First-Remission AML: Conditioning Intensity in a Marginally Fit Patient
A 68-year-old with AML in first remission is eligible for allogeneic transplant, but the trial that quantified how much conditioning intensity buys him in relapse reduction never enrolled anyone his age.
Harold T., a 68-year-old retired postal carrier, has spent most of his retirement in the garage refinishing furniture he picks up at estate sales — a hobby his wife jokes has taken over the whole first floor. He was diagnosed with NPM1-mutated, FLT3-wild-type AML five months ago after presenting with fatigue and easy bruising, and achieved a first complete remission after standard 7+3 induction followed by one cycle of consolidation. His counts have recovered, his marrow is clean, and a fully matched sibling donor — his younger brother — is available and willing. The transplant team agrees he should proceed to allogeneic transplant in CR1; what they don't agree on is how hard to condition him first.
Harold's fitness for that conversation is genuinely borderline, not clearly good or clearly poor. His HCT-CI comorbidity score comes to 3 — treated hypertension, mild COPD (DLCO 68% of predicted, no home oxygen), and a remote deep vein thrombosis on long-term apixaban — a score associated with meaningfully higher non-relapse mortality after intensive conditioning, though not high enough by itself to rule transplant out. The trial most often cited for choosing conditioning intensity in exactly this situation, BMT CTN 0901 (Scott), found myeloablative conditioning nearly halved 18-month relapse compared with reduced intensity (13.5% versus 48.3%) at the cost of more toxicity and no significant difference in overall survival — but the trial enrolled adults aged 18 to 65. Harold is 68. Applying its relapse-reduction case to him isn't reading the trial's result; it's extrapolating past the population that produced it, at exactly the age where the toxicity side of that same trade would be expected to weigh more, not less. And the relapse number it would buy him is not the number the trial measured: by 2022 European LeukemiaNet criteria his NPM1-mutated, FLT3-wild-type disease sits in the lowest-relapse-probability molecular category of any AML subtype, so the 48.3% arm of that comparison describes a risk he does not carry in full. A frailty assessment ordered because his HCT-CI sits in the range where it adds information beyond the comorbidity index put his six-minute walk at 380 meters — normal for his age, and the one number in his workup that argues for treating him as fitter than his score suggests, though it says nothing about the lungs that would absorb the alkylator.
Deciding conditioning intensity before a matched sibling transplant
Myeloablative conditioning is what actually moves relapse risk in AML, and relapse is what kills most patients who make it through transplant. BMT CTN 0901 found an 18-month relapse rate of 13.5% with MAC versus 48.3% with reduced intensity — that's not a marginal difference, it's more than triple. Harold's NPM1-mutated disease is favorable-risk by molecular testing, but favorable-risk in CR1 still isn't zero-risk, and giving up most of that relapse protection because of a comorbidity score of 3 trades a known, quantified benefit for a toxicity risk we haven't actually measured in him yet.
If he tolerates induction and consolidation the way his counts suggest he has, I don't think we should assume in advance that his lungs and kidneys can't handle a standard regimen.
You're right that the relapse numbers are real and large — I'm not disputing the trial's own result. What I'm disputing is reading it as Harold's result. BMT CTN 0901 enrolled patients 18 to 65. He's three years past that ceiling, with an HCT-CI of 3 that the trial's own eligibility criteria would likely have screened out at several sites. The same trial that found MAC's relapse benefit also found higher treatment-related toxicity in the MAC arm — and toxicity from full-dose alkylator conditioning concentrates in exactly the organs his comorbidities already compromise: his COPD-affected lungs, his borderline renal clearance.
This isn't a case where his age alone is disqualifying — his performance status is good and his cardiac function is normal. It's that the specific trial being cited to justify the more intensive regimen simply never tested anyone who looks like him.
Both of those readings are honest about the same trial's limits for him — which is exactly why I don't think the choice should be MAC-as-tested versus RIC-as-tested at all. Fludarabine plus treosulfan is a conditioning regimen developed specifically for older and comorbid AML/MDS patients considered unsuitable for standard myeloablative busulfan. Beelen's MC-FludT.14/L trial randomized it against reduced-intensity busulfan-fludarabine and found improved event-free survival, driven by lower non-relapse mortality rather than by better disease control. And look at who it let in: age 50 or over, or an HCT-CI above 2. Harold clears both. He is one of the few patients in this room who is inside a trial's entry criteria rather than three years outside them.
It doesn't promise BMT CTN 0901's full relapse-reduction number — that trial's own patients were healthier by definition — but it's evidence generated in patients who actually look like Harold, rather than a number borrowed from patients who don't.
Agreed: fludarabine-treosulfan conditioning, chosen specifically because it was developed for patients whose fitness profile falls between the two arms BMT CTN 0901 actually tested. Harold's brother proceeds as the matched sibling donor. Pulmonary and renal function will be reassessed immediately before conditioning begins.
Not agreed, and left explicitly on the record rather than smoothed over: whether this regimen genuinely closes the relapse-risk gap MAC would have addressed, or only partially does. The transplant physician would have accepted higher measured toxicity risk in exchange for BMT CTN 0901's larger number if Harold's actual pulmonary testing had come back better than it did; the geriatrician views the treosulfan data as sufficient on its own terms, not as a compromise. Both agreed the honest answer was to choose the regimen built for a patient like him rather than debate which mismatched trial to borrow from.