In the fall of 2021, Gabriel Arias felt like his body was “rotting from the inside.” He was diagnosed with acute myeloid leukemia, a form of blood cancer so aggressive that doctors had him hospitalized the day of his biopsy. In cases like his, the ideal treatment is a transplant. Arias’s cancer-prone blood cells needed to be destroyed and replaced with healthy ones taken from the bone marrow or blood of a donor who matched him biologically. Fortunately, doctors found him a match in the volunteer-donor registries—a man in Poland. Unfortunately, Arias’s single match in the entire world was no longer available to donate.

In the past, the road to transplant might have ended here, but a medical advance had dramatically expanded the pool of donors for patients such as Arias. With the right drug, Arias could now get a transplant from his brother, a partial match, or, as he ultimately chose, he could join a clinical trial in which his donor would be a stranger who shared just eight of 10 markers used in bone-marrow transplants. Under this looser standard, Arias’s registry matches multiplied from one to more than 200. “It really is a game changer,” says Steve Devine, the chief medical officer of NMDP, a nonprofit that runs a donor registry. Today, agonizing searches for a matched donor are largely a thing of the past.

The drug powering this breakthrough is actually very old. Cyclophosphamide was first developed in the 1950s for chemotherapy. Fifty years later, researchers at Johns Hopkins began studying whether it could be repurposed to prevent a common and sometimes deadly complication of bone-marrow transplants called graft-versus-host disease, where the donor’s white blood cells—which form the recipient’s new immune system—attack the rest of the body as foreign. The bigger the mismatch between donor and recipient, the more likely this was to happen. Cyclophosphamide worked stunningly well against graft-versus-host disease: The drug cut rates of acute and severe complications by upwards of 80 percent.

Cyclophosphamide has now enabled more patients than ever to get bone-marrow transplants —more than 7,000 last year, according to NMDP. (Bone-marrow transplant is still used as an umbrella term, though many of these procedures now use cells collected from the blood rather than bone marrow, which can be done without surgery. Both versions are also known, more accurately, as hematopoietic or blood stem-cell transplants.) The field has essentially surmounted the problem of matching donors, a major barrier to transplants, Ephraim Fuchs, an oncologist at Johns Hopkins University, told me. Fuchs couldn’t remember the last time a patient failed to get a blood stem-cell transplant because they couldn’t find a donor.

It wasn’t obvious that cyclophosphamide would work so well. “I’m just going to come clean,” Devine told me. “Back in 2003 and 2005, I thought it was crazy.” Derived from a relative of mustard gas, the drug is known to be highly toxic to a variety of blood cells; in fact, doctors had long used it to kill the diseased bone marrow in patients before transplant. Why would you want to give such a drug after transplant, when the new donor cells are still precious and few? It defied a certain logic.

But as far back as the 1960s, researchers also noticed that high doses of post-transplant cyclophosphamide could prevent graft-versus-host disease in mice, even if they did not know why. Over the next few decades, scientists working away in labs learned that cyclophosphamide isn’t quite carpet-bombing the blood. It actually spares the stem cells most important to successful transplant. (Blood stem cells differentiate into all the types of red and white blood cells that a patient will need.) Why cyclophosphamide works so well against graft-versus-host disease is still unclear, but the drug also seems to selectively kill white blood cells active in the disease while sparing those that quell the immune system.

By the late ’90s, doctors saw a clear need to expand the search for donors. Bone-marrow transplants are most successful when donor and recipient share the same markers, known as HLA, which are protein tags our cells use to distinguish self from nonself. We inherit HLA markers from our parents, so siblings have about a one-in-four chance of being perfectly matched. As families got smaller in the 20th century, though, the likelihood of a sibling match fell. Donor registries such as NMDP were created to fill the gap, however imperfectly.

Doctors soon began coalescing around the idea of using family members who were only haploidentical, or half matched, meaning they shared at least five out of 10 HLA markers. Every child is a half match to their parents, and every parent to their child; siblings also have a 50 percent chance of being half matches. But when doctors first tried these transplants, the “outcomes were horrible,” Leo Luznik, an oncologist at Johns Hopkins, told me. Patients had frighteningly high rates of graft-versus-host disease, and more than half died within three years.

Based on the lab findings, Luznik, Fuchs, and other colleagues at Johns Hopkins wondered if post-transplant cyclophosphamide could help. The pharmaceutical companies that made it were uninterested in funding any research, Luznik said, because “it was an old, very cheap drug.” With government grants, however, the team was able to prove that cyclophosphamide got the rate of graft-versus-disease as low as in matched sibling transplants. By the late 2000s, transplants with half-matched family members were becoming routine.

Still, not every patient will have a sibling or parent or child who can donate. Doctors began wondering if cyclophosphamide could work for unrelated donors too. If only eight of the 10 markers have to be matched, then almost everyone would find a donor, even multiple donors. This was especially important for patients of mixed or non-European ancestry, who have a harder time finding unrelated donors, because people of those backgrounds make up a smaller proportion of registry donors and because they can carry a more diverse set of HLA markers. Two-thirds of white people can find a fully matched registry donor, but that number drops to 23 percent for Black Americans and 41 percent for Asians or Pacific Islanders.

Amelia Johnson, who is half Indian and half Black, was one of the first children to get a transplant from a mismatched unrelated donor in a clinical trial in 2022. Her mom, Salome Sookdieopersad, remembers being told, “You guys need to start recruiting bone-marrow donors to help increase your chances.” When that still didn’t turn up an ideal match, Sookdieopersad prepared to donate to her daughter as a half match. But then Amelia was offered a spot in the clinical trial, and they decided to take it. Transplants with mismatched unrelated donors had already been tried in adults—that was Arias’s trial—and they offered other potential benefits. A younger donor, for example, has younger cells, which fare noticeably better than older ones. Amelia did end up with a bout of graft-versus-host disease; cyclophosphamide lowers the risk but not to zero. Still, the transplant was necessary to save her life, and her mom pointed out that some risk was unavoidable, no matter the type of donor: A friend of Amelia’s got graft-versus-host even with a perfectly matched one. Doctors were able to treat Amelia’s complications, and she returned to school last August. The pediatric trial she was part of is ongoing.

In adults, where more data are available, doctors are already moving ahead with mismatched, unrelated donors. Between this and half-matched family members, patients who once might have had zero donors are now finding themselves with multiple possibilities. Doctors can be choosier too: They can select the youngest donor, for example, or match on characteristics such as blood type. The larger pool of donors also prevents situations like Arias’s, in which a single matched donor who signed up years ago is no longer available, which happens with some regularity. Cyclophosphamide is now routinely used in matched transplants too, because it lowers the risk of graft-versus-host disease even further.

Arias’s mismatched unrelated donor in the trial was an anonymous 22-year-old man who lives somewhere in the United States. When Arias and I spoke last month, it had been almost exactly two years since his transplant. He’s cancer free. He and his wife just welcomed a baby girl. None of this would have likely been possible without the transplant, without the donor, without a 70-year-old drug that had been smartly repurposed.