criticisms are numerous.’ 13
Specific criticisms of the Mangano paper have been detailed by Charles R. Morris in his book entitled ‘The Surgeons’. 14 Most importantly, the abstract for the New England Journal article includes the conclusion that ‘aprotinin was associated with a doubling of the risk of renal failure requiring dialysis among patients undergoing complex coronary artery surgery (odds ratio 2.59)’. Morris points out that this was arguably the most damning charge against aprotinin and the one most widely picked up by the lay press (and multiple medical malpractice lawyers). The actual article, however, simply showed that patients receiving aprotinin were 2.59 times more likely to suffer a renal event which was defined as either renal failure requiring dialysis or an increase in serum creatine levels. The latter is often transitory and not necessarily an important event. As Morris states that ‘the idea of so prestigious a journal as the NEJM in a piece calculated for maximum impact would let such a howler slip by is incomprehensible’.
Second, a large group of patients from the database with a very high mortality rate were excluded from the analysis. No explanation was given as to why these patients were excluded. Finally, although the propensity scoring method should supposedly have achieved a balance of risk factors among patients receiving versus not receiving aprotinin, the reality is that the aprotinin group had a much higher incidence of risk factors, suggesting that there was not fair matching between those who did and did not receive aprotinin.
The Mangano article was received with widespread skepticism by practicing cardiac surgeons. Numerous letters to the editor at NEJM resulted and are typified by the letter by Ferraris et al. which stated ‘we believe that the study by Mangano et al overemphasizes the risk of aprotinin for the prevention of bleeding during cardiac surgery without a demonstration of its substantial benefits’. 15 What was also not mentioned in the Mangano article is the fact that multiple previous randomized controlled trials had consistently documented safety and efficacy of aprotinin. Specifically for children, there were 12 randomized controlled trials to that time which had enrolled 626 children. These trials were reviewed by Arnold et al. in a review in Anesthesia and Analgesia published in 2006. Arnold concluded that ‘aprotinin reduced the proportion of children who received red blood cell or whole blood transfusion during cardiac surgery by 33% (relative risk = 0.67)’. Neither thrombotic nor allergic complications of aprotinin were observed in any pediatric trial and the frequency of adverse events was similar between groups in all studies but one. 16
Mangano refuses to allow FDA unsupervised access to his raw data
Subsequent to the controversy raised by the Mangano article, the FDA requested that Mangano submit his data for analysis by a panel of FDA statisticians. At a September 2006 public hearing, Mangano stated that he would allow FDA staff to look at his data but only on his computer and while he was physically present. 14 According to Morris, he categorically refused to give them the data so that they could analyze it themselves. The FDA stated ‘we were offered limited access to the data. There was this qualifying expectation that our examination be chaperoned (or supervised if you will…). Our statisticians felt uncomfortable having supervised access to the data in the sense that we would not have the ability to explore the data and at least verify the mathematics and the statistical aspects’. With this statement in mind, it is difficult to understand how an NEJM editorial published November 23, 2006 could state that Mangano had ‘offered the FDA unrestricted access to the data but the offer was not accepted’.
Morris and others have speculated that the unusual treatment of the Mangano article by the NEJM can be attributed to the difficulties that the journal and its editor Jeffrey Drazen were facing at the time of review of the Mangano article. This coincided with publicity in the Wall Street Journal which revealed that the NEJM had been informed of data discrepancies in an article published earlier in the NEJM which promoted the use of Vioxx. The Vioxx paper was withdrawn in March 2006 when it became clear that data on cardiac events that later led to the drug’s withdrawal had been withheld in the paper. Criticism of editor Drazen who had previously disclosed financial relations with 21 pharmaceutical companies between 1994 and 2000 was extreme. The Journal of the Royal Society of Medicine stated that the episode was more proof ‘that medical journals are an extension of the marketing arm of pharmaceutical companies.’ 17
Subsequent to the publication of the Mangano paper, the FDA convened a panel which recommended that aprotinin should remain available. However, the prestige of the NEJM led malpractice lawyers to begin advertising widely for patients who had suffered any complications after surgery to investigate whether aprotinin had been used and whether this might be the cause of their complications. In part, because of the review by the FDA and presumably related to their medicolegal exposure, the Bayer Corporation began to limit the availability of aprotinin.
Impact of aprotinin withdrawal on pediatric heart surgery
Complete withdrawal of aprotinin from cardiac surgery operating rooms and intensive care units in May 2008 has led to an expected increase in use of blood and blood products with their attendant risks. 18 Abrupt withdrawal of the agent has allowed historical comparisons to be made between populations of patients who underwent surgery in the aprotinin era with those in the post-aprotinin era.19,20 No study of this nature has documented any significant reduction in the incidence of postoperative complications in a pediatric patient population particularly stroke, renal failure, and death as highlighted by the Mangano article. Klugman et al. 19 found that patients in the post-aprotinin era left hospital with a significantly lower hematocrit despite a trend towards greater transfusion of red cells. A large retrospective study of pediatric cardiac patients undertaken before aprotinin withdrawal found significantly greater blood product usage without aprotinin 21 consistent with the results of the many prospective pediatric cardiac trials. Studies in the adult cardiac surgery literature following aprotinin withdrawal have also failed to document any reduction in the incidence of the relevant postoperative complications. In fact, a recent report at the American Heart Association showed a consistent trend of fewer complications in aprotinin-treated patients relative to those managed without aprotinin in a more recent time frame. 22
Appeals have been made to the FDA and the Bayer Corporation by individual surgeons as well as the Congenital Heart Surgeons Society of North America. Dr Pam Cyrus, Head of the Hematology Division of the Bayer Corporation, explained to the author that Bayer was not prepared to establish a pediatric indication for aprotinin use in spite of the fact that the Mangano article was based on surgery in adults undergoing coronary artery surgery. The ‘BART’ trial which has also been cited as evidence of a higher risk of complications with aprotinin relative to alternative antifibrinolytic agents such as tranexamic acid and epsilon amino caproic acid was also a trial restricted to elderly high-risk patients undergoing coronary artery surgery. 23 Another recent report at the American Heart Association meeting found that the use of tranexamic acid as an alternative to aprotinin had been accompanied by a highly significant increase in the risk of seizures after cardiac surgery. Tranexamic acid is known to have a risk of provoking seizures. 24
In countries other than the US, the use of aprotinin has re-emerged. As numerous reports have appeared confirming that withholding of aprotinin has been associated with increased exposure to blood products with their attendant risks, countries such as India have allowed aprotinin to be distributed by companies other than the Bayer Corporation. In fact, aprotinin is used in the majority of Indian pediatric heart surgery centers today for the majority of operations. 25
The future of aprotinin in the US
As more evidence is accumulated from clinical experience in the post-aprotinin era, it has become apparent that rather than reducing the risks of heart surgery in children, withdrawal of aprotinin has increased the risk of bleeding. The time is right for another trial of aprotinin in pediatric cardiac surgery. The Pediatric Exclusivity Provision should provide incentive to conduct such a trial. Furthermore, since aprotinin is no longer covered by patent, it is also to be hoped that alternative suppliers will emerge in the US who will have the courage to distribute this important agent, which has been demonstrated in at least previous 12 randomized trials in children as well as multiple non-randomized trials to reduce the risk of complications and exposure to blood products after heart surgery in children.
Footnotes
Author’s Biographies
Dr Richard A Jonas, MD, joined the staff of Children's National Medical Center in Washington DC in September 2004 as the chief of Cardiovascular Surgery and Co-director of the Children's National Heart Institute. He undertook his surgical training in Melbourne Australia, Auckland New Zealand and in Boston at the Brigham and Women's and Children's Hospitals. In 1994 Dr. Jonas was appointed to the William E. Ladd Chair of Surgery at Harvard Medical School and became the Cardiovascular Surgeon-in-chief at Children's Hospital Boston. He has an active clinical practice in congenital cardiac surgery and maintains his own NIH-supported laboratory and clinical research programs. He is the author of over 300 peer-reviewed publications and has authored and edited five textbooks. He is a consultant to the FDA and a regular member of the CICS study group of the NIH. Dr. Jonas was President of the American Association for Thoracic Surgery 2005-6 and of the Congenital Heart Surgeons Society of North America 2009-10. He currently has 2 active RO1 grants from the NHLBI, one of which is entitled “Neuroprotection with Serpins during cardiac surgery”.