Bone marrow transplantation is at once the simplest and the trickiest of human organ transplants. Mechanically, it resembles an ordinary blood transfusion. Marrow, the clear blood-producing fluid within bones, is extracted from a donor's pelvis and infused through a vein into the patient's blood stream. The marrow finds its way to the cavities inside bones and begins making new blood cells within two or three weeks.

But if the transplant fails, death is almost certain, because a person cannot survive without healthy bone marrow. Marrow transplants are a last resort for people whose bone marrow has been damaged or destroyed by radiation or certain kinds of cancer.

The success of a marrow transplant depends on two critical factors: an exact tissue match between donor and recipient, and intensive follow-up care during the ensuing weeks, when the patient -- much like a person with AIDS -- is highly vulnerable to infection, rejection and hemorrhaging.

"It requires every resource of modern medicine to keep them alive," said Dr. Robert P. Gale, the California bone marrow expert who performed transplants this month in Moscow on 19 people exposed to high doses of radiation from the accident at the Chernobyl nuclear reactor. He described their care as "a series of very delicate balances."

Developed in the 1950s as a possible treatment for radiation poisoning, bone marrow transplantation has been used since then almost exclusively on patients with diseases of the bone marrow, such as leukemia, aplastic anemia and lymphoma.

The Chernobyl disaster, the worst nuclear accident in history, marked the first massive use of marrow transplants to treat radiation poisoning since the treatment was devised.

"We've been gearing up for this for the past several years, hoping we'd never have to use it," said Gale, associate professor of medicine at the University of California at Los Angeles. He is also chairman of the advisory committee of the International Bone Marrow Transplant Registry, which collects data on marrow transplants at 128 medical centers around the world.

Nearly 12,000 bone marrow transplants have been performed worldwide, about half of them in the past three years, according to registry officials. Three-year survival rates for patients range from 20 percent to 90 percent, depending on the type of disease treated and the degree of the tissue match between donor and recipient.

The biggest threat to a transplant patient is rejection. If the donor's tissue doesn't match the recipient's, it may be rejected as "foreign" by the body's immune system. Even more likely, the new marrow will reject the patient, a condition known as graft vs. host disease.

"Giving marrow to someone is very simple," said Dr. Dean Buckner, head of the supportive care program at the Fred Hutchinson Cancer Research Center in Seattle, the world's leading bone marrow transplant center. "The key is the supportive care afterwards."

Whether the Soviet medical facilities are equipped to provide marrow transplant patient with such intensive followup care, including massive transfusions, intravenous nutrition and antibiotics, "remains to be seen," Buckner said. An unspecified number of the 19 transplant recipients already have died.

The crucial tissue-matching of donor to recipient depends on a laboratory test for human leukocyte antigens, or HLA. Leukocytes are white blood cells. Antigens are proteins on the surfaces of cells which trigger the defenses of the body's immune system.

The HLA test determines a person's tissue type by the pattern of six different antigens. Thousands of combinations are possible.

HLA is like a cell's identification card, said Dr. Lyle Sensenbrenner, head of experimental hematology at the Johns Hopkins Oncology Center in Baltimore.

"The antigen is the ID card that each cell in your body carries around on its surface. It's the ID card that the body uses to recognize who is itself and who isn't." When it recognizes a "foreign" antigen, the body's immune system calls out antibodies to fight it. In a marrow transplant, if the "ID cards" on the transplanted marrow cells don't match the body's own cells, rejection usually occurs.

But just as a person can occasionally get by with using another's ID card, particularly if the appearance and birth date match, marrow cells sometimes can be transplanted into a patient with a closely matched tissue type. (In both situations, the chances of success are best when the two individuals are identical twins.)

The major advances since early experiments in marrow transplants in the 1950s have come in tissue-matching to lessen the risk of rejection. It was not until 1969, for example, that a medical team in Seattle performed the first successful adult marrow transplant involving a donor who was not an identical twin.

Almost all bone marrow transplants are between siblings, said Alfred Rimm, professor of biostatistics and epidemiology at the Medical College of Wisconsin in Milwaukee, where the international marrow transplant registry is based.

A patient has a one in four chance of matching a sibling's HLA type. But with the trend toward smaller families, only about 30 percent of transplant candidates find a matched donor in the family.

Recent efforts to transplant HLA-matched marrow from unrelated donors into patients who lack a sibling donor have been disappointing, Rimm said.

One reason for the lack of success may be the delay involved, said Dr. John Hansen, director of the tissue-typing lab at the Hutchinson Center in Seattle. By the time a matched donor is found outside the family, the patient often is too ill to survive even a well-matched transplant.

The Hutchinson center has sought out-of-family donors for 250 patients -- but found donors promptly for only 12, Hansen said.

No national registry of potential donors exists. Some blood centers keep lists of tissue-typed volunteers, but the largest computerized list of potential donors in the country -- in Milwaukee -- contains only about 10,000 names.

Out of a group of 20,000 potential marrow donors, statisticians say, the odds of finding of perfect HLA match for a typical patient are about 33 percent. In a group of 60,000, the odds rise to about 50 percent, and in a group of 100,000, to about 90 percent.

Leukemia patients who undergo bone marrow transplants are treated beforehand with high doses of controlled radiation to destroy diseased marrow and blood cells that are likely to reject a transplant. The Chernobyl victims, however, were gravely ill because of damage to their marrow from accidental radiation.

The first challenge for Gale's team of doctors in Moscow was to decide which patients would survive on their own, which were hopelessly ill and which needed immediate treatment. Of the 35 severely irradiated patients, 19 received marrow transplants.

The ideal candidates for bone marrow transplant were those who had received between 800 and 1,000 rads.

"Below that, you may not need a bone marrow transplant to survive," said Johns Hopkins' Sensenbrenner. "Above that, you're going to start getting other complications which are going to kill you -- mainly damage to the gut."

Since no one knew how much radiation had leaked out of the Chernobyl plant, doctors had to estimate the victims' exposures by examining the damage to particularly vulnerable tissue -- the inside of the mouth, the intestines and the marrow itself.

And time was short.

"If you don't get the patients early enough, the radiation wipes out the white blood cells, which are what you need to do the tissue typing," Sensenbrenner said. The white blood count can drop to zero in a severely irradiated person within as little as three days. Other ways of testing tissue types take too long to be used in a radiation victim.

In six of the transplants performed in Moscow, radiation poisoning was so severe that it destroyed all the white blood cells. That left doc- tors with no way of finding a matched donor.

For these patients, doctors obtained marrow from an alternative source -- liver tissue from human fetuses, which has fewer antigens and is less likely to cause rejection. Fetal liver tissue transplants are rare and have been used in the United States mainly on infants with a genetic condition known as severe combined immune deficiency syndrome, whose symptoms and prognosis are much like AIDS.

Marrow transplant experts around the country praised Gale's efforts to help the Soviets treat the Chernobyl victims and study the effects of severe accidental irradiation.

"Here was a unique opportunity to learn something -- something we've been worried about for many years," Sensenbrenner said. "The whole bone marrow transplant program got its start after World War II with an effort to find out how to protect people from lethal radiation."

Many scientists are disturbed by what Sensenbrenner called the "awful suppression" of information by the Soviets in the first days after the Chernobyl accident. The lack of candor, he said, could hamper efforts to learn what went wrong and how a recurrence can be prevented.

"The most important thing Bob Gale did for the world -- maybe even more than the lives he saved -- was to bring back firsthand information as to what really happened over there."