The authors were critical of the FDA for not requiring the companies to publish their findings, and for allowing additional trials to be conducted after the risk should have been apparent.
"At some point, somebody should have realized that we've tried it in trauma patients, we've tried it in surgical patients, we've tried it in stroke patients, we've tried many different formulations and we keep finding the same result," said Dr. Charles Natanson, lead author of the meta-analysis, a technique in which data from a number of trials are combined and re-analyzed.
"At some point, and we sort of argue in the paper that may have been the year 2000 . . . it was time to put a halt" to additional trials, Natanson said.
Canadian patients were also exposed to the risks in clinical trials of a product called Hemolink, which was made by a now defunct company called Hemosol Inc. The company conducted at least three trials of the product, though data from only two were included in the JAMA analysis.
Natanson, who is an advocate for patient safety in clinical trials, was paid $10,000 to review the data from a third trial when Hemosol was trying to decide whether to push on with the research. The company later went into receivership.
The company never published the findings and Natanson cannot comment on what he saw or recommended.
An editorial accompanying the review points out further studies of these products are underway or planned in a number of countries around the globe.
The authors, from the Ottawa Health Research Institute, argue future work on blood substitute products should not be allowed in humans until it's clear from animal studies that the products are likely to be safer.
Lead author Dean Fergusson, a clinical trials expert, said the withholding of the negative results meant ethics boards and trial participants could not accurately weigh the risks and benefits of the research.
"How can patients or their decision makers make truly involved consent without all this information? I think that's a huge message," said Fergusson, who reported he was paid a one-time fee of $500 for attending a Hemosol advisory board meeting.
The lack of disclosure suggests company stock prices were placed at a higher priority than the safety of people being asked to go into clinical trials, experts suggest.
"They didn't think beyond the four walls of their corporation, I think, quite frankly," Fergusson said of the companies.
Lexchin was more critical of Health Canada for treating unpublished studies as confidential information. "They're calling this business information, not health information," he said.
Epstein said the FDA's own analysis of the accumulating experience with the blood substitutes led it to conclude there were safety risks associated with the entire class of drugs. "It's because of FDA's safety concerns that there currently are no ongoing studies of HBOCs in the United States trials and no approved HBOC products."
But he challenged the Natanson article's contention that the risk should have been apparent by 2000, saying the FDA had access to information Natanson and his colleagues have not seen that suggested there might be benefits in some situations.
"Basically even though aware of certain safety signals, our reviewers determined that 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," Epstein said.
