Blood substitutes significantly increase the chances of heart attacks and death in patients, yet Canadian and U.S. regulators gave the green light to testing on humans, says a paper published yesterday.

A review of all published literature on human blood-substitute trials in the Journal of the American Medical Association found that people who received hemoglobin-based blood substitutes during a trial had a 30 per cent greater chance of death, and were three times more likely to suffer a heart attack.

"I became extremely concerned by the consistent pattern of increased risk that patients were being placed in," said Dr. Charles Natanson of the National Institutes of Health, who led the review.

His researchers analyzed 16 randomized controlled studies of five different blood substitutes involving 3,700 patients. Among those receiving a blood substitute, they found 164 deaths. Among the same number of patients who instead received ordinary blood products or saline transfusions, 123 died.

The risk of heart attack was nearly tripled in the blood-substitute groups, with 59 recorded compared with 16 in the control groups.

Experts speculate that hemoglobin in the blood substitutes is a scavenger of nitric oxide from the bloodstream, causing blood vessels to constrict and sticky platelets to build up – adding to the risk of a heart attack.

For years, a reliable replacement for blood has been sought as a medical breakthrough, and source of huge revenues for the producers. The potential is obvious to save battlefield lives, as one example. Ersatz blood could theoretically be stored for years without refrigeration, apply for all blood types and be shielded from infections such as hepatitis or the AIDS virus.

The downside risk, however, was often not fully disclosed either to test patients or to ethics boards approving such trials, Natanson said.

After a risk is known, "the next person should never have to undergo that same risk," said Natanson. "That risk should be made public."

In the U.S., a company that runs a clinical trial must give the results to the U.S Food and Drug Administration but is not required to publish the findings – even if there's a "toxic" outcome. In many cases, Natanson said, trials were approved by the FDA, with knowledge that there was a potential risk involved.

Natanson knew of only one clinical trial in Canada, which involved testing of the product Hemolink by Hemosol Corp., a Mississauga company that has been in bankruptcy proceedings since late 2005.

With files from The Associated Press and The Canadian Press