People treated with blood substitutes in clinical trials over the past two decades were 30% more likely to die and had almost a threefold higher rate of myocardial infarction (MI) than patients in control groups, a new meta-analysis suggests [1]. Even more damning, the authors of the study say that the US Food and Drug Administration (FDA) had ample evidence eight years ago to put a stop to ongoing studies of hemoglobin-based blood substitutes (HBBS) — some of which were used in cardiac surgery trials — but did not.
"Prompt meta-analyses of the HBBS trials by the FDA most likely would have demonstrated significant risks by 2000," Dr Charles Natanson (National Institutes of Health [NIH], Bethesda, MD) and colleagues write in paper released online in the April 28, 2008 issue of the Journal of the American Medical Association. "Had the agency placed a moratorium on trials at that point, product-related deaths and MIs in subsequent trials most likely would have been prevented," the authors write.
In a telebriefing with the press, however, director of the FDA's Office of Blood Research and Review, Dr Jay Epstein, staunchly defended the agency. According to Epstein, the FDA has always acted in the best interests of the public and was, in fact, privy to information that the authors of the meta-analysis have not had access to, which was reassuring enough to the agency to permit certain — but not all — trials of HBBS to continue in recent years. What's more, says Epstein, the FDA had independently identified concerns with HBBS and, even before Natanson et al's meta-analysis was published, had scheduled a workshop April 29-30 to discuss safety of HBBS and their further study.
Stopping studies could have saved lives
Natanson et al's study included 16 trials of five blood substitutes being tested in the setting of trauma, vascular surgery, cardiac surgery/coronary artery bypass grafting (CABG), other surgery, shock, or acute ischemic stroke. Data for the analysis came from literature searches, press releases (in the case of two trials that have never been published) and from one FDA review. The authors found a statistically significant increase in death (relative risk 1.30, 95% CI 1.05 – 1.61) and MI (relative risk 2.7, 95% CI 1.67 – 4.40), with no particular subgroup bearing the brunt of the risk.
To heartwire, Natanson added that while the analysis looked only at the end points of death and MI, he believes the toxicity of these products goes even further. "I'm concerned that there is a generalized toxicity, including renal failure, strokes, liver-function abnormalities, and pancreatic enzyme abnormalities," he said. Natanson was paid $10,000 in 2004 by one HBBS maker, Hemosol Inc, to review its study data and also sat on an FDA advisory committee review of unpublished data for another HBBS in development. While he says he cannot comment specifically on what he saw in either setting, his experience with different data sets was enough to make him worry that trials of similar products were continuing and results were not being made public in a timely fashion.
Existing legislation, he added, prevents the FDA from making public information deemed proprietary by the companies.
"Results of human clinical trials that put patients at risk and randomize them need to be fully and expeditiously disclosed to the scientific and medical communities," Natanson told heartwire. "There's great inefficiency and real risk to patients from the failure to report data promptly, because scientists and researchers at companies can't build on the results of others and [institutional review boards] IRBs will not be able to make full decisions. One of the biggest lessons from this is that regulations have to be changed, and we cannot have secret science if we're going to protect the public. The FDA and companies and their various rules and regulations need to change such that these are not trade secrets any more."
A great idea gone wrong
None of the HBBS included in the meta-analysis are available in the United States (US) for human use (although one is approved for use in dogs); one product is on the market in South Africa. The authors point out that signals of significant harm were seen as early as 1996, but five clinical trials are ongoing and at least one is in the planning stages worldwide.
"These trials are active and they're online, but I don't know how actively they're enrolling," Natanson said. While he is the first to agree that blood substitutes that don't require refrigeration or cross-matching for blood type and have a long shelf-life represent a massive, unmet medical need, he believes that companies need to go back to the drawing board and stop testing these products in humans.
"There's no question this is a great idea," he told heartwire. "Research should continue, but in the four-legged animals, not the two-legged animals."
In an editorial [2], Drs Dean A Fergusson and Lauralyn McIntyre (Ottawa Health Research Institute, ON) highlight the need for public reporting of positive and negative studies. Without this, ethical review boards for the studies included in Natanson et al's analysis, as well as the patients or proxy decision makers involved, had no access to critical information about safety from other studies to guide what in some cases were life-and-death decisions.
Epstein, for his part, insisted that the FDA is "required to follow current laws and regulations, [including] limitations on public disclosure of information that sponsors may provide to the agency." And while the FDA "fully supports" increased public access to clinical-trial information and timely publication of positive or negative clinical-trial results, Epstein is adamant that the information provided confidentially to the FDA is being put to use in a way that ensures patient safety, while leading, where possible, to new and urgently needed new products.
"Our point of view is that the FDA has been highly vigilant in its oversight of hemoglobin-based oxygen carriers and has carefully considered the potential benefits and risks of the product in our review of all proposed studies, taking into account the specific medical situation," Epstein told the press. "We can't concur with the statement in the paper that a prior meta-analysis would have changed the FDA's assessment or resulted in stopping of all studies in the year 2000. . . . Basically, even though aware of certain safety signals, our reviewers determined there were enough differences between products and their intended uses to support a careful weighing of individual clinical-trial proposals, only some of which were allowed to proceed."
Fergusson and McIntyre, however, conclude that on the basis of Natanson et al's study, along with preclinical studies pointing to potential toxicity, "further phase 3 trials of hemoglobin-based oxygen sources should not be conducted."
What's more, they continue: "Until the mechanisms and potential toxicities of hemoglobin-based oxygen-source products are better understood, patients cannot be placed at unacceptable risk. If and when phase 1 through 3 studies can be justified, initial studies should involve patient populations who could benefit the most — namely, trauma patients who require out-of-hospital resuscitation, where blood is not accessible or available."
The meta-analysis was coauthored by researchers from the NIH and Public Citizen (Washington, DC).
The Warren G. Magnuson Clinical Center of the NIH, Bethesda,Maryland, and Public Citizen, Washington, DC, provided support for this study. Dr. Natanson has disclosed receiving the 1-time payment of $10,000 from Hemosol. Dr. Fergusson also disclosed receiving a 1-time $500 honorarium for attending an advisory board meeting for Hemosol. The other study authors have disclosed no relevant financial relationships.
Sources
Natanson S, Kern SJ, Lurie P, et al. Cell-free hemoglobin-based blood substitutes and risk of myocardial infarction and death: A meta-analysis. JAMA. 2008;DOI:10.1001/jama.299.19.jrv80007. Available at: http://jama.ama-assn.org.
Fergusson DA, McIntyre L. the future of clinical trials evaluating blood substitutes. JAMA. 2008;DOI:10.1001/jama.299.19.jed80027. Available at: http://jama.ama-assn.org
The complete contents of Heartwire, a professional news service of WebMD, can be found at www.theheart.org, a Web site for cardiovascular healthcare professionals.
Clinical Context
Because HBBSs are infusible, oxygen-carrying liquids with long shelf lives and no need for refrigeration or cross-matching, they are well-suited for treatment of hemorrhagic shock in remote areas or military settings. Since 1996, however, some randomized controlled trials (RCTs) of HBBSs have shown lack of clinical benefit and increased risks.
Hemoglobin molecules used to manufacture HBBSs are not contained by a red cell membrane, and they rapidly scavenge nitric oxide when released into the vasculature, potentially causing systemic vasoconstriction, decreased blood flow, increased release of proinflammatory mediators and potent vasoconstrictors, and a loss of platelet inactivation. These conditions may cause vascular thrombosis of the heart or other organs. Despite these concerns, at least 1 HBBS is currently approved for use outside of the United States, and new clinical trials are underway or are being planned worldwide.
Study Highlights
The goal of this meta-analysis was to evaluate the safety of HBBSs in patients who have undergone surgery, have had a stroke, or sustained traumatic injuries.
The investigators searched PubMed, EMBASE, and the Cochrane Library for articles using hemoglobin and blood substitutes from 1980 through March 25, 2008.
They also reviewed FDA advisory committee meeting materials and conducted Internet searches for company press releases.
Inclusion criteria were RCTs of therapeutic use of HBBSs in patients at least 19 years of age.
Of 70 trials identified in database searches, 13 trials met inclusion criteria.
Data from 2 additional trials were reported in 2 news releases, and additional data were included in 1 relevant FDA review.
Data were extracted on death and MI as outcome measures.
The 16 included trials tested 5 different products and enrolled a total of 3711 patients in various clinical populations.
The control group varied from trial to trial (saline, red blood cells, or various plasma expanders).
A test for heterogeneity of the results of the included trials was not significant for either mortality or MI, or both. Data were therefore combined with use of a fixed-effects model.
HBBSs were associated with a statistically significant increase in the risk for death (164 deaths in the groups treated with HBBS and 123 deaths in the control groups; RR, 1.30; 95% CI, 1.05 – 1.61).
HBBSs were also associated with a statistically significant increase in the risk for MI (59 MIs in the groups treated with HBBS and 16 MIs in the control groups; RR, 2.71; 95% CI, 1.67 – 4.40).
Estimated number needed to harm was 62 patients treated for each treatment-related death.
Estimated number needed to harm was 50 patients treated for each treatment-related MI.
Subgroup analysis showed that the increased risks for death and MI were not restricted to a particular HBBS or clinical indication.
Based on their meta-analysis, the investigators concluded that use of HBBSs is associated with a significantly increased risk for death and MI.
They recommend a policy whereby any new or existing HBBSs should undergo preclinical studies in animal models that replicate the known toxicities of HBBSs in humans before additional clinical trials are permitted to proceed.
Limitations of this meta-analysis were as follows: details on some study protocols were unavailable, data for 1 trial based on a pooled FDA analysis used unclear methods, control group varied from trial to trial, and varied levels of blinding were used.
Pearls for Practice
In this meta-analysis, there was a statistically significant 30% increase in the risk for death in the groups treated with HBBS vs the control groups. Estimated number needed to harm was 62 patients treated for each treatment-related death.
The risk for MI was more than doubled in the groups treated with HBBS vs the control groups. Subgroup analysis showed that the increased risk was not restricted to a specific HBBS or clinical indication. Estimated number needed to harm was 50 patients treated for each treatment-related MI.
Review by Dr. Ramaz Mitaishvili
