A group of hospitals in Toronto has discovered principal investigators (P.I.s) among their ranks giving blood transfusions to patients in order to alter their lab values to better match inclusion criteria for clinical trials.

Jeannie Callum, director of transfusion medicine for Sunnybrook Health Sciences Center, said she and colleagues who run the hospital group’s transfusion medicine collaborative, which closely scrutinizes the comings and goings of blood products, noticed blood being ordered for cancer patients who didn’t meet the normal criteria for a transfusion. Over the course of a year, the collaborative discovered three P.I.s at three different facilities working on three different chemotherapy studies who ordered such transfusions.

“If it happens once, you think, OK, that’s just one physician,” said Callum. “When you see it at three different hospitals, though, this is a pattern.”

All of the patients had advanced cancer. The first one had an albumin level of 22 g/L but needed to be at 30 g/L to qualify for the trial and thus took a transfusion of 25% albumin. The second patient fell just under the required platelet count of 50,000 per cubic millimeter with a pre-transfusion count of 48,000. And the third patient’s doctor ordered transfusion of a unit of red cells to increase her hemoglobin level from 8.3 g/dL to the required 9.0 g/dL to be eligible for the trial.

Callum said upon making the discovery, her department immediately called each physician, and then Canada’s Research Ethics Board. When she began to see a pattern, she contacted colleagues in other parts of Canada and the U.S. to ask if they’d ever seen such a practice. They told her they hadn’t, but that their hospitals don’t closely oversee and track blood products like her group does. Transfusion medicine collaboratives like Sunnybrook are unusual.

Callum said she realized this could be a widespread practice. To get the word out, she and her Sunnybrook colleagues wrote a letter to the editor of the New England Journal of Medicine. It appeared in the Oct. 21 issue, warning of the practice.

“I don’t think this is rare, and I don’t think it’s unique to us,” she said.

Callum said the physicians ordering the transfusions believe they’re doing the right thing for their patients, who have cancer and have run out of treatment options. And the patients are fully aware
of the reason they received the transfusions. But the pharma companies sponsoring the trials were kept completely in the dark, as was the ethics board.

P.I.s feel they’re bumping up against eligibility criteria that has in recent years become narrow and seemingly arbitrary, said Callum, so they are beginning to “break rules behind the scenes” for their patients whose only hope is to get into an experimental chemotherapy trial.

“It’s an ethical dilemma,” she said.

Gretchen Miller Bowker, chief operating officer of Indianapolis-based Pearl IRB, said it is an egregious offense. “The principal investigator manipulating blood levels to allow patients to enter clinical research studies is actually putting the patients at risk, as those criteria are generally set based on safety,” she said. “So if they’re disregarding safety measures that the sponsor has put in place, then the subject is at increased risk—and the P.I.’s actions have the potential to jeopardize the entire trial, which then makes the drug unavailable for everyone.”

The answer, said Callum, lies in closer examination of inclusion and exclusion criteria in protocols, and closer oversight of investigators.

“We need more dialogue on this,” she said. “And when we set inclusion and exclusion criteria, we need to make sure we are setting them because there are medical reasons and safety reasons for them. We also need to make sure we believe in them, and pledge to follow them. I think that responsibility falls with all parties—the pharmaceutical company, the P.I. and the research ethics board. We also need to make sure there is very close oversight of trials, and that everyone is aware of the possibility that physicians will do this, so everyone keeps a high level of suspicion.”

When done surreptitiously, the downsides are numerous. A patient’s blood levels mirror his organ function, Callum explained. If the numbers are not in a range deemed stable enough to take the drug, receiving the experimental agent could make them sicker. For the drug sponsor, this could throw off the true side effect profile of the drug. And for someone who is sicker than he lets on, a drug’s intended benefit may not have as much effect, making the agent seem less helpful than it really is. Additionally, there are the general risks of transfusion.

In some cases, though, a blood infusion to meet eligibility criteria may not have any effect and may be acceptable, she said, urging drug developers to go the extra step to clarify that in protocols.

While Sunnyvale is unique in its close scrutiny and tracking of blood products within its halls, more hospitals are instituting such haemovigilance programs.

 —Suz Redfearn