Allogeneic Bone Marrow Transplant

Courage is not the variable. Two numbers are, and both of them can be estimated before anybody agrees to anything.

Number one asks how likely the disease is to come back if you do not have a transplant. Number two counts the chance the transplant kills you for reasons that have nothing to do with the disease, and everything else, the donor search, the conditioning, the year of clinic visits, follows from which of those two numbers is larger. A meta analysis of 24 trials and 6,007 patients found that for one whole risk group the first number is already small enough that transplanting makes things worse.

Two numbers

Families hear the word transplant and reach for the language of bravery. Doctors reach for arithmetic. Arithmetic does better.

An allogeneic bone marrow transplant swaps a sick blood and immune system for a donated one. It cures diseases that no amount of chemotherapy cures, and it does so through a mechanism that also damages people, sometimes permanently and sometimes fatally. That combination means the treatment cannot be judged on its own merits, so it has to be judged against the alternative, patient by patient, and the comparison comes down to two estimates that any competent hematology team can produce in a week. How likely does this disease look to come back without a transplant. How likely does this transplant look to kill or disable somebody who would otherwise have lived. Neither estimate lands exactly. Both beat a guess by a long way, and both have been studied in tens of thousands of people, which means the conversation ahead of you is not a leap of faith but a piece of comparison shopping carried out under time pressure.

This page walks through both numbers, what moves them, and what to do when they come out close together.

The question underneath the question

Patients ask whether a transplant will work. The team answers a narrower question, which is whether transplanting produces more survivors than not transplanting in people who look like you.

Number one, what happens without a transplant

Everything here comes from the disease. Its genetics, how it responded to the first course of chemotherapy, whether sensitive testing still finds it when the marrow looks clean, and how many times it has already come back, and for some diseases and some genetics that number is small. For others it approaches certainty.

Number two, what the transplant costs

Everything here comes from you. Age, heart, lungs, liver, kidneys, previous treatment, infections you are carrying, the type of donor available and the intensity of conditioning planned. It gets reported as non relapse mortality, which is the probability of dying from the procedure instead of from the cancer, and in modern series it runs anywhere from under one in ten to better than one in three.

When the arithmetic says no

Most useful in that meta analysis is the group that gained nothing.

For good risk acute myeloid leukemia in first remission, the hazard of relapse or death with a transplant came out at 1.06, and the hazard of death at 1.07. Both confidence intervals sat across one. Transplanting those patients did not help them and carried every risk that a transplant carries, which is why current practice keeps them on chemotherapy and holds transplant in reserve for a relapse that may never happen.

Read that twice, because it cuts against instinct. The intuition that more aggressive treatment must be safer treatment is wrong here in a measurable way. A transplant offered to somebody whose disease was unlikely to return is not a bonus. It subtracts.

Exactly the same logic runs in reverse for high risk disease, where declining a transplant to avoid its early mortality trades a visible risk for a larger invisible one. Both errors happen. The second error happens more, and the first hurts more to watch.

The words in the letters

Transplant correspondence runs on abbreviations, so here are the ones that matter, in the order you will meet them.

Allogeneic means the cells came from another person, which is the whole subject of this page. Autologous means your own cells were collected and given back, and it is a different operation with different risks. Conditioning means the chemotherapy, sometimes with radiation, given before the donor cells arrive. Day zero marks the day they go in, and every other day gets numbered from it. Engraftment names the day the new marrow starts producing blood cells. GVHD stands for graft versus host disease, the donor immune system attacking the person it was given to, and nRM stands for non relapse mortality, which is the second of the two numbers this page is built on. MRD stands for measurable residual disease, meaning leukemia that sensitive testing finds when a microscope cannot. HLA covers the tissue type proteins that decide how well a donor matches. Haploidentical describes a relative who matches on exactly half, which every parent and every child does by definition. And the Disease Risk Index and the comorbidity index are the two scoring systems that turn all of this into something you can compare.

What sequencing did to the first number

Until recently the first number came mostly from a chromosome count. Three buckets, favorable, intermediate and adverse, and a large middle bucket that told nobody very much.

A study of 1,540 patients treated on three prospective trials sequenced 111 cancer genes and found 5,234 driver mutations across 76 genes or genomic regions. Eighty six percent of patients carried two or more drivers. Patterns of co mutation sorted the cohort into eleven classes with distinct outcomes, and the effect of any single mutation frequently changed depending on what sat beside it.

That last clause is the important one, since a mutation that predicts a good outcome on its own can predict a poor one in company, which means a report listing mutations without interpreting their combinations is only half a report.

Eleven classes instead of three

The eleven classes did not replace the old risk groups so much as break up the middle one, which is where most patients were sitting and where the transplant decision was hardest.

Why this matters at the bedside

Three of the classes were new. Leukemia driven by mutations in chromatin or RNA splicing regulators accounted for eighteen percent of patients. Leukemia carrying TP53 mutations, whole chromosome gains and losses, or both, accounted for thirteen percent. Both of those groups did badly, and the class defining mutations added their effects together instead of overlapping, so for a patient sitting in either of them, the chance of the disease returning after chemotherapy alone is high enough that the arithmetic usually favors transplant even when the second number is not flattering. For a patient with mutated NPM1 and the right company around it, the reverse can be true. Chromosome reports show none of it, and a unit that has not sequenced your marrow is estimating the first number with an instrument that went out of date around 2016.

The test that moves the number most

Genetics describes the disease you started with. Measurable residual disease describes the disease you still have.

What it changes

Marrow that looks clean under a microscope can still carry one leukemia cell in ten thousand, and flow cytometry or a molecular test will find it. Patients who clear that test relapse far less than patients who do not, within the same genetic risk group and the same treatment, which means it follows that a positive test raises the first number sharply and can turn a borderline decision into a clear one, while a negative test can do the opposite. The practical instruction is simple. Find out whether the test was done, what it showed, and whether the answer changed anyone plan at all. In a great many centers it is not done at all, and a decision built without it is being made on older information than the patient realizes.

The Disease Risk Index

Somebody eventually had to turn all of this into one comparable number, and that is what the Disease Risk Index does. It takes the diagnosis and the disease status at the time of transplant and sorts patients into four bands. Researchers tested it on 13,131 people reported to the American registry who were transplanted between 2008 and 2010. Two year survival ran from 64 percent in the lowest band down to 24 percent in the highest, and more to the point, the index was the strongest single predictor in the analysis, and it held its power regardless of the patient age, the intensity of conditioning, the source of the graft or the type of donor. A refined version built on a randomly chosen training group of 9,849 patients then predicted better than the original in the remaining 3,282. That detail matters because a score which survives being split, refit and retested on people it has never seen is a score you can quote to a patient without embarrassment.

What the index does not know

No score describes the patient. The Disease Risk Index was built before routine sequencing and before measurable residual disease testing became ordinary, so it treats a first remission as a first remission whether or not the marrow is truly empty.

Two people can land in the same band and face genuinely different odds, one carrying a TP53 mutation and a positive flow test, the other carrying mutated NPM1 and a clean one. A good team uses the band as a starting position and then argues with it out loud, in front of you, using the sequencing report and the residual disease result, while a team that quotes the band and stops has mistaken a research instrument for a decision.

The two scores side by side

Both numbers now have a published scoring system behind them. Seeing them next to each other is the fastest way to understand what the conversation is actually weighing.

Narrow screens scroll this table sideways. Swipe or drag to reach every column.

Score two works differently, because instead of grading the disease it grades the person, by adding points for organ problems that have already been shown to raise the chance of dying from the procedure.

On a phone this table slides left and right. Drag it across to see the second column.

Reading a comorbidity score properly

Scores get misread in two directions, and both directions hurt people.

What to ask when you are given a number

Find out which conditions were counted and how many points each contributed, because the list is public and you are entitled to check it. Ask which of them can be improved before conditioning starts, since iron overload, poorly controlled diabetes, an untreated dental infection and deconditioning all move, then ask what non relapse mortality the unit itself has seen at one hundred days and at one year in patients with your score, and not what the literature says. And ask whether the score changed the plan, because a score calculated and then ignored is paperwork.

Age stopped being the barrier

For most of the history of this treatment, a number on a birth certificate ended the conversation. Sixty served as the ceiling, then sixty five. The ceiling moved because the conditioning changed and not because the patients did.

What happened to the over seventies

1. Between 2000 and 2013, 1,106 adults aged seventy or over had a first allogeneic transplant across 103 American centers.
2. Those transplants went from one in a thousand of all allogeneic transplants in 2000 to 3.85 percent in 2013, which is a thirtyfold shift in who gets offered one.
3. Two year survival rose from 26 percent in the earlier period to 39 percent in the later one, and progression free survival rose from 22 percent to 32 percent.
4. Two year transplant related mortality stayed where it was, between 33 and 35 percent, and did not improve at all across the period.
5. In the modern half of the series, a comorbidity score of three or more, a cord blood graft and full intensity conditioning each independently predicted worse survival.

How to read that honestly

The survival gain came from choosing patients better and from relapsing less, and not from making the procedure safer. One person in three transplanted over the age of seventy still died of the transplant instead of the disease, and that figure sat still for thirteen years, and so the correct summary is that selected people over seventy should be considered, which is what the authors wrote, and that being considered is not the same as being suitable. Anybody over seventy who has been told that age rules them out has been given out of date information. If somebody has told you that age no longer matters, that is wrong in the other direction.

Scores built after those two

Both of the scores above have known weaknesses, and two research groups tried to fix them by combining what each one measured.

Researchers in Japan took 2,344 people transplanted in first remission of acute myeloid leukemia, split them into a training set and a validation set, and built an index out of five items, using age, sex, performance status, the comorbidity score and the donor type. Applied to the validation patients it separated two year non relapse mortality into four levels at 11, 16, 27 and 33 percent. Its discrimination came out at 0.67. The European score managed 0.60 and the comorbidity index used alone managed 0.57.

European investigators did something similar in 812 patients having reduced intensity transplants in first remission, taking the strongest items from both existing scores and adding donor and recipient cytomegalovirus status, and their three risk groups came out at two year non relapse mortality of 8 percent, 17 percent and 38 percent. A spread that wide, inside a group of patients who all looked broadly similar on paper, is the clearest argument there is for scoring people formally instead of judging them across a desk.

Doing the sum

Putting both numbers on the table at once gives the conversation a shape.

1. Write down the chance the disease returns within two years without a transplant, using the genetics, the residual disease test and the remission number. Call it the first figure.
2. Write down the chance of dying from the transplant within two years, using the comorbidity score, the age, the donor type and the planned conditioning. Call it the second figure.
3. Compare them, and remember that a relapse after chemotherapy is not automatically fatal, so the first figure is a risk of the disease returning and not a risk of dying from it.
4. Ask what happens to both figures if you wait. Some diseases get harder to transplant with every month. Some do not move at all.
5. Ask what happens to both figures with a different donor, a different conditioning intensity or eight weeks of getting fitter first, because every one of those is a lever somebody can pull.
6. Then decide, knowing that where the two figures are close the evidence stops giving an answer and your own preferences legitimately take over.

Nobody should be asked to consent to this without having seen those two figures written down.

First remission or second

Arguments in a transplant meeting turn on when far more than on whether.

The case for each

Transplanting in first remission catches the disease at its most controlled, which is when the graft versus leukemia effect has least work to do and when the patient is at their fittest. Waiting keeps the transplant in reserve, avoids its risks entirely for the patients who were going to be cured anyway, and lets a second remission prove that the disease can still be controlled at all. The meta analysis of first remission transplants settles the question for the two ends of the risk spectrum and leaves the middle genuinely open, which is why two experienced hematologists can look at the same intermediate risk patient and reach opposite conclusions without either of them being careless. Where that happens, the residual disease test breaks the tie, and where the test has not been done the argument continues without new information.

The donor clock

Deciding to transplant is not the same as being able to, because a donor has to exist, agree, pass a medical assessment and reach a collection room, and each of those takes time that the disease does not necessarily grant.

Listen for the sentence that says the search is under way and the plan will be finalized when it finishes. That sentence should reassure you. Silence over which week the answer will arrive should not.

Getting fit for it

Several of the things that raise the second number come down, and the window to lower them is the weeks between the decision and the conditioning. Iron overload from repeated transfusions damages the liver and the heart and responds to treatment. Dental infection causes trouble afterward and takes an afternoon to find. Blood sugar that has drifted upward through months of steroids comes back down once somebody looks at it. Deconditioning from six weeks in a hospital bed responds to walking, and walking is the cheapest thing on this page, and smoking has to stop, and stopping two weeks before conditioning beats stopping the week after discharge. Nothing here turns a score of five into a score of one. What it does is stop the score drifting upward while everybody waits for a donor, and in a patient who is borderline it occasionally moves the decision. None of it is glamorous work and all of it is cheap, which is a combination that tends to get a treatment overlooked in a field full of expensive interventions. Somebody on the team should own the list and tick it off, because a job that belongs to everybody belongs to nobody.

What no score contains

Every number on this page describes populations. You remain one person, and several of the things that will decide how this goes for you have never been scored by anybody.

The unscored variables

Whether somebody can live with you for four months. Whether your household survives one wage disappearing. Whether you have a second language in common with the nurses who will be explaining things to you at three in the morning. Whether the nearest hospital to your home can manage a febrile transplant patient at two on a Sunday, or whether you are the kind of person who finds a long dangerous treatment easier to bear than a long uncertain wait, or the opposite kind. None of these counts as a soft consideration added at the end for completeness. They change outcomes, they decide whether somebody makes it to a hundred clinic appointments, and they belong in the conversation at the same moment as the hazard ratios and not afterward.

The published figures

Below sit the figures quoted on this page, with the population each one came from, so that anything a team tells you has something to be measured against.

Pull this table sideways on a small screen to reach all three of its columns.

Every figure there comes from a population, and none of them predicts anything for one person.

What a second opinion should contain

Second opinions in this field have a reputation for being polite restatements of the first one. A useful one looks different. Start with the disease risk band and which sequencing result put it there. Whether a residual disease test was done, what it showed and what would change if it were repeated all belong in it, along with a non relapse mortality estimate given as a range and the comorbidity score that produced it. Donors come next, meaning which ones are realistically available and how long each would take. Then a plain statement of whether the reviewer agrees with the plan already in front of you, and where they disagree, what they would do instead and why. All of it written down, because a transplant decision gets revisited three or four times by people who were not in the room the first time, and a conversation nobody recorded is a conversation that has to be had again from the beginning.

Ours comes in writing and costs nothing. Where it agrees with the plan you already have, it says so and recommends staying where you are.

When the answer is no

Sometimes the arithmetic comes out against transplanting, and the conversation that follows is handled badly more than any other conversation in this field.

A no can mean several different things and they are not interchangeable. It can mean the disease is favorable enough that transplant would subtract rather than add, which is good news delivered in a confusing register. It can mean the second number is too high today and might not be in eight weeks, which is a postponement rather than a refusal, or it can mean the disease is not controlled enough for a transplant to hold, in which case the next move is treatment aimed at getting it controlled and then asking again. Or it can mean that nothing available is likely to work, which is the hardest version and the one that deserves the most time and the plainest language. Anybody told no is entitled to know which of those four they have been told, and a family that leaves without knowing will spend months believing a door was closed that was only ever propped open with a date on it.

Between the decision and day zero

Once yes has been said, a fixed sequence runs, and knowing its shape stops the waiting feeling like inactivity.

1. Donor confirmation and workup, running in parallel with whatever treatment is holding the disease.
2. A full organ assessment. Heart scan, lung function, kidney and liver bloods, viral serology, dental review and a fresh marrow sample.
3. The fertility conversation, which has to happen before conditioning and never after it, and which is the step most frequently skipped when things are moving fast.
4. A central line inserted, usually a few days before conditioning begins.
5. Conditioning itself, running somewhere between a week and two days before day zero depending on the regimen.
6. Then the cells, which arrive as a bag on a drip stand and go in over an hour or two with nothing dramatic about the day at all.

What to sort out while that runs

Accommodation for a companion, income and sick pay, who is looking after children or parents, and who holds power of attorney if you cannot make decisions for a period, and families that arrange those things during the waiting weeks find the hundred days afterward far easier than families that arrange them in a corridor.

What the family decides

One person consents to a transplant. Several people live through it.

The commitments nobody lists

Companions have to be present for months, which means somebody stopping work, and if a sibling is the donor, that person carries a role nobody asked them to audition for, and the honest handling of it is to have them assessed by a separate team who can say no on their behalf without the patient ever knowing why. Children need an explanation pitched at their age and repeated, because a parent disappearing into a hospital for a month lands hard and lands differently at four, at nine and at fifteen. Somebody has to hold the paperwork, the medication list and the appointment calendar, and that person should be named out loud instead of emerging by default as whoever is least able to refuse. Deciding all of this before the admission is not pessimism. It marks the difference between a family that copes and a family that improvises for four months.

Deciding this from another country

Distance changes the arithmetic in ways that the published scores never measured.

What travel adds to the second number

Time is the first thing, since a search that would take four weeks at home takes longer when reports have to be translated, visas obtained and flights booked, and in a disease that is being held and not cured, delay is not neutral. Continuity follows. A patient transplanted in one country and followed in another is relying on a handover that has to be written properly or it does not happen at all. Money comes third, and it belongs in the open and not at the end. What drives the cost of an allogeneic transplant is the donor type and the search, the conditioning regimen chosen, how long the inpatient stay actually runs, whether complications need intensive care, how much of the outpatient period is spent here and how long a companion stays. Against all of that sits the reason people travel, which is access to a unit that performs the operation often enough to be good at it, and in particular to one that can perform a half matched transplant when the registries return nothing.

What we will not tell you

Pages like this one are usually where a hospital promises things. Here follows what we will not promise.

We will not give you a survival figure before we have read your marrow report, your sequencing, your residual disease result and your organ function, because a number produced without those is marketing. We will not tell you that transplant is your only option before we have checked whether your disease risk band says otherwise. We will not quote our own results as a single percentage without telling you which patients they came from. And we will not tell you the procedure is safe, because the published non relapse mortality across every series on this page runs between roughly eight and thirty five percent depending on who is being transplanted.

If another unit has given you a clean figure and no range, ask which patients produced it.

Having it done here

Istanbul has a practical advantage in this particular operation, and stating it precisely helps more than stating it vaguely.

Why the donor question runs differently here

The international registries are dominated by volunteers of northern European ancestry, so patients from Turkey, the Middle East, North Africa, South Asia and mixed families are matched to an unrelated stranger far less often than the size of those registries suggests. Half matched transplantation from a parent, a child or a sibling removes that problem, because a relative who matches on exactly half is available to almost everybody, and units that perform those routinely therefore offer something to these populations that units which do not perform them cannot, whatever else is true about either place. That makes the honest case for having this done in a center like ours, and the claim runs narrower than the ones usually made.

What we arrange

A transplant works as a relocation and not a trip, and the practical scaffolding around it decides whether a family gets to the end of it.

Plan on twenty to thirty five nights as an inpatient, then roughly ninety more nights within reach of the hospital while clinic visits run two or three times a week. Fitness to fly home is judged on blood counts, on whether graft versus host disease is active and on how much immunosuppression is still running, which usually puts it past the third month rather than on a fixed date.

The practical side, in order

Questions we are asked, an allogeneic transplant FAQ

References

1. Koreth J, Schlenk R, Kopecky KJ, et al. Allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission, systematic review and meta analysis of prospective clinical trials. JAMA. 2009;301(22):2349-2361.
2. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med. 2016;374(23):2209-2221.
3. Armand P, Kim HT, Logan BR, et al. Validation and refinement of the Disease Risk Index for allogeneic stem cell transplantation. Blood. 2014;123(23):3664-3671.
4. Sorror ML, Logan BR, Zhu X, et al. Prospective validation of the predictive power of the hematopoietic cell transplantation comorbidity index, a Center for International Blood and Marrow Transplant Research study. Biol Blood Marrow Transplant. 2015;21(8):1479-1487.
5. Yanada M, Konuma T, Mizuno S, et al. Predicting non relapse mortality following allogeneic hematopoietic cell transplantation during first remission of acute myeloid leukemia. Bone Marrow Transplant. 2021;56(2):387-394.
6. Versluis J, Labopin M, Niederwieser D, et al. Prediction of non relapse mortality in recipients of reduced intensity conditioning allogeneic stem cell transplantation with AML in first complete remission. Leukemia. 2015;29(1):51-57.
7. Muffly L, Pasquini MC, Martens M, et al. Increasing use of allogeneic hematopoietic cell transplantation in patients aged 70 years and older in the United States. Blood. 2017;130(9):1156-1164.