Each time health care workers grab a pint of blood for an emergency transfusion, they make sure the donor and recipient have compatible blood types. But they do not pay attention to the donor’s sex. A new study raises questions as to whether that should change.

In the first large study to look at how blood transfusions from previously pregnant women affect recipients’ health, researchers discovered men under 50 were 1.5 times more likely to die in the three years following a transfusion if they received a red blood cell transfusion from a woman donor who had ever been pregnant. This amounts to a 2 percent increase in overall mortality each year. Female recipients, however, did not appear to face an elevated risk. The study of more than 42,000 transfusion patients in the Netherlands was published Tuesday in JAMA The Journal of the American Medical Association.

The American Red Cross and the researchers themselves were quick to say the study is not definitive enough to change the current practice of matching red blood cell donors to recipients. But if this explosive finding is confirmed with future studies, it could transform the way blood is matched—and it would suggest millions of transfusion patients worldwide have died prematurely. “If this turns out to be the truth, it’s both biologically interesting and extremely clinically relevant,” says Gustaf Edgren, an expert who was not involved in the study but co-wrote an editorial about it. “We certainly need to find out what’s going on.” Edgren, an associate professor of epidemiology at the Karolinska Institute and a hematologist at Karolinska University Hospital in Stockholm, says his own research suggests the donor’s sex makes no difference to the transfused patient. “Our data is really not compatible with this finding,” he says.

But the new study is the fourth work—including a pilot study by the same authors—to find differences in the survival rates of blood transfusion recipients associated with sex mismatches. And the findings hint that potential problems extend beyond the question of whether female donors have ever been pregnant. One of the studies suggested women were at a disadvantage when they received male blood, and that the opposite was true as well.

Moreover, the three teams were from different countries, used different data sets and all had slightly different findings. The direction of each of their results, however, was the same: biological sex matters, says Henrik Bjursten, a professor in the department of cardiothoracic surgery, anesthesia and intensive care at Lund University/Skane University Hospital in Stockholm. Bjursten, who helped lead the study that found male-to-female transfusions were also problematic, was not an author on the latest JAMA work.

To definitively prove there is a problem, Bjursten says scientists would have to find a plausible biological mechanism to explain these differences—and then run two randomized controlled trials designed to look at whether the donor’s sex and pregnancy history affect the recipient. Still, the Dutch study raises enough red flags that he would like to see transfused red blood cells matched male-male and female-female now, even before a connection can be confirmed. “My personal opinion is yes…I would want to have it sex-matched,” Bjursten says, adding it would not be difficult to implement such a change. “There are millions of lives at risk here. Do we want to take the risk or do we want to go the safe route and try to avoid the harm?”

Bjursten’s own research found risk to both male and female cardiac surgery patients who received blood from someone of the opposite sex. It suggests gender-mixed transfusions may, on average, take about a year off a patient’s life. With 100 million transfusions per year worldwide, if 10 million to 40 million of those cause harm, he says, “the numbers start adding up.”

It may be difficult to arrive at a solid conclusion about whether sex matters in red blood cell donations. The ethics of running a randomized trial, in which some patients receive sex-mismatched blood products, may also be questionable now that so many doubts have been raised, Bjursten notes. But finding definitive answers without such trials will be tough. Existing data sets, like the one used by the Dutch group, often have holes. “It’s not clear that currently available databases will ever be able to answer this,” says Ritchard Cable, a scientific director with the American Red Cross, who co-wrote the editorial with Edgren and is hoping to compile a reliable database. Researchers in France are also planning a follow-up study, says Maxime Desmarets, a public health physician and epidemiologist at the University of Franche-Comté in France, whose own research suggests no gender difference in blood transfusions.

Desmarets and Cable, along with the American Red Cross, say current research does not justify a change in the way red blood cell donors are matched with patients. The study “needs confirmation as conflicting studies also exist,” Mary O’Neill, interim chief medical officer of the American Red Cross, said in a prepared statement. “As further research is required, we do not anticipate a change to the standard blood donation criteria or current conservative transfusion practices at this time. The Red Cross will closely examine subsequent studies on this subject to ensure the ongoing safety and availability of the blood supply.”

Scientists speculate women who have been pregnant could have some immune factor in their red blood cells that causes more rejection among younger male recipients. The main theory is that perhaps women who had sons developed antibodies to proteins in the Y chromosome of male DNA, as an immune reaction to their pregnancies. But that is a hypothesis the new study could not test, because the researchers did not have information about the sex of the women’s offspring. It is also possible the male and female immune systems are fundamentally different in some way or the men are reacting to sex differences in RNA found in the women’s blood, Bjursten says.

Until a smaller study on sex mismatches by the same Dutch team six years ago, no one had thought to look at the pregnancy history of red blood cell donors, says Rutger Middelburg, an epidemiologist with Sanquin Research in the Netherlands, who helped lead that pilot work and the study published Tuesday. The differences in mortality are difficult to detect unless researchers know what to look for, Middelburg wrote via e-mail. “Even now, we find that in our data set simply looking at all patients can dilute the effect to the level where it becomes undetectable,” he added. “We had to specifically look at the right patient group.” He does not know why the team saw a survival difference only in younger men.

It is conceivable, he says, that younger men might have different diseases triggering their need for a transfusion than older men, which might make them more vulnerable to problems incorporating women’s red blood cells.

The data was not perfect, Middelburg says. The team examined records of patients who had received transfusions years earlier, and the researchers did not know the pregnancy status of all the women donors. The researchers disqualified data from patients who received blood from both men and women—and because men can donate blood more often than women, the pool was already skewed male, he notes. Women who had been pregnant at any point accounted for just 6 percent of the donors the team studied, although the association was still statistically valid. “We are very confident of our results," Middelburg says.

He is continuing his research, and now hopes to get additional funding. “My priorities would be [to] look into more detailed pregnancy histories and causes of death,” he says, “but much other relevant research could still be done with sufficient resources.”