FDA Faulted for Approving Studies of Artificial Blood

By Rob SteinWashington Post Staff Writer
Tuesday, April 29, 2008; A02

A new analysis concludes that the Food and Drug Administration approved experiments with artificial blood substitutes even after studies showed that the controversial products posed a clear risk of causing heart attacks and death.

The review of combined data from more than 3,711 patients who participated in 16 studies testing five different types of artificial blood, released yesterday, found that the products nearly tripled the risk of heart attacks and boosted the chances of dying by 30 percent.

Based on the findings, the researchers questioned why the FDA allowed additional testing of the products to go forward and why the agency is considering letting yet another study proceed.

"It's hard to understand," said Charles Natanson, a senior investigator at the National Institutes of Health who led the analysis, which was released early by the Journal of the American Medical Association so the data could be presented at an FDA meeting on the subject. "They already had data that these products could cause heart attacks and evidence that they could kill."

An FDA official defended the agency, saying it had carefully weighed the risks and benefits of each study individually and had convened this week's two-day meeting to address the very concerns raised by the analysis.

"FDA independently was aware of essentially the same concerns that have been raised, and indeed that is the reason we have convened this scientific workshop and is the reason why we have made careful decisions about allowing some studies to proceed and others not to proceed," said Jay S. Epstein, director of the Office of Blood Research and Review. "Our point of view is that FDA has been highly vigilant in its oversight."

An artificial blood substitute that has a long shelf life and does not need refrigeration could save untold lives by providing an alternative to trauma patients in emergencies, especially in rural areas and in combat settings.

But attempts to develop such products have been marred by repeated failures and fraught with controversy, in part because some products have been studied under rules allowing researchers to administer them without obtaining consent from individual patients. Such trials were permitted based on the argument that there was no alternative because trauma patients are often unconscious and time is often too limited to obtain consent from a family member.

Natanson conducted the analysis after becoming concerned about the consistent risks emerging from studies of various versions of products known as hemoglobin-based blood substitutes. After the Washington-based consumer group Public Citizen sued the FDA to gain access to data submitted to the agency, Natanson and colleagues at NIH and Public Citizen pooled data from studies conducted between 1998 and 2007.

"It didn't matter what type of patient you studied. There was no one product that was responsible for this. It was similar regardless of the patient population studied, the company that manufactured the product, whether the study was published or unpublished, or the chemical characteristics of the individual products," Natanson said. "The effect was robust."

Based on the available data, Natanson and his colleagues said, the FDA could have been aware of the risk as early as 2000.

"Since this time they did five more trials," he said, including a 2004 study involving 714 patients in which 11 patients receiving an artificial blood had heart attacks and 47 patients died.

The earlier findings should have been disclosed so that doctors at hospitals considering whether to participate in the studies would be better informed about the potential risks.

"Keeping data from being public represents real risks to patients," Natanson said. "If secret science is allowed, other companies can't build on the successes and failures and [outside reviewers] won't be able to fully assess the risk."

But Epstein said the FDA did block some studies from proceeding and allowed them to proceed only when officials were satisfied that the potential benefits outweighed the risks.

"We have viewed each product in its own right. We have needed to consider the extent to which different products and different clinical circumstances warranted an independent assessment of the relative risks and benefits. We have done that in every case," Epstein said.

Although none of the products have been approved in the United States, at least one has been approved in South Africa. Five studies are ongoing in eight other countries, and the FDA is considering a request by the Navy to conduct another study of Hemopure, an artificial blood product made by Biopure Corp. of Cambridge, Mass., on 334 trauma patients.

Biopure condemned the analysis as fundamentally flawed.

"There are vast differences among these products that make any pooling of data flawed, especially across different clinical experiences," A.G. Greenburg, the company's vice president of medical affairs, said in an e-mail. "Moreover, the analysis of Biopure's experience, based on pooling of heterogeneous trials, we believe to be significantly flawed as it fails to meet the homogeneity criteria of meta-analysis, thus invalidating the conclusions."

But a former Biopure official said yesterday that he agreed with the analysis.

"The risk appears to be a class risk. It appears to be present for all products," said William D. Hoffman, director of the cardiac intensive-care unit at Massachusetts General Hospital in Boston, who was Biopure's medical director and chief medical officer from 1998 to 2000. "They should all be on hold until they figure out what is causing the toxicity."

Hoffman said he tried to get the company to halt an earlier study when he became concerned about the product's safety.

"I went to the leadership at the company at the time and was outvoted," he said.

View all comments that have been posted about this article.

© 2008 The Washington Post Company